Citation
Controlling the indirect exchange coupling in anisotropic magnetoresistive sensors with ruthenium

Material Information

Title:
Controlling the indirect exchange coupling in anisotropic magnetoresistive sensors with ruthenium
Creator:
Osminer, Teresa Lee Wagar
Publication Date:
Language:
English
Physical Description:
xv, 78 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Ruthenium ( lcsh )
Magnetoresistance ( lcsh )
Detectors ( lcsh )
Ferromagnetism ( lcsh )
Detectors ( fast )
Ferromagnetism ( fast )
Magnetoresistance ( fast )
Ruthenium ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaf 78).
General Note:
Department of Electrical Engineering
Statement of Responsibility:
by Teresa Lee Wagar Osminer.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
526477008 ( OCLC )
ocn526477008
Classification:
LD1193.E54 2009m O75 ( lcc )

Full Text
/
CONTROLLING THE INDIRECT EXCHANGE COUPLING IN
ANISOTROPIC MAGNETORESISTIVE SENSORS
WITH RUTHENIUM
by
Teresa Lee Wagar Osminer
B. S., University of Colorado Denver, 2000
A thesis submitted to the
University of Colorado Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Electrical Engineering
2009


This thesis for the Master of Science
degree by
Teresa Lee Wagar Osminer
has been approved
by
Dr. Hamid Z. Fardi
/
Dr. Miloje Radenkovic
/?/ <0
--1--
*w
Date


Osminer, Teresa Lee Wagar (M.S., Electrical Engineering)
CONTROLLING THE INDIRECT EXCHANGE COUPLING IN ANISOTROPIC
MAGNETORESISTIVE SENSORS WITH RUTHENIUM
Thesis directed by Professor Hamid Z. Fardi
ABSTRACT
The Advanced Magnetic and Quantum Materials program at the National Institute of
Standards and Technologies (NIST) specializes in designing, fabricating, and testing
devices such as magnetic sensors. One specific topic of investigation is the use of
ruthenium (Ru) as a spacer material in a trilayer structure composed of a
ferromagnetic-Ru-antiferromagnetic layering strategy to control the exchange
coupling between the ferromagnetic (FM) and antiferromagnetic (AFM) layers. The
purpose of the research is to examine whether such a trilayer design can be
engineered to tune the sensitivity and dynamic range of the magnetic sensor. The
devices used for the project were composed of AFM layers of iridium manganese
(IrMg), nonmagnetic spacers of Ru and were topped with FM layers of Permaloy
(nickel iron (NiFe)). The primary differentiator between the samples was the
thickness of the spacer layers. Five samples, with Ru thicknesses of 0.25nm, 0.5nm.
0.75nm, lnm, and 2nm were analyzed and compared. All devices under test were
fabricated prior to 2008 by members of the Advanced Magnetic and Quantum
Materials for the purpose of examining the newly hypothesized effects of Ru on
indirect exchange coupling. This report focuses on the measurement method for
establishing the experimental results rather than on the material science of the AMR
sensor design and fabrication. The results of this testing revealed that this method of
sensor tuning improves the sensitivity of AMRs without improving the dynamic
range. They illustrate that the measurement methods are important and useful for


engineering these devices. Further testing on a greater sample set would be required
to draw any definitive conclusions.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
Signed:
Hamid Z. Fardi


DEDICATION
I dedicate this thesis to my husband, Matt, who is my keystone, inspiration and
motivation. His selfless support and encouragement has made this journey as
rewarding as the pursuit of a graduate degree could be.


ACKNOWLEDGEMENTS
I would like to thank Dr. Hamid Fardi, who has been teacher, advisor and colleague
throughout the pursuit of my graduate degree. I would also like to thank my other
colleague, Fabio da Silva and associate, Sean Haloran, and especially my supervisor,
David Pappas, at the National Institute of Standards and Technology (NIST), for the
outstanding research opportunity entrusted to me, their knowledgeable guidance and
generous tutelage. In addition, I am grateful to the members of my review committee
for their valuable time and advice in evaluating my thesis.


TABLE OF CONTENTS
Chapter
1. Introduction...........................................................16
2. Defining Parameters....................................................18
2.1. Electromagnetic Units of Measurement...................................18
2.2. Noise Theory...........................................................19
2.2.1. Causes of Intrinsic Noise........................................19
2.2.2. External Noise Interface.........................................23
2.3. Sensitivity............................................................24
2.4. Field Noise............................................................24
3. Magnetic Sensors.......................................................26
3.1. Ferromagnetic and Antiferromagnetic Layering...........................28
3.2. Magnetoresistive Effect................................................30
3.3. Anisotropic Magnetoresistive Sensors...................................31
3.4. Barber pole Differential Bridge Biasing Method.........................33
4. Measurement Strategy...................................................36
4.1. Magnetoresistive Sensor Noise Metrology................................36
4.2. Electronic Design Considerations.......................................38
vii


TABLE OF CONTENTS (Cont.)
Chapter
4.3. Design Validation.................................................40
4.4. Design Limitations................................................41
4.4.1. Sensitivity Measurement Process..............................41
4.5. Noise Measurement Process.........................................42
4.6. From Theory to Practice...........................................42
5. Data Processing....................................................43
5.1. Analysis.........................................................43
5.2. Results...........................................................43
5.3. Conclusions.......................................................45
6. Discussion.........................................................46
APPENDIX A................................................................48
APPENDIX B................................................................53
viii


LIST OF FIGURES
Figure
2.1 Johnson Noise (flat Power Spectral Density).............................20
2.2 Flicker Noise (1/f noise).............................................21
2.3 Domain Walls and Magnetic Pockets in Inhomogeneous Magnetic
Materials............................................................22
2.4 Brownian Noise (1/f2 noise)..............................................22
3.1 Sensitivity Ranges for Common MR Sensors.................................26
3.2 Required Level of Sensitivity for Many Applications of
Magnetoresistive Sensors.............................................27
3.3 Relationship Between Sensor Noise and Size...............................28
3.4 a)Ferromagnetic and b)Antiferromagnetic moments..........................29
3.5 Magnetization Curves.....................................................29
3.6 Fundamental Magnetoresistive Effect......................................30
3.7 a) Longitudinal Resistivity (pO and b) Transverse Resistivity (p,)......32
3.8 Anisotropy axis for ferromagnetic an antiferromagnetic materials........32
3.9 Change in magnetoresistive resistance versus magnetization angle.........33
3.10 Shift in MR transfer function...........................................33
3.11 Barber Poles............................................................34
3.12 Trilayer Structure......................................................35
IX


LIST OF FIGURES (Cont.)
Figure
3.13 Sensitivity Versus Ruthenium Thickness..................................35
4.1 Sensitivity measurement test set.......................................36
4.2 Noise measurement test set.............................................37
4.3 Inrush Current Limit Circuit...........................................38
4.4 Modeled transient suppression of inrush current limit protection circuit.39
4.5 DAC Bridge Voltage Command.............................................39
4.6 Printed circuit board design noise mitigation measures.................40
5.1 Measured Sensitivities at 60A for all Five AMR Sensors.................44
5.2 Field noise measured at 60pA for AMR sensors...........................44
5.3 Johnson noise (white) and 1/f noise (pink) characteristics evident in
experimental data..................................................45
A. 1.1: 0.25nm Ru Sample Field Noise.......................................49
A. 1.2: 0.25nm Ru Sample Normalized Measured Noise.......................49
A. 1.3: 0.50nm Ru Sample Field Noise.......................................50
A. 1.4: 0.50nm Ru Sample Normalized Measured Noise.......................50
A. 1.5: 0.75nm Ru Sample Field Noise.......................................51
A. 1.6: 0.75nm Ru Sample Normalized Measured Noise........................51


LIST OF FIGURES (Cont.)
Figure
A. 1.7: l.Onm Ru Sample Field Noise.....................................52
A. 1.8: 1 .Onm Ru Sample Normalized Measured Noise.......................52
A. 1.9: 2.Onm Ru Sample Field Noise.....................................53
A. 1.10: 2.Onm Ru Sample Normalized Measured Noise.......................53
B. 1.1: Measured Sensitivity for 0.25nm Ru Sample when Bias Current is
0.059mA = 1.302 pV/G..............................................55
B.1.2: 0.25nm Ru Sample Noise with 0.059 Bias Current.....................55
B.1.3: Measured Sensitivity for 0.25nm Ru Sample when Bias Current is
0.030mA = 646.34 nV/G.............................................56
B.1.4: 0.25nm Ru Sample Noise with 0.030 Bias Current.....................56
B.1.5: Measured Sensitivity for 0.25nm Ru Sample when Bias Current is
0.015mA = 290.17 nV/G.............................................57
B.1.6: 0.25nm Ru Sample Noise with 0.015 Bias Current.....................57
B. 1.7: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is
0.121mA = 11.198 pV/G.............................................58
B.1.8: 0.50nm Ru Sample Noise with 0.121 Bias Current.....................58
B. 1.9: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is
0.060mA = 5.498 pV/G..............................................59
B.1.10: 0.50nm Ru Sample Noise with 0.060 Bias Current....................59
XI


LIST OF FIGURES (Cont.)
Figure
B. 1.11: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is
0.030mA = 2.700 pV/G..............................................60
B. 1.12: 0.50nm Ru Sample Noise with 0.030 Bias Current...................60
B. 1.13: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is
0.015mA = 2.700 pV/G...............................................61
B. 1.14: 0.50nm Ru Sample Noise with 0.015 Bias Current...................61
B. 1.15: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.914mA = 17.464 mV/G..............................................62
B.1.16: 0.75nm Ru Sample Noise with 0.914 Bias Current....................62
B.l .17: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.462mA = 8.907 mV/G...............................................63
B. 1.18: 0.75nm Ru Sample Noise with 0.462 Bias Current...................63
B. 1.19: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.230mA = 15.654 pV/G..............................................64
B. 1.20: 0.75nm Ru Sample Noise with 0.230 Bias Current...................64
B. 1.21: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.107mA = 9.502 pV/G...............................................65
B. 1.22: 0.75nm Ru Sample Noise with 0.107 Bias Current...................65
B. 1.23: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.092mA = 1.805 mV/G...............................................66
B. 1.24: 0.75nm Ru Sample Noise with 0.092 Bias Current...................66
xii


LIST OF FIGURES (Cont.)
Figure
B. 1.25: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.046mA = 4.000 pV/G...............................................67
B. 1.26: 0.75nm Ru Sample Noise with 0.046 Bias Current....................67
B.1.27: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.030mA = 594.39 pV/G..............................................68
B.1.28: 0.75nm Ru Sample Noise with 0.030 Bias Current....................68
B.1.29: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is
0.015mA = 1.228 pV/G...............................................69
B.1.30: 0.75nm Ru Sample Noise with 0.015 Bias Current....................69
B. 1.31: Measured Sensitivity for 1 .Onm Ru Sample when Bias Current is
0.250mA = 52.905 pV/G..............................................70
B. 1.32: 1 .Onm Ru Sample Noise with 0.250 Bias Current....................70
B.1.33: Measured Sensitivity for l.Onm Ru Sample when Bias Current is
0.170mA = 25.113 pV/G..............................................71
B. 1.34: 1 .Onm Ru Sample Noise with 0.170 Bias Current....................71
B.1.35: Measured Sensitivity for l.Onm Ru Sample when Bias Current is
0.076mA = 11.273 pV/G..............................................72
B. 1.36: 1 .Onm Ru Sample Noise with 0.076 Bias Current....................72
B. 1.37: Measured Sensitivity for 1 .Onm Ru Sample when Bias Current is
0.030mA = 4.378 pV/G...............................................73
B. 1.38: 1 .Onm Ru Sample Noise with 0.030 Bias Current....................73
xiii


LIST OF FIGURES (Cont.)
Figure
B.1.39: Measured Sensitivity for l.Onm Ru Sample when Bias Current is
0.015mA = 2.063 pV/G................................................74
B.1.40: l.Onm Ru Sample Noise with 0.015 Bias Current......................74
B. 1.41: Measured Sensitivity for 2.0nm Ru Sample when Bias Current is
0.060mA = 1.972 pV/G................................................75
B.1.42: 2.0nm Ru Sample Noise with 0.060 Bias Current......................75
B. 1.43: Measured Sensitivity for 2.0nm Ru Sample when Bias Current is
0.030mA = 754 nV/G..................................................76
B.1.44: 2.0nm Ru Sample Noise with 0.030 Bias Current......................76
B. 1.45: Measured Sensitivity for 2.0nm Ru Sample when Bias Current is
0.015mA = 429 nV/G..................................................77
B.1.46: 2.0nm Ru Sample Noise with 0.015 Bias Current......................77
XIV


LIST OF TABLES
Table
2.1 Basic units and conversion factors for magnetic measurements............18
2.2 Four Types of electromagnetic interference..............................23
3.1 Comparison of MR sensors sensitivity to size............................27
XV


1.
Introduction
Sensors which detect magnetic fields have become an integrated part of modem
life. The ingenuity of scientists and engineers has made these devices the heart of an
ever expanding catalog of products used daily by average people, as well as
professionals in specialized fields. Currently, magnetic sensors of various qualities
and sensitivities are used in such industries as automotive, information/knowledge
management, geophysical and space exploration, health care, military and
commercial applications, law enforcement and archeology, among others. With the
increasing number of uses for this technology, the market for these sensors has been
forecasted to exceed $1.5 billion world wide by 2010 [1], Such lucrative and
pervasive electronic components certainly warrant development and refinement. As
with almost all previous improvements in technological products, the demand for
smaller, lighter, less expensive but more effective products drives the direction of
research.
The focus of this research was on a variety of magnetic field transducer known as
an Anisotropic Magnetoresistive (AMR) sensor. The sensors used in these
experiments were fabricated in 2007 by members of the Advanced Magnetic and
Quantum Materials at the National Institute of Standards and Technology (NIST) for
the purpose of examining the effects of an internal Ruthenium (Ru) layer on indirect
exchange coupling in a tri-layer structure. The five devices used in this project were
composed of an Antiferromagnetic (AFM) layer of iridium manganese (IrMg), a
nonmagnetic spacer of Ruthenium (Ru) and a ferromagnetic (FM) layer of Permaloy
(nickel iron (NiFe)). Recent theories suggest the product line of AMRs may benefit
from the unidirectional stabilization of free layers (Halloran et al., 2007), (also
called uniaxial anisotropy, which causes the magnetic moments of the magnetic
material to align along one axis). It is believed this will improve the overall
performance by increasing the sensitivity of the devices without effecting the
dynamic range or, conversely, increasing the dynamic range without losing
sensitivity. For this project, unidirectional stabilization is achieved through the
indirect exchange coupling between the FM and AFM layers sandwiching a layer of
Ru, which inhibits the direct exchange coupling strength of FM and AFM layers in
direct, molecular contact with each other. Samples were designed with different
thicknesses of Ru layers in order to compare the performance of each sensor based on
that parameter.
Because the samples under test were previously fabricated, this report
concentrates on the development of the measurement method for establishing the
experimental results. It does not delve heavily into the material science of
16


magnetoresistive (MR) sensor design and fabrication, although a minimal explanation
of the theory of AMR design is provided for background.
To begin, a brief explanation of the electronic phenomenon of noise and its
limiting affect on the performance of magnetic sensors is presented. This topic is
covered in chapter two. Chapter three provides an overview of the different types of
magnetic sensors commonly used in modem products. The approach to testing the
AMRs, including the test equipment and electronic circuit board design, is discussed
in chapter four, followed by a review of the analysis, results and conclusions in
chapter five. Chapter six presents a discussion of the conclusions drawn from the
experimental results along with some closing thoughts.
17


2. Defining Parameters
As with all families of electronic components, MR sensors are categorized by
specific characteristics which are the principal constraints considered when
differentiating between the varieties of these devices. In addition to operating and
quiescent currents and voltages, input and output noise as well as sensitivity and
dynamic range are the primary defining parameters of interest when incorporating a
magnetoresistive sensor into any electronic design.
2.1. Electromagnetic Units of Measurement
The language of electromagnetic science is suffused with certain specialized units
which should be defined in order to fully understand the topic of this project. The
International System (SI units) and Centimeter-Gram-Second (cgs) units are used
interchangeably throughout this paper. The basic units and conversion factors most
associated with this subject matter are summarized in Table 2.1 [2],
Table 2.1 Basic units and conversion "actors for magnetic measurements
Quantity cgs units SI units cgs to SI Conversion
Magnetic flux density (B) Gauss (G) Tesla (T) 1 gauss=0.0001 T,
Field intensity (H) Oersted (Oe) A/m 1 oersted = 103 / 4ti A/m
Magnetization (M) emu/cm3 A/m 1 emu/cm3 = 10'3 A/m
(4jtM) Gauss (G) A/m 1 gauss = 103 / 4k A/m
Sensitivity - V/T -
Noise - v/VHz -
Field Noise - t/Vhz -
A good rule of thumb that most magnetic field scientists remember is: lOe = 1 G (in
vacuum) = 1 OOuT = 0.1 mT = 0.796A/cm = 79.6A/m [3]
18


2.2.
Noise Theory
The term noise in this document refers to an interfering signal imposed on
electronic devices and circuitry, which influences the quality of the circuitrys output
signal. Noises are due to uncertainties in physical phenomena and need to be treated
statistically [4]. The most common three sources of noise arise from thermal
fluctuations, material impurities and random fluctuations of current. These are
inherent in the electronic components and are very difficult to mitigate. Other
undesirable noise sources originate outside of the circuit. Such sources are
responsible for radiated emissions, known in the reliability and standards field as
electromagnetic interference, and can usually be minimized through careful design
practices. Integrated circuits (ICs) are affected by all of these to varying degrees
depending on their construction and function. It is important when studying highly
sensitive devices to understand both sources of noise, and their possible effect on the
experiment, in order to accurately analyze experimental results.
One standard of referring to the different modes of noise generation is by the
distinctive frequency responses of the Power Spectral Densities (PSD) which is
characteristic of each mode. A PSD is a statistical function of frequency commonly
used in signal processing to quantify components of power with respect to a spectrum
of frequencies. Without extensively exploring the derivation of power spectral
density as it is implemented by electronic instrumentation, recognize that the units of
PSDs are in watts/Hz. Though the vernacular is to refer to the power spectral density
of a noise source, power is not a quantity that is directly measurable in electronic
circuitry. Voltages and currents are the elements of power that can be directly
measured from an electronic component, so PSDs are simplified into units of voltage
rather than watts. Power is proportional to the square of the voltage applied to a
circuit; therefore the voltage or, more specifically, the root-mean-square of the
voltage (Vrms), which is the value obtained from ordinary electronic measurement
instruments, is commonly used as shorthand for PSDs. The units of noise voltage
spectral density are V/VHz .
2.2.1. Causes of Intrinsic Noise
The sources of noise which are inherent in any electronic component with a
resistive element show distinctive spectral density characteristics. These
characteristics can be correlated to certain color spectra. Figures 1 thru 3 are
examples of the colors of noise:
19


2.2.1.1.
Johnson Noise
100 1000 10000
frequency (Hz)
Figure 2.1 Johnson Noise (flat Power Spectral Density)1
Johnson noise is a phenomenon common to all conductive materials and is caused
by fluctuations of the electric current due to the random thermal motion of the charge
carriers. These fluctuations occur even if there is no external power source and are
not to be confused with self-heating produced by current flowing through a lossy, or
non-ideal, material.
Johnson noise exhibits a PSD typical of white noise. The spectral density remains
constant across frequency bands below 80 GHz; frequencies up to and including the
RF band. It does not increase or decrease with frequency within this range; a trait
shared with white light. Consequently, Johnson noise bounds the lower limit of the
noise voltage in electronic devices [5]. The only method for eliminating it is through
cooling the noise source to the point that molecular motion ceases. For commercial
applications, this is not practical and has other electro-mechanical ramifications that
nullify the white noise reduction benefits. The Johnson noise voltage (rms) in a
particular frequency range can be calculated with the equation
1 Image retrieved from http: "en. wikipedia. ora/wiki/Filc: White noise spectrum .png
20


V4kBTRAf [V],
2-1
V,
noisc(rms)

where kB is the Boltzmann Constant, T is the temperature in Kelvin, R is the
resistance value of the noise source and A / is the bandwidth. Because the Power
Spectral Density does not increase or decrease with frequency, the frequency
component can be ignored and the equation for Johnson noise is simplified to [5].
V(=V4kTR [v/n/Hz
2-2
2.2.1.2. Flicker Noise
100 1000 10000
Frequency (Hz)
Figure 2.2 Flicker Noise (1/f noise).2
Flicker noise can be attributed to fabrication anomalies and impurities in the
materials that a conductive device is constructed from [4]. These impurities and
fabrication flaws make the magnetization of the material inhomogeneous, causing
magnetic as well as physical deformities. As a result, pockets of higher or lower
magnetization may be found in the material. As with any un-uniform surface or
volume, these pockets of varying magnetic denseness cause the transfer of charge
carriers to flow in spurts and eddies. Figure 2.3 illustrates the effect of
inhomogeneity on a magnetic material as it applies to the formation of magnetic
pockets, or domains.
2 Image retrieved from http: en.vvikipedia.ora vviki/File:Pink noise spectrum.pna
21


Figure 2.3 Domain Walls and Magnetic Pockets in Inhomogeneous Magnetic Materials
The only method of minimizing the generation of this type of interference is through
rigorous control of the manufacturing process. Apart from strict fabrication
processes, placing a high-pass filter after the noise source (i.e. the sensor) is the
principal method of reducing flicker noise.
Flicker noise, which shares a PSD typical of pink noise, has a frequency spectrum
that falls off at a constant rate as frequency increases. It is also known as Barkhausen
noise [6] and, even more commonly as 1/f (spoken as one-over-f) noise because the
PSD is proportional to the inverse of the frequency. While pink noise arises from a
variety of causes, flicker noise results from effects related to direct current (DC)
flowing through the noise source. 1/f noise in magnetic materials that are relatively
coarse grained is higher than that in materials that are made up of smaller,
homogeneous molecules [7], The amount of power per decade of frequency is
constant [5],
2.2.1.3. Other Recongnizable Noise Sources
Brown Noise
Frequency (Hz)
Figure 2.4 Brownian Noise (1/f' noise)3
3 Image retrieved from http:, en.vvikipedia.org/w iki/File: Brown noise spectrum.pnu
22


Though Johnson noise and 1/f noise dominate the frequency response of most
electronics, there are lesser offenders to be aware of. Brownian noise, or brown
noise, is produced by Brownian motion and exhibits a PSD predominantly within the
lower frequencies, or red spectrum. This type of signal interference has a PSD
-)
proportional to 1/f", therefore attenuates at 6dB/octave; twice as quickly as flicker
noise [5]. Unless the effects of flicker noise are successfully mitigated, therefore,
brown noise will not be a significant factor in electronic devices. Shot noise is caused
by random fluctuations of current which tends to flow in an electrical conductor in
clumps rather than in a smooth, ideal stream. Similarly, burst noise occurs when
random, sudden, step-like transitions between two or more energy levels occurs due
to trapped charge carriers transitioning at unpredictable times rather than
instantaneously, as in the ideal model.
2.2.2. External Noise Interface
The intrinsic sources of noise discussed in the previous section are all limiting
factors which must be understood and recognized, but are extremely difficult to
mitigate. They set unavoidable restrictions on the effectiveness of all physical
components. The influence of noise which impinges on the electronics from external
sources must not be disregarded, though. A sub-specialty of electrical engineering is
growing rapidly which specifically addresses electromagnetic interference. This field
has gained importance as physically miniaturized, micro-powered electronics have
become mass producible. Table 2.2 below is a matrix of the four types of EMI and
their definitions [8]:
Table 2.2 Four Types of electromagnetic interference
Emission Susceptability
Conducted The transmission of noise through paths such as wires, traces or other conductive media. The reception of noise from other sources through paths such as wires, traces or other conductive media.
Radiated The transmission of noise through the air or free space. The reception of noise from other sources through the air or free space.
23


2.3.
Sensitivity
One of the principal parameters used to characterize detectors and sensors is
sensitivity. For the EM field transducers used in this project, sensitivity refers to the
amount of voltage derived from a sensor within a magnetic field at a given excitation
current as compared to the voltage derived in the absence of a magnetic field for the
same excitation current. The sensitivity coefficient, S, can be calculated by dividing
the output voltage of a sensor by the magnetic field the sensor is measuring [3]
where Hx is the field being measured. The magnetic field is also often expressed in
units of teslas. When these units are preferred, recall from chapter one that the
conversion factor is 1.26pT = lA/rn.
Sensitivity is also often discussed, within the magnetic materials discipline, in
terms of ratios used as coefficients such as AV/V, AR/R, as well as Ap/p, which is the
magnetoresistivity coefficient and can be calculated by[3]
where p/ is the lateral resistivity of the material when the bias current flows parallel to
the magnetization and p, is the transverse resistivity of the material when the bias
current flows perpendicular to the magnetization. The AV/V, AR/R and Ap/p
coefficients are unit-less and often expressed as a percentage. These coefficient
forms are also regularly used instead of the sensitivity coefficient, S, because they
represent a method of quickly comparing relative responses to magnetic fields rather
than calculating specific values.
2.4. Field Noise
Now that an overview of noise and sensitivity has been provided, the relationship
of the two to this research is important to recognize. The experimental results that
were analyzed and compared were in terms of field noise (a term used to differentiate
noise voltage PSD from the same noise PSD converted into field units). It was
calculated by dividing the measured noise voltage PSD of a sensor at a specific bias
current by the measured sensitivity at the same bias current:
A/mJ
V
2-3
Ap _ P/+P,
P SPz+AP
2-4
24


V
l-'icldNoise
T ^Noisc 'vA/hT
_>/Hz_ Sensitivity Vrals/T
2-5
Field noise is of even greater interest than simply comparing intrinsic noise levels
or sensitivities alone because it determines the boundaries of the noise floor for the
sensors. When choosing an MR sensor for any given application, the approximate
range of field magnitudes to be measured should be estimated. Once that range is
known, the field noise of the sensor chosen for the design should be at least an order
of magnitude below the estimated magnetic field range, in order to ensure accurate
results within the linear operating range of the sensing device.
25


3. Magnetic Sensors
There are several different types of magnetic sensors which function in
overlapping ranges of field sensitivity and frequency response. These devices operate
as transducers, converting magnetic field intensity into voltage. Some of the most
common magnetic sensors used in modem products are hall effect, Giant
Magnetoimpedance (GM1), Giant Magnetoresistor (GMR), Anisotropic
Magnetoresistor (AMR), Tunnel Magnetoresistor (TMR) and Fluxgate sensors [3].
Figure 3.1 below illustrates a general range of field sensitivity these sensors are able
to detect. Figure 3.1 shows the required level of sensitivity for many common
applications of Magnetoresistive sensors. The earths magnetic field is about 1 gauss,
or 0.1 mT.
Figure 3.1 Sensitivity Ranges for Common MR Sensors
26


Frequency (Hz)
Figure 3.2 Required Level of Sensitivity for Many Applications of Magnetoresistive Sensors
Table 3.1 provides some interesting average statistics on the most prevalent magnetic
sensors used today.
Table 3.1 Comparison of MR sensors sensitivity to size
Magnetic Sensor Variety Region of Sensitivity Device Volume
Hall Effect Hundreds of nT/>/Hz 0.001mm3
GM1 nT/VHz 0.01mm3
Magneto- electric nT/yfHz 1mm3
GMR Hundreds of pT/VHz 0.001mm3
AMR
TMR
Flux Gate Tens of pT/VHz 1cm3
SQUID* Tens of ff/VHz 1cm3
Superconducting Quantum Interferometer Device
To illustrate the significance of size, Figure 3.3 graphs the relationship between
sensor noise and volume.
27


noise VS- Volume
1.0E+04 Hall + GM1
1.0E+02 MR
br i.on too
jx. £ i 1 0E-02 1 0E-04
4 ME
Proton
+ CSAM Fluxgate
M04 Hc4
^ Hybrid
w + SQUID
SERF
1.0E-06
1.0E-07 1.0F.-05 1.0E-03 E0E-01 1.0E+01
Volume (cm3)
Figure 3.3 Relationship Between Sensor Noise and Size
All of the information above is necessary to develop an effective magnetic field
sensor. AMRs were chosen for the focus of this study because, among other reasons,
they exhibit the best sensitivity in the smallest package. The primary objective of this
project is to understand how to extend the effective region of the AMR sensors down
toward the picotesla sensitivity zone while maintaining a useable bandwidth.
3.1. Ferromagnetic and Antiferromagnetic Layering
Before discussing AMR theory, the subjects of ferromagnetism and
antiferromagnetism must be understood. These two magnetic properties are
temperature dependant and indicate a magnetic materials molecular response to an
external magnetic field. A ferromagnetic (FM) material has the ability to become
permanently magnetized, exhibiting a characteristic pattern of unidirectional
magnetic moments illustrated in Figure 3.4a. Conversely, antiferromagnetic (AFM)
materials without the influence of an external magnetic field exhibit a bidirectional
magnetic moment pattern typical to that shown in Figure 3.4b.
28


t * t * t t I t t t + 4
1 * t * * * i * t t
t * t * + + t * * t t 1
Figure 3.4 a)Ferromagnetic4 and b)Antiferromagnetic5 moments
FM materials are further subcategorized into soft and hard ferromagnets. Soft
FM materials are very susceptible to external magnetic fields (having a high
sensitivity), where hard FM materials require a stronger magnetic field to be affected.
Coercivity (Hc) is the magnetic field required to flip the direction of magnetization in
a magnetic material. It is used to measure the resistance of a ferromagnetic material
to becoming demagnetize. Soft FM materials have a coercivity of less than 10 Oe.
Hard FM materials have coercivities over 100 Oe [4], Antiferromagnetic materials
are also considered hard because their natural spin orientation is minimally effected
by external magnetic fields.
When a soft FM layer is connected to a layer of AFM material at the molecular
level, direct exchange coupling between the two magnetic materials occurs [9], The
effect of this coupling can be observed as a shift in the characteristic magnetization
curve. Sample magnetization curves for FM, AFM and exchange coupled FM/AFM
bi-layer are shown in Figure 3.5.
M MM
a} b) c)
Figure 3.5 Magnetization Curves for a) FM, b) AFM and c) direct exchange coupled FM/AFM bi-
layer illustrating the shift in the magnetization curve6
The direct exchange coupling between an FM and AFM layer is much stronger
than many of the low EM fields that modem devices are supposed to detect. The
sensitivity of a device made from two directly coupled magnetic materials is very
Image retrieved from http://en.wikincdia.ort;. wiki'File: Ferromagnetic ordering.svt
5 Image retrieved from httn: en.wikincdia.on; wiki File: Anti ferromagnetic ordering, svt;
6 Image retrieved from httn:.'/en.wikipedia.org/wiki/File:Exchangebias.png
29


small; the coercivity is too high to be useful for sensitive measurements. Indirect
exchange coupling between two FM materials separated by a nonmagnetic spacer
material, such as Ru, Cu, Rh or Ir, has also been well documented in the 1990s [9],
Wang, et al. first reported on the theoretical indirect exchange coupling between an
FM and AFM layer separated by a nonmagnetic spacer in 2005 [9], The
ramifications of this hypothetical tri-layer structure lead to the supposition that
sensitivity tailoring could be accomplished by adjusting the thickness of the spacer
layer in order to tune an AMR sensor to a sweet spot where the noise to dynamic
range ratio is optimal [10],
3.2. Magnetoresistive Effect
The magnetoresistive effect causes the resistance in some magnetic materials to
be dependant on the state of the materials magnetization. If an external magnetic
field is imposed on the magnetic material, the magnetization of the material could be
altered, causing the resistance to also change. In addition to the first two conditions,
if a bias current is supplied to the magnetic material, the Lorenz force deflects the
path of the charge carriers, affecting an even greater change in the measured
resistance of the magnetic material. This property, which is fundamental to the
operation of MR sensors, is illustrated in Figure 3.6 [3]. The application of a
Figure 3.6 Fundamental Magnetoresistive Effect, a) Current path in a conductor with no external
magnetic field present, b) The effect of the Lorentz force when a magnetic field is applied to the
current carrying conductor, c) The effect of the Lorentz force on the current paths in the conductor.
30


magnetic field alters the length of the current path. This is the mechanism by which
the resistance of the material, and subsequently the resistivity coefficient Ap/p, are
changed. The direction of deflection can be determined using the right hand rule.
3.3. Anisotropic Magnetoresistive Sensors
The term anisotropic is defined in the Merriam-Webster dictionary as exhibiting
properties with different values when measured in different directions. Resistivity in
an AMR is not only affected by the presence of a magnetic field, but also on the
orientation of the current path relative to the magnetic field. This relationship is
described by the Voigt-Thomson formula:
p(i9) = p, sin2 5 + p, cos2i9 = p, + Ap cos2i9 3-1
where 3 is the angle between the direction of current flow and the direction of
magnetization [3].
As discussed in chapter two, the sensitivity of these magnetic field transducers is
a function of the physical dimensions and three-dimensional orientation of the device
detecting the field, as well as the materials chosen to make device and the design
technology employed. Calculation of the sensitivity coefficient can be expanded to
S =
= 2 J L Ap
( \
1 V
V W J A m
3-2
where J is the current density in A/cnT, L is the sensor length in meters, Ap has the
units of Q-m, Hk0 is the anisotropy field of the sensor material in A/m, Ms is the
saturation magnetization in A/m and t/w is the ratio of the thickness to the width. [3]
As mentioned in chapter two, the symbol p/ represents the resistivity of a material
when a bias current flows parallel to the magnetization orientation. The symbol p,
represents the resistivity of a material when the bias current flows perpendicular to
the direction of magnetization. Figure 3.7 illustrates these two resistivities The
directional dependence implied by the sensitivity coefficient calculation, in equation
3-2, is a feature of anisotropic magnetoresistors. Magnetization in such devices
31


Figure 3.7 a) Longitudinal Resistivity (pi) and b) Transverse Resistivity {p,)
usually shows a tendency toward one direction, designated the easy axis or anisotropy
axis, and an adverse tendency away from the direction known as the hard axis [4],
Figure 3.8 Anisotropy axis for ferromagnetic an antiferromagnetic materials.
Figure 3.9 below is a graph of the change in resistance caused by the magnetic
field applied perpendicular to the flow of current. Note that this function is parabolic
(until the inflection point at 45), therefore it has two valid solutions to any given
resistance coefficient. Not only is this a non-linear response, but the direction of the
field can not be determined based on the output of a device like this. To linearize the
transfer curve, the angle of anisotropy is tilted relative to the flow of current (but not
relative to the direction of magnetism).
32


Figure 3.9 Change in magnetoresistive resistance versus magnetization angle.
Applying a magnetic field at an incident angle of 45 from the direction of the current
flow serves to shift the parabolic function, allowing a relatively large linear region
which is nonsymmetrical about the y-axis (Figure 3.10).
AR
Figure 3.10 Shift in MR transfer function when the angle between the direction of B and I is 45.
The slope of this linearized region of operation is used to define the sensitivity of
the device. In addition, the direction of the magnetic field relative to the sensors
orientation can now be described with a unique value.
3.4. Barber pole Differential Bridge Biasing Method
The design topography chosen for the sensors used in this study was a bridge
configuration of four sensors with aluminum barber pole stripes inclined at 45 to
the easy axis. Figure 3.11 shows the AMR devices and the design approach used.
33


Figure 3.11 a) Xic7 CAD drawing of the barber pole design b) microscope image of one of the AMR
sensors fabricated for this study c) Biasing approach for barber pole bridge. The long, solid arrows
indicate the direction of initial magnetization in each of the four bridge elements. The long, dotted
arrows indicate the deflection of the magnetization in each bridge element caused by the external
magnetic field, Hext. The short arrows indicate the direction of current flow as defined by the barber
pole orientation on each bridge element [12].
Placing the conductive strips at a 45 angle to the easy axis causes the bias current to
flow at the same angle, creating the biasing condition discussed at the end of the
previous section, producing a linearized region of operation in the sensor. Making
use of the additional design measure of the alternating bridge configuration allows a
differential measurement and in effect doubles the signal response to the magnetic
field. This design approach, as a rule, produces
7 Xic Graphical Input Editor by Whiteley Research Inc., w ww.wrcad.com
34


The five devices used for this project were composed of AFM layers of iridium
manganese (IrMg), nonmagnetic Ru spacers of varying thicknesses and were topped
with FM layers of Permaloy (nickel iron (NiFe)). Permaloy is a popular alloy to use
for the FM layer because the resistivity coefficient is relatively high [11], The AMRs
had Ru thicknesses of either 0.25nm, 0.5nm, 0.75nm, 1 .Onm or 2.0nm.
Figure 3.12 Trilayer Structure
It was anticipated that the sensitivity of each sample would improve as the
thickness of the ruthenium increased until the Ru layer became so thick that the
exchange coupling no longer created uniaxial anisotropy. Figure 3.13 demonstrates
preliminary results from sensitivity testing performed on the samples in 2007. [10]
300.0
Ru Thickness (nm)
Figure 3.13 Sensitivity Versus Ruthenium Thickness
35


4.
Measurement Strategy
It is critical when collecting telemetry on highly sensitive detectors to have a
clean, low-noise, EMI and EMC compliant electronic test set. The significance of
noise generated within the electronic board itself was explained in chapter two. Each
component of the design which has a first-order effect on the load (i.e. the magnetic
sensor) must be considered for intrinsic noise contribution to the system. In addition,
special care must be taken to protect the elements of the circuit from coupling to
noise sources external to the system. This chapter covers the precautions taken to
ensure the successful collection of data from the devices under test.
4.1. Magnetoresistive Sensor Noise Metrology
Figure 4.1 is a block diagram of the sensitivity measurement system and Figure
4.2 is a block diagram of the noise measurement system used to study the affect of
ruthenium layers on FM/AFM exchange coupling.
Custom LabView Data Acquisition/Control Program
Helmholtz Coils
Figure 4.1 Sensitivity measurement test set
36


Custom LabView Data Acquisition/Control Program

SR785 Signal Analyzer
Figure 4.2 Noise measurement test set
The data collection equipment consisted of a computer, running a National
Instruments LabVIEW Virtual Instrument (VI) Graphic User Interface (GUI) to
command the various electronic components and collect the telemetry from the
device under test (DUT). For measuring Sensitivity, the VI controls a Kepco power
supply capable of supplying 20 amps to a helmholtz coil. The VI also commands
the electronic circuit board to supply voltage to the AMR bridge located within the
helmholtz magnetic field. The bridge input current and voltage and bridge
differential voltage are collected and amplified in the electronic circuit and fed back
to the VI for processing. The results are saved as an Excel file for later analysis.
Next the bridge differential voltage output is connected to the SR785 spectrum
analyzer. The VI GUI is set up to command the spectrum analyzer to a default
configuration optimal for measuring the noise voltage spectrum of the AMR under
test. The entire electronics circuit board is placed in a magnetically shielded
enclosure to isolate the sensor from magnetic fields since the goal is to measure the
amount of noise generated within the AMR regardless of any impinging fields. The
Power Spectral Density (PSD) is sent to the computer VI for processing and is saved
37


as an excel file for analysis. The telemetry from the spectrum analyzer has the units
of Vrms/-/Hz In addition to these basic pieces of equipment, an isolation transformer
will help remove AC (particularly 60Hz) components from the PSD results.
4.2. Electronic Design Considerations
Several revisions were made to the electronic circuit board with the intent of
focusing the function of the circuitry on specific measurements required for
characterizing MR sensors. Information, stored electronically, and the observations
of colleagues who had personal experience with the former versions of the electronics
were gathered and analyzed before a revision plan was developed. Specific attention
was focused on minimizing intrinsic noise within the circuit board and its
components, as well as the addition of certain features designed to expedite the test
and data gathering process.
One prevailing issue with the beta design of the electronics board was a tendency
for the ICs to become damaged when the 12VDc battery power source was
connected. According to reports from the scientists using the electronics, the
connector would usually arc when the power connector was plugged in. Protective
circuitry, shown in Figure 4.2, was added to limit the current surge from the supply
batteries.
Figure 4.3 Inrush Current Limit Circuit.
Insertion of the n-channel, enhancement-mode MOSFET, Ql, into the return path
serves to slow the turn-on time of the circuit boards main power. At the time, t = 0,
when the switch, SI, is turned on, the time constant of R8 and C19 in parallel slows
the gate voltage rise time, in effect placing a high-pass filter on the gate-to-source
turn-on voltage, removing any transient current surges caused by the equivalent
capacitance within the board. The on-transient suppression is demonstrated in
Figure 4.3.
38


140 E-3
Power Siflply Transient Response
-20 E-3 -------------------------------------------------------------------------------------
45 E-6 47E-6 49 E-6 51 E-6 53 E-6 55 E-6 57 E-6 59 E-6
line (s)
Figure 4.4 Modeled transient suppression of inrush current limit protection circuit.
The next major addition to the circuitry was the Digital to Analog Converter
(DAC). This feature, shown in Figure 4.4, was added in order to remotely control the
voltage on the sensor bridge.
Figure 4.5 DAC Bridge Voltage Command.
39


Optocouplers were inserted between the computer interface and the DAC circuitry to
isolate the circuit board from any switching noise conducted from the PC. They are
open collector, hence the pull-up resistors.
EMC concerns were also addressed through careful design of the grounding
scheme. Isolated ground planes were employed for the circuitry powered by different
power supplies. Ground loops were avoided and differential trace routing was
adhered to.
Figure 4.6 Printed circuit board design noise mitigation measures
4.3. Design Validation
Before any of the valuable samples, limited in quantity and irreplaceable, were
installed into the newly designed circuit board, the functionality of the board was
thoroughly vetted. Dummy load bridges were assembled using common carbon fdm
resistors for this testing. Because the intrinsic noise characteristics of ordinary
resistors is well understood, the results of this bring-up testing could be used to
endorse the electronics design. Such verification testing is mandatory in most
electrical engineering design procedures and is always a recommended practice when
using first generation electronics for any experimental data collection and processing.
40


4.4.
Design Limitations
The design of the electronics board incorporated several protection measures;
however it is not completely immune to damage through injudicious commanding.
The amplifier which controls the bridge voltage, and subsequently the bridge
excitation current, was only capable of sourcing 30mA before risking damage. The
maximum voltage commandable from the computer was 8.32V and the lowest bridge
voltage was measured to be about 100Q. This sample load at the maximum
command voltage would demand 83mA from the amplifier almost three times the
rated maximum output current for that part. The bridge resistances of all loads to be
tested with the electronics were measured and the maximum command voltage for the
bridge was limited to ensure the amplifier is not required to deliver more than 30mA
to the load.
Another, less hazardous limitation to the electronics performance was the use of a
finite power source. There was not time to design a voltage monitor for the power
supply, so it was necessary to charge the batteries before every test session and
measure the voltage levels at the beginning of the test day and again at the end. The
DACs required a minimum of 11.4VDc for proper operation, so if the voltage at the
end of the day was found to be less than that, the data results would be suspect.
The last major design limitation which had to be kept in mind was the differential
bridge voltage. The bridge output amplifier on the electronics board had a minimum
gain of 10 which could not be altered or reduced. With a power supply of 12VDC,
the maximum differential voltage across the bridge was IV before the output
amplifier would saturate. A monitoring feature was added to the GUI to ensure the
amplifier remained within its linear operation region to avoid incorporating corrupted
data into the analysis process.
4.4.1. Sensitivity Measurement Process
Sensitivity data was collected using the LabVIEW graphic user interface
developed to control the excitation current to the helmholtz coil via a 20A power
supply, provide a commandable bias current to the sample, and store the resulting
output telemetry from the sensor.
The GUI allows the experimenter to set the field strength produced by the
helmholtz coils. This value is converted to a command signal to the helmholtz power
supply. The number of data sets to be stored can also be selected, determining the
resolution possible for the experiments. All command and telemetry values are
collected and stored by the GUI. These values can be output and stored as an Excel
spreadsheet to use for further analysis.
41


4.5.
Noise Measurement Process
The electronics board is powered by two 12V Lead-Acid batteries configured to
supply 12VDc- A sample is installed at the electronic boards interface connector
and the entire assembly is inserted into a triple-layered isolation chamber. The entire
measurement sequence is controlled by a LabVIEW GUI. This GUI commands the
voltage level applied to the bridge, sets up the SR785 Spectrum Analyzer to measure
the Power Spectral Density of the sample and retrieves the data from the Spectrum
Analyzer for manipulation and analysis.
4.6. From Theory to Practice
Most of the bridge resistances were not as balanced as anticipated by the design
process for the bridge configuration [12] and the minimum excitation current possible
with the eight-bit DAC input design still created a differential voltage greater than the
amplifier could measure. The imbalances of all of the samples were analyzed and the
design modified to accommodate the greatest imbalance. As a result, the excitation
current was reduced from a range of 30mA 300pA (bounded on the upper limit by
the output capacity of the amplifier and the lower limit by the least significant digit
commanding the DAC) to a range of 4mA to 15pA. Recall from equation 3-2 that the
sensitivity coefficient of a magnetic sensor is directly proportional to the current
density through the device. Because the area of a sensor is constant once the device
is made, a lower bias current results in a lower current density and, consequently, a
lower sensitivity.
42


5.
Data Processing
The raw data outputs from the sensitivity measurements and the noise
measurements were processed using Microsoft Excel. All experimental results for
each sample at various bias currents were graphed and can be view in the appendices.
5.1. Analysis
The sensitivity data collected for each sample at specific commandable excitation
currents were processed in Excel. The slopes of the measured volts per gauss equate
to the sensitivity of the samples at specified bias currents. This value was easily
computed using the SLOPE(yi:yn, xi:xn) function.
Next the noise output from the signal analyzer was plotted. The noise voltage
was normalized to the bias current applied at the time of measurement and this value
was plotted on the graph with the raw noise voltage data for comparison. The field
noise was calculated by dividing the noise voltage by the computed sensitivity. As
with the normalized noise data, the field noise normalized to current was added to this
graph.
5.2. Results
After all of the data from the five samples was collected, the highest excitation
current possible which was common in all of the samples was chosen for comparison.
Unfortunately, because of the imbalances in most of the sensors, the highest current
common to all DUTs was only around 60pA. Unquestionably this reduction in
excitation current impacted the overall results of the study.
Sensitivity versus Ru thickness was graphed for each DUT. Figure 5.1 presents
the results of this data analysis. With the exception of one anomalous result, the
sensitivities of the samples did appear to increase with increased Ru thickness, until
the 2.0nm Ru sample apparently exceeded the limits of the indirect exchange
couplings ability to impose uniaxial anisotropy.
43


Sensitivity Versus Ru Thickness
30E-6
25E-6
____ 20E-6
|
& 15E-A '
£
Z
I
^ 10E-6
5E-6
0
Sensitivity
Ru thickness |nm|
Figure 5.1 Measured Sensitivities at 60A for all Five AMR Sensors.
The results of the field noise comparison between the samples are presented in
Figure 5.2.
WK Bridge Field Noise
10E-6
£
I
IE-6
UX1E-9
0.25tmRu (a()A1bm\
0.50rmRu^'0.0ftm\
----0.75miRu (a().045nv\
1 (IXmRu@0.076nv\
lUEr-9
KJM-V3 1E-+0 MM-X) 100E-K) 1 F>3 10E+3 100E+3
Frequency (It/)
Figure 5.2 Field noise measured at 60|aA for AMR sensors with Ru layers of 0.25nm. 0.50nm,
0.75nm. l.Onm, and 2.0nm.
44


As with the results of the analysis of the sensitivity and Ru thickness data in
Figure 5.1, the apparent trend for field noise versus Ru thickness demonstrates
improvement (i.e. lower field noise) for thicker insulation layers until the transition
between 1.0 nm and 2.0 nm Ru, at which point, the presumed lack of exchange
coupling appears to degrade the noise performance of the AMR sensor.
5.3. Conclusions
Within the limits of this project, the two sources of noise which dominate the
experimental results were clearly 1/f noise and Johnson noise. The effects of other
sources of interference cannot be extracted from the data with the experimental
equipment and data processing method used. The effects of the two overriding noise
sources are visible in the example processed noise data, highlighted in Figure 5.3.
AM* Bridge Noise
025frnRu'j 0.06m\
0 50nrtRufcf OOfrrA
----------075f¥nRufa 0045m\
1 (XkiiiRufci 0.07(*iA
2 (XlnnRn f'O.Ofro\
Figure 5.3 Johnson noise (white) and 1/f noise (pink) characteristics evident in experimental data.
The results of the sensitivity analysis and the field noise analysis indicate that the
l.Onm Ru insulator thickness is optimal for this particular sensor design. The
dynamic range for this device was still well above the picotesla range, though, so
while the sensitivity improved, the overall performance of the sensors did not
demonstrate the projected enhancement. For future studies on these AMR sensors,
closer examination of the noise floor of the entire system may reveal the source of the
excess interference, or demonstrate that this is an inherent limitation of this particular
design.
45


6.
Discussion
The one inconsistent measurement in the sensitivity versus Ru thickness analysis
(Figure 5.1) was on the 0.75 nm Ru sample. This was interesting because this sample
initially appeared to be the most appropriate for such analysis. The AMR bridge was
the best balanced; therefore the bias current possible through this sample was limited
by the maximum allowable power rating of the die rather than the limitations
described for all of the other samples in section 4.4. The maximum bias current for
this sample was 9.11 mAoc- As a result, the highest sensitivity measured for this
device was 17.5 mV/G or, converted to SI units, 175 V/T. The measured field noise
at the maximum bias current was one to two orders of magnitude lower than at the
common bias current level available for the evaluation analysis of this study. Graphs
of these results were included in Appendix B as points of interest.
Consider what it would mean if the test conditions and samples were only
vulnerable to first order effects. It could then be inferred that the change in sensitivity
between the different devices would track linearly. The sensitivity of the 0.75nm Ru
sample improved from 4pV/G with a bias current of 60pA to 17.5mV/G with a bias
current of 9.11mA; a 4300% improvement. If the increase is similar for the 1 .Onm Ru
sample, the sensitivity of that device could, theoretically, approach 50mV/G (or
500V/T, or 625mV/(kA/m)). Unfortunately, due to the imbalance in the l.Onm Ru
samples bridge and the limitations of the test equipment available, this measurement
can only be speculated about. For comparison, however, the sensitivity of the
common Philips Semiconductor brand AMR sensor, KMZ10B, is about 48mV/(kA/m)
and a highly sensitive AMR available from that company boasts a sensitivity of
200mV/(kA/m); 1 /3rd that of the theoretical results for the 1 .Onm Ru sample tested [3].
The dynamic range of the KMZ10B is advertised to be 0-1 MHz. The noise
performance at low frequencies, however, is not published in the datasheet provided
by Philips Semiconductor. Without actually subjecting one of those sensors to the
same testing that the Ru samples underwent, this value is, to use an engineering
analogy, floating (without common reference). Further testing to corroborate these
hypothetical deductions would be very interesting.
In addition, testing on samples with Ru thicknesses between l.Onm and 2.Onm
may reveal an even better sensitivity. Unfortunately there are no points of reference
between the apparent maximum sensitivity with a 1.0 nm Ru layer and the unusable
2.0 nm Ru sample, so it is neither known at what point the sensitivity of the device
ceases to improve, nor how quickly the sensitivity deteriorates after the maximum
insulation layer thickness is exceeded.
46


The topic of noise in electronics is extremely relevant in todays ever increasingly
technological society. With the requirements for faster processor speeds comes the
need for electronic circuits which do not generate excessive noise and/or are resistant
to outside noise sources. Advances in circuit fabrication methods have not only
allowed for the miniaturization of electronics, but also drastically decreased the
power required to operate these new devices. As the required power levels decrease
further, however, the normal operating region of these tiny circuits is overpowered by
environmental charge fluctuations known as noise. When the signal-to-noise ratio
decreases toward one, the electronic component is no longer useable. Research, such
as this study, which advancing the understanding of noise issues and of methods for
noise mitigation, is critical the further development of magnetic field sensing
technology.
47


APPENDIX A
Field Noise and Normalized Noise By Device
48


Noise ((\'(rinsV\ll/)/A) Noise (Icshl'll/)
0.25nm Ru Sample Data
1.00H-05
l.OOh-06
1 .(X)PX)7
1.00E-01
1 .(X1E-05
l.(X)H-<)6 ;
l.(X)E-07
1.00E-08
1 .(X)E-01
AMR Bridge //<'/ -----C<£(>.059m\
(a;0.030m\
(o:0.015rrA
l.(X)E(X) l.(X)E 01 l.(X)H<02 I.OOE 03 1.00E04 I.OOE-OS
Frequency (Hz)
Figure A.1.1: 0.25nm Ru Sample Field Noise.
AMR Bridge V*'*'(Sanplc 0.25nm Ru, Normalixd to Currents)
-----@0.059nA
' ( (^0.015rrA
I ^
l.(X)E (X) l.OOE-OI l.(X)E ()2 1 .OOF.-03 I.OOE-04 I.OOE 05
Frequency (Hz)
Figure A.1.2: 0.25nm Ru Sample Normalized Measured Noise.
49


Noise ((\(rms)/\lt?)/A) Noise (Tesla/VHz)
0.50nm Ru Sample Data
l.(X)E-06
1 .(X)E-<)7
1. i.oon-<)9
l.OOE-Ol
l.(X)E-()6
l.(X)E-<)7
l.(X)E-OX
l.(X)E-<)9
l (X)E-Ol
AMR Bridge Field Soise (Sarrple 0.50nm Ru)
------(a>0.121rrA
-----(a>0.()60rrv\
(£t;0.03fliTA
(t^O.OLSnvX.
l.(I)E Frequency (Hz)
Figure A.1.3: 0.50nm Ru Sample Field Noise.
AMR Bridge Sense (Sanple 0.50nm Ru. NornwIiAxi to C urrents)
-----(a.0. l21rrA
-----(2t<().060m\
(t£0.03()rrA
<&>0.015itA
l.(X)E Frequency (Hz)
Figure A.1.4: 0.50nm Ru Sample Normalized Measured Noise.
50


Noise ((V(rms)/\H/.)/A) Noise (Tesla/\1 b)
0.75nm Ru Sample Data
1.00K-06
l.(Mvfl7
1 .OOl.-OK :
1 ,(X)h-09
l.(X)E-H) :
ux)E-ii
l.(X)h-12
lOOE-OI
uxhxxs -
UXtXT?
1 , l.(X)E-09
i. l.(X)E01
AMR Bridge Heid Stnse (Sanple 0.75nm Ru)
(O/0.914rrA
(£0.462m\
(d0.230rrA
-----(o-0.107trA
----(a;0.092nV\
-----C40.(H6n>\
----(4'0.03nA
^0.015irA
1.00EHX) l.(X)trX)] ).(X)EH)2 1.00hH)3 l.(X)li+ Frequency (I b)
Figure A. 1.5: 0.75nm Ru Sample Field Noise.
AMR Rridge ,Vv'(Sanpic 0.75nm Ru, Nomwibed to Currents)
-----Ca;0.914nv\
----(a^).462n>\
| rtt0.230m\
! -----(a;0.H)7n>\
l.(X)Ht(X) l.(X)E*Ol ).EKC l.(X)HK)3 l.OOEMM l.(X)EK)5
Frequency (Hz)
Figure A. 1.6: 0.75nm Ru Sample Normalized Measured Noise.
51


Noise ((V(rins)/\1 1/VA) Noise (Tcsla/vl Iz)
1 .Onm Ru Sample Data
l.<)Oh-06
l.00h-07
UX)MK
UJOMW
l.OOB-Ol
1 ,(X)1>()6
i oot- 1 .(X)h-OK
1.<*>|--CW
I ,OOL-f)l
AMR Bridge f ield.\oise (Sairple I.Omn Ru)
(o.0.250rrA
-----(a.O.lTOrrA
(a 0.076rrA
(a 0.015irA
I.(KH-.HX) l.OOL-K)! !.(XvH)2 l.OOlXtf l.OOhHM 1.00t-K)5
Frequency (Itz)
Figure A.1.7: l.Onm Ru Sample Field Noise.
AMR Bridge V^(Sanpie l.Onm Ru. Nomulized to Currents)
(a-0.250rrA
----(a 0.170rrA
(a 0.076rrA
(a 0.030rrA
(a0.015trA

l.h+<>0 1.00h-H>l !,(X)h+<)2 l.OOfc+03 1 OOIHU 1.00h+O5
Frequency (I tz)
Figure A. 1.8: l.Onm Ru Sample Normalized Measured Noise.
52


Noise ((VirnKt/VHzVA) Noise (Tesla/VHz)
2.0nm Ru Sample Data
1.00E-05
l.(X)E-06
l.(X)E-07
l.(X)Er(>l
I .OOH-05
l.(X)E-06
I.OOE-07
). l.(X)Fr()l
AMR Bridge Held \wr.*'(Sanple lOnmRu)
-----@0.()60rrA
(a/<).030rrA
(a-O-OlSirA
l.(X)EKX) l.(X)EK)l l.(X)E02 l.(X)HK)3 l.OOEKW l.(X)E)5
Frequency (Hz)
Figure A.1.9: 2.0nm Ru Sample Field Noise.
AMR Brieve Awf.v£(Sanple 2.0nm Ru Normalized to C urrents)
(^0.060ny\
(a/0.030rrA
@0.015rtv\
l.(X)EKX) l.(X)EH)] UXIEKC l.(X)HK)3 l.(X)F.M>4 l.(X)Ei05
Frequency (Hz)
Figure A.1.10: 2.0nm Ru Sample Normalized Measured Noise.
53


APPENDIX B
Sensitivity and Noise Data by Device and Bias Current
54


0.25nm Ru Sample Data with 0.059mA Bias Current
SensitiMtv
(BHcfee Current = O.US>m\)
-24.0E-3
-24.0E-3
-24.0E-3 "
-24. IE-3 -
-24. IE-3 -
£ -24.IE-3 -
&
£ -24.1E-3 '
-24.IE-3
-24.2E-3 -
-24.2E-3 1
-24.2E-3
-24.2E-3 -
-40 -30 -20 -10 0 10 20 30
Magnetic Held (Ciauss)
40
Figure B.1.1: Measured Sensitivity for 0.25nm Ru Sample when Bias Current is 0.059mA
= 1.302 pV/G.
I0.00F-6 ;
i .oof-6 :
a
IOO.(X)F-9 ;
AMR Noise (0.25nm Ru (a 0.059m\)
Fidd Noise
[J/'H']
I.00E-9..........................
1.0F-0I I .OF (*)
I .OF Ot 1 .OF 02
Frequency (Hz)
Figure B.0.1: 0.25nm Ru Sample Noise with 0.059 Bias Current.
55


0.25nm Ru Sample Data with 0.030mA Bias Current
Sensithity
(Bridge Current = (ML30rrv\)
-12.12E-3
-12.I4E-3
-12.16F.-3
£ -I2.18E-3
|
J -12.20K-3
-12.22E-3
-I2.24E-3
-12.26H-3 _t-------------1------------
-40 -3() -2() -10 0 10 20 30 40
Magnetic Held (Gaum)
Figure B.1.3: Measured Sensitivity for 0.25nm Ru Sample when Bias Current is 0.030mA
= 646.34 nV/G.
AMR Noise (0.25nm Ru (a 0.030nv\)
- Field Noise
[T7^]
t
i.of oy
I.OF-Ol
I .OF 00 1.0FO1
I OR 02 I .OF. 03 I .OF CM
I .OF 05
Frequency (H^)
Figure B.1.4: 0.25nm Ru Sample Noise with 0.030 Bias Current.
56


Voltage (V)
0.25nm Ru Sample Data with 0.015mA Bias Current
Sensitivity
(Etridge Current = 0.015mA)
-6.14F.-03
-6.16E-03
-6.18E-03
-6.20E-03
-6.22E-03
-6.24E-03
-6.26F.-03
-6.28E-03
-40 -30 -20 -10 0 10 20 30 40
Magnetic Held (Gams)
Figure B.1.5: Measured Sensitivity for 0.25nm Ru Sample w hen Bias Current is 0.015mA
= 290.17 nV/G.
I.0E-05
AMR Noise (0.25nm Ru (a 0.015m\)
Field Noise
[T7VFU]
I 0F-06
----Vfcasircd Noise
[Mrm>)/VHz]
----Vfcasired Noise
NonraliAxi to
C'uren
[(Mrm>AHz)/A]
1.0H-09
I OF 00 I OF 01 I .OF 02 1.0F 03 1 OF 04 1.0F 05
Frequency (Hz)
Figure B.1.6: 0.25nm Ru Sample Noise with 0.015 Bias Current.
57


0.50nm Ru Sample Data with 0.121mA Bias Current
Sensitivity
(Bridge Current = O.I21m\)
24.2E-3
24.0E-3
7 23.8E-3
I
% 23.6E-3
23.4E-3
23.2E-3 -
23.0E-3 -
22.8E-3 -
-70
-50 -30
Magnetic Meld (Gauss)
Figure B.1.7: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is 0.121mA
= 11.198 pV/G.
Frequency (H^)
Figure B.1.8: 0.50nm Ru Sample Noise with 0.121 Bias Current.
58


0.50nm Ru Sample Data with 0.060mA Bias Current
Senati\itv
(Bridjjt* Current = 0.060mA)
11.7B-3 |
11.6E-3
11.5E-3 p
11.4E-3 p
11.3E-3
n.OE-3









-10 10
Magnetic Field ((iauss)
&

Figure B.1.9: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is 0.060mA
= 5.498 pV/G.
AMR Noise (0.50nni Ru (a 0.060mA)
Fidd Noise
[TAA-fc]
I 0F.-09
I .OR 10
I OROI
I OF 00 I OF 01 I .OF 02 I .OF 03 I .OF 04 1 OF 05
Frequency (Hz)
Figure B.1.10: 0.50nm Ru Sample Noise with 0.060 Bias Current.
59


0.50nm Ru Sample Data with 0.030mA Bias Current
5.3E-3
5.3E-3
5.2Er3
5.2E-3
7 5.1E-3
I
3 5.1E-3
5.0E-3
5.0E-3
4.9E-3
4.9E-3
Senstisitv
(Bridge Current = (M130m\)
-70 -50 -30 -10 10 30 50 70
Magnetic Held (Gauss)
Figure B.1.11: Measured Sensitivity for O.SOnm Ru Sample when Bias Current is 0.030mA
= 2.700 pV/G.
*
t
I OE-Ofc I
l.OF-IW
I .OB
AMR Noise (O.SOnm Ru (a 0.030mA)
Field Noise
[T/'iHt]
~ Vfcasirod Noise
[Wm*y>/Ki]
' Vfcasirod Noise
Norrrnliasdlo
Clrrcrt
[(Wm*y>*tyA]
I OF < I .OF 01 l .OF <12 I .OF 03 t.OF Frequency (Hz)
Figure B.1.12: 0.50nm Ru Sample Noise with 0.030 Bias Current.
60


0.50nm Ru Sample Data with 0.015mA Bias Current
Sensiti\it\
(Bridge Current = 0.015m\)
5.3E-3
5.3E-3
5.2E-3
5.2E-3
£ 5.1B-3 '
I
5 5.1E-3
5.0E-3
S.OE-3
4.9E-3
4.9E-3
-70 -50 -30 -10 10 30 50 70
Magnetic Held (Gbiks)
Figure B. 1.13: Measured Sensitivity for 0.50nm Ru Sample when Bias Current is 0.015mA
= 2.700 pY7G.
I OF 06
1.0F-07 -
*
2
AMR Noise(0.50nmRu(S0.0I5m\)
Fidd Noise
----Measured Noise
[MrmsyVFL']
----Meastred Noise
Norrmli^Ed to
C'inert
[(MrrrKVN^VA]
l.OF 09 - .... .......................................... ....................... ..................... .......................
I.0E-01 I OF 00 I OF 01 1 OF 02 l.OF 0.1 l.OF 04 l.OF OS
Frequency (H/.)
Figure B.1.14: 0.50nm Ru Sample Noise with 0.015 Bias Current.
61


0.75nm Ru Sample Data with 0.914mA Bias Current
Sensiti\itv
(Bridge C urrent = 0.914n v\)
0.60
0.40
0.20
(>.)
-0.20
| -0.40
T
* -0.60
-o.so
-1.00
-1.20
-1.40 ----
-60 -40 -20 0 20 40 60
Magnetic Held ((muss)
Figure B.1.15: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.914mA
= 17.464 mV/G.
AMR Noise (0.75nm Ru (a 0.914mA)
Field Noise
[TA/Hc]
1 OF 00 l.OF 01
Vfcasired Noise
Normal i^xi lo
Qrrert
[(MrrrB)/VH^)/A]
I OF 04 l.OF 04
Frequency (Hy)
Figure B.1.16: 0.75nm Ru Sample Noise with 0.914 Bias Current.
62


0.75nm Ru Sample Data with 0.462mA Bias Current
Sensitivity
(Bridge Current = 0.462rrv\)
0.30
0.20--------------------------------
0.10---------------------------------
0.00 1-------------------
-0.10 -
| JIM

^ -0.30
-0.40 - - ----------- ---------
-o.5o :--------------------1-------------------
-0.60 - ----------------------------j----------------------------------------
-0.7() - ------------------------------------------------------------------------------------------------------------------
-60 -40 -20 0 20 40 60
5Held (Ciauss)
Figure B.1.17: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.462mA
= 8.907 mV/G.
AMR Noise (0.75nm Ru (a 0.462rrv\)
l.OF-Ol 1.0E 00 I (IF 01 1.0F.02 I .OF 03 I .OF. 04 l .OF 05
Frequency (Hz)
Figure B.1.18: 0.75nm Ru Sample Noise with 0.462 Bias Current.
63


0.75nm Ru Sample Data with 0.230mA Bias Current
ScfRitiut)
(Britlge Current = 0.230m\)
1.50E-03
1.00E-03----------------------------- - -
5.00E-04 ----r-------
O.OOE-OO -
£
I
-5.00E-04 ----------------
-1.00E-03 i - -----------------------
-1.50E-03 - _ ------------------------------
-2.00E-03 ' '
-15() -100 -50 0 50 100 150
Magnetic Held (< ihuss)
Figure B.1.19: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.230mA
= 15.654 pV/G.
AMR Noise (Q.75nm Ru (a 0.230nv\)
Fidd Noise
[T/VR']
1.0E-09 - - -.................-------------'+ ^ 1..........................*--' *->-***m--------'1
1.0F-01 I.OEOO l.OFOl 1.0F.02 I 0E 0.3 I.OF. 04 l.OF. 05
Frequence (H/.)
Figure B.1.20: 0.75nm Ru Sample Noise with 0.230 Bias Current.
64


0.75nm Ru Sample Data with 0.107mA Bias Current
-I00.0E-6
-3(X).OE-6
-500.0E-6
-7(X).OE^6
- -1.1E-3
-1.3E-3
Senti\itY
(Ciirrent = 0.107m\)
-60 -40 -20 0 20 40 60
Magnetic Held ((iiitss)
Figure B.1.21: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.107mA
= 9.502 pV/G.
AMR Noise (0.75nm Ru (a 0.107 mA)
&
*
2
----Field Noise
Masirod Noise
[WnruyVH:]
Measired Noise
Norrralized to
Cirrcrt
[(M rrrsV'^fcyAJ
I .OB-09 * ------------------------ ' .....
1.0B0I I.OF. OO I OF 01 1.0F 02 1 OF 05 I OF m 1.0E-05
Frequency (Hz)
Figure B.1.22: 0.75nm Ru Sample Noise with 0.107 Bias Current.
65


Noise
0.75nm Ru Sample Data with 0.092mA Bias Current
80.0E-3
60.0E-3
40.0E-3
20.0E-3
(XXXOE-O
-20.0E-3
-40.0E-3
-60.0E-3
-SO.OE-3
-l(X).()E-3
-120.0E-3
-140.0E-3
Senatnity
(Bridge Current = 0.092nv\)

-60 -40 -20 0 20 40 60
INfagnetic Reid ((jIuirs)
Figure B. 1.23: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.092mA
= 1.805 mV/G.
AMR Noise (0.75nm Ru (a (l.(W2ni\)
I.0F-07 :
Fidd Noise
i oe-10 :

I .OB- 11.................................-.......- 1 ' "------ -----------------'
1.0B-01 I.OB 00 I.OF-OI I .OF. 02 1.0E 03 I OF. Frequency (Hz)
Figure B.1.24: 0.75nm Ru Sample Noise with 0.092 Bias Current.
66


0.75nm Ru Sample Data with 0.046mA Bias Current
Se nativity
(Bridge Current = 0.(U6nv\)
& -X00.0E-6





-20 0 20
Magnetic Field (t>uuvs)
Figure B.1.25: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.046mA
= 4.000 pV/G.
AMR Noise (0.75nm Ru (a 0.(M6mA)
1 .OF-06
Field Noise
[T/VFfc]
-----Vfcastrcd Noise
[MrmiVVFk]
I.OF-07
*
2
----jVfcasircd Noise
NonraliAxj to
Clrrctl
[(WmsyAtyA]
l.OR-OS 1
1 0B-09
I.OF.
Figure B.1.26: 0.75nm Ru Sample Noise with 0.046 Bias Current.
67


0.75nm Ru Sample Data with 0.030mA Bias Current
Sereatmty
(Bridge Current = 0.092mA)
40.0E-3
30.0E-3 ------------------------- ---- -
20.0E-3 I ---------------- *
IO.OE-3 '
(KXXOE-K) - -
-IO.OF.-3 *

-20.0E-3 - !
i

-30.0E-3

-60 -40 -20 0 20 40 60
Magnetic Meld (Ciauss)
Figure B. 1.27: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.030mA
= 594.39 pV/G.
AMR Noise (0.75nm Ru (6 0.03itv\)
I OF-Oh .
----Fidd Noise
[T/\/Hz]
I OH-07 :
I.OF-OK
U
X
1
I OF 10 :
i

Vtastrcd Noise
[MrrrsyVHz]
-----Vfcasurcd Noise
NorTrali^cd to
Cirrcrt
I .OF 01
l.OF
1 OF. 03 I OF. m l .OF. 05
Frequency (Hz)
Figure B.1.28: 0.75nm Ru Sample Noise with 0.030 Bias Current.
68


0.75nm Ru Sample Data with 0.015mA Bias Current
-750.0E-6
Senstnitv
(Bridge Current = 0.015m\)
-8(X).()E-6
-850.0E-6
I
-XKl.OE-^
-950.0E-6
-l.OE-3
-60 -40 -20 0 20 40 60
Magnetic Field
Figure B.1.29: Measured Sensitivity for 0.75nm Ru Sample when Bias Current is 0.015mA
= 1.228 pV/G.
Field Noise
Vfcasircd Noise
[MrrrKVVFt^]
*
*
Maisired Noise
NomHliAd lo
Cirrert
[(MnTKVVFL)/A]
Figure B.1.30: 0.75nm Ru Sample Noise with 0.015 Bias Current.
69


Noise
l.Onm Ru Sample Data with 0.250mA Bias Current
-17.0E-3
Sensitnity
(Bridge Current = 0.25mA)
-18.0E-3
-19.0E-3
& -20.0E-3
-21.0E-3
-22.0E-3
-23.0E-3 * - -------1-------------1-------------1-------------1-------------
-40 -30 -20 -10 0 10 20 30 40
Magnetic Held(Cjiauss)
Figure B.1.31: Measured Sensitivity for l.Onm Ru Sample when Bias Current is 0.250mA
= 52.905 pV/G.
l.OF-06 -
AMR Noise (l.Onm Ru (ft 0.25nv\)
Fidd Noise
[T/Vl-k]
Vfcasired Noise
I.0E-09--------- -------------------'------- ------ ----------*.......................................... J.........1--------1---------- 1'
I.OE-OI I OF. 00 I OF. 01 1 OF 02 I OF 03 l.OF Frequency (H/)
Figure B.1.32: l.Onm Ru Sample Noise with 0.250 Bias Current.
70


l.Onm Ru Sample Data with 0.170mA Bias Current
Sensitixitv
(Bridge Current = 0.17mA)
Figure B.1.33: Measured Sensitivity for l.Onm Ru Sample when Bias Current is 0.170mA
= 25.113 pV/G.
AMR Noise (l.Onm Ru (a0.17nv\)
&
t
Mastrod Noise
[WrmoMfc]
Vfcasired Noise
Nnrrralisdto
Cirrerl
[ Frequency (I l/i
Figure B.1.34: l.Onm Ru Sample Noise with 0.170 Bias Current.
71


I.Onm Ru Sample Data with 0.076mA Bias Current
-4.4E-3
-4.5E-3
-4.6E-3
-4.7E-3
-4.8E-3
| -4.9E-3
-5.0E-3
-5.1E-3 -
-5.2E-3
-5.3E-3 '
-5.4E-3 *
Sensitivity
(Bridge Current = 0.076nvV)
.:,sr
-10 0 10
Magnetic Held ((Tauss)
Figure B.1.35: Measured Sensitivity for I.Onm Ru Sample when Bias Current is 0.076mA
= 11.273 pV/G.
AMR Noise (1 .(him Ru (a 0.076nv\)
I .OE-06
Field Noise
VLastred Noise
[Mnreys/Hz]
Frequency (H/)
Figure B.1.36: I.Onm Ru Sample Noise with 0.076 Bias Current.
72


Noise
l.Onm Ru Sample Data with 0.030mA Bias Current
Senstivitv
(Bridge Current = ().030nv\)
Figure B.1.37: Measured Sensitivity for l.Onm Ru Sample when Bias Current is 0.030mA
= 4.378 pV/G.
AMR Noise (I .(him Ru <& 0.030mA)
Field Noise
[T'ltfc]
----Vfcasircd Noise
[Mm^yVHz]
-----Mcasirod Noise
NorrraliiEd to
Cirrcrt
[(WnreV>*feVA]
l .OB-09 ---------
1 OBOI I OF 00
OF 01
Frequency (H/.)
Figure B.1.38: l.Onm Ru Sample Noise with 0.030 Bias Current.
1.0E 05
73


l.Onm Ru Sample Data with 0.015mA Bias Current
-1.45E-3
Sensiti\ity
(Bridge Current = 0.015mA)
-1.50E-3
-I.55E-3
I
^ -I.60E-3
-1.65E-3
-1.70E-3
-40 -3() -20 -10 0 10 20 30 40
Magnetic Held (Cik)
Figure B.1.39: Measured Sensitivity for l.Onm Ru Sample when Bias Current is 0.015mA
= 2.063 pV/G.
AJV1R Noise (l.Onm Ru (a 0.015m\)
Field Noise
[T/VFk]
1
*
z
NormiliAxl to
C'unert
[(MrrrK)/'sH')/A]
10F-09................
I.OFOI l.OF <*)
Figure B.1.40:
I .OF 01 I OF 02 I OF. 03 I OF. 04 I .OF 05
Frequency (H/.)
.Onm Ru Sample Noise with 0.015 Bias Current.
74


Noise
2.0nm Ru Sample Data with 0.060mA Bias Current
Sereathity
(Bridge Current = 0.060m\)
-23.32E-3 ' .... -----------------------------------------------1......."j-----------i---------
-23.34E-3 ' * ------ --------------;---- ---------------
-25 -20 -15 -10 -5 0 5 10 15 20 25
Magnetic Held (Gauss)
Figure B.1.41: Measured Sensitivity for 2.0nm Ru Sample when Bias Current is 0.060mA
= 1.972 pV/G.
AMR Noise (2.0nm Ru (a. 0.060m\)
Fidd Noise
[T/^fc]
1.0F-01 I.OF- 00 I .OF Ol I OF 02 I .OF.-03 I OF Frequency (Hz)
Figure B.1.42: 2.0nm Ru Sample Noise with 0.060 Bias Current.
75


2.0nm Ru Sample Data with 0.030mA Bias Current
Senrithity
(Bridge C urrent = 0.(130nv\)
Figure B.1.43: Measured Sensitivity for 2.0nm Ru Sample when Bias Current is 0.030mA
= 754 nV/G.
AMR Noise (2.0nm Ru (a 0.030m\)
I.0E-G5 ;
Fidd Noise
[T/'A-t-]

1.0E09.................. ...................... .....-1---------~----------i * .....
l.OF-OI I .OF 00 1 OF 01 1.0F. 02 1.0F 03 1 OF Frequency (Hz)
Figure B.1.44: 2.0nm Ru Sample Noise with 0.030 Bias Current.
1 .OF 05
76


2.0nm Ru Sample Data with 0.015mA Bias Current
Sensitivity
(Bridge Current = 0.015m\)
-4.94E-03
-4.%l>03
-4.9XK-03
p -5.00H-03
1
Z -5.G2E-03
-5.04H-03
-5.06F-03
-5.0KH-03
-25 -20 -15 -10 -5 0 5 10 15 20 25
Magnetic Field ((imss)
Figure B.1.45: Measured Sensitivity for 2.0nm Ru Sample when Bias Current is 0.015mA
= 429 nV/G.
A\1R Noise (2.0nmRu(a 0.015mA)
Fidd Noise
[T/'lHt]
&
l
Vfcasircd Noise
[MrrrByVFk]
-----Vfcasircd Noise
NorrraliAxi to
Cirrcri
l7\ffrvcV-\/l-GV A!
I OF.-OK
l.OFi-tW................. ................ ................... ........ ...................
I.0F-0I 1.0F 00 1 OF Ol 1.0F 02 I .OF 03 I .OF (W
Frequency (Hz)
Figure B.1.46: 2.0nm Ru Sample Noise with 0.015 Bias Current.
I OF. 05
77


REFERENCES
[ 1 ] World Magnetic Sensor Components and Modules/Sub-systems Markets,
Frost & Sullivan, Palo Alto, CA, Market Report #F670-32, 2006.
[2] J. C. Mallinson, Magneto-Resistive and Spin Valve Heads, 2nd ed., l.Mayergoyz, Ed.
San Diego: Academic Press, 2002, p. 174.
[3] S. Tumanski, Thin Film Magnetoresistive Sensors, in Series in Sensors, B. E. Jones,
Ed. Philadelphia: Institute of Physics Publishing, 2001, pp. 1,5, 15, 116, 326.
[4] S. X. Wang and A. M. Taratorin, Magnetic information Storage Technology,
l.Mayergoyz, Ed. San Diego: Academic Press, 1999, pp. 20, 21, 233, 257.
[5] P. Horowitz and W. Hill, The Art of Electronics, 2nd ed., New York: Cambridge
University Press, 1989, pp. 431-433.
[6] L. V. Meisel and P. J. Cote, Power Laws, Flicker Noise, and the Barkhausen Effect,
Physical Review B, vol. 46, no. 17, pp. 10 822-10 828, Nov. 1992.
DOI: 10.1103/PhysRevB.46.10822
[7] A. Hubert and R. Schafer, Magntic Domains, the Analysis of Magnetic
Microstructures. New York: Springer-Verlag, 2000, p. 434.
[8] J. R. Brauer, Magnetic Actuators and Sensors, M. E. El-Hawary, Ed. New Jersey:
Wiley-Interscience, 2006, p. 209.
[9] D. Wang, J. Daughton, C. Nordman, P. Eames, and J. Fink, Exchange Coupling
Between Ferromagnetic and Antiferromagnetic Layers via Ru and Application for a
Linear Magnetic Field Sensor, Journal of Applied Physics, vol. 99, pp. 256-257, Apr.
2006. doi: 10.1063/1.2162507.
[10] S. T. Halloran, F. C. S. da Silva, H. Z. Fardi, D. P. Pappas, Permanent-magnet-free
Stabilization and Sensitivity Tailoring of Magnetoresistive Field Sensors, Journal of
Applied Physics, vol. 102, pp. 46-49, Aug. 2007. DOI: 10.1063/1.2764243.
[11] E. M. Williams, Design and Analysis of Magnetoresistive Recording Heads. New
York: Wiley-Interscience, 2001, p. 21.
[12] F. C. S. da Silva, S. T. Halloran, A. B. Kos, and D. P. Pappas, 256-channel Magnetic
Imaging System, Review of Scientific Instruments, vol. 79, pp. 128-134, Jan. 2008.
DOI: 10.1063/1.2829883.
78