12"»
$1:
.- 3
”it

. ,. . . .. alumni: ....WJWAM .. ... x . ..
u A 1 'gbv.n. H >-. . . v. .-

.. 1 J v v .. , .9 A

. . .. ..obwav:w.utfltr\. WV. .30 . r.

A . .4 a... .z

. pr”, .35.. . ...n. ukh.......u.r...k¥.!...,

.... nL .
.55....»35... .... _ . .. . , . . . .
. . . . . . . . . 3.....-

i.

...“. . “1.5%.... ..
A;5..m. . u.
... Jain...
F . t . .
...... -uwfiwmn
. . . Julia.
. .12..

43:2
%

*4:
,I‘

’ 22:12:;
. 323? ' . 2%
‘1',“ -A [3!

D. ‘.v .
. :mnaémufi...»
LLB. .Wmflxfi ......
n , Kn Wigwam.» .2. fin.
. . ....5......\o..~m.w w... Jun,

.
'4‘...
’-

5. . ...,
1.4%....

. ....pnufla}: .... n

.....«r .. 1.. .14. 1:... -1... I
. pl 11.; 15‘ an.

.. ...”.uuflumhm.l. 1.1%.... ...».

H.)l..ln.¢.vr.u+...x.r

, -.ylfi....-..\l

rqu.I .\

. . . Cw
-- l
..v- ......»ha.
42“»... . .
. n... .. -.. .
. . ....uramé ....fim... m... .
.. . .. 6.1. 5......» finish.“ .
w..-.........4..li. . _.
.fifiud.....h.¢...,. . . .
:1 . 1:" l“)
. ... \ll

:I.v
.

. .. .4 ..
. ...... . Wfihka
......A......l.o ... u...
4.... .....ka¢

'
IO

."-

....
....v... .. Z... ..

.... $14.55....1uu. .tfi.Fn.R.

, . .... . .Mfl... gymnmnukwmummufiuwfiubhw. . . .

I4. ‘1 (gourd... l .. .. . .l . .

.3 .... ... flan . .. .. .. . .. ,. flu- ...,

,. . . ... .. . . . mfufiuu .15

Y “5......

‘ .
Wild-=4?
.‘_,' ..
”kW-Hf} '

.i‘

.
i».
.....Huzmmwfit...» .3. - ..u.
3.... . 4...... 2.4.5»... . . . . . . .. I.
. L. o ‘nflufififlzvlmmmu-Wfi. . . . . . . , , . ..., 1.3.3.» thruav-lkhfln.
, .ba.ten..:-u§um.il.v. . , -qhivvfinflhv,
. 53'. .I». K.
.
.. ”tun" Av.¢....!.|.

L53... ...
o . .11.}..mqgflhm: t . T . .
I . . ... . .. . . . ... I . . ...! ....-. «l ,

. .... . . . . . .. 5.1.5.5.“ --.“...r

}
3.! .
, ... - .
.. ..c.....».rliwl 2. . . .

0.1.3:.51-.. v.....r......v.lk.uc.x 9.4. .

’1...» ...
.--. ....vv‘ -..... ..ttv... ... Z 4
l .. 4' .
II. 4.}. “....” broil. ......DVII. ... . .
A SI‘A‘W-IO|D «0.! Vi.

.. ..Imr .H.

'A\.cd

V

. ID‘A‘I t '5‘ tb I . .. r a
......T . .54...- .. 1... . .
.0 a. . .
. . {$33- .. arr...-
uufic..- ‘tinb

h," t‘. .

.

Him». in? .

. foo... 0L.o..l..!n...\.|{

. ......4-.!..-........u!

. $14.1».HN...‘

. -.4. ..... {I sl
13......i...
,. ..l i

...]. .I y

:6. ....

I

h. t .
4‘19 1"
I. .1

' ‘

2". ,,

. .,
'09
i 'A V

{.9}. ' ..rw 04,1... ... .IL.‘ .

.H. '. A .-
. ..-. .--
b .....u 1 H3421] 90.0.. n. ... .1

. 5 t! 1- rain...“

.. .l; i -
._ 1414:1954 £53.... .. .. -

(In. . .t . l’htl ll
..th ”91.1. -..:o . Yuvvt
vl, .blv‘..>\1 y.\1..!...

ukulllt..v. ..-..-
All ..... I A .|

 

. I. . .
5.3K... cu...
. 5.. -1 : 4
1'1 . "‘.v‘ \ O. O. .‘a .t \ O .4. u I
....» .2:\...tk:.n. ...}...Nmfimiha . ... . t , z ..
3 .il‘x.n.nflw .. v . . . . . . .
. . . . .v . (£541 . v . u .. l. . w. . I
, b---‘.2\........¥ . . 13.....- I» .. . . . .
. . . . ..Dlxuuyv..uu..lu.\:|3:w\-f.p.l .. 1!”... .
1 ll .41- AI\ .1 b’r INVA- U?nl.b.|l
. Iii.-. .....qulfllnuhwhvnHPW.

 

w . ‘l t. ...

:7an

.1] ’1 I

‘5. I .

¢ .. .v . .5. HI.» use;
. {LI

1. V‘IIII.

F .il

 

 

THEStS

075W?)

ICHIGAN T TE RSITY LIBRARIES

lllllllllllllll)WWII!Hill/Ill 1mm

3 1293 01570 7569

W

This is to certify that the

dissertation entitled

"Irreversibility and Vortex Pinning in
Nb/CuX Multilayers"

presented by

Carl Norman Hoff

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Physics.

\ i. ' '
J jjor professor

MS U is an Affirmative Action/Sq ual Opportunity Institution 0-12771

 

 

Date December 1Q, 1996

 

 

 

 

 

 

PLACE IN RETURN BOX to remove We checkout from your record.
TO AVOID FINES return on or before dete due. . -

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WWI

Irreversibility and Vortex Pinning in Nb/CuX Multilayers
By

Carl Norman Hoff

A DISSERTATION

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

DOCTOR OF PmOSOPHY

Department of Physics and Astronomy

1996

II?

SUSC€
Cu, 4
susce
Meas
have

700%
belley
data h
COnsis-
also be
sputter

either I

perpem

ABSTRACT
Irreversibility and Vortex Pinning in Nb/CuX Multilayers

By

Carl Norman Hoff

The irreversibility between the zero field cooled (ZFC) and field cooled (FC)
susceptibilities of sputtered Nb/CuX multilayers has been measured. CuX is either pure
Cu, CuGe(2%), CuGe(5%) or CuMn(O.3%). The temperature T," below which the
susceptibility is irreversible is measured as a fimction of field perpendicular to the layers.
Measurements with the field parallel to the layers had irreproducible results. All samples
have Nb layers 280A thick with the normal metal layer thickness (A; between 20A and
700A with T,-,, measured in fields from 25G to 30006. The curve T,~,,(H) or H,,,(T) is
believed to be the phase boundary between the vortex solid and the vortex liquid. The

data has been fit to the form H,,,(7) = H,,,0(1 - T/TCO)P and values of P are found that are

consistent with theory. The critical current density Jo and pinning force density Fp have
also been measured in patterned Nb/CuX multilayers. These samples were fabricated by
sputtering through a trilayer photolithographic mask. In these measurements CuX is

either pure Cu, CuGe(2%), CuGe(lO%) or CuMn(O.3%). Measurements with the field

perpendicular and parallel to the layers were performed, both with the current in the plane

th

p6

fo'

the

of the multilayers and perpendicular to the field. Again all samples have Nb layers 280A
thick but the normal metal layer thickness (in was varied between 25A and 100A. In the
perpendicular orientation evidence of a transition from 3D line to 2D pancake vortices was
found in the pinning force density data. In the parallel orientation it is demonstrated that
the pinning strength of the N layers is affected by the addition of Ge, with no detrimental
effects on T c. Study of these superconducting/normal metal multilayers proves to be a
good approach to understanding similar process in high-Tc superconductors which have

limited the useful devices that can be made fiom these materials.

to my parents

iv

 

and su

CHCOUI

Sputter
early ir
Jim Mu
the sho;
he‘ll wit

SUPport 1
and than}
enCOUrag
Protjding

Fir
Phl’sics ar

ACKNOWLEDGMENTS

Many thanks go to my thesis advisor, Professor Jerry A. Cowen for his guidance
and support for the duration of this project. I am especially grateful for his patience and
encouragement during the writing and editing of this dissertation.

A special thanks goes to Reza Loloee for his help with the “Squids” and the
sputtering system. Thanks are due Mike Wilson for the use of some of his samples in the
early irreversibility measurements. Thanks are also due the shop guys, Tom Palazzolo,
Jim Muns, and Tom Hudson for the many times they rushed one of my projects through
the shop with sometimes very short notice. A special thanks goes to Carol F legler for her
help with the EDS studies. I’d also like to thank all the graduate students and staff at
MSU who helped me maintain my sanity over the past five and a half years.

My parents, J. M. and Onalee Hoff, deserve special recognition since without their
support for the last 28 years, I could never have made it this far. My greatest appreciation
and thanks must extend to my wife Patricia and my son Scott. I thank my wife for all her
encouragement, support, and love these past few years. I thank my son for always
providing a necessary diversion, and for reminding me of the really important things in life.

Financial support for this project was generously provided by the Department of
Physics and Astronomy, the Center of Fundamental Materials Research, and the National

Science Foundation.

LII

L15

C HA
SAM

33

CHAPTE
EXPERu

4.1

TABLE OF CONTENTS

LIST OF TABLES .................................. viii

LIST OF FIGURES .................................. ix

CHAPTER 1

INTRODUCTION .................................. 1

CHAPTER 2

SAMPLE PREPARATION AND FABRICATION ................. 4
2.1 Substrates .................................. 5
2.2 Patterning (for Critical Current Measurements) .............. 6
2.3 The Sputtering System ........................... 11

CHAPTER 3

SAMPLE CHARACTERIZATION .......................... 23
3 1 XRD .................................... 24
3.2 EDS ..................................... 24
3.3 Surface Profiles of Patterned Samples ................... 25

CHAPTER 4

EXPERIMENTAL PROCEDURES ......................... 29
4.1 Magnetization Measurements ....................... 30
4.2 Critical Current Measurements ....................... 36

CHAPTER 5

THE MAGNETIZATION IRREVERSIBILITY LINE ............... 42
5.1 Theory ................................... 43

 

I
"'7' ,

 

1
.
t, ....

C HA
F L'TI

LIST

5.2 Experimental Results ............................ 50

5.2.1 Perpendicular ............................ 53

5.2.2 Parallel ............................... 68

5.3 Conclusions ................................. 68
CHAPTER 6

CRITICAL CURRENT AND VORTEX PINNING ................. 72

6.1 Theory ................................... 73

6.2 Experimental Results ............................ 76

6.2.1 Perpendicular ............................ 77

6.2.2 Parallel ............................... 92

6.3 Conclusions ................................. 98
CHAPTER 7

FUTURE DIRECTIONS ............................... 99

LIST OF REFERENCES ............................... 101

vii

'.
r
‘.
I
P
. r-IV-V"

 

 

 

‘ 4‘, v»! ‘

Ta
Ta

Tat

Tabl

LIST OF TABLES

Table 1 - Typical gun setting and deposition rates used in this study .......... 21
Table 2 - EDS results of all CuX film types. ..................... 25
Table 3 - All samples measured with H perpendicular to the multilayers ........ 55

Table 4 -Fitting parameters for all samples measured in the perpendicular
orientation. ................................. 56

Table 5 - All samples measured in this study ...................... 76

viii

 

FQWc

Figure

Emmi

me4

meS

Hgne6

“@m7
Hflm8

Flé’llrefil

 

Figum 10 .
F

F1sure H

 

LIST OF FIGURES

Figure 1 - A typical Nb/CuX multilayer which is produced by sputtering on a Si
substrate. The normal metal layer thickness dN was varied from 25A
to 700A and the superconducting layer thickness was fixed at 280A for
this study. Typical in-plane dimensions are of the order of 1cm. . . . .

Figure 2 - The pattern (actual size) that was photographed to create the contact
film mask. The actual photolithographic pattern is reduced by a factor
of 15.4 ...................................

Figure 3 - The trilayer photoresist mask used to create the patterned samples. . .

Figure 4 - An illustration of a patterned sample. The sample wire is 50pm wide
and 5mm long while the contact pads are approximately 1mm by 2mm.
The top two contact pads are for the voltage leads while the bottom
two are for the current leads ........................

Figure 5 - Diagram of the substrate holders used in this study. The shutters could
be rotated by pulling on one of the pins with an externally controlled

wobble stick ................................

Figure 6 - Diagram of the Sample Positioning and Masking Apparatus (SPAMA)
plate ....................................

Figure 7 - Schematic cross section of the cooling system with sample holder. . . .
Figure 8 - A cross section of the triode magnetron sputtering gun ..........

Figure 9 - A schematic of the sputtering system (only three of the four guns are
shown for clarity). ............................

Figure 10 - A SEM image of a wire in a 5000A Nb film. The wire is 50pm wide. .

Figure 11 - A profile of a wire in a 5000A Nb film obtained on a Sloan
Dektak IIA .................................

Figure 12 - The core of the MPMS. .........................

ix

3

6

9

10

12

l3

16

17

20

27

28

31

 

Figure ‘

Figure 1
Figure 1
Figure 1
Figure 1

Figure 1

Figure l‘

Figure 21

Figure 2‘.

Figure 2

Figure 13 -

Figure 14 -
Figure 15 -
Figure 16 -
Figure 17 -

Figure 18 -

Figure 19 -

Figure 20 -

A screen capture from the MPMS control software. The negative squid
response as the sample moves through the center pickup coil, the 2cm

mark, indicates diamagnetic behavior ................... 32
Samples mounted for magnetization measurements ............ 35
A typical set of ZF C and FC magnetization curves ............ 37
Samples mounted for critical current measurements. .......... 39
A typical current sweep, this one taken at 5.6K and 600G ........ 41
The three dominant processes for a Cooper pair entering a normal

metal layer ................................. 46
The 3D/2D/3D transition seen in H 3.2 (T) ................. 49

The differences between successive data points on a FC magnetization
curve graphed to illustrate the sharp rise at T c ............... 51

Figure 21 - The difference between FC and ZFC data points measured at the same

Figure 22 -

Figure 23 -
Figure 24 -
Figure 25 -

Figure 26 -
Figure 27 -
Figure 28 -
Figure 29 -

Figure 30 -

temperature graphed to illustrate the sharp drop at T," .......... 52

H; (T) determined fi'om magnetization measurements (A) and

transport measurements (V) along with H;(T) detemiined
magnetically (O) .............................. 54

The H T phase diagram for the Nb 280A / CuGe(2%) 150A sample. . . 57
The H T phase diagram for the 5000A Nb film .............. 58

The H T phase diagram for the Nb 280A/ CuGe(5%) 100A sample. . . 59

H620 , H,”o , PC and Pm versus dN for the Nb/Cu samples ......... 60
He20 , H,”o , PC and Pm versus dN for the Nb/CuGe(2%) samples. . . . . 61
cho , H,”0 , PC and P,” versus dN for the Nb/CuGe(5%) samples. . . . . 62
H€20 , H,”0 , PC and P,-,, versus die for the Nb/CuMn(0.3%) samples. . . 63
7:0 versus dN for all sample types ..................... 65

 

Figure 3

Figure 3

Figure 3

Figure 3

Figure I

Figure I

 

" Figure i

Figure

Figure 31 - Pc, Pm and T graphed as a firnction of Ge concentration for the

co

Nb 280A / CuGe(x%) 20A samples ....................

Figure 32 - PC, P1,, and T o graphed as a firnction of Ge concentration for the

6‘

Nb 280A / CuGe(x%) 100A samples ...................

Figure 33 - The H T phase diagram for the Nb 280A / Cu 400A sample with the
field applied parallel to the layers .....................

Figure 34 - An illustration of a kinked vortex region. ................

Figure 35 - Several of the I-V curves collected for the Nb/Cu 280A/100A sample.
The right most curve is taken at 5K with each successive curve taken at
higher temperatures in 0.2K increments ..................

Figure 36 - The Hc2(T) curve for the Nb/Cu 280A/100A sample. ..........

Figure 37 - The critical current versus reduced field h at various temperatures for
the Nb/Cu 280A/ 100A sample .......................

Figure 38 - The pinning force density versus reduced field h at various temperatures
for the Nb/Cu 280A/100A sample .....................

Figure 39 - Scaling of Fp(h) for the Nb/Cu 28OA/100A sample ............

Figure 40 - Prefactors as a function of reduced temperature for the Nb/Cu
zsoNlooA sample. ...........................

Figure 41 - Scaling of Fp(h) for the Nb/CuGe(10%) 280A/25A sample ........
Figure 42 - Scaling of F p(h) for the Nb/CuGe(2%) 280A/25A sample. .......
Figure 43 - A graph of x versus (In for all sample types ................

Figure 44 - F, versus h at fixed reduced temperature for the Nb/CuGe(2%)
samples. .................................

Figure 45 - Fp versus h at fixed reduced temperature for the Nb/CuGe(10%)
samples. .................................

Figure 46 - Fp versus h at fixed reduced temperature for the Nb/Cu samples. . . . .

Figure 47 - Attempt at scaling Fp versus h" for the Nb/CuGe(10%) 280A/ 100A
sample ...................................

66

67

69

75

78

79

80

81

83

84

86

87

88

89

9O

91

93

 

 

Figure 4

Figure 4

Figure 5

Figure 5

Figure 48 - Attempt at scaling Fp versus 11" for the Nb/CuGe(10%) 280A/25A
sample ................................... 94

Figure 49 - The critical current versus reduced field h" at various temperatures for
the Nb/CuGe(lO%) zsoA/soA sample. ................. 95

Figure 50 - The critical current versus reduced field 11" at various temperatures for
the Nb/CuGe(10%) 280A/100A sample .................. 96

Figure 51 - F ,, versus h at three difl‘erent temperatures for the Nb/CuGe(2%) and
Nb/CuGe(lO%) 280A/ 100A samples ................... 97

xii

 

 

 

ther
are 1
unde

rdat

Vibe
mOV'l
drill:
Steac
aPpe
dusu
the u
dean
dfivu
affect

densh

Chapter 1
INTRODUCTION

The properties of anisotropic superconductors are particularly important in light of
the discovery of high-Tc superconductivity in anisotropic metallic oxides. Though these
are the best known anisotropic superconductors, they are the least understood. The
understanding of the physical properties of these high-Tc compounds is limited by the
relatively small size of parameter space that is explorable.

In all type-II superconductors, a vortex system is formed above the Meissner state.
When an external current density j is applied to the vortex system, the flux lines start to
move under the action of the Lorentz force. Within a perfect homogeneous system the
driving Lorentz force is counteracted only by a frictional force proportional to the
steady-state velocity of the vortex system. The consequence of this flux motion is the
appearance of a finite electric field and dissipation. The desired property of
dissipation-free current flow is then lost. It is precisely this problem that severely limits
the usefulness of high- T c superconductors in practical applications. In order to recover the
desired dissipation-free current flow, the flux lines have to be pinned. In this case the
driving Lorentz force is counteracted by a pinning force. Fortunately, any static disorder
affecting the superconducting order parameter will contribute to a finite pinning force

density, thereby reestablishing the technological usefulness of type-II superconductors.

l

 

 

 

 

How:
optim
force

depinr

high-7
focuse.
same r
Space i
either p
COOper
study 01
the thicl
Concentl

illustrate

SUPercor
critical c
melting .
dissipatio
mOIIOn p

imponan

2
However, the issue of dissipation-free current flow has now become a matter of
optimizing the pinning force density, since increasing the current density until the Lorentz
force overcomes the pinning force causes dissipation to reappear as the vortices become
depinned.

Much work1 both theoretical and experimental has been done on vortices in
high-Tc superconductors, but there is still much that is not well understood. This study
focuses on low-Tc superconductor / normal metal multilayers which show many of the
same properties as the high-Tc superconductor but the range of explorable parameter
space is much larger. All samples in this study are Nb/CuX multilayers, where CuX is
either pure copper, CuGe, or CuMn. Ge is a nonmagnetic impurity which doesn’t break
Cooper pairs, and Mn is a magnetic impurity which does break Cooper pairs. Unlike the
study of high-Tc compounds where the number of variable parameters is very limited, here
the thickness of the superconducting layer ds, the normal metal layer thickness die, and the
concentration and type of impurity in the normal metal layer can all be varied. Figure 1
illustrates all these variable parameters.

This study focuses on the investigation of two properties common to all type-II
superconductors, the irreversibility line seen in magnetization measurements, and the
critical current density. The irreversibility line, which is proposed to arise from the
melting of a vortex solid to a vortex liquid, is of great interest to the study of
dissipation-free current flow as only the vortex solid phase can be effectively pinned as the
motion present in the liquid phase leads to dissipation. The critical current density is

important as it allows one to calculate the pinning force density directly.

Cu
Cu X CuGe(x%)
CuMn(y°/o)

 

gu A typic Nb oduced by sp
The norm kn varied fi'om A A an
superco kn ed at 280A fo yp
' plane 1cm.

7"

Suue

Thec

 

desig
depos
run al
base ;
purity
backgi
lmed L

SUbstral

Process (

Spullellng

Chapter 2
SAMPLE PREPARATION AND FABRICATION

The films and multilayered samples used in this study were fabricated at Michigan
State University, using the CMP (Condensed Matter Physics group) sputtering system.
The computer controlled ultra high vacuum (UHV) DC magnetron sputtering system was
designed primarily by Dr. William Pratt Jr. and built by Simard Inc.2 The system allows
deposition from up to four different L. M. Simard “Tri-Mag” sputtering guns in a single
run and holds up to 16 half-inch-square substrates or 32 smaller patterned substrates. The
base pressure of the system prior to sputtering is typically 3 2.0x10’8 torr. The ultra high
purity Ar used in the sputtering process, coupled with the low base pressure, limits
background contamination of the gas in the chamber to less than 10 ppm. All samples
used in this study were deposited onto polished, single-crystal, (001)-oriented Si
substrates.

The remainder of this section will detail substrate preparation, the patterning
process (used for critical current measurements only), mounting of the samples in the

sputtering system, and preparation and general aspects of the sputtering system.

 

 

2.1

used:
squar
each

dehyt

areli

rnour
cut it
hexar
593 1.

at 9Cf
houra
Inflkyi
healing
lirh dei'
SUbsnaze:
dehydrate

are blOLm ‘

2.1 Substrates

All substrates are polished, single-crystal, (GOD-oriented Si. For samples to be
used in magnetization measurements, 3 inch diameter wafers3 are cleaved into half inch
squares. The square substrates are then sequentially washed ultrasonically for 5 minutes
each in the following: alconox and deionized water, deionized water, acetone, and
dehydrated ethyl alcohol. As soon as the substrates are removed from the alcohol, they
are blown dry with compressed nitrogen and mounted in the sputtering system.

For samples to be used in the critical current measurements, a 3 inch wafer is
mounted with paraffin in a surface grinder. A diamond impregnated wheel is then used to
cut it into 0.16 inch by 0.25 inch rectangles. These substrates are washed ultrasonically in
hexanes for 1 hour to remove any remaining paraffin, and then cleaned ultrasonically in a
5% MICRO4 solution for 5 minutes. While still in the solution, they are heated for 1 hour
at 90°C, cleaned in the ultrasonic bath for 5 more minutes, and heated for an additional
hour at 90°C. By the end of the cleaning process, the 5% MICRO solution turns from a
milky color to water clear. If any color remains after 1 hour into the final heating, the
heating continues until all color is gone. The substrates are rinsed 5 times ultrasonically
with deionized water for 5 minutes each time to remove all the cleaning solution. The
substrates are then sequentially washed ultrasonically for 5 minutes each in acetone, and
dehydrated ethyl alcohol. As soon as the substrates are removed from the alcohol, they

are blown dry with compressed nitrogen.

6

2.2 Patterning (for Critical Current Measurements)

Before the fabrication of a lithographic mask on the substrates can be completed,
one needs a contact film mask. It is called a contact mask, because the film is placed in
direct contact with the spun photoresist when the pattern is transferred. Therefore the
transferred pattern has dimensions essentially identical to the film pattern. The first step in
creating the contact film mask, is generating the black and white pattern using a computer
drawing package. The second step is printing the pattern on high quality paper, using a
good laser printer to assure a high contrast image. Figure 2 shows the pattern used in
these experiments. The third step is photographing the image, to transfer the pattern to
film. For this step a Nikon FM2 camera is loaded with Kodak T-Max ISO 100
Professional B&W film.’ Several pictures are taken with different parameters to assure

that a good film mask will result. The film is developed using standard dark room

 

 

Figure 2 - The pattern (actual size) that was photographed to create the contact film mask.
The actual photolithographic pattern is reduced by a factor of 15.4.

7

procedures and allowed to dry before it is inspected under a microscope to find the best

film mask. For this experiment, the best mask was produced using a F-stop of 5.6 with a

shutter time of 1/4 second.

A clean rectangular substrate is now ready for fabrication of the trilayer mask that

will be used to create the patterned sample as illustrated in Figure 3. There are 10 steps in

the creation of the mask. They are:

1.

Spin on 1813 photoresist‘5 at 4000 rpm for 30 seconds. This creates a uniform layer
approximately 1.4um thick.

Bake the sample for 1 hour at 90°C. If the temperature goes above 104°C the
photoresist becomes so hard that it cannot be easily removed with acetone and

therefore may not be used.

. Expose the whole photoresist layer with an ultra-violet (UV) lamp for 8 seconds.

Evaporate a 340A layer of Al at no more than 2 A/second. The evaporator reaches a
base pressure of 6x1045 tort. The slow deposition rate is required to avoid excessively
heating the photoresist and hardening the film. The thickness and deposition rate are
monitored by a quartz crystal film thickness monitor.

Repeat step 1. This creates a second 1.4um layer.

Repeat step 2.

Expose the photoresist layer covered with the desired 35mm film mask with a
UV lamp for 9 seconds.

Place the sample in 452 developer6 until the top layer of photoresist, the Al layer, and

finally the photoresist beneath the Al layer is removed in the exposed areas. This step

8
takes about 2 minutes with new photoresist but can take up to 5 minutes with older
photoresist. Photoresist older than 2 months should not be used. The progress of the
developing can be checked under a microscope. If the developing is not complete (an
undercut of about 2pm in the bottom photoresist layer should be seen) the sample may
be placed back in the developer. Once developing is complete the sample is removed
from the developer, rinsed in deionized water and blown dry with compressed N2.

9. Repeat step 2. This bake is solely to help dry the samples and minimize outgassing
when the samples are placed in the ultra high vacuum of the sputtering system.

10. Place the sample in a Plasma Therm7 Batchtop System VII Reactive Ion Etcher (RIE).
This etch is intended to remove any photoresist lefi on the Si substrates in the exposed
areas. The system is pumped and purged several times until a base pressure of
13 mtorr or less is reached. Then a 20 watt oxygen etch is run for 30 seconds. The
pressure set point is 50 mtorr and the oxygen flow rate is 7.5 sccm.

The samples are now ready to be mounted in the sputtering system. After the film
or multilayer is deposited through the mask, the sample is removed from the sputtering
system and the trilayer mask is removed by soaking the sample in acetone for a few

minutes. See Figure 4 for an illustration of the final patterned sample.

 

35 mm Film Mas '

1.4umPR

 

m ... .
same <-. i:
s s

->.~.

4 ..
e. e

t

3 nm A1 Mamie”:
xv “a“ *-

 

 

 

 

Exposed

 

 

 

,5

I.

u
4%
“gm

2%?

u

Removed by Developer

WM:

at:

#1:):va

<5.
,

r‘
«-

mistake; -

fifth-3.»
5.-

u; -

-.-.\
v~

 

Figure 3 - The trilayer photoresist mask used to create the patterned samples.

”95% W: A ‘53.“, H¢w5y5 um. )yyn u. surges“-

u

”.5.,
#“fimfi’

Wayne‘s,

.
:2?

5?
95
are}.

.-.r
I.-<

.5

w?
3235.?

.5... “5 5555 555 5-
335-3. 5......555555 5555555555> .5.

\..
-5 5 5 ~55 5 555
5252:\ h- ;.t-55555.%~.:5: 555a Mix “35.5 «.5
. 5 5 55555555 55 5 .5555 . 5.55.555 5.5 5.5.5.-
iezi‘rfis‘i ‘25: 55.5 ‘2 5&3: Ezimu 3:55:35“: 55.33.3352“ ‘5 ”eh-5:555:55 m 5:55:25
3:52‘1535555
5555

-5 ““25

55
55
35.35555

stdh’fl. v M“

55 5
3‘5??? K553555555 were “ _5
“-555.53
.

 

 

 

4.1 mm

Figure 4 - An illustration of a patterned sample. The sample wire is 50pm
wide and 5mm long while the contact pads are approximately
1mm by 2mm. The top two contact pads are for the voltage
leads while the bottom two are for the current leads.

11

2.3 The Sputtering System

Both the square substrates and the patterned substrates are mounted in an
aluminum holder as shown in Figure 5. The system can accommodate up to 8 substrate
holders, each having 2 locations designed to hold a square substrate. A half inch square
stainless steel support can be placed in each location to hold two of the patterned
substrates. The holders are arranged on the SPAMA (Substrate Positioning And Masking
Apparatus) plate as shown in Figure 6. Each holder employs one of two types of rotatable
stainless steel shutters. The first type has one opening so that only one substrate location
at a time can be used for deposition. The second type has two openings so that both
substrate locations can be deposited on simultaneously. Both types of shutters also allow
both substrate locations to be covered simultaneously to protect the substrates or already
grown samples from unwanted deposition while a sample is being grown in another
holder. The two holed shutters are good for making multiple samples with the same
parameters. In addition to the substrate holders, the SPAMA plate also holds two quartz
crystal film thickness monitors, one each for Nb and CuX.

The sputtering chamber is capable of reaching an ultimate base pressure of less
than 2.0x10’8 torr. To achieve this pressure a high speed CTI Cryo Torr 8 cryopump is
used to pump the system down, which takes roughly 2 days. The system is baked at 60°C
during the first 10 to 12 hours of pumping. Once the base pressure is reached the partial
pressures of the background gases inside the chamber are measured using a Dycor

Electronics M100 Quadrapole Gas Analyzer. The gasses checked included He, H20, N2,

      
  
   

pin for rotating
shutter

substrate exposed
only with the two
substrate holed shutters
exposed

Figure 5 - Diagram of the substrate holders used in this study.
The shutters could be rotated by pulling on one of the
pins with an externally controlled wobble stick.

Sample holder
positions (SH)

  
  

Film thickness monitor

positions (FTM) Substrate

radius

Figure 6 - Diagram of the Sample Positioning and Masking Apparatus (SPAMA) plate.

14

02, and Ar. The partial pressures of H20 vapor and N2 gas are always the highest with
partial pressures between 10'8 and 10'9 torr. All other gases have partial pressures at least
an order of magnitude lower.

The substrate temperature is stabilized during sample deposition by using a
Cu block and foil to thermally link the substrate to the Al holder. Since the Cu blocks are
pressed against the back of the substrates they also keep the substrates from moving as the
SPAMA plate rotates. The Al holders are in turn thermally linked to an internal liquid N;
reservoir located at the bottom of the shafi holding the SPAMA plate as shown in
Figure 7. The internal reservoir is kept full by passing high pressure (1000 psi) N2 gas
through a fine stainless steel capillary cooled by an external heat exchanger. The heat
exchanger consists of the capillary welded inside a length of 1/4 inch copper tubing,
through which liquid N; is constantly flowing. The high pressure N2 gas is then liquefied
and continues through the fine stainless steel capillary to the internal reservoir. With this
cooling system, heat that is deposited in the substrate by the incoming sputtered atoms is
conducted away from the sample to the Al sample holder by the Cu foil. Heat is then
progressively transferred to the SPAMA plate, to four OFHC (Oxygen Free High
Conductivity) Cu rods, and finally to the internal liquid N2 reservoir where evaporated N2
gas carries the heat out of the system through a second stainless steel capillary. An
important part of this design is the use of the small capillaries which permit the whole
assembly of the internal reservoir and the SPAMA plate to rotate. The Cu tube carrying
liquid N; in the external heat exchanger continues on to a large internal ring or Meissner

trap around the top of the chamber before being vented to the outside. The Meissner trap

15

is used to freeze out impurities inside the chamber. A thermocouple placed on this ring is
used to confirm that liquid N; is flowing. With the cooling system in place, thermocouples
placed on top of the Cu blocks are used to monitor the temperature throughout the
sputtering run. Since each thermocouple is separated fiom the actual deposition surface
by the Cu block and the Si substrate, the samples certainly get hotter than the monitored
temperatures. For the typical square substrates, the thermocouple temperatures are kept
between -20°C and +20°C for the duration of a run. With the patterned substrates, these
temperatures are kept between -60°C and 40°C. The reason for the lower temperatures
for the patterned samples is two fold. First, it minimizes outgassing from the photoresist
and second, it avoids excessive heating and hardening of the photoresist.

The sputtering system contained four different L. M. Simard “Tri-Mag” or triode,
magnetron sputtering guns. A schematic of one gun is show in Figure 8. These four guns
are housed in the stainless steel cylindrical vacuum chamber of the sputtering system
placed 14cm from the cylinder center (as measured from the target center) and spaced 90°
apart. In the present experiment, all samples are either films or repeated bilayers
(multilayers), so at most only two guns are required per sample. The guns are covered by
a manually controlled rotating chimney assembly. Each gun requires two chimneys. The
first is completely covered with Al foil to block the deposition beam from reaching the
substrate, thereby protecting an exposed substrate from deposition of unwanted material
before its shutter can be closed, for example immediately afier a sample has been
deposited. The second chimney which has a 5cm diameter hole, is rotated into place while

the sample deposition takes place.

N2 gas

‘—
—->

16

thermocouple

    
 

 

 

pressure
screw

 

  
  

 

 

 

T? L—HK

 

Internal
Reservoir

Figure 7 - Schematic cross section of the cooling system with sample holder.

_ _ bridge
capillaries
)/ sam 1 substrate
p e
' Exchanger v
k _| . ,, I
\ |‘,_pin for
rotating
SPAMA plate shutter
Cu block
Stainless steel shutter
Cu rod support for the smaller
patterned substrates
- OFHC Cu

     
   
 

Magnetic Aluminum gun
confinement housing
shield

 

.AIl

 

EH

Tantalum TCooling waterl
filament

Figure 8 - A cross section of the triode magnetron sputtering gun.

18

In the sputtering process, ionized Ar atoms are accelerated toward the target. On
impact with the target, the kinetic energy of the ion is transferred to target atoms, ejecting
them fi'om the target surface. This process results in a diffuse beam of target atoms. The
starting gas is 99.999% Ar, passed through a liquid N2 cold trap, then a gas purifier, and
finally forced through several small closely spaced openings at the gun base. The Ar
atoms are ionized by establishing an electric current between the anode and the filament.
A permanent magnet produces the field used to confine the Ar plasma to the region just
above the target. A negative potential is applied between the Ar plasma and the target to
accelerate the ions toward the target. During sample deposition, the chimney with the
hole limits the ejected material to a beam approximately 5cm across. This constraint on
the beam limits unnecessary deposition on system components. Finally a substrate placed
12cm above the target collects some of the target material.

Although material is readily ejected from the target surface, most of the kinetic
energy of the Ar ions is deposited in the target as heat. Therefore the target is mounted
on a Cu base which is attached to a Cu block containing circulating water to remove the
accumulated heat buildup. Once the Ar ions have struck the target, they acquire an
electron from the metal surface and are re-emitted as neutral Ar atoms. To provide these
electrons in a DC sputtering system, electrically conducting targets must be used.

The sputtering system is computer controlled. The computer directly controls the
angular position of the SPAMA plate and performs all the calculations needed to deposit
films or multilayers with the proper constituent thicknesses. The computer also monitors

the voltage applied to, and the current passing through, the target. The deposition rates

19
are measured several times during each run using the FTM’s, then manually entered into
the computer. This information is used by the computer to determine the proper
deposition times for the constituent layers.

Before sputtering, the pressure is brought up to 2.5 mtorr by backfilling the
chamber with the high purity Ar gas. The pressure is held in dynamic equilibrium
throughout the run by flowing Ar at a constant rate through each gun and by opening the
gate valve to the pump just enough to maintain 2.5 mtorr. The deposition begins by
turning on the necessary sputtering sources and letting them equilibrate, which takes less
than 10 minutes. Since no more than two guns are on at any one time, heat buildup on the
SPAMA plate is minimized. The target voltages and currents are fixed from run to run for
each material so that the deposition rates are similar for samples made in different runs. A
schematic of the sputtering system is shown in Figure 9 with typical gun settings and
deposition rates listed in Table 1. Afier the deposition rates are measured and entered into
the computer, the shutter below the substrate ready for deposition is manually opened.
The sample is rotated to a safe position away from any guns that are on. Then the
manually controlled chimney assembly is rotated so the open chimneys are above the
targets. When a film is to be grown, the computer rotates the substrate to a position
above the desired gun for a time determined by the entered deposition rate and the desired
film thickness. Multilayers are grown by rotating the substrate back and forth between
two sources. After the sample has been grown, it is again rotated to a safe position and
the chimney assembly is rotated so the covered chimneys are above the targets. The

shutter is then closed and the sample is protected until it is ready to be removed from the

 

1
,. . . . n.—-
‘ ~—-.

3%"

Fig

20

stepping motor

substrate holder

atmosphere

 

SPAMA plate
\

 

 

Chimney _ Chimney
with hole ’ without
hole
, sputtering
gun

 

 

 

 

 

 

 

Figure 9 - A schematic of the sputtering system (only three of the four guns are shown for
clarity).

21
sputtering system. Once removed fi'om the sputtering system, samples are stored in a
vacuum dessicator to limit surface oxidation.

Sputtering targets are cut to the desired cylindrical shape, 5.72cm in diameter by
0.64cm thick, by the stafi‘ of the MSU Physics Shop. Pure Nb and Cu targets are cut from
0.64cm thick plate obtained from Angstrom Sciences.8 The purity of these bulk materials
is 99.95% for Nb and 99.999% for Cu. The CuX alloy targets are cut fi'om alloyed ingots
fashioned at MSU using an rf-induction furnace. The alloy targets are made from
99.9999% pure Cu, 99.9% pure Mn, and 99.999% pure Ge, obtained from Aesar/Johnson
Matthey Inc.9 The final Ge and Mn concentrations are determined using EDS on thick

sputtered films as described in section 2 of Chapter 3.

Table 1 - Typical gun setting and deposition rates used in this study.

Metal T V T Current A
Nb 0.7
Cu 0.7

0.7

0.7
0.7
0.7

 

22

An acid etch consisting of a mixture of 50% I'INO3 and 50% deionized water is
used to clean most of the parts of the sputtering system. These include the sample
holders, sample shutters, SPAMA plate, sample supports (for the patterned substrates),
gun bodies, and the Al target confinement rings. Deposits not readily removed by the acid
are scrubbed vigorously with an iron wire brush and then re-etched. Nb proved to be very
diflicult to remove from the gun parts because the Al gun parts etched away nearly as fast
as the unwanted Nb deposits. Therefore a razor blade is used to scrape away most of the
Nb, after which approximately 5% HF is added to the acid solution to help remove the
rest. Since stainless steel is unaffected by HF, all stainless steel parts are etched in the
I'INO3, deionized water, and HF solution. After the acid etch, all these parts are rinsed in
deionized water and sequentially washed ultrasonically for 5 minutes in acetone, and
dehydrated ethyl alcohol. The SPAMA plate is too large to fit in the ultrasound bath so it
is rinsed by hand with acetone and alcohol. Each gun part is only used with a designated
type of target to minimize the possibility of cross contamination. The chimneys are never
cleaned directly, but before each run they are covered with Al foil which is discarded at
the end of the run. The magnetic confinement plates, sample bridge assemblies, Cu heat
sink blocks, and thermocouples are also not cleaned before each run, but kept clean
between runs. These components are not cleaned because they are either, never in contact
with the sputtered metal beam (and therefore have no significant deposits on them), are
not in close proximity with the polished face of the substrate, or are highly reactive in all
acids and so are discarded when their deposits become severe, which is the case for the

magnetic confinement plates, which are made of iron.

Chapter 3
SAMPLE CHARACTERIZATION

All of the samples used in these experiments were either films or multilayers. The
multilayers were of the form Nb/CuX, where CuX was either pure Cu, CuGe(2%),
CuGe(5%), CuGe(lO%), or CuMn(O.3%). The purpose of sample characterization was to
determine or verify the structure of the samples. What is the actual Ge or Mn
concentration in the Nb/CuGe and Nb/CuMn multilayers? Were the layer and film
thicknesses as expected fi'om the sputtering conditions? These were just a few of the
questions addressed in the analyses. To answer these and other questions posed about the
structure of the samples three basic apparatuses were used. The CFMR (Center for
Fundamental Materials Research) Rigaku powder x-ray difl‘ractometer was used to check
the layer thicknesses. EDS (Energy Dispersive Spectroscopy) measurements were taken
on a JOEL ISM-6400 SEM to check the concentration of impurities in the CuX layers.
Measurements of the surface profile of the patterned samples were made with a Sloan
Dektak IIA.

In this chapter, the results of these measurements are discussed.

23

24

3.1 XRD

Most of the x-ray diffraction data were taken and analyzed by M. L. Wilson.
Details of the difl'ractometer and several data sets are found in his dissertation”. He found
the bilayer separation A of the multilayers to be within 5% of their expected value, and the
rms deviations of the Nb and CuX layer thicknesses to be no larger than 3A. The Nb and
CuX layers are crystalline and preferentially oriented with the Nb bcc (110) and the CuX
fcc (111) planes parallel to the substrate surface to within 5°.

All of the x-ray studies were made on non-pattemed samples. Also the bilayer
separation had to be S 100A to identify the Bragg peaks fiom the layering of the sample.
Since the Nb layer was fixed at 280A, such measurements could not be performed on the
multilayers used in this study but similar results were assumed for these thicker layered

samples.

3.2 EDS

EDS is used to determine the concentration of desired and undesired impurities in
the layers. The EDS measurements are performed on a JOEL ISM-6400 SEM. EDS
measures the elemental composition of a sample by exciting the atoms in the sample with a
high energy electron beam. For this study a 15 kV accelerating voltage is used. The
relative fluorescent x-ray line intensities are then compared using a standardless analysis

program to obtain the relative atomic percentages. Table 2 lists the EDS results of all

25

Table 2 - EDS results of all CuX film types.

 

 

 

 

 

 

 

 

 

 

CuX Type Cu Wt% Impurity Wt%

Cu 100% None

CuGe(2%) 98.00i0.05 Ge 2.00:0.05
CuGe(5%) 95. 101:0. 10 Ge 4.90i0.10
CuGe(l 0%) 89.85i0.11 Ge 10.151011
CuMn(O.3%) 99.73i0.05 Mn 0.27:0.05

 

 

CuX type films used in this study. The EDS data showed no evidence for unwanted
impurities in either the Nb or CuX films above the detection limit of the method (less than

0.1 at.%). However a beryllium window located inside the collimator absorbs lower

energy x-rays and prohibits the detection of elements with atomic number s 11.

3.3 Surface Profiles of Patterned Samples

A Sloan Dektak IIA was used to obtain surface profiles of the patterned samples
to determine the width of the wire. In Figure 10 is a SEM image of a wire in a 5000A Nb
film. The same wire is profiled in Figure 11. The profile shows that the wire width at half

height is 50pm. The SEM image shows an edge roughness of about flum. All patterned
samples similarly checked are consistent with a wire width of 50:3 um.

The surface profiles also provide a check on the film thicknesses expected from the

sputtering conditions. As seen in Figure 11, the film is measured to be 5000A thick,

26
exactly as expected from the sputtering conditions. All similarly checked films were well

within 5% of their expected value, in agreement with the XRD results.

27

 

is 50pm wide.

wire in a 5000A Nb film. The wire

Figure l0 - A SEM image of a

28

 

 

 

 

 

 

 

5000 b ‘35:.“ V‘:_:‘:_:“_- #‘ ..V‘f _
4000 r _

i
3000 - 3 0: -

"i k 50 pm +3
2000 - f i .
1000 5 i _

0 10 20 30 40 50 60 70
pm

Figure 11 - A profile of a wire in a 5000A Nb film obtained on a Sloan Dektak IIA.

Chapter 4
EXPERIMENTAL PROCEDURES

The experimental results reported in this thesis were obtained from magnetization
and critical current measurements. Both measurements were made using a Quantum
Designll Magnetic Property Measurement System (MPMS). The magnetization
measurements used the system’s SQUID magnetometer circuits and the MPMS control
software. The critical current measurements used the MPMS as a computer controlled
gas flow cryostat and magnetic field platform. The control sofiware was also used to run
External Device Control (EDC) programs that controlled a Keithley12 224 Programmable
Current Source and a Keithley 182 Sensitive Digital Voltmeter.

The mounting of the samples for magnetization and critical current measurements,
and the data taking procedures will be described in the remainder of this chapter.

Illustrative data will also be shown.

29

30

4.1 Magnetization Measurements

Magnetization measurements are made using the MPMS’s superconducting
quantum interference device (SQUID) system. The SQUID is coupled to a series of
pickup coils, wound in a second derivative configuration in which the upper and lower
single turns are counterwound with respect to the two-tum center coil. This configuration
strongly rejects interference from nearby magnetic sources. The separation between the
top and bottom coils is 3cm. A measurement is taken by moving the sample through the
pickup coils, in a 4cm scan centered on the two-tum center coil. Throughout the scan the
SQUID voltage is read and a SQUID response versus sample position curve is generated.
This scan is generally repeated 3 times and the 3 curves averaged. The average curve is fit
with a theoretical form and the best fit is used to calculate the sample magnetization in
EMU. The standard deviation is determined from the variation in the 3 scans. Figure 12
illustrates the core of the MPMS. Figure 13 is a screen capture fiom the MPMS control
software illustrating the SQUID response versus sample position curve.

The above calculations are all internal to the MPMS control software. The control
software is also used to generate a list of system commands such as setting the
temperature or field. A generated list of these commands is called a sequence. Once a
sample is mounted, loaded into the system, and a sequence started, the system can be left
unattended until the sequence finishes. The user will return and find his data (T, H, z, A 1,

etc.) tabulated in a file on the system’s computer.

31

Superconducting R Superconducting
Solenoid Solenoid
+ O o
+ 0 OR
A
-o / o
R

 

 

 

 

 

 

 

 

 

Sample / \\\ Squid Pickup Coils
Vacuum Region
Vacuum Tube
Jacket Tube He Gas Flow
Sample Tube Thermometers

 

/ Gas Heater

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 12 - The core of the MPMS.

32

 

 

 

 

 

 

Set Paraueters

 

 

 

 

 

 

flagnetic Field (6) 188.88
Temperature (K) 5.88
§
g '85 Tenperature Undercool Off
- Nunber Data Points 58
§ 1 Scans to average 3
-1sl Scan Length (on) 4.88
' Long. filgorithn Iterative Reg.
_2 . L . . . . . Trans. algorithm Iterative Reg.
0 as 1 L5 2 z: 3 as
:amu<am>
Date 82-16-95 Tine 14:39:88 Heasurenent axis Longitudinal

Longitudinal Range 8.625888 anu [0R 1
Transverse Range 8.888358 8"“ [RR ]

 

 

 

 

 

Tenperature 5.88 R Stable
Field 188.88 6 Stable
He Level 76.88 x Hagnet Charging Node No Overshoot
Node Ready to accept Connands High Resolution Node Enabled
Data File Name CH123_§8
, D DC Data Block Type Summary

 

Figure 13 ~ A screen capture from the MPMS control software. The negative squid
response as the sample moves through the center pickup coil, the 2cm mark,

indicates diamagnetic behavior.

 

33

For temperature control, the sample is placed in 1 mtorr He exchange gas, which
determines the sample temperature by coupling the sample to the He gas from the main
reservoir of the system. This gas is pumped past a heater and around the sample chamber.
The temperature is monitored by a pair of thermocouples in a different exchange gas also
separated from the pumped gas. The field is controlled by regulating the amount of
current supplied to a superconducting solenoid that encompasses the cylindrical sample
chamber. This solenoid generates a field directed along the central axis of the sample
chamber.

This measurement technique can cause problems in extracting the magnetization of
a superconducting sample'3’”, because the sample is moved during the measurements. A
small inhomogeneity in the applied field or temperature over the sample path can produce
very misleading results. These concerns determined the 4cm scan length, since the
“apparent” magnetization proved to be independent of scan length for scans of 4cm or
less.

The sample is mounted inside a suprasil quartz tube15 suspended from the bottom
of a sample rod supplied by Quantum Design. The tube has an inner diameter of 6.7Smm,
an outer diameter of 7.5mm, and is 20cm long. Kapton tape" is wrapped around the base
of the sample rod to match the inner diameter of the tube. The tube is slid over this tape
for support. About 6.5cm from the bottom of the tube are 8 small holes, through which
white thread is passed to support the sample. The holes, which were cut by the MSU
glass shop, are divided into 2 planes with 4 holes each. The 2 planes are separated by

0.35mm or about the thickness of the sample substrates. For insertion into the tube, a

34
4.5mm square sample is cleaved from the half inch square sample created in the sputtering
system. Once in the tube, the sample is tied into place with white thread. The sample is
now ready to be loaded into the MPMS. For this experiment two sample orientations are
used, one with the applied field perpendicular to the plane of the sample and the other with
the applied field in the sample plane. Figure 14 shows a sketch of a sample mounted in
both orientations.

Before the sample is lowered down the sample column, it is put into a loading
airlock which is purged several times with He gas. The superconducting solenoid is also
soft quenched to remove any remnant fields left by the last user of the system. The sample
is lowered down the sample column and the temperature is allowed to stabilize at 5K. The
4cm scan length is then centered on the center SQUID pick up coil.

With the sample in place, a sequence to determine the irreversibility temperature as
a function of applied field, T,,,(H), can be started. The sequence repeats a series of zero
field cooled (ZFC) and field cooled (FC) temperature sweeps for a number of difi‘erent
fields. The sequence is programmed to start by warming the sample to above Tc and
cooling the sample to well below T c in zero field. The field is then applied and
magnetization measurements are made in increments of 0.05K as the temperature is raised.
These measurements make up the ZFC temperature sweep. Once the temperature is
several tenths of a degree above T c, magnetization measurements are again made every
0.05K as the temperature is lowered. These measurements make up the FC temperature
sweep. The FC measurements continue until a clear divergence is seen between the two

temperature sweeps. The field is then turned off, the temperature set to above T c, and the

35

Sample Rod A

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

4 l
———Kapton Tape
W)
U U\
9 V >
(U
Perpendicular
/ Orlentatlon
. Sample
Whlte Thread
V4; ———————————————————————— : -----
.1 \
h [l/ 2 \\
\
FY // \\
. / ‘ \
\ l _ gr \
\ :1 / \l
\ \ : 5 l
\ E I
\ \ \ ‘\ ‘ O > I;
\ \ \\ v ' ll
. \\ “ , 'l
A \ \ \\ (b " //
\ \ \\\\ l/
\ Parallel /

 

Suprasil Quartz Tube \\\Orientation /,.//

‘ ,’

Figure 14 - Samples mounted for magnetization measurements.

36
sample again zero field cooled. The field is then raised to the next desired field value and
the process is repeated. The desired field values are always set in increasing order to
minimize the effect of remnant fields.
Typical ZFC and FC data are displayed in Figure 15. T c is defined as the onset of

diamagnetism and Tm is the temperature below which the magnetization is irreversible.

4.2 Critical Current Measurements

All critical current measurements are also taken in the DC MPMS which allowed
measuring fields up to 5.500x104G. Since the same apparatus is used for these
measurements, the temperature and field is controlled exactly as described in the previous
section. There is the added benefit that the sample is stationary during the measurement,
meaning the sample is not exposed to the small inhomogeneity in the applied field or
temperature over a sample path. For a critical current measurement, the MPMS control
software is used to run external device control (EDC) programs. The two external
devices controlled by these programs are a Keithley 224 Programmable Current Source
and a Keithley 182 Sensitive Digital Voltmeter. The EDC programs set the current from
the current source and read the voltage from the digital voltmeter. The user will return
and find his data (T, H, I, and V) tabulated in a file on the system’s computer.

The sample is mounted at the bottom of a special EDC sample rod supplied by
Quantum Design, which contains small Cu wires for the electrical measurements. Two

such sample rods are used in this study, one for parallel measurements and another for

37

 

 

 

 

Nb/Cu 280A/100A
Field = 2506
0.0000 ~ a mu nun ummmmql
,. ‘__I:s=__ 1’
g -0.0004;*********t**fi -
e is? l
a T. = 7.725K
.3 -0.0000 ~ 4: ’5‘ —
d
{-3 E T”. = 7.45K
a -0.0012 ~ E -
51
a? 5+
-0.0016 ~ +51 zpc ~
E
_0.0020 . r L L r r r 1 4 r . 1
6.8 7.0 7.2 7.4 7.6 7.8 0.0

Temp (K)

Figure 15 - A typical set of ZFC and FC magnetization curves.

38
perpendicular measurements. Figure 16 shows a sketch of a sample mounted in both
orientations. The sample is heat sunk to a large metal platform at the bottom of the
sample rod with a small amount of Apiezon M Grease". Electrical contact is made to the
sample using silver paint to connect the two current leads to the current contact pads, and
the two voltage leads to the voltage contact pads. Once the silver paint has dried, the
sample is ready to be loaded into the MPMS.

As before, the sample is put in the loading airlock, which is purged several times,
and the superconducting solenoid is sofi quenched to remove any remnant field. The
sample is lowered down the sample column slowly to avoid rapid sample cooling which
may cause the silver paint electrical contacts to come loose. The sample is then stabilized
at 5K and positioned close to the center SQUID pick up coil. The SQUID is not used in
these measurements, but the center pick up coil also marks the center of the
superconducting solenoid.

With the sample in place, an EDC program is started. This program records the
measured voltage for an applied 0.1 mA current in zero field as the temperature is
increased in increments of 0.02K. The program starts at a temperature below T c and
continues increasing temperature until a sharp rise appears in the recorded voltage well
above the noise level (typically ~ 0.5uV) of the measurement. The onset of this sharp rise
characterizes T c of the sample. The measured 7'. is then put into the second EDC
program, which measures the critical current as a function of applied field and
temperature. This procedure helps maximize the efficiency of the program, which can

take up to 3 days to collect a complete set of critical currents. This second program starts

39

ll;— Sample Rod

.> Current Leads <
>Voltage Leads <

 

 

 

 

@@®®@@@@@®

 

7?]

p——.‘

l E Sample \@

Parallel Orientation Perpendicular Orientation

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 16 - Samples mounted for critical current measurements.

40
in zero field and measures the critical current in 0.2K steps from 5K up to T 0. Once the
first temperature sweep is done, the field is set to the first desired field setting and the
temperature sweep repeated. This time measurements are only made up to T C(H). The
next field is then set and the process repeated until T C(H) is less then 5K. At any given
temperature and field setting, the critical current is determined by increasing the applied
current until a specified voltage is reached, in this case 10 uV. The current is increased in
0.1 mA increments from 0.1 mA to 10 mA and in 1 mA increments after that if necessary.
The critical current is defined as the largest current which produced a measured voltage of
less then 0.5 uV. Figure 17 shows a typical current sweep with the critical current

indicated.

41

Nb/Cu 200A/100A
707-6

 

V (IN)

 

 

—
b
p.

 

 

0 l 2 3
I (mA)

Figure 17 - A typical current sweep, this one taken at 5.6K and 600G.

Chapter 5
THE MAGNETIZATION IRREVERSIBILITY LINE

Irreversibility has been seen in many superconductors, from the high- T c

”"930 to Niobium films". It can be defined simply as the separation of the zero

ceramics
field cooled (ZFC) and the field cooled (FC) magnetization curves. For the ZFC
magnetization curve, the sample is cooled to below T, in zero magnetic field and a
measuring field is applied to measure the magnetization as a function of increasing
temperature. For the FC magnetization curve, the sample is cooled in the measuring field
as the magnetization is measured. Figure 15 is an example. In the ZFC case, there are no
vortices in the sample before the measuring field is applied. As the field is applied, flux
enters the sample in the form of vortices which become pinned due to the presence of
defects or other pinning centers present in the sample. These pinning sites prevent more
vortices from entering the sample, and are responsible for the large diamagnetic signal of
the sample. In the FC case since the flux was already in the sample as it was cooled, the
sample is less diamagnetic. The same pinning sites now prevent vortices from being
expelled. This irreversibility will go away at a temperature 7),, < T c where the thermal
energy is sufficient to unpin the vortices.

In this chapter, the theory behind the field dependence of Tm, known as the

irreversibility line, 7),,(H), will be outlined. Experimental data will be used to address the

42

43
ongoing debate over whether this history—dependent property is associated with flux creep,
a glass transition, or vortex-lattice melting. The results of irreversibility measurements

performed on Nb/CuX multilayers in two sample orientations will be shown.

5.1 Theory

The three most significant theoretical models regarding the irreversibility line,
T,-,,(H) or HAT), involve flux creep, a transition fiom a vortex glass to a vortex liquid, or
a transition fi'om a vortex lattice to a vortex liquid. The flux creep model, presented by
Yeshum and Malozemofl‘ 18, says that the irreversibility line is the result of the thermally

activated depinning of the vortices and predicts:

 

87gf2BkT ln(BdQ/E ) 2”
— t = c c . Equation [1]

2.56Hc220<l>0§
Here I = T / T c is the reduced temperature, B the magnetic induction, d the average distance
between pinning centers, (2 the oscillation frequency of a flux line in a pinning well, E, the
electric field determined from the lowest measurable voltage divided by the sample length,
f the ratio of the penetration depth to the flux lattice spacing, (Do the flux quantum, and 6
the superconducting correlation length. For parameters typical of conventional
low-temperature superconductors, ln(BdQ/Ec) is about 30. Near T c, where the field B is

approximately H, this equation is of the form:

3/2
H,” (T) = H,”0 [I — 17,—.) Equation [2]

C

where H,"o is a constant.

The same 3/2 power law behavior is expected for the melting of a vortex glass.
This behavior is expected by analogy with the Almeida-Thouless21 line in spin-glasses,
describing the melting of a similar glass-like state. Both systems are fi'ustrated. The
spin-glass system is fiustrated by a random coupling constant. The frustration in the
vortex glass is due to the competition between the desire to form a triangular array of
vortices which minimizes the vortex-vortex interaction energy, and the randomness of the
pinning sites.

Houghton, Pelcovits, and Subdra22 have derived an expression for the vortex lattice

melting line of a uniaxially anisotropic material:

, J; 4(JE — 1) _ ,
filTb[—J—l————b— +1} _ a Equation [3]

where a, the degree of susceptibility of the vortices to thermal motion, is given by:

x' M

2

2 H 1/2
a = 2 x 105(3) [7,42% i] . Equation [4]

Here t= T/Tc, b =H/Hc2(T), Hc2(7) = Hc20[l - T/Tc], x; 11/; is the Ginzburg-Landau

parameter, the lengths x1 and 5 are the penetration depth and correlation length
respectively, and M/Mz is the ratio of the in-plane and out-of-plane effective masses. The

Lindemann criterion for melting is governed by the parameter c, which is the ratio of the

45
mean-square thermal displacement of the vortices from their equilibrium positions to the
vortex lattice spacing. For T near T c it can be shown that the melting curve is given by:

2
H,,,(T) = H,”0 [1— 77;) Equation [5]

C

which has the same functional form as the prediction made by the two previous models,
with a different exponent.

The transition at T." is believed to be highly sensitive to disorder, being either first
order or continuous depending on the number, type, and geometry of impurities.
Although there is a good grasp10 on how these various impurities affect T c, there is a lack
of theoretical predictions for their effects on Tm. For thick Cu interlayers, T c is reduced
from what it would be for an isolated Nb film because Cooper pairs leak into the
interlayers by the proximity effect. As the Cu layers get thinner the T c of the multilayer,
T cg”, goes up because some of the Cooper pairs pass all the way through the Cu layer and
enhance the pair density in adjacent Nb layers. T c can also be raised by adding Ge
impurities to the Cu. These don’t destroy the Cooper pairs, but they tend to reflect the
pairs back into the Nb layer from which they came, again enhancing the pair density in Nb.
The addition of Mn destroys most of the Cooper pairs that enter the interlayer and
dramatically reduces T cm. In Figure 18 is shown a schematic version of the three
dominant processes for a Cooper pair entering a normal metal interlayer.

A theoretical interest addressed in this study is the proposed23 3D to 2D transition

that occurs in high-Tc materials at high field with the applied field oriented perpendicular

46

 

 

o
o
1" . .
Transmrssron
. Spin-Scattering
/
o

/ . Reflection

o

o
o

S N S

Figure 18 - The three dominant processes for a Cooper pair entering a normal metal layer.

to the CuO planes. At high fields, as the density of vortices gets high, it is thought that
the vortex-vortex interaction within the layers overcomes the interaction between the
layers, causing the 3D line vortices to breakup into 2D pancake vortices. The high field
destroys or severely weakens the proximity induced superconductivity in the N layers. If
this happens, increased pinning, which would normally increase Tm, might not have much
of an effect, since a pinned pancake has no effect on pancakes in other layers, whereas a
pinned line vortex afl‘ects many layers. If this effect can be seen in the Nb/CuX
multilayers, it should be possible to control the field at which this transition takes place.
By adding impurities, either Ge or Mn, the interaction between vortices in adjacent layers

can be weakened causing the 3D line vortices to breakup into pancake vortices at lower

47
fields. There are however currently no theoretical predictions on how this transition will

affect the shape of T...(H) in the phase diagram.

Another theoretical question which arises is what to expect for the behavior of T...

when the multilayer sample is parallel to the field. It is known10 that in the parallel

orientation a 3D/2D/3D crossover can be seen in H :2(T ) as the temperature is lowered.

He; in bulk materials is reached when the area available per flux quantum approaches 27%,
twice the area of the normal region surrounding a vortex. Therefore H62 =(<1>o/2mfz)
where (Do is a flux quantum and 5 is the coherence length perpendicular to the field. Since
multilayers are not isotropic, when the field is applied parallel to the layers one must
consider two 5’s, a 5” parallel to the layers, and a :1 perpendicular to the layers. In these
multilayers where H62 0: 105...; the temperature dependence follows one of three trends.
Near 7', both 6’s diverge as 5 0c (1 - TIE-SN)“2 so 3D behavior is observed with
H52 cc (1 - T/Tcs~)1. As T decreases, 5. is limited by ds and 2D behavior is seen with
H52 0: (1 - T/Tm)”2 and finally at a still lower temperature where both 6’s are small

compared to d3, He; behaves as in the bulk S material:
T2._3<T<T..w ngocO-T/Tcsw)
T3“, < T < T2H HE, oc(1— 77730)” u E 1/2. Equation [6]

T < T.. H:. H::*(T)

Here T 293 is the 3D to 2D dimensional-crossover temperature, T 21) is the effective T c for

the 2D regime, T 293 is the crossover temperature from 2D to bulk behavior, and H31“ (T)

is He; for bulk S.

48

The H), (T) curve of a Nb 280A / CuGe(5%) 200A multilayer as determined from

resistance measurements is shown in Figure 191°. No theoretical work has been done to
determine how this dimensional crossover will affect the behavior of T,-,,.

Another controversy involves how best to experimentally determine the
irreversibility line”. One method is to use the field cooled (FC) and zero field cooled
(ZFC) dc magnetizations. Upon increasing the temperature, the joining of the ZFC and
FC curves (recorded at a field Ho) takes place at a temperature T 0 giving a point on the
irreversibility line, H.,,T(To) = Ho. The obvious experimental limit of this technique is the
noise in the data”, since T o is the temperature at which the difference in the ZFC and FC
magnetizations becomes greater then the noise. As the two magnetizations are smoothly
joining curves, the error bars in T o are important and difficult to determine experimentally.

Another experimental technique26 used to determine the irreversibility line is the
maximum in the out-of-phase component [(7) of the ac susceptibility at a temperature
T m, recorded at a low frequency and at some small ac field amplitude superimposed on a
large dc field Ho. Here Tm determines H,,,(T) by the relation H,,,(Tm.,.) = H0. Others27
have used the onset temperature of ['(T) instead of Tm in the above relation. The
absorption 1" is proportional to the ac magnetization hysteresis loop area, which is
non-zero in the presence of irreversibility. Therefore irreversibility should exist above
Tm. Based on the above arguments, the use of the ['(T) onset criteria should be more
consistent than the use of the [(7) maximum criteria. However, skin depth penetration

effects also give non-zero ['(T) values. All these complications produce frequency and ac

amplitude dependencies in ['(T) which shouldn’t be present if the irreversibility line

49

 

 

 

 

Cu CuGe 57.
z/aoi/zodi )
30000 _ v 1 . 1 . r r I I ' I T I ' 1 _
. Tao-8 ‘
25000— 1
20000 — -
Q
_% 15000- ~
:1: _
10000 ~ Tm ‘
5000 T l "
O 1 . 1 1 1. 4
0 1 2 3 4 9 10

 

Temperature (K)

Figure 19 - The 3D/2D/3D transition seen in H :2 (T) .

50
probes a critical phenomena like the melting of a vortex solid. Therefore the physical
meaning of the ac irreversibility line remains unresolved. While the dc technique lacks
accuracy, there is less dispute over the validity of the dc irreversibility line. There exist
other determinations of H.,,(T) using electric transport properties, but the presence of

Lorentz forces adds further complications to those of the magnetic studies.

5.2 Experimental Results

For this study the “dc” technique was used. The ac technique was tried but
produced irreproducible results and was abandoned. Transport measurements were also
made to look for H,,,(7) but with the available system only Hc2(T) could be seen. Below
H.2(7) the measured resistance of the sample dropped below the inherent noise of the
system (about 10'6 ohms) and no irreversibility could be seen. One concern that can be
raised about measuring the dc magnetization in a Quantum Design SPMS is the motion of
the sample during the measurement. Since the MPMS requires the sample to move
through a non-uniform field, it can be thought of as imposing a small amplitude ac field on
the measurement with a period of the measurement time (about 2 minutes).

Figure 15 is a graph of the ZFC and FC magnetization curves of a
Nb 280A/Cu 100A sample in an applied perpendicular field of 250G. Figure 20 is an
illustration of the difference between successive data points on the FC magnetization

curve, which shows the characteristic sharp rise at T c. Figure 21 illustrates the difference

51

 

 

 

 

 

 

 

Nb/Cu 200A/100A
Field = 2500
1E-4 F E '1
’5 _ _ .
8 BE 5 I E
8 0E—5 - —— -
:1
d
.N __
*3 4E-5 ” A I 4
T. = 7.725K .
so . j
=1 212—5 ~ /
C $ .1...
OEOmIIII;-s-— db T—t—fi—n—a—i—H

 

8.8 7.0 7.2 7.4 7.8 7.8 8.0
Temp (K)

Figure 20 - The differences between successive data points on a PC magnetization curve
graphed to illustrate the sharp rise at T ...

52

 

 

 

 

 

 

Nb/Cu 280A/100A
Field = 250G
A 4» I
a 0.0000 I . i W
a T
3 1
a -0.0002 ~ E —
O
*3 1
5“ —0.0004 r I T," = 7,45 ~
0
g —0.0006 - E -+
g 1
. 1
Q -0.0008 - —
‘é‘ 1
_0.0010 1 41- 1 1 1 1 1 L 1 1 1 1
8.8 7.0 7.2 7.4 7.6 7.8 8.0

Temp (K)

Figure 21 - The difference between F C and ZF C data points measured at the same
temperature graphed to illustrate the sharp drop at T1”.

53
between FC and ZFC data points measured at the same temperature versus measuring

temperature. Here the characteristic sharp drop is seen at T,-,,.

The T C(H) and T...(H) data for all samples were then fit with:

T
H T 2H l— c
02( ) €20[ T

‘0

 

P. P."
J and Hm”) = H1”,[1‘ 37:?) Equation [7]

co

with H

020 7

Pa. H -

mo ’

P1,, and TCo as fitting parameters. Both data sets were fit

simultaneously since T was a parameter in both equations.

H.2(T) was also determined from transport measurements for a few samples.
Good agreement was found between the transport and magnetically determined

parameters. See Figure 22 for a graph of both Hf,(T) curves and the H,fi,(T) curve of

the Nb 280A/ Cu 100A sample.
5.2.1 Perpendicular

All samples measured in the perpendicular orientation are listed in Table 3.
Table 4 contains the previously mentioned fitting parameters obtained fi'om the data for
these samples. Figure 23 through Figure 25 show the H-T phase diagrams for several of

these samples. Figure 26(a) is a graph of cho and H .,,o versus the normal metal layer

thickness, d”, for the Nb/Cu multilayers. The dN = 0 point was obtained fiom a
measurement on the 5000A film. Figure 26(b) is a similar graph for the PC and P...

parameters. Figures 27 through 29 are similar to Figure 26 except they display the

54

Nb /Cu
zeal/100A

 

Him“ = 25000:1: 1200C "
Pém‘“ = 1.05:t0.02

   
  
  
 
    

3000

His. = 25000i 1300G

8
'U
5 L
In
1000 r _

Ht", = 2750013000
Pt, = 1.50:1:0.02

 

o . 1 . 1 1 1 1 1 1 1 1 1 1 1 1 ~ ':
6.0 6.2 6.4 6.6 6.8 7.0 7.2 7.4 7.6 7.8 8.0
Temp(K) '1'... = 7.83et0.01K

 

Figure 22 - H $(T ) determined from magnetization measurements (A) and transport
measurements (V) along with H;(T) determined magnetically (O).

55

Table 3 - All samples measured with H perpendicular to the multilayers.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Number Sample Description

456-16 5000A Nb Film

360-7 Nb 280A / Cu 2011

360-9 Nb 280A / Cu 4011

360-11 Nb 280A / Cu 1004

360-13 Nb 280A / Cu 200A

440-7 Nb 280A / Cu 400A

360-12 Nb 280A / Cu 700A

622-3 Nb 280A / CuGe(2%) 2011
6224 Nb 280A / CuGe(2%) 40A
622-5 Nb 280A / CuGe(2%) 60A
6226 Nb 280A / CuGe(2%) 80A
622-7 Nb 280A / CuGe(2%) 100A
6228 Nb 280A / CuGe(2%) 150A
622-9 Nb 280A / CuGe(2%) 200A
622-10 Nb 280A / CuGe(2%) 30011
62241 Nb 280A/ CuGe(2%) 400A
382-2 Nb 280A / CuGe(S%) 2011
382-4 Nb 280A / CuGe(5%) 40A
382-8 Nb 280A / CuGe(5%) 100A
382-10 Nb 280A / CuGe(5%) 200A
389-9 Nb 280A / CuGe(5%) 400A
382-7 Nb 280A / CuGe(5%) 7004
344-10 Nb 280A / CuMn(O.3%) 15A
412-9 Nb 280A / CuMn(O.3%) 20A
344-9 Nb 280A / CuMn(O.3%) 30A
344-7 Nb 280A / CuMn(O.3%) 60A
344-6 Nb 280A / CuMn(O.3%) 100A
344-4 Nb 280A / CuMn(O.3%) 400A

 

 

56

Table 4 - Fitting parameters for all samples measured in the perpendicular orientation.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample H3. (G) P: H.:, (G) P:- T; (K)
Number

456-16 24100:}:2800 1.00:|:0.06 255001 1700 1.35i004 8.93:1:002
360-7 28400125000 0.99:1:007 31100:}: 3400 1.27:1:006 8.25zt0.02
360—9 254001-2600 1.00i0.05 39900:t 3100 1.44:1:004 8.26:1:001
360-11 25000i1200 1.053002 27500:h 800 1.50i002 7.83:1:001
360-13 17300i1600 1.09i0.06 23700:}: 1800 1.70i005 7.37:t0.02
440-7 13700:):1100 1.14.2006 16100:): 1300 1.463006 6.64i002
360—12 7400i 500 0.92:1:005 7100i: 400 1.19:1:004 6.002h0.02
622-3 2080028200 0.90:1:007 413001: 7800 1.48i0.08 8.39:1:002
622-4 31900:l:2600 1.1133004 295003: 2100 1.38:1:004 8.32:t0.01
622—5 25100i3700 1.03:h0.07 19900:t 1900 1.23fl.05 8.172t0.02
622-6 2240028000 0.99:1:007 3590021: 3500 1.50i005 8.03zt0.02
622-7 27100:t6600 1.13:1:0. 12 55400:|:10600 2.05:1:0. 12 8.033004
622-8 2260012800 1.11:1:007 49600i 4900 1.93:1:007 7.831002
622-9 2030028100 1.07:1:008 249001: 2500 1.51:t0.06 7.71:1:002
622-10 18500:t1300 1.14:1:004 23700i 2100 1.57i005 7.41i002
622-11 1670012400 1.08i008 21500i 2400 1.59i0.08 7333:003
382—2 21600i3400 0.95:1:007 124000:l:24200 2.05:l:0. 10 8431-002
382-4 2230028100 0.97:1:006 54600i 5900 1.80i006 8.27:1:002
382-8 1970011900 0.962005 639001: 5300 2.00:1:005 7.91:1:001
382-10 22600i3700 1.09i009 19100i 1200 1.692t005 7.472003
389-9 15400i1700 0.98:1:006 275001: 2000 1.87i005 7063:002
382-7 15300i2500 0.94:1:009 2040021: 1800 l.80:‘:0.07 6.89i0.03
344- 10 29200:):6100 1.06:1:008 56500:):10200 l 663:0.08 7.93i001
412-9 2240015700 0.93:1:0. 12 88000i20200 2.01:t0.13 7.73i0.03
344-9 2540043200 1.08i006 63500i 6500 1.90i006 7.51:);002
344-7 22900:}:5600 1.32:120. l3 65300115200 2.23zt0.15 6.98i0.04
344-6 13200i1300 0.98zh006 133005- 1400 1.42i0.07 6.10i0.02
344-4 4900i2000 0.86i0.17 llOOOi 3200 1.82:1:O.17 4.52i002

 

 

S7

 

Nb/CuGe(27.)
26011/15011
3000 l l r 1 1 l ‘1 l r r 1 r T r . I r 1 r
11:... = 22600i28006
Pi = 1.113007
2000 ~ -
£5.
2 .
.0.)
r:
1000 — _.
at... = 496001149006
131.. = 1.934007 K. _ -1,

 

 

 

6.0 6.2 6.4 6.6 6.8 7.0 7.2 7.4- 7.6 7.8
Temp (K) To, = 7.83:1:002K

Figure 23 - The H T phase diagram for the Nb 280A/ CuGe(2%) 150A sample.

58

Nb Film
50004

 

   
  
 
   

 

3000 1 1 1 1 . 1 . 1 . 1
Hip‘ = 24100i2800G
Pi = 1.00:t0.06
2000 r ‘
g
E r
.91
h l
1000 “
Hf”. = 25500i1700G
Pt..- = l.35:l:0.04

 

 

o 1 l A l 1 l 1 l 1 l 1 l 1 l 1 l
7.0 7.2 7.4 7.6 7.8 8.0 8.2 8.4 8.6 8.8 9.0
Temp (K) T. = 8.9340.02K

Figure 24 - The H T phase diagram for the 5000A Nb film.

59

 

 

 

 

Nb/CuGe(57.)
ZBOA/IOOA
3000 l V r 1 l l I ' T 1 l ' l ' I
Hi... = 19700419000
P: = 0.964005 1
2000 1 _
9’: .
Id 1
.9
r:
1000 - -
110.. = 65900453000 ~. ‘
p1,. = 2.004005 .3 -. a
O 1 1 1 1 1 1 1 #1 L 1 1 1 1 1 "53—... E;—
60 6.2 6.4 6.6 6.6 7.0 7.2 7.4 7.6 7.8 8.0
Temp (K) T... = 7.91:1:0.01K

Figure 25 - The H T phase diagram for the Nb 280A / CuGe(5%) 100A sample.

6O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nb/Cu
50000....-..,.T..-r
(a)
40000- E o H...
‘ H02
A 30000— I
u = E
v D I
18 t 1
20000— I
10000 ~
i
0 1 l 1 l 1 l 1 l 1 l 1 l 1 l
0 100 200 300 400 500 600 700
2.4 ' I ' I ' I v 1 r I v I . fl
2.2L (b)
. . P
2.o~ ‘ Pi:
1.8~ E
1.61
01 » I
1.4;- 5 E
1.2 [E I E
l
1.0 E; I I
l I
0.8
l
0.6

 

 

 

0 100 200 300 400 500 600 700
419(4)

IIr'rro ’

Figure 26 - H

€20 ’

P. and P1,, versus die for the Nb/Cu samples.

Nb/Cu0e(27.)
70000 1 r 1 1
h w a
60000 7 H ( ) 7
O
1 9 ‘ Id:

50000 1 1 —
E; 40000 - E -
no 30000 ~ g ( 4

’ f
.....1 11 1 i 1 .
10000 1 _
o l 1 l 1 l l
100 200 300 400
204 f I ' I f l
i
2.2 ~ (b) 1
1- . P 4
2'0 h E E A Pin 7
1.8 — -
1.6 - ..
0. r E E E E
1.4 — E 1
1.2 ~ 5 j
_ 1 1 1 .
1 0 f T I E E
o L I I
0.8 r -
0.6 4 l 1 l 1 l 1_ l
0 100 200 300 400
(111(4)

Figure 27 - H

61

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

020 2

H.

mo 2

PC and P1,, versus dN for the Nb/CuGe(2%) samples.

62

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nb/CuGe(57.)
1r— I I I I I If I I f
140000 — (8)
120000 1“ . H1...
. ‘ ch
100000
Q 00000 ~
m0 .
60000 ~ 1 E
40000 —
' I
20000 111 x I I If
0 1 l 1 l 1 l 1 l 1 l 1 l 1 l
0 100 200 300 400 500 600 700
204’ ' I ' I ‘ T ' I I I I I ' I
2.2 — (b)
. . P
2.0 E E ‘ Pin
11 1 1 1 i i
1.6 -
Q,‘ 1
1.4 ~
1
1.2 —
1 1
*r-T — T T
0.0 —
0.6 1 J 1 l 1 l 1 l 1 l 1 l 1 l
0 100 200 300 400 500 600 700
(111“)
Figure 28 - cho , H,”0 , PC and P1,, versus aim for the Nb/CuGe(5%) samples.

63

Nb /CuMn(O.37. )

 

110000 __ 1 1 1 1 1 1 1 1
100000 - (a)
90000; 1 . H11.
00000 L
70000 1
00000 L
50000 L
40000 L
30000 11$

20000 L I; I

10000 ~

 

 

 

 

 

Ho (G)

 

fl
~l>'|l-O‘l

 

 

o 1 L 1 l 1 1
0 100 200 300 400

 

204 ' I ' r ' I v I

21: (b)

 

I ' I '

24)

 

 

 

1.0;
1.6 11
114 1 1 §[

113-

1 1 1 l J_

 

1111’H
1

013*

H4
l——->—-Hl

 

 

 

l
0.6 l 1 l 1 l 1
0 100 200 300 400
dN(A)

Figure 29 - H H . PC and P1,, versus 6111 for the Nb/CuMn(O.3%) samples.

c2o ’ 1'70 9

64
parameters for the Nb/CuGe(2%), the Nb/CuGe(5%), and the Nb/CuMn(O.3%) samples
respectively.

H€20 and HMO decrease with increasing normal metal layer thickness, due to the

decrease of T

Co with increasing (1”. Figure 30 illustrates this dependence of 7;) on dN for
all sample types.

From these graphs one can conclude that PC is a constant and equal to one within
the noise of the data. This is the expected result for H120). Although there is
significantly more scatter in the P1,, data, it can be argued that for each interlayer type P,”
is a constant.

In an attempt to investigate the proposed 2D to 3D transition the Nb/CuGe(x%)

parameters were also graphed as a function of Ge concentration. P1, P1”, and 72.0 are

graphed as a function of Ge concentration for the Nb 280A / CuGe(x%) 20A samples in
Figure 31. These samples should represent strongly proximity coupled S layers. For all

samples PC and T are constant but P1,, shows an interesting trend. P1,, is less then 1.5 for

the 0% and 2% samples but jumps to over 2.0 for the 5% sample. This may indicate the
formation of 2D pancake vortices in the Nb layers, the 3D line vortices being destroyed by
the added impurities in the N layers. The impurities severely weaken the
superconductivity in the N layers, thereby destroying the coupling between S layers. A
similar graph for the Nb 280A / CuGe(x%) 100A samples, Figure 32, where the S layers
are now moderately proximity coupled is consistent with this picture. Here the jump in
P1,, is now occurring between 0% and 2%, it being easier to destroy the 3D line vortices in

the thicker N layers. To test this hypothesis the 5% sample was irradiated at the

65

Nb/CuX

 

I ' I

X = Pure Cu
X = Ge 22

X = Ge 52

X = Mn(0.32)

To. (K)

 

 

1

 

 

1 g

 

0 100 200 300 400 500 600 700
dn (A)

Figure 30 - 720 versus dN for all sample types.

66

 

 

 

 

 

 

 

 

 

 

Nb/CuGe x) ‘ P.
2001/20 0 P111
I To.
2.1 ~ ‘ 9
1.9 -
I I
1.7 '
. ~ 0
1.5 1 A
DE 1 :4
a ’- V
a: 1.0 . I:
1.1 — - 7
0 0 - 1 I 1
L
0.7 1
‘ - 0
0.5 l 1 l 1 l
02 22 51

Figure 31 - P1, P1,, and 72° graphed as a function of Ge concentration for the
Nb 280A / CuGe(x%) 20A samples.

67

 

Nb/CuGe x) ‘ Po
2001/100 : $1..

 

 

 

 

 

 

 

 

 

 

S A
a: 1.3 ~ l a
n. » 9‘

1.1 — I I _ T — 7

0.9 r i i

+ .
0.7 1
0.5 1 1 1 1 1 1 1 I 6
oz 22 52 Irrad 52

Figure 32 - PC, P1,, and 72.0 graphed as a function of Ge concentration for the
Nb 280A / CuGe(x%) 100A sammes.

68
MSU-NSCL cyclotron to create columnar defects in the sample and see if the 3D line

vortices could be restored. A total fluence of 5x1011 Kr+13

nuclei, each having an energy
of 1.64 GeV, irradiated a sample area of about 0.25 cmz. As seen in Figure 32, P1,, was
indeed restored to a value close to 1.5.

One would expect that as dN continued to increase the jump in P,,, from 1.5 to 2
would continue to occur at lower and lower concentrations. This appears not to be the
case as can be seen in Figure 27. Here it is noted that although P1,, is near 2 for the two
samples with dN values of 100A and 150A, the samples with dN values of 200A, 300A, and
400A all have P,,, values near 1.5. However there are some concerns about the quality of
the data collected for these three samples. The 200A and 400A samples showed
uncharacteristic paramagnetic bumps in both the ZFC and FC magnetization curves just
below T 1. Also the irreversibility between the ZFC and FC magnetization curves
unexpectedly disappeared near 2500G for all three samples. This behavior was not seen in

any of the other samples. Due to these concerns the P,-,, values reported for these three

samples may be misleading.

5.2.2 Parallel

All the Nb/Cu samples were measured in the parallel orientation. Figure 33 shows
the H T phase diagram for the Nb 280A/ Cu 400A sample. This was the only sample that
showed a clear irreversibility in the FC and ZFC magnetization curves. For the other

samples the data was just too noisy to extract meaningful T1,, values. Even this parallel

Field (G)

69

 

 

 

 

Nb/Cu
2001/4004
3000 1 1 1 1 1 1 1 1 1
11'... = 1070040000
1 P: = 14040.04
2000 ~ _
1000 — 1 1
m... = 1710047000 1 1 . .
P1,... = 1.474003 "59:.
O l 1 l 1 l 1 "'T-p—.
4.5 5.0 5.5 0.0 6.5 7.0
Temp (K) To, = 6.74:1:0.01K

Figure 33 - The H T phase diagram for the Nb 280A / Cu 400A sample with the field

applied parallel to the layers.

70
data proved to be difficult to reproduce because small misalignments produced very

difi‘erent results.

5.3 Conclusions

In the perpendicular orientation one sees an interesting trend in H ”a and H620.

H . and H620 decrease with increasing normal metal layer thickness which is expected

I ”0

due to the similar behavior of Tc,- It is surprising however that H "o and cho are

essentially equal to each other in all samples as there are no theoretical predictions that
this should be the case. The one case where this breaks down is for small (111 in the
Nb/CuGe(5%) and Nb/CuMn(O.3%) samples. For dN s 100A in the Nb/CuMn(O.3%)

samples and dN < 100A in the Nb/CuGe(5%) samples, H,"o is about a factor of 3 greater
than H020. Since these are the two samples types where the coupling between S layers
should be the weakest, this may be an indication of the 3D to 2D transition. The increase
in H,”o is consistent with the increased pinning strength expected with the 3D line
vortices, however since both Hm0 and He20 are extrapolated back to zero temperature

from data taken at temperatures near Tc, , the actual low temperature values of H," and

H02 are sure to be quite different than those found. This is evident form the fact that the

actual irreversibility line can never cross H120). Therefore the extrapolated values H,”0

and H620 are dependent on the temperature range over which data was taken, which

71

makes the comparison between different samples rather difficult. The corresponding
values of P1,, and PC are essentially constant for each sample type. PC is roughly 1 for all
samples as indicated by the solid line on the graphs, which is the expected value. P1,, is
approximately 1.5 for the Nb/Cu and the Nb/CuGe(2%) samples which agrees with the
value predicted by the flux creep model or the vortex glass model. P1,, is approximately 2
in the Nb/CuGe(5%) samples which agrees with the value predicted by the vortex lattice
melting model. However considering the scatter present in the data, no firm conclusions
can be drawn from these two observations. There is so much scatter in P,-,, for the
Nb/CuMn(O.3%) samples that it would be hard to pick between 2 and 1.5, but all values
lie between 1.3 and 2.3. The strongest indication of the 3D/2D transition can be seen in
the Nb/CuGe multilayers as a fiJnction of Ge concentration. The value of P1,, is 1.5 in a
region where 3D line vortices are expected and the value jumps to 2 in a manner
consistent with the picture of the formation of 2D pancake vortices.

In the parallel orientation, reproducible measurements could not be made with the
available WMS. The difficulty of measuring a consistent and reproducible irreversibility
line in all orientations has been encountered by others”, most of whom conclude that the

study of transport properties has fewer inherent problems and is potentially more usefiil.

Chapter 6
CRITICAL CURRENT AND VORTEX PINNING

Critical current density and vortex pinning are two closely related properties of
superconductors. The study of these properties has intensified recently due to the interest
in making practical high-current applications of these materials. The fastest growing area
of this type of research is the study of the highly anisotropic high-Tc ceramics, but many of
the same physical processes take place in the low-Tc materials. With the greater control
one has over the various parameters in low-Tc S/N multilayers (d3, dN, S, N, etc.) it should
be easier to study these physical processes in the low-Tc materials and gain insight which
should be transferable to the high- T. superconductors. The critical current density, Jc, is
defined as the current density at which an arbitrarily small voltage is observed. The vortex
pinning force density, F p, is equal to the critical value of the Lorentz force chB per unit
volume.

In this chapter, the theory of vortex pinning will be outlined. The results of critical
current measurements on numerous Nb/CuX multilayers in two sample orientations will be
shown. The agreement between theory and experiment will be discussed and conclusions

presented.

72

73

6.1 Theory

The pinning of vortices in the vortex state by various imperfections in type-II
superconductors is responsible for the existence of a critical current density, .11, defined as
the current density at which an arbitrarily small voltage is observed. The critical value of

the Lorentz force per unit volume, F p, is determined from Jc.

Fp = J c x B Equation [8]

In the perpendicular orientation of an anisotropic superconductor there are two possible
vortex states — 3D vortex lines and 2D vortex pancakes. As the vortex-vortex interaction
within the layers overcomes the interaction between the layers, the 3D line vortices may
break up into 2D pancakes. Severely reduced pinning results since pinning in one layer
now has little or no effect on other layers.

F rietz and Webb29 were the first to note that the pinning force density appeared to

obey a scaling law of the form
2.5 .
F, = [ch (7‘)] f(h), Equation [9]

where flh) is a function only of the reduced magnetic field h = H/ch. The importance of
such a scaling law is that one could measure Fp at one temperature to get f(h) and then

simply scale the results by [He2 (T )]2'5 to predict Fp at any other temperatures. Though

this first work was done on a number of Nb-based alloys, similar scaling laws of this form

Fp = [Hc2(T)]xf(h)1 Equation [10]

74

30,31 -
, wrth

have been shown to apply to a wide variety of other superconducting systems
typical values of x being between 2 and 3.

The qualitative approach to understanding the features of Fp(h,T) can be
reproduced by a simple model presented by Kramer”. His model has the four following
properties. First, at low reduced fields Fp(h,T) is given by a fimction F1(h,T). Where F1 is
a pinning force computed on the assumption that all pinning interactions are broken,
meaning that the maximum pinning forces have been exceeded upon reaching the critical
Lorentz force. Second, at high reduced fields Fp(h,7) is given by a function F2(h,T).
Where F2 is a pinning force computed on the assumption that strong pins do not break but
rather the vortex state shears plastically around them. Third, the peak in F ,,(h,T) will be
reached when F1(h,7) z F2(h,T). Since F2 is only weakly dependent on pinning strength,
increasing the pinning strength and thus increasing F1, will shift the peak in Fp(h,7) to

lower h and increase the peak value of Fp(h, 7), while leaving the high h values of F p01,?)

essentially unchanged. Fourth, the scaling law requires that both F1 and F2 depend on
temperature as H :2(T ) .

A very interesting model that is still lacking a theoretical description is proposed
by Koorevaar, Maj, Kes and Aarts”. The model they proposed was to explain the
nonmonotonic behavior of J. seen in Nb/NbZr multilayers as a function of field applied
parallel to the multilayer. For their model they present the following simple phase
diagram. For all T < T c a Meissner (M) phase exists at very low fields. In fields just above
the Meissner phase there is an anisotropic Abrikosov (AA) region where the vortices are

straight, as in an ordinary anisotropic superconductor. The modulation of the order

75
parameter is weak in this region. In fields above the AA region there is a kinked (K)
region where the vortices now consist of disks with cores perpendicular to the layers and
strings with cores along the layers, Figure 34. The field where a peak in J. occurs, which
they call Hp, is proposed to separate the AA and K regions. In the regime above Hp, Jc is
detemiined from the motion of the disks along the layers and not the motion of the strings
perpendicular to the layers as the intrinsic pinning related to the layered structure is too
strong. Below Hp movement of the vortices normal to the layers is possible. One may
expect that in the ideal case with the field applied parallel to the multilayers no vortices
perpendicular to the planes (or disks) would be present. However, besides the fact that
the sample can always have a small misalignment due to mounting, more important is the
fact that the plane of the layers can vary in multilayers not grown epitaxially, making it

impossible to align the sample exactly with the field.

Vortex
String

Vortex Disk

 

Figure 34 - An illustration of a kinked vortex region.

76

6.2 Experimental Results

For this study patterned samples had to be used as the large currents needed for
the unpattemed samples produced heating in the contacts. Table 5 contains a list of all
measured samples and their measurement orientation. Chapter 4 describes the actual
procedures used to make the measurements. Once all the I-V curves for the various
temperatures and fields are collected from the MPMS, the data is analyzed using a
program that extracts all the critical currents. Also the H.2(T) curve is generated by

extrapolating to the point where the critical current goes to zero. The 1112(7) curve is fit

with the same functional form used in Chapter 5.

Table 5 - All samples measured in this study.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample Number Sample Description Measurement Orientation
707-2 3000A Nb Film 1 and I I
782-1 280A Nb Film I I
707-6 Nb 280A / Cu 100A 1 and II
739-2 Nb 280A / Cu 50A 1 and I1
. 7390 Nb 280A / Cu 25A 1
770-5 Nb 280A / Cu 25A ||
707-7 Nb 280A / CuGe(2%) 100A 1 and I I
770-6 Nb 280A / CuGe(2%) 50A ||
739-5 Nb 280A / CuGe(2%) 25A 1 and II
723-5 Nb 280A/CuGe(10%) 100A 1
770-7 Nb 280A / CuGe(lO%) 100A | I
770-8 Nb 230A / CuGe(lO%) 50A I I
739-7 Nb 280A / CuGe(lO%) 25A 1 and II
707-8 Nb 280A / CuMn(O.3%) 100A 1 and II
771-5 Nb 280A/ CuMn(O.3%) 50.4 H
771-6 Nb 280A / CuMn(O.3%) 25A II

 

 

77

6.2.1 Perpendicular

In Figure 35 is shown a superposition of the I-V curves of the Nb/Cu 280A/100A
sample collected at an applied field of 600G and a series of temperatures between S and
7K in 0.2K increments. As can be seen in the figure all the noise in the superconducting
regime is below the 0.5 11V level. Therefore this value was chosen for the determination
of 1.. Figure 36 shows the H120) curve extracted for this sample and Figure 37 displays
the critical current versus the reduced field h at various temperatures. The top curve is
taken at SK, the next curve at 5.2K, and so on in 0.2K increments. Note all curves are
monotonically increasing with decreasing field with a sharp rise at low field where the
sample enters the Meissner state. Figure 38 shows the pinning force density versus the
reduced field h. Again the top curve is taken at 5K, the next curve at 5.2K, and so on in
0.2K increments.

If one compares the 72° of the patterned film shown in Figure 36 with that of a
similar non-pattemed film shown in Figure 22, it is noted that Tc, of the patterned is

significantly lower. This is true for all patterned films. This difference is probably due to
the patterned film not being as clean as the non-pattemed film because carbon impurities
could be present in the patterned film due to the photoresist mask used during the
sputtering process. Since carbon is atomic number 6, these impurities would not show up

in the EDS analysis.

78

Nb /Cu 200A/100A
707—6

 

 

 

 

 

 

Figure 35 - Several of the I-V curves collected for the Nb/Cu 280A/100A sample. The
right most curve is taken at 5K with each successive curve taken at higher
temperatures in 0.2K increments.

79

Nb/Cu 200A/100A n=7

 

  
   
 
 

 

 

 

707-8
8000 7
1 H... = 26800117006

’6 7000 — 7.: 7.491—0.04K
v - p. = 1.091005
3 0000 —
.2 ~
In
1. 5000 —
a 1
.2 4000 r
g .
o 3000,
E 2000 — P
n. . HQKT)=ILn&1"T/Tq)'

1000 -

O l l J A l
5 6 7
Temp (K)

Figure 36 - The H120) curve for the Nb/Cu 280N100A sample.

80

Nb/Cu 2001/1005. 11:?
707—6

 

30»- _

20- ~

1. (mA)

 

10

 

 

Figure 37 - The critical current versus reduced field h at various temperatures for the
Nb/Cu 280A/100A sample.

81

Nb/Cu 2001/1001 n=7
707—6

 

 

 

 

10 L l l l l l l l l l
9 ~ _
0 L -
... 7 ~ _
ma _ _
\ 6 .
1.2 5 ~ 1
o . .
5. 4" ‘1
In 3 ~ ~
2 — _
1 — -
.Z” 0
0 11—15: W 1 . 1 .
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.0 0.9 1.0

h (Ii/11320))

Figure 38 - The pinning force density versus reduced field h at various temperatures for
the Nb/Cu 280A/100A sample.

82
As predicted, the pinning force data scales in agreement with the model presented

by Kramer”. To check this, each Fp curve was fit with a fiinction of the form:

 

 

F p = prefactorK m; n) 11]”I [( m;- n) (l — h)]n Equation [1 1]

where m, n, and the prefactor are the fitting parameters. All curves were fit
simultaneously with the same m and n values. This function was chosen arbitrarily as it
had approximately the same shape as the data. It went to zero at h = O and 1 as needed
and has a peak between 11 = 0 and l. The only really important parameter in the above fit
is the prefactor. The advantage of using this method to extract a scaling prefactor as
opposed to extracting only the maximum of each Fp curve is the fact that the whole curve
is being considered rather than just the peak value of the curve. Figure 39 is a graph of
the scaled data for the Nb/Cu 280A] 100A sample. The scaling prefactors were then fit to

the form

TV

prefactor = const[l -— — Equation [12]

Tc.)

where const and P’ were fitting parameters. Tao was fixed to the same value found in the

determination of H.2(7). Figure 40 is a graph of the prefactors for this same sample. To
determine the value of x in Equation [10] you need H02(T) = He20 (l — T / 71.0 )P‘ along with

the above equation. This produces the following relationship

x = — Equation [13].

83

 

Nb/Cu 200A/100A n=7
707-0

1.1 L- r T ' l ‘ l ' l ' 1+ ' +1 I l l d

1.0 - ""‘Mfi .
fl ’ ‘.o .aov‘ ‘
a 0.9 r ‘ v “
E 0.0 - ii" ‘ :23 ‘
'3 0.7 1 " 'a —
T: 0.6 T .020 ...
g 0.5 h I, €01 + o ‘
'H _ . _
E 0'4 . + . ° m = 1244:0022
\" 0-3 j . 3 n = 120010.025 o. “
h 0.2 i °O? .0 4

 

0.1 F, = Prefactor{(m+n)h/m}"{(m+n)(1 -h)/n}" ".

L A_

-4

 

 

0.0 * ‘ ‘
0.0 0.1 0.2

h (H/HMT»

Figure 39 - Scaling of Fp(h) for the Nb/Cu 280N100A sample.

++<<>Obo¢0l>o
51:1999995’9'5’9'9'
fiODOhNOOO-fihfic

0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

NRNNNNNKNNN

84

Nb/Cu 200A/100A n=7

 

  
   

 

 

 

707-6
101.......,........ - . - , - T - T .
I f3
o I Prefactor = Const(1 - T/TJ" Z
a F‘
4 E
3 E
u 2 T
H .
o
a ,
3 10° :
“It -
3 . :
n. 5 _
4 z
3 i: Const= 174¢2 ;
2 7 T.,= 7.49 (Fixed) ~
i ' = 2.603.001 ‘
10-1 - i - . i ..... 1 i J - . J A
0 7 a 0 10-1 2 3

1 — T/Tc,

Figure 40 - Prefactors as a function of reduced temperature for the Nb/Cu 280A/ 100A
sample.

85
All the samples measured with H perpendicular to the layers showed good scaling.
Figures 41 and 42 illustrate the scaling of E. for several other samples. Figure 43 is a
graph of x versus dN for all sample types. All multilayers have x values that fall in the
expected range, between 2 and 3, with most clustered around 2.5.

The best evidence of the presence for a transition from 3D vortex lines to 2D
vortex pancakes is seen when we graph Fp versus )1 for several samples of the same type at
a fixed reduced temperature and compare them with what is seen in a thick (bulk) Nb film
prepared in the same manner. Figures 44 and 45 are such graphs for the Nb/CuGe(2%)
and Nb/CuGe( 10%) samples respectively. In both figures it is seen that the (1.; = 25A
samples have stronger pinning than in the bulk Nb sample where we know there are 3D
line vortices. This makes sense if there are still 3D line vortices and the Ge serves as good
pinning sites. However in the (IN = 10011 samples the pinning is greatly reduced indicating
that the sample is now in the 2D vortex pancake state where the vortices in one layer are
decoupled from the surrounding layers, meaning pinning in one layer has little to no efi‘ect
in the other layers. Figure 46 is a similar graph for the Nb/Cu samples. Here we see a
monotonic decrease of pinning strength with dN. This could indicate that the 3D/2D
crossover occurs at a dN value less then 25A or more likely that the pure Cu just doesn’t

serve as an effective pinner.

1.1
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1

F,/Prefactor (1 0" N/m’)

86

Nb/CuGe( 107.) 200A/25A n=7
739-7

 

 

m = 1.422i0.019
' ,. n = 1.274zt0.018

F, = Prefactor{(m+n)h/m}"{(m+n)( l -h)/n}III

l x

0.0 ‘
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.3 0.9

h (Ii/H3010)

Figure 41 - Scaling of Fp(h) for the Nb/CuGe(lO%) 280N25A sample.

ixu++¢oo>o¢00>o

Pfifififl$9999999999

A
V

v 1
U
A?

1.0

 

OOIF

Snunfifififlfififiuunfifi

OOIF

87

Nb/Cuce(2Z) 2001/2511 n=7
739-5

 

1.1 I I I I I I I I I I I I I I I I I I
1.0 '
0.0 ’
0.0 h
0.7 ". ‘
0.0
05 _x i

004' 0‘
" = 1.248:l:0.016 .
0.3 m

'x "' = _ , (S
0.2 My1,. n 1 14710 010 ‘1,
0.1 '

F,/Prefactor (1 0' N / m’)
3
Q

l

 

F, = Prefactorflm+n)h/m}"{(m+n)(1-h)/n}‘ ”0’, -.

l++<OODOQOCDO

Pssssspeeeeppppm

1 X

 

 

0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.7 0.0 0.9
h(H/H.+a(T))

Figure 42 - Scaling of Fp(h) for the Nb/CuGe(2%) BOA/25.3. sample.

1.0

fifififififififififififiafififi

 

88

Scaling Exponent 3: vs. (1.. Nb [CuGe 22
Nb/Cuceiio

F’=f(h){n"(T)}z Nb/CuMn(0. 3 )
20
2.7; L
2.6; L
2.5; L
2.4;

CID.

 

2.2 -_ -
2.1 ~ 1
2.. L :
... : i :

1.8 l J 1 . l A 1 1 l L 1 1 1 1 1 1 l 1 I . l -
0 10 20 30 40 50 60 70 80 90 100

dx (A)

 

 

 

Figure 43 - A graph of x versus (114 for all sample types.

89

. 707— 7 Nb/CuGezz again/10011 n=7 '1'... =7. 4910. 04K H...=2. 00:10. 1M

 

 

 

 

 

I 739-5 Nb/CuGez 280 “/25 11:7 8. 16:10. 03K 3. 06:20. 071‘
O 707- 2 Nb Film 3000A 8.60:1:0.04K 3.06:1:0.07T
40 _ Reduced Temperature: _
t = 036710.007
SE 30 - -
\ I
z
E
3 20 r "l
“I
10 ‘ I) d
,J
1’ l
0 . 1 l 1 l 1 l 1 1 1 l 1 l 1 l 1 1 1 1 1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.9 0.9 1.0

h (H/HMT»

Figure 44 - F p versus h at fixed reduced temperature for the Nb/CuGe(2°/o) samples.

90

I 730-7 Nb/CuGe 1oz 2001/2511n=7 0.2040041: 3.074010'r

0 723—5 Nb/CuGe 102} 2001/1001 n=7 T..=7.0040.05K Ht..=3.3040.21'r
0 707-2 Nb Film 3000 0.004004K 3.0040071'

 

l ' l ' l ' l ' l ' I ' l ' l T 1
Reduced Temperature:

4° ' t = 007040002 ‘

:7 30~ -

E

\

z

E 20— 1\ .

Ii: ~ / t 1
10* ,/ \ *

 

 

 

0 1 l 1 l L l 1 l 1 l 1 l 1 1 1 1 1 . 1
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.7 0.0 0.0 1.0
h (H/H£1(T))

Figure 45 - F I, versus h at fixed reduced temperature for the Nb/CuGe(10%) samples.

91

. 707—0 Nb/Cu 2001/1001 11:7 Tc,=7.4940.04K H..,=2.6040.17T
1 730—2 Nb/Cu 2001/1301 11:? 7.004004K 3.0440151‘
- 730—3 Nb/Cu 2001/251 n=7 000400411 3.3340. 121'
0 707—2 Nb Film 30001 0.604004K 3.064007T

 

4O _ Reduced Temperature:
t = T/Te, = 0.66740007

1 W.

20-

F, (10' N/m’)

10r

    

 

0 1 l 1 l 1 1 1 1 1 l 1 I 1 I 1 L 1\‘\l. 1
0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.7 0.0 0.9 1.0
h(H/H%1(T))

 

Figure 46 - F ,, versus h at fixed reduced temperature for the Nb/Cu samples.

92

6.2.2 Parallel

Here the exact same procedure was used for extracting the critical currents and the
H020) curve as in the perpendicular case. However here the scaling fails. Figures 47 and
48 illustrate the best attempt at scaling on two of the parallel samples. However if one
looks at the critical current, Figure 49, behavior similar to that described by Koorevaar,
Maj, Kes, and Aarts32 is seen. However instead of seeing a peak in 1., more of a plateau is
seen. This seems to indicate the structural vortex lattice transition they proposed is also
present in our Nb/CuX samples. This may also account for why the data doesn’t scale.
Figure 49 and 50 are two 1, versus h diagrams. Again the top curve is taken at 5K, the
next curve at 5.2K, and so on in 0.2K increments. Figure 50 includes two runs at 5K, the
solid and open symbols, showing the reproducibility of the data. As in the perpendicular
case a sharp rise is seen at low h as the sample enters the Meissner state.

However a more interesting property is seen in the pinning force density as a
fiinction of Ge concentration. Here the effects which reduce coupling between S layers
and produce weak pinning in the perpendicular orientation, cause the N layers and their
interfaces to be good pinning sites. This is exactly what is seen in Figure 51 where F I, is
plotted versus h for two multilayers at three different temperatures. In the Nb/CuGe(2%)
sample, pinning is comparable to that of a Nb film. In the Nb/CuGe(10%) sample, the
pinning has increased by an order of magnitude and there is even a slight enhancement of

TCo . The peak has shifted to a much lower value of h, which is explained by the model of

flux pinning presented by Kramer”.

1.3
1.6
1.4

F,/Prefactor (10' N/m’)

93

Nb/CuGe(102) 2001/1001 n=7
770-7

 

1.2

l

‘ F, = Prefactor{(m+n)h/m}"{(m+n)(1-h)/ n}"

l

L m = 1.00440030
*. + n = 2.17040054

 

 

l++<ODDOCQCOO

71.713199999999951“

 

 

0.0 L l 1 l 1 l 1 l 1 l 1 1 1 l 1 1 1 l
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.3 0.9

h (Ii/HUT»

Figure 47 - Attempt at scaling FI, versus h" for the Nb/CuGe(10%) 280N100A sample.

fifififififififififlfififi

94

 

. 5.0x
Nb/CuGe(107.) 2001/251 11:7 . gigfi
739-7 v .
F, = Prefactor{(m+n)h/m}"‘{(m+n)(1—h)/n}ll 1 333%
' l I ' l ' 1 ' l ' l ' l ' l T l D 6‘4K
1 2 * O 8%
- 1 . . m= 157440.032 .* 72011
3 v v ,. 1‘: :3," n = 297440003 .: 3f;
“ 1 7.8K
~ 7.01:
v 8.0K

F,/Prefactor (1 0" N/m’)

 

 

 

J_

.0 1 141 1 1% 1 1 1 L 1 l 1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

h (Ii/HEAT»

Figure 48 - Attempt at scaling F, versus H' for the Nb/CuGe(10%) 280N25A sample.

95

Nb/CuGe(107.) 2001/501 11:7
770-8

 

  
    

l ' T l

gaming-01,1...

  
  

    
    
    
   

F

(4

I. (am)
8

w
%
4;"

...—-

   

’—

,5

f/

  

 
  

' I
at;

 

O
o
"l
l
ol
N
O
to
.0"
.0
C311!
0'3.
O
03
O
‘1
.0
a:
.0
co
y-a
0

Figure 49 - The critical current versus reduced field 12" at various temperatures for the
Nb/CuGe(lO%) 280A/SOA sample.

96

Nb /CuGe(107.) 2001/1001 11:7
770—7

 

T ' l ' l ' l l l ' l ' l V T fi'

I. (111.1)
8

  

0 .. ;_
0.0 0.1 0.2 0.3

 

 

0.4 0.5 0.0 0.9 1.0
h(H/H'..(T))

Figure 50 - The critical current versus reduced field h‘l at various temperatures for the
Nb/CuGe(10%) 2801/1001 sample.

97

 

 
   
  

Nb/CuX 2001/1001 . 2.2%
- 5:4K
200 _ T T T l l l l l 1 fi
180 —

x = Ge(1oz) Tc, = 7.94K

F, (10" N/m’)

 

   

0 1 L . ‘1 . 1 . 1 . 1 . 1 1 1 1 1 _:
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
h (H/H11(T))

 

 

Figure 51 - FIp versus h at three different temperatures for the Nb/CuGe(2%) and
Nb/CuGe(10%) 2801/1001 samples.

98

6.3 Conclusions

In the perpendicular orientation F, scales and the expected scaling exponent is
found. There is also strong evidence for a transition from 3D vortex lines to 2D vortex
pancakes as a fimction of normal metal layer thickness found in the Nb/CuGe(2%) and the
Nb/CuGe(10%) multilayers. In both samples this transition occurs between 25A and
100A. For the Nb/Cu multilayers this same analysis proves inconclusive, possibly due to
the lack of strong pinning sites in the pure Cu.

In the parallel orientation F, does not scale, but there does appear to be evidence
for a structural vortex lattice transition from a anisotropic Abrikosov region to a kinked
region where the critical current density is now determined by vortex disks with their
cores perpendicular to the layers. Since J, is determined by two vastly different physical
mechanisms in these two regions, it is not too surprising that F, does not scale. It has also
been demonstrated that the pinning strength of the N layers is affected by the addition of
Ge, with no detrimental effects on T c. In going from N layers of CuGe(2%) to

CuGe(lO%), the pinning strength increases by an order of magnitude.

Chapter 7
FUTURE DIRECTIONS

There is still a lot of theoretical work that needs to be done on the irreversibility
line problem. The transition at T1,, is believed to be highly sensitive to disorder, being
either first order or continuous depending on the number and geometry of impurities.
There are however no theoretical predictions for how the nature of the impurities should
affect the shape of the irreversibility line. There are also no theoretical predictions for
how the 3D line / 2D pancake vortex transition should affect the shape of the
irreversibility line. There are also several interesting properties that may be seen in the
parallel orientation, like the 3D/2D/3D transition discussed earlier, which may influence
the shape of the irreversibility line, but again there are no theoretical predictions. This is
not to say that there is no direction for the experimental work to take. The use of a
SQUID magnetometer which allows for the direction of the applied field to be aligned
with the sample should greatly improve the reproducibility of the parallel measurements of
the irreversibility line.

Due to the inherent difficulty in obtaining good and reproducible measurements of
the irreversibility line with magnetization measurements, most future work will probably
be directed toward the study of transport properties. The next step in the vortex pinning

problem has already begun. Dr. P. Duxbury and his student Z. Zhou have begun a

99

100
detailed study into the vortex pinning energy present in layered superconductors. They
started from Ginzburg—Landau theory and are using a six parameter model. Their model
has two layer types and each is characterized by a coherence length, a penetration depth,
and a thickness. They have started by doing some numerical calculations of the energy of
a single vortex parallel to the layers as a function of displacement. Comparing the change
in energy from one layer to the next allows one to calculate the pinning energy. They plan
to add additional vortices to their model to take into the account the vortex-vortex
interactions. They will also investigate the perpendicular orientation by looking at vortices

perpendicular to the layers.

LIST OF REFERENCES

LIST OF REFERENCES

1 G. Blatter, M. V. Feigel’man, V. B. Geshkenbein, A. I. Larkin and V. M. Vinokur, Rev.
Mod. Phys. 66, 1125 (1994).

2 I. M. Slaughter, W. P. Pratt, Jr., and P. A. Schroeder, Rev. Sci. Instrum. 60 (1), 127
(1989)

3 Si wafers were purchased from Silicon Quest International, 2904 Scott Blvd., Santa
Clara, CA 95054.

4 MICRO is a liquid laboratory cleaner purchased from International Products
Corporation, P. O. Box 70, Burlington, NJ 08016.

5 Kodak’s part number is 1208966.

6 Photoresist and developer were purchased fi'om Shipley, 455 Forest Street,
Marlborough, MA 01752.

7 Plasma Therm Inc., 9509 International Ct, St. Petersburg, FL 33716.

8 Angstrom Sciences, P. O. Box 18116, Pittsburgh, PA 15236.

9 Aesar/ Johnson Matthey Inc., P. O. Box 8247, Ward Hill, MA 01835-0747.

‘0 M. L. Wilson, Ph. D. dissertation, Michigan State University (1994).

” Quantum Design, 11578 Sorrento Valley Road, Suite 30, San Diego, CA 92121.
‘2 Keithley, 28775 Aurora Road, Cleveland, OH 44139.

'3 M. Suenaga, D. O. Welch, and R. Budhani, Supercond. Sci. Technol. 5 SI (1992).

1‘ Matthew. F. Schmidt, N. E. Israelofi‘, and A. M. Goldman, Phys. Rev. B. 48, 3404
(1993).

101

 

102

‘5 Purchased from Wilmad Glass Company, Inc., Route 40 & Oak Road, Buena, NJ
083 10.

‘6 Purchased fi'om Electrical Insulation Suppliers, Inc., 19769 15 Mile Road, Mt. Clemens,
NH 48043.

‘7 Purchased from Apiezon Products, M&I Materials Ltd., PO. Box 136, Manchester
M60 IAN, England.

‘8 Y. Yeshurun and A. P. Malozemoff, Phys. Rev. Lett. 60, 2202 (1988).

19 P. de Rango, B. Giordanengo, R. Tournier, A. Sulpice, J. Chaussy, G. Deutscher,
J. L. Genicon, P. Lejay, R. Retoux and B. Raveau, J. Phys. France 50, 2857 (1989).

2° Z. J. Huang, Y. Y. Xue, R. L. Meng and C. W. Chu, Phys. Rev. B 49, 4218 (1994).
21 J. R. L. de Almeida and D. J. Thouless, J. Phys. A: Math Gen. 11, 983 (1978).

22 A. Houghton, R. A. Pelcovits and A. Subda, Phys. Rev. B 40, 6763 (1988).

23 Y. M. Wan, s. E. Hebboul and J. C. Garland, Phys. Rev. Lett. 72, 3867 (1994).

2‘ Robert A. Hein, Thomas L. Francavilla and Donald H. Leibenberg, ed., Magnetic
Susceptibility of Superconductors and Other Spin Systems, Plenum Press (1991).

2’ Youwen Xu and M. Suenaga, Phys. Rev. B 43, 5516 (1991).

26 J. H. P. M. Emmen, G. M. Stollman and W. J. M. De Jonge, Physicia C 169, 418
(1990)

27 L. Civale, T. K. Worthington and A. Gupta, Phys. Rev. B 43, 5425 (1991).

28 M. C. Frischherz, F. M. Sauerzopf, H. W. Weber, M. Murakami and G. A.
Emel’chenko, Supercond. Sci. Technol. 8, 485 (1995).

29 w. A. Fietz and w. w. Webb, Phys. Rev. 178, 657 (1969).
3° Edward J. Kramer, J. Appl. Phy. 44 1360 (1973).
3‘ A. M. Campbell and J. E. Evetts, Adv. Phys. 21, 372 (1972).

32 P. Koorevaar, W. Maj, P. H. Kes, and J. Aarts, Phys. Rev. B 47, 934 (1993).

   

"111111111111

IES

ljl