Influence of cation choice on magnetic behavior of III-N dilute magnetic semiconductors

MISSING IMAGE

Material Information

Title:
Influence of cation choice on magnetic behavior of III-N dilute magnetic semiconductors
Physical Description:
x, 92 leaves : ill. ; 29 cm.
Language:
English
Creator:
Frazier, Rachel Marian
Publication Date:

Subjects

Subjects / Keywords:
Materials Science and Engineering thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Materials Science and Engineering -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2005.
Bibliography:
Includes bibliographical references.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Rachel Marian Frazier.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 028077350
oclc - 847496254
System ID:
AA00022649:00001

Table of Contents
    Title Page
        Page i
    Acknowledgement
        Page ii
    Table of Contents
        Page iii
        Page iv
    List of Tables
        Page v
    List of Figures
        Page vi
        Page vii
        Page viii
    Abstract
        Page ix
        Page x
    Chapter 1. Introduction to dilute magnetic semiconductors
        Page 1
        Page 2
        Page 3
    Chapter 2. Theoretical and experimental background and investigations
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
    Chapter 3. Survey of potential dilute magnetic semiconductors with ion implantation
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
    Chapter 4. Thin film alluminum nitride dilute magnetic semiconductor growth and optimization of magnetic properties
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
    Chapter 5. Thermal stability investigation
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
    Chapter 6. ALN and GAN based dilute magnetic semiconductor applications
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
    Chapter 7. Investigation of gadolinium as a magnetic dopant
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
    Chapter 8. Conclusion
        Page 82
        Page 83
        Page 84
        Page 85
    Appendix. Superconducting quantum interference device magnetometry
        Page 86
        Page 87
        Page 88
    List of references
        Page 89
        Page 90
        Page 91
    Biographical sketch
        Page 92
        Page 93
        Page 94
Full Text










INFLUENCE OF CATION CHOICE ON MAGNETIC BEHAVIOR OF III-N DILUTE
MAGNETIC SEMICONDUCTORS














By

RACHEL MARIAN FRAZIER


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005














ACKNOWLEDGMENTS

Many people helped support the research presented in this dissertation, too many to

list individually. I appreciate all that each person contributed. In particular, I thank each

member of my research group, especially those who taught me the basics, requirements

for success in my research. I also thank each member of my committee for the inspiring

thoughts and insightful conversations. I have met many and learned much during my

graduate career at UF.















TABLE OF CONTENTS

page

ACKNOWLEDGMENTS .................................................................................................. ii

L IST O F T A B L E S ............................................................................................................... v

L IST O F FIG U R E S ........................................................................................................... vi

CHAPTER

1 INTRODUCTION TO DILUTE MAGNETIC SEMICONDUCTORS....................... 1

2 THEORETICAL AND EXPERIMENTAL BACKGROUND AND
IN V ESTIG A TIO N S ..................................................................................................... 4

Experimental Progress in DMS .................................................................................... 4
Focus on Ill-A s...................................................................................................... 5
Extension to III-N s ................................................................................................ 5
Theoretical Background................................................................................................ 7
Free Carrier Mediated Model of Ferromagnetism in DMS................................... 7
Percolation Picture of Ferromagnetism in DMS ................................................... 8

3 SURVEY OF POTENTIAL DILUTE MAGNETIC SEMICONDUCTORS WITH
ION IMPLANTATION.............................................................................................. 10

Ion Implantation Method Description ........................................................................ 10
Survey of Ion Im planted A IN ..................................................................................... 11
Experimental Characterization and Results of Ion Implantation................................ 12

4 THIN FILM ALUMINUM NITRIDE DILUTE MAGNETIC
SEMICONDUCTOR GROWTH AND OPTIMIZATION OF MAGNETIC
PR O PE R T IE S ............................................................................................................. 22

Growth by Molecular Beam Epitaxy.......................................................................... 22
Growth of Thin Film AlMnN..................................................................................... 24
Crystallinity and Phase Structure ........................................................................ 25
M agnetic Properties............................................................................................. 26
M agnetic M echanism .......................................................................................... 27
Growth of Thin Film AlCrN....................................................................................... 28
Crystallinity and Phase Structure ........................................................................ 29








Role of V/III on Growth and M agnetic Properties........................................... 30
Effect of Dopant on Ferromagnetism ......................................................................... 30

5 THERM AL STABILITY INVESTIGATION ........................................................... 42

6 ALN AND GAN BASED DILUTE MAGNETIC SEMICONDUCTOR
APPLICATIONS........................................................................................................ 50

Spin Filter Applications of A1N DM S ........................................................................ 50
Device Structure and Design............................................................................... 50
Device Testing and Results ................................................................................. 52

7 INVESTIGATION OF GADOLINIUM AS A MAGNETIC DOPANT................... 64

Introduction of a New M agnetic Dopant.................................................................... 64
Growth of GaN:Gd ..................................................................................................... 64
M magnetic Properties of GaN:Gd ................................................................................. 65
Thermal Stability Investigation of GaN:Gd ............................................................... 67

8 CONCLUSION........................................................................................................... 82

AIN-based DM S Survey............................................................................................. 82
Impurity Comparison.................................................................................................. 83
Thermal Instability...................................................................................................... 84
Device Applications.................................................................................................... 84
Summ ary and Future W ork ........................................................................................ 85

APPENDIX

SUPERCONDUCTING QUANTUM INTERFERENCE DEVICE
M AGNETOM ETRY .................................................................................................. 86

Sample M easurement M ethod.................................................................................... 87
Background Subtraction ............................................................................................. 88

LIST OF REFERENCES................................................................................................... 89

BIOGRAPHICAL SKETCH ........................................................................................ 92















TABLE


Table page

4.1. Table featuring the lattice constant and electrical properties of AlMnN and
A lC rN film s.............................................................................................................. 32















LIST OF FIGURES


Figure page

3.1 Powder x-ray diffraction scan of A1N implanted with Mn. No additional phases
are observed as compared to undoped AIN .............................................................. 15

3.2 Powder x-ray diffraction scan of AIN implanted with Cr. Additional phases as
compared to the A1N are shown in the figure .......................................................... 16

3.3 Powder x-ray diffraction scan of AIN implanted with Co. Additional phases as
compared to the A1N are shown in the figure .......................................................... 17

3.4 Magnetization versus applied field trace of Mn+-implanted A1N at a temperature
o f 10 0 K ..................................................................................................................... 18

3.5 Magnetization versus applied field trace of Cr+-implanted A1N at a temperature
of 300 K ..................................................................................................................... 19

3.6 Magnetization versus applied field trace of Co+- implanted AIN at a temperature
of 300K ..................................................................................................................... 20

3.7 Magnetization as a function of temperature for Co+-implanted A1N. A magnetic
field of 500 Oe was applied during the measurement.............................................. 21

4.1 Estimated saturation magnetization and estimated remanent magnetization of
A1MnN. The film was grown with Mn cell temperature of 650C and under a
nitrogen flow of 1.3 sccm ......................................................................................... 33

4.2 Magnetization versus applied magnetic field for AlMnN at temperatures of 10K
and 300K. The film was grown with a Mn cell temperature of 650C and under a
nitrogen flow of 1.3 sccm ......................................................................................... 34

4.3 Magnetization vs. applied field of the undoped AIN showing paramagnetic
beh av ior .................................................................................................................... 35

4.4 Magnetization versus temperature for films of either undoped A1N, single phase
A lM nN or M n4N ..................................................................................................... 36

4.5 Reflection high energy electron diffraction photo of AlCrN film grown at a Cr
cell temperature of 992C.The photo depicts a 2D/3D pattern with one by three
reconstruction ........................................................................................................... 37








4.6 Atomic force microscopy image (1 pm by 1 pm) of AlCrN with an rms roughness
value of approxim ately 11 nm .................................................................................. 38

4.7 Magnetization versus nitrogen flow for AlCrN films grown at a substrate
temperature of 780C and with a Cr cell temperature of 987C .............................. 39

4,8 Magnetization vs. applied field for the optimal AlCrN and AlMnN films..............40

4.9 Magnetization vs. temperature in the temperature range from O10K-50K for
AlMnN and AlCrN. A difference in curvature is seen for the two materials...........41

5.1 Estimated saturation magnetization for AlCrN at each anneal temperature............46

5.2 Magnetization versus applied field measurements comparing the as-grown
AlCrN to the post-anneal AlCrN. Very little magnetization is left after an anneal
of the A lC rN at 700C .............................................................................................. 47

5.3 Magnetization versus temperature taken under an applied field of 250 Oe
comparing as grown AlCrN to post-anneal AlCrN .................................................. 48

5.4 Powder x-ray diffraction scans comparing as grown AlCrN to post-annealed
AlCrN. No second phases are apparent after the 700C anneal............................... 49

6.1 Schematic of the all semiconductor tunneling magneto-resistance stack. The
reference stack contained undoped A1N in place of the A1MnN layer, and all
thicknessed remained the same. The dark squares represent ohmic contacts made
to the top GaN:Si layer and to the underlying MOCVD GaN buffer....................... 55

6.3 Mask design used for fabrication of all semiconductor device. Alignment marks
are found at each comer and in the middle of the mask. The larger dark bars
represent the area of the mesa of the device. The open, or light areas, represent
where ohmic contact was made to the device .......................................................... 56

6.4 Scanning electron micrograph of all semiconductor device. Top view shows top
and bottom Ti/Au ohmic contact, the top of the mesa and the etched valley
showing the M OCVD GaN buffer ........................................................................... 57

6.5 Current-voltage measurement of all semiconductor tunneling magneto-resistance
device with and without and applied magnetic field. Open circles represent the I-
V measurement taken without an applied field. Dark squares represent I-V
measurement taken after application of a 4000 Oe field. Note that tunneling
increases after the field is applied ............................................................................ 58

6.6 Resistance vs. applied field measurement taken at 5K for the all semiconductor
reference device........................................................................................................ 59

6.7 Resistance vs. applied field taken at 300K for the all semiconductor tunneling
m agneto-resistance device........................................................................................ 60








6.8 Dark field ZSTEM image taken of the all semiconductor tunneling magneto-
resistance device. The dark AlMnN layer shows roughness indicating poor
grow th quality at the interface.................................................................................. 61

6.9 Selected area diffraction pattern tunneling electron micrograph taken of the all
semiconductor tunneling magneto-resistance device. The AlMnN spin filter
layer is indicated with an arrow. Strain can be seen within the AIMnN layer as
indicated in the photo ............................................................................................... 62

6.10 Scanning electron micrograph of a tunneling magneto-resistance device with
FeNi as a spin injector. From the photo, the degradation of the FeNi is visible on
the contact pad. This degradation represents nearly 90% of the devices. The
FeNi degraded during photolithography, most likely due to the use of solvents
during processing ..................................................................................................... 63

7.1 Reflection high energy electron diffraction pattern of GaN:Gd, TGd = 1050C.
The picture shows a 2D/3D pattern with 1 x3 reconstruction................................... 70

7.2 Atomic force microscopy image representing GaN:Gd with rms roughness of
1.54 1 nm ................................................................................................................... 7 1

7.3 Magnetization vs. applied field loop taken at 50K for GaN:Gd, with TGd =
950C. Hysteresis is observed at 50K, but not necessarily at 300K or 350K..........72

7.4 Magnetization vs. applied field taken at 50K for GaN:Gd corresponding to TGd =
1000C -1100C ........................................................................................................ 73

7.5 Estimated saturation magnetization vs. inverse of Gd cell temperature plot
w which show s the optim al T aGd ................................................................................... 74

7.6 Magnetization vs. applied field loops taken at 50K, 300K, and 350K for
G aN :G d w ith TGd = 1050C ..................................................................................... 75

7.7 Magnetization vs. temperature for GaN:Gd with Tod = 1050C. Note that the
magnetization is significantly decreased near room temperature ............................ 76

7.8 Estimated saturation magnetization vs. anneal temperature for GaN:Gd with TGd
= 10 50 C .................................................................................................................. 77

7.9 Magnetization vs. applied field comparing the as grown GaN:Gd to that
annealed at 600C ..................................................................................................... 78

7.10 Magnetization vs. temperature of GaN:Gd after 700C anneal ............................... 79

7.11 Magnetization vs. applied field at 50K and 350K for GaN:Gd annealed at
70 0 C ........................................................................................................................ 80

7.12 Magnetization vs. applied field at 300K for GaN:Gd annealed at 700C ................ 81













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

INFLUENCE OF CATION CHOICE ON MAGNETIC BEHAVIOR OF III-N DILUTE
MAGNETIC SEMICONDUCTORS

By

Rachel Marian Frazier

August 2005

Chair: Cammy Abemrnathy
Major Department: Materials Science and Engineering

With the increasing interest in spintronics, many attempts have been made at

incorporating spin-based functionality into existing semiconductor technology. One

approach, utilizing dilute magnetic semiconductors (DMS) formed via introduction of

transition metal ions into III-Nitride hosts, would allow for integration of spin based

phenomena into current wide bandgap device technology.

To accomplish such device structures, it is necessary to achieve single phase

transition metal doped GaN and AIN which exhibit room temperature magnetic behavior.

Ion implantation is an effective survey method for introduction of various transition

metals into AIN. In ion implanted AIN, the Co and Cr doped films showed hysteresis at

300K while the Mn doped material did not. However, it is not a technique which will

allow for the development of advanced spin based devices. Such devices will require

epitaxial methods of the sort currently used for synthesis of III-Nitride optoelectronics.








One such technique, Gas Source Molecular Beam Epitaxy (GSMBE), has been

used to synthesize A1N films doped with Cr and Mn. Room temperature ferromagnetism

has been observed for AlMnN and AlCrN grown by GSMBE. In both cases, the magnetic

signal was found to depend on the flux of the dopant. The magnetization of the AlCrN

was found to be an order of magnitude greater than in the AlMnN. The temperature

dependent magnetic behavior of AlCrN was also superior to AlMnN; however, the

AlCrN was not resistant to thermal degradation.

An all-semiconductor tunneling magnetoresistive device (TMR) was grown with

GaMnN as a spin injector and AlMnN as a spin filter. The resistance of the device should

change with applied magnetic field depending on the magnetization of the injector and

filter. However, due to the impurity bands found in the AlMnN, the resistance was found

to change very little with magnetic field.

To overcome such obstacles as found in the transition metal doped AIN, another

dopant must be used. One viable dopant is Gd, which due to the low concentration

incorporated in the semiconductor matrix should provide a single impurity level within

the DMS instead of an impurity band. The incorporation of Gd into GaN and AIN may be

the ultimate dopant for these III-N based DMS.













CHAPTER 1
INTRODUCTION TO DILUTE MAGNETIC SEMICONDUCTORS

The evolving field of spin transport electronics (spintronics) has drawn great

interest within the past decade. The main goal of spintronics is the utilization of the

electron spin in conjunction with that of electron charge. 1-7 The potential dual role of spin

and charge in spintronics is interesting not only from a fundamental solid state physics

standpoint, but also from a practical device implementation view. Many new device

capabilities and functionalities may be added when the electron spin is manipulated.'

There are numerous driving forces behind the field of spintronics. Some of these

include the advantages of spin over charge. Spin may be easily manipulated by an

external magnetic field, which results in less work required to change spin as compared

to charge3, and there is potential for long coherence time with spin whereas charge is

easily destroyed by scattering effects such as collisions with defects or other charges.2

The use of spin in electronics may lead to devices unattainable by systems based on

charge alone. These devices are proposed to be much smaller than current electrical

devices, use less electricity, and may be powerful enough to compute what charge-based

devices cannot.2 The aforementioned advantages take use of the two channel conduction

associated with spin transport4 which is based on Mott's spin channel model8. Another

possible advantage is the integration of electronic, optoelectronic and magnetoelectronic

functions on a single chip, which would result in more variety of performance than

current microelectronics.2








Spintronics is currently utilized in existing technology. In fact, the 1995

prediction of a huge industry5 became reality in 2001 in the form of a $1 billion per year

industry.2 The technology utilizes magnetoresistance, or resistance which depends on the

relative orientation of spins.6 The giant magnetoresistance (GMR) sandwich structure

makes up a read head and memory storage cell which IBM uses to sense changes in

magnetic fields in current memory technologies .6The GMR sandwich

structure is a stack of ferromagnetic and nonmagnetic metal layers in which the resistance

relies on the alignment of the spins within the magnetic layers. When the magnetic layers

are aligned, the resistance is low and conversely, if the layers are antiparallel, the

resistance is high. GMR effects are also incorporated into sensors and switches which

require low power.

Other memory devices based on spin dependent tunneling are being investigated.

Similar to GMR cells, these use an insulator in place of the nonmagnetic metal which

amplifies the magnetoresistance. Such tunneling magnetoresistive (TMR) devices benefit

from the increase in magnetoresistance, however they rely on high uniformity of

amorphous oxide layers over a large area.

Commercialized devices such as the GMR read head, rely on ferromagnetic

metals for their functionality. Devices which require gain, such as light emitting diodes

and transistors, cannot be made solely of metal, therefore, a major breakthrough in

spintronics will require semiconductor-based devices.9 Much work has been done on spin

transport from a ferromagnetic metal into a semiconductor to try and reproduce the GMR

effect in semiconductors.9-12 Most of these attempts used a semiconductor sandwiched

between two ferromagnetic metal contacts. In all metal devices, magnetoresistance is








measured to be greater than 100%;10 however, semiconductor-based devices achieve only

_1%.12 It has been suggested that the low value of magnetoresistance found in the

semiconductor device arises from the inefficient spin transport across the

metal/semiconductor interface13 which results from a conductivity mismatch at the

interface. To alleviate this mismatch, researchers have turned to dilute magnetic

semiconductors.13-18 Dilute magnetic semiconductors (DMS) differ from ordinary

semiconductors in that they have local magnetic moments at the cation sites of the crystal

structure.14 The advantages of DMS include bandgap tunability, which allows for

operation over a wide range in the electromagnetic spectrum, from far infrared to deep

UV.15 The use of DMSs would also allow for lattice matching. Furthermore, both of these

advantages enable the possibility of spintronics incorporation into compound

semiconductor technology,15 something that is unattainable with the use of all metal

devices.













CHAPTER 2
THEORETICAL AND EXPERIMENTAL BACKGROUND AND INVESTIGATIONS

Experimental Progress in DMS

The first materials studied for use as DMS include chalcogenides and the II-VI

semiconductors. Both the II-VIs and the chalcogenides have very low Curie temperatures

(Tc), below 100K. 19 Device structures incorporating these materials prove the potential

for spintronics, regardless of the low operating temperature (limited by the Tc). One

important example utilizes EuS as a spin filter by relying on tunneling magnetoresistance

(TMR). During operation of this device, the magnetization of a layer of Gd and a layer of

EuS is controlled by the application of an external magnetic field. The EuS acts as a

tunnel barrier with the barrier height being dependent on the applied magnetic field

through spin splitting of the conduction band at temperatures below the Curie point (Tc =

17K). Tunneling is exponentially dependent on barrier height and therefore one spin

channel has a higher probability of tunneling than the other. By reversing the

magnetization within the EuS tunnel barrier, the tunneling current can be controlled,

leading to TMR. Below Tc, the measured TMR is as large as 130%, indicating high spin

efficiency. However, at 30K almost no TMR is observed, demonstrating the inability to

incorporate EuS into current technology due to the extremely low operating

temperatures.20-22

The first III-V semiconductor system studied for DMS material was the InAs

system. It was found that a low growth temperature was necessary for the creation of a

homogenous InMnAs alloy.20'21 A higher growth temperature results in the segregation of








the Mn dopant, or clustering of the Mn.20'23 N-type InMnAs is known to be

paramagnetic2 and p-type InMnAs is ferromagnetic.21'2 However, the Curie

temperature of InMnAs is only 7.5K,21'22 much too low for the incorporation of the

material into devices.

Focus on III-As

The focus turned to the GaAs24-28 system due to the extensive use of GaAs in

microwave and telecommunications technology.24 The GaAs system is more promising

than the InAs system because the Curie temperature is higher. A range of Curie

temperatures have been found for p-type GaMnAs, from 50K to 110K.24-27 The Mn

concentration was varied within GaMnAs and it was found that the increase in Mn

content did not result in the expected increase in magnetization. 6 It has been established

that this is due to self-compensation of the Mn acceptors. The structure of the Mn in the

GaMnAs lattice, from a homogeneous system to a semiconductor matrix with transition

metal precipitates, affects carrier type as well as magnetic properties.26 Most GaAs

studies have been limited to Mn as the transition metal dopant. However, GaAs doped

with Cr has also been studied.28 Preliminary results indicate that GaCrAs is

superparamagnetic.28 The acceptor level of Cr is much deeper than that of Mn in GaAs

(0.89 eV for Cr as compared to 0.11 eV for Mn), which results in low mobility and low

Hall voltages in GaCrAs.28

Extension to III-Ns

Despite the fact that GaMnAs is more promising than InMnAs, the Curie

temperature is still too low for efficient device operation. Therefore, it is necessary to

develop a DMS with a Curie temperature above room temperature. One way is to explore

the potential of other semiconductor host systems.








The predicted Curie temperature for GaMnN is greater than room temperature.29

This prompted interest in GaN based DMS along with the potential spintronic

applications into current GaN technology such as high temperature, high power devices

and visible light emitting devices.30 In contrast to the GaAs system, the GaN and A1N

systems are difficult to dope, so in order to create a DMS, the material must be fabricated

far from equilibrium.31 Given the high melting point (TMp) of GaN, even non-equilibrium

growth results in a higher growth temperature than used for GaN.31 Many Curie

temperatures have been reported for GaMnN. These range from about 25K to greater than

room temperature. 32-35 One group has reported a Curie temperature of 940K, although

this temperature was extrapolated from lower temperature data, and was not physically

measured.3-34

Although Mn is by far the most popular transition metal incorporated into GaN,

another impurity has drawn considerable interest recently. Unlike GaCrAs, which was
28 37-38
found to be superparamagnetic,8 GaCrN is strongly ferromagnetic.37-38 In fact, recent

experimental investigation shows that GaCrN is ferromagnetic to above 900K.38 In

addition to Cr, there has been one report of strong ferromagnetism in GaN:Gd above

room temperature,39 which indicates yet another promising DMS.

Despite the experimental progress in bulk III-N DMS systems, there has been little

experimental progress in device structures. There is evidence that quantum well

structures of GaMnN separated by A1N enhance the magnetic signal as compared to bulk

GaMnN.40 However, other than first principle studies of spin injection at GaCrN/AIN

interfaces, 41 there is no indication of the ability to incorporate the III-N DMS into

spintronic devices.








In parallel to GaN, another system to note is that of AIN. Current applications of

A1N include MIS heterostructures, gas sensors, heterojunction diodes and ultra-violet

(UV) light emitting diodes (LEDs) and lasers. The fabrication of an AIN-based DMS has

the potential of integrating spintronics into the previously mentioned technologies, the

most important being UV applications. In fact, there is current interest in the Al-V DMS

systems. Though AlMnAs was found to be paramagnetic42, the predicted Curie

temperature of AlMnN is greater than room temperature.43 AlCrN has been grown by

reactive sputtering44 and by molecular beam epitaxy.45 In both cases, the Curie

temperature was reported to be greater than room temperature.

Theoretical Background

Although much research into DMS materials has been done,'9-45 there is not one

single theoretical model which accurately fits the ferromagnetic behavior for all systems

studied. Thus, it is necessary to review the main theoretical models and correlate them to

the DMS systems which they apply.

Free Carrier Mediated Model of Ferromagnetism in DMS

In the 1950s, Zener proposed a model of ferromagnetism appropriate for

metals.46'47 This model described ferromagnetism as driven by the exchange interaction

between carriers and localized spins. However, it was later discovered that the Zener

model left out critical criteria for the origin of ferromagnetism in metals.29 The model did

not take into account the itinerant character of magnetic electrons and it also neglected

Friedel oscillations of electron spin polarization around localized spins. In an attempt to

model ferromagnetism in DMS, Dietl48 applied the Zener model to III-V and II-VI

systems, arguing that the Friedel oscillations average to zero because in a DMS, the

distance between the carriers is greater than the distance between the spins. This model is








the equivalent of the Ruderman-Kittel-Kasuya-Yosida (RKKY) interaction model. By

taking into account the anisotropy of carrier-mediated exchange interaction associated

with the spin-orbit coupling in the host material, the Curie temperature (Tc) was found to

be dependent on the hole concentration.29 This is the basis for free-carrier mediated

ferromagnetism. Much research has been done to try to prove that the RKKY interaction

is most likely responsible for ferromagnetism in DMS.48-49

Percolation Picture of Ferromagnetism in DMS

The free carrier mediated ferromagnetic model assumes very high carrier density,

approximately 1020 cm3.49 However, the model does not apply to low carrier systems. An

alternative approach, which applies to both low and high carrier density GaAs systems,

uses the percolation theory to calculate the ferromagnetic-to-paramagnetic transition

temperature.50 In order to explain this theory, a brief review of the polaron percolation

model is necessary. Bound magnetic polarons (BMPs) form from the exchange

interaction between localized carriers and magnetic impurities. BMPs consist of many

magnetic impurities which interact with a localized carrier, say a hole. The direct

exchange interaction between the impurities may be antiferromagnetic (for example,

between Mn); however, the indirect exchange interaction between the BMPs can result in

ferromagnetism.51 The ferromagnetism arises due to the induced field between the

magnetic impurities when the localized hole acts on these impurities. If the localized

holes are parallel and the magnetic impurities are aligned with the holes, then the induced

magnetic field is maximized and there is a ferromagnetic interaction.51 When disorder of

the hole positions is incorporated, there is a resulting enhancement of the ferromagnetic-

to-paramagnetic transition temperature.50 Numerical and analytical simulations suggest








that free carriers are not the only parameter affecting ferromagnetism in GaAs-based

DMS.52

This percolation theory has also been applied to the III-Nitride system. The basis of

the model relies on the difficulty of attaining a high concentration of free carriers in wide

bandgap materials. In this situation, the RKKY interaction is not valid, and

ferromagnetism arises from the indirect interaction between two magnetic ions.43

Essentially, clusters of spins are formed and ultimately coalesce at what is known as the

percolation threshold, which determines the ferromagnetic phase transition.43














CHAPTER 3
SURVEY OF POTENTIAL DILUTE MAGNETIC SEMICONDUCTORS WITH ION
IMPLANTATION

Ion Implantation Method Description

Ion implantation is used in integrated circuit production lines to incorporate

dopants into semiconductors with high uniformity and control. This is achieved by

rastering a beam of ions at fixed energy across the semiconductor surface. Due to the

surface bombardment at high energy, these ions are able to penetrate the semiconductor

and incorporate into the host lattice. The projected range of the ions depends nearly

linearly on the implantation energy. The dose of implanted dopant is determined by the

flux of incident ions and is given by Q = Ft where Q is the dose, F is the flux of incident

ions/cm2s and t is the implantation time.

During ion implantations, the semiconductor host lattice is damaged by the

energetic ions. The host lattice atoms become displaced and crystal structure becomes

disordered. For this reason, and to allow substitutional incorporation of the dopant ion,

implanted semiconductors are thermally annealed at elevated temperatures. This

generates sufficient energy to restore most of the crystal structure and allow dopant ions

to move to substitutional sites within the host lattice. Although the post-implant anneal

helps to recover lattice order, many implant-induced defects remain within the

semiconductor lattice.

The advantages of ion implantation allow for applications in dilute magnetic

semiconductors. Most prevalent is the use of ion implantation as a survey method in the








determination of choice of dopant and host semiconductor. Ion implantation provides

useful insight into the prospect of specific choice of semiconductor lattice and dopant in

terms of magnetic efficiency. It is a cost effective and quick method to incorporate a

predetermined amount of dopant and examine the plausibility of a given DMS. Implanted

samples may be easily characterized by magnetic techniques to verify ferromagnetic

behavior of a DMS. For this reason, ion implantation into A1N is addressed.

Survey of Ion Implanted AIN

Epitaxial layers of 1 pm A1N were grown by metal organic chemical vapor

deposition (MOCVD) on sapphire (A1203) substrates. Three dopant ions (Cr+, Co+, and

Mn+) were implanted into the MOCVD A1N. The implantation energy was 250 keV

which corresponds to a projected range of 150 nm for each dopant ion. The dosage was

set to 3E16 dopant ions per square cm (dopants/cm2) for each case, corresponding to peak

concentrations of dopant ions of approximately 3 atomic percent (at %) in the AIN.

During the implantation, the substrates were heated and held around 300C to promote

dynamic annealing. To further reduce lattice disorder and encourage dopant incorporation

onto substitutional Al sites, post-implant annealing was performed at 950C for two

minutes in a Heatpulse 610T rapid thermal anneal (RTA) system under a nitrogen

ambient.

The implanted A1N samples were characterized to determine the phase structure,

magnetization and defect inclusion arising from the implantation for each dopant ion. The

phase of each material was determined via powder x-ray diffraction (XRD), the

magnetization via superconducting quantum interference device (SQUID) magnetometry,

and the defect inclusion via photoluminescence (PL). The results of the material








characterization determine the effectiveness of AIN as a DMS, and give a rapid

comparison of potential dopants. These results are discussed in the next section.

Experimental Characterization and Results of Ion Implantation

Photoluminescence (PL) spectra were taken at a temperature of 10K of implanted

A1N samples after thermal anneal. Comparing the undoped AIN to Cr+, Co+, and Mn+

doped A1N gives insight into the defects arising from implantation. The undoped AIN

shows strong band-edge emission at 6.05 eV and two broad emission bands at 3.0 and 4.4

eV. The two emissions not corresponding to the band gap of the material are most likely

related with deep level impurities within the band gap of the A1N. The implanted A1N all

show an absence of band edge emission, which suggests that the point defect

recombination centers created during the implantation are stable against the thermal

anneal. Each of the implanted AIN shows a band at 5.889 eV which represents lattice

disorder induced by the implantation, since it is independent of the Cr+, Co+, and Mn+

species introduced into the host lattice. Essentially, there is little difference between the

Cr+, Co+, and Mn+ implanted AIN, and expected dissimilarities between the undoped

and implanted AIN, which correspond to recombination centers and lattice disorder

arising from the implantation.

Powder x-ray diffraction (XRD) scans (Figures 3.1-3.3) of the ion implanted AIN

reflect the phase structure of the material, with single phase material being the optimal

situation. The XRD scan for the undoped AIN provides a useful reference to eliminate

any peaks observed in the implanted films as host lattice contributions. The Mn+

implanted A1N shows no additional peaks as compared to the undoped AIN (Figure 3.1).

This suggests that this material is single phase. Both the Co+ and Cr+ implanted AIN

show additional peaks compared to the reference scan. These peaks correspond to AlxCry








and AlxCoy phases for the Cr+ implanted and Co+ implanted A1N, respectively. It is

important to note that neither AlCry nor AlxCoy phases are ferromagnetic, which

suggests that the additional second phases contributed no ferromagnetic signal to the

magnetic study.

Superconducting quantum interference device (SQUID) magnetometry data show

magnetic properties of films. Two important traces, the magnetization dependence on

applied field (M vs. H) and the magnetization dependence on temperature (M vs. T), are

most frequently used to demonstrate ferromagnetism (FM) and the temperature range

across which FM exists, respectively. Hysteresis observed in M vs. H loops demonstrates

material with FM properties. Persistent FM through all temperatures indicates the

material's retention of FM, and implies that the Curie point (Tc) has not been reached. In

all cases of magnetic dopants, the implanted A1N show hysteresis. The Mn+ implanted

A1N M vs. H trace at 100K shows FM (Figure 3.4). The 10K M vs. H also shows clear

hysteresis, however, no clear hysteresis was observed at 300K. This suggests that the FM

is severely diminished or nonexistent at room temperature. However, both Cr+ and Co+

implanted A1N demonstrate FM at 300K (Figure 3.5). For each implanted ion, clear and

well-defined hysteresis is observed in the M vs. H. Also, the M vs. T trace for Cr+ and

Co+ (Figure 3.7) further suggests FM to room temperature by the persistence of magnetic

signal to 300K. The contrasting magnetic behavior at 300K of the Mn+ implanted A1N

and the Cr+ and Co+ implanted AIN suggests that Mn+ is the least optimal choice of

dopant in terms of magnetization for an AIN-based DMS. The optimal dopant would

exhibit FM at room temperature, thereby allowing for incorporation of the DMS into

current technology.








Thus, the implantation of Cr+, Co+, and Mn+ into A1N provides useful insight into

the prospects of AIN-based DMS. Each of Cr+, Co+ and Mn+ implanted AIN

demonstrates an absence of band-edge emission in PL spectra. Mn+ implanted AIN

shows single phase material, whereas Cr+ and Co+ implanted AIN show multiphase

material. However, the Cr+ and Co+ implanted materials exhibit FM at room

temperature, but the Mn+ implanted A1N does not. It is not expected that the additional

phases found in the Cr+ and Co+ implanted A1N give rise to the FM at 300K since the

secondary phases are not FM. In general, the use of ion implantation to introduce

magnetic dopants into A1N suggest that Cr+ and Co+ are more magnetically active than

Mn+, however, Mn+ is more readily incorporated to produce singe phase material.

Hence, further investigation is necessary to confirm the initial survey.















Mn-implanted AIN









A. l


20 40 60
28 (degrees)


80 100


Figure 3.1. Powder x-ray diffraction scan of AIN implanted with Mn. No additional
phases are observed as compared to undoped AIN.


800


600


400














I1 U -- ,- - -- --


Cr-implanted AIN


20 40 60
2e (degrees)


80 100


Figure 3.2. Powder x-ray diffraction scan of AIN implanted with Cr. Additional phases as
compared to the A1N are shown in the figure.


800



600


U -- 1 -- -- 1 -- -- 1 -- -- 1 -- --


















Co-implanted AIN







I i


20 40 60
2e (degrees)


80 100


Figure 3.3. Powder x-ray diffraction scan of AIN implanted with Co. Additional phases
as compared to the A1N are shown in the figure.


800


600


400






18





,X,0-5.
1x10`5
Mn-implanted AIN

E 5x10O- T=100K


"-- i li"8'""
ai

"5x10 n
0)
CO GE
N


-lx10-
I I I I I I I I
-1000-750 -500 -250 0 250 500 750 1000
H(Gauss)


Figure 3.4. Magnetization versus applied field trace of Mn+-implanted AIN at a
temperature of 100K.













1.0x10-5
:3
E 5.0x10'6



*IU~
N
S-5.0x10
0)
(a
-1.0x105


Cr'-implanted AIN




I I I I I I I I
T= 300K ,!

=!

ling
so


0i i
L ----U ,----
I I I I I I*


-1000-750 -500 -250 0 250 500 750 1000
H(Gauss)


Figure 3.5. Magnetization versus applied field trace of Cr+-implanted A1N at a
temperature of 300K.













1.0XI0"5 -II
Co*-implanted AIN 1
T3 T 300K
E 5.0x106 T0
0.)
C m
4 0.0
N T .
-5.0x10"- U:


-1.0x105 -

-1000-750 -500 -250 0 250 500 750 1000
H(Gauss)


Figure 3.6. Magnetization versus applied field trace of Co+- implanted AIN at a
temperature of 300K.











-1.0xl .5 Co0-implanted AIN

-1.0x 10.5
5 m U Field-cooled
-1.1x10 Zero field-cooled
=3 1.1x10'5-
E 11x H = 500 0e
'-1.2xl 10 .5T

-1.2x10-5

-1.3x105
0 50 100 150 200 250 300
Temperature(K)

Figure 3.7. Magnetization as a function of temperature for Co+-implanted AIN. A
magnetic field of 500 Oe was applied during the measurement.













CHAPTER 4
THIN FILM ALUMINUM NITRIDE DILUTE MAGNETIC SEMICONDUCTOR
GROWTH AND OPTIMIZATION OF MAGNETIC PROPERTIES

Growth by Molecular Beam Epitaxy

In order to produce single crystal material with minimum defects, an epitaxial (epi)

growth method is required. For this reason, the main growth method used in the

fabrication of thin films is molecular beam epitaxy (MBE). During MBE growth,

molecular beam constituents react with a heated crystalline substrate under ultra-high

vacuum (UHV) conditions. The molecular beams are produced when solid, high purity

metals are heated within a Knudsen cell (k-cell). In the growth of the epi III-Vs discussed

in this paper, the group III and magnetic impurities were metal sources with 7N purity.

The group V source was produced by inducing a plasma from a 7N purity nitrogen gas

source, so the specific growth method is dubbed gas source molecular beam epitaxy

(GSMBE). MBE, hence GSMBE, provides advantages over other epitaxy methods due to

the ability to precisely control growth conditions such as substrate temperature, and

group III and group V atoms impinging on the substrate surface. This allows for 1) a low

growth rate, 2) the ability to abruptly initiate and/or terminate the growth, and 3) sharp

interfaces and smooth surfaces, all of which provide the possibility of producing devices

such as quantum wells, dots, and other structures relying on control of interlayers and

interfaces. These MBE features also provide opportunity to monitor and characterize the

growth surface in situ, including techniques such as reflection high energy electron

diffraction, surface probe microscopy, and x-ray photoelectron spectroscopy.








The initial growth surface plays a role in the quality of the epitaxially grown

overlayer. Hence, the choice of substrate impacts the MBE-grown film. For III-Ns (such

as GaN, A1N, and AlGaN) substrates include sapphire (A1203), metal-organic chemical

vapor deposition (MOCVD) GaN buffers on A1203, and silicon (Si) and silicon carbide

(SiC). Each of these substrates has a different lattice match to the III-Ns. Impact of

substrate choice on material properties such as magnetization will be presented later,

however the preparation of each substrate prior to growth is slightly different. MOCVD

GaN buffers are chemically treated prior to growth with a 1:1 HCI:DI H20O 3 minute dip, a

25 minute ultraviolet ozone (UVO3) exposure, and a 5 minute buffered oxide etch (BOE)

dip. Each MOCVD GaN buffer is treated with this recipe prior to mounting to remove the

native oxide on the GaN surface. Both Si and SiC receive a BOE dip to remove the native

oxide, however the A1203 substrate does not require a chemical treat prior to mounting.

After any necessary chemical preparation, the substrates are mounted to Mo blocks with a

thin In layer to provide thermal contact. The substrates are then loaded into the MBE

system. The in situ treatment of the substrates differs due to the different lattice

mismatch, and constituents of the substrate. A1203 substrates require a nucleation layer of

the III-N to initialize a decent growth surface, where as MOCVD GaN buffers do not,

however both require an exposure to the N species prior to growth. Si and SiC substrates

must not be exposed to N species prior to growth, or a silicide will form at the substrate

surface and provide a poor growth surface.

During growth, the molecular beam species impinge on the substrate and undergo

absorption and migration. Growth occurs when the impinging group III and group V

atoms collide and stick to the surface, defined by the sticking coefficient which describes








how effectively the atoms stick to the surface. The three main growth conditions that

affect the rate of growth are the flux of the group III species, the ratio of group V to

group III flux of atoms (V/Ill) and the substrate temperature (Ts). The effect of these

growth conditions on the magnetic properties of the material under investigation are

presented later in the chapter. Three distinct situations are possible during growth, which

are affected by the growth conditions (V/Ill and Ts). These growth modes are Frank-van

der Merwe (layer by layer), Stranski-Krastanow (layer plus island) and Volmer-Weber

(island). The growth mode of the film is monitored in situ by reflection high energy

electron diffraction (RHEED), and represents the previously-mentioned growth modes by

two dimensional (2D = line), combination of two dimensional and three dimensional

(2D/3D = line and spots), and three dimensional (3D = spots) patterns, respectively.

The specific bulk epitaxial films grown and studied include A1N and GaN. The

magnetic impurities introduced during growth include Mn, Cr, and Gd. The rest of the

chapter serves to describe the different materials and the structural and magnetic

properties arising due to the different growth conditions.

Growth of Thin Film AlMnN

Growth of the films presented occurred in a Varian Gen II by gas-source molecular

beam epitaxy. Solid AI(7N) and Mn(7N) sources were heated in standard effusion cells.

Gaseous nitrogen was supplied by an Oxford rf plasma head. All films were grown on

(0001) oriented sapphire substrates, indium mounted to Mo blocks. AlMnN and A1N

films were grown at a temperature of 780C, as indicated by the substrate heater

thermocouple. A nitrogen flow of 1.3 standard cubic centimeters per minute (sccm)

corresponded to a chamber pressure of 2.3x 10-5 Torr during the growth of the films.








Sapphire substrates were first nitridated for 30 minutes at a substrate temperature of

1000C under 1.1 seem nitrogen (chamber pressure = 1.9x10-5 Torr). Nucleation at 575C

for 10 minutes and a 30 minute buffer layer at 950 C followed nitridation, both under

1.1 sccm nitrogen. Both A1N and AlMnN films were grown with a substrate temperature

of 780C and an Al effusion cell temperature of 1150C. The Mn cell temperature (TMn)

was varied from 635C to 658C. The growth rate of the AIN was 0.2g.m/hr and the

growth rate of the AlMnN films was 0.16pm/hr.

In situ reflection high energy electron diffraction (RHEED) was used to monitor

films during growth. AIN demonstrated 2D growth and AlMnN films demonstrated

2D/3D growth. Determination of the phase composition of the layers was carried out by

x-ray diffraction in a Phillips APD powder diffractometer. AlMnN grown with a Mn cell

temperature of 635C was found to be single phase. The AlMnN with TMn=658C formed

AIMn as detected by powder x-ray diffraction (XRD). A Mn cell temperature of 650C

was found to be the upper limit of single phase AlMnN under previously mentioned

growth conditions. For comparison, a layer of Mn4N was also grown on sapphire. Auger

Electron Spectroscopy (AES) showed Mn to be present in all of the Mn films, however

accurate determination of the Mn concentration in AlMnN was hindered by the small Mn

signals. It is estimated that AlMnN films contain no more than 1% Mn.

Crystallinity and Phase Structure

Crystalline quality of the films was inspected using high resolution x-ray

diffraction (HRXRD) in a Philips X'pert diffractometer equipped with a Cu Ka source.

Rocking curves were performed on films in order to determine the FWHM and lattice

constants of the films. Lattice constants were calculated using Bragg's law from data








obtained by the HRXRD investigation. Table 1 shows results of those calculations for

A1N and AlMnN grown using two different Mn cell temperatures. The lattice constant

was found to decrease as the Mn cell temperature increased for single phase material. A

similar pattern was observed for single phase GaMnN films grown in the same system

under different conditions. GaN implanted with Mn has been reported to exhibit

substitutional or near substitutional incorporation. It is expected that the incorporation of

interstitial Mn should either increase or have no effect on the lattice constant. The

observation of a decrease in the lattice constant of the AlMnN films suggests that the Mn

occupies a substitutional site. This is further confirmed by Hall analysis, also given in

Table I, which shows pure A1N to be highly resistive as expected and material containing

an AlMn second phase to be highly conductive n-type. By contrast, single phase AlMnN

was found to be p-type. If Mn incorporates substitutionally, one would expect by analogy

with its behavior in other III-V materials that it would behave as a deep acceptor. The

observation of p-type behavior fits this explanation.

Magnetic Properties

Magnetic measurements were performed on samples using a Quantum Design

MPMS SQUID magnetometer. Magnetization as a function of magnetic field and

temperature was measured with the applied field parallel to all films. Magnetic

remanence and coercivity indicating hysteresis was observed in ternary AlMnN films at

10, 100, and 300K. As a rough estimate, saturation magnetization was extracted from the

hysteresis loops as the magnetization at 1000 Gauss (G). This estimated saturation

magnetization was found to decrease at 300K compared to 100K for AlMnN grown at

TMn=650C. Also, the remanent magnetization was extracted from the hysteresis loops








and was found to decrease with temperature. The values of temperature dependent

estimated saturation magnetization and remanent magnetization of AlMnN are shown in

Figure 4.1. A comparison of the M vs. H loops at 10K and 300K of the AlMnN shows a

large decrease in observed hysteresis, as shown in Figure 4.2. This decrease in the overall

magnetization as seen in the M vs. H loops possibly indicates that the Curie temperature

of the AlMnN is just beyond 300K. However, clear hysteresis is evident at room

temperature. The coercive field (approximately 70 Gauss) was found to be independent

of temperature. Undoped A1N grown under the same conditions as AlMnN demonstrated

paramagnetic behavior, as shown in Figure 4.3. This indicates that hysteresis arises with

the addition of Mn. The diamagnetic background due to the sapphire substrate was

subtracted from the raw data and the subsequent corrected data was used for analysis.

The magnetization was not normalized to the Mn concentration due to the inability of

AES to precisely detect the small amount of Mn in the films.

Magnetic Mechanism

Clusters of second phases, undetectable by methods mentioned above, are not

thought to be the cause of hysteresis observed at 300K. This is supported by magnetic

analysis of material containing the most likely cluster phases, AIMn and Mn4N.

Magnetization as a function of temperature for A1N, single phase AlMnN, AlMnN with

an AIMn phase present, and Mn4N show substantially different behavior, as shown in

Fig.4.4. The reason for the low T(10-50K) behavior seen in A1N and AlMnN films is

still unknown. The M vs. T of Mn4N clearly indicates ferromagnetic behavior, and the

formation of clusters has been proposed as the cause of hysteresis in some ferromagnetic

III-V materials. However, the formation of Mn4N clusters does not influence the

magnetization above 50K, since clearly the magnetization drops to zero at that








temperature. Also, the magnetization vs. temperature indicates that the formation of

AlMn clusters is not the cause of the ferromagnetism observed, evidenced by the order of

magnitude difference between the values of magnetization over 150K. Hence, the

incorporation of Mn into the A1N lattice forming the ternary AlMnN is most likely the

reason for the observed hysteresis.

In conclusion, room temperature ferromagnetism has been observed in AlMnN

grown by gas-source MBE. The lattice constant decreased with increasing Mn cell

temperature for single phase material, indicating constant site occupation, probably

substitutional. Hysteresis in M vs. H at room temperature was observed in single phase

material and the magnetization as a function of temperature suggests ferromagnetism

caused by AlMnN, not clusters.

Growth of Thin Film AICrN

AlCrN films were grown by Molecular Beam Epitaxy (MBE) on c-plane sapphire

in a similar fashion to the AlMnN films discussed in the previous section. The same Al

and nitrogen sources were used, and the same nitridation, nucleation, and high

temperature buffer layers were grown prior to DMS growth. Solid Cr (7N) was used as

the dopant source in the AlCrN films by heating in a standard effusion cell. AlCrN films

were grown at a substrate thermocouple temperature reading (Ts) of 780C. The AlCrN

films were grown with a Cr cell temperature thermocouple reading (Tcr) from 982-

1005C. The Cr concentrations were in the range 1-3 at. %, as measured by x-ray

microprobe and secondary ion mass spectrometry.

Reflection high energy electron diffraction (RHEED) was observed in situ and used

to monitor the film growth. The AlCrN films for the most part demonstrated 3D growth,

as evidenced by spotty RHEED patterns. However, the AlCrN grown at TCr = 982C








demonstrated 2D growth with a pattern indicative of a 1x3 reconstruction (Figure 4.5),

which was not observed for any other AlCrN films. Atomic force microscopy (AFM), a

widely-used tool for investigating the topology of materials, was used to investigate the

surface morphology of the AlCrN films. As expected from the RHEED patterns, the

AlCrN films were considerably rougher than undoped A1N grown under similar

conditions (Figure 4.6).

Crystallinity and Phase Structure

Single phase and multi-phase AICrN films were investigated. Powder x-ray

diffraction was used to determine the phase composition of the films. The materials

corresponding to Tcr = 982-992C were found to be single phase while material grown at

Tcr = 1005C was found to contain multiple phases. Both Cr2N and AlCry (most likely

corresponding to A13Cr2 and AlCr3) were found in the higher Cr content film. High

resolution x-ray diffraction was used to determine the lattice constant of the AlCrN films.

As shown in Table 4.1, the lattice constant (ao) of the single phase AlCrN was smaller

than for A1N. This behavior was also observed in the similarly grown AlMnN films, as

reported earlier in the chapter. The presence of second phases increased a0 relative to the

single phase films. A decrease in lattice constant with the introduction of a dopant

indicates substitutional incorporation of the dopant. The rise in lattice constant with the

presence of multiple phases suggests that more of the dopant is incorporating

interstitially. The single phase AlCrN was found to be semi-insulating, while the multi-

phase material was found to conduct via an electron hopping mechanism with activation

energy for conduction of 0.19 eV. For non-optimized growth, second phases of Cr2N and

AlxCry are produced in the AIN and the material is conducting (-1000 Ohm-cm), see

Table 4.1.








Role of V/Ill on Growth and Magnetic Properties

The nitrogen flux was varied from 1.2-1.5 sccm in the growth of the AlCrN films to

determine optimal V/III ratio. The significance of the V/III ratio lies in the relationship

between incorporation of the transition metal ion and the resulting magnetization.

Without a change in group III species or growth temperature, a change in group V species

will affect the number of substitutional sites for dopants to occupy. At high V/III,

however, the excess N at the surface may interfere with the transition metal

incorporation, resulting in a reduction of the magnetic ordering. This results in an optimal

V/III which will incorporate the maximum number of substitutional dopants. It follows

that this will lead to the maximum magnetic signal arising from the substitutional

dopants. Figure 4.7 shows the variation of the 300K saturation magnetization with respect

to the flow rate of the nitrogen for the AlCrN films grown at Tcr = 987C. High N2 flows

appear to reduce the sticking coefficient of the Cr, resulting in a reduced Cr concentration

and hence lower Ms.

Effect of Dopant on Ferromagnetism

The ferromagnetic signal was found to be much stronger in the optimized AlCrN

film than in the optimized AlMnN film. Figure 4.8 compares the magnetization vs.

applied field for the optimal AlCrN and AlMnN films. The signal strength of the

optimized AlCrN is -2.5 times that of the optimized AlMnN. This suggests that the

ferromagnetic interaction in the A1N between the Cr ions is stronger than that of the Mn

ions. The coercive fields of the AlCrN and AlMnN are nearly identical. This suggests that

the domain structures of the materials are similar.

Strong ferromagnetism persists to 350K (the temperature limit of the SQUID

magnetometer), in the optimized AlCrN. Moreover, the saturation magnetization remains









unchanged in the temperature region from 10K-350K. By contrast, A1MnN films grown

under similar conditions were found to decrease in magnetization at 300K as compared to

100K, as shown in Figure 4.1. This suggests that AlCrN has a Curie point above that of

AlMnN. Moreover, the Curie temperature of AlCrN is well above room temperature

whereas the Curie temperature of AlMnN is in the vicinity of room temperature. Another

difference between the AlMnN and the AlCrN is in the shape of the magnetization vs.

temperature trace. Interestingly, the paramagneticc tail" seen in the AlMnN in the 10-50K

region is less pronounced in the single phase AlCrN, as shown in Figure 4.9. This may be

another indication of a stronger ferromagnetic interaction between Cr dopant ions and Mn

dopant ions. Overall, Cr tends to be a better dopant than Mn in ferromagnetic A1N.









Table 4.1. Table featuring the lattice constant and electrical properties of AlMnN and
AlCrN films._________ _____
Lattice Constant FWHM Carrier Type/ Resistivity
(Ang) Concentration (Qcm)
________________________(cm-3)____
A1N 4.89592 0.0919 insulating
AlMnN 4.89292 0.0928 -
(TMN=635
C)_____
AlMnN 4.89169 0.0902 p-type/-1E18 1.6
(TMN=650
C)_____
AlMnN/Al 4.89326 0.0984 n-type/-2E20 5.7E-3
Mn
AlCrN

A1CrN/AlxC 1000
ry+Cr2N













0.60
S0.55 ---M AIMnN Tn=650

? *M TM AIMnN T =650
0.50 Mn

S0.45

0.40
C
0.35
Ca
.N 0.30
(D
c 0.25
CD
S0.20

0.15
S0 0
CU
E 0.10

lU 0.05 o
0 50 100 150 200 250 300
Temperature (K)


Figure 4.1. Estimated saturation magnetization and estimated remanent magnetization of
AlMnN. The film was grown with Mn cell temperature of 650C and under a
nitrogen flow of 1.3 sccm.












0.8 1 1 i*

0.6 A AIMnN,T=10K
o AIMnN,T=300K
0.4


0.2E -
0 o.o --------_ _--
S0.0
E
-0.2 4,4S^ 8'

-0.4

-0.6

-0.8 I -- I
-1000 -500 0 500 1000
H (Oe)


Figure 4.2. Magnetization versus applied magnetic field for AIMnN at temperatures of
10K and 300K. The film was grown with a Mn cell temperature of 650C and
under a nitrogen flow of 1.3 sccm.












S2.0x10
E
_ 1.0x10
E
a)
" 0.0
0
CO
.N -1.0x10

0)
CZ -2.0x10


i -
* UndopedAIN I iiB

U U
I
U

Eul




I I I I


-1000 -500 0 500 1000
H(Gauss)


Figure 4.3. Magnetization vs. applied field of the undoped AIN showing paramagnetic
behavior.












-0.4

-0.6

-0.8

-1.0

-1.2

-1.4

-1.6


0 50 100


150 200 250 300


Temperature (K)



Figure 4.4. Magnetization versus temperature for films of either undoped AIN, single
phase AlMnN, or Mn4N.


*\A
I I I I I *



0-nn MEn amai n n nnnnmmm mnmn.0nn an m nnm Rnm


-s A IM n N
SAIN, undoped
Mn4 N

A H=500 0e




I I I I




























Figure 4.5. Reflection high energy electron diffraction photo of AlCrN film grown at a Cr
cell temperature of 992C.The photo depicts a 2D/3D pattern with one by
three reconstruction.









1.00





-0.75




__ _-0.50






0.25






0 0.25 0.50 0.75 1.00 JM
Figure 4.6. Atomic force microscopy image (1pm by 1pm) of AlCrN with an rms
roughness value of approximately 11 nm.












3.5 I -I

3.0 AICrN, TC = 987C

2.5 T = 300K

S2.0
E

1.5
E
U) 1.0

0.5


I I I I
0.50 ^ -


1.2 1.3 1.4 1.5
N2 Flow (sccm)


Figure 4.7. Magnetization versus nitrogen flow for AlCrN films grown at a substrate
temperature of 780C and with a Cr cell temperature of 987C.












A Optimal AlCrN 4*
* Optimal AIMnN
T=300K
4,4p
T^T
i#.,1i i i *nn0ii0
nii.-- ** -^
A'

: .... **- **


-1000


1000


H (0e)


Figure 4.8. Magnetization vs. applied field for the optimal AlCrN and AlMnN films.






41





0.85


0.80 o-- '-- --*--*--- o---o---i-n


0.75
E
P -0- Field-cooled AICrN
= 0.70 Z-o- Zero Field-cooled AlCrN
E
.. A. Field-cooled AIMnN
S-0.70 A Zero field-cooled AIMnN
"N-0.75
S-075 H = 250 Oe
-0.80 4 .
S -0.85 .... .. A .. .........
-0 .90 4 ..... .. ..... .... ....... ........... 4 ...
-0.95 -
I I I n *
0 20 40 60 80 100
Temperature (K)


Figure 4.9. Magnetization vs. temperature in the temperature range from IOK-50K for
AlMnN and AlCrN. A difference in curvature is seen for the two materials.













CHAPTER 5
THERMAL STABILITY INVESTIGATION

The prospects for integrating a dilute magnetic semiconductor (DMS) into current

technology depend upon the ability for the material to withstand certain chemical,

thermal, and mechanical processes that occur during the processing of the thin film

material into devices. Typically, III-N semiconductors are exposed to chemicals, in the

form of solvents such as methanol and isopropyl alcohol, during the fabrication of

devices including high electron mobility transistors (HEMTs) and light emitting diodes

(LEDs). Standard procedures utilizing solvents are performed to mask the underlying

material in order to provide a particular pattern to either remove the semiconductor or to

deposit metal contacts. The strong bonding found in the III-Ns made these DMS resistant

to such chemical exposure. However, some of the metal contacts to the III-Ns require a

heat-activiating intermixing of various metal layers. This thermal anneal procedure is

necessary to produce optimal electrical contact for III-N ohmics. Thermal anneal

temperatures may easily exceed 700C, or the low growth temperature required for III-N

DMS growth. For this reason, it is necessary to investigate the thermal stability of DMS,

to be certain that the material will withstand the elevated temperatures encountered

during processing.

In order to effectively determine the material's thermal stability with respect to

magnetic ordering, a systematic approach is applied. The magnetization of thin film epi-

A1N DMS, is probed after a rapid thermal anneal at temperatures increasing in 100C

increments. The magnetization is extracted from SQUID measurements taken at 300K








after each rapid thermal anneal (RTA) until no hysteresis is observed. The RTA

temperature at which the magnetization is lost is taken to be the upper thermal limit of the

material. In this manner, the thermal stability of III-N DMS may be systematically

studied.

The optimal epi-AICrN used in the investigation of thermal stability was grown at a

substrate temperature (Ts) equal to 780C with an Al cell temperature (TAO) of 1150C

and a Cr cell temperature (Tcr) of 987C. The material was grown under a nitrogen

plasma with a flow of 1.3 sccm. The AlCrN was annealed from 300C-700C in 100C

increments. Figure 5.1 shows the estimated saturation magnetization for each of the

anneal temperatures. The estimated saturation magnetization was taken to be the

magnetization at an applied field of 1000 Oe. From the as-grown condition (no anneal) to

the initial anneal at 300C, the magnetization decreases from 3.12 emu/cm3 to 1.11

emu/cm3. This drastic drop by almost a third indicates that the magnetic ordering is being

affected by even a low temperature anneal at 300C. The magnetization then decreases at

the subsequent anneal at 400C, nearly an order of magnitude below the as-grown value,

but remains stable until Tanneal = 700C. This indicates that the magnetic ordering is

unaffected in the temperature range from 400C up to nearly 700C. However, at Tanneal =

700C nearly all magnetic ordering is lost, as shown in Figure 5.2. The raw

magnetization vs. applied field (M vs. H) after annealing at 700C shows evidence of

some hysteresis remaining, however the data is noisy and the loop is nearly flat implying

that most of the magnetic ordering is destroyed by the 700C anneal. The magnetization

vs. temperature (M vs. T) before and after the 700C anneal (Figure 5.3) corroborates the

observation that the magnetization is lost as compared to the as-grown AlCrN. The noisy,








yet consistently near zero, field-cooled signal is well below the signal obtained in the as-

grown case.

The upper limit of thermal resistance to magnetic degradation of the epi-AlCrN

appears to be very near, yet lower than, the growth temperature of the material. The

apparent discrepancy between the temperature at which the material's magnetic

properties degrade and that of the magnetic properties of the as-grown AlCrN is

explained as follows. During growth, active nitrogen species are delivered to the surface

which is at elevated temperatures. This exposure to the active nitrogen species continues

after growth during the cool down of the material. This is routine procedure which

prevents the disassociation of nitrogen from the material while the material is still heated,

ultimately resulting in no change of material during cool-down. However, during the

RTA, the ambient gas is N2, not the atomic nitrogen found in the plasma. It is plausible

that during the anneal the quality of the AlCrN degraded due to the elevation of

temperature without an atomic nitrogen overpressure, which lead to the loss of nitrogen

and degradation of magnetic ordering. The extent of degradation could range from slight

deterioration in crystallinity of the AlCrN to destruction creation of second phases.

The investigation of the AlCrN material phase after the 700C anneal would give a

clear indication of phase degradation. Figure 5.4 shows the powder x-ray diffraction

(XRD) scans of AlCrN before and after the anneal at 700C. No second phases are

observed after the 700C anneal. The main difference between the XRD scans is the

broad peak around 20 degrees, which corresponds to the substrate, and gives no insight

into the epi layer. The absence of second phases implies that the extent of material

degradation from the anneal is not sufficient enough to change the phase of the material.








However, as seen in the SQUID data, the effect of a 700C anneal leads to a degradation

of magnetic properties.

The thermal stability of optimal epi-AlCrN was investigated and was found to be

poorer than that of GaCrN. The upper limit of anneal temperature before destruction of

magnetic properties was found to be 700C. No indication of second phases was

observed after the 700C anneal, which implies that the RTA was not significant enough

to destroy the single phase AlCrN. However, although the single phase AlCrN withstood

the RTA, the magnetic interaction within the material did not. This suggests that the

mechanism for magnetization within this particular DMS is too weak to endure routine

processing procedures during fabrication, such as the annealing of ohmic contacts.

Hence, processing procedures must be altered for incorporation into current technology

or alternate materials will need to be used in place of AlCrN.














2.5


2.0


1.5

I?
E 1.0

E
- 0.5



0.0


-0.5


0 -#- Saturation Magnetization
AlCrN, TC, = 9870C


^S



*-




*--*0


I I I I I I I


-100 0 100 200 300 400 500 600 700 800

Anneal Temp (C)



Figure 5.1. Estimated saturation magnetization for AlCrN at each anneal temperature.


. I













4.0x10




2.0x10


-2.0x10




-4.0x10


AICrN, Tcr = 987C as grown
A AICrN, Tc =9870C Tanneal =7000C I see

U.||

Em
me

Eu
.... AA :AAAAAAAA*4*


Em
EU
EU
mli
iI *
I-- A I I I4' ^*^ '^


-1000


-500


1000


H (Gauss)


Figure 5.2. Magnetization versus applied field measurements comparing the as-grown
AlCrN to the post-anneal AlCrN. Very little magnetization is left after an
anneal of the AlCrN at 700C.






48





I I


0.8 -mmm mmmmm m m m


E U AlCrN as grown

SzA AlCrN, post 700C anneal
E
CD 0.7
C:
.o Field-cooled at H = 250 Oe
Ca
N
CT
D 0.0 6AA^A j




-0 .1 I I I I
0 50 100 150 200 250 300
Temperature (K)


Figure 5.3. Magnetization versus temperature taken under an applied field of 250 Oe
comparing as grown AlCrN to post-anneal AlCrN.











800 AICrN, T, = 987C '

700 -- as grown
............... annealed,
600 AIN (002) 7000C

500 AIN (004)

400 A 3 (006) 012
C,-
20010 i I (0012)
0 300 I""
200 substrate


100

0 1
20 40 60 80 100
2 0 (degrees)


Figure 5.4. Powder x-ray diffraction scans comparing as grown AlCrN to post-annealed
AlCrN. No second phases are apparent after the 700C anneal.













CHAPTER 6
ALN AND GAN BASED DILUTE MAGNETIC SEMICONDUCTOR
APPLICATIONS

Spin Filter Applications of AIN DMS

The most obvious application of an AIN-based DMS into an active device is by its

use as a ferromagnetic barrier. A ferromagnetic insulator shows promise as a spin filter

by allowing tunneling of aligned electrons and preventing tunneling of antiparallel

electrons. The effect is a change in resistance with applied magnetic field and is dubbed

tunneling magneto-resistance (TMR). For example, EuS was used as a spin filter and

showed evidence of magnetoresistance (MR) exceeding 130% at temperatures below the

Curie temperature of the material (-70K). A1N has the potential of being applied in the

same manner, however with an operable device at room temperature, since the Curie

point is above 300K.

Device Structure and Design

The all semiconductor device structure schematic is shown in Figure 6.1. Two

device stacks were grown, one with a ferromagnetic insulator and a reference with a non

ferromagnetic insulator. The tunneling magneto-resistance (TMR) stack includes two

ferromagnetic layers: a spin injector which serves to provide electrons with aligned spin

states and a spin filter which controls the amount of charge transport via tunneling. The

reference stack contains only one ferromagnetic layer, the spin injector. In the case of the

TMR stack, tunneling is expected to increase if both the spin injector and spin filter are

magnetically aligned by the application of an external magnetic field. However, in the








case of the reference device, there is no change expected in tunneling with the application

of an external magnetic field.

In the TMR stack, optimal GaMnN was used as a spin injector and optimal AlMnN

was used as a spin filter. In the reference stack, the A1N was grown under the same

conditions as the TMR stack, but without the addition of Mn. Both structures were grown

on an MOCVD-GaN buffer from the same wafer, which was chemically pretreated with a

3 minute 1:1 HCI:H20 dip, a 25 minute UVO3 exposure, and a 5 minute BOE dip

followed by a DI rinse and N2 dry. The substrate was heated in situ under N plasma to

700C before growth and a streaky RHEED pattern was obtained prior to growth to

ensure a clean growth surface. A 50 nm thick GaN:Si layer was grown, followed by a 7.5

nm AlMnN layer (A1N, for the reference stack). The GaMnN spin injector was 10 nm

thick and another 50 nm GaN:Si layer was grown to provide better electrical contact to

the ohmics.

The devices were then fabricated into thin bars with large area contact pads on

each end. The mask design is shown in Figure 6.3. The bar bell type structures allow for

incorporation of anisotropic ferromagnets such as Co, for which the aspect ratio of 10:1

would force the coercive field to higher values. This shape is useful for creating large

differences in the values of the coercive fields to separate magneto-resistance maxima in

measurements. The devices were fabricated using standard photolithography to pattern

and etch mesas via ICP and to pattern then evaporate ohmic contacts. An SEM image of

the fabricated all semiconductor TMR device is shown in Fig. 6.4. Ohmic contacts

consisted of 50 nm Ti and 500 nm Au. Ohmic contacts were not annealed to improve








contact resistance due to the thermal instability of the DMS, which is addressed in

Chapter 5.

Device Testing and Results

Electrical and magneto-resistance measurements were performed to determine the

functionality of the devices. All measurements were done on both the TMR stacks and

the reference stack. Room temperature current-voltage (I-V) measurements of the TMR

stack without an applied magnetic field and with an applied magnetic field compare

device operation in the randomized state vs. device operation in the parallel state. Figure

6.5 shows an increase in slope and exponential behavior of the TMR device after an

applied field of H = 4000 Oe. The device in the randomized state shows less of the

expected tunneling behavior (as indicated by the decrease in slope). This suggests that the

tunneling increases when the GaMnN and AlMnN are aligned, which agrees with the

expected operation of the device. The room temperature I-V measurements provide a

useful survey method to determine the promise of the device.

Magneto-resistance (MR) measurements were performed in a Quantum Design

Physical Properties Measurement System (PPMS) by measuring the resistance of the

samples while sweeping an applied magnetic field. The PPMS measurements are limited

to a maximum of 95 mV and 999pA. These restrictions hinder measurements of highly

resistive material. The room temperature measurement of the reference sample showed

no change in resistance with respect to the magnetic field. This complies with the

expected behavior with no spin filter. The resistance vs. applied magnetic field (R vs. H)

at 5K is shown in Figure 6.6 for the reference device. A slight change in resistance

appears at high field, which most likely comes from the small change in resistance of the

GaMnN. Since the change is so small, it is most likely due to the thin spin injector's








inherent magnetoresistance. The origin of the splitting of resistances between 3000 Oe to

-500 Oe and on the up sweep between -500 Oe to 3000 Oe is unknown. However, the

same splitting is found in the R vs. H of the TMR device at 300K, shown in Figure 6.7.

No MR can be seen at 300K in the TMR device. Comparison measurements at 5K could

not be performed due to the high contact resistance found in the TMR device at low

temperature.

Structural investigation of the TMR device was performed by TEM of the cross

section of the device to determine the growth quality at the interfaces. Fig. 6.8 is a dark

field image where the TRM layers can easily be seen. Interfacial roughness between the

GaN:Si and the AlMnN layer is evident by the wave-like shape of the AlMnN layer.

Upon closer examination of the AlMnN layer, defects can be seen (Fig. 6.9). This

indicates that the growth of the device is not optimized, and is a plausible reason for the

inability of the device to operate as expected at room temperature.

In conclusion, room temperature I-V measurements showed promising results for

the TMR device. Tunneling increased when the spin injector and spin filter were aligned,

as compared to the randomized state. However, PPMS measurements were not quite so

promising, in that no TMR was observed at room temperature in the TMR device.

Unfortunately, due to the limits of the PPMS, no MR measurements could be made at

low temperature. Through TEM imaging, high amounts of strain were observed in the

AlMnN spin filter. This is one possible reason for no MR at 300K. Another possibility is

that the Mn impurity introduces impurity bands into the DMS (as discussed in Chapter 4).

This distribution of levels may cause the MR to be less defined.






54


Other spin injectors were also grown, namely FeNi and MnAs, which replaced the

GaMnN in the TMR stack. FeNi did not withstand the photolithography, and oxidized

and degraded before testing was accomplished, see Figure 6.10. The MnAs survived

processing, but was too soft for wire bonding, making MR measurements impossible.

Improvements on TMR devices can be made by using alternative tunnel layers which

contain an impurity level instead of impurity distribution. One such promising impurity is

Gd, which would allow for a single impurity level in III-N DMS.









50 nm GaN:Si
50Onm GaN:Si


10 nm GaMnN

7.5 nm AlMnN


50 nm GaN:Si


MOCVD GaN buffer
MOCVD GaN buffer


Figure 6.1. Schematic of the all semiconductor tunneling magneto-resistance stack. The
reference stack contained undoped A1N in place of the AlMnN layer, and all
thicknessed remained the same. The dark squares represent ohmic contacts
made to the top GaN:Si layer and to the underlying MOCVD GaN buffer.





Skeleton of Ohmic Window to Mesa and Valley


ED


55
55
99r


99
99


Figure 6.3. Mask design used for fabrication of all semiconductor device. Alignment
marks are found at each comer and in the middle of the mask. The larger dark
bars represent the area of the mesa of the device. The open, or light areas,
represent where ohmic contact was made to the device.


i


w


ED


0





































U 90x I
Figure 6.4. Scanning electron micrograph of all semiconductor device. Top view shows
top and bottom Ti/Au ohmic contact, the top of the mesa and the etched valley
showing the MOCVD GaN buffer.












0.050 -,

0.045

0.040 0 GaMnN/AIMnN/GaN, H = 4000 Oe
o GaMnN/AIMnN/GaN, H = 0
0.035 -

0.030 -

CL 0.025 -
E
-0.020

0.015

0.010

0.005 o o oo00000000000000000ooooooooaooo
0.005

0.000 0"_
0.0 0.1 0.2 0.3 0.4 0.5
V (Volts)


Figure 6.5. Current-voltage measurement of all semiconductor tunneling magneto-
resistance device with and without and applied magnetic field. Open circles
represent the I-V measurement taken without an applied field. Dark squares
represent I-V measurement taken after application of a 4000 Oe field. Note
that tunneling increases after the field is applied.













* GaMnN/AIN/GaN I
I = 999 pA
5K


3.750 -


3.745


3.740 I I
-3000-2500-2000-1500-1000 -50


N
'A

0%


. I


0 0 500 10001500200025003000
H (0e)


Figure 6.6. Resistance vs. applied field measurement taken at 5K for the all
semiconductor reference device.


3.755


. . . . . .


I I I I


, I i - ,


! i





60





-2.500 1 1 1 1 I I

GaMnN/AIMnN/GaN

-2.525 I = 999 A
T =300K

-2.550 -


0 | S
S-2.575 .o ^:".J ,,

0.0. : **e* *
5" ,, 0

-2.600 -


-2.625 I I I I
-3000-2500-2000-1500-1000-500 0 500 1000 1500 2000 2500 3000
H (0e)


Figure 6.7. Resistance vs. applied field taken at 300K for the all semiconductor tunneling
magneto-resistance device.






















_ 200,000 x
Figure 6.8. Dark field ZSTEM image taken of the all semiconductor tunneling magneto-
resistance device. The dark AlMnN layer shows roughness indicating poor
growth quality at the interface.



































Selected area diffraction pattern tunneling electron micrograph taken of the all
semiconductor tunneling magneto-resistance device. The AlMnN spin filter
layer is indicated with an arrow. Strain can be seen within the AlMnN layer as
indicated in the photo.






















Figure 6.10. Scanning electron micrograph of a tunneling magneto-resistance device with
FeNi as a spin injector. From the photo, the degradation of the FeNi is visible
on the contact pad. This degradation represents nearly 90% of the devices.
The FeNi degraded during photolithography, most likely due to the use of
solvents during processing.













CHAPTER 7
INVESTIGATION OF GADOLINIUM AS A MAGNETIC DOPANT

Introduction of a New Magnetic Dopant

The incorporation of transition metals (such as Mn and Cr) into III-Ns results in

ferromagnetic semiconductors. However, as discussed in previous chapters, the resulting

DMS is not optimal for incorporation into devices. Very recently, another potential

dopant for III-Ns has been reported.39 A large magnetic moment was found for GaN:Gd.

The strong ferromagnetism arose from a low concentration level (<1016 cm-3) of Gd in

GaN. This was attributed to a long range spin polarization of the Gd atoms within the

GaN. The importance of this finding lies in the ability to dope a DMS with carriers of a

concentration greater than the Gd concentration.39 Therefore, it is possible to effectively

utilize GaN:Gd in spintronic devices without relying on an impurity band or pinning the

Fermi level within the DMS. For this reason, the magnetic properties of epitaxial GaN

doped with Gd are explored.

Growth of GaN:Gd

Epitaxial GaN doped with Gd was grown by gas source molecular beam epitaxy

(GSMBE). Samples were grown on metalorganic chemical vapor deposition (MOCVD)

GaN buffers, chemically treated prior to growth with the standard preparation discussed

in Chapter 4. The substrates were heated to 700C under 1.6 sccm nitrogen plasma

before growth to verify a streaky reflection high energy electron diffraction (RHEED)

pattern, which indicates a clean growth surface. During growth initiation, the substrates

were first exposed to Ga and N, before the Gd shutter was opened. Each of the GaN:Gd








samples was grown at a substrate thermocouple reading (Ts) of 700C, a Ga cell

temperature (TGa) of 785C and a nitrogen plasma flow of 1.6 sccm. The Gd cell

temperature (TGd) was varied between 900C- 125C.

RHEED was used to monitor the growth in situ. Samples corresponding to TGd =

900-950C demonstrated spotty (3D) RHEED patterns. GaN:Gd grown at TGd = 1050C

demonstrated 2D/3D RHEED patterns with a 1x3 reconstruction. Figure 7.1 shows the

I x3 reconstruction observed during growth of the GaN:Gd. Atomic force microscopy

(AFM) was used to investigate the roughness of the films after growth. An AFM image

of Gd-doped GaN is shown in Figure 7.2 for TGd = 950C. The rms roughness value was

2.934 nm, which was typical for all GaN:Gd.

Magnetic Properties of GaN:Gd

Superconducting quantum interference device (SQUID) magnetometry was carried

out in a Quantum Design Magnetic Properties Measurement System (MPMS) to

investigate the magnetic behavior of the GaN:Gd. Background magnetization was

subtracted out as described in the appendix.

Magnetization vs. applied field (M vs. H) was scanned for each of the GaN:Gd

samples. For TGd = 900C, no hysteresis was observed at any temperature. At TGd =

950C, ferromagnetism is observed at 50K, as shown in Figure 7.3. However, the

magnetization level is very low at 50K and becomes noisy and questionable at 300K and

350K. For TGd in the range of 1000C-1100C, ferromagnetism is clearly evident in the

M vs. H loops. Figure 7.4 shows the M vs. H loops at 50K for the samples corresponding

to Tod = 1000-1100C. At TGd = 1125C, there is no evidence of ferromagnetism in

GaN:Gd at 50K, 300K, or 350K. Therefore, the Gd cell temperature range of 100C

yields ferromagnetism in GaN:Gd.








A comparison of the estimated saturation magnetization (corresponding to the

magnetization at 1600 Gauss) depicts the dependence of ferromagnetism on TGd. The

optimal Gd condition is TGd = 1050C, as shown in Figure 7.5. This trace of Ms and MR

at 300K vs. TGd implies that there is a certain amount of dopant which yields the

maximum magnetization. This behavior is similar to that found for the transition metal

dopants, Mn and Cr, as discussed in Chapter 4. However, unlike the transition metal

dopants, the Gd incorporation is at a much lower level. The flux readings for Mn and Cr

during growth are on the order of 10-8, but the Gd flux reading is much lower, below

10-10. The presence of an optimal cell temperature for the transition metals can be

attributed to the concentration level and incorporation site of the dopants. At too low of a

cell temperature, not enough dopant is incorporating substitutionally which results in a

low magnetization level. At too high a cell temperature (corresponding to a few atomic

percent dopant) the transition metals begin to occupy interstitial sites which results in a

degradation of the magnetic interaction. However, in the case of Gd, it is not probable

that the amount of Gd incorporated into the GaN can be high enough for the Gd to

incorporate interstitially. Therefore, a different mechanism must be driving the

magnetization in GaN:Gd.

The magnetization vs. applied field for TGd = 950C was found to decrease from

50K to 350K. However, this behavior was not observed at higher TGd. Figure 7.6 shows

the M vs. H loops at T = 50K, 300K, and 350K for TGd = 1050C. The magnetization

appears stable from 50K to 350K for TGd = 1050C as indicated by the loops. However,

the magnetization vs. temperature (M vs. T) seems to indicate a decrease in

magnetization, as shown in Figure 7.7. The zero field-cooled and field-cooled traces are








widely separated until near room temperature. The instability around room temperature

may be a sign that this temperature is in the vicinity of the Curie point (Tc) of the

material. This behavior is reminiscent of that of bulk Gd, which has a Tc 300K.

Conversely, the magnetization of the transition metals (which are antiferromagnetic) do

not correspond to the magnetization in III-N:TM. This is yet another indication that there

is a different mechanism of ferromagnetism in the GaN:Gd.

Thermal Stability Investigation of GaN:Gd

As described in detail in Chapter 5, the thermal stability of the optimal GaN:Gd

(TGd = 1050C) was investigated in a systematic way. The GaN:Gd was annealed from

300C to 7000C to determine the temperature at which magnetization is destroyed. M vs.

H measurements taken at 300K for each anneal temperature indicate the change in

magnetization with anneal temperature. Figure 7.8 shows the saturation magnetization vs.

the anneal temperature (Tannea). After the first anneal at 300C the saturation

magnetization immediately decreases from the as grown value and generally decreases as

the anneal temperature increases. This is a possible indication of the destruction of the

magnetic properties of the film. However, the M vs. H depicts a different scenario. Figure

7.9 compares the M vs. H of the as grown GaN:Gd to that annealed at 600C. As shown

in figure 7.8, the saturation magnetization decreases by a factor of greater than one half.

Although the maximum magnetic signal decreases, the overall hysteresis increases. The

loop becomes more square, with an increase in magnetic remanence and coercivity. This

broadening of the remanence and coercivity indicates a "hardening" of the

ferromagnetism. One possible explanation of the hardening involves a change in domain

structure. The high anneal temperature may alter the domain structure in such a way as to

make the domains less easy to manipulate, thereby requiring a stronger magnetic field to








produce zero magnetization. The increase in remanence with anneal temperature

indicates the retention of more magnetic alignment within the material. Therefore, it

seems plausible that annealing the GaN:Gd creates an alteration within the domain

structure producing a larger coercivity and remanence. Also, the decrease in saturation

magnetization indicates a decrease in the strength of the magnetic interactions within the

GaN:Gd, however the magnetization remains even after a 700C anneal. This retention of

magnetization after a high temperature anneal is an improvement in thermal stability

compared to the Cr-doped AIN.

The M vs. T corroborates the existence of magnetism after the 700C anneal.

Figure 7.10 shows the separation in the field-cooled and zero field-cooled loops up to

about 320K. Above 320K, the M vs. T suggests the magnetization is reduced greatly.

This is also seen in the M vs. H at 350K as compared to 50K, shown in Figure 7.11. The

overall hysteresis is decreased at 350K, with a lower coercivity and field at which the

material saturates. Although the hysteresis is beginning to vanish at 350K, the GaN:Gd

annealed at 700C remains ferromagnetic at 300K, as shown in Figure 7.12. This result

was not found for the material doped with transition metals, as discussed in Chapter 5.

This suggests that the magnetic interaction within the GaN:Gd is not destroyed after a

high temperature anneal. Moreover, it implies that Gd is a superior dopant to the

transition metals.

The incorporation of Gd into GaN results in an overall improvement of magnetic

properties. The magnetization is increased although the dopant concentration is lower.

More importantly, the ferromagnetism persists even after a 700C anneal. This results in

the possibility of incorporating GaN:Gd into spintronic devices which require a high








temperature anneal without destroying the magnetic properties. The application of

GaN:Gd into spintronic devices will rely on other material characteristics such as

transport, which will require investigation. Also, the impact of Gd on conductivity

requires further investigation. One possible method is to use current-voltage

measurements to probe the impurity within a quantum well. This should show whether

the impurity is a distribution of levels or a single level. Although more investigation is

needed to probe the effectiveness of GaN:Gd, the initial findings suggest it is superior to

other III-N DMS.


































ire 7.1. Reflection high energy electron diffraction pattern ot 6aN:(jd, TGd =
The picture shows a 2D/3D pattern with 1x3 reconstruction.










-1.00





-0,75





-0.50





-0.25


-0
0 0.25 0.50 0.7 5 1. 00 JM

Figure 7.2. Atomic force microscopy image representing GaN:Gd with rms roughness of
1.541 nm.


















D050427-1
GaN:Gd, TGd = 950C











1 T=50K




II I .. I I I


-1500 -1000 -500


0
H (Gauss)


500 1000 1500


Figure 7.3. Magnetization vs. applied field loop taken at 50K for GaN:Gd, with TGd =
950C. Hysteresis is observed at 50K, but not necessarily at 300K or 350K.


6.0x10"1


4.0x10"1


2.0xl 0"'


-2.0x101


-4.0x101


-6.0x1 0'














I I I *I I I
10 GaN:Gd

8 TUd= 1000C
6- TG=10500C I I

4 A T^ =1075C T
2 0 T G=1100C |

S-2 *^ ..^v


-4- ,, t T=50K

-8-

-10 -
I I I I I I
-1500 -1000 -500 0 500 1000 1500

X Axis Title


Figure 7.4. Magnetization vs. applied field taken at 50K for GaN:Gd corresponding to
TGd= 1000C-1100C.

















S5I 1 I05 C -
- TGd = 1050C- J


m M. at 300K
--- M. at 300K


8

7
C.,
E
C- 6

E
-5
C
0

N
a)3
0)
CU
22


I '1
E


0.00090 0.00092 0.00094 0.00096 0.00098 0.00100

1/TGd


Figure 7.5. Estimated saturation magnetization vs. inverse of Gd cell temperature plot
which shows the optimal TGd.


0
0 0
p I I I I


I















1.0xlO i i *

8.0x10 --T=50K i
-^---T = 300K
6.0x10 T = 350K

4.OX100

2.0x10-

E o.o0

S-2.0X100



-4.Oxi 00

-8.0x10

-1.0x10 ,---, I I
-1500 -1000 -500 0 500 1000 1500
H (Gauss)


Figure 7.6. Magnetization vs. applied field loops taken at 50K, 300K, and 350K for
GaN:Gd with TGd = 1050C.













1.2x 10

1.1x10

1.1x10

1.1x10

1.1x10


1 .lx10
1.0xl0

1.0xI0

1.0x10

9.8x101

9.6x 101

9.4x 101'


0 50 100 150 200
Temp (K)


250 300 350


Figure 7.7. Magnetization vs. temperature for GaN:Gd with TGd = 1050C. Note that the
magnetization is significantly decreased near room temperature.


D050511-1
GaN:Gd, TGd = 1050C



--o- Field-cooled
S---- Zero field-cooled
0





H 2..E0

i,.., -" 1H= 200 Oe
I I I I I I *















4.0 I I I I

E 3.8 _____-
M
E 3.6 -Ms of GaN:Gd
__ TGd = 1050C
c 3.4
0
N 3.2 T=300K

0) 3.0

c 2.8
0
T 2.6

U) 2.4 -

S2.2
E
".c 2.0 I I I I I i *
W 0 100 200 300 400 500 600

Tanneal (0C)


Figure 7.8. Estimated saturation magnetization vs. anneal temperature for GaN:Gd with
TGd = 1050C.






78







-*- as grown --
4 600C anneal --"

0o!-:,o 0-"
.o/ n^ .O::::.*o:*"*:-o.:':o ..--.o .-
", ../1 *0 ""

E .0 o
00



-2 .

-2 o, .....o T=300K


-4

I I I I I I
-1500 -1000 -500 0 500 1000 1500
H (Gauss)


Figure 7.9. Magnetization vs. applied field comparing the as grown GaN:Gd to that
annealed at 600C.












7.0x1 01
6.5x10'
6.0xlO'-
5.5x10'
5.0x 101
4.5x10.1
4.0x 10'
3.5x10'
3.0xl 0'
2.5x10'
2.0x10'
1.5x10-'
1.0x10!
5.0x10,2
0.0
-5.0x10'2
-1.0x1 0'
-1.5x10'
-2.0x10'


0 50 100 150 200
Temp (K)


250 300 350 400


Figure 7.10. Magnetization vs. temperature of GaN:Gd after 700C anneal.


GaN:Gd, T = 1050C
* annealed 7000C







t FC
0 ZFC




I I I I I I 2 I






80





I I

4.Oxl0
--T=50K I
0 -T=350K
2.0x100 --

E
0.0
E .


-2.0x10 I
i "GaN:Gd annealed 700C


-4.0x10 -0

-1500 -1000 -500 0 500 1000 1500
H (Gauss)


Figure 7.11. Magnetization vs. applied field at 50K and 350K for GaN:Gd annealed at
700C.











6.0x10 -

4.0x10 GaN:Gd annealed 700C
2"X10 i i t i^

2.0x100 -
E
*0.0
E
-2.x10o i
.4,0x0 ii T=300K

-4.0x10 -

-6.0x10 I I I ,- I I I
-1500 -1000 -500 0 500 1000 1500
H (Gauss)


Figure 7.12. Magnetization vs. applied field at 300K for GaN:Gd annealed at 700C.













CHAPTER 8
CONCLUSION

The potential technological importance of dilute magnetic semiconductors ranges

from an additional degree of freedom within a device to allow a multifunctional chip and

the distant goal of quantum computing. However, in order to realize such possibilities, a

few requirements must be met.

The most fundamental responsibility of the dilute magnetic semiconductor (DMS)

is to exhibit ferromagnetism in thin film form. This provides evidence that the material

has potential function within DMS-based device prospects. The ferromagnetism must

exist at or beyond room temperature in order for the DMS to operate efficiently within

current technology. Another requirement which must be met is that the DMS be stable

against high temperatures which are found during the processing and fabrication of

devices. This is an obvious requirement since the destruction of the DMS during

fabrication into a device would lead to the inability of the fully fabricated device to

operate. The more obscure requirements arise during the application of the DMS into

currently used device schemes. For example, active devices (devices which depend on

electron movement through the material) normally utilize single energy levels within a

given layer of the device. These issues with respect to the effectiveness of A1N as a DMS

have been addressed and are summed up in the following paragraphs.

AIN-based DMS Survey

Three dopants, namely Mn, Cr, and Co, were surveyed as effective magnetic

impurities in A1N by ion implantation. This is a quick method to incorporate a








predetermined amount of impurity and determine the plausibility of a given DMS. A1N

substrates, grown by metal organic vapor deposition, were implanted with approximately

three atomic percent of each impurity ion. Powder x-ray diffraction showed that the Mn-

implanted A1N was single phase; however, both the Cr- and Co-implanted A1N exhibited

second phases. Superconducting quantum interference device magnetometry showed that

the single pahse Mn-implanted A1N exhibited ferromagnetism at temperatures below

300K. Both the Cr- and Co-implanted A1N exhibited ferromagnetism at 300K. The

additional phases found in the Cr- and Co-implanted material do not contribute to the

ferromagnetic signal at 300K since the phases themselves are not ferromagnetic. In

general, the use of ion implantation to introduce magnetic dopants into AIN suggests that

Cr and Co ions are more magnetically active than Mn ions. However, Mn is more readily

incorporated into AIN to produce single phase material. For these reasons, epitaxial

AlMnN and AlCrN were investigated.

Impurity Comparison

Two AIN-based DMS were grown by molecular beam epitaxy. The two different

magnetic impurities introduced into the A1N host lattice were Mn and Cr, and the

magnetic properties were investigated to determine the optimal dopant. The impurity cell

temperature and the V/Ill were varied for both epitaxial AlCrN and AlMnN to optimize

the magnetic signal. Both the AlMnN and AlCrN were ferromagnetic at room

temperature. However, the magnetic signal versus applied field of the AIMnN decreased

by a factor of 2 from 100K to 300K. The magnetic signal of the AlCrN showed no

evidence of lessening over the same temperature range. It is possible that the Curie

temperature of the AlMnN is lower than that of the AICrN. Also, the magnetization of the








AlMnN was overall smaller than that of the AlCrN. This suggests that Cr is a better

magnetic impurity than Mn in an AIN-based DMS.

Thermal Instability

In order to determine the thermal stability of the AlCrN in an environment similar

to that encountered during device processing, the film was thermally annealed in a

systematic way. The magnetization of the AlCrN was found to decrease by almost one-

third after a fairly low temperature anneal at 300C. The magnetization was nearly

completely destroyed by an anneal at 700C. A very small amount of hysteresis

remained, but the magnetization was nearly equal to zero. The AlCrN showed no

evidence of second phases after being annealed, which suggests that the magnetization

was not destroyed due to the formation of nonmagnetic phases. Most likely, the magnetic

interaction between the impurities was destroyed by the 700C anneal.

Device Applications

The AlMnN was incorporated into an all-semiconductor tunneling

magnetoresistance structure. The device was designed so that the resistance would

decrease when the magnetization of the two ferromagnetic semiconducting layers was

aligned. However, the device was found to demonstrate no dependence of resistance on

magnetic field. It is possible that current limitations of the measurement system are

responsible for the lack of resistance response to the applied field. However, it is likely

that the distribution of levels available for tunneling created a passageway for electrons

which does not depend strictly on magnetic field. Therefore, the impurity band would

desensitize the magnetoresistance and prevent the device from operating in the expected

manner.








Summary and Future Work

Although the Cr was found to be a better magnetic dopant than the Mn in AIN, it

was determined that neither Cr nor Mn are viable impurities for an AIN-based DMS. The

incorporation of such a DMS into current technology is prevented by two things. The

AICrN is unable to withstand the high temperatures encountered during the processing of

devices. Also, the introduction of one-three atomic percent of an impurity such as Mn or

Cr introduces an impurity band into the semiconductor. This proves to be detrimental to

an active device which relies on the transport of carriers through a single impurity level,

not a distribution of impurity levels. Therefore, an alternative magnetic impurity should

be investigated.

One such potential dopant includes Gd. It is possible to incorporate less than an

atomic percent of Gd into a host semiconductor lattice. Therefore, Gd would act as a

"true" dopant by introducing a single impurity level. If the incorporation of Gd into AIN

results in ferromagnetism at room temperature, then it would be possible to incorporate

an AIN-based DMS into an active device and potentially into current technology.

Initial investigations of GaN:Gd grown by gas source molecular beam epitaxy

indicate that the material is ferromagnetic at room temperature. Also, the ferromagnetism

appears to resist thermal anneals at 700C. A quantum well consisting of 2.5 nm GaN:Gd

layers separated by lOnm A1N spacer layers has been grown to determine the impurity

level associated with the Gd dopant. The viability of Gd-doped III-N DMS rests in the

proof that Gd introduces a single impurity level. This would allow the incorporation of

the DMS into current semiconductor technology and would prove the capability of

spintronics.













APPENDIX
SUPERCONDUCTING QUANTUM INTERFERENCE DEVICE MAGNETOMETRY

Magnetization measurements were performed in a Quantum Design's Magnetic

Properties Measurement System which uses a superconducting quantum interference

device (SQUID) to provide the capability to sense the magnetic moment of materials

under investigation. The SQUID consists of two Josephson junctions in a closed

superconducting loop, which allows the SQUID to detect extraordinarily small changes in

external magnetic fields. The SQUID itself does not directly sense the magnetic moments

of the material under investigation. The movement of the sample through

superconducting rings that are connected to the SQUID with superconducting wires

induces a current through the detection coils. The change of magnetic flux through the

superconducting rings results in a change in electrical current which is proportional to the

change in magnetic flux. The superconducting rings are inductively coupled to the

SQUID detector, and therefore the change in the current through the rings alters the

SQUID output voltage which is proportional to the magnetic moment of the material

under investigation. The SQUID magnetometer is routinely calibrated with a Palladium

sample of known mass. The SQUID is sensitive to very small current variations in the

superconducting coils which allows measurement of samples with very small magnetic

moments. This is the most sensitive of the magnetometry techniques, which is useful in

investigating dilute magnetic semiconductors with very low magnetic moments.








Sample Measurement Method

Typically samples are measured within a plastic drinking straw (by suggestion of

the SQUID manufacturer, Quantum Design). The plastic straw is not known to have

ferromagnetic signal associated with it. However, there is a small signal (mostly

diamagnetic) which arises due to both the straw and also the substrate. This signal is

subtracted as explained in the next paragraph. The general size of each sample measured

in the SQUID is less than one centimeter squared. The sample is situated in the straw

perfectly parallel to the straw so that the magnetic signal parallel to the sample is

measured. The sample is centered with respect to the coils prior to measuring the

magnetization vs. field or temperature. Also, the magnet itself is degaussed prior to

measurement to ensure that no trapped magnetic flux remains in the magnet. Zero field

cooled (ZFC) traces were performed by cooling the sample to 10K then applying a small

magnetic field before sweeping the temperature. Field cooled (FC) measurements were

performed by cooling the sample to 10K under a small applied field then sweeping the

temperature. The net result of the zero field cooled and field cooled measurement is the

determination of the temperature range over which irreversibility in the magnetization

exists. In other words, determining the temperature over which the sample remains

ferromagnetic. This range of irreversibility can also be seen by performing M vs. H loops

at varying temperature. Both measurements together provide sufficient evidence to the

type of magnetization of the material. If both hysteresis in the M vs. H loop and a

difference in the FC and ZFC traces can be observed, then the material is deemed

ferromagnetic.








Background Subtraction

Another important aspect of the magnetization measurements is the subtraction of

the background magnetization. The step is performed for each of the measurements

(magnetization vs. applied field, or M vs. H) to eliminate any extraneous magnetic signal.

At the end of the M vs. H loop, the magnet is swept from 1 Tesla (T) to 5T to bring out

the weaker diamagnetic and paramagnetic signals within the sample and sample holder.

The slope of this high field sweep is the susceptibility of the background magnetization.

The susceptibility multiplied by the magnetic field is then subtracted from the raw data.

This is the method used to subtract the background magnetization.















LIST OF REFERENCES

1. Wolf, Stuart A., J. of Superconductivity: Incorporating Novel Magnetism 13, 195-
199 (2000)

2. Das Sarma, S., American Scientist 89, 516-523 Nov.-Dec. (2001)

3. Zorpette, G.. IEEE Spectrum December (2001) 30-35

4. J. Gregg, W. Allen, N. Viart, R. Kirschman, C. Sirisathitkul, J-P. Schille, M.
Gester, S. Thompson, P. Sparks, V. Da Costa, K. Ounadjela, M. Skvarla, J. of Mag.
and Mag. Mat. 175, 1-9 (1997)

5. J. L. Simonds, Physics Today 26-32 April (1995)

6. S. Wolf, D. Treger, IEEE Transactions on Magnetics 36 2748-2751 (2000)

7. S. Das Sarma, J. Fabian, X. Hu, I. Zutic, Sol. St. Comm. 119 207-215 (2001)

8. N. F. Mott, Proc. R. Soc. 153 633 (1936)

9. M. Oestrich, J. Hubner, D. Hagele, P. J. Klar, W. Heimbrodt, W. W. Ruhle, D. E.
Ashenford, B. Lunn, Appl. Phys. Lett. 74 1251-1253 (1999)

10. G. Prinz, Physics Today 58-63 April (1995)

11. D. D. Awshalom, J. M. Kikkawa, Physics Today 33-38 June (1999)

12. M. Osofsky, J. of Supercon.: Incorporating Novel Magnetism 13 209-219 (2000)

13. C. Gould, G. Schmidt, G. Richter, R. Fiederling, P. Grabs, L. W. Molenkamp,
Appl. Surf. Sci. 7640 1-8 (2000)

14. S. von Molnar, D. Read, J. Mag. And Mag. Mat. 242-245 13-20 (2002)

15. B. T. Jonker, X. Liu, W. C. Chou, A. Petrou, J. Wamock, J. J. Krebs, G. A. Prinz, J.
Appl. Phys. 69 6097-6102 (1991)

16. G. Schmidt, L. W. Molenkamp, Physica E 10 484-488 (2001)

17. H. Ohno, F. Matsukura, Y. Ohno, Sol. St. Comm. 119 281-289 (2001)

18. G. Schmidt, L. W. Molenkamp, J. Appl. Phys. 89 7443-7447 (2001)





90


19. T. Dietl, H. Ohno, Physica E 9 185-193 (2001)

20. H. Munekata, S. von Molnar, A. Segmuller, L. L. Chang, L. Esaki, Phys. Rev. Lett.
63 1849-1852 (1989)

21. H. Ohno, H. Munekata, S. von Molnar, L. L. Chang, J. Appl. Phys. 69 6103-6108
(1991)

22. H. Ohno, H. Munekata, T. Penney, S. von Molnar, L. L. Chang, Phys. Rev. Lett. 68
2664-2667 (1992)

23. H. Munekata, T. Abe, S. Koshihara, A. Oiwa, M. Hirasawa, S. Katsumoto, Y. lye,
C. Urano, H. Takagi, J. Appl. Phys. 81 4862-4864 (1997)

24. H. Ohno, A. Shen, F. Matsukura, A. Oiwa, A. Endo, S. Katsumoto, Y. lye, Appl.
Phys. Lett. 69 363-365 (1996)

25. T. Hayashi, M. Tanaka, K. Seto, T. Nishinaga, H. Shimada, H. Hayashi, K. Niihara,
Physica E 2 404-407 (1998)

26. A. Van Esch, L. Van Bockstal, J. De Boeck, G. Verbanck, A. S. van Steenbergen,
P. J. Wellmann, B. Grietens, R. Bogaerts, F. Herlach, G. Borghs, Phys. Rev. B 56
13103-13112 (1997)

27. F. Matsukura, Phys. Rev. B 57 R2037 (1998)

28. H. Saito, W. Zaets, R. Akimoto, K. Ando, Y. Mishima, M. Tanaka, J. App. Phys.
89 7392-7394 (2001)

29. T. Dietl, H. Ohno, F. Matsukura, J. Cibert, D. Ferrand, Science 287 1019-1022
(2000)

30. C. Liu, E. Alves, A. R. Ramos, M. F. da Silva, J. C. Shares, T. Matsutani, M.
Kiuchi, Nuc. Inst. Meth. Phys. Res. B 191 544-548 (2002)

31. E. Kulatov, H. Nakayama, H. Mariette, H. Ohta, Y. A. Uspenskii, Phys. Rev. B 66
1-9(2002)

32. M. E. Overberg, C. R. Abernathy, S. J. Pearton, N. A. Theodoropoulou, K. T.
McCarthy, A. F. Hebard, Appl. Phys. Lett. 79 1312-1314 (2001)

33. H. Hori, S. Sonoda, T. Sasaki, Y. Yamamoto, S. Shimizu, K. Suga, K. Kindo,
Physica B 324 142-150 (2002)

34. S. Sonoda, S. Shimizu, T. Sasaki, Y. Yamamoto, H. Hori, J. Cryst. Gr. 237-239
1358-1362 (2002)