Zr‘s
ﬁg

k.
A.

1‘

. 4.5... J
”mquwauyrb. .. .r
2...“..2???

23:0. 3.. 1
« 2|..—

-. o. of
s. a}
th‘ﬂ.‘

gt”.
a 3.  
%.2
. ‘
. .. Hanna.” 5... . ., . w
1. . . x. .
. . , . . . #5.?
.. 1.5 ...r.. . 3m»:

to.

z"

, $2:
up :2:

i,“

3!

‘1'?

"e

. .
1:1“. .
d. “a.” , :. ._
r) . . “WM.“
r , , V Lam.“ ,
i . .
mm". . . ,
V “Hi ., ‘

15....
.540.

v.

0 r1. aﬁlﬁ

.1. ., _

>

is...

.Jm....w.x..\ ‘ . ‘

~ 1 . . .2. 1..., Q

‘ , 5.51.3.3...u
‘ S .. .

. V .

.-

L

d

a ‘

. A
:2

a. .I .t J :4: _ w: n

. ‘ was} 1....“phvﬁﬂuﬁ ”may

a, w. ”my... a}: £5.36... ‘ 3.: .

..._...\,;.\.r...~ .., 9:51.“ 4&3»

.Vi:~v. .4 w xx .915

.
..

..
Ark: . .
, 4 . v ‘
. “u? :23:
, ‘ 93. x...”
. ,5 It... A: .
X!
.?
in: :

:— a

‘\ nuthv:

1%:

. p‘ u I .

2.” (“\AQvlu
3...,

‘1

v
. . ‘11
. F1: . 1‘
3:5...
I:
.4ro‘lAc
. . L .

 

 

_ 2 .
I);

2.

.
3...?

. , q , ..|‘ ~
galsﬁlmm
h

>
. .. any
. . .31?! ..

 

11"?

 

 

 

(It .55., x
. {@5941 a... ; ”urn
.. vv‘ .XVVAI Inu‘

:Il‘laalc. 1|

 

 

THESlS

200 i

This is to certify that the
dissertation entitled
Growth and Characterization of Epitaxial Films and
Magnetic Multilayers Grown By DC Sputtering

presented by

Reza Loloee

has been accepted towards fulﬁllment
of the requirements for

Ph.D. Materials Science and

degree in
Engineering

 

 

ﬂ)

‘ﬁajor professor I

Date December 13, 2000

 

MS U i: an Afﬁrmative Action/Equal Opportunity Institution 0- 12771

 

 

 

 

LIBRARY

Michigan State
University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

Anna
"- O gum.
lhni! 0 n “'
Via

 

 

 

 

 

 

 

 

 

 

 

 

6/01 excuncmm.pes-p.1s

 

GROWTH AND CHARACTERIZATION OF EPITAXIAL FILMS AND
MAGNETIC MULTILAYERS GROWN BY
DC SPUTTERING.

By

REZA LOLOEE

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Materials Science and Mechanics

2000

ABSTRACT

GROWTH AND CHARACTERIZATION OF EPITAXIAL FILMS AND
MAGNETIC MULTILAYERS GROWN BY
DC SPUTTERING.

By

REZA LOLOEE

Epitaxial Nb—Cu-Co and Nb-Cu-PermalloyTM (‘Py’) multilayers have been grown on
(1 l 5 O) sapphire substrates by sputter deposition. Electron backseatter diffraction
patterns (EBSPs) have been used to characterize individual layers, while atomic force
microscopy (AFM) was used for studying surface topography. EBSP is a valuable
technique for characterization of these thin sputtered ﬁlms and shows that the Nb ﬁlms
grow as high-quality epitaxial single crystals with (110)bcc orientations. Cu ﬁlms grown
on the epitaxial Nb display two in-plane epitaxial variants corresponding to two stacking
sequences of {111}fcc planes. These Cu variants take up the Nishiyama-Wasserman
orientation relationship with the underlying Nb. Subsequent sputtering of Co or Py on
the epitaxial Cu ﬁlms results in growth of two variants of {111}fcc layers. Orientation
maps of the epitaxial Cu, Co, and Py ﬁlms grown at elevated temperatures, which
illustrate the size and distribution of the growth variants, are presented. The EBSP
analysis of the epitaxial Cu, Co, and Py samples grown at 90°C or room temperature

revealed that two different orientation variants are present. However, due to degradation

of the associated EBSP patterns, mapping of the orientation distributions was not
possible.

In this dissertation, the effects of epitaxial growth of sputter-deposited multilayers and
spin-valves in a chosen orientation on the magneto-transport properties of these systems
are discussed. In particular, sputter-deposited (Py-Cu-Py-FeMn) exchange-biased spin-
valves (EBSV) structures were chosen for epitaxial (111)-orientated giant
magnetoresistance (GMR) measurements. Optical lithography was used to prepare
samples allowing current ﬂow perpendicular to the layer plane (CPP), which provides the
best access to the ﬁmdamental GMR parameters. The results obtained from magneto-
transport measurements revealed that the deposition rate affects the exchange anisotropy
between ferromagnetic and antiferromagnetic layers, the CPP-magnetoresistance, and the
coercive ﬁeld of the epitaxial EBSV. Studies of pinning the ferromagnetic layer to the
adjacent antiferromagnetic layer along different in-plane axes provided information about
the magnetocrystalline and shape anisotropy of both pinned and free ferromagnetic

layers. Results have been compared to those for polycrystalline samples.

TO MY PARENTS, MY WIFE AND MY SONS

ACKNOWLEDGMENTS

There are many people in both Department of Materials Science and Mechanics, and
Department of Physics and Astronomy that I would like to thank them for their help and

support during this study.

Department of Materials Science and Mechanics:

Many thanks go to my thesis advisor, Professor Martin A. Crimp for his guidance and
support for duration of this project. I am especially grateful for his patience and
encouragement during the writing of this dissertation. I would also like to thank my
guidance committee, Professor David Grummon and Professor Thomas R. Bieler for their
helpful insights and suggestions.

I would like to thank Steve Ng and Ben Simkin for their help with Cam Scan, Hong
Geng for teaching me TEM sample preparation, Dr. John Heckman and Dr. Richard

Schalek for editing a part my dissertation.

Department of Physics and Astronomy:

I would like to thank Professor Jack Bass and Professor William P. Pratt for their
patience, endless encouragement and support, helpful insights and enthusiastic
discussions. I would also like to thank my colleagues, Professor Phillip Duxbury and
Professor Thomas Kaplan for helpful discussions. Many thanks are due to Tom
Palazzolo, Jim Muns, and Tom Hudson for their help and input on design of sputtering

equipment. A special thanks goes to Nate Verhanovitz for helping me with using

V

AutoCAD drawing software, editing and for writing such wonderful LabView programs
for automating our Quick Dippers and sputtering system. I would also like to thank Dr.
Juan. A. Caballero for fabricating some of the CPP Spin-valves, training me for
microfabrication using photolithographic technique, and having discussion that
contributed greatly to my understandings; Dr. Baokang Bi for AF M training; Dr’s. Ruby
Ghosh and Lowell McCann for editing and great comments. I would like to thank Bob
Slater for helping me with some of the sample preparation for GMR measurement. Also
our collaborator in Leeds University, Didier Bonzec has made a few samples and many
good comments on this project.

I am the most thankful to my wife Hora, my sons, Arshia and Avesta for their
patience, support and generous love over the past few years.

I would like to acknowledge and thank the Department of Physics and Astronomy for
their support and encouragement for employees seeking higher education.

The ﬁnancial support for this project was generously provided by the MSU CF MR
and the US NSF under grants DMR 98-20135 and MRSEC DMR 98-09688. I also
acknowledge the use of the W. M. Keck microfabrication facility at MSU, an NSF

MRSEC facility.

vi

TABLE OF CONTENTS

LIST OF TABLES.

LIST OF FIGURES.

1 INTRODUCTION.
1.1 The GMR EFFECT.
1.2 GMR MEASUREMENT.
1.3 GMR AND EPITAXIAL MAGNETIC SYSTEMS.
1.4 DEPOSITION TECHNIQUES.
1.5 CHARACTERIZATION TECHNIQUES.
1.6 REFERENCES . .

2 EXPERIMENTAL PROCEDURES.

2.] INTRODUCTION.

2. 2 APPARATUS DESIGN AND DEVELOPMENT.
2.2.1 Sputtering.
2. 2. 2 Room Temperature Chamber Top Assembly
2. 2. 3 High Temperature Chamber Top Assembly

2.3 THE GROWTH TECHNIQUES. .
2.3.1 Polycrystalline Films and Multilayers. .
2.3.2 Epitaxial Films and Multilayers.

2.4 STRUCTURAL CHARACTERIZATION .
2. 4. 1 Atomic Force Microsc0py
2. 4. 2 Electron Backscatter Pattern. .

2.5 MAGNETO- TRANSPORT CHARACTERIZATION.

2.5.1 Magneto-Resistance (CPP- GMR measurements).

2.5.2 Magnetization. .
2.6 REFERENCES

3 GROWTH OF Nb FILMS ON SAPPHIRE.
3.1 INTRODUCTION.
3.2 BACKGROUND
3.2.1 Growth of Nb Film on AL203. .
3.2.2 Grain Structure of Sputtered Metal Films
3.3 RESULTS
3.3.1 Grth Process
3. 3. 2 AFM Analyses. .
3. 3. 3 EBSP Analyses.

3. 3. 4 Orientation Relationship Between Nb and A12O3.

vii

xi

NOOON

l3

17

20
20
21
25
25
33
33
36
37
38
4o
44
44
45
46

47
47
51
52

54
54
58
63

3.4 DISCUSSION
3.5 SUMMARY .
3.6 REFERENCES.

4 GROWTH OF Cu FILMS ON EPITAXIAL Nb.
4.1 INTRODUCTION. .
4. 2 BACKGROUND: NUCLEATION AND GROWTH.
4.2.1 Orientation Relationship Between fcc and bcc
4.2.2 Nucleation
4.2.3 Twin Formation
4.2.4 Nature of Twin Boundaries
4.3 SPUTTER DEPOSITION OF Cu FILMS
4.4 RESULTS
4.4.] AFM ANALYSES.
4.4.1.1 As Grown Cu Films .
4.4.1.2 Effect of Film Thickness
4.4.1.3 Effect of Post Deposition Annealing.
4.4.1.4 Effect of Deposition Rate
4.4.1.5 Effect of Deposition Techniques
4. 4. 2 EBSP ANALYSES. . .
4.4.2.1 EBSP Patterns of Cu ﬁlms

4. 4. 2. 2 Effect of Growth Parameters on the Variants

Size and Their Distributions.

4.4.2.3.The Effect of Film Thickness
4.4.2.4.The Effect of Post Deposition Annealing
4.4.2.5 The Effect of Deposition Rate

4.4.2.6 The Effect of Deposition Techniques

4.5 ORIENTATION RELATIONSHIP BETWEEN Cu AND Nb. .

4.6 SUMMARY
4.7 REFERENCES.

5 GROWTH OF Co AND Py FILMS ON EPITAXIAL Cu.

5.1 INTRODUCTION .

5.2 GROWTH OF Co FILMS ON EPITAXIAL Cu(l l 1).
5.2.1 Nucleation and Growth -- a Review
5.2.2 Sputter Deposition of Co Films
5.2.3 AFM Analysis
5.2.4 EBSP Analyses. .
5.2.5 Growth Process

5.3 GROWTH OF Py FILMS ON EPITAXIAL Cu(111)
5.3.1 Nucleation and Growth --- a Review
5.3.2 AFM Analyses. .
5.3.3 EBSP Analyses. .

5.4 DISCUSSION.

5.5 SUMMARY

5.6 REFRENCS

viii

66
66
69

72
73
75
81
83
84
88

88
89
95
98
101
101
103
105

114
119
121
123
127
130
132
138

141
141
141
145
147
150
156
157
157
159
162
169
170
176

6 MAGNETO-TRANSPORT PROPERTIES OF EPITAXIAL MAGNETIC
FILMS AND SPIN VALVES.

6.1 MAGNETIC PROPERTIES OF FEROMAGNETIC FILMS 177
6.2 MAGNETIC PROPERTIES OF EPITAXIAL Py FILMS 180
6.2.1 Introduction 180
6.2.2 Experimental Pocedures 180
6. 2. 3 Results 181
6. 3 MAGNETIC PROPERTIES OF EPITAXIAL Py BASED EBSV 186
6.3.1 Introduction 186
6.3.2 Experimental Procedures 190
6.3.3 Results 190
6.4 SUMMARY. . . . . . . . 206
6.5 REFERENCES. . . . . . . . 211

7 SUMMARY AND CONCLUSIONS. . . . . 214

ix

LIST OF TABLES

Table 3-1 Crystallographic data for both Nb and A1203 is summarized. 49
Table 4-1 Crystallographic data for both Nb and Cu is summarized. 74
Table 4-2 Surface energy of selected elements and materials. 76
Table 4-3 Summary of Average Grain (twin) Sizes. 128
Table 4-4 Summary of the sputtered Cu ﬁlms grown on epitaxial Nb 133
Table 5-1 Crystallographic data of Co is summarized. » 146
Table 5-2 Crystallographic data of Py and Cu is summarized. 158

Table 5-3 Comparison of Lattice Mismatch, Atomic Radius, and Surface Energy of Nb,
Cu, Ag, Co, and Py 169
Table 5-4 Crystallographic data of Nb and Ag is summarized. 172

Table 6-1 The list of symmetrical exchange-biased spin-valves samples prepared with
photolithographic technique. 193

LIST OF FIGURES

Figures identiﬁed with asters in this dissertation are presented in color.

*Figure 1.1 a) A representation of anti- ferromagnetic coupling (external ﬁeld is zero)
and the effect of the applying magnetic ﬁeld in a magnetic multilayer. b) The ﬁeld
dependence of a Co/Ag multilayer (GMR effect) [10]. . . . . . . . 4

*Figure 1.2 Illustration of the scattering of spin-up and spin-down electrons in a
magnetic multilayer. E represents a ferromagnetic layer and N represents a non-magnetic
layer. . . . . . . . . . . . . . . . . 5

*Figure 1.3 a) Two magnetic states (antiparallel and parallel) of a spin valve consisting
of an antiferromagnetic layer (FeMn), pinned ferromagnetic layer (Py), nonmagnetic
layer (Cu), and free ferromagnetic layer (Py). b) The magnetoresistance of a spin valve
shows the GMR effect with small applied ﬁeld [8].. . . . . . . . . . 7

*Figure 1.4 (a) Schematic diagrams of the CPP and CIP conﬁgurations. Drawings are
notinscale....................10

*Figure 1.4 (b) The GMR effect in a [Co(6nm)/Ag(6nm)]6o magnetic multilayer
measured at 4.2Kusing the CPP and CIP conﬁgurations [10].. . . . . . . 11

*Figure 1.5 A non scaled schematic diagram of the MSU designed CPP spin-valve
sample. The up-right comer shows a picture of a sputtered sample with the CPP
conﬁguration..................12

*Figure 2.1 Front views of the MSU sputtering system with four dc-triode guns and a
magnetron gun capability, including the related power supply and the residual gas
analyzer..................23

*Figure 2. 2 Top views of: a) an unassembled triode gun, b) four assembled guns with
FeMn, Cu, Py, and Nb targets, c) four chimneys 1n the open position. Note that after
deposition a computer controlled stepper motor rotates the chimneys to the closed
position to protect the samples. . . . . . . . . . . 24

*Figure 2.3 (a) A top view of the room temperature chamber top assembly, (b) a bottom
view of the SPAMA plate, including the CPP and CIP masks, ﬁlm thickness monitors
(FTM), secondary heat exchanger, and the cold trap, (c) and a top view of the same
SPAMA plate, including the CPP and CIP substrate holders and FTMs. . . . 26

*Figure 2.4 A top view of the high temperature chamber top assembly including the
wobble sticks, FTM ampliﬁer, stepper motor, and some of the feed throughs. . . 28

xi

*Figure 2.5(a). A schematic of the high temperature heater assembly. The Cu heat shield
is water-cooled and slide into the Al guide, on top of the substrate holder. . . 30

*Figure 2.5(b) A schematic of the low temperature heater assembly. This heater is
capable being cooled form 600°C to about room temperature within a few minute, using
cooled-high pressure nitrogen gas. . . . . . . . . . . . 32

*Figure 2.6 Side views of high and low temperature heaters, sliding shutter, cooling
water line, and other accessories. The ﬁgure on the top right hand side comer shows an
unassembled low temperature heater. . . . . . . . . . . . 34

*Figure 2.7 (a) A top view of the Molybdenum substrate holders (b) and a schematic top
view of the Molybdenum mask for deposition of magnetic multilayers and spin-valves.
35

*Figure 2.8 Block diagram of the components used in D1 atomic force microscope with a
tapping mode capability. (Courtesy of Digital Instruments [5]) . . . . 39

Figure 2.9 a) Schematic representation illustrating the inelastic collisions, Bragg
diffraction, and resultant formation and imaging of EBSPs as a result of an incident
electron beam. Adopted from [6]. b) A diagram of three set of band from {220} plane
that are projected onto the phosphor as a pair of cones. The cones intersect the screen as
hyperbolae and deﬁne the edges of the Kikuchi bands. c) Experimental set-up for EBSP
measurements. Figures (b) and (c) are adopted from [10]. . . . . . 41

*Figure 2.10 (a) A view of the CamScan 44FE electron microscope column, (b) a front
view of main panel, (c) and PC monitor displaying an EBSP before and after indexing are
shown. . . . . . . . . . . . . . . . . . 43

*Figure 3.1 (a) A portion of the corundum cell and (b) atomic packing of a (1 120) plane
of sapphire. One of the forms of alumina, A1203, has the corundum crystal structure,
which is similar to a hexagonal close-packed structure. There are 12 Al ions and 18 0
ions associated with each unit cell. The lattice parameters of alumina are a = 0.475 nm
and c = 1.299 nm. The ionic radii of the Al 3+ = 0.051 and the O 2' = 0.132. There are 6
A1203 groups with 12 Al ions and 18 0 ions present in a hexagonal prism having the
given lattice parameters. . . . . . . . . . . . . . 50

Figure 3.2 (a) Affect of argon pressure and substrate temperature on microstructure of
sputtered metallic ﬁlms (adopted from [26]). (b) Zone model for evaporated metal ﬁlms
(adopted from [26]). . . . . . . . . . . . . . 53

Figure 3.3 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of Nb ﬁlms deposited at substrate temperatures of 25°C (No PDA),
450°C (PDA=1100°C), 660°C (PDA =1100°C), and 750°C (PDA=1100°C), with image
RMS roughness of 3.17, 2.52, 1.25, and 0.35 nm, respectively. . . . . 55

X11

Figure 3.3(e) Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of Nb ﬁlms deposited at substrate temperatures of 1100°C with
image RMS roughness of 33.1 nm. The epitaxial Nb ﬁlm grown at substrate temperature
of 1100°C was found to be very rough, often containing many holes. . . . . 56

Figure 3.3(1) Atomic-force-microscopy images showing the surface topography of two
Nb ﬁlms deposited at substrate temperature of 750°C and post deposition annealed at (a)
”00°C and (b) 950°C with image RMS roughness of 0.35 and 0.16 nm, respectively.

Figure 3.4 Electron backscatter pattern obtained from 250 nm Nb ﬁlms grown onto
(1 1 2 0)Al203 at a substrate temperature of : a) 25°C, b) 450°C, c) 650°C, (1) 750°C.

Figure 3.5 Electron backscatter diffraction pattern (EBSP) obtained from an epitaxial-
bcc Nb ﬁlm grown on (1 1 2 0) A1203 at a substrate temperature of 750°C. 0n left-hand
side, a few bands were chosen automatically by software. 0n right-hand side the pattern
wereindexed. . . . . . . . . . . . . . . . 62

Figure 3.6 Pole ﬁgures of the epitaxial Nb ﬁlm showing the distribution of <110>,
<111>, <112>, and <100> poles. . . . . . . . . . . 64

Figure 3-7. Stereographic projections of the epitaxial Nb ﬁlm and sapphire substrate
showing the epitaxial relationship between the Nb ﬁlms and A1203 substrates is
(1 10).,chb ”(1 1 2032,33203 with [111]bchb H[0001]hcpA1203. . . . 65

Figure 4.1 Atomic arrangement of the (a) NW and (b) KS (0 = 526°) relationships.
With 0, atoms on (110)bcc and O, atoms on (1 11)fcc [6]. . . . . . 77

*F1gure 4. 2 (a) Atomic arrangement of bcc Nb (110). The atomic radius of Nb 18 0.143
nm Wlth d(|10)=0.2337 nrn. . . . . . . . . . . 78

*Figure 4.2 (b). Atomic arrangement of fcc Cu (111). The atomic radius of Cu is 0.128
nm With (1(111) = 0.2078 nm. . . . . . . . . . . . . 79

*Figure 4- 2 (c) Unrelaxed overlaps of Cu(111) on Nb(110) in the N-W orientation
relationship (Cu[101] I I Nb[100]). For simplicity, perfect (defect free) Nb and Cu layers
are assumed... . . . . 80

*Figure 4-2 ((1) Unrelaxed overlaps of Cu(l l l) on Nb(110) with consideration of K-S
relatlonsh1p(Cu[011] Il Nb[111]).. . . . . . . . . . . 82

*Figure 4.3 Schematic representations showing top and side views of the stacking
sequences (twins) in(lll) fcc.. . . . . . . . . . . . 85

xiii

Figure 4.4 a) HRTEM images of incoherent 2(3) twin boundaries in Cu, imaged along
[011]. The 9R stacking faults are labeled, containing four or ﬁve layer defects. b)
Embedded atom method atomistic simulations of the structure of the Cu boundaries [36].

Figure 4.5 Atomic force-microscopy images and line trace surface analyses showing the
surface topography of 300 nm thick Cu ﬁlms deposited onto 20 nm thick Nb ﬁlm at
substrate temperatures of a) 25°C, b) 150°, and c) 350°C. The line trace surface analyses
shows the image RMS roughness are about 4, 4.5, and 20 nm, respectively. . 90

Figure 4.6 Atomic-force—microscopy images and line trace surface analyses showing the
surface topography of 150 nm thick Cu ﬁlms deposited at substrate temperatures of: a)
150°C , b) 350°C; and c) 475°C with image RMS roughness of about 1.3, 23, and 2.6 nm,
respectively................ 92

Figure 4.7 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 75 nm thick Cu ﬁlms deposited at substrate temperatures of: a)
150°C , b) 250°C; and c) 430°C with image RMS roughness of about 1.5, 12, and 20 nm,
respectively. . . . . . . . . . . . . . . . 93

Figure 4.8 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 20 nm thick Cu ﬁlms deposited at substrate temperatures of: a)
250°C and b) 350°C with image RMS roughness of about 5.5 and 10 nm, respectively.

Figure 4.9 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of: a) 20 , b) 100, c) 170, and d) 300 nm Cu ﬁlms; grown at 350°C on
a 250 nm epitaxial Nb buffer layer. The line trace surface analysis shows the image RMS
roughness are about 10, 28, 23, and 28 nm, respectively. . . . . . . 96

Figure 4.10 Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of: a) 20, b) 200, c) 250 mm Cu ﬁlms; grown at 150°C and post
deposition annealed at 350°C. The line trace surface analysis shows the image RMS
roughness are about 4.5, 5.7, and 8.5 nm, respectively. . . . . . 97

Figure 4.11 The effect of post deposition annealing on the surface topography of 300
nm thick Cu ﬁlms. The post deposition annealing temperatures were a) as grown
=150°C, b) 250°C, c) 350°C, and (1) 500°C, respectively. The line trace surface analysis
shows the image RMS roughness are about 4.8, 8.8, 6.3, and 3.6 nm, respectively. 99

Figure 4.12 Atomic force-microscopy images and line trace surface analyses showing
the surface topography of 300 nm thick Cu ﬁlms deposited at a substrate temperatures of
350°C and post deposition annealed at: a) as grown and b) 500°. The line trace surface
analyses shows the image RMS roughness are approximately 28 and 6.5, respectively.100

xiv

Figure 4.13 Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of 200 nm epitaxial Cu ﬁlms grown at 150°C and post deposition
annealed at 350°C with deposition rate of: a) 1, b) 4.1, and c) 7 A/s on a 250 nm epitaxial
Nb buffer layer. The line trace surface analyses shows the image RMS roughness are
about 1.5, 5.7, and 4.8 nm , respectively. . . . . . . . 102

Figure 4.14 Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of: a) 300 nm Cu ﬁlm sputter deposited at 350°C on a 250 nm
epitaxial Nb buffer layer, b) 300 nm Cu ﬁlm; grown at 350°C by MBE technique on a
100 nm epitaxial Nb buffer layer. The line trace surface analyses shows the image RMS
roughnesses are about 23 and 46 nm, respectively. The deposition rates were a) start with
lA/s for ﬁrst few monolayers and then 4.1 A/s and b) 0.5 A/s. . . . . 104

Figure 4.15 Indexed EBSPs and corresponding pole ﬁgures from a 300nm epitaxial Cu
ﬁlm sputtered at 350°C onto a 250 nm of epitaxial Nb substrate showing the two (111)
orientation variants. . . . . . . . . . . . . 106

Figure 4.16 Changes in the Kikuchi bands of EBSP patterns obtained from 300 nm thick
epitaxial Cu ﬁlm as the sample was moved by 0.1 pm steps. In part (a) and (b) the EBSP
patterns are from one of the variants. In parts (c) and (d) the Kikuchi bands contributed
from both variant were present. In parts (e) and (1), only the Kikuchi bands from other
variant are distinguishable. . . . . . . . . . 107

*Figure 4.17 Effect of tilting on electron beam size on the surface of the sample
(effective beam size) and on the interaction volume within the sample. . . . 109

*Figure 4.18 EBSP (111) pole ﬁgures and orientation map of the two (111) orientation
variants present in the 300 nm epitaxial Cu ﬁlm. In ﬁgure (a), the orientation map as
recorded by computer is shown. In the ﬁgure (b), the same map after noise was removed
by Channel 4 software. . . . . . . . . . . . 110

*Figure 4.19 Orientation map of the distribution of two twin variants in a 200000 um2
area (2000 um x 100 pm) of a 300 nm epitaxial Cu ﬁlm grown onto 250 nm of epitaxial
Nb. The green areas are from unsolved patterns, and all other color spots are improperly
indexed. Note that these maps were connected in one piece in the original map ........ 112

Figure 4.20. Orientation map of the 300 nm thick epitaxial Cu ﬁlms grown onto 250 nm
epitaxial Nb at a substrate temperature of a) 350°C and b) 150°C. . . . 115

Figure 4.21 Electron backscatter pattern obtained from a 300 nm Cu ﬁlm grown onto
epitaxial Nb ﬁlm at 90°C. This pattern shows some degradation compared to patterns
from Cu ﬁlm grown at 90°C. . . . . . . . . . . . 117

Figure 4.22 Changes in the Kikuchi bands of EBSP patterns obtained from a 300 nm
thick epitaxial Cu ﬁlm deposited at elevated temperature as sample was moved by 0.1 pm

XV

steps (see ﬁgure 4.16) are illustrated in (a), (b), and (c). (a) and (0) reveal two (111)
orientation twin variants. The pattern in (b) contains bands from both variants. In ((1),
the EBSP pattern of a Cu ﬁlm grown at 90°C 18 very similar to pattern (b) an indication
of presence of both twin variants. . . . . . . . . . 118

Figure 4.23 An AFM image of a 300 nm thick Cu ﬁlm deposited at 350°C,
superimposed on the orientation map of the same sample. This comparison shows there

is no correlation between the size islands (on AFM image) and size of twin variants (on
EBSP map). . . . . . . . . . . . . . 120

*Figure 4.24. Orientation maps of (a) 250, (b) 100, and (c) 20 nm Cu ﬁlms grown on a
250 nm epitaxial Nb ﬁlm at a substrate temperature of 350°C. The orientation maps in
0)) and (c) displayed a few different variants as indicated on maps (as different color),
which shows contributions from the underlying Nb. . . . . . . 122

Figure 4.25 a) Orientation map of the 300 nm Cu ﬁlm grown onto 250 nm epitaxial Nb
at a substrate temperature of 350°C. b) Orientation map of the 300 nm Cu ﬁlm grown
onto an epitaxial Nb ﬁlm at a substrate temperature of 350°C and post deposition
annealing of 500°C for 30 min. . . . . . . . . . . . 124

Figure 4.26 a) Orientation map of the 300 nm Cu ﬁlm grown onto 250 nm epitaxial Nb
at a substrate temperature of 150°C. b) Orientation map of the 300 nm Cu ﬁlm grown
onto an epitaxial Nb ﬁlm at a substrate temperature of 150°C and post deposition
annealing of 500°C for 30 min. . . . . . . . . . . . 125

Figure 4.27. Orientation maps of two 300 nm thick epitaxial Cu ﬁlms grown at 150°C
and post deposition annealed at 350°C with deposition rates of: a) lA/s and b) 7A/s. The
white areas on both maps are due to unsolved patterns (mostly on grain boundaries). 126

Figure 4.28 Orientation map of the distribution of two variants in 300 nm epitaxial Cu
ﬁlm grown onto. a) A 250 nm of epitaxial Nb using dc sputtering. b) A 100 nm of
epitaxial Nb using the MBE technique. (The MBE facility is located at University of
Leeds England)... . . . . . . . . . . 129

Figure 4. 29 Stereographic projections of the epitaxial Nb and Cu ﬁlms are shown. The
epitaxial relationship between the bee Nb and fcc Cu variants was found to be the
Nishiyama-Wasserman (N-W) relationship: (110)bchb| 1(lll)fccCu with [1 l
O]bchbI|[1 121.com. . . . . . . . . . . . . . 131

*Figure 4-B.1 (a) Comparison of orientation maps obtained form 20 um x 1 pm area of
300nm thick epitaxial Cu ﬁlm, using different electron beam scanning step size. A 50pm
aperture was used for these scans. (b) 10 11m x 1 um area of different location on the
same ﬁlm and without aperture. All other parameter were the same during scans. 136

xvi

*Figure 5.1. A superimposed atomic arrangement (unrelaxed) of Cu/Co with about 2%
lattice mismatch. The lighter areas are coincidence sites between Co and Cu atoms. The
darker areas are coherent with a large coherency length and are convenient sites for
nucleation of coherent islands. . . . . . . . . . . . 143

Figure 5.2 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 20 nm thick Co ﬁlms grown on 20 nm epitaxial Cu(111) (grown at
350°C) at substrate temperature of: a) 460°C and b) 80°C. The line trace surface
analyses shows the image RMS roughnesses of both ﬁlms are about 16 nm. . 148

Figure 5.3 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 100 nm thick Co ﬁlms grown on 200 nm epitaxial Cu(111) (grown
at 350°C) at substrate temperature of: a) 460°C and b) 80°C. The image RMS
roughnesses are about 11 and 4 nm, respectively. . . . . . . . 149

Figure 5.4 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 150 nm Co ﬁlms grown on 20 nm epitaxial Cu(111) (grown at
150°C and post deposition annealed at 350°C) at substrate temperatures of: a) 460°C and
b) 80°C. The image RMS roughnesses of about 2.6 and 3.1 nm, respectively. . 151

Figure 5.5 EBSPs from 150 nm epitaxial Co ﬁlms deposited on 20 nm epitaxial Cu ﬁlms

(grown at 150°C and post deposition annealed at 350°C) at substrate temperatures of: (a)
460°Cand(b)80°C. . . . . . . . . . . . . . . 153

Figure 5.6. Orientation map of a 150 nm epitaxial Co ﬁlm grown at 460°C on 20 nm
epitaxial Cu ﬁlm (grown at 150°C and post deposition annealing of 350°C) showing the
distribution of the twin variants. . . . . . . . . . . . . 155

Figure 5.7 AF M images and line trace surface analysis of 200 nm thick Py ﬁlms
deposited on 20 nm epitaxial Cu ﬁlms at substrate temperatures of 350°C and 90°C. The
epitaxial Cu ﬁlms were deposited at 150°C and post deposition annealed at 350°C. 160

Figure 5.8 AF M images and line trace surface analyses showing the surface topography
of 200 nm Py ﬁlms grown on 20 nm epitaxial Cu(111) (grown at 150°C and post
deposition annealed at 350°C) at substrate temperature of: a) 350°C and b) 90°C. The
image RMS roughnesses are about 2 and 9 nm, respectively. . . . . 161

Figure 5.9 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 200 nm Py ﬁlms grown at 90°C on a) 20 nm of epitaxial Cu(111)
(grown at 150°C and post deposition annealed at 350°C and different substrate than in
ﬁgure 5.6). b) 8 nm of FeMn. The RMS roughnesses are about 2 and 0.9 nm,
respectively. . . . . . . . . . . . . . . . 163

xvii

Figure 5.10 EBSPs from 200 nm thick epitaxial Py ﬁlms deposited on 20 nm thick
epitaxial Cu ﬁlms (grown at 150°C and post deposition annealed at 350°C) at substrate
temperatures of: (a) 350°C and (b) 90°C. . . . . . . . . . 164

Figure 5.11. Orientation map of 200 nm thick epitaxial Py ﬁlm grown on a 20 nm thick
epitaxial Cu ﬁlm (grown at 150°C and post deposition annealing of 350°C) at a substrate
temperature of 350°C showing the distribution of the twin variants in the Py. . 165

Figure 5.12 EBSPs from 200 nm epitaxial Py ﬁlms deposited on 20 nm epitaxial Cu
(grown at 150°C and post deposition annealed at 350°C) at substrate temperature of: (a)

350°C and (b) 90°C. (c) a EBSP pattern of a 200 nm epitaxial Py ﬁlm deposited on 8 nm
of FeMn (grown at 90°C) at a substrate temperature of 350°C. . . . . 167

*Figure 5.A.1. A superimposed atomic arrangement (unrelaxed) of fcc Ag (111) on bcc
Nb (110) (with N-W relationship and 1.1% mismatch) is shown. . . . 173

Figure 5. A. 2 EBSP patterns of a 300 nm epitaxial Py ﬁlm deposited on 200 nm epitaxial
Nb at substrate temperature of: (a) & (b) 350°C and (c) 90°C. Patterns 1n part (a) & (b)
are the EBSPs of twin areas. . . . . . . 174

*Figure 5.A.3. Orientation maps of 200 nm epitaxial Ag ﬁlms grown on epitaxial Nb at

substrate temperature of 350°C. The green areas are unsolved points (loss of EBSP
signal). . . . . . . . . . . . . . . . 175

*Figure 6.1. (a) A typical hysteresis loop of a ferromagnetic material (such as Fe, Co, Ni
or their compounds). (b) Magnetization hysteresis loops for a free Py layer (red) and a
Py-FeMn bilayer (blue) after pinning. The exchange-bias ﬁeld (HE) and the coercive
ﬁeld (He) of the pinned Py are indicated in the ﬁgure. . . . . . . . 179

*Figure 6.2 The magnetization hysteresis loops of 200 nm thick epitaxial Py ﬁlms grown
on 20 nm thick epitaxial Cu at substrate temperatures of 90°C (red) and 350°C (green). a)
The coercivity (Hc = half width of magnetization loop) increases from 12 to 50 Oe for
samples grown at low temperature (90°C). b & c) Comparison of the hysteresis loops of
both samples when the samples are rotated 90 degrees with respect to the ﬁrst orientation
within the sample plane. The hysteresis loops overlapped for both samples, therefore
there are no magnetocrystalline anisotropy in theses two epitaxial Py ﬁlms. . . 182

*Figure 6.3 The magnetization hysteresis loops of a 200nm epitaxial Py ﬁlm grown onto
8 nm epitaxial FeMn at room temperature a) As grown, b) Pinned. The green loops
correspond to the applied ﬁeld along a selected axis and the red loops are with the sample
rotated 90 degrees with respect to the ﬁrst within the sample plane. Note that zero degree
in (b) is along the easy axis (pinning direction) and applied ﬁeld. . . . . 185

xviii

*Figure 6.4 Schematic diagram of an exchange-biased hysteresis loop of an ideal spin-
valve. (a) Illustration of the orientation of the spins in both the pinned and free
ferromagnetic layers for both the parallel and antiparallel state. (b and c) illustrate the
scattering of spin up and spin down electrons in both the free and pinned ferromagnetic
layers..................188

*Figure 6.5 a) Sample geometry of the microfabricated exchange-biased spin-valve
samples. b) A cross-sectional diagram showing the current ﬂow and connections to the
samples. The contact area between the top Nb, with I+ and VI leads (current ﬂow), and
spin-valve is the area of the sample being measured [17]. . . . . . . 191

Figure 6.6 CPP Magnetoresistance hysteresis loop of a symmetrical Py based EBSV. In
this sample, an anti-parallel state is not observed, suggesting that there is magnetic
coupling between the free Py and pinned Py layers resulting in both of layers switching at
the same time. The coupling between Py layers 18 probably caused by growth of pinholes
due to a dirty substrate. . . . . . 196

Figure 6.7 CPP-MR hysteresis loop of an epitaxial symmetrical Py based EBSV sample
for which the substrates were cleaned in a clean room. AAR is 2.6 ﬁlm2 in this sample.
198

Figure 6.8 CPP-MR hysteresis loops of three epitaxial Py spin-valves: Nb250-Cu2o-Py24-
Cu24-Py24-FeMng-Au15-Nb250 nm. AAR are 2.8, 2.6, and 2.3 me2 for sample deposited
with Cu deposition rates of 1.2, 3.4, and 9.8 A/s, respectively. All layers were deposited
at 90°C except for the ﬁrst Nb and Cu layers. . . . . . . . 199

*Figure 6.9 CPP-MR hysteresis loop of an epitaxial symmetrical Py based EBSV:
Nb2so-Cu2o-Py24-Cu24-Py6-FeMn3-AuIs-Nb250nm. AAR is about 1.4 me2. . . 202

Figure 6.10 Minor hysteresis loops of an epitaxial EBSV (Py(24)—Cu(24)-Py(6)-FeMn(gnm)).
The epitaxial EBSV sample was pinned at a) 0°, b) 45°, and 90°. Zero and 90°,
corresponds to the [110] and [112] directions in the Py(111) plane. The similarity of
coercive ﬁelds suggests that there is no magnetic anisotropy (except some shape effect)
in this epitaxial EBSV. . . . . . . . . . . . . . 203

*Figure 6.11 Magnetization hysteresis loops of a small section (blue) of the epitaxial
EBSV (Cu2o-Py24-Cu24-Py6-FeMng-Cu20nm) pinned at a) 0° and b) 90° angles. The
coercive ﬁeld (30 0e) and saturation ﬁelds are equal in both directions. . . 204

Figure 6.12 The CPP- MR hysteresis loops of two polycrystalline Py spin-valves: Nb250-
Cu2o-FeMng-Py24-Cu24-Py24- Auls-szso nm and one epltaxial sz50-CU2o-Py24-CU24-Py24-
FeMng-Auls-Nb250 nmo . . . . . . . . . . . . . . . 207

Figure 6.A.1. The resistance and magnetization of a 250 nm thick epitaxial Nb ﬁlm as a
function of temperatures. The Tc of this ﬁlm is about 9.1 K. . . . . 210

xix

CHAPTER 1

INTRODUCTION

In recent years there has been rapid progress toward the development of a new
generation of magnetic devices such as magnetic head readers, magnetic position or
velocity sensors, and magnetic random access memory (MRAM). These magnetic
devices consist of layered structures with a (ferromagnetic / non-magnetic)N= number ofrepcms
conﬁguration, referred to as a magnetic multilayer, e.g. ([Co/Cu]N), or with a
(antiferromagnetic / ferromagnetic / non-magnetic / ferromagnetic) conﬁguration,
referred to as spin-valve, e.g. (FeMn/Py/Cu/PermalloyTM (‘Py’)). The operation of these
devices is based on the large changes in electrical resistance (AR) of a magnetic
multilayer upon experiencing an external magnetic ﬁeld, referred to as giant
magnetoresistance (GMR). Large changes in resistance upon application of a very small
magnetic ﬁeld (Oe), small ﬁeld dependence hysteresis (small coercive ﬁeld), and small
saturation ﬁelds are needed to optimize the performance of such devices. GMR in these
devices depends on structural details such as the thin-ﬁlm growth direction, interfacial
roughness and interrnixing, and grain boundaries. However, the exact effects of these
variables on given systems are not yet well understood. Consequently, it is necessary to
further characterize the growth and ﬁlm structure of these magnetic systems to explore
the effect of crystal structure on the magneto-transport properties for advanced
applications. This dissertation describes the development of sputter deposition
techniques for growing epitaxial magnetic multilayers and spin-valves. Furthermore, the

dissertation describes the results of crystallographic and magnetic characterization of

sputter-deposited epitaxial samples and compares these with polycrystalline samples

made by sputtering.

1.1 THE GMR EFFECT

The operation of some of magnetic devices such as head readers was originally based
on a small change (a few percent) in resistivity of a magnetic layer (generally a NiF e
alloy) during application of an external magnetic ﬁeld. The change in the resistance of
this type of single magnetic ﬁlm depends on the angle between the electric current and
the direction of applied ﬁeld (thus, the direction of magnetic moments). This behavior is
known as the anisotropic magnetoresistance effect (AMR). AMR is usually expressed by
the relation between resistivity and the angle between the magnetization in the
ferromagnetic layer.

p = p0 + Ap - c0320
where p0 is the resistivity at zero ﬁeld, Ap the change in resistivity upon applying
magnetic ﬁeld, and 0 is the angle between the magnetization vector and the applied
current [1].

In 1988 two European scientists, P. Gruenberg of Germany and A. Fert of France,
discovered a different magnetoresistance effect [2, 3]. Independently, they observed that
the application of an external magnetic ﬁeld to ultra-thin engineered multilayers
decreased the resistance drastically. This behavior is called the giant magnetoresistance
(GMR).

The GMR effect was ﬁrst seen in single crystal Fe/Cr superlattices grown by

molecular beam epitaxy (MBE), and in sputter deposited polycrystalline Fe/Cr

2

multilayers [2]. These multilayers consisted of alternating thin layers of magnetic and
non-magnetic materials, where with the thickness of the non-magnetic layer chosen so
that the magnetic moment of the Fe layers (magnetic layers) align anti-ferromagnetically
(anti-parallel = AP state) to minimize the energy. Under application of an external
magnetic ﬁeld, these moments aligned parallel (P state) to each other in the external ﬁeld
and the resistance decreased. The change in resistance in the magnetic multilayer is
explained by the two-current series-resistor model. According to this model, there are
two spin dependent conduction channels in ferromagnetic materials corresponding to the
spinT (majority) and spinJv (minority) electrons. In the simplest case the change in
resistivity is given by the following equation:
AER = SR Ti —91 M

when and 93M: mtmt

were 91 TJV= m

are the ﬁnal resistivities at AP state and

P state, respectively, and 91 T and 91 NL are the resistivity of the spin up and spin down
channels in ferromagnetic layers [4]. Figure 1.1 shows a schematic diagram of a
magnetic multilayer and the effect of the external ﬁeld on the resistance (GMR effect).
In Figure 1.2, the scattering of spin up and spin down electrons in a multilayer ﬁlm
system is shown. When the magnetic moments of the layers are antiparallel (ﬁg. 1.2(a)),
the electrons with one spin sense will scatter in the lower layer while the electrons with
the opposite spin sense will scatter in the upper layer, resulting in all of the electrons
being scattered and maximum resistance. If the moments are parallel, only electrons with
one spin sense will be scattered, the other electrons being free to move, resulting in

minimum resistance.

   

4.__

No Field Magnetic Field

I Ferromagnetic (Fe, Co, Ni)
Non-magnetic (Cr, Cu, Ag)

 

(a)

on
.vv'vr-vl .- lvvvvl'v—rvlv

 

“(3)3 (CIP)

 

 

 

'.l....l....l....l....l.
-0.5 -0.25 0 0.25 0.5

 

H (kOe)
MR% = 100 x |R(AP) — R(P)| / 12(1))
0’)

Figure 1.1 a) A representation of antiferromagnetic coupling (external ﬁeld is zero) and
the effect of the applying magnetic ﬁeld in a magnetic multilayer. b) The ﬁeld
dependence of a Co/Ag multilayer (GMR effect) [10].

4

Two-Current Channel-Model

i2

    

23'? ZZHQEZ 5!

 

mmﬂﬁﬂ “mm—T9”;
4 mtml

Figure 1.2 Illustration of the scattering of spin-up and spin-down electrons in a magnetic
multilayer. E represents a ferromagnetic layer and N represents a non-magnetic layer.

One of the major problems with magnetic multilayers of the original type [2, 3] was
the requirement of a high magnetic ﬁeld (more than 2000 0e for Co/Cu and more than
10000 0e for Fe/Cr) for alignment of the magnetic moments parallel to the external ﬁeld.
The high ﬁeld is not a problem in some applications. However, it is a problem in other
areas such as read-heads and MRAMs. This problem was solved by Dieny and co-
worker [5, 6] who succeeded in producing the GMR effect in an alternative exchange-
biased spin-valve (EBSV) (such as F eMn/Py/Cu/Py) conﬁguration. There is also another
type of spin-valve, hybrid spin-valve (such as Co/Cu/Py) [7]. In this study, only Py based
exchange-biased spin-valves (EBSV) are considered.

An EBSV consists of two ferromagnetic (F) layers (such as Py), one of which is
pinned to an antiferromagnetic (AF) layer (such as FeMn), separated by a non-magnetic
(N) layer (such as Cu) usually thick enough so exchange coupling between the Py layers
is weak (see Figure 1.3(a)). The role of the AF layer is to pin the in-plane direction of
magnetization of the adjacent ferromagnetic layer. In general, spin-valves have the
advantage of needing a very small external magnetic ﬁeld (low switching ﬁeld) to realign
the unpinned (free) magnetic layer to produce the GMR effect. This property makes the
spin-valves suitable candidates for low ﬁeld applications, such as those mentioned above.
Figure 1.3(b) shows the GMR effect of a typical Py-based symmetric EBSV. In this
ﬁgure, the switching from the parallel state to the antiparallel state is relatively fast and

requires only 20 (Oe) compared to approximately 450 (Oe) in the Figure 1.1(b).

   

”- It mmvgff
- ’y w 7’ Py
{—— , d— _,
g é E ,

 

 

 

l FeMn FeMn
‘ _
AP-state P-state
Magnetic Field
(a)
“R 3
(9‘) <—
—>
35
3
25
2
1.5
A 1 c
<—— 05 —>
-1co o 10) 21) an

Applied Field (0:)

Figure 1.3 a) Two magnetic states (antiparllel and parallel) of a spin valve consisting of
an antiferromagnetic layer (FeMn), pinned ferromagnetic layer (Py), nonmagnetic layer
(Cu), and ﬁee ferromagnetic layer (Py). b) The magnetoresistance of a spin valve shows
the GMR effect with small applied ﬁeld [8].

1.2 GMR MEASUREMENT

In regular magnetic multilayers and spin-valves, GMR is typically measured with a
four-probe current-in-plane (CIP) technique (current and magnetic ﬁeld are parallel). As
is described above, the GMR effect arises from spin-dependent scattering at the interfaces
of the multilayers and CIP—MR depends on the electron mean free path. Therefore, if the
electron mean free path is smaller than the thickness of the non-magnetic interlayer, a
weak GMR effect is observed [9]. In the CIP geometry, the current distribution is not
uniform, which adds to the complexity of modeling the GMR effect. In the early 19903,
W. Pratt et a1. [10] developed a new measurement technique in which the ﬂow of
electrons is forced perpendicular through the layers. This conﬁguration is called the
current perpendicular to the plane (CPP) geometry. In the CPP geometry,
superconducting Nb contacts on the top and bottom of the CPP-MR samples are used to
obtain uniform current ﬂow through the layers. Another main difference between CIP
and CPP-MR is that the mean free path is not an important factor in the CPP geometry
and the spin diffusion length (typically a few hundreds of nm) is the major factor
affecting CPP-MR [10]. Because of the high symmetry of the CPP conﬁguration, CPP-
MR is theoretically less complex and easier to understand than the CIP equivalent and
provides the best access to the fundamental parameters in modeling GMR [11].
In the CPP conﬁguration, the area which electrons ﬂow across is large, about 1 mm2
(WxT z 1 mm x 1 mm) and the resistance is measured across less than 1000 nanometers
(L z 0.0007 mm), compared to WxT z 1 mm x 0.0007 mm and L z 3 mm in CIP
samples. This results in very small CPP resistances (chp z 10'7 Q, compared to Rap z 1

0) which require sensitive measuring devices, such as a superconductive quantum

8

interference device (SQUID). Figure 1-4(a) illustrates both the CIP and CPP
conﬁgurations, as well as corresponding GMR measurements (Figure 1-4(b)). Note the
much larger GMR response in the CPP conﬁguration. In Figure 1-5, an MSU CPP
conﬁguration (lmmz) sample is shown. For study reported in this thesis, however,
optical lithography was used to prepare micron-sized pillar samples in the CPP

conﬁguration.

1.3 GMR AND EPITAXIAL MAGNETIC SYSTEMS

There are two kinds of interaction between ferromagnetic layers which are separated
by a non magnetic layer (metallic). The ﬁrst is ferromagnetic coupling due to exchange
coupling (see Appendix l-A and [12, 13]) between ferromagnetic layers which depends
on the thickness of the non magnetic layer. This type of coupling referred to as
Ruderrnan-Kittel-Kasuya—Yosida (RKKY)-like coupling. This coupling is due to an
exchange mechanism with an oscillatory behavior of GMR as a function of nonmagnetic
spacer thickness. The second is a magnetostatic coupling known as Neel’s orange-peel
effect, which is due to the interfacial roughness [14].

In a typical spin-valve, an interfacial interaction (see Appendix l-A) between the
antiferromagnetic and ferromagnetic layers results in an exchange anisotropy, which is a
unidirectional anisotropy or ‘exchange-bias’ [15]. These types of spin-valves, such as
FeMn/Py/Cu/Py, are referred to as exchange-biased spin-valves. The exchange-bias is a
key feature in spin-valves to achieve the antiparallel and parallel states between the two
ferromagnetic layers, one of which is adjacent to the antiferromagnetic layer (details are

given in chapter 6). Exchange bias has been studied for both polycrystalline and single-

 

CPP
Geometry

 

 

 

 

 

CIP
Geometry

 

 

 

 

 

, WxT
Rrppz 10 " Q (L ~07 um, W ~1000 um, T ~1000 um)
R ('w z IQ (L z3000 um, W z 1000 um, T z 0.7 pm)

Figure 1.4 (a) Schematic diagrams of the CPP and CIP conﬁgurations. Drawings are not
in scale

 

118(11): (CPP)

 

 

 

 

118(11): (CIP)

 

 

 

 

CPP—MR:450/0 > CIP-MR: 9%

Figure 1.4 (b) The GMR effect in a [Co(6nm)/Ag(6nm)]6o magnetic multilayer measured
at 4.2 K using the CPP and CIP conﬁgurations [10].

11

Anti-feroma_ . .

 

Figure 1.5 A non scaled schematic diagram of the MSU designed CPP spin-valve
sample. The up-right comer shows a picture of a sputtered sample with the CPP
conﬁguration.

crystal (MBE grown) systems [16, 17].

One of the advantages of the CPP geometry over CIP is that the CPP-MR should be
less sensitive to ﬁlm inhomogeneities such as grains, impurities, and structural defects
[14]. Comparable data for Co/Cu CPP-MR samples prepared by standard sputtering,
MBE, and electrodeposition seem to support this expectation [11]. In general, epitaxially
grown samples (typically grown by MBE) should have fewer structural defects. This is
due to the lower kinetic energy of the depositing atoms, very small deposition rate, and
ultra-high vacuum (~10‘ll torr) in such deposition systems. Therefore, epitaxially grown
samples should be useful for assessing fundamental GMR parameters and for exploring
anisotropic properties with respect to the ﬁlm crystal structure, e.g. magnetocrystalline
anisotropy in magnetic multilayers and spin-valves. At present, epitaxial sputter
deposited multilayers have mostly been studied by CIP-MR [18-21]. Only one group has
used MBE grown epitaxial Cu/Co(111) multilayers for CPP-MR studies [22]. However,
less expensive sputter deposition offers higher deposition rates, and a larger number of

substrates in the chamber at same time for growing epitaxial CPP samples.

1.4 DEPOSITION TECHNIQUES

In typical sputter deposition, the substrates are kept at ambient temperature, which
results in the growth of polycrystalline ﬁlms. By selecting a proper single crystal
substrate, the growth of a highly textured ﬁlm is possible [23]. To grow an epitaxial
metal ﬁlm, a proper single-crystal substrate and a high temperature environment (hot

substrate) is required [18, 24, 25].

13

For this study, a new sputtering top assembly with proper accessories, including high
and low temperature heaters, was designed and constructed to achieve epitaxial growth of
metallic ﬁlms. This new high temperature assembly top has been used in conjunction

with a pre-existing dc magnetron sputtering system.

1.5 CHARACTERIZATION TECHNIQUES

Epitaxial ﬁlms have been characterized using a wide range of techniques including x-
ray diffraction (XRD) [18, 26, 27], reﬂection high-energy electron diffraction (RHEED)
[26-29], low-energy electron diffraction (LEED) [26, 29], neutron diffraction, high-
resolution electron microscopy (HREM) [27, 30], scanning tunneling microscopy (STM)
[31, 32], and atomic force microscopy (AF M) [33]. However, it is very difficult to obtain
enough crystallographic information from an individual technique to form a complete
picture of the microstructure (in large scale). For example, x-ray diffraction or neutron
diffraction, provide information about spatial average area of the sample [34] but do not
provide detailed information about the size and shape of grains, existence and distribution
of different variants, and strain in the samples. Therefore, the use of a better
characterization technique to examine the relationship between morphology, orientation,
and orientation distribution is desirable.

A complimentary approach to many of the above characterization techniques is
acquisition of electron backscatter patterns (EBSP), which are made with a scanning
electron microscopy (SEM) technique and can measure crystallographic orientations,
grain distribution of crystalline materials, and strain in the samples. EBSP is one of the

few characterization technique that provides a practical means for coupling between the

14

orientational and morphological aspects of microstructures in a quantitative and
statistically reliable manner [35]. The backscattered electrons form patterns by near-
surface diffraction upon exiting the crystals. Therefore, the patterns are dependent on
near-surface crystallographic structure. The formation of EBSP has been reviewed in
detail by Dingley [36], and a number of computer image processing procedures for
analysis of these patterns have been developed [34-38]. However, the application of
EBSP and associated orientation mapping to the study of epitaxial metal ﬁlms and
superlattices has only been dealt within a cursory manner.

In this study EBSPs, AF M, and CPP-GMR have been used to characterize the
morphologies, crystallite size, orientation distribution (variants distribution), and
magneto-transport properties of epitaxial sputter deposited Nb, Cu, Co, and Py ﬁlms,
bilayers, trilayers, multilayers and spin-valves. The relative growth orientation
relationships (bcc/fcc) of the substrate and ﬁrst few layers have been characterized using
pole ﬁgures obtained from the EBSPs. Studies of ﬁlms with varying thickness have also
allowed assessment of the minimum effective ﬁlm thickness at which these ﬁlms may be
readily characterized by EBSP. For the ﬁrst time, orientation mapping has been used to
assess grain size and twin morphologies of each epitaxial layer in these magnetic
systems. The relationships between structural properties (such as misﬁt dislocations and
point defects), magneto-transport properties (such as, coercive ﬁeld, GMR value, and

magnetization), and magnetocrystalline anisotropy have been observed and are discussed.

15

APPENDIX l-A
Exchange Coupling: In most ferromagnetic materials, such as cobalt, nickel, iron,
and many alloys of these and other elements, a quantum interaction between neighboring
atoms tends to align the atomic magnetic dipoles in parallelism. At low temperature, they
become lined-up (coupled) due to their interactions and at high temperature (Curie
temperature) randomizing effect overcome the coupling [12, 13]. This quantum

interaction is called exchange coupling.

Exchange Anisotropy and Interfacial Interaction: “Exchange anisotropy refers to the
magnetic manifestations of an exchange coupling at the interface between two diﬁferent
magnetically ordered systems” [39]. For example, when a ferromagnetic ﬁlm, such as
Co, Py, or Fe is deposited on an antiferromagnetic ﬁlm such as F eMn, CoO, or NiO, by
heating the deposited sample to above Néel temperature of the antiferromagnetic and then
cool to room temperature in some external magnetic ﬁeld, the antiferromagnetic ﬁlm will
have a uniform spin direction at the interface. This is due to the inﬂuence of the
neighboring ferromagnetic spins. The inﬂuences of the ferromagnetic layer on the
neighboring antiferromagnetic layer due to interfacial interaction is called exchange
anisotropy. A simple model of antiferromagnetic-ferromagnetic interaction is shown in

ﬁgure below.

 

Ferromagnetic

 

 

 

 

 

 

. -1 Antiferromagnetic

 

l6

1.6 REFERENCES

l) D. J. Mapps, Applied Magnetism, Kluwer Academic Publisher, 377 (1994).

2) M. N. Baibich, J. M. Broto, A. Fert, F. Nguyen Van Dau, F. Petroff, P. Etienne, G.
Creuzet, A. Friederich, and J. Chazelas, Phys. Rev. Lett., 61, 2472 (1988).

3) G. Binash, P. Grﬁnberg, F. Saurenbach, and W. Zinn, Phys. Rev. B, 39, 4828 (1989).

4) T. Valet and A. Fert, Phys. Rev. B 48 7099 (1993).

5) B. Dieny,V.S. Speriosu, S. S. P. Parkin, B. A. Gurney, D. R. Wilhoit, and D. Mauri,
Pyhs. Rev. B 43, 1297 (1991). B. Dieny, V.S. Speriosu, S. Metin, S. S. P. Parkin, B.
A. Gurney, and P. Baumgart, J. Appl. Phys., 69 4774 (1991).

6) A. Fert and P. Bruno, and SS. Parkin, Ultra —Thin Magnetic Structures, Vol. 2,
Springer-Verlag, Berlin (1993).

7) T. Shanjo, Surf. Sci. Rep., 12, 49 (1991).

8) M.F. Gillies et al., J. Appl. Phys., 78, 5554 (1995).

9) RM. Levy, Solid State Physics Series, Academic Press, 47, 367 (1994).

10) W. P. Pratt Jr., S.F. Lee, J.M. Slaughter, R. Loloee, P.A. Schroeder, and J. Bass,
Phys. Rev. Lett. 66, 3060 (1991).

11) J. Bass, W. P. Pratt Jr., and P. A. Schroeder, Com. Cond. Mat. Phys., 18, 223 (1998).

12) H. P. Myers, Solid State Physics, Talor & Franncis, Bristol PA USA, p 374 (1990).

13) D. Halliday and R. Resnick, Fundamentals of Physics, 3ed edition, John Wiley &
Sons, Inc. New York p 791 (1988).

14) M. A. M. Gijs and G. E. W. Bauer, Advance in Phys, 46, 285 (1997).

15) X. Portier, A. K. Petford-Long, P. Bayle-Guillemaud, A. C. Anthony, and J. A. Brug,

J. Magn. Magn. Mater. 198, 110 (1999).

17

16) A. E. Berkowitz, and Kentaro Takano, Magn. Magn. Mater. 200, 552 (1999).

17) S. Maat, L. Shen, and G. J. Mankey, Phys. Rev. B, 60, 10252 (1999).

18) G. R. Harp and S. S. P. Parkin, Appl. Phys. Lett. 65, 3063 (1994).

19)E.E. Fullerton, M.J. Conover, J .E. Mattson, C.H. Sowers, and SD. Bader, Appl.
Phys. Lett. 63, 1699 (1993).

20) R. E. Somekh and C. S. Baxter, J. Crystal Growth 76, 119 (1986).

21)D.J. Smith, AR. Modak, T.A. Rabedeau, and S. Parkin, Appl. Phys. Lett., 71, 1480
(1997).

22) NJ. List, W.P. Pratt Jr, M.A. Howson, J. Xu, M.J. Walker, and D. Greig, J. Magn.
Magn. Mater. 148, 342 (1995).

23) I. K. Shuller, Phys. Rev. Lett. 41 336 (1980).

24) M. Ohering, The Materials Science of Thin Films, Academic Press In., San Diago,
(1992)

25) C. P. Flynn, J. Phys. F (Metal Phys.) 18, L195 (1988).

26) E. Bauer, J. Phys: Condens. Matter 11, 9365 (1999).

27) T. Wagner, J. Mater. Res., 13, 693 (1998).

28) P. Bodeker, A. Abromeit, K. Brohl, P. Sonntag, N. Metoki, and H. Zabel, Phys. Rev.
B, 47, 2353 (1993).

29) M. T. kief and W. F. Egelhoff, Jr, Phys. Rev. B, 47, 10785 (1993).

30) J. Mayer, G. Gutekunst, G. Mobus, J. Dura, C. P. Flynn, and M. Rtihle, Acta Metall.
Mater., 40, S217 (1992).

31) J. de la Figuera, J. E. Prieto, C. Ocal, and Miranda, Phys. Rev. B, 47, 13043 ( 1993).

18

32) J. Carnarero, L. Spendeler, G. Schmidt, K. Heinz, J. J. de Miguel, and R. Miranda,
Phys. Rev. Lett. 83, 2448 (1994).

33) B. Wolﬁng, K. Theis-Brohl, C. Sutter, and H. Zabel, J. Phys: Condens. Matter, 11,
2669 (1999).

34) B. L. Adams, S. 1. Wright and K. Kunze, Metallurgical Transaction, 24A, 819 (1993).

35) s. 1. Wright, J. Computer-Assisted Microscopy, 5, 207 (1993).

36) DJ. Dingley, Scanning Elect. Microsc., IV 273 (1981).

37)N.C. Krieger Lassen, D. Jul Jensen and K. Conradsen, Scanning Microscopy, 6, 115
(1992)

38) A.J. Wilkison and DJ. Dingly, Acta Metall. Mater., 40, 3357 (1992).

39) A. E. Berkowitz and K. Takano, J. Magn. Magn. Mater. 200, 552 (1999).

19

CHAPTER 2

EXPERIMENTAL PROCEDURES

2.1 INTRODUCTION

The focus of this study was to design and develop a sputtering system to grow
epitaxial magnetic multilayers and spin-valves, and to optimize techniques for epitaxial
growth. The sputter deposited samples were characterized using different techniques
including AF M, EBSP, CPP GMR, and magnetometry. These techniques have provided
information about the orientation, orientation distribution, and magneto-transport
properties of epitaxial ﬁlms, multilayers, and spin-valves.

This chapter is divided into four sections: apparatus design and development, growth

techniques, structural characterization, and magneto-transport characterization.

2.2 APPARATUS DESIGN AND DEVELOPMENT

There are a number of different ﬁlm deposition techniques. Molecular beam epitaxy
(MBE), thermal evaporation, metal organic chemical vapor deposition (MOCVD),
chemical vapor deposition (CVD), and sputter deposition are among the most widely
used. Sputter deposition is divided to four branches: ion deposition, ion-assisted
deposition, DC & RF magnetron deposition, and RF & DC triode—magnetron deposition.

An ultra-high vacuum, dc-magnetron (triode) sputtering system, located in the Physics
and Astronomy Department at Michigan State University, has been modiﬁed for this

study. An existing room temperature chamber top assembly has been used in the past to

20

grow polycrystalline and amorphous ﬁlms and multilayers. A new high temperature
chamber top assembly has been designed and built for this study to grow epitaxial ﬁlms
and multilayers and regular polycrystalline ﬁlms and multilayers. The main emphasis
was on the design of two substrate heater assemblies. One has both cooling and heating
capability, with substrate temperatures ranging from below room temperature to about
600°C. The second one was designed for high temperature sputter deposition, with
substrate temperatures ranging from 600°C to 1200°C. Each heater assembly can be
lowered onto given substrate when heating is required. A brief description of de-
magnetron sputter deposition technique and the design of both tops is included in this

chapter.

2.2.1 Sputtering:

A typical dc-magnetron (triode) sputtering gun system consists of three electrodes: an
anode, a cathode, and a target. The sputtering chamber must be evacuated usually to a
base pressure of 1x10'8 torr or lower. Lower pressure will provide cleaner sputtering
environment (lower partial pressure of water vapor, N2, 02 and etc.) and therefore cleaner
ﬁlms (free from impurities). Sputtering takes place in puriﬁed argon gas at various
pressures, typically 2-5 xlO'3 torr. Upon applying a high current to the ﬁlament
(cathode), the heated ﬁlament emits electrons that accelerate toward the anode. These
accelerating electrons ionize the Ar gas molecules in their paths creating positive ions.
The target is located between the anode and cathode and the whole assembly is coupled
to a pair of parallel magnets. Applying a negative voltage to the target creates a potential

difference between the positive ions and the target, and the ions accelerate toward the

21

target (ion bombardment). During bombardment, the ions physically remove atoms from
the target surface. The neutral ejected atoms ﬂow toward the substrate where they
adhere. Each sputtering gun assembly is water-cooled. Figure 2.1 give an overall view
of the triode-sputtering system.

The following are advantages of triode sputtering over other deposition techniques:

0Both conductive and semiconductive materials can be deposited.

OHigher deposition rates are possible.

oThe small ratio of residual gas molecules to sputtered atoms results in higher
ﬁlm densities and fewer impurities.

0The ﬁlm composition is uniform and usually very similar to the
target composition.

OPrecise control of the deposition rate is possible by independent regulation of the
plasma and/or target voltage.

OMaximum usage of the target material (up to 90%) is possible.

The sputtering system used for this study contain four independent Simard triode
sputtering guns with a 2.25" x 0.25" (diameter x thickness) target size and a regular diode
magnetron gun with a one inch target size. The ﬁve gun assemblies provide a capability
of making a variety of multilayer samples. Figure 2.2 shows top views of the gun
assemblies. A mechanical pump and a cryo-pump are used for evacuation. The plasma
voltage, target current, and deposition rate of all sputtering guns are monitored by a PC

running a LabVIEW program.

22

 

Figure 2.1 Front views of the MSU sputtering system with four dc-triode guns and a
magnetron gun capability, including the related power supply and the residual gas
analyzer.

23

   
 

Small Magnetron

Wobble Stick

 

(C)
Figure 2.2 Top views of: a) an unassembled triode gun, b) four assembled guns with
FeMn, Cu, Py, and Nb targets, c) four chimneys in the open position. Note that after
deposition 3 computer controlled stepper motor rotates the chimneys to the closed
position to protect the samples.

24

2.2.2 Room Temperature Chamber Top Assembly:

The room temperature chamber top assembly is hemispherical in shape and is made of
stainless steel. This chamber top consists of the following components: cold trap, cooling
system, thermocouples, shaﬁ connected to a stepper motor, sample positioning and
masking apparatus (SPAMA) plate, ﬁlm thickness monitor (FTM), and substrate holders.
Figure 2.3 shows photographs of the overall system and an inside view of the room
temperature chamber top assembly. The cold trap is essential for freezing any remaining
water vapor. Liquid nitrogen (LN2) for the cold trap ﬂows continuously during sputter
deposition. The cooling system consists of a capillary tube that passes through the LN2 in
the cold trap. Dry nitrogen gas at 1000 psi pressure ﬂows through the capillary tube. By
controlling the dry nitrogen gas pressure, the substrate can be cooled to -100°C. Three
K-type thermocouples (chromel-alumel) monitor the substrate temperatures during
deposition. The SPAMA plate assembly is connected to a computer-controlled-stepper
motor through a shaft. The stepper-motor is on the outside of the chamber, and the shaft

goes into the chamber through a rotary feed-through.

2.2.3 High Temperature Chamber Top Assembly:

The high temperature chamber top assembly was designed and assembled by
Professor W. Pratt of Physics and Astronomy Department and the author. As brieﬂy
described in the introduction (section 4), two substrate heater assemblies were designed.
One with both cooling and heating capability, ranging from below room temperature to
about 700°C. The second, for high temperature growth, ranging from 700°C to 1200°C.

The general principle behind the design of these heaters was based on two heating

25

 

Figure 2.3 (a) A top view of the room temperature chamber top assembly. (b) A side
view of the SPAMA plate, including the CPP and CIP substrate holders and FTMs.

26

techniques: (1) Heating by radiation (photons) for 700°C -1200°C and (2) Heating by
conduction (phonons) for below 700°C. The detail of design and operation of both
heaters will be discussed in next sections.

Most of the chamber top parts were manufactured in the machine shop of the Physics
and Astronomy Department at Michigan State University. The high temperature chamber
top assembly was based on a ﬂat plate made of stainless steel. The top provides the
following:

OOne 1.5" UHV viewport.

0Two wobble sticks; each mounted on a 2%" ‘ConF lat’ (CF) ﬂange, for lowering
the heaters onto the substrates or raising the heaters when not in use.

OOne 2%" CF ﬂange which contains a BNC connection for a ﬁlm thickness
monitor and two high-pressure gas lines for cooling the low temperature heater.

OOne 2%" CF ﬂange which contains electrical feed-through providing current
to both heaters.

OOne 2%" CF ﬂange which contains liquid Nitrogen lines for cold trap.

OOne 2%" CF ﬂange which contains water lines for cooling the Cu heat shield
of the high temperature heater.

OOne 2%" CF ﬂange which contains feed-through for both C-type and
K-type thermocouples

OOne stepper motor for rotating the SPAMA plate.

Figure 2.4 shows some of these accessories of the high temperature chamber top

assembly.

27

 

Figure 2.4 A top view of the high temperature chamber top assembly including the
wobble sticks (WS), FTM ampliﬁer (FTM), stepper motor (SM), and some of the feed
throughs.

28

High Temperature Heater:

The high temperature chamber top assembly was designed for growing epitaxial ﬁlms
and multilayers. In epitaxial growth, substrate temperatures in the range of about 200 to
1200°C are required, depending on the sputtered element [1]. In this study, an epitaxial
Nb ﬁlm must be ﬁrst deposited on a sapphire substrate. However, a pre-deposition

annealing of the sapphire substrate at 1200°C was required before the Nb ﬁlm was

deposited at about 700-900°C. Therefore, the high temperature heater (HTH) was
designed and dedicated to annealing sapphire substrates and depositing epitaxial Nb ﬁlms
by radiation technique. The heater element (VXPBNlA) was purchased from Kurt J.
Lesker Company, Clairton, PA and consisted of a ﬁlm of carbon (graphite) sandwiched
between two thin layers of pyrolytic boron nitride. At full power, this heater could reach
1700°C. A basic description of HTH assembly is provided in Figure 2.5 (a). As shown
in this ﬁgure, the heater element was held by two long molybdenum (Mo) screws
(insulated), which also act as the electrodes. There were two heat-shields, made out of
0.005” Mo foil, surrounding the heater element and the substrate holder assembly (when
the HTH was in its lowered position). There were two C-type thermocouples, one on the
back of heating element and one in front. When the heater was lowered, the front
thermocouple came in direct contact with the back of the substrate,. Upon lowering, the
separation between the heating element and the substrate was about 1-2 mm. The heating
then took place through radiation from the white-hot heater (photons). However, the
sapphire is transparent to the photons. Therefore, a thin layer (~200 nm) of Nb must be
deposited on the backside of substrate to block the photons and provide a uniform

temperature during deposition on the front (polished side) of the sapphire substrates. The

29

High Temperature Heater Assembly

Thermocouples (C-type)

 
  
  
 
 
     

+1 _' Mo Screw
Heater Holder _ . /-> Mo Nut
(Insulator) “ :7
-------- pun-b Insulators
,"I
I
/ cl
5
Water Line § — _ J

Water-cooled
Cu Heat Shield

'4.

 

 

“ -- H . .
Mo. Heat Shields¢:::::;l_-_:| > enter

1
....I'...
Guide (A1) 4 ------

V
Mo substrate assembly

Figure 2.5(a). A schematic of the high temperature heater assembly in its lowered
position. The assembly slides into the AI guide, on top of the substrate holder.

30

temperature of substrates during high temperature annealing and deposition of Nb ﬁlms
were measured by using the front thermocouple. In addition, an optical pyrometer was
used for calibrating and checking the accuracy of this thermocouple for temperatures
above 700°C. The optical pyrometer could be aimed onto the front of the heated
substrate through a sapphire window located on the bottom of the sputtering chamber.
Having a Nb ﬁlm on the backside of sapphire substrates ensured that the pyrometer
accurately measures the temperature of the substrates. Due to the high temperatures
(high T 2 700°C), a water cooling system was provided for HTH assembly. The water
lines were ﬂexible but limit the mobility of the heater in its lowered position to i 90°
rotation of the SPAMA plate. A vertical two-stage wobble-stick was used for lowering
and raising the HTH. For deposition of epitaxial multilayers, the heater had to be able to
move with the substrate over several targets. Therefore, a second heater with larger

mobility and lower temperature range was desired.

Low Temperature Heater:

A second heater (a low temperature heater = LTH) was designed with an angular
mobility of 270° for the SPAMA plate. The LTH is dedicated for growing epitaxial ﬁlms
and multilayers that required lower growth temperatures (below room temperature to
600°C). LTH heater is more accurate than the HTH in this temperature range, because
the substrate is heated by conduction (phonons) via direct contact between the heater and
substrate. A basic description of LTH assembly is illustrated in Figure 2.5 (b). As shown
in this ﬁgure the heating element (same as the heating element in the HTH) is in direct

contact with the stainless steel (SS) jacket. The two Mo screws (insulated) act as heater

31

Low Temperature Heater Assembly

Thermocouples (C-type)

Mo Screw (elecgode)

   
  
 
   

Insulator 4 ----------------- ,0

High Pressure _ _
N2 Gas Inlet

 

- - ‘> Heater

SS Jacket
with ‘
Gas Reservoir

 

I-

SS Heat Shield

v
Mo substrate assembly

Figure 2.5(b) A schematic of the low temperature heater assembly in its lowered
position. This heater is in direct contact with the substrate and capable of being cooled
from 600°C to about room temperature within a few minutes, using cooled-high pressure
nitrogen gas.

SS = Stainless Steel

32

electrodes and hold the heating element in direct contact with the jacket. The LTH SS
jacket could be cool down by high-pressure nitrogen gas (capillary tube) from 600°C to
about room temperature in a few minutes. There were two c-type thermocouples in LTH
assembly, one right on the top of heater element and the other was inside the bottom wall
of the jacket. The thermocouple in the lower part was calibrated by using a third
thermocouple (K-type), which was attached to the front of the test-substrate (using silver
paint). A second vertical, two-stage, wobble-stick eas used here to lower and raise the
LTH. Figure 2.6 shows two different side views of the HT top, including pictures of both
types heaters. In Figure 2.7, a top view of the substrate holders, including the heat shield
guide, and a high temperature molybdenum (Mo) mask is shown. The substrate holders
were designed to avoid heating the SPAMA plate during annealing and deposition. Four
Mo rods (0.035” OD) are used to keep each substrate holder 1/2 inch above the SPMA
plate. All the holders are also made from Mo plate (0.042”). A stainless steel sliding-
shutter keeps the substrate and ﬁlms from being contaminated by sputtering. Four C type
(tungsten-rhenium) thermocouples monitor the temperatures of the heaters, substrate, and

SPAMA plate. The thermocouples are either 0.005" or 0.01" in diameter with a useful

temperature ranges of 0-2320°C: +Lead: W(5% Re) and -Lead W(26% Re).

2.3 THE GROWTH TECHNIQUES

2.3.1 Polycrystalline Films and Multilayers:

All of the polycrystalline ﬁlms and multilayers were sputter deposited in the
temperature range of —30°C -to- +30°C. The deposition rates varied from 5 tol3 A/s,

depending on the target material. Several polycrystalline ﬁlms and multilayers were

33

HTH
water line

 

(b)

Figure 2.6 Side views of high and low temperature heaters, sliding shutter, cooling water
line, and other accessories.

34

Cold trap

   
   
 

  

Substrate holder
(Polycrystalline samples)

 

 

Figure 2.7 (a) A top view of the Molybdenum substrate holders, (b) and a schematic top

view of the Molybdenum mask for deposition of magnetic multilayers and spin-valves.

35

grown at low temperature for comparison to similar epitaxial ﬁlms and multilayers. The

results obtained form these comparisons are discussed in chapters (4-6).

2.3.2 Epitaxial Films and Multilayers:

To achieve fully epitaxial growth in multilayers and spin valves containing
Nb(110)bcc, CO(111)fcc, CU(111)fcc, and Py(111)rcc layers, it is necessary to establish
optimum deposition conditions for each subsequent layer. Because of reactivity between
Co (and possibly Py), with Nb, Co and Py should not be deposited directly on Nb [2,3].
Therefore the sequence of Nb-Cu-[Co-Cu]N or Nb—Cu-Py-. .. is preferred.

Growth conditions in a preferred orientation: To grow epitaxial CPP multilayers
and spin-valves in the preferred (111) orientation (such as Cu/Co(111) multilayers or

FeMn-Py-Cu-Py(l l 1) spin-valves), the epitaxial Nb contact layer must ﬁrst be grown in

a (110).,cc orientation [4]. (1120) oriented A1203 has been reported to be a suitable

substrate for growing an epitaxial (110) Nb ﬁlm [5]. In the present study, sapphire

substrates with one side polished to within 0.3 degree of the (1120) A-plane were
purchased from Union Carbide.

Substrate cleaning: Most of the substrates were prepared in the sputtering room,
cleaned with acetone and then alcohol in an ultrasonic bath. Others were ﬁrst cleaned in
a combination of boiling deionized (DI) water and detergent (Alconox) in an ultrasonic
bath, then rinsed in boiling deionized water, rinsed again and mechanically cleaned in
acetone using cotton swabs, rinsed in isopropyl alcohol, and ﬁnally rinsed with DI water
and dried with industrial grade nitrogen gas. These latter substrates process were all in a

clean room.

36

Substrate annealing: After reaching the desired vacuum (~1 x 10'8 torr), the
substrates were pre-deposition annealed at 1200°C for twenty minutes to remove any
residual organic layer and to decrease the surface roughness.

Deposition process: To determine the optimum growth temperature, Nb ﬁlms were
sputtered using substrate temperatures ranging from 25 to 1200°C. To improve the
surface smoothness, some of the sputtered ﬁlms were post-annealed at 950°C for ten
minutes. The high temperature heater (HTH) was then turned off and allowed to cool-
down while the low temperature heater (LTH) was heated up to the desired temperature
in range of ISO-500°C. Upon removal of the HTH, the SPAMA plate was rotated to the
position for loading the LTH. Cu ﬁlms were then deposited onto the Nb at temperatures
ranging from ambient temperature to 500°C. Some of these samples (grown at low
temperatures) were then post-annealed in a temperature range of 250 to 500°C. Finally,
the Co, Py, and FeMn ﬁlms were deposited at desired temperatures ranging from room
temperature to 460°C.

Deposition rates: The deposition rate for each element was measured using a quartz
ﬁlm thickness/rate monitor. A typical deposition rate was 3-4 A/s for most samples.
However, samples were also grown with different deposition rates for comparison

purposes.

2.4 STRUCTURAL CHARACTERIZATION
Epitaxial ﬁlms were structurally characterized using atomic force microscopy (AF M)
and electron backscatter diffraction patterns (EBSP). EBSP is capable of measuring

crystallographic orientations of single-crystal and/or large grain polycrystalline materials

37

with a spatial resolution of approximately 1° and a detection depth of approximately 200
nm [6-9].
In this section a brief description of both instruments that were used for structural

characterization of epitaxial samples are reviewed.

2.4.1 Atomic Force Microscopy

Description: A Digital Instruments atomic-force microscope (AFM) with tapping
mode capability was used to image the surfaces of the ﬁlms. This microscope has three
operational modes: contact, tapping, and lateral force modes. Each mode measures
different forces that act between the tip of cantilever and atoms on the surface of
specimen. One of the most important forces is the interatomic force called the van der
Waals force. This force is strongly dependent on the distance between two atoms and
consists of a repulsive force and an attractive force. The operation of the contact and
tapping modes are based on the repulsive part of the interatomic force as the cantilever
approaches the surface of the specimen. In the tapping mode, the cantilever alternately
contacts the specimen surface and lifts the cantilever off the surface to avoid dragging the
tip across the surface of the specimen. A schematic of a AFM microscope is shown in
Figure 2.8 (Courtesy of Digital instruments [10]).
Sample preparation: Samples for atomic force microscopy did not require any speciﬁc
sample preparation. The samples must be clean and free of dust. A new cantilever

enhances the resolution of the images.

38

 

 

     

Oscillation Piezo

    
 
   
 

Photo-
detector

 

Silicon Cantilever

Sample with Integral Tip

 

 

 

Figure 2.8 Block diagram of the components used in D1 atomic force microscope with a
tapping mode capability. (Courtesy of Digital Instruments [5])

39

2.4.2 Electron Backseatter Diffraction Pattern:

Description: Electron backscatter patterns (EBSP) is a scanning electron microscope
(SEM) diffraction technique for measuring crystallographic orientations of single-crystal
and/or polycrystalline materials. An EBSP consists of many intersecting, linear features
called Kikuchi bands. These bands have slightly higher intensities with respect to a non-
uniform background. This contrast is due to the diffraction of backscattered electrons by
the crystal [8].

When an electron beam is incident on a sample, different interactions take place. The
emission of secondary electrons, backscattered electrons, x-rays, and Auger electrons are
among the most important results of these interactions. The electrons that impinge on the
specimen surface (in all directions) go through inelastic scattering, losing about ~1% of
their energy, and establish an interaction volume in the specimen. Near the surface, those
backscattered electrons that satisfy the Bragg diffraction condition for a particular set of
planes are channeled differently than the non diffracted ones and produce a change in
intensity. These diffracted electrons are called the backscattered diffracted electrons
(Figure 2.9 (a)). The diffracted backscattered electrons from each set of planes form a
cone with its axis normal to the plane. The number of backscattered electrons depends on
the atomic number, Z, of the specimen. The relationship between these two variables is
given by the series:

it = 11355 / 713 = -0254 + 0.0162 — [1.86x10”] 22 + .......
where n is the backscatter coefﬁcient, 11335 is the number of backscatter electrons, m3 is

the number of incident electrons, and Z is the atomic number [6].

40

mernsm
IEAI
1. I.

wn.:
SURFACE

 
 
   

  

2 Silicon
Unit Cell

 

 

 

 

 

Figure 2.9 a) Schematic representation illustrating the inelastic collisions, Bragg
diffraction, and resultant formation and imaging of EBSPs as a result of an incident
electron beam. Adopted from [6]. b) A diagram of three set of band from {220} plane
that are projected onto the phosphor as a pair of cones.. c) Experimental set-up for EBSP
measurements. Figures (b) and (c) are adopted from [10].

41

The number of backscattered electrons escaping from the surface of specimen, is also

1
(1+ costly/2"“2 .

 

dependent on the tilt angle of specimen: 11(0) =

Usually, a tilt angle of approximately 0 = 70° with respect to the electron beam is
optimum. Figure 2.9(b) shows a schematic illustration of the diffraction cone and the
pattern formation on the phosphor screen placed in front of the sample. A diagram of the
EBSP system set-up is shown in ﬁgure 2.9(c).

A CamScan 44FE FEG-SEM was used to produce backscattered electrons. An
accelerating voltage of 25 kV, emission current of approximately 120 uA, objective
aperture of 50pm, and a working distance of 22 or 21 mm was used for this study. A
CamScan ORTEX CCD video camera images the phosphor screen. The imaged patterns
were then transferred to a Harnamatsu ARGUS-lO image processor which subtracts the
background noise from the diffracted patterns. The processed patterns were then saved
and analyzed using an interfaced PC. The detection and indexing of the Kikuchi bands
can be performed either manually or automatically using the Channel 4 software package
[11]. This software determines the crystal orientations and displays the orientation data
in the form of pole ﬁgures and inverse pole ﬁgures. The Channel 4 orientation software
was also used to display crystal grain orientation maps of desired areas, which have been
scanned by moving either the electron beam or the sample (stage). The time required for
the collection of a set of data varied from two hours to six days, depending on the length
and step size of the scan. The spatial resolution of the EBSP technique is approximately
1 urn, but it is limited to the volume a few tens of nanometers below the surface [9].

Figure 2.10 shows the CamScan FE44 SEM and its accessories.

42

EBSD
camera

 

 

 

Figure 2.10 (a) A view of the CamScan 44FE electron microscope column, (b) a front
view of main panel, (c) and PC monitor displaying an EBSP before and aﬁer indexing.

43

Sample preparation: No speciﬁc sample preparation is necessary for EBSP.
However, since EBSP is very sensitive to the sample surface quality; a dirty, rough,
and/or heavily oxidized surface can easily disturb the pattern. Therefore, care must be
taken to keep samples clean, as well as dust and oxide free. During this study, all
samples were kept in desiccators ﬁlled with a mixture of Ar 98% H2 2% gas.

Sample holders: Two different kinds of aluminum sample holders, one a half inch
round plate (which holds one sample) and a one inch by one inch square plate (capable of
holding up to four 1/2 inch x 1/2 inch samples) were used. Since the sapphire substrates
were insulators, silver paint was used to make electric conduction between the samples
and the Al sample holders. A piece of p-type (100) Si wafer, placed along side the

sample, was used to calibrate the EBSP system with respect to the X-Y-Z stage.

2.5 MAGNETO-TRANSPORT CHARACTERIZATION:

Special equipment was required for the measurement of the magnetic and magneto-
transport properties of magnetic multilayers and spin-valves (such as magneto-resistance,
magnetization, and coercivity of magnetic ﬁlms). Sections 2.5.1 and 2.5 .2 contain brief

discussions of how these magnetic properties were measured.

2.5.1 Magneto-Resistance (CPP-GMR measurements)
Description: CPP measurements require Nb superconducting contact layers on the
top and bottom of the multilayers and spin valves. To measure the very small resistance

(~10’7 Q) in CPP samples, a superconducting quantum interference device (SQUID) was

44

used. A detailed description of the CPP-GMR measurement apparatus has been reported
elsewhere [12].

Sample preparation: A molybdenum mask was used to sputter the epitaxial
multilayers and spin—valves in a L-shape (1 mm x 8 mm) pattern as was shown in ﬁgure
2.7(b). After sputtering the epitaxial samples, the top surface is often gold coated for
protection. A series of 75 and 6 pm square pillars are then fabricated using
photolithographic techniques. The details of the photolithography fabrication have been
reported elsewhere [13]. The spin-valve samples are then pinned in an oven ﬁlled with
Ar 98%H2 2% gas. In a typical pinning process, samples are heated up to ~200°C in the
presence of about 30 Oe magnetic ﬁeld. After cooling the sample, the ferromagnetic
layer is pinned to the adjacent antiferromagnetic layer through exchange-biased coupling.
The pinned direction, which is referred to as the easy axis, is along the applied ﬁeld. The
details of the pinning process are given in chapter 6.

2.5.2 Magnetization

Description: A Quantum Design SQUID magnetometer is used to measure the
magnetization of multilayers and spin-valves. This magnetometer has the capability of
making fast, precise measurements of the dc magnetic properties of magnetic multilayers
and spin-valves over a wide range of temperatures (2-350 K) and applied ﬁelds
(:1 Tesla, and 1 Tesla = 10000 06)

Sample preparation: The sample chamber of this magnetometer was cylindrical,
approximately 7 mm in diameter. The diameter of epitaxial ﬁlms were approximately
11mm x 11mm. Therefore, these samples were cut in half (~5 mm) and placed vertically

into the sample chamber. No other speciﬁc preparations were required.

45

2.6 REFERENCES:

1)
2)

3)

4)

5)

6)

7)

3)

9)

C. P. Flynn, J. Phys. (metal Phys), 18 L195 (1988).

K. Brohl, P. Bodeker, N. Metoki, A.Stierle, and H. Zabel, J. Crystal Growth, 127, 682
(1993)

R. E. Somekh and C. S. Baxter, J. Crystal Growth 76, 119 (1986).

W. P. Pratt Jr., S.F. Lee, J.M. Slaughter, R. Loloee, P.A. Schroeder, and J. Bass,
Phys. Rev. Lett. 66, 3060 (1991).

S. M. Durbin, J. E. Cunningham, M. E. Mochel, and C. P. Flynn, J. Phys: Met. Phys,
11, L223 (1981).

A. W. Gibson, M.S. thesis, Michigan State University, (1995).

DJ. Dingley, Scanning Elect. Microsc. IV, 273 (1981).

NC. Krieger Lassen, D. Jul Jensen and K. Conradsen, Scanning Microscopy 6, 115
(1992)

A.J. Wilkison and DJ. Dingly, Acta Metall. Mater. 40, 3357 (1992). 8) S. 1. Wright,

J. Computer-Assisted Microscopy 5, 207 (1993).

10) Digital Instrument, Instruction manual (1997).

ll) CHANNEL 4 orientation mapping manual; HKL Technology, (1997).

12) P. Holiday, Ph.D. dissertation, Michigan State University, (1996).

13) R. D. Slater, J. A. Caballero, R. Loloee, and W. P. Pratt, Jr., Submitted for publication

in J. Appl. Phys. (2001).

46

CHAPTER 3

GROWTH OF Nb FILM ON SAPPHIRE

3.1 INTRODUCTION

As outlined in chapter one, it is necessary to establish optimum growth conditions for
each subsequent layer to achieve epitaxial multilayers and spin-valves of CO(111)fcc,
CU(111)fcc, and PY(111)fcc. Furthermore, to measure GMR in magnetic multilayers and
spin-valves with current ﬂow perpendicular to the layer planes (CPP), a Nb contact layer
must be grown prior to the deposition of the multilayer [1]. A (111) growth direction in
the fcc magnetic multilayers and spin-valves can be achieved by using epitaxial
Nb(110)bcc substrate [2-4]. Therefore, the epitaxial Nb layer is the foundation of the
epitaxial magnetic layer system, and the growth and characterization of this Nb ﬁlm is an
important step towards building high quality epitaxial magnetic structures. In this
chapter, the growth of epitaxial Nb on a single crystal of sapphire is discussed. This
discussion is followed by the crystallographic characterization of sputter deposited Nb

ﬁlms, using AFM and EBSP.

3.2 BACKGROUND

The epitaxial growth of Nb ﬁlms on A1203 substrates (metal-ceramic), using several
different techniques such as MBE and sputter deposition, has been extensively
investigated [2-20]. In these studies different characterization techniques have been used
to investigate the nucleation, the growth, and the interfaces of the Nb ﬁlms grown on

different orientations of sapphire.

47

Nb and A1203 have very similar thermal expansion coefﬁcients and Nb thin ﬁlms
grow on A1203 almost strain free with a highly quality crystal structure and large
coherence lengths [11,13]. Crystallographic data and some of the thermal properties for
both Nb and A1203 are summarized in Table 3-1. Orientation relationships are deﬁned by
the interface planes and two parallel directions in the two materials. In the Nb/Al203

system, different orientation relationships have been observed (for examples see [2,9]):
(0001)Al203 ll (111)Nb with [2II0]AI2O3 || [lOI]Nb
(0110)Al203||(I21)Nb with [0001]Al2O3 || [111]Nb

(1120)Al203||(101)Nb with [0001]Al203 || [111]Nb
(01I2)Al203 || (010)Nb with [21101711203 || [lOI]Nb

Because of the above orientation relationships, the growth direction of epitaxial Nb
ﬁlms can be selected by choosing the properly matched sapphire substrate.

In this study an epitaxial Nb(110) underlayer (contact layer) was required for growth
of the subsequence epitaxial magnetic multilayer structures and spin-valves in a (111)fcc
orientation. It has been shown that bcc Nb(110) planes can be grown epitaxially on
A1203 (1120) substrates [2, 3, 9]. The in-plane epitaxial relationships are:

[0001]Al2O3 H [111]Nb
[1010]A12O3||[121]Nb
A portion of the corundum (sapphire) cell and atomic packing of a (1120) plane of

sapphire are shown in Figure 3.1 [21, 22].

48

Table 3-1

 

 

 

 

 

 

 

 

 

 

CRYSTALLOGRAPHIC DATA

Element: A1203 - Corundum - Sapphire Nb—Niobium-Columbium
Crystal system: Hexagonal-Rhombohedric * Cubic — bcc

= 0.4763 nm 0.33067 nm
b = 0.4763 nm 0.33067 nm
e = 1.3003 nm 0.33067 nm
Unit cell volume: 0.2555 nm3 0.0359 nm3
# of atoms per unit 12(Al) + 18(0) = 30 2
cell:
Atomic radius: A1 = 0.143 nm Nb = 0.143 nm
Ionic Radius .413+ = 0.051 nm Nb+4 = 0.074 nm

02- = 0.132 nm

Symmetry: A single 6-fold rotation axis Four 3-fold rotation axes
Melting Point: 2052 °C - 2469 °C
Thermal expansion 5.4 x IO'WC'I || c-axis RT
coefﬁcient: 6.2 x 10'6 C'1 I c-axis RT 7.1 x 10'6 Cl

 

 

 

 

*Oxygen ions in a hexagonal close packed arrangement.

*Aluminum ions occupy the octahedral interstices.

49

'Oxxgrn
<—’/— \ /

Al"

 

 

/
4

62/

 

(a)

Figure 3.1 (a) A portion of the corundum cell and (b) atomic packing of a (1120) plane
of sapphire. One of the forms of alumina, A1203, has the corundum crystal structure,
which is similar to a hexagonal close-packed structure. There are 12 A1 ions and 18 0
ions associated with each unit cell. The lattice parameters of alumina are a = 0.475 nm
and c = 1.299 nm. The ionic radii ofthe Al 3+ = 0.051 and the O 2' = 0.132. There are 6
A1203 groups with 12 Al ions and 18 0 ions present in a hexagonal prism having the
given lattice parameters.

50

3.2.1 Growth of Nb Film on A1203

Theoretically, growth and nucleation of Nb on sapphire should follow one of three

basic growth modes. These three well-known modes are [23]:

1)

2)

3)

The Volmer—Weber (VW) mode (island), which occurs when the smallest
stable clusters nucleate on the substrate and grow in three-dimensions to form
islands. The growth of islands is more common when there is stronger
bonding between the deposited atoms than deposited atoms to the substrate.
The Frank-van der Merwe (FM) mode (layer-by-layer), which occurs when
the smallest stable clusters nucleate on the substrate and grow in two
dimensions to form a planar sheet (continuous layer). This type of growth
mode occurs when there is strong bonding between the atoms and the
substrate and is commonly seen in single crystal epitaxial growth of
semiconductor ﬁlms.

The Stranski-Krastanov (SK) mode (layer+ island): In SK growth mode, after
the formation of a few monolayers (due to strong bounding between the atoms
and the substrate), the surface energy changes and bonding between
subsequent layers decreases with simultaneous increases in bonding between

the atoms, resulting in three-dimension island formation.

Metal atoms typically do not wet insulating substrates and bond more strongly to each

other. Due to this bonding, nucleation and growth of the metal on the insulator usually

occurs in the VW (island formation) mode [24, 25]. When three-dimension islands form,

the island size increases in three dimensions, until coalescence of the islands occurs,

leading to a continuous ﬁlm. If all of the islands grow with the same orientation or rotate

51

to the same orientation during coalescence, a single crystal will form. In practice, a
number of factors affect the nucleation and growth of crystalline ﬁlms. Among the most
important factors are the nucleation rate, the substrate temperature, deposition rate, and

the surface quality of the substrate or buffer layer [23].

The nucleation and growth of Nb(110) on the Al2O3(1120) substrates seems to be

unusual and different from the typical growth of metals on insulator substrates. Mayer et

al. have shown that the Al2O3(1120) surface is oxygen terminated and creates adsorption
sites for the Nb atoms [1 l]. The results reported by a few other investigators indicate that
a strong anisotropic bonding exists between Nb ﬁlms and A1203 substrates, which is due

to one adsorption per site for the Nb atoms in the ﬁrst monolayer [15].

3.2.2 Grain Structure of Sputtered Metal Films

The temperature of the substrate and the argon pressure in the sputtering systems will
affect both the sticking coefﬁcient (the ratio of the number of atoms adhering to the
number of atoms striking the substrate) and ﬁlm structure. The structural morphologies
of deposited ﬁlms at different temperatures were described by Movchan et a1. [26] using
a zone structure model [23, 26]. In ﬁgure 3-2(a), the inﬂuence of both argon pressure
and substrate temperature on microstructure of sputtered metallic ﬁlm is shown [26]. In
part (b) of this ﬁgure, four different zones as a function of temperature are illustrated
[23]. In this ﬁgure, the horizontal axis is Ts/T M where T5 is the substrate temperature and

TM is the melting temperature of the deposited metal.

52

Transition Structure columnar Grains

   
   

Consisting of Extending Through
Densely Packed Coating Thickness
WW °I 739"“ Grain Structure

Crystallitee Separated
byVoids . '
it? 201»: m .
El?" ‘7'," 1
' 333,7, 4 i. '1. ‘1‘: U;
H.“ ‘ ‘ ! j‘ll’ 1.0
. , .l o.
,l. 0.8
'I 0.7
4 Substrate
Argon Temperature (T /T 1
Pressure 1
lmtorr)

(a)

zone 1: : zone In
I

20115 I: 2014: r
I

 

 

 

 

 

l
l
I
GRAIN: I onser or [cxrcuswe
RENUCLEATION: ' GRANULAR .GRAm
. : EPITAXY lGROWTH
I . l
I l l
l 1 I ——
‘ 0' 6" «0‘4" 0" 9 ﬂ
33‘ an" 5:. 39,. l
10'35'53'21‘ '32-“ 59.1"
‘1‘ 9 3" ..;‘:'n ‘
‘9'“ “.!‘...-ﬂu\..g'!£. Lemon . b _..-. c“. 9 A. o-.-- I

 

DJ 0.2 0.3 0.5 0.7
0T
S/TM
(b)

Figure 3.2 (a) Effect of argon pressure and substrate temperature on microstructure of
sputtered metallic ﬁlms (adopted from [26]). (b) Zone model for evaporated metal ﬁlms
(adopted from [26]).

53

3.3 RESULTS
3.3.1 Growth Process

As was described in section 2.3.2 the sapphire substrates used in the present study
were pre-deposition annealed at 1200°C. Sputtering of the Nb ﬁlms was carried out
using substrate temperatures ranging from 25 to 1200°C. Surface smoothness of some of
the sputtered ﬁlms was improved by post-annealing the ﬁlms at 950°C and 1100°C. All
ﬁlms were grown in 2.5><10'3 Torr Ar, with a chamber base pressure of ~1 x10"8 Torr and

the deposition rate for Nb was 3 A/s. A list of epitaxial Nb ﬁlms is tabulated in appendix

3.A.

3.3.2 AFM Analyses

A series of 250 nm thick Nb ﬁlms was grown on Al2O3(1120) at different substrate
temperatures. The surface topography of these ﬁlms was imaged using a Digital
Instruments atomic force microscope (AF M). In general, the quality and topography of
sputtered Nb ﬁlms grown on A12O3(1 120) is affected by the substrate temperature.
Figure 3.3 shows AF M images and line trace surface analyses of a Nb ﬁlm grown at: (a)
room temperature, (b) 450°C, (c) 660°C, (d) 750°C, and (e) 1100°C which illustrates the
roughness of these ﬁlms. Note that samples deposited at substrate temperatures of 450°C
and higher were post deposition annealed at 1100°C for 10 minutes.

The samples deposited at ambient temperature (not annealed with Ts/TM = 0.01) were
characterized by a rough granular topography suggestive of a polycrystalline structure.

The average size of the topographic futures (grains) in the Nb ﬁlms grown at room

temperature were approximately 125 nm x 50 nm. The image root-mean-square (RMS)
54

 

 

 

5.0
2.50

 

Mﬂvvﬂvnl WHEN“ n WVRUAWA AMI/“ill U flu/Vivi“

ﬂ>

 

0

-2.50

 

 

 

 

 

 

 

1.00

 

 

 

-l.00

 

 

 

 

 

-2.00

 

n
a
N
M
a
8
9
q-
m
n
8
_
31

(C) (d)
Figure 3.3 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of Nb ﬁlms deposited at substrate temperatures of 25°C (No PDA),
450°C (PDA=1100°C), 660°C (PDA =1100°C), and 750°C (PDA=1100°C), with image
RMS roughness of 3.17, 2.52, 1.25, and 0.35 nm, respectively.
(PDA = Post Deposition Annealed)

55

 

 

 
  
  
   
 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
 
 

       
  
  
  
 
 

I. 7.731 III
MS 11.845 M

In DC
llama) 8.836 he
than 88.613 run
I: 31.084 nu
R1 CM 8

Iadlul 5.356 In:
Si’ua 113.85 nu

  

 

“2.2.--“ _._. _.__..
Surface dlatame 7 745 :

Hertz distance“) 7.734 m- f
Uert distance 3.303 me i

Surface distance [
Hort: distance 1
Uert distance '
M910 ‘
Surface distance '
Maria distance .
Uert “stamp

Spectral period DC
Spectral free 0 Hz ,
Spectral MS any 5.535 M i

h!

    
 

Figure 3.3(e) Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of Nb ﬁhns deposited at substrate temperatures of 1100°C with
image RMS roughness of 33.1 nm. The epitaxial Nb ﬁlm grown at substrate temperature
of 1100°C was found to be very rough, often containing many holes.

56

roughness was about 3.8 nm. Note that roughness measurements were performed over
entire individual images and are referred to as image RMS roughness. The line traces in
all AFM images are only representative of a single trace not the entire image.

At a substrate temperature of 450°C (Ts/TM = 0.18), the surface topography suggested
some alignment and possibly collimated grain formation. This could be due the post
deposition annealing and a TEM cross-sectional study can clear up this issue. At this
point, however, the effect of growth temperature seems to be the main factor, in
agreement with zone T in Figure 3.2. The average topographic futures in this ﬁlm varied
from 45 to 55 nm with an image RMS roughness of 2.52 nm. At a substrate temperature
of 660°C (Ts/TM = 0.26), oriented steps, which may be structural ledges, run across the
surface with an image RMS roughness of 1.25 nm. The steps have an average height of
approximately 0.50 nm, equivalent to two (110) interplaner spacing of Nb (\l 2 x 0.331
nm = 0.47 nm) The misorientation (0) in Figure 3.3(e) was calculated using the height
of the structural ledges and number of them in one row within 1250 11m x 1250 nm
scanned area:

8 (steps in 1250 nm length) x 0.5 = 4 nm

Sin (0) = 4 nm /1250 nm .'.0 = 0.18 degrees

This misorientation is most likely due to miscut in Al2O3(1120).

At a substrate temperature of 750°C (Ts/TM = 0.3), the surface topography appeared to
be very smooth, free of any kind of grains and structural ledges (Figure 3.3(d)) resulting
in an image RMS roughness of 0.35 nm. At a substrate temperature of 1100°C, the ﬁlms

were found to be very rough, containing many holes. Figure 3.3(e) shows an AF M image

57

and line trace surface analyses of a 250 nm thick ﬁlm grown at a substrate temperature of
1100°C (Ts/TM = 0.46). The image RMS roughness is about 33 nm.

As shown in Figure 3.3(f) a sample deposited at 750°C and post deposition-annealed
at 950°C for 10 minutes had a better surface topography (smoothness) as a sample
annealed at 1100°C. Therefore, a substrate temperature of 750°C with post deposition
temperature of 950°C was chosen as an optimum growth temperature for Nb ﬁlms used
as a foundation for subsequent ﬁlms grown in this study.

It should be noted that the optimal growth temperature for growing epitaxial ﬁlms was

proposed by Flynn [27] to be determined by the following empirical equation:

3
T epitaxial growth (K) = (’8') T melting temperature (K)

which results in T epitaxial growth = 755°C for growth of epitaxial Nb ﬁlm. In practice,
however, there are other factors such as growth technique and deposition rate that can
affect the predicted growth temperature. The optimum growth temperature (750°C) of
the epitaxial Nb in this study was very close to the predicted growth temperature

The AFM results for these ﬁlms reveals only information about the surface
morphology. The AF M results do not yield crystallographic information about the
sputter deposited Nb ﬁlms. To obtain information about the orientation and crystalline

quality of these ﬁlms, the EBSP technique was used.

3.3.3 EBSP Analyses
EBSP is very useful for characterizing the orientation of sputtered ﬁlms. The patterns

from the Nb ﬁlms grown at a substrate temperature of 600°C or higher were very sharp

58

 

 

 

 

 

 

 

 

 

 

 

-5.0

(ii)

Figure 3.3(f) Atomic-force-microscopy images showing the surface topography of two
Nb ﬁlms deposited at substrate temperature of 750°C and post deposition annealed at (a)
1100°C and (b) 950°C with image RMS roughness of 0.35 and 0.16 nm, respectively.

59

and easily indexed as bcc. Due to the Nb thickness (th = 250 nm), the patterns did not
suffer from contributions from the underlying A1203. Figure 3.4 shows representative
EBSPs from Nb ﬁlms. The patterns obtained from a Nb ﬁlm grown at ambient
temperature [Figure 3.4(a)] did not have any evidence of EBSP bands. This lack of
Kikuchi bands along with the surface topography suggests that the grains in the deposited
ﬁlm are randomly oriented highly defective and this ﬁlm is polycrystalline. At an
ambient temperature, the growth directions of Nb islands were not inﬂuenced by
substrate orientation (no substrate wetting). Patterns obtained from a sample grown at
450°C showed a weak signature of a few bands [Figure 3.4(b)]. This EBSP result and
those of AF M surface topography, suggested that either the grains in this ﬁlm are lightly
textured, or all of them are collimated and have the same orientation (highly textured),
but contain a lot of dislocations and/or other defects that cause strong pattern degradation.
Electron backscatter diffraction patterns obtained from ﬁlms grown at 660°C and 750°C
were very sharp and are shown in ﬁgure 3.4(c & d). EBSP of the Nb ﬁlm deposited at
1100°C was similar to those grown at 750°C. An EBSP obtained from one of these ﬁlms
(grown at 750°C) is shown before and after indexing in Figure 3.5.

Stage control scanning (controlled movement of the motorized specimen stage) was
used to study and characterize the orientation distribution of Nb in the ﬁlms grown at
750°C. Depending on the analysis, scan sizes ranged from 3 pm x 3 pm with 1 pm steps,
to 1 mm x 1m with 5pm steps. The resulting pole ﬁgures indicated that the entire ﬁlm
had the same orientation, within the 1° accuracy of the measurement [28]. Therefore, the
ﬁlms are epitaxial single crystals. The Nb ﬁlms display normals of [110], which is

consistent with the growth occurring on the closest packed {110} planes in the bee Nb.
60

 

Figure 3.4 Electron backscatter pattern obtained from 250 nm Nb ﬁlms grown onto
(1 1 2 0)Al2O3 at a substrate temperature of : a) 25°C, b) 450°C, c) 650°C, d) 750°C.

61

.coxowﬁ one?
£015 Doomn mo 058383 banana

E- 953?... Tie

a 32 con. 338:3 cm 80¢ 8:850 Emma 5033 ”BREE—w auuwomxomn 5585 m. m Saw—m
3: alpha .5 Si

E033 2.: 22m USE Em: :O oBBHMOm .3 33.23883 comozo 295 means Bum a at; USE
ﬁlluﬂﬁe'i. 0.4.2.58 dil—
oeocsm ._ 3 .533. mm

 

 

 

.M .8830 52, M—

o 362:: a : 8.92:» E
l 2&3 m2

 

    

nm em nan ”3.9m
. a2 2.5
25:00 263

62

3.3.4 Orientation Relationship Between Epitaxial Nb and Al203

EBSP generated pole ﬁgures of the bee Nb and single crystal A1203 substrate were
used to determine the epitaxial relationship between the sputtered Nb and substrate. Pole
ﬁgures, which show the distribution of selected crystallographic directions in the sample
relative to a given sample axis, were produced from the EBSP data. Pole ﬁgures of the
epitaxial Nb ﬁlm showing the distribution of <110>, <111>, and <100> poles are shown
in Figure 3.6. These pole ﬁgures were then superimposed to create the stereographic
projection for the epitaxial Nb ﬁlm. Stereographic projections of the (110)bchb and
(1150) A1203, collected from the same sample (which had a small portion of the
substrate masked during sputtering), are shown in Figure 3.7. The epitaxial relationship
between Nb and A1203 was found to be (110)bchb || (1 1§0 )th A1203 with [111]bchb ||
[OOOl]hcpA1203, which agrees with HREM observations [11,12,19]. Generally, the
epitaxial relationships determined using EBSP should be signiﬁcantly more accurate
(within ~1°) than those deduced from HREM, where structural images can be obtained
when samples are tilted signiﬁcantly off-axis leading to inaccuracies up to 4° [30]. In the
epitaxial Nb ﬁlms deposited for this study, the resulting pole-ﬁgure data from adjacent
sections of Nb and A1203 did not also reveal any misorientation (within the accuracy of

the measurement).

63

Pole Figure of an Epitaxial Nb Film

 

 

{111} {110}
1-11 -1-11 on
101 -101
' I
011
111 -111 1
I
110

 
 

 

ll-l -ll-l

 

{112} {100}

b001

 

 

 

 

 

Figure 3.6 Pole ﬁgures of the epitaxial Nb ﬁlm showing the distribution of <110>,
<111>, <112>, and <100> poles.

64

0001

1120

0001

 

 

Figure 3.7. Stereographic projections of the epitaxial Nb ﬁlm and sapphire substrate
showing the epitaxial relationship between the Nb ﬁlms and A1203 substrates is

(1 TO)bchb ”(1 1 2 0),“, 11.1203 with [111],,,Nb ll [0001],,,,,A1203.
65

3.4 DISCUSSION

In general the results of AFM and EBSP suggests that in the sample grown at an
ambient temperature there are strong bonding between Nb atoms and they form a
granular polycrystalline ﬁlm. At temperature of 450°C, there must be better bonding
(compared with those grown at an ambient temperature) between the Nb atoms and the
substrate (better surface topography and weak appearance of some Kikuchi bands).
Therefore, the Nb islands could have grown in a partially preferred orientation, without
any signiﬁcant coalescence between the Nb islands (not enough thermal energy for atoms
to move around). However, the possibility of partial effect from post deposition
annealing at 1100°C can not be ruled out. At higher temperature, in particular at 750°C,
a very smooth surface topography and well deﬁned Kikuchi bands suggests a strong

bonding between Nb ﬁlm (atoms) and A1203 substrate in agreement with predicted

growth mode [15, 23].

3.5 SUMMARY

In this chapter the growth and structural characterization of sputter deposited Nb ﬁlms
on (11:20 )hcp A1203 substrate using AFM and EBSP have been discussed. Nb ﬁlms
deposited at ambient temperature without post deposition annealing were characterized
by a rough granular topography, suggestive of a polycrystalline structure. For those
deposited at a substrate temperature of 450°C and post deposition annealed at 1100°C,
the surface topography suggested some alignment and possibly collimated grain. At a
substrate temperature of 660°C post deposition annealed at 1100°C, oriented steps, which

may be structural ledges, run across the surface. The optimum growth temperature was

66

found to be 750°C with post deposition annealing at 950°C. EBSP and AFM results
revealed that the Nb ﬁlms deposited at 750°C grow as (110)bcc single crystals with a very
smooth surface topography of about ~0.2 nm. It has been shown that bcc Nb(110) planes
can be grown epitaxially on A1203 (11.2-0) substrates. The in plane epitaxial relationships
were:
[0001]A1203 || [111]Nb
and

[1010]».1203 n [121]Nb.

67

Appendix 3.A

Summary of Sputtered Nb Films

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Run # / Predeposition annealing Growth Temperature / Post deposition
Sample # Temperature Annealing Temperature

7-1 1200°C 750°C/ 1100°C
7-2 1200°C 660°C/ 1100°C
7-3 1200°C 450°C / 1100°C

7-4 1200°C 1100°C/ 1100°C
7-6 None Room Temperature
8-6 1200°C 760°C/ 1110°C
11-1 1200°C 765°C / 950°C
12-2 1200°C 750°C / 950°C
17-1 1200°C 750°C / 950°C
35-3 1200°C 750°C / 950°C

 

A11 substrates were A1203 (1150).

Deposition rate = 3 A/s

Deposition pressure = 1-2.5 x 10'3 torr.

68

 

3.6 REFERENCES

1) W. P. Pratt Jr., S.F. Lee, J.M. Slaughter, R. Loloee, P.A. Schroeder, and J. Bass,
Phys. Rev. Lett. 66, 3060 (1991).

2) S. M. Durbin, J. E. Cunningham, M. E. Mochel, and C. P. Flynn, J. Phys.:Met. Phys.,
11, L223 (1981).

3) S. M. Durbin, J. E. Cunningham, and C. P. Flynn, J. Phys.:Met. Phys., 12, L75
(1982)

4) R. E. Somekh and C. S. Baxter, J. Cryst. Growth, 76, 119 (1986).

5) J. Kwo, M. Hong, and S. Nakahara, Appl. Phys. Lett., 49, 319 (1986).

6) K. Burger, W. Mader, and M. Riihle, Ultramicroscopy, 22, 1 (1987).

7) W. Mader and M. Riihle, Act. Met., 37, 853 (1989).

8) M. Kuwabara and J. C. H. Spence, J. Mater. Res., 4, 972 ( 1989).

9) J. Mayer, C. P. Flynn, and M. Riihle, Ultramicroscopy, 33, 51 (1990).

10) C. H. Lee and K. S. Liang, Acta Metall. Mater., 40, 8143 (1992).

11)J. Mayer, G. Gutekunst, G. Mobus, J. Dura, C. P. Flynn, and M. Rﬁhle, Acta Metall.
Mater., 40, $217 (1992).

12) D. X. Li, P. Pirouz, A. H. Heuer, S. Yadavalli, and C. P. Flynn, Acta Metall. Mater.,
40,8237(l992)

13) P. M. Reimer, H. Zabel, C. P. Flynn, and J. A. Dura, Phys. Rev. B, 45, 11426 (1992).

14) T. Wagner, M. Lorenz, and M. Riihle, J. Mater. Res., 11, 1255 (1996).

15) G. Song, A. Remhof, K. Theis-Brohl, and H. Zabel, Phys. Rev. Lett., 79, 5062
(1997)

16) G. Gutekunst, J. Mayer, and M. Ruhle, Phil. Mag. A, 75, 1329 (1997).

69

17) G. Gutekunst, J. Mayer, V. Vitek, and M. Riihle, Phil. Mag. A, 75, 1357 (1997).

18) M. Wagner, T. Wagner, D. L. Carroll, J. Marien, D. A. Bonnell, and M. Rﬁhle, MRS
Bulletin, 42 (1997).

19) T. Wagner, J. Mater. Res., 13, 693 (1998).

20) B. Wolﬁng, K. Theis-Brtihl, C. Sutter, and H. Zabel, J. Phys.: Condens. Matter, 11,
2669 (1999).

21) D. R. Askeland, The Science and Engineering of materials, PWS-Kent publishing Co.
Boston page 459 (1989).

22) EMS On Line, Center Interdepartmental de Microscoie Electronic, EPFT, Lausanne.
Pierre-Henri Jouneau and Pierre Stadelmann. http://cimesgl .epﬂ.ch/CIOLS/.

23) M. Ohering, The Materials Science of Thin Films, Academic Press In., San Diago,
(1992)

24) E. Bauer, Z. Kristallogr. 110, 423 (1958).

25) J. H. v. d. Merwe, Interf. Sci. 1, 77 (1993).

26) R. J. Hill, Physical Vapor Deposition, Temescal, Ca, Page 144 (1986). BA.
Movchen and A. V. Demchishin, Phys. Met. Metallogr. 28. 83 (1969). H. T. G.
hentzell, C. R. M. Grovenor, and D. A. Smith, J. Vac. Sci. Tech. A2, 218 (1984). J.
A. Thornton, Ann. Rev. mater. Sci. 7, 239 (1977).

27) C. P. Flynn, J. Phys. F: Met. Phys., 18 L195 (1988).

28) B.D. Cullity, Elements of X-Ray Diffraction Addison-Wesley, Reading
Massachusetts, Second edition (1978).

29) S. 1. Wright, J. Computer-Assisted MicrOSCOpy, 5, 207 (1993).

70

30) H. Geng, J. W. Heckman, W. P . Pratt Jr., J. Bass, J. F. Espinosa, S. D. Conradson, D.

Lederrnan, and M. A. Crimp, J. Appl. Phys., 86, 4166 (1999).

71

CHAPTER 4

GROWTH OF Cu FILMS ON EPITAXIAL Nb

4.1 INTRODUCTION

In the growth process of typical CPP magnetic multilayers or spin valves, a thin layer
(10-20 nm) of Cu ﬁlm must be ﬁrst deposited on the Nb contact layer. The presence of
the Cu layer eliminates the reaction between the Co or Py layer with the Nb contact layer
[1]. In the growth of epitaxial magnetic systems, this Cu layer has an additional role; it is
the foundation for growing epitaxial multilayers and spin valves in a preferred
orientation.

A series of Cu ﬁlms with different thickness (20-300 nm) were grown on 250 nm
thick epitaxial Nb ﬁlms at several substrate temperatures. As described in Chapter 3,
high quality epitaxial single crystal (1T0) Nb ﬁlms have been grown on (1120) sapphire
substrates at an optimum temperature of 750°C (post deposition annealed at 950°C) using
dc magnetron sputtering. A list of the epitaxial Cu ﬁlms grown in this study is given in
appendix 4.A.

In the past, researchers have used various fabrication techniques, such as MBE and
sputter deposition, to grow epitaxial Cu ﬁlms [2-24]. Most of the epitaxial Cu ﬁlms
fabricated using MBE and sputter deposition systems were grown on Nb substrates, using
substrate temperatures between 150°C and 350°C.

In this chapter, background information on the nucleation and the grth of fcc
materials (i.e. Cu) on bcc materials (i.e. Nb), including the epitaxial relationships

between bee and fee, twinning in (111)fcc metals, and twin boundaries, are discussed.

72

This chapter includes a detailed explanation of sputter deposition of Cu ﬁlms on epitaxial
Nb. Also included are the results of AF M and EBSP characterization of the sputter-
deposited Cu ﬁlms (grown at 350°C, 150°C, and 90°C). A discussion of the growth
process in sputter deposited ﬁlms, based on the AFM and EBSP analysis, concludes this

chapter.

4.2 BACKGROUND: NUCLEATION AND GROWTH

The crystallography and physical properties of Nb and Cu are well-known.
Crystallographic data for both Nb and Cu are summarized in Table 4-1.

The nucleation and growth of Cu on epitaxial Nb differs from the epitaxial growth of
Nb on A1203. In the Cu/Nb case, the surface free energies of the substrate, interface, and
overlayer are important factors that determine the growth mode (Three possible growth
modes, Volmer-Weber (VW), Frank-van der Merwe (FM), and Stranski-Krastanov (SK)
were described in Chapter 3). Details of the growth modes for metal/metal materials
have been discussed by Bauer and van der Merwe [3].

Theoretically, the free energies (y) of the growing surface and substrate and the
interface between the substrate and growing material play dominant roles in determining
the growth mode. For example, if the surface energy of the substrate is greater than the
growing overlayer surface energy, that is if y, 2 an + Yin‘, then the FM growth mode (layer
by layer) will occur. In this growth mode m, 7;", and y, are the surface energy of the
overlayer ﬁlm (thickness = n), the surface energy of the interface (thickness = n’), and the

surface energy of the substrate, respectively [24].

73

Table 4-1

CRYSTALLOGRAPHIC DATA FOR Nb AND Cu

 

 

 

 

 

 

Element name: Nb—Niobium-Columbium Cu—Copper
Crystal system: Cubic — bcc Cubic — fee

a = 0.33067 nm 0.36148 nm
Unit cell volume: 0.0359 nm3 0.0467 nm3

# of atoms in unit cell: 2 4

Atomic radius: Nb = 0.143 nm Cu = 0.128 nm

 

d (h,k.l)

dam) = 0.2337 nm

dull) = 0.2078 nm

 

Symmetry elements:

Four 3-fold rotation axes

Four 3-fold rotation axes

 

 

Melting point:

 

2469 °C

 

1083 °C

 

74

 

There are two basic conditions for growing epitaxial ﬁlms. a) The surface energies of
the two metals (overlayer and substrate) must be compatible. b) There must be only one
crystallographic orientation relationship between the ﬁlm and the substrate. It will be
shown that the Nb and Cu can satisfy these requirements. Table 4-2, lists the surface

energies for some metals and insulators.

4.2.1 Orientation Relationship Between fee and bcc: A (110) bcc plane contains two
<111> directions (111 and 11T : close-packed directions), one <100> direction, and
one <011> direction. A (111) fcc plane contains three <110> directions (110, 101, and
011: close-packed directions) and three <112> directions (112, 121, and 211). The
epitaxial relationships between bcc and fcc materials were originally studied in the early
19305 [25-27]. Nishiyama and Wasserrnan independently showed that one of the

possible atomic arrangements between (110)],cc || (111)fcc, is [1T0]bcc || [1 T2]fcc, which

also results in [100]bcc || [01T]fcc (this will be referred to as the NW relationship) [25,

26]. Kurdjumov and Sachs showed another atomic arrangement between (110)t,cc ||

(111»... with [1 T 1 1b,, n [01 T 1., (the KS relationship) [27]. Figure 4.1 shows the atomic
arrangements of both the NW and KS relationships. These two relationships differ by
5.26° on the (110),“ || (111)fcc planes. The orientation relationship can actually vary
continuously from the NW to the KS to again the NW as r (r = abcc/afcc) increases from
0.70 to 0.86 [6]. The NW relationship has been reported to be the most common in
(110)Nb/(111)Cu cases [2, 3, 6, 28].

To observe the geometrical NW relationship between (110)bchb and (111)fccCu, the

atomic arrangements of these two planes (see Figure 4.2 (a), (b)) have been superimposed

75

in the NW orientation relationship. The result of this superposition is shown in Figure

4.2 (c). The two alignments directions Cu[101] || Nb[100] and Cu[l T2] || Nb[T10], are
labeled. Observe that none of the close-packed directions aligns with each other. The
areas where Cu atoms coincide with Nb atoms appear in the diamond-shaped area and

form a broad band along the [100]bcc and [011]]‘cc (see A and B in Figure 4-2(c)).

Table 4-2

SURFACE ENERGY OF SELECTED ELEMENTS AND MATERIALS

 

Transition metal Ti V Nb Mo Ru Rh Pd Ta W Pt

2.6 2.9 3.0 2.9 3.4 2.8 2.0 3.0 3.5 2.7 J/m2

 

Noble metal A1 Ag Cu Au

1.1 1.3 1.9 1.6J/m2

 

Magnetic metal Cr Mn Co Fe Ni Gd

2.1 1.4 2.7 2.9 2.5 0.9J/m2

 

 

Semiconductors Diamond Si Ge GaP GaAs
1.7 1.2 1.1 1.9 0.9J/m2
Insulator LiF A1203 Can MgO NaCl
0.34 1.4 0.45 1.2 0.3.l/m2

 

 

 

 

Note that these are approximate values, which are difﬁcult to measure and depend on

surface orientation and reconstruction [24].

76

 

O O O o Q
[1121,

11101,,
O O 0 O O

 

 

 

 

o 9 ~
0 '53 8

O
O, o 0
um,

“—51:

071101.,

 

q o
' (b)

 

 

 

Figure 4.1 Atomic arrangement of the (a) NW and (b) KS (0 = 5.26°) relationships.
With 0, atoms on (110)ooc and O, atoms on (1 1 1)fcc [6].

77

 

Figure 4.2 (a) Atomic arrangement of bcc Nb (110). The atomic radius of Nb is 0.143
nm With dam) = 0.2337 nm

78

     
  
  

           
 
 

   
   

    
  
   
 

         
 

 
 

          
      
   

o
'0 o 0.. o,‘ :0.
o 9 ’6‘ 0'0
. .9.0.0 .0
' 0.0.0..
o a o
. o o o
o

o o
o o o o o o o 0
0.0.0.0.... 0.0.
o o o o o o o o o
0.0.... o o 0.0...
..p.-~ .

    

  
 

      
 

 

Figure 4.2 (b). Atomic arrangement of fee Cu (111). The atomic radius of Cu is 0.128
nm Wlth do | l) = 0.2078 nm.

79

     
  

      
     

999-9999999 99999999999-99999999
9‘9‘9’9’9’9'9z9:9:9:9‘9 9:9'9'9‘9'-‘-‘-""°”-5.9.”"9'9’9’9

000000 0000 0000000100
0.0.0.0.09090. .0.0. .0. .0.0.0.0 0 0.0.0.0.0.0
0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0 0.0.0.0 0.0
0000 000000000 0 0 000
0 . 0 0 0 0 .0 0 .1
. 0 0 0.0 '0. 0 0

  
     
   
     

 

      
 
 

       
 

   
 
 
    
 
     
  
 
  
 

0.0.0.0.0.0. . 0.0’0.0 0.0.0 . . .0.0.0.0 .0.
0000000 0 00 0 0 0
00 0 00 0000 0 0 90
00 00 .00 000000000 0
.00 0 00000,.00 00 00
00 00 00000,, 90‘0 00 0
00 00 000000-.00.- 00 D0
00 00 000000.000. 00 0
9999999999999,99,9 99 I0
0000000000 ‘0, 00000 0
0000000000.-.I.000 .0
000000....00090 0...
00000\\0 _000000000
0 09"... 0000.0
0 09‘000 0 00
0 00
, 00 00
I 0
0

  

  
 
  
   
       

    

      
 

          
     
    
    

    

      

     
   

 
 
 

 
   
    
     
     

   

     
 

      
 
 

 
 

   

  

   
   
  
  
   

  
    
   
  
 

 
 
  

 
   
 

  
 
       
  

    

9: . 99 *9.
9.9 9 9.9 .9.9.9.9.9.9. .9. . .9
99999999999999999
9999 9999999999999999999
99999999999999999999999 99
999999999999999999999999 9999999
9999999999999,99999999999999999
9999999999999999999999999999999
79999999999999999999999999999999
99999999999999999999999999999999
99999999999999999999999 999999
999999999999999999999 9999999999
9999999999999999999999999999999
9999 9 9 999 9999 9999999999
9999 99 9 999999999999999999
9999 9 9 9 999999999999999999
9999 99 999999999999999999999
999999999 9999999999999999999999
9999999999999999999999999999999
99999999999999999999999999999999
9999999999999999999999999999999
999999999999999999999 999 9 9
999999999999999999999 9
999999999999999999999999
999999999999 999999
999999999999999999999
99999 99999999999999
999999 999999 99999
99999 999999 999
999999 999999999999 9
99999999999999999999
999999999999999999999
99999999999999999999
9999999 9999999999999
99999999999999999999
999999999999999999999
9 9999 9999999999999
.'..... .Q...‘
999 999999999999999
999 9999 999 99
999 999
99999 9999
' ‘ '9 9 9’9’9’9’9 9 9 9 9'9'9‘9 9 9 9
99999-9999 9999999999 9
9 99-999999999999999 9
9999999999999999999999 7
999999999999999999999999999999
9999999999999999999999999999999
99999999999999999999999999999999
99999 9999999999999999999999999
999999 9999999999999999999999999
99999 999999 999999999999999999
99999 99999 9999999999999999999
99999 9999 999999 999999999
999999 99999 999999 999999999
99999 999999 9999999 99999999
999999 99999 9999 9 9999
99999999999999999999 9 999
9999-9999999999999999 999999
9999999999999999999
9999999999999999999
9999999999999999999
9999999999999999999
9‘9’9‘9‘9'9’9’9’9-‘9'9‘9‘9’9’9’9‘9'9'9'
9999999999999-999999

  

   

Figure 4-2(c) Unrelaxed overlaps of Cu(111) on Nb(110) in the N-W orientation
relationship (Cu[101] I I Nb[100]). For simplicity, perfect (defect free) Nb and Cu
layers are assumed.

80

These diamond-shaped areas have higher energy compared with other areas and are
unfavorable areas for nucleation. The regions between the diamond-shaped areas form
zig-zag fringes (see C on the ﬁgure). In these areas, the Cu atoms sit in or near the
troughs (valleys) between three Nb atoms (good atomic matching = favorable areas for
nucleation). There are also fringes formed between the corner of these diamonds, parallel
to [110b,cc and [112]rcc direction (see D on ﬁgure 4.2 (c)) with relatively good atomic
matching.

In Figure 4—2 ((1), the Nb and Cu planes (Figure 4.2 (a), (b)) have been superimposed
in the KS orientation relationship for comparison to the NW orientation relationship.
Similar to the NW orientation relationship, the diamond-shaped areas are unfavorable
areas for nucleation. In both cases the fringe patterns are created due to existence of
misﬁt dislocations (edge, screw, and mixed dislocations) [6, 28]. However, there are
more misﬁt dislocations per unit area (.'. higher misﬁt strain energy) in an epitaxial ﬁlm
with the KS orientation relationship than those with the NW orientation relationship
(same epitaxed metal pair: see 4 c&d). Bruce and Jaeger [28] have shown that favorable
epitaxial relationships are those having a smaller misﬁt strain energy (.'. fewer misﬁt
dislocations per unit area). In (111)fcc / (110)bcc (similar to (111)Cu on (110)Nb)
epitaxed metal pairs, the NW orientation relationship is generally favorable over the KS

orientation relationship.

4.2.2 Nucleation: The nucleation and growth of fee materials on bcc substrates have
been investigated by many researchers [2, 3, 6, 28]. However, the epitaxial nucleation of

fee thin ﬁlms on bcc substrates is not well understood. Bauer and van der Merwe [3],

81

’.‘9:.:9:9:.:9:9:9:.:9
9 9 9 . 9 9 9 9 9
. . ... .. . ..9-9-9.9-9-9-9-.-
9 9
9 9‘9’9’:'9°9 9 9':'9‘9'9'9’9 NbUIUIICuWIU
.‘9‘9’9’9'9’9’9‘9’9'9°.°.°9’9
9...9...9.9.9.9.9.9.9.9.9.9...9.9;
9 9 9 9 9 9 9 9 9 9 9 9 9
9 9 9 9 9 9 9 9 9 .
9 9 9 9.9 9

9 . 9 9 . . 9 . . 9 .
9 9 9 9 9 . . 9 9 9 9 9
9.9:9:9°9°9°9:9:9'9’ '9'.

         
     
  
  

 
     

   
   

9 9
9 9’.’.’9°9°
.‘9‘.’.‘9’9‘9‘9
. 9 9 . 9 .

      
    
   

  

. .9 9
9.9.9:9:9:9:9:9:9. '
9 9 9 . . 9 9 9
9 9 9 9 9 9 9 9

0
0.0000000000000000

0000000000"0O0

001N0000HI000’

        
  
  

   

. 9 9
9 9 9 9 9 9 .‘9’9’.’.‘ .
9 . . . 9 9 .‘:‘:‘:‘:‘.‘. .
9 . . 9 9 9 9 9 . . . 9 9
9 9 9‘...‘9‘.‘9'9'.'9 9
. . ..9 9 9 9 . . . 9

      
   
 
   

. 9 .
9 9 . 9 . .
’9'.’9'9°.°9°9’9‘9'9’9’9 9'.
. 9 9 9 9 9 . 9 9 9
9'9’.‘.’.°9°9 . .‘9‘9'9‘ ‘
9.9 9.. 9 . 9 ....9.... 9 .
. 9.. 9 .

9°: 9‘. 9 9 . 9
9 9 9 9 9 9 9 .
9 9 9 . 9 9 9
9 9 . 9 9 9 9 9
9 9 9 9 9 9 9
9 . . 9 . . 9 9 9 9
. . 9 9 . . 9 9 9
. . . 9 9 9 9
9......

        

9

. 9 9 9

9.9.9...
9

   
 

.‘9‘: 9'. 9 9 9 9
.:.:.‘.’.°.‘.'.’.:.
.

0 0 0 0 0

0
0 0
0.0.0.0... 0

0 0

0 0 0 0 0 0

0 0 0 0 0 0 0 I 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0
0 0.0.0.0 0.0.0.0.0.0.0.0.0.0.0.0 .0.0.

 

 

Figure 4-2 ((1) Unrelaxed overlaps of Cu(l l 1) on Nb(110) with consideration of K-S
relationship (Cu[Oll] H Nb[lll]).

82

have theoretically predicted that at elevated temperatures (in structures similar to Nb/Cu),
the ﬁrst (and possibly second) Cu layers grow pseudomorphically, aligning exactly with
the top Nb layer. Starting with the second or third layer, the atomic distributions of Cu
are similar to the (111)fcc plane with the NW orientation relationship [3]. From the third
layer, three-dimensional (3-D) nucleation occurs within the SK growth mode (layer +
island) [3]. At room temperature, quasi-F M (layer by layer) growth has been
experimentally observed and reported [3, 29].

In the SK growth mode (at elevated temperature), formation of coherent (dislocation
free) islands are energetically favored over incoherent (dislocated) islands [28]. These
coherent islands are lattice-matched near their bases, but are largely strain-relieved near
their top and sidewalls. Further deposition leads to the growth of islands and eventual
coalescence to a continuance ﬁlm [30]. Theoretical work by Drucker et a1. [31] reported
that coherent islands are energetically favorable up to a critical size (<100 nm). Further
growth of these islands can cause a coherent-to-incoherent transition via the introduction

of misﬁt dislocations.

4.2.3 Twin Formation: In the growth of fcc(1 1 l) metals on bcc(110) metals, formation
of twins within the fcc(111) layer is a common phenomenon and has been reported by
many researchers [3, 8, 9, 23, 32]. As was discussed in the pervious section, the ﬁrst
fcc(111) layer (i.e. Cu(111)) grows pseudomorphically on the top bcc(110) (i.e. Nb(110))
layer, then starting with the second or third layer, the atomic distributions of Cu are
similar to the (111) plane [3]. This ﬁrst fcc(111) Cu layer becomes a buffer layer for

arriving Cu atoms. The formation of twins will take place on the top of this fcc(111) Cu

83

layer. The formation of twins in Cu(111) was observed using scanning tunneling
microscopey (STM) [13, 15], which showed that the Cu(111) coherent islands form at
one of two possible three-fold sites [12, 15, 22]. One of these three-fold islands follows
the correct stacking (ABCABC) while the second island forms a different stacking
(ACBACB). Schematic representaions of the side and top views of the stacking sequence
(twins) are shown in Figure 4-3.

The orientation relation between the Cu(111) twins and the Nb substrate is the NW
orientation [3]. Upon further ﬁlm growth, no additional twin orientations form. This is
due to the fact that the rotation angle is a rational multiple of 21: [3]. The presence of two
variant twin morphologies in crystalline superlattices (such as Cu/Mo, Cu/Nb) has been
discussed theoretically by Bauer and van der Merwe [3]. These twins have also been
experimentally observed by a number of researchers in MBE or sputter deposited
Co/Cu(111) multilayers using XRD, LEED, STM, and HREM [8, 12, 15, 19, 23, 31].
These papers, however, do not report on the distribution and size of twins in the epitaxial

Cu ﬁlms.

4.2.4 Nature of Twin Boundaries: To study boundaries between the two twin variants
in epitaxial Cu ﬁlms, high-resolution transmission electron microscopy is required. This
requirement is beyond the scope of this study; however, there have been theoretical and
experimental studies on twin grain boundaries in copper [33-36]. These studies will be
used in this section to illustrate the relationships between twin and grain boundaries.
Twinning is intrinsic to the growth of fcc(1 1 1) Cu and occurs with 2 = 3 coincidence-site

lattice (CSL) orientation. The twin boundaries that form during growth can be

84

 

Figure 4.3 Schematic representations showing top and side views of the stacking
sequences (twins) in (111) fcc.

85

categorized into three groups [33]:

a) Coherent twin boundary (CTB): The grain boundaries align themselves with a
common {111} plane of both grains. These are relatively low interface
energy (~ 0.022Jm'2) boundaries with no mobility.

b) Incoherent twin boundary (ITB): The grain boundaries are not parallel to any
{111} plane in either adjacent grains. This is an incoherent 2(3) twin
boundary. The interface ﬁlm energy of an ITB is around 0.5 Jm'z.

c) Symmetric incoherent twin boundary (SITB): This is a special case where the
grain boundary plane coincides with a common {211} plane of both grains.
The energy of SITB is around 0.84 Jm'z.

Medlin et a1. [36] have investigated the nature of twin boundaries in copper by
depositing epitaxial Cu ﬁlms on a single crystal Cu substrates. The results showed that
the twin boundaries in this homo-epitaxial Cu are incoherent 2(3) twin boundaries.
However, the boundaries are complex, containing nanometer (1.4 to 3 nm) thick slabs
with 9R stacking. The 9R slabs are rhombohedral, generated from the fcc structures by
creating an intrinsic fault on every 3rd {111} plane. This produces a stacking sequence of
ABC-BCA-CAB [34]. Atomistic models and HRTEM images of the incoherent 2(3)
twin boundaries of copper are shown in Figure 4.4 [36]. The 9R structure is very ﬁagile
and defects in the form of stacking faults can be generated during HRTEM observation.

The above review of the fundamental concepts should provide a foundation for the
understanding of experimental characterization and the results obtained from AFM and

EBSP characterization analysis.

86

’o-~99..-9—9~~c .
.. -.-.

’ 9-0....g...

  

99 9 9p.""
09M , ' .

. ‘) ) . .
.. , .w. " 5 ~ = O
(a . (3 .9 ”Q g0) )8 9 )'. 9' 131‘ 9.
«3%: ,9 .‘ (339%? O 0 09+. 3 211'

 

 

Figure 4.4 a) HRTEM images of incoherent 2(3) twin boundaries in Cu, imaged along
[011]. The 9R stacking faults are labeled, containing four or ﬁve layer defects. b)
Embedded atom method atomistic simulations of the structure of the Cu boundaries [36].

87

4.3 SPUTTER DEPOSITION OF Cu FILMS

As described in the Chapter 2, the high temperature heater was removed aﬁer post
deposition annealing of the Nb contact layer to cool the substrate to desired temperatures
(90°C - 350°C). Meanwhile, the low temperature heater was warmed-up to the desired
deposition temperature and mounted on the substrate. Typically, the heater switching
processes took approximately 10 minutes. After the target deposition temperature was

achieved, the Cu ﬁlm was deposited.

4.4 RESULTS
4.4.1 AFM Analysis
Cu samples with ﬁlm thicknesses ranging from 20 nm to 300 nm were grown on 250

nm of epitaxial Nb at substrate temperatures of 90°C, 150°C, and 350°C. The majority of
the Cu ﬁlms were subsequently post deposition annealed at 350°C, with a few annealed
at 250°C and 500°C. The surface topography of each ﬁlm was examined using AFM.
Note that the surface topography measurement is a confusing issue and depends how one
will deﬁnes a roughness. The following are examples of different kinds of surface
roughnesses:

a) Small grains (islands) with shallow terraces.

b) Small grains (islands) with deep terraces.

c) Large grains (islands) with shallow terraces.

(1) Large grains (islands) with deep terraces.
There are two parameters in the given examples; depth of terraces and size of grains
(islands). If the depth of terraces is the important parameter, then the Root Mean Square

88

(RMS) Roughness, which is the standard deviation of the Z (Height) values, becomes the
critical parameter (see Appendix 4.B). In this study, while the image RMS roughnesses
for all samples have been measured, the sizes of grains (or islands) and the depth of
terraces have also been considered for deﬁning a surface roughness. For example, in case
of having two surfaces with similar depth terraces but different island size, the surface
with smaller size islands will be considered rougher.

In general, the growth temperature, ﬁlm thickness, deposition rate, and substrate
topography affected the morphology of the sputtered Cu ﬁlms. The results of AFM
analysis are categorized as follows:

OSurface topography of the as-grown samples.

OEffect of post deposition annealing on the ﬁlm surface topography.
OEffect of ﬁlm thickness on the ﬁlm surface topography.

OEffect of deposition rates on the ﬁlm surface topography.

9Effect of deposition techniques on the ﬁlm surface topography.

4.4.1.1 As-Grown Cu Films
The as-grown ﬁlms consisted of eighteen Cu ﬁlms which were grown at 90°C, 150°C,
350°C, 400°C, and 475°C. Selected AFM results from these ﬁlms are categorized as a
function of ﬁlm thickness.
300 nm thick samples: AFM images and line trace surface analyses of 300 nm
thick Cu ﬁlms grown at 90°C, 150°C, and 350°C are shown in Figure 4.5. The image
RMS roughness in images a, b, and c are about 4, 4.5, and 20 nm, respectively. The Cu

ﬁlm grown at 90°C has a very granular appearance, suggesting small collimated islands
89

The Effect of Deposition Temperature on Surface Topography of
300 nm Thick Cu Films as Grown at 25°C, 150°C, 350°C.

 

_ 5.00
‘, 9 HI

    

 

 

10.0

 

 

 

 

 

 

mi
4
.4
- i.
' ?

 

 

 

j
3

 

 

 

 

~19.0 (
{4
q

Figure 4.5 Atomic force-microscopy images and line trace surface analyses showing the
surface topography of 300 nm thick Cu ﬁlms deposited onto 20 nm thick Nb ﬁlm at
substrate temperatures of a) 25°C, b) 150°, and c) 350°C. The line trace surface analyses
shows the image RMS roughness are about 4, 4.5, and 20 nm, respectively.

90

(~ 80 nm) in this ﬁlm. At higher temperatures, the ﬁlms have different surface proﬁles.
Smoother surfaces are observed, but some deep troughs exist, suggestive of better
coalescence of islands (than those grown at lower temperature) and formation of large
size islands on the surfaces of these ﬁlms.

150 nm thick samples. Three Cu ﬁlms with a thickness of 150 nm were
compared. These ﬁlms were deposited at substrate temperatures of 150°C, 350°C, 475°C
on 250 nm thick epitaxial Nb. AFM images and line trace surface analyses of these Cu
ﬁlms are shown in Figure 4.6. The image RMS roughness of these ﬁlms are 1.3, 23.3,
and 2.6 nm, respectively. It appears that the high temperature deposition produced much
larger islands and that island coalescence is occurring more often than that in the lower
temperature deposited ﬁlms.

75 nm thick samples. The AFM images and line trace surface analyses of three
Cu ﬁlms with thickness of 75 nm are shown in Figure 4.7. These Cu ﬁlms grown at
substrate temperatures of 150°C, 250°C, 430°C with image RMS roughness of 1.5, 12,
and 20 nm, respectively.

20 nm thick samples: There were two 20 nm thick Cu ﬁlms that were in the as-
grown category. These were grown at substrate temperatures of 250°C and 350°C. The
AFM images and line trace surface analyses of these two-samples are shown in Figure
4.8. The image RMS roughness of these are 5.5 nm and 10 nm, respectively.

The AFM analysis results from the as-grown ﬁlms revealed that sputter deposition of
Cu ﬁlms with lower substrate temperature generally results in a smoother surface
topography. However, samples deposited at higher substrate temperature generally have

deeper valleys (trough). The exception was in the case of Figure 4.6(c) (grown at 475°C),

91

The Effect of Deposition Temperature on Surface Topography of
150 nm Thick Cu Films as Grown at 150°C, 350°C, and 475°C.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘5
3
v I

H W
a

6]

U"

I

r x no 2 oo 3.bo 43a 5 no

 

Figure 4.6 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 150 nm thick Cu ﬁlms deposited at substrate temperatures of: a)

150°C , b) 350°C; and c) 475°C with image RMS roughness of about 1.3, 23, and 2.6 nm,
respectively.

92

The Effect of Deposition Temperature on Surface Topography of
75 nm Thick Cu Films as Grown at 150°C, 250°C, and 430°C.

5.00

2.50

 

 

 

10.0
I

 

 

-10.0
1

 

 

 

 

 

 

n
...
O
O
N
O
O
N
on
O
A
0|
0
O

 

 

 

10.0

WWW t
l Ulljllljjl U

 

E

-10.0

 

 

 

 

i.'oo 2. o 4.bo 5.00

(C)

Figure 4.7 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 75 nm thick Cu ﬁlms deposited at substrate temperatures of: a)
150°C , b) 250°C; and c) 430°C with image RMS roughness of about 1.5, 12, and 20 nm,
respectively.

93

The Effect of Deposition Temperature on Surface Topography of
20 nm Thick Cu Films as Grown at 250°C, and 350°C

 

 

 

 

 

 

10.0

 

(

 

 

-l0.0

 

 

 

Figure 4.8 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 20 nm thick Cu ﬁlms deposited at substrate temperatures of: a)
250°C and b) 350°C with image RMS roughness of about 5.5 and 10 nm, respectively.

94

which has larger sized islands, and very few shallow troughs. These islands had a better

chance to coalesce in this temperature range.

4.4.1.2 Effect of Film Thickness

In this section, the effect of ﬁlm thickness on the surface topography of the sputter
deposited Cu ﬁlms is examined. The ﬁrst group consisted of four Cu ﬁhns deposited at a
substrate temperature of 350°C on epitaxial Nb. AF M images and surface analysis of
these Cu ﬁlms with thicknesses of 20, 100, 170, and 300 nm are shown in Figure 4.9.
The image RMS roughness are 9.9, 28, 23.3, and 28.7, respectively. The image RMS
roughness results suggest that the thinnest Cu ﬁlm (20 nm) had smoother surface
morphology. However, the AF M images suggest that thinner ﬁlms have smaller islands
(~0.3 um compared to ~10 pm), but by considering the number of island per unit area,
the thinner ﬁlms are rougher. As examples, on thicker ﬁlms (100 and 170 nm) the island
sizes increased while the troughs become deeper (especially in 100 nm thick ﬁlm). The
images of the 300 nm ﬁlm in part (d), however, had the roughest surface topography
compared to rest of the samples with a mixed island size and many deep troughs. The
second group consisted of three Cu ﬁlms deposited at 150°C and post deposition
annealed at 350°C. Figure 4.10 shows AFM images and the line trace surface analyses
of: a) 20 nm, b) 200 nm, 250 nm Cu ﬁlms deposited on epitaxial Nb at 150°C. The
image surface roughnesses were 4.5, 5.7, and 8.5 nm, respectively. In this group of Cu

ﬁlms, the island sizes and trough depths increased with increasing ﬁlm thickness.

95

The Effect of Film Thickness on Surface Topography of
Cu Films Deposited at 350°C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

-10 0 t
:
-==::
.— a
G
'_ G
.q‘
u ‘=::::::::
__c=
ﬁ

-10 0

 

 

 

 

10.0

 

 

 

 

 

 

 

 

 

 

‘10.0 I
l
' =:::::£:::::
l
'10 0

p.
.—
a
o
N
8
5"
8
.9
o
o
a!
o
o

(C) ((1)

Figure 4.9 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of: a) 20 , b) 100, c) 170, and d) 300 nm Cu ﬁlms; grown at 350°C on
a 250 nm epitaxial Nb buffer layer. The line trace surface analysis shows the image RMS
roughness are about 10, 28, 23, and 28 nm, respectively.

96

The Effect of Film Thickness on Surface Topography of
Cu Film Deposited at 150°C

 

 

 

 

 

-10 0
1
~10 0

 

 

 

 

 

 

n
p
8
N
8
(9)
O
O
J
8
U1
8
n
9‘
D
O
N
D
O
9:
O
O
A
8
0|
8

 

10.0
1

 

 

 

-10.0
I

 

 

A
p
D
G
N
O
D
u
a-
O
A
O
0
U1
0
D

”I

(C)
Figure 4.10 Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of: a) 20, b) 200, c) 250 nm Cu ﬁlms; grown at 150°C and post

deposition annealed at 350°C. The line trace surface analysis shows the image RMS
roughness are about 4.5, 5.7, and 8.5 nm, respectively.

97

4.4.1.3 Effect of Post Deposition Annealing

Some of the Cu ﬁlms were subsequently post deposition annealed at several
temperatures to examine the effect of annealing on the topography of the sputtered ﬁlms.
The ﬁrst group of samples consisted of four 300 nm thick Cu ﬁlms deposited at 150°C.
Three of these ﬁlms were post deposition annealed at 250°C°, 350C°, and 500C° for 30
minutes. Figure 4.11 shows the AFM images and line trace surface analysis of these
ﬁlms. The image RMS roughnesses are approximately 4.5, 8.8, 6.3, and 3.6.nm,
respectively. Films with AF M images in part (b), (c), and (d) were deposited during the
same sputter deposition run using similar substrates (from same wafer). These images
appeared to be rougher than image in part (a). The AF M image of the sapphire substrate
used for these three ﬁlms showed rough surface topography and pre-deposition annealing
of A120; at 1200°C did not smooth the surface. However, by comparing these three
samples, it seems that when the annealing temperature increased to 500°C, the surface
roughness decreased.

Two additional samples were deposited at a substrate temperature of 350°C and one of
them was post deposition annealed at 500°C. In Figure 4.12, AF M images and surface
analysis of the 300 nm thick Cu ﬁlms grown at 350°C before and after post deposition
annealing at 500°C for 30 minute are shown. The ﬁlm surface roughness decreased to
about 6 nm from 28 nm, as shown in Figure 4.12 (a). The AFM images of these two
ﬁlms shows a clearer picture of the effect of post-deposition annealing on the Cu ﬁlms.
In image (b), there are large, continuous, smooth islands with some troughs. This is an

indication of the effect of post-deposition annealing.

98

The Effect of Post Deposition Annealing on Surface Topography of 300 nm thick Cu
Films Grown at 150°C

—5.00 .'

  

.‘ 5.00

    

nu

_'a',"'.-:‘v1\""‘
" '.--J', ‘ .". I"
'1 - ~'-..‘:'-\ ' ,'
v.'.' -‘ ‘ ‘:|*
. In; '9); -"-..P :1”... if}.
nu. I V
C
. 811 ll
1 b
. , l.

 

 

 

 

 

 

 

 

 

 

O O
s
t 100 zoo 300 400 I 1.00 200 300 400 500
”I I!!!
(a) (b)

 

 

 

-10.0

 

 

 

 

 

-10.0

 

 

 

9
9-
S
N
8
N
S
A
8
U!
8
a
n
D
D
N
O
O
m
8
:5
O
O
In
8

Figure 4.11 The effect of post deposition annealing on the surface topography of 300
nm thick Cu ﬁlms. The post deposition armealing temperatures were a) as grown
=150°C, b) 250°C, c) 350°C, and d) 500°C, respectively. The line trace surface analysis
shows the image RMS roughness are about 4.8, 8.8, 6.3, and 3.6 nm, respectively.

99

The Effect of Post Deposition Annealing on Surface Topography of
300 nm thick Cu Films Grown at 350°C

 

{0.0
1

 

 

-10.0

 

 

 

 

 

 

 

 

(b)

Figure 4.12 Atomic force-microscopy images and line trace surface analyses showing
the surface topography of 300 nm thick Cu ﬁlms deposited at a substrate temperatures of
350°C and post deposition annealed at: a) as grown and b) 500°. The line trace surface
analyses shows the image RMS roughness are approximately 28 and 6.5, respectively.

100

4.4.1.4 Effect of Deposition Rate:

As discussed in the Chapter 1, epitaxial ﬁlms fabricated by MBE have very smooth
surfaces and these ﬁlms are typically grown with a deposition rate of 0.1-0.5 A/s.
Therefore, in this study the effect of the deposition rate (DR) on the surface topography
of epitaxial ﬁlms grown by sputter deposition was of interest. Three Cu ﬁlms were
deposited with different deposition rate conditions:

a) Start and ﬁnish with a low DR = l A/s.

b) Start and ﬁnish with DR = 4.1 A/s.

c) Start and ﬁnish with a high DR = 7 A/s.
These ﬁlms were grown at 150°C and post deposition annealed at 350°C. Figure 4.13
shows AFM images and line trace surface analyses of 200 nm Cu ﬁlms with the
respective image RMS roughness of approximately 1.5, 5.7, and 4.8 nm for 1 A/s, 4.1
A/s, and 7 A/s. The Cu ﬁlm deposited with low deposition rate (1 A/s) appeared to have
large size islands and some holes. The Cu ﬁlms grown with 4.1 A/s, and 7 [Vs deposition
rate had rougher surface topographies with smaller islands, and possibly collimated grain
formation in Cu ﬁlm with 7 A/s deposition rate. These results are consistent with
theoretical predictions [37], which indicate that increasing the deposition rate results in

smaller islands.

4.4.1.5 Effect of Deposition Techniques
Compared to MBE, sputtering deposition is a relatively high-energy deposition
technique. To allow a comparison of the epitaxial sputtered samples grown in the present

work, a 300 nm thick Cu ﬁlm, grown by MBE, was provided by the University of Leeds-

101

The Effect of Deposition Rate on Surface Topography of
Cu Films Deposited at 150°C

 

 

 

 

 

 

 

 

 

 

 

 

.. W L
't 100 2 W ”a 3"“ ‘ 0° 5 0° 712 1.00 2.00 3.00 4 00 5 00
(a) (b)

 

10.0

 

 

I

 

-10.0

 

 

5.00

9
.-
N
a.
O
on
O
D
&
D
O

”I
(C)
Figure 4.13 Atomic-force-microscopy images and line trace surface analyses showing
the surface topography of 200 nm epitaxial Cu ﬁlms grown at 150°C and post deposition
annealed at 350°C with deposition rate of: a) 1, b) 4.1, and c) 7 A/s on a 250 nm epitaxial
Nb buffer layer. The line trace surface analyses shows the image RMS roughness are
about 1.5, 5.7, and 4.8 nm , respectively.

102

England. This sample was grown on a 100 nm epitaxial Nb ﬁlm (grown at 950°C prior
to deposition of the Cu ﬁlm) at a substrate temperature of 350°C. The surface
topography of this sample was compared with a 300 nm thick sputter deposited Cu ﬁlm
onto 250 nm epitaxial Nb ﬁlm at the same temperature of 350°C. The deposition rate for
the sputter deposited sample was 4.1 A/s (1 [Us for ﬁrst few monolayers and gradually
increased to 4.1 A/s). The growth rate for the MBE grown ﬁlm was 0.5 A/s. AFM
images and surface analyses of these two ﬁlms are shown Figure 4.14. The image RMS
roughnesses are about 23 and 46 nm, respectively. The AFM image in Figure 4.14(b)
indicates that the MBE grown ﬁlm had larger islands when compared with the sputter

deposited sample, with smoother island tops and deeper grooves between these islands.

4.4.2 EBSP Analysis

Up to this point the surface topographies of sputter deposited Cu ﬁlms have been
discussed; however, the surface topography does not provide information about the
orientation and epitaxy of the deposited ﬁlms. To generate orientation information, a
second characterization technique, EBSP, was carried out. The EBSP technique was
used to examine the crystallographic nature of the Cu ﬁlms that were investigated using

AFM. The results of the EBSP data analysis have been categorized in a similar way to

the AFM results.

103

Sputter Deposition vs MBE

 

 

 

 

-10.0

 

 

 

 

10.0

 

 

 

-10.0

 

0 .00

1.00 M 3.00 4.00 '5
”I!
(b)
Figure 4.14 AF M images and line trace surface analyses showing the surface topography
of: a) 300 nm Cu ﬁlm sputter deposited at 350°C on a 250 nm epitaxial Nb buffer layer,
b) 300 nm Cu ﬁlm; grown at 350°C by MBE technique on a 100 nm epitaxial Nb buffer
layer. The image RMS roughnesses are about 23 and 46 nm, respectively. The deposition
rates were a) start with IA/s for ﬁrst few monolayers and then 4.1 A/s and b) 0.5 A/s.

104

4.4.2.1 EBSP Patterns of Cu Films

EBSP and different variants EBSP data was collected randomly from different
locations on each sample. The pole ﬁgures obtained from the EBSP data of the Cu ﬁlms
revealed that two different orientation variants were present in all of the ﬁlms analyzed.
Both variants display <111> growth normals, consistent with growth occurring on the
close-packed {111} planes in the fcc Cu. The two variants differ by a multiple of 60°
rotation about this (111) normal, indicating that while the ﬁrst epitaxial Cu plane is the
same for each variant, the subsequent stacking of additional {111} planes occurred by
two different stacking orders, ABCABC versus ACBACB. Thus, the two variants are
oriented in a 2(3) twin relationship, but have incoherent boundaries (ITB). In Figure
4.15, an EBSP and corresponding pole ﬁgures of a 300nm thick sputter deposited Cu ﬁlm
are presented. In this ﬁgure, each of the patterns was captured from an individual variant
(twin). Because of the orientation relationship between these two variants, a number of
relatively strong Kikuchi bands (such as {131} and {311}) in the EBSP pattern are
common to both orientations, thus the two orientations are distinguished by differences in
other Kikuchi bands or poles such as I = [101], II = [011], and III = [001].

EBSP and twin boundaries. After observation of twin variants in the epitaxial Cu ﬁlms,
it was of interest to determine the effective resolution of the EBSP (effective interaction
volume) on the twin boundaries. To study this issue, an approximate location of a twin
boundary was selected; the electron beam was then slowly scanned over the selected
sample area. A series of patterns were collected during the scan, as shown in Figure 4.16.
These patterns show the transition from one variant to the other. The Kikuchi bands in

the EBSP pattern (Figure 4.16 (a)) from one the variants (only I = [101] pole can be seen)

105

WO>UUO>

 

OW>OW>

 

Figure 4.15 Indexed EBSPs and corresponding pole ﬁgures from a 300nm epitaxial Cu
ﬁlm sputtered at 350°C onto a 250 nm of epitaxial Nb substrate showing the two (111)
orientation variants .

106

EBSPs of Cu(111) Twins and Twin Boundaries

 

(a) (b)

 

(C) (d)

 

(C) (0

Figure 4.16 Changes in the Kikuchi bands of EBSP patterns obtained ﬁ'om 300 nm thick
epitaxial Cu ﬁlm as the sample was moved across a twin boundary. In part (a) and (b)
the EBSP patterns are from one of the variants. In parts (c) and (d) the Kikuchi bands
contributed from both variant were present. In parts (e) and (i), only the Kikuchi bands
from other variant are distinguishable.

107

showing no evidence of bands from the other variant. As the electron beam approached
the twin boundary, a collection of Kikuchi bands from both variants, with relatively the
same intensities, appeared (all three poles I, II, and III can be seen in Figures c and d).
By moving further away from the ﬁrst variant and boundary, only the Kikuchi bands that
originated from the second variant (II and III poles) are distinguishable in the EBSP
pattern (Figure 4.16(f)). Therefore, scanning the electron beam over a ﬁxed area (for
example 10 pm) of a sample and using several different step sizes (such as 0.1, 0.2, 0.5,
0.8, 1.0, 1.5, and 2 pm) could provide a relative effective interaction volume (resolution
of the EBSP) within the epitaxial Cu ﬁlms. A typical effective interaction volume for Cu
in this study was found to be approximately 0.8-1.2 pm. After tilting the sample about 70
degrees from its horizontal position, the effective interaction volume increased to
approximately 2.3 to 3.5 pm in one direction as shown in Figure 17. As the result of this
measurement, a step size of 1 pm was chosen to be used for most of this study. The
detail of step size measurement is given in appendix 4.B.

Indexing of the EBSP patterns. All patterns which originated from individual variants
were easily indexed using the Channel 4 software. However, EBSP patterns which
contained bands from both variants (at twin boundaries) usually remain unsolved during
indexing, resulting in blank points on the orientation map at the twin boundaries. An
example of this behavior is shown in Figure 4.18. In this ﬁgure, an orientation map and
corresponding pole ﬁgure showing two (111) orientation variants. The light green areas
(blank points) around the twin boundaries are unsolved orientations. These unindexed

areas (noise) could be removed using the Channel 4 software which substitutes the

108

Effective Electron Beam Size

Electron Beam

 

 

1 Sample

70 Tilt

Sample

Figure 4.17 Effect of tilting on electron beam size on the surface of the sample (effective
beam size) and on the interaction volume within the sample.

109

'L

Orientation Map of 300 nm Thick Epitaxial Cu Film

 

 

shaded areas

Cu (111) pole figure at" ligit 7T

.. [’K- _

. I . ’

P1 is. f . A,

s“ l ,v" Cu (111) pole figure of dark ,I’
x i . r shaded areas ~-.\'

 

 

 

 

. * - - .J
—-25 um; Map1:Step-l um; Grid-101x53
(b)

 

Figure 4.18 EBSP (111) pole ﬁgures and orientation map of the two (111) orientation
variants present in the 300 nm epitaxial Cu ﬁlm. In ﬁgure (a), the orientation map as
recorded by computer is shown. In the ﬁgure (b), the same map aﬁer noise was removed
by Channel 4 software.

110

 

unindexed point with the orientation of a nearest neighbor (with larger variant size.
Figure 4.18(b)).

EBSP and Orientation map of variants distribution To measure the size and
distribution of the variants in the epitaxial Cu ﬁlms, the samples were mapped using
stage control scanning. To resolve the changes in variants size and their distribution
within the Cu ﬁlms, the effective interaction volume and scanning step size become
crucial factors. By choosing a step size larger than the effective interaction volume, some
of the details such as variants size, twin boundaries, and their distribution will be missed
during the scan and will result in a meaningless map. As described in previous section, a
step size of 1 mm was chosen to be used for most of this study.

Film uniformity. To determine a representative scanning size, the uniformity of the
variants distribution in sputter deposited epitaxial Cu ﬁlms was investigated by
performing long range scans. The result of a 2000 pm x 100 pm scan size of one of the
epitaxial Cu ﬁlm is shown in Figure 4.19. By comparing these long-range scans with
many smaller (100 um x 100 um) scans size on the epitaxial Cu ﬁlms, it was observed
that the variants displayed a similar behavior with a random distribution without any
particular morphological patterns evident. Therefore, a smaller scanned area (for
exampleIOO pm x 100 pm) could be a representative of the variants distribution within
the epitaxial ﬁlm. It should be noted that the data taken for Figure 4.19 totaled 144 hours
(6 days and nights) of electron microscope time. The smaller scanned areas usually took

between 5 to 10 hours of electron microscope time.

111

Long Range Distribution of Twin Variants

Starting
Point of
. Stage Scan

 

'3' I
1

again??? 3?:

$15.5}:

 

-«

Figure 4.19 Orientation map of the distribution of two twin variants in a 200000 pmz
area (2000 pm x 100 pm) of a 300 nm epitaxial Cu film grown onto 250 nm of epitaxial
Nb. The green areas are from unsolved patterns, and all other color spots are improperly
indexed. Note that these maps were connected in one piece in the original map.

112

Long Range Distribution of Twin Variants

 
      
 

 

.-.-' I'm-J .35" ,9".
.= P" if;

-. ﬂit-a"

 

 

 

 

 

113

4.4.2.2 Effect of growth parameters on the variant size and their distributions.
In this section, the effects of the following parameters will be presented.
OGrowth temperature.
OFilm thickness.
0Post deposition annealing.
ODeposition rate.

ODeposition technique.

4.4.2.2.] The Effect of Growth Temperature

Cu ﬁlms which were deposited at a substrate temperature ranging from 150°C to
475°C displayed similar variant distributions to those shown in Figure 4.17. Cu ﬁlms
grown at lower temperature, however, display different variant distributions. In this

section, the effect of growth temperature on size of twin variants and their distribution for

300nm thick Cu ﬁlms is discussed.

300 nm thick samples grown at 350°C, 150°C, and 90°C The orientation maps
of 300 nm thick epitaxial Cu ﬁlms grown at substrate temperatures of 350°C and 150°C
are shown in Figure 4.20. The variants in these two maps did not display any particular
pattern or morphology. The variants sizes, determined by the linear intercept method,
showed that these ﬁlms have an average twin variant size of 4 :t 0.5 pm. The relative

areas of the two twin variants varied between 45% and 55%.

114

The Effect of Post Deposition Annealing On the
Orientation Distribution of Epitaxial Cu Film

 

(b)

Figure 4.20. Orientation map of the 300 nm thick epitaxial Cu ﬁhns grown onto 250 nm
epitaxial Nb at a substrate temperature of a) 350°C and b) 150°C.

115

In a typical magnetic multilayer or spin-valve, only the ﬁrst Nb and consecutive Cu
layer are deposited at elevated temperatures (to ensure the epitaxy). The rest of magnetic
and non-magnetic layers (such as Cu spacer) must be deposited at low temperature to
reduce the interdiffusion. Therefore, to study the quality of epitaxial Cu ﬁlms grown at
low temperature, a few Cu ﬁlms were deposited onto epitaxial Nb at 90°C and the EBSP
result was analyzed. A typical EBSP pattern (unindexed) from these samples is shown in
Figure 4.21. In this ﬁgure, the Kikuchi bands in the EBSP pattern do not have very sharp
band edges (compared to those in ﬁgure 4.15) and has slight degradation. By comparing
Figures 4.22 (b & (1); some of the Kikuchi bands in (b) (Cu ﬁlm grown at high T) are
clearer and the band edges are sharper than those in (d) as pointed by arrows. This
difference suggests the existence of strain in the Cu ﬁlm that was grown at room
temperature [3 8]. To measure the size and distribution of the variants in Cu ﬁlms grown
at 90°C, stage control scanning was used. However, aﬁer careful observation of the
patterns during the scanning, it was noticed that the patterns did not change during the
scanning processes and had relatively the same appearance. It was also noted that these
EBSP patterns are very similar to those that were collected from grain boundaries that
was shown in section 4.5.1 (see Figure 4.22 (b & (1). As explained in that section, the
analysis of these types of EBSP patterns revealed that both variants coexist within the
EBSP interaction volume.

Therefore, without consistent band indexing, the patterns are unsolvable and it is
impossible to use EBSP to fully characterize the variant size and distribution in the Cu

ﬁlms grown at this temperature (90°C) range. However, it can be concluded that the Cu

ﬁlms grown 90°C onto epitaxial Nb substrate were epitaxial, possessing two different

116

EBSP of a Cu Film Grown
at 90°C Temperature

 

Figure 4.21 Electron backscatter pattern obtained from a 300 mm Cu ﬁlm grown onto
epitaxial Nb ﬁlm at 90°C. This pattern shows some degradation compared to patterns
from Cu ﬁlm grown at 350°C.

117

Comparison of EBSP Pattern of Epitaxial Cu Films Deposited at
an Elevated and at an Ambient Temperatures

 

(C)
Figure 4.22 (a), (b), and (c) changes in the Kikuchi bands of EBSP patterns of a 300 nm
thick epitaxial Cu ﬁlm deposited at 350°C as sample was moved across a boundary (see
ﬁgure 4.16). (a) and (c) are individual twin variants. (b) and (d) contains bands from
both variants

118

(111) orientation variants. The results also suggest that the sizes of these variants (twins)
are smaller than the effective interaction volume and they simultaneously contribute
crystallographic information to EBSPs [39].

Grain sizes --- AF M vs. EBSP. In section 4.4.1.1 (AF M results), the surface
topography of the epitaxial Cu ﬁlms grown at room temperature and 350° were shown in
Figure 4.5-(a) and (c). It was a point of interest to establish a relationship between the
size and distribution of islands from AF M images and size and distribution of the twin
variants in the EBSP orientation map of a sample. An AF M image of a 300 nm thick Cu
ﬁlm grown at 350°C was superimposed on a corresponding EBSP orientation map. In
Figure 4.23, the AF M image and EBSP variants distribution map of this sample is shown.
The result of this comparison revealed that there is no correlation between the size and
distribution of the islands in the AFM image and the size and distribution of the twin

variants in the EBSP orientation map.

4.4.2.3 The Effect of film thickness - Limit of EBSP

As noted in section 4.4.1.2 of the AF M results, the surface roughness of epitaxial Cu
ﬁlms typically increase as the ﬁlm thickness increases (regardless of growth
temperature). Despite this increase, the twin variant morphology and size does not
appear to be a strong function of ﬁlm thickness for ﬁlms between 150 to 300 nm.
However, as the Cu ﬁlm thickness decreases, the backscattered electron interaction
volume extends into the Nb buffer layer. Thus, the diffracted electrons from the Nb

signiﬁcantly contribute to the electron backscatter diffraction patterns of the thin ﬁlms,

119

GRAIN SIZE
AF M vs. EBSP

. -5.00

.; . .. " 2.50
I .

10 [mm -. '
2' I: o

o
5.00 m.
Figure 4.23

An AFM image of a 300 nm thick Cu ﬁlm deposited at 350°C,
superimposed on the orientation map of the same sample. This comparison shows there

is no correlation between the size islands (on AFM image) and size of twin variants (on
EBSP map).

   

 

  

120

   
   

creating confusion in the indexing processes. This behavior is illustrated in Figure 4.24,
which shows orientation maps of 250, 100, and 20 nm thick ﬁlms.

In general, epitaxial Cu ﬁlms with a thickness greater than 150 nm did not have
signiﬁcant EBSP contributions from the underlying Nb, and only two Cu twin variants
are present. EBSPs from 100 nm thick samples were partially affected by backscatter
contributions from the underlying Nb, and some of the patterns could not be indexed. In
Fig. 4.24(b), the pink and gray areas are the two Cu twin variants, the green, black, dark
blue, blue, and orange areas are either unsolved or improperly indexed due to band
mixing. Note that it is possible that the improper indexed patterns could also come from
a few Cu clusters nucleated with different orientations, but only a TEM and HREM study
can conﬁrm or reject this possibility. A few brown spots that were indexed as Nb also
appear. The EBSPs from the 20 nm Cu ﬁlms had signiﬁcant contributions from the
underlying Nb.

These contributions resulted in some of the patterns being unresolvable, but others
being indexed as either Cu or Nb. In Fig. 4.24(c), the brown areas correspond to properly
indexed Nb, while the blue areas were solved as Nb, but were indexed in the different

orientation.

4.4.2.4 The Effect of Post deposition annealing

In section 4.4.3, it was shown that post annealing affects the surface topography of the
sputtered Cu ﬁlms. However, EBSP results from all of the post deposition annealed
samples showed that post deposition annealing generally does not affect the twin

morphology and variant orientation distributions. This was the case regardless of the

121

.nZ 33.53: 05 Bow 2033380 950% 523
A860 thmtc m3 32: so 33065 3 35¢? “55$“. Bum a coxﬂnmmv A8 EB 3V 8 238 couﬁcoto BE. .Uoomm mo chagnEB
89533 a 3 EE 9,2 333:5 E: on a no :38» v.85 :0 S: cm 9v 98 .2: 3v 6? 3 no awe coussto .¢~.v oSwE

2 I :3 5'
85.2 I E: 5

 

sputter deposition temperature. As an example, orientation maps for two 300 nm thick
epitaxial Cu ﬁlms grown at a substrate temperature of 350°C, one of which (b) had been
post deposition annealed at 500°C, are shown in Figure 4.25. These ﬁlms have an
average twin variant size of 4.1 and 3.7 pm, respectively. The relative areas of the twin
variants for both ﬁlms varied between 45% and 55%.

EBSP analysis of the epitaxial Cu ﬁlms grown at 150°C and post deposition annealed
at 500°C also revealed two different (111) orientation variants. However, the distribution
map of these sample shows some grain (variant) growth compared to an as-grown ﬁlm (at
150°C). In Figure 4.26, orientation maps for two 300 nm thick epitaxial Cu ﬁlms grown
at a substrate temperature of 150°C, one of which (b) had been post deposition annealed
at 500°C, are shown. These ﬁlms have an average twin variant size of 5.6 and 7.1 pm,

respectively.

4.4.2.5 The Effect of Deposition Rate

As discussed in section 4.4.1.4, the surface topography of the sputtered Cu ﬁlms was
affected by deposition rate. A low deposition rate (1 A/s) produced larger islands with
smoother ﬁlm surfaces (compared to those grown at higher deposition rate). EBSP results
of these epitaxial ﬁlms grown at deposition rates of 1 A/s, 4.1 A/s, and 7 A/s, however,
revealed similar orientation maps in all samples. Orientation maps of 200 nm Cu ﬁlms
deposited with a deposition rate of 1 A/s and 7 [Vs at a substrate temperature of 150°C
and post deposition annealed at 350°C are illustrated in Figure 4.27. The quality of the
EBSP patterns of sample with deposition rate of 1 A/s was very good, but slight

degradation was noted in sample with growth rate of 7 A/s. Grain sizes in the ﬁlms
123

THE EFFECT OF POST DEPOSITION ANNEALING ON THE
Orientation Distribution of Epitaxial Cu Film

 

Figure 4.25 a) Orientation map of the 300 nm Cu ﬁlm grown onto 250 nm epitaxial Nb
at a substrate temperature of 350°C. b) Orientation map of the 300 mm Cu ﬁlm grown
onto an epitaxial Nb ﬁlm at a substrate temperature of 350°C and post deposition
annealing of 500°C for 30 min.

124

The Effect of Post Deposition Annealing On the
Orientation Distribution of Epitaxial Cu Film

 

(b)
Figure 4.26 a) Orientation map of the 300 nm Cu ﬁlm grown onto 250 nm epitaxial Nb
at a substrate temperature of 150°C. b) Orientation map of the 300 nm Cu ﬁlm grown
onto an epitaxial Nb ﬁlm at a substrate temperature of 150°C and post deposition
annealing of 500°C for 30 min.

125

THE EFFECT OF DEPOSITION RATE ON THE ORIENTATION MAPS
of Epitaxial Cu Films ‘

 

 

 

(b)

Figure 4.27. Orientation maps of two 300 nm thick epitaxial Cu ﬁlms grown at 150°C
and post deposition annealed at 350°C with deposition rates of: a) lA/s and b) 7A/s. The
white areas on both maps are due to unsolved patterns (mostly on grain boundaries).

126

deposited with different deposition rates were determined by the linear intercept method.
These ﬁlms have an average twin variant size of epitaxial ﬁlm grown with deposition rate
of 1 A/s and 7 A/s are 2.8 and 4.9 pm, respectively. The relative areas of the twin
variants were similar to the previous cases, which varied between 45% and 55%. Table
4.3 summarizes the grain size and relative areas of the twins in all the epitaxial Cu ﬁlms

studied.

4.4.2.6 The Effect of Deposition Technique

The epitaxial Cu ﬁlm grown by MBE (at Leeds University) also revealed the existence
of two variants. Figure 4.28 shows the similarity of the orientation maps of 300 nm thick
epitaxial Cu ﬁlms (MSU and Leeds) grown onto epitaxial Nb substrates. The average
grain sizes in these ﬁlms were 4.1 and 3.5 pm, respectively. The relative areas of the

twin variants that varied between 45% and 55%, were similar to the previous cases.

127

Table 4.3

Summary of Average Grain (twin) Sizes

(Statistical: Intercept Method)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Run & Ele Thic- Growth Scan length % Twins Area Grain size
Sampl me kness /Anneal (pm) (pm
e # nt (nm) T°C X X Y X Step Size i 1 pm)
17-2 Cu 300 350/500 10x10x1 4
18-1 80 350/No lele1 N/A TS
18-2 80 43 ON 0 50x40x0.1 N/A TS
18-3 20 250/No N/A TS
18-4 20 350/No N/A TS
19-1 20 150/250 N/A TS
21-1 300 350/No 100x52x1 51%, & %48,& 4.3
%lUnSolved (US)
21-2 300 350/No 100x36x1 55%, & 45% 4.5
2000x26x0.5 49%, %32, %19US 5.3
21-3 250 ISO/350 50x82x1 55%, & 45% 4.7
22-3 300 250/350 100x50x1 -----
22-4 100 350/No 100x70x.5 44%, &36%, 4.8
20% US
23-1 300 150/No 50x50x1 53%, & 46% 7
11% US
23-2 300 150/250 50x50x1 59%, & 41% -----
23-3 300 150/350 2000x23x1 57%, & 43% 3.5
23-4 300 150/500 50x50x1 70%, & 30% 4-6
26-1 72 ISO/No TS
26-2 150 150/No 50x20x1 4.7
26-3 300 350/500 50x80x1 4.6
29-1 200 150/350 100x100x1 2.6
29-2 200 150/350 200x49x1 6.6
29-3 200 150/350 100x200x1 4.8
30-1 300 350 200x73x1 4.5

 

 

 

 

 

 

 

TS = Sample is Ihinner than 150 nm and there are band contributions from the buffer
NA = In all ﬁlms grown at low temperature, the grain distributions are not

layer.

detected by EBSP techniques, thus the grain sizes are _N_ot Available.

128

 

Orientation Maps of Sputter Deposited vs MBE Sample

 

 

l.-

(b)
Figure 4.28 Orientation map of the distribution of two variants in 300 nm epitaxial Cu
ﬁlm grown onto: a) A 250 nm of epitaxial Nb using dc sputtering. b) A 100 nm of

epitaxial Nb using the MBE technique. (The MBE facility is located at University of
Leeds England).

129

4.5 ORIENTATION RELATIONSHIP BETWEEN Cu AND Nb
fcc materials typically grow with their (111) plane parallel to (110)bcc substrates (see
section 4.2.1). The most favorable epitaxial relationships between (110)bcc and (111)fcc
are:
The NW relationship
[1 TOlhec ll [1 T21m and ITOOIbcc H [01 T lrce,

or the KS relationship

[ITIIbcc || [01 Tlrec

The epitaxial relationship between the sputter deposited epitaxial bcc Nb and fee Cu
ﬁlms was found by masking or removing part of the Cu ﬁlm in several samples (grown at
different deposition runs)
and collecting EBSP data from adjacent sections of Nb and Cu. The resulting {100},
{110}, and {111} pole ﬁgures were used to plot the stereographic projections of both the ‘
Nb and the Cu. Figure 4.29 shows a typical example of the stereographic projections of
epitaxial (110)Nb and (111)Cu ﬁlms. By superimposing the Cu stereographic projection
on the Nb, it can be seen that (110)hchb || (111)fccCu, with [1T0]hech || [1T2]fcc Cu and

[I00]bchb || [01 I ]fcc Cu, which is the NW relationship. This is different from that
observed in the polycrystalline Nb-(Cu-Co)x multilayers sputter deposited at room
temperature in our lab [40]. In that study, HREM image analyses of polycrystalline
multilayers showed that the Nb—Cu orientation relationship was close to the KS
relationship (110).,chb u (111)..,,Cu with [1111mm u [01 ‘1’ ],~,,Cu. The KS relationship
was also observed in polycrystalline Nb/Cu multilayers by Mitchell et a1. [41] using

selected area electron diffraction patterns.
130

sue: w a _ _ez§8 w a a? 535 _ E _ e2§§ : aséésa Sie aﬁemaaeeeﬁaz 2: he 9 ease as ageg
:0 8m use 32 can 05 5253 Q2838? E5330 2E. .mEE :0 we.» 92 333mg 05 mo maouooaoa oEmeoEBm mmé onE

131

 

 

 

 

 

4.6 SUMMARY

In this chapter, the nucleation and growth of epitaxial Cu ﬁlm was reviewed. Cu ﬁlms
fabricated on epitaxial Nb substrates using sputter deposition were characterized using
AF M and EBSP techniques. The AF M results showed that the surface topography is
strongly affected by the temperature and surface quality of the substrates. Most of AF M
results were reproducible. EBSP results revealed that all Cu ﬁlms grown at a temperature
higher than 150°C display 2 in-plane epitaxial variants corresponding to two stacking
sequences of {111}fcc planes. The distributions of these twin variants were random.
There was no correlation between the relative size of the twin variants and their
distributions in the EBSP orientation map and the size of the islands and their
distributions in the AFM image. These epitaxial Cu ﬁlms take on the NW epitaxial
relationship with respect to the Nb underlayer. The EBSP analysis of Cu ﬁlms grown
onto epitaxial Nb at room temperature revealed that two different orientation variants are
present with a very small grain size. However, owing to degradation of the associated
EBSPs and presence of the Kikuchi bands from both (111) twins variants in EBSP

patterns, mapping of the orientation distributions was not possible.

132

Appendix 4.A

Summary of the sputtered Cu ﬁlms grown
on epitaxial Nb ﬁlm in this study

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Run #/ Type of Film Growth Temperature
Sample #
8-03 Nb / Cu 750 °C / 350°C
250 / 120 nm Both deposited using HTH
8-05 Nb / Cu 750 °C / 450°C
250 / 120 nm Both deposited using HTH
15-01 Nb / Cu 658 °C / 350°C
250/ 150 nm HTH / LTH
Nb deposited using HTH
Cu deposited using LTH
15-02 Nb / Cu 654 °C/ 475°C
250/ 150 nm HTH / LTH
15-03 Nb / Cu 654 °C / 250°C
250/ 150 nm HTH / LTH
16-02 Nb / Cu 730°C / 400°C
250 / 100 nm HTH / LTH
17-02 Nb / Cu 730 °C / 400°C
250/ 150 nm HTH / LTH
18-01 Nb / Cu 725°C / 250°C
250 / 80 nm HTH /LTH
18-02 Nb / Cu 725 °C / 430°C
250 / 80 nm HTH /LTH
18-03 Nb / Cu 750°C /250°C
250 / 20 nm HTH /LTH
18-04 Nb / Cu 750°C / 350°C
250 / 20 nm HTH /LTH
19-01 Nb / Cu 750°C / (150°C-Annealed at 250 °C)
250 / 20 nm HTH /LTH
21-01 Nb / Cu 735 °C / 350°C
250 / 300 nm HTH /LTH
21-02 Nb / Cu 735 °C / 350°C
250 / 200 nm HTH /LTH
21-03 Nb / Cu 735°C / 150°C-Annealed at 350Ann
250 / 250 nm HTH /LTH
22-03 Nb / Cu 750°C / 250 °C-Annealed at 350°C
200-300 nm HTH /LTH

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

22-04 Nb / Cu 750°C / 350°C-Annealed at 350°C
200-100 nm HTH /LTH
23-01 Nb/ Cu 750°C / 150 °C-Annealed at 150°C
250 / 300 nm HTH /LTH
23-02 Nb / Cu 750°C / 150°C - Annealed at 250°C
250 / 300 nm HTH /LTH
23-03 Nb / Cu 750 °C/ 150 °C-Annealed at 350°C
250 / 300 nm HTH /LTH
23-04 Nb / Cu 750 °C / 150°C-Annealed at 500°C
250 / 300 nm HTH /LTH
26-01 Nb / Cu 750°C/ 150 °C
200 / 80 nm HTH /LTH
26-02 Nb / Cu 750 °C/ 150°C
200 / 150 nm HTH /LTH
26-03 Nb / Cu 750°C / (350-Annealed at 500) °C
200 / 300 nm HTH /LTH
29-01 Nb / Cu 750°C - (ISO-Annealed at 350) °C
200/200 nm DR (4,] A°/s)
29-02 Nb / Cu 750°C - (ISO-Annealed at 350) °C
200/200 nm Low DR (1 A°/s)
29-03 Nb / Cu 750°C - (1 50-Annealed at 350) °C
200/200 nm High DR (7 A°/s)
30-01 Nb / Cu 750°C - 350°C
250/300 nm Start with low DR
Leeds3K Nb / Cu 950 °C / 350°C
20 / 300 nm
Leedle Nb / Cu 950 °C /350°C
20 / 300 nm
Test Cu on Si Room temperature
300 nm

 

All of Nb buffer layers are post deposition annealed at 950°C for 20 minutes.
HTH = High temperature heater

LTH = Low temperature heater

DR = Deposition rate

 

Appendix 4.B
Effective Interaction Volume and Scanning Step Size
To obtain orientation distribution maps from the sputtered ﬁlms, stage scanning or
electron beam scanning was required. However, for scanning a large ﬁlm area, stage
scanning is more convenient. To collect reliable and relatively accurate orientation
distribution maps, the scanning step size must be small enough to overlap the effective
interaction volume during the scan. The proper scan size was found using the following

procedure.

a) The electron beam was used to scan a selected sample area (an electron beam
scan is more reproducible than a stage scan).

b) An area (10 um x 1 pm ) that contained a twin boundary was selected.

c) This area was scanned several times with different scanning step sizes (0.1— 2
pm) using the Channel 4 software.

(1) The orientation maps that corresponded to each scanning step size were

compared (size of each twin and location of the twin boundary).

In this study, 10 pm of an epitaxial Cu ﬁlm was ﬁrst scanned with step sizes of 2, 1.5,
1, 0.8, 0.5, 0.25, and 0.1 pm. The resulting orientation maps are shown in Figure 4.C.1.
In all cases, four twin boundaries were identiﬁed. The size of twins was similar in the
maps with the step size of 0.1 to 1 pm. However, in the scans with small step size (0.1
pm), Kikuchi bands from both variants were present at twin boundaries. At these twin
boundaries, the Kikuchi bands could be indexed as either one of the twins. This causes

uncertainty on the size of immediate twins. The orientation maps with step sizes of 2 and

135

Comparison of Scanning Step Size
using Electron-Beam Scan

 

 

 

 

 

 

 

 

 

 

 

 

 

 

{— - - - --_ — " '> Grid Step size
I' - eh .- ~~ *1 W1 11x1 2.01m
If '7 7....W....’_. m I I 14x1 1.5

i ' i i 21x1 1.0
LN... .. 1 “If—:6 , 1 26x1 0.8

I 1 41x1 0.5
P—‘i . ' ' 1‘ 81x1 0.25

.. ,i 201xl 0.1

 

 

 

 

 

 

Grid Step size

E, _ E a 11X] 1.011.!!!

“new... ”i.

I ., , - . I 21x1 0.5

[Izmirnissuatt « _ . . a 101X1 0.1

uvn en t. — >q. 4‘3" w. .vm

 

 

 

 

 

(b)

Figure 48.1 (a) Comparison of orientation maps obtained form 20 mm x 1 pm area of
300nm thick epitaxial Cu ﬁlm, using different electron beam scanning step size. A 50pm
aperture was used for these scans. (b) 10 um x 1 pm area of different location on the
same ﬁlm and without aperture. All other parameter were the same during scans.

136

1.5 pm showed different sizes of twins, suggesting that the actual twin boundaries were
missed. Using these results and many visual observations of the EBSP patterns during
these scans, the step size of 1 pm was selected for this study. However, a few areas were
scanned using 1 and 0.5 pm step sizes to ensure reproducibility of the orientation

distribution maps.

Note that these orientation maps are qualitatively correct and they do not represent the

exact twin distribution in these epitaxial Cu ﬁlms.

137

4.7 REFERENCES:

l.

10.

11.

12.

13.

Metallurgy and metallurgical engineering series, second addition, P.M. Hansen,
McGraw-Hill book company, New York, (1958).

Ivan K. Schuller, Phys. Rev. Lett. 44, 1597 (1980).

E. Bauer and Jan H. van der Merwe, Phys. Rev. B 33, 3657 (1986).

A. Cebollada, J. L. Martinez, J. M. Gallego, J. J. de Miguel, R. Miranda, S. F errer, F.
Batallan, G. Fillion, and J. P. Rebouillat, Phys. Rev. B 39, 9726 (1989).

A. Cebillada, R. Miranda, C. M. Schneider, P. Schuster, and J. kirschner, J. Magn.
Magn. Mat. 102, 25 (1991).

M. Kato, Mat. Sci. engr. A146, 205 (1991).

W. F. Egelhoﬁ' Jr. and M. T. Kief, Phys. Rev. B 45, 7795 (1992).

D. Greig, M. J. Hall, C. Hammond, B. J. Hickey, H. P. Ho, M. A. Howson, M. J.
Walker, N. Wiser, and Wright, J. Magn. Magn. Mat. 110, 1239 (1992).

K. Brohl, P. Bodeker, N. Metoki, A.Stierle, and H. Zabel, J. Crystal Growth, 127, 682
(1993)

A. Schreyer, K. Brc'ihl, J. F. Anker, C. F. Majkzak, Th. Zeidler, P. Bodeker, N.
Metoki, and H. Zabel. Phys. Rev. B 47, 15334 (1993).

M. J. Hall, B. J. Hickey, M. A. Howson, M. J. Walker, J. Xu, D. Greig, and N. Wiser.
Phys. Rev. B 47, 12785 (1993).

P. Bodeker, A. Abromeit, K. Brohl, P. Sonntag, N. Metoki, and H. Zabel, Phys. Rev.
B, 47, 2353 (1993).

J. de la Figuera, J. E. Prieto, C. Ocal, and Miranda, Phys. Rev. B, 47, 13043 (1993).

138

14

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25

26

27

28

M. A. Howson, B. J. Hickey, J. Xu, D. Greig, P. Rhodes, and N. Wiser, Phys. Rev. B
49, 9560 (1994).

J. Camarero, L. Spendeler, G. Schmidt, K. Heinz, J. J. de Miguel, and R. Miranda,
Phys. Rev. Lett. 83, 2448 (1994).

G.R. Harp and 8.8. P. Parkin, Appl. Phys. Lett. 65, 3063 (1994).

NJ. List, W.P. Pratt Jr, M.A. Howson, J. Xu, M.J. Walker, and D. Greig, J. Magn.
Magn. Mater. 148, 342 (1995).

G. R. Harp and S.S.P. Parkin, Thin Solid Films 288, p. 315, (1996).

J. Camarero, T. Graf, J. J. de Miguel, G. Miranda, W. Kuck, M. Zharnikov, A.
Dittschar, C. M. Schneider, and J. Kirschner, Phys. Rev. Lett. 76, 4428 (1996).

H. T. Hardner, M. B. Wiessman, B. Miller, R. Loloee, S.S.P. Parkin, J. Appl. Phys.
79, 7751 (1996).

D. J. Smith, AR. Modak, T.A. Rabedeau, and S.S.P. Parkin, Appl. Phys. Lett., 71.
1480(1997)

J. Cmarero, J. Ferrion, V. Cros, L. Gomez, A. L. Vazquez de parga, J. M. Gallego, J.
E. Prieto, J. J. de Miguel, and R. Miranda, Phys. Rev. Lett. 81, 850 (1998).

X. W. Zhoh and H. N. G. Wadley, Acta Mater. 47, 1063 (1999).

F. J. Himpsel, J. E. Ortega, G. J. Mankey, and R. F. Willis, Advances in Phusics, 47
511 (1998)

Z. Nishiyama, Sci. Rept. Tohoku Univ. 23, 368 (1934).

G. Wasserrnan, Arch. Eisenhuttenwesen 16, 647 (1933).

G. Kurdjumow and G. Sachs, Zeir. Physikl 21, 961 (1930).

L. A. Bruce and H. Jaeger, Phil. Mag. A, 38, 223 (1978).

139

29. A. Bauer, J. Phys: Condens. Matter 11, 9365 (1999).

30. H. M. Koduvely and A. Zangwill, Phys. Rev. B, 60, R2204 (1999).

31. J. Drucker, Phys. Rev. B, 48, 18203 (1993).

32. S. C. Jain, A. H. Harker, and R. a. Cowley, Phil. Mag. A, 75, 1461 (1997).

33. U. Wolf, F. Ernst, T. Muschik, M W. Finnis, and H. F. Fischmeister, Phil. Mag. A,
66,991(1992)

34. T. Muschik, W. Laub, U. Wolf, M. W. Finnis, and W. gust, Acta metall. mater. 41,
2163 (1993).

35. A. Scmidt, F. Ernst, M. W. F innis, and V. Vitek, Phys. Rev. lett. 75, (2160 (1995).

36. A. L. Medlin, G. H. Campbell, and C. B. Carter, Acta Mater. 46, 5135 (1998).

37. M. Ohring, The Materials Science of Thin Films, Academic Press, Inc. San Diego CA
(1992)

38. A. J. Wilkison, and D. J. Dingly, Acta metall. Mater. 40, 3357 (1992).

39. R. Loloee, P. Duxbury, Private communications.

40. H. Geng, J. W. Heckman, W. P. Pratt Jr, J. Bass, F. J. Espinosa, S. D. Conradson, D.
Lederman, and M. A. Crimp, J. Appl. Phys., 86, 4166 (1999).

41. T. E. Mitchell, Y. C. Lu, A. J. Grifﬁn, M. Nastasi, and H. Kung, J. Am. Ceram. Soc.,
80, 1673. (1997).

42. P. RAI-Choudhury, Handbook of Microlithography, Micromachining and
Microfabrication, , V.1, page 205, SPIE Press, London, (1997).

43. J. S. Greeneich, J. Appl. Phys., 45, 5264 (1974).

44. N. Samoto, R. Shimizu, H. Hashimoto, I. Adesida, E. Wolf, and S. Nama, J. Vac. Sci.

Technol. B. l 1367 (1983).

140

CHAPTER 5

GROWTH OF Co AND Py FILMS ON EPITAXIAL Cu

5.1 INTRODUCTION:

In chapters 3 and 4, the details of the growth of epitaxial Nb(110) and Cu(111) were
presented. The next step towards making epitaxial magnetic multilayers and spin-valves
in a preferred (111) orientation is to deposit Co or Py ﬁlms onto the epitaxial Cu(111)
layer.

In this chapter, the nucleation and growth of Co and Py are ﬁrst reviewed, followed
with the details of sputter deposition of the Co and Py ﬁlms. Also included in this
chapter are the results of the AFM and EBSP characterization of the sputter-deposited Co
and Py ﬁlms (grown at 460°C and 90°C). A discussion of the growth process of these

sputter deposited ﬁlms, based on the AF M and EBSP analysis, concludes this chapter.

5.2 GROWTH OF Co FILMS ON EPITAXIAL Cu(lll)
5.2.1 Nucleation and Growth --- a Review:
Background Co is an allotropic metal that can grow with either a fcc or hcp crystal
structure, depending on the nucleation sites and growth conditions. At room temperature,
bulk Co exists with an hcp crystal structure, while at 417°C, it undergoes structural phase
transition to an fcc crystal structure.

In thin ﬁlms, the situation is more complicated. During hetero-epitaxial growth of Co

on Cu(111), the growing Co layers have the freedom to continue to stack either with the

141

substrate in an fcc sequence or in the hcp sequence. The hcp sequence is more favorable
energetically at room temperature.

Before reviewing the nucleation and growth of Co on a Cu(111) substrate, a
demonstration of the geometrical relation between Co and Cu(111) will help form a
biases for the nucleation and growth discussion. In Figure 5.1, a superimposed atomic
arrangement (unrelaxed) of Co(1 1 1) and Cu(111) with a 2% lattice mismatch is shown.
In this ﬁgure an orientation relationship of Co(111)<110> || Cu(111)<110> is shown.
The areas where Co atoms coincide with Cu atoms appear with complete circles (light
areas) and they are unfavorable areas for nucleation (as described in chapter 4). The

darker areas are coherent and are convenient sites for nucleation of coherent islands.

Nucleation, growth, and twin formation: The nucleation, growth, and structural
characterization of Co ﬁlms on epitaxial CU(111)fcc substrates (grown at room
temperature) have been investigated by many researchers [1-12].

Heinz et a1. [2] have investigated the crystallography of ultra-thin Co ﬁlms grown
epitaxially on Cu(111). They reported that from the lattice misﬁt point of view (2%), a
pseudomorphic growth of Co (C0(111)f¢c) was expected. However, their results did not
agree with the expectation due to the freedom of the fee or hcp stacking. LEED studies
showed that up to 2 ML of cobalt ﬁlms largely copies the fcc stacking of the substrate.
The growing ﬁlm is increasingly dominated by hep stacking when this thickness is
exceeded.

Tonner et al. [5] have also reported similar results using angle-resolved x-ray-

photoelectron diffraction and LEED. They found that the initial growth of Co (with

142

 

Figure 5.1. A superimposed atomic arrangement (unrelaxed) of Cu/Co with about 2%
lattice mismatch. The lighter areas are coincidence sites between Co and Cu atoms. The
darker areas are coherent with a large coherency length and are convenient sites for
nucleation of coherent islands.

143

the substrate maintained near room temperature) is in the metastable fcc phase. The ﬁrst
two monolayer Co atoms continue the stacking sequence of the fcc(1 l 1) Cu substrate
through a layer-by-layer and pseudomorphic growth mode. After these two layers up to
10 ML (~2.5 nm), about 50% of the Co atoms still continue the fcc(l 1 1) (ABCABC...)
and 50% undergo a change of stacking to hop (ABAB. . .).

The growth and structure of Co thin ﬁlms on Cu(111) substrates have also been
investigated by Kief and Egelhoff [6] using x-ray-photo electron, Auger electron forward
scattering, LEED, and RHEED. They noted that, growth of Co on Cu(111) at room
temperature is generally not a layer-by-layer FM growth mode. Also, there was much
larger portion of fee stacking than hcp stacking in the ﬁlms grown at room temperature.

In a STM study of the growth of fee Co on Cu(111) at room temperature, de la
F iguera et al. [7] reported that the growth is not layer-by-layer and the ﬁrst Co monolayer
does not wet the Cu substrate. Rather it grows in the form of bilayer islands. The
existences of large (30 A) triangular Co islands were found to nucleate at each atomic
terrace with two orientations, rotated 60° with respect to each other. Both orientations
have equal probability and similar island size. These Co twins could coincide on Cu
twins or grow independently.

Brohl et al. [8] have used the RHEED to study epitaxial growth of Co ﬁlms on
Cu(111). They reported that at a substrate temperature below 100°C the Co ﬁlms grow
on Cu(111) pseudomorphically in the metastable fcc phase up to 20 nm.

Unfortunately, these reviews do not provide a clear picture of the nucleation, growth
and twin formation of the epitaxial Co ﬁlms fabricated at low temperature (below

100°C). Nucleation and growth are simpler at a high temperature (460°C). Co ﬁlms

144

grow with a fee structure, however, there is signiﬁcant interdiffusion and alloying
between Co and Cu (about 10% miscibility at 400°C) [2, 13].

Most of the samples in the reviewed articles were fabricated using MBE. There are a
few reports in which the samples were fabricated by different techniques such as
sputtering deposition or pulsed laser deposition [1,6]. Similar to those discussed above,
there were no clear and consistent conclusion of how the epitaxial CO(111)fcc nucleates
and grows as well as how far the fcc layer can grow.

For this study, EBSP has been used to identify the orientation (in fee phase or possible
hcp phase) of 20 nm to 150 nm sputter deposited Co ﬁlms. In the next two sections, the
experimental results of AFM surface topography analysis and EBSP orientation
distribution analysis are presented. Based on the results from these two sections, the
growth process in the sputter deposited Cu ﬁlm for this study will be discussed. Note
that due to limitation of EBSP for thin ﬁlms, the Co ﬁlms that were deposited for this
study are much thicker than those that have been reviewed.

The crystallographic and physical properties of Co are summarized in Table 5.1.

5.2.2 Sputter Deposition of Co Films

As described in the Chapter 4, after post deposition annealing of a Cu layer, the low
temperature heater was either removed to cool the substrate to about room temperature or
warmed up to a higher temperature (460°C) for high temperature deposition of Co or Py

ﬁlms.

145

Table 5.1

 

 

 

 

 

 

CRYSTALLOGRAPHIC DATA
Element: Co- Cobalt Co-Cobalt
Crystal system: Hexagonal — hcp Cubic — fee
a = 0.2507 nm 0.3544 nm
b = 0.2507 nm 0.3544 nm
e = 0.4069 nm 0.3544 nm
Unit cell volume: 0.0221 nm3 0.0445 nm3
# of atom in unit cell: 4 4
Atomic radius: Co = 0.125 nm Co = 0.125 nm

 

d (h.k.l)

d(0001) = 0.2171111!)

dull) = 0.2046 nm

 

Symmetry elements:

A single 6-fold rotation axes

Four 3-fold rotation axes

 

 

Melting point:

 

1495 °C

 

1495 °C

 

146

 

5.2.3 AFM Analysis:

Co samples with ﬁlm thicknesses ranging from 20 nm to 150 nm were grown on 20
nm of epitaxial Cu ﬁlms at substrate temperatures of 80°C and 460°C. The surface
topography of each ﬁlm was examined using AF M. In general the growth temperature,
ﬁlm thickness, and substrate topography affected the morphology of the sputtered Co
ﬁlms. The results of AFM analysis are categorized as follows:

oEffect of substrate temperature on surface topography of the Co ﬁlms.

OEffect of ﬁlm thickness on surface topography of the Co ﬁlms.

5.2.3.1 Effect of substrate temperature on surface topography of the Co ﬁlms

All of the Co ﬁlms were grown on 20 nm epitaxial Cu(111) substrates. One set of Cu
substrates was deposited at 350°C and the second set was deposited at 150C° and post
deposition annealed at 350C°.

20 nm thick samples The surface morphology of the 20 nm Co ﬁlms deposited
at 460°C and 80°C was hardly affected by the substrate temperature. Figures 5.2(a) and
(b) show the atomic-force-microscopy images and line trace surface analyses of 20 nm
thick Co ﬁlms. These Co ﬁlms have been grown on a 20 nm epitaxial Cu ﬁlm (grown at
350°C) at substrate temperatures of 460°C and 80°C. The image RMS roughness of Co
ﬁlms deposited at above temperatures were similar, approximately 16 nm. The surface
topography of these two samples is very similar, with similar island sizes.

100 nm thick samples Despite the fact that the 100 nm thick Co ﬁlms were
deposited on a similar substrate to that used for the 20 nm thick samples, they show a

signiﬁcantly different surface topography. In Figure 5.3, 100 nm Co ﬁlms grown at
147

AFM Images of 20 nm Thick Co Film
Deposited at Substrate Temperature of 460°C and 80°C.

2.00

1.00

 

 

10.0

- mymvmvv

r.'so 2.00

 

-10.0

 

 

 

P
g.
e

 

10.0

 

-l0.0
r

 

 

 

 

0 0.30 1.110 1.50 -2.00
W
(b)

Figure 5.2 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 20 nm thick Co ﬁlms grown on 20 nm epitaxial Cu(111) (grown at
350°C) at substrate temperature of: a) 460°C and b) 80°C. The image RMS roughnesses
of both ﬁlms are about 16 nm.

148

AFM Images of 100 nm Thick Co Film
Deposited at Substrate Temperature of 460°C and 80°C

 

 

-10.0

 

 

 

 

10.0

 

-10.0

 

 

 

 

Figure 5.3 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 100 nm thick Co ﬁlms grown on 200 nm epitaxial Cu(111) (grown
at 350°C) at substrate temperature of: a) 460°C and b) 80°C. The image RMS
roughnesses are about 11 and 4 nm, respectively.

149

460°C and 80°C are shown. The images RMS roughness are about 11 nm and 4 nm,

respectively. The islands are similar in shape but larger in the ﬁlm deposited at 460°C.
150 nm thick samples Also, two 150 nm thick Co ﬁlms were deposited on

epitaxial Cu ﬁlms at substrate temperatures of 80°C and 460°C. AFM images and line

trace surface analyses of 150 nm Co ﬁlms grown at 460°C and 80°C are show in Figure
5.4. The image RMS roughness of these two Co ﬁlms were about 2.6 and 3 nm,
respectively. Note that these two samples were fabricated during a separate run and on a
different sapphire substrate (with different surface proﬁle) than the previous samples. The
150 nm thick Co ﬁlm deposited at high temperature (460°C) has a smoother surface than
low temperature deposited Co ﬁlm.

The overall results of the AFM studies (for all thiknesses) revealed that the ﬁlms
grown at 460°C mostly have larger island sizes and smoother surfaces. This is possibly
due to larger nucleation sites and coalescence between the islands. In section 5.2.3, the
orientation of these ﬁlms will be investigated using EBSP.
5.2.3.2 Effect of ﬁlm thickness on surface topography of the Co ﬁlms.

Since these Co ﬁlms were deposited at separate sputtering runs on substrate with
different surface proﬁle, the comparison of the surface topographies as a function of

thickness cannot be completely assessed and has therefore been omitted.

5.2.4 EBSP Analysis
The EBSP technique was used to examine the crystallographic nature of the sputtered
Co ﬁlms. The results of the EBSP data analysis have been categorized in a similar way

to the AFM analysis.
150

AFM Images of 150 nm Thick Co Film
Deposited at Substrate Temperature of 460°C and 80°C.

5.00

 

 

 

 

 

 

2.50
‘3
2'
" IWV'MVMAWMV‘M vv "WV MW‘UV" ”IPA/I)
D
8'.
I 2 50 5 DO

 

10.0

 

 

-10.0

 

 

 

Figure 5.4 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 150 nm Co ﬁlms grown on 20 nm epitaxial Cu(111) (grown at
150°C and post deposition annealed at 350°C) at substrate temperatures of: a) 460°C and
b) 80°C. The image RMS roughnesses of about 2.6 and 3.1 nm, respectively.

151

5.2.4.1 Temperature dependence, 80°C vs. 460°C

As discussed in Chapter 4 (Section 4.5.2.2 Limit of EBSP), EBSPs from 100 nm thick
(and thinner) samples were partially affected by backscatter contributions from the
underlying Nb, and some of the patterns could not be indexed. In the case of Co ﬁlms,
this same problem exists for 20 and 100 nm thick samples. Presumably, for ﬁlms thinner
than 150 nm the underlying epitaxial Cu (and Nb in 20 nm Cu thick ﬁlms) will contribute
signiﬁcantly to the patterns, but since Co and Cu may have the same crystal structure and
very similar lattice parameters, the EBSP patterns from the Co and underlying Cu can
merge and be indistinguishable. Therefore, only the EBSP study of 150 nm thick Co
ﬁlms is discussed.

EBSP data was collected randomly from different locations on both 150 nm thick Co
ﬁlms. Similar to the epitaxial Cu ﬁlms, the pole ﬁgures obtained from the EBSP data of
the 150 nm thick Co ﬁlms grown at 460°C revealed that two different orientation variants
were present. Both variants display <111> growth normals, consistent with the growth
occurring on the close-packed {111} planes in the fcc Co. Electron backscatter patterns
of the Co ﬁlms were very sharp for samples grown at a substrate temperature of 460°C.
The pole ﬁgure obtained from the EBSP data of the 150 um Co ﬁlms grown at 80°C,
however, revealed only one orientation variant. There was also substantial pattern
degradation for the samples grown at a substrate temperature of 80°C. Figure 5.5 shows
EBSP patterns of these two 150 nm thick epitaxial Co ﬁlms grown on 20 nm epitaxial Cu
at substrate temperatures of 460°C and 80°C. In ﬁgure (a) a sharp pattern from one of
the variants is shown. The two orientations were distinguished by a change in the

Kikuchi bands in the pattern. However, after careful observation of the patterns during

152

' "XV" .
Bill: 2“ 5‘.

1h! (,unlml

 

(b)
Figure 5.5 EBSPs from 150 nm epitaxial Co ﬁlms deposited on 20 nm epitaxial Cu ﬁlms

(grown at 150°C and post deposition annealed at 350°C) at substrate temperatures of: (a)
460°C and (b) 80°C.

153

the scanning, it was noticed that the patterns did not change during the scanning
processes and had relatively the same appearance (as shown in ﬁgure 5.5(b)). The
presence of these bands in the EBSP patterns indicate that both variants coexist within the
EBSP interaction volume and the indexing software does not reliably distinguish between
these bands.

All patterns from the 150 nm thick Co ﬁlms grown at 460°C were easily indexed as
consistent with the fcc structure and, under no circumstances, did any of the Co ﬁlms
display EBSPs consistent with the hcp structure. While the two variants could not be
distinguished in 150 nm ﬁlm grown at 80°C, the observed bands were consistent with
two fcc variants and no obvious evidence of the hcp structure was found.

To measure the size and distribution of the variants in the epitaxial Co ﬁlms grown at
460°C, the samples were mapped using stage control scanning. An orientation
distribution map of 150 nm thick Co ﬁlm grown on a 20 nm thick epitaxial Cu ﬁlm with
a substrate temperatures of 460°C is shown in Figure 5.6.

The epitaxial Co ﬁlms grown at 460° have a similar twin distribution to epitaxial Cu
(grown at an elevated temperature). However, based on the reviewed articles [1-12] for
the case of Co ﬁlms grown at 80°C, there is a possibility of the existence of hcp Co
clusters along with two fcc(lll) variants in these ﬁlms. To investigate this issue,
HRTEM is needed to identify and to characterize the sputter deposited Co ﬁlms, which is
beyond the scope of the present study.

The average grain size of the twin variants in the epitaxial Co ﬁlm grown at 460°C,

was approximately 5 pm with a relative grain area varying between 42% and 58%.

154

 

 

-=10 |.lrn; Mapl.‘ Step=l um; Grid=5

 

Figure 5.6. Orientation map of a 150 nm epitaxial Co ﬁlm grown at 460°C on 20 nm
epitaxial Cu ﬁlm (grown at 150°C and post deposition annealing of 350°C) showing the
distribution of the twin variants.

155

5.2.5 Growth Process

Based on the information obtained from the literature review and experimental AF M
and EBSP results, the following conclusions can be made regarding the growth processes
in the sputter deposited Co ﬁlms:
At high temperature (460°C): This is a hetero-epitaxial growth (Co grows on Cu with
same orientation but has different lattice parameter). The size of islands appear on the
surface topography (from AFM analysis, for example ﬁgure 5.3(a)), and size and
distribution of the twin variants (from EBSP analysis, for example ﬁgure 5.6), are
suggestive of the SK growth mode. The formation of two variants (twin) is similar to the
epitaxial Cu ﬁlms. According to literature review, these twins could follow the
underlayer Cu twins structures or form independently. A HRTEM study is needed to
resolve this issue.
At low temperature (80°C): At low temperature, the arriving atoms do not have enough
time to ﬁnd or move toward a step edge or nucleated sites. Rather nucleation take place
spontaneously as an island on top of each other. This causes the formation of a misﬁt
distortion network within the ﬁlm and thus a rougher surface and strain within the ﬁhn.
The smaller size islands seen in surface topography (see ﬁgure 5.3 (b)) and the existence
of very small twin variants (not resolved in EBSP) as well as pattern distortion, are
supportive of the given conclusion which was obtained from the literature review. The
distortion of the EBSP patterns would be the result of such strain. Again HRTEM study

is needed for possible clariﬁcation.

156

5.3 GROWTH OF Py FILMS ON EPITAXIAL Cu(lll)
5.3.1 Nucleation and Growth --- a Review:
Background Permalloy (Py) is often substituted for Co in magnetic multilayers and is
one of the ideal and frequently used ferromagnetic materials in both types of hybrid and
exchange-biased spin-valves. Py is an alloy of Ni and Fe (NIgFCg) with a fcc derivative
L12 structure. A typical exchange-biased spin-valve can be grown in either of the
following orders.
ONb-Cu-FeMn-Py-Cu-Py : note the ﬁrst Py must be deposited on FeMn.
ONb-Qw-Cu-Py-FeMn : note the ﬁrst Py must be deposited on Cu.
In polycrystalline samples, the order is not very critical. However, in epitaxial samples
the substrate (buffer layer) may be a crucial issue.

The growth of Py on Cu(1 l 1) (or FeMn(111)) should accrue by simple, hetero-
epitaxial growth of an fcc metal on top of another fcc metal, with relatively small lattice
mismatch (1.3%). Therefore, a well-deﬁned pseudomorphic layer-by—layer growth up to
a large number of monolayers is expected [2]. Unfortunately, the nucleation and growth
of Py has not been investigated by researchers.

In the next two sections, the experimental results of the AF M surface topography
analysis and EBSP orientation distribution analysis are presented. The growth process
then will be discussed based on the outcomes of the AF M and EBSP analyses.

The crystallographic and physical properties of Py and Cu are summarized in Table 5.2.

157

Table 5.2

 

 

 

 

 

 

 

 

 

CRYSTALLOGRAPHIC DATA
Alloy or Element: Pym-Permalloy Cu-Copper
Crystal system: Cubic — fcc L12 Cubic — fcc
a=b=c 0.352nm 0.36148nm
Unit cell volume: 0.0447 nm3 0.0467 nm3
# of atom in unit cell: 4 4
Atomic radius: Py = *NA Cu = 0.128 nm
(I am) dun) = 0.205 nm dun) = 0.2078 nm
Symmetry elements: Four 3-fold rotation axes Four 3-fold rotation axes
Melting point: 1480 °C 1083 °C

 

 

 

 

 

*NA : Py (NigoFe2o) has a fcc derivative L12 structure and the dose not have a single

value for the atomic radius.

158

5.3.2 AFM Analysis

For this study, a few 200 nm thick Py ﬁlms were deposited on 20mm epitaxial Cu
ﬁlms at substrate temperatures of 350°C and 90°C. The surface topography of each ﬁlm
was examined using AF M. The surface topographies of the sputtered Py ﬁlms were
affected by the growth temperature and substrate topography. The results of the AF M

analysis are categorized as follows:
OEffect of the substrate temperature on the surface topography of the Py ﬁlms.

OEffect of the buffer layer on the surface topography of the Py ﬁlms.

Effect of the substrate temperature on surface topography of the Py ﬁlms.

The AF M analysis showed that Py ﬁlms had surface morphologies similar to epitaxial
Co ﬁlms. Two sets of 200nm thick Py ﬁlms were deposited at 350°C and 90°C on 20 nm
thick epitaxial Cu ﬁlms. The epitaxial Cu ﬁlms were grown at a substrate temperature of
150°C and post deposition annealed at 350°C. Figure 5.7 shows atomic-force-
microscope images and corresponding line trace surface analyses of 200 nm thick Py
ﬁlms grown on 20 nm epitaxial Cu ﬁlm at substrate temperatures of 350°C and 90°C.
The Py sample grown at 350°C has relatively large islands, while the sample grown at
90°C seems to be very rough with very small grains. In Figure 5.8, the AF M images and
the corresponding line trace surface analyses of 200 nm thick Py ﬁlms grown on 20 nm
epitaxial Cu ﬁlms at substrate temperatures of 350°C and 90°C (deposited during
different sputtering runs) are shown. The Py sample grown at 350°C has relatively
smooth surface topography, while the sample grown at 90°C seems to be rougher with

many small interconnected islands.
1 59

AF M Images of 200 nm Thick Py Films
Deposited at Substrate Temperatures of 350°C and 90°C.

5.00

 

 

 

 

 

 

 

 

 

 

Figure 5.7 AFM images and line trace surface analysis of 200 nm thick Py ﬁlms
deposited on 20 nm epitaxial Cu ﬁlms at substrate temperatures of 350°C and 90°C. The
epitaxial Cu ﬁlms were deposited at 150°C and post deposition annealed at 350°C.

160

AFM Images of 200 nm Thick Py Films
Deposited at Substrate Temperatures of 350°C and 90°C.

 

 

 

-10.0
I

 

 

 

 

 

 

-10.0

 

 

 

n
p
O
O
N
D
O
D
:4
D
A
O
O
"I

(b)

Figure 5.8 AFM images and line trace surface analyses showing the surface topography
of 200 nm Py ﬁlms grown on 20 nm epitaxial Cu(lll) (grown at 150°C and post
deposition annealed at 350°C) at substrate temperature of: a) 350°C and b) 90°C. The
image RMS roughnesses are about 2 and 9 nm, respectively.

161

Effect of buffer layer on surface topography of the Py ﬁlms.

To investigate the effect of buffer layer on epitaxial Py ﬁlms from crystallographic
point of view one sample was grown with 8 nm of FeMn deposited on a 20 nm epitaxial
Cu ﬁlm at a substrate temperature of 90°C prior to the deposition of 200 nm thick Py
ﬁlm. The AF M results showed that the 200 nm Py deposited on the 8 nm of FeMn had
a very smooth surface compared to those deposited directly on a Cu substrate. An AF M
image and line trace surface analysis of 200 nm thick Py ﬁlms deposited on Cu ﬁlm
(grown at 150°C and post deposition annealed at 350°C) (a) and FeMn ﬁlm (b) are shown

in Figure 5.9.

5.3.3 EBSP Analysis

Similar to the EBSP analysis of Co ﬁlm, the EBSP analyses of Py ﬁlms grown at a
substrate temperature of 350°C revealed that two different variants are also present in
these ﬁlms. Both variants display <111> grth normals, consistent with growth
occurring on the close-packed {111} planes in the fcc Py. These epitaxial Py ﬁlms have
similar variant distributions to those of the epitaxial Cu and Co ﬁlms. Figure 5.10 shows
EBSP patterns of these two 200 nm thick epitaxial Py ﬁlms grown onto 20 nm epitaxial
Cu at substrate temperatures of 350°C and 90°C. In ﬁgure (a), a sharp pattern from one
of the variants is shown. Similar to Cu and Co, the two orientations were distinguished
by a change in the Kikuchi bands in the pattern. In ﬁgure 5.11 an orientation map of the
distribution of the two variants in a 200 nm thick epitaxial Py ﬁlm deposited at 350°C is
shown. The average twin size in epitaxial Py ﬁlms grown at high temperature was about

5 pm, and a typical relative grain area varies between 44% and 56%.

162

AF M Images of 200 nm Thick Py Films
Deposited on Different Buffer Layer

 

10.0
1

 

 

-10.0
1

 

 

 

( 2:50 5. 00

 

10.0

 

 

f.

 

-10.0

 

 

(ii)

Figure 5.9 Atomic-force-microscopy images and line trace surface analyses showing the
surface topography of 200 nm Py ﬁlms grown at 90°C on a) 20 nm of epitaxial Cu(1 l 1)
(grown at 150°C and post deposition annealed at 350°C and different substrate than in
ﬁgure 5.6). b) 8 nm of FeMn. The RMS roughnesses are about 2 and 0.9 nm,
respectively.

163

.rted Donna

 

 

   

     
     

yrlr 1 IlII

' 41.1th-0- oimolon 3:35:14!

4WN' Eel-v: n 3! n
M. W W ”5
inn-mud ’ MIN"

WW. 3 W112"; HLWW.

   
  

 

 

 

(Ivar ( umml

 
    

b
> in- mo- Oriana”!
> Sm: Inna- :

"Q'.'.’.‘1.“"" i Educ m « 22

' Wad MAD: OJ!

.- s.“_"."',m : M a] <54
_.__ l __'ﬂ99: 0.34 a“

   

 

 

(b)

Figure 5.10 EBSPs from 200 nm thick epitaxial Py ﬁlms deposited on 20 nm thick
epitaxial Cu ﬁlms (grown at 150°C and post deposition annealed at 350°C) at substrate
temperatures of: (a) 350°C and (b) 90°C.

164

 

-=10 1.1m; Mapl.‘ Step=1 um; Grid=5

 

Figure 5.11. Orientation map of 200 nm thick epitaxial Py ﬁlm grown on a 20 nm thick
epitaxial Cu ﬁlm (grown at 150°C and post deposition annealing of 350°C) at a substrate
temperature of 350°C showing the distribution of the twin variants in the Py.

165

The EBSPs of Py ﬁlms grown at a substrate temperature of 90°C showed signiﬁcant
pattern degradation, which is similar to that observed in the Co ﬁlms (ﬁgureS-lOCb».
However, all of the patterns were easily indexed for the fee structure. The distortion and
degradation of patterns is again likely caused by strains in the ﬁlm resulting from an
increase in the numbers of misﬁt dislocations, threading dislocations, and point defects.
Since the two orientation variants are distinguished by a change in the Kikuchi bands in
the patterns, mapping of the two orientation variants was not successful for 90°C ﬁlms,
however, because of pattern degradation and presence of Kikuchi bands from both

variants.

Effect of the buffer layer on Py ﬁlm EBSPs of Py ﬁlms grown on Cu at 350°C and

90°C are shown in part (a) and (b) of Figure 5.12. An EBSP of Py ﬁlm deposited on
FeMn at room temperature is shown in part (c) of the same ﬁgure. In the case of
Py/FeMn, EBSP patterns from the Py ﬁlm grown onto 8 nm of FeMn at room
temperature were relatively clear and sharp. This increase in Py ﬁlm pattern sharpness
indicated that there was less strain in the Py ﬁlm grown on FeMn ﬁlm than those grown
on Cu ﬁlm [14].

The released of strain in Py ﬁlm deposited on FeMn buffer layer had been also
observed by Tang et al.[14], using HRTEM and XRD. Their results revealed that the
misﬁt strain between the F eMn layer and Py is released by the formation of a dome shape
on the FeMn surface rather than by the formation of misﬁt dislocations at the interface

between the two layers.

166

EBSP of Epitaxial Py ﬁlm on Epitaxial Cu and FeMn

 

(a)

 

(b)

 

(C)
Figure 5.12 EBSPs from 200 nm epitaxial Py ﬁlms deposited on 20 nm epitaxial Cu
(grown at 150°C and post deposition annealed at 350°C) at substrate temperature of: (a)
350°C and (b) 90°C. (c) a EBSP pattern of a 200 nm epitaxial Py ﬁlm deposited on 8 nm
of FeMn (grown at 90°C) at a substrate temperature of 350°C.

167

5.4 DISCUSSION

Unfortunately, in all three cases of Cu, Co, and Py ﬁlms grown at an ambient
temperature (below 100°C), the indexing software does not reliably distinguish between
mixed Kikuchi bands, which originated from both variants. There was also some degree
of distortion in all three cases. However, the distortions of the EBSP patterns of Co and
Py grown at 90°C were more severe than those in Cu ﬁlms grown at 90°C. There are a
few possible factors that could cause such a distortion. Among them are surface energy
of each layer, lattice mismatch between two consecutive layers, and atomic radius of each
deposited material. Each of these factors could affect the interface energy, the nucleation
and growth rates of the islands, and the density and type of dislocations and point defects.

The surface energy, lattice mismatch, and atomic radius of Nb, Cu, Co, Py are
tabulated in table 5.3 for comparison. To have more data for comparison, two Ag ﬁlms
have also been grown on epitaxial Nb substrate, similar to Cu ﬁlms, one at 350°C and the
second one at 90°C. Ag has different surface energy, lattice mismatch (with Nb), and
atomic radius than Cu or C0. The value of surface energy, lattice mismatch, and atomic
radius of Ag are also included in the table 5.3. The details of this experiment are given in
appendix 5.A. EBSP analyses of the epitaxial Ag ﬁlms grown at 350°C revealed two
different (111) orientation variants, and the NW epitaxial relationship with the Nb
substrate. The variant distributions were compared to those found in the epitaxial Cu
ﬁlms grown at 350°C. Again, the variants did not display any particular pattern or
morphologies. The EBSP patterns of Ag ﬁlm grown at a substrate temperature of 90°C

showed some degradation, more similar to Co and Py cases.

168

Table 5.3

Comparison of Lattice Mismatch, Atomic Radius, and Surface Energy

 

 

 

 

 

of Nb, Cu, Ag, Co, and Py
System Surface energy Lattice mismatch Atomic radius
Nb-Cu 3—1.9 J/m2 12% 0.143nm-0.128nm
Nb-Ag 3—1.3 Mn2 1% O.l43nm-0.l44nm
Cu-Co 1.9—2.7 Mn2 2% 0.128nm-0.125nm
Cu - Py 1.9 — ~2.7 J/m2 1.3% 0.128 nm -

 

 

 

 

 

 

Comparisons of different variables in Table 5.3 suggest that the surface energy of a
given material are effecting the EBSP patterns. It seems that when the surface energy of

the overlayer is greater than surface energy of the substrate, the degradation in the EBSP

will appear.

169

5.5 SUMMARY

In this chapter, the nucleation and growth of two ferromagnetic materials (Co and Py)
on epitaxial Cu ﬁlm were reviewed. All of the sputter deposited ﬁlms were characterized
using AF M and EBSP techniques. The AF M results demonstrated that the surface
topography is affected strongly by the temperature and surface quality of the substrates.
EBSP results revealed that all Co and Py ﬁlms grown at high temperature display two in-
plane epitaxial variants corresponding to two stacking sequences of {111}fcc planes.
The EBSP analysis of Co and Py samples grown at 90°C revealed that two different
orientation variants are present. However, due to degradation of the associated EBSP
patterns and presence of Kikuchi bands from both variants, mapping of the orientation

distributions was not possible.

170

APPENDIX 5.A

Ag has a fee structure and the Ag/Nb lattice mismatch is 1% with an atomic radius
difference of 0.7% (see table 5.4). Ag can grow with the NW or KS epitaxial relationship
onto an epitaxial Nb substrate. A properly superimposed atomic arrangement (unrelaxed)
of Ag on Nb (with the NW relationship) is shown in Figure 5.A.1.

Similar to growth of epitaxial Cu ﬁlms, a few Ag ﬁlms were grown on 250 nm
epitaxial Nb at substrate temperatures of 350°C and 90°C. The EBSP results of Ag ﬁlms
grown at 350°C and 90°C were similar to the epitaxial Cu. Figure 5.A.2 illustrates the
EBSPs of epitaxial Ag grown at 350°C and 90°C. Figure 5.A.3 shows the orientation
distribution map of 300 nm epitaxial Ag grown on 250 nm epitaxial Nb. The epitaxial
relationship between the sputter deposited epitaxial bcc Nb and fcc Ag ﬁlms, was found
by removing part of the Ag ﬁlm and collecting EBSP data from an adjacent section of Nb
and Ag. The resulting (100), (110), and (111) pole ﬁgures were used to plot the
stereographic projection of both Nb and Ag. By superimposing the Ag stereographic
projection on the Nb, the result suggests that (110)bchb || (111)fccAg, with [1 T O]bchb ||

[1 T 2]rccAg and [ T 0 O]bchb || [0 1 T]fccAg, which agrees with the NW relationship.

171

Table 5.4

 

 

 

 

 

 

CRYSTALLOGRAPHIC DATA
Element: Nb—Niobium-Columbium Ag-Silver
Crystal system: Cubic — bcc Cubic — fee
a = 0.33067 nm 0.4086 nm
b = 0.33067 nm 0.4086 nm
c = 0.33067 nm 0.4086 nm
Unit cell volume: 0.0359 nm3 0.0682 nm3
# of atom in unit cell: 2 4
Atomic radius: Nb = 0.143 nm Ag = 0.144 nm

 

d (h.k,|)

(1(110) = 0.2337 nm

0“”) = 0.2359 nm

 

Symmetry elements:

Four 3-fold rotation axes

Four 3-fold rotation axes

 

 

Melting point:

 

2469 °C

 

962 °C

 

172

 

 

 

 

        

   
  
 

    
  
 

 

.0 .0 :0. .0 .0.0.0 _ 0.0.0. 0 .07 0. .0:0:0.0.0.::0.0.0.0.0.0.0.0.0.0 .
m000.0:0 _0 00 ”0 “000 000 0
.0 00 .0000 .00

W00000 000.000 0..:0000 00.:
0.0000000000 0. 0.000~00 0 0

0.0.:.00 00000 0. .00 0.000 00000
0.00 000000000000 00:..000000
000.0 0m:.000000000000000 00 0~0
0.000.0'00 000000 0000 .00n’.000."000 00
.0 0. 0.0 0.0.0 00 0.0.0.0. 0.0.0.0.0 .0.0.0 .0.0 .0.0:0

"00 '00 00000 00000 00 ”0

0 _0.0. 0

    
  

            
           
 
     
 

   
 

  

-0

0‘:‘ 0. 1

'0:0':.:::. .0. .0:0-’0’
'0 0 0
.0 0 0 0

 
 
    
 
 

’00‘
0:‘0 0’ 0 0. 0 0 0
0000000 000000000
0'0:0.0:0’0:0:0:0:0° ’0 ‘- ‘0:0:0:0.0:0
0'0 "00 0 ..00 0 00 ’0
000 0000 0.0:0:0000
0 0 0
0.0. 0 0:0: ' ’ ‘

 
 
        
  
 
 
 

’0
0000000°'.000
0.0:0:0:0 0’ 0:0°0‘0‘ .0“ 0:0

000

  

 
   

   

0
00.0 0:0:0: .E:::0:0 .0.0
“00"”:0 00.0 0
.0- 0 :0.0 .0: 0 .0.0.00

90.0.
“0'0 0

    

:' :0’0 0'0. 0

0:0 .0:0: 0:0:0:0:0’0 ”0.0 0

:0.0.0:::0:: ‘0 0 0
0'

   

”0 0.

’0'0‘0:'00’$: :-0000:0::’ 00.0.0.

0 0.0.0.0.0:0:0:0.0:0:0:0.0 0:0°0:
0: 0'0 00 0 0 0 0 0 0 0' 0'

0° 0 0.0. 0

       
  

   
 
 
 
 

  

    

    
 
  

     
 

  
 

0.0:0:0.:. . ...
00-0 ’0 00" :0°0’00°°"
000 ".0000'0 0'0 0.0
000 :000000 00 '0 00
0000090:00'0 W000 :0000
0:0‘0-00000 00° ‘0’.“00000 000
00 "’°‘00°':0"0:"°’ 00 00:000
0.0.0 0 0 0.0 :0 ..0 .0 . 0:0. 0 .0:0.0.0.0.0 .0.0.0
0.000 0.00000.. :0.0.0
::00
0'0 ‘0‘0' 0’

    

      

   
   
 

‘. .'0‘:‘0’0'0'0 "00 .0'0' :'0:::0' '. 00' 0°:'00 00’ 0 0 0
00000000000000000000000
00. 00 000m 00 00

0:0:0 :0’000 ”00 .0.000 °°”"'00'
0.000000 :..000’.00000 0
000,-00 0.00 000 0m0000

. 0 0‘ 00~°00000000‘
0000000 .0000

0000 0:::0000 000
0_ 0000
“0:000:0:

     
                
 

        
 
 

 
  
 

  
     
  

      
 

     

.0. 0:0 ,.
7 2 _ '00” .0: 0.0 0 0 0. 0 ..00
’0- 0'0. 090.0:0' 0' 0" 0° 0':'0°:':'0° "" °:0 0 0:0 ’0’0’0 ’.° 0:0°:‘0
.0. 0:0 ., 0:. 0. 0 :0.0.0.0.0.0. 0.0.0.0.0.0.0.0.0.0.0.0.0.0.0. .0.0.
0'0 000 50.0.0” .0. .0000 '.00
0 0 0: .0 0‘ 0‘ 00. 0‘ 0:

0.0 0 0 0. 0: 0 0
0 ~ .0” , - .00 .0. 0. 0.0 00
0' 0 , , . 0.0.0:0 .0.0.:.0 . 0:0. 0.0.0.0
0 0 0:0.0.

      

       
     
  

   
  
 

  
 

          

        
   

.0 0 0 0 0
0 0 0:0: 0:0 0 0.0.0.0. 0 . 0 0'0
0'0 0 0 0.0.0 0’0 0 0’0‘0 ’0'0’0‘0'0’0‘0'0 0‘0‘0-0 00‘ ‘ 0. 0'0
00000-0 °‘0,’00.00.°000 0000000'H000’0:0’0::00 0
‘0‘0’0’0-‘0’ 0'0 0’0. 0’0 ’0‘0 ’0 ° 0°0°0°0°0°0'0:0 0’0 00' '0’0’0 ’0’0’0’
0.0.0. 0- .0.0.0.0.— W0 0’ 0 0 0. 0 0- .0 0.0:0. 0.0:0.0.0.0
000”.00.000' °:000000000

     

Figure 5.A.1. A superimposed atomic arrangement (unrelaxed) of fcc Ag (111) on bcc
Nb (110) (with N-W relationship and 1.1% mismatch).

173

 

 

(b)

 

(C)

Figure 5.A.2 EBSP patterns of a 300 nm epitaxial Py film deposited on 200 nm epitaxial
Nb at substrate temperature of: (a) & (b) 350°C and (c) 90°C. Patterns in part (a) & (b)
are the EBSPs of twin areas.

174

_-10 um; Mapl; Step-0.5 um; Grid-101x101

 

 

Figure 5.A.3. Orientation maps of 200 nm epitaxial Ag ﬁlms grown on epitaxial Nb at
substrate temperature of 350°C. The green areas are unsolved points (loss of EBSP
signal).

175

5.5 REFRENCES:

1) M. Zheng, J. Shen, J. Barthel, P. Ohresser, Ch V Mohan, and J. Krischner, J.
Phys.:Condens. Matter 12 783 (2000).

2) K. Heinz, S. Muller, and L. Hammer, J. Phys.: Condens. Matter, 11, 9437 (1999).

3) D J. Smith, AR. Modak, T.A. Rabedeau, and S.S.P. Parkin, Appl. Phys. Lett., 71.
1480 (1997).

4) GR. Harp and S.S.P. Parkin, Thin Solid Films 288, 315 (1996).

5) B. P. Tonner, Z.L. Han, and J. Zang, Phys. Rev. B, 47, 9723 (1993).

6) M. T. kief and W. F. Egelhoff, Jr, Phys. Rev. B, 47, 10785 (1993).

7) J. de la Figuera, J. E. Prieto, C. Ocal, and Miranda, Phys. Rev. B, 47, 13043 (1993).

8) K. Brdhl, P. Bodeker, N. Metoki, A.Stierle, and H. Zabel, J. Crystal Growth, 127,
682 (1993).

9) P. Bodeker, A. Abromeit, K. Brohl, P. Sonntag, N. Metoki, and H. Zabel, Phys. Rev.
B, 47, 2353 (1993).

10) J. Camarero, L. Spendeler, G. Schmidt, K. Heinz, J. J. de Miguel, and R. Miranda,
Phys. Rev. Lett. 83, 2448 (1994).

11) J. Camarero, T. Graf, J. J. de Miguel, G. Miranda, W. Kuck, M. Zharnikov, A.
Dittschar, C. M. Schneider, and J. Kirschner, Phys. Rev. Lett. 76, 4428 (1996).

12) Ch. Rath, J. E. Prieto, S. Mﬁller, R. Miranda, K. Heinz, Phys. Rev.B, 55, 10791
(1997)

13) T. Lyman, Metals Handbook, The American Society for Metals, (1958)

14) L. Tang, D. E. Laughlin, and S. Gangopadhyay, J. Appl. Phys., 81, 4906 (1997).

15) A. J. Wilkison, and D. J. Dingly, Acta metall. Mater. 40, 3357 (1992).

176

CHAPTER 6

MAGNETO-TRANSPORT PROPERTIES OF EPITAXIAL

MAGNETIC FILMS AND SPIN-VALVES

This chapter focuses on the magnetic and magneto-transport properties of epitaxial Py
based spin-valves. The effect of spin-valve growth conditions on important device
properties, including: (a) exchange bias ﬁeld, (b) coercive ﬁeld, and c) establishment of a
well deﬁned antiparallel state, have been investigated. To develop a better understanding
of the magneto-transport properties of epitaxial Py based spin-valves, this chapter will
ﬁrst present an overview of magnetic properties of single Py ﬁlms with and without an
FeMn buffer layer. After this foundation is established, the magnetic characterization of
single epitaxial Py ﬁlms and the magneto-transport and magnetic characterization of
epitaxial spin-valves are presented. In each of these sections, a brief introduction will be
followed with the experimental procedures and results. This chapter concludes with a

brief summary.

6.1 MAGNETIC PROPERTIES OF FEROMAGNETIC FILMS
Hysteresis Curves:

In ferromagnetic materials, the application of a sufﬁciently large external ﬁeld in the
direction opposite to the initial direction of magnetic domains will force the domains to
reverse their direction. By continuously changing the external magnetic ﬁeld from a

large positive value to a large negative value and back to the original positive ﬁeld, a

177

magnetic hysteresis loop will be created. Figure 6.1(a) depicts a typical ferromagnetic
hysteresis loop. The coercive ﬁeld, He, (points C and E) is the ﬁeld at which the total
magnetization M is zero. If the hysteresis loop is not symmetric about zero, Hc is usually
deﬁned as half the width of the hysteresis loop measured at M=O.

Pinning of the Ferromagnetic Layer:

If a ferromagnetic layer is placed next to an antiferromagnetic layer, the magnetization
of the ferromagnetic layer can be “pinned” in a chosen direction by cooling the bilayer
from above the Neel temperature of the antiferromagnet to room temperature in a
sufﬁcient large magnetic ﬁeld. The pinning direction of the pinned ferromagnetic layer is
referred to as the easy axis. The exchange anisotropy (known as ‘unidirectional
exchange-bias’ [1]) due to pinning between the antiferromagnetic and ferromagnetic
layers leads to a shift of the magnetization hysteresis loop along the magnetic ﬁeld axis
by HE the exchange—bias ﬁeld, and also usually increases the coercive ﬁeld, He, in the
hysteresis loop (see Figure 6.1(b-blue) [1-4]. After pinning, it is energetically favorable
for the ferromagnetic layer to point in one particular direction, imposed by interfacial
coupling with the adjacent antiferromagnetic layer. To overcome this energy, an external
magnetic ﬁeld, equal to HE is required. Deﬁning H1 and H2 as shown in Figure 6.1 (b),

the exchange bias ﬁeld, HE, and the coercive ﬁeld, He, of the pinned layers are given by:

 

 

: HI+H2 and HC: th-HZ-

HE
Berkowitz et a1. [2] showed that the exchange-bias ﬁeld is inversely proportional to

the thickness of the pinned ferromagnetic layer: HE at 13:.

178

Magnetization

H
ll

 

 

 

—> ->
-> —> ->
4.1:
B
1" <—,
\‘ C E Magnetic Field
+HC
(a)
Magnetization
(arb. units)

 

H (Oe)

 

 

 

 

 

(b)

Figure 6.1. (a) A typical hysteresis loop of a ferromagnetic material (such as Fe, Co, Ni
or their compounds). (b) Magnetization hysteresis loops for a free Py layer (red) and a
Py-FeMn bilayer (blue) after pinning. The exchange-bias ﬁeld (HE) and the coercive
ﬁeld (Hc) of the pinned Py are indicated in the ﬁgure.

179

Several theoretical models predict that the exchange-biased ﬁeld should increase if the
antiferromagnetic layer contains small domains [5-9]. The random-ﬁeld model by
Malozemoff [8, 9] shows that small domains are typically produced by crystal disorder
within the antiferromagnetic layer. This model also predicts that HE increases with
interfacial disorder, while a perfect crystalline antiferromagnetic layer and smooth

interface with the ferromagnetic layer can reduce the exchange bias [9].

6.2 MAGNETIC PROPERTIES OF EPITAXIAL Py FILMS
6.2.1 Introduction

The crystallographic characterization of 200 nm thick epitaxial Py ﬁlms sputter
deposited on epitaxial Cu and FeMn ﬁlms was discussed in chapter 5. These Py ﬁlms
were grown at 350°C or 90°C on 20 nm thick epitaxial Cu buffer layers or on 8 nm thick
FeMn ﬁlms In the present section, the magnetic properties of these Py layers are
discussed.

From a magnetic point of view, Py is a soft ferromagnetic alloy because only a small
external ﬁeld is needed to rotate its magnetic moment. If a thin Py ﬁlm is deposited on
an antiferromagnetic material (such as FeMn), it can be pinned as described in section 6.1

(see section 6.1 and Figure 6.1(b- blue) [2-4].

6.2.2 Experimental Procedures
As described in chapter 2, an MPMS Quantum Design magnetometer was used to

measure the magnetization of the epitaxial Py ﬁlms deposited in this study to carry out

180

these measurements. A small portion (5 mm x 5mm) of a sample was cut and placed in
the magnetometer, with the plane of the ﬁlm parallel to the applied ﬁeld.

Magnetic anisotropy in epitaxial Py ﬁlms was studied by measuring the magnetization
of these samples at zero and 90 degrees to the direction of the applied ﬁeld. Note that the
samples had to be taken out of magnetometer to be rotated 90 degrees around the axis
perpendicular to the surface of the ﬁlm.

The magnetization measurements were carried out at a temperature of 12 K. At this
temperature the Nb buffer layer is not superconducting and does not affect the
magnetization measurements. Details about the transport and magnetic properties of

epitaxial Nb buffer layer are given in appendix 6.A

6.2.3 Results
6.2.3.1 Magnetic Properties of Epitaxial Py Films deposited on Epitaxial Cu Film.
The magnetization hysteresis loops of two sputter deposited epitaxial Py ﬁlms (200
nm thick) grown at 350°C and 90°C as a function of magnetic ﬁeld are shown in Figure
6.2(a). The coercive ﬁeld of the epitaxial Py ﬁlm grown at high temperature is small
(approximately 12 0e), while the Hc of the sample grown at ambient temperature (90°C)
is much larger, about 50 Oe. This difference may suggest that the epitaxial Py ﬁlm
deposited at 90°C has lower crystalline quality and/or may have signiﬁcant strain most
likely due to structural defects. Recall that the electron backscattered patterns (EBSPs) of
epitaxial Py ﬁlms grown on epitaxial Cu ﬁlms at 350°C were very sharp, while the
EBSPs of samples grown at 90°C were distorted in a manner consistent with strain

caused by misﬁt dislocations, threading defects, and point defects (see section 5.3.3).

181

Nb—Cu-Py
250-20-200 nm

0.0% 'IIIIIIIIIIIIIIIIIIIIIIIIIIIVII

 

0.00375 T = 350°C

\i
l | I
I

lllllllllIllllllllllllLllll

0.0025

 

0.00125

-0.00125

Magnetization {emu}
0

-0.0025

llllllvlllllllllllIIIIIIIYT

t.

-0.00375

(a)
'o.ow’L‘Lll“lll'lllllllllllllriilii
-500 -375 -250 425 0 125 250 375 500
Magnetic Field (09)

 

 

..

 

 

 

 
    

0005 ...,...,.r.T . 0.005_...,...,...........,...,...,..r
i I q.
“”375 . 1‘ 0.00315 — . (ltlcglcc
I 1 : ( . .n 1
0.0025 _— _‘ 0.0025 __ . )(ltlcgicc d

0.00125 '

Magnetization (omui
O

41.00125

 

 

 

 

 

 

 

43.00125
0.0025“ 0.0025 J
. b l I
- ) 1 (c) ,
000375 , 0.00375 d
no“'...1...l...1L._L1...|...I.lli.L. am’...i...1...1...|...i...|...|...
.500 (575 450 -1 25 0 13 250 375 500 -500 415 450 -1 28 0 1 3 2” 375 500

Magnetic Pick! (00) Magnetic Field (00)

Figure 6.2 The magnetization hysteresis loops of 200 nm thick epitaxial Py ﬁlms grown
on 20 nm thick epitaxial Cu at substrate temperatures of 90°C (red) and 350°C (green). a)
The coercivity (Hc = half width of magnetization loop) increases from 12 to 50 Oe for
samples grown at low temperature (90°C). b & 0) Comparison of the hysteresis loops of
both samples when the samples are rotated 90 degrees with respect to the ﬁrst orientation
within the sample plane. The hysteresis loops overlapped for both samples, therefore
there are no magnetocrystalline anisotropy in theses two epitaxial Py ﬁlms.

182

Magnetic anisotropy:

The magnetic anisotropy (in particular, magnetocrystalline anisotropy) of these two
epitaxial Py ﬁlms was measured by comparing the magnetization of each ﬁlm for two in-
plane orientations. The difference in magnetic properties in different directions in
magnetic materials is called magnetic anisotropy. Magnetocrystalline anisotropy is the
tendency of the spins in the ferromagnetic materials to align in a particular direction
imposed by the crystal lattice. Similarly, shape anisotropy is the tendency for these spins
to align due to sample shape. In cubic magnetic materials, there should be no in-plane
magnetocrystalline anisotropy in the (111) plane [10]. Therefore, the magnetic properties
of a magnetic ﬁlm with [111] growth direction should not be different if measured in the
[110] and [1T2] directions. The results of such measurements are shown in Figure 6.2(b
&c). The overlapping hysteresis loops for 200 nm thick epitaxial Py ﬁlms grown at

either 350°C or 90°C reveal no magnetocrystalline anisotropy for either epitaxial Py ﬁlm.

6.2.3.2 Magnetic Properties of an Epitaxial Py Film deposited on an Epitaxial FeMn
Film.

In Figure 5.12(c) of section 5.3.2, the effects of the epitaxial Cu and FeMn buffer
layers on the crystallography of the epitaxial Py ﬁlms were discussed. It was shown that
the strain in the Py-FeMn bilayer is lower than that observed in the Py-Cu bilayer. Here
it will be shown that the epitaxial Py ﬁlm grown on a FeMn ﬁlm also exhibited different
magnetic behavior than that grown on a Cu ﬁlm. Magnetization measurements were

performed at 12 K on a 200 nm thick epitaxial Py ﬁlm grown at a substrate temperature

of 90°C on 8 nm thick FeMn. The magnetization hysteresis loops for this sample are

183

shown in Figure 6.3(a). The coercive ﬁeld (He) was 7.5 Oe in one direction but
decreased to 4 0e when it was rotated 90 degrees. These low coercivity ﬁelds suggest
that the ferromagnetic (Py) layer grown on an FeMn ﬁlm contains less strain compared to
the Py grown on Cu ﬁlm (compare 4 0e and 7.5 Oe for Py on F eMn to 50 Oe for Py on
Cu) but also suggest the possibility of a magnetic anisotropy.

The data of Figure 6.3(a) have two distinguishing characteristics. (1) The hysteresis
loop collected at zero degrees (green) is slightly shifted from zero ﬁeld. (2) The
magnetizations are the same in the two different directions, but the saturation and
coercive ﬁelds are different. Generally, the presence of different saturation and coercive
ﬁelds in hysteresis loops for different directions in the ﬁlm is an indication of magnetic
anisotropy. A shift in the hysteresis loop could also be caused by an exchange-bias ﬁeld
between the Py and FeMn layers (as described in section 6.1). However, as no external
magnetic ﬁeld was applied and the Py layer is relatively thick, HE should be very small
[2]. One possible source of this anisotropy is that the small ﬁinge magnetic ﬁeld of the
sputtering gun magnets produced a weak exchange-bias interaction between the FeMn
and Py layers, resulting in some degree of inadvertent pinning occurred during growth.
To determine if this in fact occurred, this sample was taken through the usual pinning
procedure (pinned along the zero direction), and the magnetization measurements were
again performed. The resulting magnetization hysteresis loops are shown in Figure
6.3(b). The hysteresis loop collected at zero degrees again has a small shiﬁ, but the
coercive ﬁelds (about 4 0e) are now the same in both perpendicular directions.
Considering the thickness of this Py ﬁlm, this small shiﬁ suggests a very weak local

exchange bias ﬁeld between the epitaxial Py and its epitaxial FeMn buffer layer.

184

szso'Cuzo'FeMﬂs‘PY2oo nm

0.0012
0.0009
0.0006

I 0 degree
°-°°°3 90 degree

-0.0003

Magnetization {emu}
0

-0.0006

-0.0009

 

-0.0012
-100 -50 0 50 100

Magnetic Field (Ce)

0 degree
0 90 degree

Magnetization {emu}

 

-100 -50 0 50 100
Magnetic Field (0e)
Figure 6.3 The magnetization hysteresis loops of a 200nm epitaxial Py ﬁlm grown onto
8 nm epitaxial FeMn at room temperature a) As grown, b) Pinned. The green loops
correspond to the applied ﬁeld along a selected axis and the red loops are with the sample
rotated 90 degrees with respect to the ﬁrst within the sample plane. Note that zero degree
in (b) is along the easy axis (pinning direction) and applied ﬁeld.
185

Summary:

The results of the magnetic characterization of the epitaxial Py ﬁlms reveal that:

1) The Py samples grown at an ambient temperature (90°C) on an epitaxial Cu buffer

2)

3)

layer have much larger coercive ﬁelds than those grown at an elevated temperature
(350°C). This is likely due to the presence of signiﬁcant strains in the epitaxial Py
ﬁlms grown at low temperature.

The Epitaxial Py ﬁlm deposited at 90°C on an FeMn buffer layer shows signiﬁcantly
smaller coercive ﬁeld than a similar ﬁlm grown at 90°C on Cu. This difference may
be associated with lower strain in the epitaxial Py ﬁlm, which is in agreement with
the results obtained from crystallographic characterization.

There is no magnetic anisotropy in the epitaxial Py ﬁlms grown on Cu buffer layer
regardless of growth temperature. However, the magnetic anisotropy seen in the
epitaxial Py ﬁlm grown on FeMn buffer layer is due to existence of the exchange bias

anisotropy.

6.3 MAGNETIC PROPERTIES OF EPITAXIAL Py BASED SPIN-VALVES

6.3.1 Introduction

In section 6.2, the magnetic properties of epitaxial Py ﬁlms were studied. In this

section the results of CPP-MR magneto-transport and magnetization measurements will

be presented for the epitaxial Py-based exchange-biased spin-valve samples (EBSV).

One of the advantages of magnetic spin-valves over magnetic multilayers is that in

spin-valves only a small ﬁeld is needed to produce the GMR effect with a sharp

switching response. There are two types of spin-valves, hybrid spin-valves (such as

186

Co/Cu/Py) and exchange-biased spin-valves (such as FeMn/Py/Cu/Py). A brief
discussion of hybrid and exchange-biased spin-valves is given in Appendix 6B. In this
study, magnetic and magneto-transport properties of epitaxial Py-based exchange-biased
spin-valves (EBSV) are discussed.

An EBSV consists of two ferromagnetic layers (such as Py, NiCoFe, or C0), separated
by a non-magnetic layer (such as Cu, Ag), with one of the ferromagnetic layers directly
exchange-coupled (pinned: see section 6.1) to an adjacent antiferromagnetic layer (such
as FeMn, NiO, C00). The pinning direction is referred to as the easy axis. The unpinned

ferromagnetic layer separated by the non-magnetic layer is referred to as the free layer.

Parallel and Antiparallel States: Major and Minor Loops: In a typical CPP-MR
measurement, the pinned spin-valve is mounted on a measuring device inside a magnet
coil in such a way that the easy axis is in the direction of the negative ﬁeld of the coil.
By measuring the resistance of the spin valve as a function of magnetic ﬁeld (changing
the applied ﬁeld from the initial negative ﬁeld continuously, toward zero and then to the
positive direction, and then reversing it back to the initial negative ﬁeld) an exchange-
biased hysteresis 100p is obtained. There are two types of hysteresis loops: minor loops
and major loops. The difference between major and minor hysteresis loops can be best
described by referring to the schematic diagram of an ideal exchange-biased hysteresis
loop, shown in Figure 6-4(a). This ﬁgure is divided into several regions, which are
described below:

Regions 1 and 10: In the presence of a few hundred Oe of negative magnetic ﬁeld

(parallel to the biasing direction) the magnetic moments (spins) of both the pinned and

187

Resistance

 

Free Layer

Pinned Layer

 

 

 

   

 

 

Parallel State Parallel State
-H(Oe) 0 +H(Oe) Magnetic Field
(a)
Spin Down Spin Up Spin Down Spin Up

   
   
 

Free Layer NiFe

Cu

Pinned Layer NiFe

 

Parallel State Antiparallel State
(b) (C)

Figure 6.4 Schematic diagram of an exchange-biased hysteresis loop of an ideal spin-
valve. (a) Illustration of the orientation of the spins in both the pinned and free
ferromagnetic layers for both the parallel and antiparallel state. (b and c) illustrate the
scattering of spin up and spin down electrons in both the free and pinned ferromagnetic
layers.

188

free ferromagnetic layers line up parallel in the direction of the easy axis (biasing
direction). This is a parallel state. Electrons in one of the two electron channels (spin up
or spin down, see Figure 6.4(b)) experience relatively little scattering, thus the resistance
of the sample is a minimum.

Regions 3 and 8: Upon changing the magnetic ﬁeld continuously from the biasing
direction, (ﬁrst toward zero and then positive) the magnetic moments in the free
ferromagnetic layer will switch from parallel to antiparallel with respect to the magnetic
moments in the pinned ferromagnetic layer at a few Oe. Upon completion of the rotation
of the moments in the free layer, the spin-valve will be in the antiparallel state and the
resistance of the spin-valve should remain constant to form a plateau. This is now an
intermediate scattering situation for electrons with either spin up or down (scattering in
both channels, see ﬁgure 6.4(c)) and the resistance of the sample is a maximum. At this
stage one can investigate either a minor loop or a major loop.

Minor loop: By changing the magnetic ﬁeld from the middle of the antiparallel state
plateau back to the original negative ﬁeld, a minor hysteresis loop will result. A minor
loop is a hysteresis loop produced purely by the rotation of the magnetic moments in the
free ferromagnetic layer and is a way to magnetically evaluate that layer. In Figure
6.4(a), the path through points 1—>2—>3—)8—->9—>10 represents a minor hysteresis loop.
Major loop, regions 5 and 6: Starting in region 3, further positive increases in the
magnetic ﬁeld will rotate the pinned ferromagnetic layer until the moments in that layer
align themselves parallel to the moments in the free ferromagnetic layer. This process
creates a second parallel state of opposite orientation to the ﬁrst one. In this second

parallel state, the spin-valve again has minimum resistance (see ﬁgure 6.4(b)). Cycling

189

through the steps l—>2—)3—>4—)5-—>6—)7—>8—>9—>10 (Figure 6-4(a)) yields a GMR spin-
valve hysteresis “major loop” [1-4, 10-13].

Exchange-biased ﬁeld (HE): As described in section 6.1, the exchange-bias ﬁeld (HE) is
due to the effect of an antiferromagnetic layer on an adjacent ferromagnetic layer which
leads to the magnetization or magnetoresistance hysteresis loop shifting away from the
origin. To calculate HE and the coercive ﬁeld (He) of the pinned layer for an exchange-
biased spin-valve, two ﬁeld values need to be measured; H1 and H2, as shown on Figure
6.4(a).

The following sections discuss the experimental procedures and the results of
magneto-transport (CPP-MR), magnetic (magnetization) measurements, and magnetic
anisotropy of a few epitaxial spin-valves. These results will be discussed and compared
to those obtained from polycrystalline spin-valves in this and another [14] study. Note
that, structural characterization studies on some of these spin-valve samples (one each
with thick and thin layers) have been done by H. Geng et a1. [16]. The results of those

studies conﬁrm that the EB SV layers were epitaxed, but with some defects.

6.3.2 Experimental Procedures
Fabrication of Epitaxial Spin-Valves for CPP Measurement:

The epitaxial Py based spin-valves were sputter deposited in an L-shape (1 mm x 8
mm) pattern as was shown in Figure 2-7(b) (see also ﬁgure 6.5). The top surface of each
sample was usually gold coated for oxidation protection. A series of six 75 um square
pillars were then fabricated on each sample using photolithographic techniques (see

Figure 6.5(a)). This microfabrication technique involved optical lithography, ion-milling

190

Sample Geometry and Measurement

In con cts

 

(b)

Figure 6.5 a) Sample geometry of the microfabricated exchange-biased spin-valve
samples. b) A cross-sectional diagram showing the current ﬂow and connections to the
samples. The contact area between the top Nb, with F and W leads (current ﬂow), and
spin-valve is the area of the sample being measured [17].

191

or reactive ion etching (RIE), silicon monoxide evaporation, lift-off, and Nb sputtering.

Table 6.1 lists all of the exchange—biased Py based spin-valve samples sputter
deposited in this study. A few samples were damaged during microfabrication; therefore,
the results of only ﬁve samples are discussed. The CPP-MR of each pillar was measured
separately (see Figure 6.5(b)) at liquid helium temperature (4.2 K). Note that the size of
these pillars (75 um x 75 um) was much larger than the size of the twin variants seen in
the epitaxial Cu and Py ﬁlms identiﬁed using EBSP. Thus, each pillar is expected to
contain a large number of twin variants.

All of the measured spin-valve samples (except one with a thinner pinned layer“) had
the following sequence and thicknesses:

szso-C1120-PY24-CU24-PY24-F6Mns-Auls-Nb250 nm
** sz50-CUzo-PY24-C1124-Py6-F6Mn8-Auls-szso nm

The Nb ﬁlm was deposited at a substrate temperature of 750°C and post deposition
annealed at 950°C. The ﬁrst Cu layer (20 nm thick) was then deposited at 150°C, and
post deposition annealed at 350°C. The remaining layers were deposited at ambient
temperature (90°C). The standard deposition rates for Nb, Cu, Py, and FeMn were 3, 3.4,
3.3, and 2.1 A/s, respectively. For the ﬁrst few samples, the A1203 substrates were
cleaned in the sputtering room in an ultrasonic bath using acetone, then alcohol. Failure
to achieve a good antiparallel state led to preparing the remaining samples in a clean
room. In this new sample preparation technique, the substrate cleaning was changed to
the following. The substrates were ﬁrst cleaned in a combination of boiling deionized
(DI) water and detergent (Alconox) in an ultrasonic bath. After rinsing in boiling

deionized water, the substrates were rinsed and mechanically cleaned in acetone using

192

Table 6.1

Symmetrical EBSV Samples Prepared Using Photolithography

 

 

 

 

 

 

 

 

 

 

 

 

Sample Type of Spin-valves Growth Temperature
27-01 Nb-Cu-[Py-Cu-Py-FeMn]-Au-Nb 750-(150 /350)-[100°C].
Damaged 250-20-[24-20-24- 8 ]-15-250 nm HTH-LTH-[No Heater]
27-02 Nb-Cu-Py-[Cu-Py-FeMn]-Au-Nb 750-(1 5 0/ 3 50)-3 50-[100°C]
Damaged 250-20-24-[20-24- 8 ]-15-250 nm HTH-LTH-[No Heater]
33-01 Nb-Cu- [Py-Cu-Py-FeMn]-Au-Nb 75 0-(1 50/ 3 50)-[80°C].
250-20-[24-25-24- 8 ]-15-250 nm HTH-LTH-[No Heater]
33-02 Nb—Cu-Py-[Cu-Py-FeMn]-Au-Nb 750—(150/350)-350-[90°C]
Damaged 250-20-24-[25-24- 8 ]-15-250 nm HTH-LTH-[No Heater]
33-03 Nb-Cu-[Py-Cu-Py-FeMn]-Au~Nb 750-(1 50/3 50)-[90°C]
1‘0;- i"::d¢d 250-20-[12-20-12- 8 ]-15-250 nm HTH-LTH-[No Heater]
m IS 5 y
33-04 Nb-Cu—[FeMn-Py-Cu-Py]-Au-Nb 75 0-( 1 50/350)-3 50-[80°C]
Damaged 250-20-[ 8 -12-20 -12]-15-250 nm HTH-LTH-[No Heater]
37-01 * Nb-Cu-[Py-Cu-Py-FeMn]-Au-Nb 750-(150/350)-[90°C].
250-20-[24-25-24- 8 ]-15-250 mm HTH-LTH-[No Heater]
37-02 * Nb-Cu-[Py-Cu-Py-FeMn]-Au-Nb 750-(150/350)-[90°C]
250-20-[24-25-24- 8 ]-15-250 nm HTH-LTH-[No Heater]
37-03 * Nb-Cu-[Py-Cu—Py-FeMn]-Au-Nb 750-(150/350)-[90°C].
250-20-[24-25-24- 8 ]-15-250 nm HTH-LTH-[No Heater]
40-04 * Nb-Cu-[Py—Cu-Py-FeMn]-Au-Nb 750-(150/350)-[90°C].
250-20-[24-25-6- 8 ]-15-250 nm HTH-LTH-[No Heater]

 

 

 

 

Samples 27-01, 27-02, 33-02, and 33-04 were either damaged during lithographic
fabrication or preparation.
*Deposition rates of Cu in sample 37-1, 37-2, and (37-3 & 40-4) are 9.8, 1.2, and 3.4A/s,

respectively.

193

cotton swabs, then rinsed in an isopropyl alcohol. The substrates were then washed with
DI water and dried with ﬂowing industrial grade nitrogen gas. The cleaned samples were

transported to the sputtering room in closed containers.

Magnetization Measurements (Magnetic Anisotropy and Strain):

Similar to the magnetization measurement of epitaxial Py ﬁlms, a small portion of an
epitaxial sample was cut and placed in the magnetometer with the plane of the ﬁlm
parallel to the applied ﬁeld. Note that the easy axis of the spin-valve must be parallel to
the direction of applied ﬁeld. To look for any magnetic anisotropy in the spin-valves, the
sample was removed from the SQUID magnetometer and pinned at different angles (90-

degree intervals), and the magnetization (minor loop) was measured again.

6.3.3 CPP-MR and Magnetization Measurements on Py-Based EBSV.

In this section the results of CPP-MR magneto-transport and magnetization
measurements will be presented for the epitaxial Py-based exchange-biased spin-valve
samples (EBSV), mostly from szso-Cuzo-PY24-Cuz4-Py24-FeMns-Au15-Nb250nm. Here the
FeMn is used to pin the magnetization of the adjacent Py layer. As described previously
(see 6.3.1) such an EBSV should allow reliable production of both parallel and
antiparallel magnetic states. The CPP-MR results will be compared with previous results

2

for polycrystalline samples with 1 mm area and examined as a function of deposition

rate.

194

 

Samples with substrates cleaned in the sputtering room.

The substrates for the ﬁrst set of EBSVs of the speciﬁed form were prepared in the
sputtering room without an extensive cleaning (see section 6.3.2). The CPP-MR (AAR)
data from one of these samples is shown in Figure 6.6, where AAR(H) is deﬁned as:

AAR(H) = AR(H) — AR(P).
Where A is the area of the pillar, R(H) is the resistance of sample at a particular magnetic
ﬁeld, and R(P) is the resistance of sample at the parallel state.

These data have two disturbing characteristics. (1) As the ﬁeld is swept to positive
values, the data initially rise sharply as expected, but the data fall quickly back to the
value of the parallel state, instead of forming a plateau when an antiparallel state is
established. (2) AAR (1.3 me2) is only about half as large as the range of values (~2.7
me2) previously found for polycrystalline EBSVs of the same composition with area ~
1mm2 [14]. Taken together, these behaviors strongly suggest that the sample has not
reached a true antiparallel state.

Two phenomena might be responsible for failure to reach an antiparallel state.

a) Undesired ferromagnetic coupling between the pinned and free Py layers.
Such coupling could occur, for example, due to the presence of pinholes in
the Cu layer induced by either growth defects or the presence of residual dirt
particles on the substrate surface before deposition.

b) Lack of good exchange-bias coupling between the pinned ferromagnetic

layer and the FeMn.

195

CPP-MR of Py based EBSV

 

   

 

 

 

  

 

 

1.29‘I' l"'lf'(_'_l"'i‘

tub-eu-Py-Cu-Py-FeMn-Au-Nb —’ ;

- 250-20-24-25-24-8-15-250 nm .

1.01 - —
.4“ I
5 0.73 —
V I
at .
g 0.45 .-
0.17 1
i.
0.00

400 -200 O 200 400
Magnetic Field (Ce)

Figure 6.6 CPP Magnetoresistance hysteresis loop of a symmetrical Py based EBSV.
The standard deposition rates for Nb, Cu, Py, and FeMn were 3, 3.4, 3.3, and 2.1 A/s,
respectively. In this sample, an anti-parallel state is not observed, suggesting that there is
magnetic coupling between the free Py and pinned Py layers resulting in both of layers
switching at the same time. The coupling between Py layers is probably caused by
growth of pinholes due to a dirty substrate.

196

To attempt to distinguish between these two possibilities, two steps were taken. First a
new procedure for substrate cleaning was devised, involving a more rigorous cleaning
process and carrying out this process inside a clean room. Details were given in section
6.3.2. Second, to see if the lack of antiparallel state might be related to growth rate, a set

of samples was prepared with different deposition rates.

Sample prepared with substrate cleaned in a clean room.

Using these cleaned substrates, a new set of epitaxial symmetrical EBSVs were
grown. One of the samples (with standard Cu sputtering rate 3.4 A/s) had the same
sequence, deposition rates, and thicknesses as the previous sets. The result of CPP-MR
measurement of this sample is illustrated in Figure 6.7. The ﬁrst feature of the new data
to note is that the falloffs from the nominal antiparallel states are much slower than in
Figure 6.6. The second feature is that values of AAR for this sample: 2.6 fﬂmz,
comparable to the values ~2.3 me2 previously found for polycrystalline samples with A
~ 1 mm2 [15]. Together, these slower falloffs and agreement in AARs suggest that the
problem of coupling between the free and pinned Py layers has been eliminated, and that
its main source was inadequate cleaning of the substrates.

To investigate the effect of the deposition rates on the antiparallel state, three EBSV
sample were deposited (including the previous sample). Figure 6.8 shows AAR for these
three EBSV samples for which the substrates were also cleaned in a clean room. All
three were prepared in a single run, but with different sputter deposition rates for Cu: 1.2
A/s, 3.4 A/s, and 9.8 A/s. The deposition rates for the Nb, FeMn, and Py were kept the
same as before. The values of AAR for all three samples: 2.8 me2 for Cu sputtering rate

197

 

 

3.00

2.50

2.00

1.50

AAR (lnm’)

 

 

 

 

Magnetic Field (0e)

Figure 6.7 CPP-MR hysteresis loops of an epitaxial symmetrical EBSV:
Nb;so-Cuzo-Py24-Cuz4-Py24-FeMn3-AuI s-szso nm epitaxial spin-valves. Both major and
minor (red) loops are shown. AAR is about 2.6 me2.

198

 

ITIITTj’I

3.00
DR{Cu) =1.2

AAR «an)

}—
r-
}-
i-
:-
i-n
D
h
b
h
i.—
P
h
}-
r-
L—
P
}-
F
P
j—
j-
r-
i-
l-
.—
h
-
h
-
i-
I-
b

 

111114411_

YIT!E!IIUYIYIIUIIIU

 

111 111111111111111111111111111111

1111111111111114444

 

 

 

III'IYWII

2.50 DR (Cu) = 3.4
2.00
1.50

1.00

AAR (m2)

0.50

IIII'IU‘UIYUUUIIUVI'I'I'1j7

0.00 1

['1'

111111111

ITIIIIIII'IFI'IIIII
<———
—>

W:

 

 

 

111111111111111111111111111

i

'1

41111

 

AlllJ—LIilLLillllJll

 

 

30W UrTrTIIIT

2.50

2.00

1.50

DR (Cu) = 9.8 NS

IITT'I‘TfTIIIII’IIUU

——->
<———

 

1.00

AAR (m2)

0.50

’IjIIII'IIIIFIIIIITIIIY'I‘II'I

 

IIII

411111111

fQ\ H '

0.00 vvvv'"

H

4111111111111111111111111111111111

 

 

1111111111111111111

 

-100 -50

0 50 100 150 200

Magnetic Field {00)

Figure 6.8 CPP-MR hysteresis loops of three epitaxial Py spin-valves: szso-Cuzo-Pyu-
Cu24-Py24-FeMn3-Au15-Nb250 m... AAR are 2.8, 2.6, and 2.3 mez for sample deposited
with Cu deposition rates of 1.2, 3.4, and 9.8 A/s, respectively. All layers were deposited
at 90°C except for the ﬁrst Nb and Cu layers.

199

1.2 A/s; 2.6 me2 for rate 3.4 A/s; and 2.3 meZ for rate 9.8 Ads, are similar to each other

and also comparable to the values ~2.3 ﬁlm2 for polycrystalline samples with A ~ 1
mmz. Now looking at the difference between the CPP-MR behavior of these three
samples in Figure 6.8, the ‘standard’ sputtering rate of 3.4 A/s gives the best deﬁned and
ﬂattest antiparallel ‘plateau’. The free layer rotates in smallest ﬁeld, about 15 0e for rate
= 3.4 A/s, versus about 25 Oe for rate = 1.2 A/s and about 70 Oe for rate = 9.8 A/s. Also,
the antiparallel state persists to higher ﬁeld, about 130 Oe for rate = 3.4 A/s, versus about
100 Oe for rate = 1.2 A/s and only about 70 0e for rate = 9.8 A/s.

Deﬁning H1 and H2 at the ‘half-heights’ of AAR as shown in Figure 6.8, and following
the standard procedure shown in section 6.1 to deﬁne the exchange bias ﬁeld, HE, and the

coercive ﬁeld, He, gives values that are listed in Table 6.2.

 

 

 

 

 

Table 6.2
Deposition rate (A/s) 1.2 3.4 9.8
AAR ( fQ m2) 2.8 2.6 2.3
AP state persist ﬁelds (Oe) 100 130 70
Exchange-bias ﬁeld (HE, 0e) 51 73 60
Pinned layer coercive ﬁeld (Hc, 0e) 61 67 80

 

 

 

 

 

 

Sample ion-milled with better ‘cooling procedure’, including search for magnetic
anisotropy in the free Py layers.

The possibility that heating of the sample during the ion-milling process could affect
the coercive ﬁeld for the free layer, led to preparing and ion-milling of one set of samples

using a procedure that was expected to keep the sample cooler during the ion-milling.

200

This procedure involved using a thin layer of high thermal-conductive silver grease
between the back of the substrate and the Cu heat sink. The silver grease is UHV
compatible and originally was provide for mounting the regular magnetron sputtering
(Mini-Mac sputtering gun) target on the target holder for better electrical and thermal
conductions.

Since the exchange-bias ﬁeld decreases with increasing thickness of the ‘pinned’ layer
[2], to achieve enhanced pinning the thickness of the pinned layer was reduced to 6 mn,
giving an EBSV of the form (Nb2so—Cu2o-Py24-Cu24-Py6-FeMng-Au15-Nb250 nm).

Figure 6.9 shows the CPP-MR data from one such sample. This curve shows AAR =
1.4 me2 comparable to the value ~ 1.3 film2 previously found from a 1mm2
polycrystalline EBSV with the same layer thickness [15]. To look for the possibility of
magnetic anisotropy in the EBSV, this samples was pinned in three different directions
(0° ([1T0]), 45°, and 90° ([112])) and AAR was measured. The results of these
measurements are shown in Figure 6.10. The coercive ﬁeld of the minor loop is about 40
Oe when the sample is pinned along the [1T0] and the 45° directions. While the
coercive ﬁeld is about 46 Oe when the sample is pinned along the [112], slightly larger
than coercive ﬁeld along [1T0] direction. This may be due to sample geometry. To
determine if this was in fact a shape effect, the magnetization of a small 1 mm x 1mm
section of this sample was measured in the [1T0] and [112] directions (pinned before
each measurement in the [1T0] and [112] directions, respectively). The magnetization
hysteresis loops, as shown in Figure 6.11, are essentially identical (note overlap) showing

that there is no evidence of magnetic anisotropy in this sample.

201

1.50

0.00

 

-500 -250 O 250 500 750 1000 1250 1500
Magnetic Field (0e)

Figure 6.9 CPP-MR hysteresis loop of an epitaxial Py based EBSV:
sz50-CU2o-Py24-CU24-Py6-FeMng-Alh5-Nb250nm. AAR is about 1.4 me2.

202

 

 

- dqladqqdd-

 

0

I 'TTTI
AAAuA
I
l
l
I
l

V'VV'v

LA
I
l
1
I
l

 

IIU'UVY

.
IIIJIII

qdi4d‘ddddﬁ1‘d-

(b)

 

 

 

1

I
l
I
----r---r-P-r---+---+-‘
I
I
l

VII'TTUI'I

Tj'Tj] I

 

h
l-
p
b
.-
.-
b
P
D

UUYVIYTY'rY'YIIUIUUI‘V'YIUYUI‘Y'Y'IIll

 

 

 

L50----L---L---L_-_L-_-4---

qu— 4‘14114

. . _

. _ _

_ _ _
1llJlll4lll1lili
. . .

_ _ _

_ . .
TliJlll4lll1lill
. . .

0 A. 0

. . _
TllJIll+lll1ll
_ . _

_ _ _

_ _ _
TllLlllelllrll
_ _ _

_ . .

_ . .
TllLlllelllrli
. _ .

_ r .

_ W _
FliLiilplllril
. . _

. . .

_ . _
rllLlllhlilr

. _ _

_ . .

_ _ _
—-bhb_bbbbPP~bbh--
0 0

0. 5

1 0

E.
a:
3

  

 

100

75

50

 

 

0 25

-25
Magnetic Field (00)

 

Ill 1 III III llll
IIIIlTTIT'UﬁWIT—meYTTImVTTYIIII!II

 

 

 

.
111111llllllllJlLlLllllLlllll

-75 -50

 

 

1.50 ;-

page u

<<

_ _
_ . .
. _ .
_ _ .
_ JIIIﬁIIIWIII.
. . _ _ .
. . . . .
. _ _ . .
WllJillJillalllﬁlIJ
_ . _ _ .
_ al ﬂ . i.
. . _ . _
TllJillJilla ll.
. _ _ . .
_ . . . .
. . _ . _
TllllllxilltlllTli.
_ . _ _
. . . _
_ . . .
TlililltlllrlllTli
_ . .
1 _ . V.
. . _ 1.
TliLlllLlllrlilrll
_ . _ .
_ . . .
. . . .
FllLlllLlllrlllr Ill
. . . .
. . . .
_ . _ .
hrhnn—h-Ph—P bub-hhbpbbel-b
0. .0 .0
0. s 0.
0V

0
W 5
1

page MS.

0.00 j
-100

Zero and 90°,

Figure 6.10 Minor hysteresis loops of an epitaxial EBSV (Py(24)-C11(24)-Py(6)-FeMn(gnm)).
corresponds to the [110] and [112] directions in the Py(111) plane. The similarity of

coercive ﬁelds suggests that there is no magnetic anisotropy (except some shape effect)

The epitaxial EBSV sample was pinned at a) 0°, b) 45°, and 90°.
in this epitaxial EBSV.

203

 

.5
15 10 I I I I I I I I ITI I I I I I I I I I I I I I I I I I I I I I
' I T I I I I I a

q

110‘

510‘

510‘

410‘

Magnetization (emu)

O
IIIIlI—III'IIIIIIIIIIIIIIIIII

  

A
0|
.5
O
‘1

11111111111111111111111111L1111

IIII

 

 

.210" 11111LL11_1111111IIIIHILILLIIILLLJIIILI

400 -75 -50 -25 0 25 50 75 100
Magnetic Field {Co}

(a)

 

 

   

 

 

.5

1.510 plrrl[IIIUIIIYITIIIIIIrr11rﬁ1Tliro1ﬁ11q
‘ b

110 — —
- -4
P d l d
e... g : :
g 510 _ -———>» a
P -l
H " cl
c L J
§ 0 ' ‘
N .g ' "
:3 -510 - j
c - Superimposed:
a .5 _ ..
:g -110 : i" -n-———~
: 1‘ ‘ E f / 3
" . .__./ I t‘ I I ‘
.1.5 10‘ 5:1:— «rr-HH‘W” l i _
Z i . ‘
1- 5M“ :
.2104 -1_14_1 1 1+1 L1 1 1 1 111 1 1 1 L1 1 1 111 1 11L1r‘r‘r‘1“1‘1*r*r“r1‘1

 

-100 -75 -50 -25 0 25 50 75 100
Magnetic Field (0e)

(b)

Figure 6.11 Magnetization hysteresis loops of a small section (blue) of the epitaxial
EBSV (Cu2o-Py24-Cu24—Py6-FeMng-CU20nm) pinned at a) 0° and b) 90° angles. The
coercive ﬁeld (30 0e) and saturation ﬁelds are equal in both directions.

204

Summary: A set of Py based exchanged-biased spin-valve samples was prepared using
improved substrate cleaning and improved sample cooling during the ion-milling. The
data from these samples featured a sharp switch from the parallel to the antiparallel states
with a well deﬁned and ﬂat antiparallel state plateau. The values of AAR were also

comparable to similar polycrystalline EBSVs.

6.3.3.3 Comparison of Polycrystalline and Epitaxial Samples:

The behavior of the epitaxial EBSV sample with a layering sequence of (Nb250-Cu2o-
Py24-Cu24-Py24-FeMng-Au15-Nb250nm) (also see ﬁgure 6.7) was compared to two
polycrystalline samples with identical layering sequences. One of the polycrystalline
samples was sputter deposited at a substrate temperature of 90°C in the millimeter-sized
CPP geometry (see Figure 1.5). The second polycrystalline sample was sputter deposited
at a substrate temperature of -3 0°C in an L shape pattern and then photolithographicly
fabricated, with a pillar area of 75 um x 75 um [14]. The CPP-MR results of both
samples, along with the epitaxial sample, are shown in Figure 6.12. The values of AAR,

HE, and Hc are tabulated in Table 6.3.

 

 

 

 

Table 6.3
AAR HE HC
f Q m2 06 06
Polycrystalline,
75 um x 75 mm, 2.4 175 55
Tgrowth = -30°C
Polycrystalline,
1mm2 2.5 40 40
Tgrowth = 90°C
Epitaxial
75 um x 75 um 2.6 73 67
Tgmmh = 90°C

 

 

 

 

 

 

205

While the values of AAR for all three samples are similar, the coercive ﬁeld, He, in
the pinned Py layer of the polycrystalline sample deposited at -30°C is about three times
smaller than that in the exchange—bias ﬁeld in that sample. This polycrystalline sample
displays a broad antiparallel state plateau for positive magnetic ﬁelds and returns just
before the ﬁeld come to zero (Figure 6.12(a)). In contrast, the values of He in both
polycrystalline and epitaxial samples deposited at 90°C are very close to H5. These two
samples have well deﬁned antiparallel state plateaus, but did not display return
antiparallel state plateaus. This indicates that a growth temperature higher than 30°C
may affect the quality of the FeMn and Py layers and thereby weaken the exchange-bias
coupling between these two layers. As a result, the antiparallel state will not form during

the return part of major hysteresis loops.

206

 

   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

     
  

  

 

 

 

 

 

 

 

 

 

-..---_I .......... l.---’..:..
2.50 E- (a) .
2.00 *— -------------------- J
' i CCP-MR of a 75 x 75
'4; 1.50 ............. --..-l_-----_' “m2 pillarofa
Q l polycrystalline EBSV.
E 1.00 ~--—- ------ ,-------E ----------- 1 TGrowth='30°C
2 : g ‘ AAR = 2.4 ﬁlm2
0.50 ----------- T 1 ‘
0.00 t. . E" - I . :1
: Easy Axis I i l j
l l l 44 l l l L l J 1 1 L1 1+.
-100 0 1 00 200 300 400
Magnetic Field (0e)
..............,............W
2.50 E ' ' ' : ‘b’ i
2.00 j
’7‘ E CCP-MR of a
a 1.50 _ millimeter size
I; 1 o 0 E polycrystalline EBSV.
3 ' g r GM, = 90°C
0.50 ; AAR = 2.5 mm2
0.00 C . .
r I I I I I
”???1'1 11111111 1.1-1"?1'1'311'1"1I1"1-1"1-?T'1
-100 -50 0 50 100 150 200
Magnetic Field (0e)
3,00 -. n . q .- -'-'-'-I-'-'-‘ n +1 e .. J-l-‘-'-'-LFWL: ‘
(c) 3
2.50 ' ' ' ' ‘ "j
"A 2.00 """"" V"? """ : "m", CCP-MRofa75x75
a 1.50 l----:L-—---E--- - -----j umzpillarofan
at : g ‘ epitaxial EBSV.
3 1.00 --"r-----'---- TGrowth=9O°C
_ AAR = 2.6 mez
0.50 """"""""""""""
0.00 l
. . . . . l
100 150

Magnetic Field (Go)

Figure 6.12 The CPP-MR hysteresis loops of two polycrystalline Py spin-valves: Nb250-
CU20-F6Mng-Py24-CU24-Py24- Auls-szso nm and one epitaxial sz5o-CU2o-PYZ4'CUZ4-Py24-
FeMns-Auls-szso nm.

207

6.4 SUMMARY

The important issues in studying the magneto-transport properties of epitaxial single
magnetic ﬁlms and spin-valves are growth temperature, deposition rates, type of buffer
layer, and surface quality of the substrates. It has been shown that the coercive ﬁeld in
epitaxial single Py ﬁlms grown at 90°C is much larger than that observed in ﬁlms grown
at 350°C, most likely as a result of a large number of dislocations and defects which
create strain in the deposited ﬁlms. Depositing on a different buffer layer might reduce
the number of these dislocations and therefore dissipate some of the strain. The surface
quality of the substrates is also an important factor, especially for magneto-transport
performance of epitaxial EBSV. The presence of residual dirt particles on the substrate
surface was suspected to create pinholes in the Cu layer which resulted in an undesired
coupling between the two ferromagnetic (Py) layers. Changes in the substrate cleaning
procedures eliminated the problem of coupling between Py layers. In the CPP-MR
hysteresis loops from epitaxial EBSV samples prepared with substrates cleaned in a clean
room, showed sharp switching from the parallel to antiparallel states, with a long range
and ﬂat antiparallel state plateau. In studying the effect of deposition rate, an
intermediate Cu deposition rate of 3.4 A/s gave the best deﬁned and ﬂattest antiparallel
plateau. AAR values of epitaxial EBVS samples were comparable to those for similar
polycrystalline EBSVs. The deposition temperature was also an important factor in the
fabrication of epitaxial spin-valves in particular its effect on the exchange-bias coupling
between the pinned Py and the adjacent FeMn layers. Magnetization data of these EBSV

showed no evidence of magnetic anisotropy in the free layer of the EBSVs.

208

Appendix 6.A

Transport Properties of Epitaxial Nb Films

Nb is one of the many metals that become superconducting at low temperature. Bulk,
high purity Nb has a resistivity of about 13 uQ-cm at room temperature, with a
superconducting critical transition temperature (Tc) of 9.2 K. However, depending on the
growth condition, Nb ﬁlms can have different resistivity and Tc values than bulk. A
polycrystalline sputter deposited Nb ﬁlm (~300 nm thick) has resistivity of about 15-18
uQ-cm at room temperature and about 6 uQ-cm at 10 K, with Tc = 8.9 K, using both van
der Paaw and four probe measurements. Single crystal epitaxial Nb ﬁlms are expected to
have better transport properties than polycrystalline ﬁlms (less dislocations and point
defects). The epitaxial Nb ﬁlms grown by sputter deposition had a residual resistivity of
about 15-17 uQ—cm at room temperature and approximately 06-10] uQ-cm at 10 K,
using van der Paaw measurements and approximately 0.75-1.56 uQ-cm using four probe
measurements. The superconducting transition temperature of these epitaxial Nb ﬁlms
was approximately Tc E. 9.1 K, as shown in Figure 6.A.1(a).

Magnetic Property of Nb:

From a magnetic point of view, Nb is a strong diarnagnet below its transition
temperature and a very weak pararnagnet above its transition temperature. Therefore,
measurement of the magnetization as a function of temperatures can also determine the
transition temperature of the Nb ﬁlm. Figure 6.A.1(b), shows magnetization of the same

epitaxial Nb ﬁlm as in Figure 6.A.1(a).

209

The critical temperature of 250 nm thick epitaxial Nb ﬂlm

 

 

0.1E- .ewee. “1le .eq
r» T I
E c f”- -1~o—o——o—o—o— E
0.08 f a:
E i
i
‘ E
E 0.00 E E E
'5 . E .'
g E l l
0.04 , '
a , E l
.5 » E
a E l
n: 0.02 E l 7'
. , E
i E E
0 ~ 0 H Hw-HO—HaH—O—O—O—O—H—i
i i
-0.02 E: e A %Me—L _. # ~e~—e—’—i—i—44~ewegej
8 8 5 9 9 5
Temperature (K)
The diamagnetic magnetization of epitaxial Nb film
of . l p . rm 1 f . .# fv r . . 1 r . . - 1
. H=1000e E'””"""«.
E / E
-1 10" :~
E r E
A " E E
3 -2 10“ ,» ~
8 E y =
"’ l
s . E
8 .3 10“ ' I
g E /
E ,/’/ E
4 ‘10 5 T——— —*” “W'ITMTT. E
E E
.5 10.5 _. __l __. _. __.—l—I__L__.ﬂ_i ._.____l _-__.l _l ._. __L______ _l_~,L _
5 6 7 8 9 10

Temperature (K)

Figure 6.A.1. The resistance and magnetization of a 250 nm thick epitaxial Nb ﬁlm as a
function of temperatures. The Tc of this ﬁlm is about 9.1 K.

210

APPENDIX 63
TYPES OF SPIN-VALVES

There are several types of spin-valve structures. Among them, hybrid spin-valves
(such as Co/Cu/Py) and exchange-biased spin-valves (such as FeMn/Py/Cu/Py) are the
most important ones.
Hybrid spin-valves: The idea behind the hybrid spin-valve is to be able to
independently rotate the direction of the magnetization of both ferromagnetic layers.
This independent control of the ferromagnetic layer could be achieved by using either the
same type of ferromagnetic material, with one of the ferromagnetic layers is much thicker
than the other, or by using a different ferromagnetic material for the second magnetic
layer.
Examples: Pylz - Cuzo — Py24 nm and Py 12 — Cuzo — C06 "m.
Exchange-biased spin-valves (EBSV): In the EBSV structure, both ferromagnetic
layers could either be the same or different materials, or have different thicknesses.
However, one of the two ferromagnetic layers must be pinned to an anti-ferromagnetic
layer. The pinning is due an interfacial interaction between antiferromagnetic and
ferromagnetic layers that leads to anisotropy, a unidirectional anisotropy or ‘exchange-
bias’ [1]. The pinning direction is along the applied ﬁeld and is referred to as the easy
axis. The second ferromagnetic layer is known as the free layer. If the thicknesses of
both magnetic layers are identical, the EBSV structure is referred to as symmetrical
EBSV.
Example: F eMng —— Py24 - pinned layer) — Cu24 — Py24 nm— free layer) and

FCMDg— Co6— Cuzo— Py24 (nm).

211

6.4 REFERENCES.

1) R. Jungblut, R. Coehoom, M.T. Johnson, Ch. Sauer, P.J. van der Zaag, A. R. Ball,
Th. Rijks, J. aan de Stegge, and A. Reinders, J. Magn. Magn. Mater. 148, 300 (1995).

2) E. Berkowitz and K. Takano, J. Magn. Magn. Mater. 200, 552 (1999).

3) J. Nogués and Ivan K. Schuller, J. Magn. Magn. Mater. 192, 203 (1999).M. F.

4) Gillies, J. N. Chapman, and J. C. Kools, J. Appl. Phys. 78, 5554 (1995).

5) W. H. Meiklejohn and C. P. Bean, Phys. Rev. 105 904 (1957).

6) B. D. Cullity, Introduction to magnetic Materials (Addison-Wesley Publishing
Company, Reading, MA, 1972).

7) D. Mauri, H. C. Siegmann, P. S. Bagus, and E. Kay, J. Appl. Phys. 62, 3047 (1987).

8) P. Malozemoff, Phys. Rev. B 35, 3679 (1987).

9) P. Malozemoff, Phys. Rev. B 37, 7673 (1988).

10) D. Sander, A. Enders, and J. Kirschner, J. Magn. Magn. Mater. 200, 439 (1999).

11) C. Tsang, N. Heiman, and K. lee, J. Appl. Phys. 52, 2471 (1981).

12) W. P. Pratt Jr., S. D. Steenwyk, S.Y. Hsu, W. C. Chiang, A. C. Schaefer, R. Loloee,
and J. Bass, IEEE Trans. Magn. 33, 3505 (1997).

13) S. D. Steenwyk, S. Y. Hsu, R. Loloee, J. Bass, and W. P. Pratt Jr., J. Appl. Phys. 81
4011(1997)

14) RD. Slater, J .A. Caballero, R. Loloee, and WP. Pratt, Jr. to be published in J. Appl.
Phys (2001).

15) L. Vila, W. Park, J.A. Caballero, D. Bozec, R. Loloee, W.P. Pratt, Jr., and J. Bass, J.

Appl. Phys. 87 8610 (2000).

212

16) H. Geng, R. Loloee, J .W. Heckman, J. Bass, W.P. Pratt, Jr, and MA. Crimp, accepted

for publication in Magnetic Material, Structures, and MRS proceedings (2000).

213

CHAPTER 7

SUMMARY AND CONCLUSIONS
The objective of this dissertation was to design and develop a sputter deposition
system for growing epitaxial ﬁlms, multilayers, and spin-valves. Furthermore, the results
of crystallographic characterization of sputter deposited Nb, Cu, Co, and Py have been
correlated to the magnetic and magneto-transport properties of the sputter deposited Py

ﬁlms and Py based spin-valves.

This study resulted in the successful growth of epitaxial Nb(110 ), Cu(111), Co(11 1),
Py(111) ﬁlms, and Py based spin-valves. The results of AF M and EBSP characterization
revealed that Nb grew as high quality epitaxial bcc single-crystal ﬁlms on A1203
(1120)th Later in the study, epitaxial Cu ﬁlms were grown on these epitaxial Nb layers.
The crystallographic characterization of these Cu ﬁlms revealed two in-plane twin
variants corresponding to two different stacking sequences of {111} fcc planes.
Orientation distribution maps revealed that the area of the twin variants varies widely and
did not display any particular patterns. Epitaxial Co (111)fcc and Py (111)fcc layers were
grown on epitaxial Cu ﬁlms at high temperature (2 350°C). Both Co and Py ﬁlms also
displayed twin variants with orientation maps similar to those observed for epitaxial Cu
ﬁlms. The crystallographic characterization results of Co and Py samples grown at 90°C
revealed that two different orientation variants were also present, but in smaller sizes.
However, due to the degradation of the associated EBSP patterns and the resolution of

EBSP, mapping of the orientation distributions was not possible.

214

Growth temperature, deposition rates, type of buffer layer, and surface quality of the
substrates are important factors that can affect the magnetic properties of single magnetic
ﬁlms and Py based spin-valves. Regardless of the growth temperature and type of buffer
layer, no magnetic anisotropy, including magnetocrystalline and magnetoelastic
anisotropy, was observed in single epitaxial Py ﬁlms. In exchange-biased spin-valves, a
moderate deposition rate and an ultra-clean substrate resulted in samples with sharp
switching of the magnetic moment from the parallel to the antiparallel state and a very
ﬂat antiparallel plateau. Studies of the effect of pinning along different in-plane axes on
magneto-transport properties of one of the EBSVs revealed the existence of anisotropy in
the free ferromagnetic layer, which may be related to differently oriented defects in the
Py ﬁlm which were created during sample preparation (ion milling). Further study of a

similar EBSV sample displayed no magnetic anisotropy in the free Py layer.

CONCLUSION:

The growth and subsequent crystallographic nature of metal ﬁlms and superlattices
play important roles in determining the fundamental physical properties of the ﬁlms.
Epitaxed growth of Nb, Cu, Co, and Py ﬁlms from the basis for growing magnetic
multilayers (such as [Cu-Co]N) and spin-valves (FeMn-Py-Cu-Py [Pyz Ni34Fe16]) which
have applications as GMR (Giant Magneto-Resistance) position sensors, magnetic head
readers, and magnetic random access memories. GMR in these devices depends on
structural details such as the thin-ﬁlm growth direction, interfacial roughness and
interrnixing, and grain boundaries. However, the exact effects of these variables on given

systems are not yet well understood. Consequently, it has been necessary to further

215

characterize the growth and ﬁlm structure of these magnetic systems to explore the effect
of crystal structure on the magneto-transport properties for advanced applications. This
project was designed to develop a sputtering deposition system to grow and develop
epitaxial ﬁlms, magnetic multilayers and spin-valves in a preferred orientation.

In conclusion, a sputtering deposition system has been successfully designed and
developed to grow epitaxial ﬁlms. The following projects were carried out in this

dissertation:

(1) Surface topography of deposited metallic films (Nb, Cu, Co, and Py) using an atomic
force microscope (AFM) in tapping mode.
The AFM results revealed that except for Nb ﬁlms, which grow with very smooth
surfaces (depending on the growth conditions), the rest of the materials (Cu, Co, and Py)
always have rough surfaces with islands of different sizes. In general, samples which

have been grown under similar conditions, displayed very similar AF M images.

(2) Measurement of crystallographic orientations and grain distributions in deposited
ﬁlms using electron backscatter patterns (EBSP).

The EBSP results revealed that, at high temperature, all fcc materials (Cu, Co, and

Py) display two in—plane variants corresponding to two stacking sequences of {111}

fcc planes. Orientation maps revealed that the twin variants displayed a random

distribution, with no evidence of any particular morphological patterns. The average

variant sizes in ﬁlms grown at elevated temperatures, as determined by the intercept

method, were found to be 4 to 6 um. There was no correlation between the size and

216

 

distribution of the twin variants in the EBSP orientation map and the size and

distribution of the islands in the AF M images. The relative areas of the twin variants

are similar in all materials (Cu, Co, Py) and ranged between 45% and 55%. Epitaxial
ﬁlms of all three materials (Cu, Co, and Py) deposited at room temperature revealed
that they grow in small, coherent islands that do not coalesce. Epitaxial ﬁlms grown
at room temperature therefore contain large numbers of grain boundaries, misﬁt
dislocations, and other types of defects.

(3) Evaluation of magneto-transport properties of sputter deposited epitaxial magnetic
ﬁlms and spin-valves by using CPP-MR and magnetization measurements.

The important issues in studying the magneto-transport properties of single magnetic
ﬁlms and spin-valves were growth temperature, deposition rates, type of buffer layer and
surface quality of the substrates. It was shown that the coercive ﬁeld in epitaxial single
Py ﬁlms deposited on Cu at 90°C is much larger than those grown on Cu at 350°C,
presumably due to the presence of a large number of dislocations and defects which
create strain in the deposited ﬁlms. Growth of Py ﬁlms on FeMn buffer layers resulted in
lower coercive ﬁelds, possibly as a result of a reduction of the density of dislocations.
Studies of the effect of deposition rate show that moderate Cu deposition rates resulted in
sharper switching from the parallel state to antiparallel state, with a wider range of the
antiparallel state up to higher ﬁelds. The EBSV sample deposited at a high Cu deposition
rate produced a signiﬁcantly different MR loop and displayed slower switching from the
parallel to the antiparallel state and larger coercive ﬁeld. The high coercivity was
suggestive of a large number of dislocations and defects in the spin-valve, producing

strain in this sample. The deposition substrate temperature was also an important factor

217

in the fabrication of epitaxial spin-valves. It was observed that the Py (ferromagnetic)
layer adjacent to the FeMn (antiferromagnetic) in spin-valves deposited at a substrate
temperature of 90°C did not have a strong exchange coupling (pinning) to the FeMn layer
and the exchange-biased ﬁeld weakened as spin-valve exposed to high magnetic ﬁeld,
opposite to the easy axis. Magnetic anisotropy was observed in one symmetrical EBSV
which appeared to be related to existence of some oriented defects and strain induced
during sample preparation. There were two noticeable differences between the CPP-MR
hysteresis loops in the epitaxial and polycrystalline exchange-biased spin-valves with an
identical layer structure. The epitaxial samples exhibited fast switching ﬁ'om the parallel
to the antiparallel state for a small applied ﬁeld while the antiparallel state plateau was
much ﬂatter (higher resistance state remain up to larger magnetic ﬁeld). The CPP-
resistance in the epitaxial sample (in parallel state) was almost twice that of a similar

polycrystalline spin-valve; however the values of GMR were similar.

218

FUTURE STUDIES:

Crystallographic Characterization: EBSP is a convenient and accurate
characterization technique for the determination of the orientation distribution in epitaxial
ﬁlms. These orientation variants are distinguished by a change in the Kikuchi bands in
the patterns. As explained in the Chapter 4, the analysis of EBSP patterns from epitaxial
Cu, Co, and Py deposited at 90°C revealed that both variants coexist within the EBSP
interaction volume. However, in the ﬁlms grown at room temperature the patterns were
unsolvable, and without consistent band indexing, it is impossible to fully characterize
the variant size and distribution. Therefore, different crystallographic characterization
techniques, such as TEM and HRTEM, are needed to study these samples to observe the
existence of small grains, as well as to study the interface roughness of these samples.
Magneto-transport characterization: The results of magneto-transport studies showed
that the second antiparallel state in the CPP-MR hysteresis loop was not established due
to distortion of the exchange-bias coupling (pinning) between the antiferromagnetic
(FeMn) layer and the adjacent ferromagnetic (Py) layer as sample was exposed to high
magnetic ﬁelds (at the second parallel-state). It was concluded that the growth of
epitaxial Py and FeMn at moderate temperature (90°C) did not allow a strong exchange
bias coupling between the ferromagnetic and antiferromagnetic layers. Similar behavior
was also seen in the standard and micron-sizes polycrystalline EBSVs. Therefore,
additional micro-size epitaxial spin-valve samples need to be fabricated at low

temperature (30°C > T > -3 0°C) to conﬁrm the temperature dependence issue.

219