3::
H. mi” 0

.h: .L
, 1:11;?

< I
.1.

x.
.1
a, s

f:

.1»: 1."...

 

km”? 3&1
Fﬂgﬁ: .

wﬁﬁﬁ
. : :h .

, : :
“$313.1.“

11$»: . 1P.»

.
aziﬁuﬁmu ;
“in“?! .1

|V

. 2...”. : : : . . , : . : : Unn\h.sv.»..‘
. . . : . , . . , 1.4..

a. . : f .42: p .5: «'3 #3. 2:11.- : : 3: . v" ..
1:11:52». . .. : ,1: 35:2 .1: 15%»..m.... :_1»: ﬁyi: 1%.”. ﬁg ﬁg»; sﬁﬁﬁg
: . 1 . Li: . s. 1:: 22...! : 11.33: an: n .

 

This is to certify that the
thesis entitled

FABRICATION OF SUBMICROMETER-SIZED ELECTRICAL
PROBES FROM GLASS-METAL COMPOSITE USING A
LASER-BASED MICROPIPETTE PULLER

presented by

Nickolas M. Meyer

has been accepted towards fulﬁllment
of the requirements for the

Master of degree in Materials Science and
Science Engineering

 

’ 7"
a/ ‘

Major Professor’s Signature
11/03/05

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

v : ' T

vw-» -

 

 

 

LIBRARY
Michigan State
University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p.15

FABRICATION OF SUBMICROMETER-SIZED ELECTRICAL PROBES F ROM
GLASS-METAL COMPOSITE USING A LASER-BASED MICROPIPETTE PULLER

By

Nickolas M. Meyer

A THESIS
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
MASTER OF SCIENCE

Department of Chemical Engineering and Materials Science

2003

ABSTRACT

FABRICATION OF SUBMICROMETER-SIZED ELECTRICAL PROBES FROM
GLASS-METAL COMPOSITE USING A LASER-BASED MICROPIPETTE PULLER

By

Nickolas M. Meyer

Submicron electrical probes were fabricated from glass capillary tubes and thin metal
wires. These probes have applications in measuring the electrical properties of materials.
The probes were fabricated in a C02 laser based micropipette puller. Empty silica tubes
have been pulled to outer diameters of less than 100 nm (outer diameter) and wire-ﬁlled
glass tubes have been pulled to less than 500 nm (outer diameter). The glass absorbs the
laser energy and heat is transferred to the metal wire. When both the glass and metal had
softened sufﬁciently, the clamps at each end of the tube apply tensile force. This process
draws the glass and metal down together, and results in two tapered micropipettes with a
core of metal inside. The metal acts as an electrically conductive channel, which is
insulated by the surrounding glass. After the fabrication method of the microprobes was
established, the probes were integrated into a modiﬁed four-point probe measurement
system. In this system, the microprobe is mounted onto a translation stage, and is
scanned across the surface of the material, while the other three probes remain static.
This allows for measurement of the electrical conductance of the sample and the contact
resistance between the sample and the current leads. The fabrication issues of these

probes and their applications are presented.

Dedicated to Marten and Sophia Meyer,

my loving parents.

iii

ACKNOWLEDEMENTS

I would like to take this opportunity to extend my most humble gratitude to all those
who supported me throughout my graduate studies. First and foremost I’d like to thank
my advisor, Dr. Tim Hogan, without whom this opportunity would not have been
possible. I can’t give thanks enough for the dedication of your time, sharing of
knowledge, and sound advice that I received from you.

I’d like to thank my fellow team members that assisted me along the way.
Thank you from the bottom of my heart to Dr. Sim Loo, Dr. Fu Guo, Dr. Hyun Park,
Chun-I Wu, Jarrod Short, Adam Downey, Luke Niewiadomski, Dr. Roger Booth, Arjun
Jayaraman, and Akiko Kubota for all of the help you gave me. Your support and
friendship truly made this all a positive experience. I’d also like to give thanks Dr. Jun
Nogami and Dr. K. N. Subramanian for serving on my thesis defense committee.

I would like to especially acknowledge the love and support of my parents,
Marten and Sophia Meyer, and my brothers, Tony and Alex. You are the reason that I
made it this far. Thank you for everything you’ve given me and for being the
wonderful people that you are. Thank you for your prayers and for your guidance. I
could never repay the debt that I owe you.

I would also like to thank a very special person in my life, Andrea Baird. Thank
you for your tireless support of me and for being there whenever I needed you. You
helped me more than you’ll ever know.

Finally, I’d like to thank almighty God for guiding my path and keeping my

loved ones in his hands.

iv

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................... vii
LIST OF FIGURES ....................................................................................................... viii
Chapter 1
Introduction
1.1 Research Objective .................................................................... 1
1.1.1 History of Probe Fabrication ......................................................... 2
1.2 Survey of Submicron Probe Applications .......................................... 3
1.2.1 Imaging P-N junctions by NSOM ................................................... 4
1.2.2 Measurements of MEMS in Microprobe Station .................................. 6
1.2.3 Peltier-effect Module for Highly Localized Temperature Manipulation ...... 8
1.2.4 Scanning Nanolithography Using a Material-ﬁlled Nanopipette .............. 12
Chapter 2
Theory
2.1 Composite Material Mechanics .................................................................... 14
2.1.1 Mechanisms of Deformation of Fiber-Reinforced Composites in Tension. 18
2.2 Viscosity of Glass-Metal composites ........................................................... 19
Chapter 3

Material Selection and Experimental Method

3.1 Material Selection ......................................................................................... 22
3.1.1 Mechanical Property Considerations ............................................................ 24
3.1.2 Temperature Effects on the Mechanical Properties of Metal and Glass ...... 25

3.2 Probe Fabrication .......................................................................................... 32
3.2.1 Sutter P-ZOOO Micropipette Puller. ............................................................... 34

3.3 Probe Measurement Technique .................................................................... 44

Chapter 4
Results

4.1 Submicron Probe Fabrication ....................................................................... 48

4.1.1 Critical Volume Fraction .............................................................................. 49

4.1.1 Fabrication of Probes From Borosilicate Glass and Pt Wire ................... 50
4.1.3 Fabrication ofProbes From Silica Glass and Pt Wire. . . ..51
4.1.4 Fabrication of Probes From Borosilicate Glass and Gold Wire. . . . . . . . . .57

4.2 Analysis of Probe Fabrication ..................................................... 64
4.2.1 Calculation of the Elastic Modulus ............................................... 64
4.2.2 Viscosity of Probe Fabrication ..................................................... 66

4.3 Electrical Conductivity Measurement ............................................ 69
4.3.1 The One-Point Scanning Probe Method .......................................... 70
4.3.2 Electrical Conductivity Measurements Using Pt Microprobe. . . . . . . . . . . .......71

Chapter 5
Conclusion
5.1 Submicron Probe Scanning Measurement System ............................. 75
5.2 Suggestions For Future Work ..................................................... 75
5.2.1 Mount Probes in Microprobe Station ............................................ 75
5.2.2 Continued Improvement of Probe abrication Procedure ...................... 76
Appendix A ......................................................................................... 78
References ......................................................................................... 80

vi

Table 1.1

Table 3.1

Table 3.2

Table 3.3

Table 3.4

Table 4.1

Table 4.2

Table 4.3

Table A]

Table A2

Table A3

LIST OF TABLES

Dimensions of the MEMS devices measured in micro probe station ......... 8
List of materials used in research along with select properties ............... 23
Temperature dependant mechanical properties of nickel ...................... 26
Temperature dependent mechanical properties of silica glass ................ 28
Viscosity versus temperature for silica and borosilicate ....................... 29

Table of values where viscosity is calculated for different HEAT values...67

Values of viscosity for different values of VELOCITY ...................... 68
Collection of electrical conductivity data of BizTe3 using the scanning
probe conductivity technique .................................................... 72
Temperature dependant properties of pure Pt ................................... 78
Temperature dependant properties of Pt 10%Rh .............................. 78
Temperature dependant properties of Pt 10%Ir ............................... 79

vii

Figure 1.1
Figure 1.2

Figure 1.3

Figure 1.4
Figure 1.5

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5

Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7

Figure 3.8

LIST OF FIGURES

Schematic of the cross-section of the sample .................................... 5
Micro probe station used in space application measurements. . . . . . . . . ...7

Cooling device schematic with thermoelectric chip and metallic
microprobes .......................................................................... 9

Schematic diagram of the LIGA technique ...................................... 11

Nanopipette depositing a dot of photoresist onto a gold-coated glass
substrate ............................................................................ 12

Schematic of area of glass tube of diameter, dm, pulled down around the
metal wire of diameter df. This is the area where the glass and metal are

considered act as a composite material mechanically. . . . . . . . . . . . . .....14

Dimensional changes between glass tube and metal wire a). before the
initial pull and b). aﬁer the initial pulling step (not to scale) ............... 15

Total force acting on composite .................................................. l7

Brittle fracture of a borosilicate glass capillary tube that was not heated
during deformation. The force was applied axially in tension. . . . . . . . .....20

Image of glass tube pulled down into a micropipette with a platinum wire

inside ............................................................................... 20
Effect of rising temperature vs. mechanical properties of metal ............ 25
Effect of temperature on mechanical properties of nickel. . . . . . . . . . . . ...27

Plot of elastic modulus versus increasing temperature for silica glass ..... 28

Plot of log viscosity versus temperature ......................................... 30
Wire melted inside pulled micropipette ......................................... 31
Schematic illustration of wire threaded through glass tube .................. 32
Resulting probe tip shape from the fabrication procedure .................... 33
Sutter P-ZOOO micropipette puller ................................................ 34

viii

Figure 3.9 Sample of the display where the program is entered into the P-2000. . .....36

Figure 3.10
Figure 3.11
Figure 3.12

Figure 3.13

Figure 3.14

Figure 3.15

Figure 3.16

Figure 3.17

Figure 3.18

Figure 3.19

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5

Image of the carriage system of the P-2000 .................................... 37
Pull cycle when a DELAY greater than 128 is selected. . ....38
Pull cycle when a DELAY less than 128 is selected .......................... 39
Sketch of cross-sectional view of the dimensional mismatch between the
inner wall of glass capillary tube and the metal wire .......................... 40
Example of microprobe that was pulled with the wire breaking
prematurely. As seen, the tapered glass extends far beyond the point
where the wire ended ............................................................... 41
Glass tube collapsed around the metal wire as the ﬁrst pulling step in the
probe fabrication procedure ....................................................... 42
Cross-sectional view of middle portion of Figure 3.15. The glass is pulled
down so that the air gap that existed between the glass and the metal is
either minimal or eliminated completely ....................................... 43
Illustration of fabricated micro probe as a result of the pulling procedure
executed by the P-2000 ............................................................ 44
Schematic of traditional four-point probe arrangement ........................ 45
Illustration of one-point scanning probe setup. The micro probe is

mounted on a xyz transnational stage and acts as the positive voltage probe.
This probe is free to scan over the entire surface of the sample. . . . . . ...46

Illustration of one-point scanning probe setup. The micro probe is
mounted on a xyz transnational stage and acts as the positive voltage probe.
This probe is free to scan over the entire surface of the sample ............. 53

Result of pull with 0.3 mm inner diameter silica tube and 0.08 mm
diameter Pt wire. The diameter of the tip is approximately 0.02 mm.... . .54

Image of probes made of 0.3 mm inner diameter silica glass and 50 ,um Pt
wire ................................................................................... 55

Zoomed in view of one of the probes depicted in Figure 4.3. The diameter
of the tip is less than 5 pm ........................................................ 56

Initial pull that shows a borosilicate tube pulled down around a gold
wire ................................................................................. 58

Figure 4.6

Figure 4.7

Figure 4.8

Figure 4.9
Figure 4.10

Figure 4.11

Figure 4.21

Figure 4.13

Figure 4.14

Figure 5.1

Figure A]

This ﬁgure illustrates where the laser beam would strike the glass normally
and where it strikes the tube after it has been shifted over ................. 59

These images show the difference in the outcome of the pull when a). the
glass was shifted after the ﬁrst step and b) the glass was not shifted after

the ﬁrst step ....................................................................... 60
SEM image of gold probe pulled using Sutter P2000 micropipette

puller ............................................................................... 61
SEM image of gold and borosilicate probe .................................... 62
SEM image of gold micro probe ................................................ 63

Plot of the data in Table 4.]. Values of HEAT between 365 and 375 result
in the lowest viscosity of the composite ....................................... 68

Plot of viscosity versus VELOCITY for Au/Borosilicate microprobe
fabrication ......................................................................... 69

Scanning one-point probe conﬁguration for conductivity measurement on a
given sample ....................................................................... 70

Plot of AC conductivity measurement on BizTe3 sample using Pt micro
probe ............................................................................... 73

Illustration of Si nanowire grown off tip of silica-gold submicron
electrode ........................................................................... 77

Comparison of elastic modulus vs. temperature between pure Pt, Pt
10%Rh, and Pt 10%Ir ........................................................... 79

1. INTRODUCTION

1.1 Research Objective

The purpose of this research was to fabricate electrical probes of micron and below sized
tip diameters from glass capillary tubing and metal wire with a laser-based micropipette
puller. The microprobes are a glass micropipette with an electrically conducting metallic
core, which is continuous to the very tip. The glass capillary tube is drawn down with a
thin metal wire inside. The amount of tapering that will occur is dependant on the
dimensions of the glass and metal, and the parameters programmed into the micropipette
puller.

The Sutter P-2000 micropipette puller was selected for this project, mostly
because of its ability to pull silica. The P-2000 employs a C02 laser, which heats a glass
tube in the middle and then pulls it apart into two micropipettes. This puller uses ﬁve
different parameters to pull micropipettes including: power of the laser, scanning length
of the laser, the level of viscosity of the glass, the pull force, and the time between laser
deactivation and pull initiation. These parameters are varied until a combination is
discovered that will taper the glass and the metal in a uniform fashion down into small,
sharp tips.

The probes were then integrated into a measurement system to explore the
electrical conductivity of different samples. A variation on the four-point probe
conﬁguration was developed for these probes, which will be discussed in later sections.

The micro probe fabrication procedure and results will be presented throughout
the course of this thesis; along with electrical conductivity data measured using the

probes.

1.1. I History of Probe Fabrication

In 1953, researchers at Johns Hopkins University described creating a new “low-
impedance-microelectrode” to replace glass micro capillaries ﬁlled with conducting
electrolyte solution [1]. These electrolyte ﬁlled micropipettes had low signal-to-noise
ratio if their tips were large and had large impedance if the tips were small. Therefore,
they set out to produce a metal-ﬁlled micropipette that would have a small tip and low
impedance.

They describe the fabrication of a metal-ﬁlled microelectrode by melting an alloy
of 50% indium and 50% tin on a hot plate. The melted metal alloy was then sucked up
into a Pyrex capillary tube about halfway up the length and allowed to solidify. Then,
using a micropipette puller, a tip was drawn just past the point of the solidiﬁed metal.
After the tip was drawn, the micropipette was placed back onto a hot plate for a time until
the solidiﬁed alloy inside melted again. An ordinary sewing pin was then inserted into
the open end of the micropipette and pushed the metal to the tip [1]. This pin was then
used as electrical contact. This was an early method of producing a microelectrode.

In one of the earliest publications found where the metal wire and glass capillary
tube were pulled together, Bradford et a1. discussed the importance of the ductility of the
metal wire [2]. In their experiment, they utilized a 75 gm diameter platinum wire along
with a borosilicate glass tube. They claimed that the success of the electrode fabrication
was dependent on the fact that the platinum wire they were using was annealed and not
hard tempered. The increased ductility of the platinum wire enabled it to be drawn down

to extremely small dimensions.

Hard drawn Pt wire was tested, but they reported that this wire would not reduce
in diameter during the pulling procedure. They also mention that gold wire would be a
satisfactory core material though did not investigate it extensively. The ease and
quickness of fabrication, small tip dimensions, and reproducibility of the electrodes
(about 42% reproducibility) were three advantages of this procedure over the procedure
developed by the team at Johns Hopkins.

Nanonics Imaging Ltd. reported creating micro and nanoprobes from these same
group metals [3]. The process used by this company is of course proprietary, but in an
article out of Review of Scientific Instruments by a group from the Hebrew University,
they shed some light on how to create these micropipettes with a core of metal. They
treated the glass-metal combination as a composite material, with the glass acting as the
matrix and the metal wire acting as the reinforcement ﬁber.

This metal wire acts as an electrically conductive channel that is accessible from
the open end of the pipette and is continuous to the tip. Thus electrical contact can be
made in extremely localized areas where a vast number of applications lie in the ﬁeld of
material investigation and property measurement. Some examples of more current

applications of microprobes will be presented in the next section.

1.2 Survey of Submicron Probe Applications

As interest continues to grow in research of physical process on microscopic and smaller
levels, microprobes will become of greater and greater importance. This section will

offer some examples of how these types of instruments are being used in different

research situations. A brief synopsis of four different applications utilizing microprobes

or micropipette-based instruments will be presented.

I .2. I . Imaging P-N junctions by Scanning Near-ﬁeld Optical Microscopy (SNOM)

One of the goals of this research project was to obtain the capability of taking
measurements in extremely localized areas using fabricated microprobes. An article by
G. Tallaria [4] working with Nanonics Imaging Ltd discusses the characterization of p-n
junctions based on scanning probe microscopy and conductive microprobes. The
measurements presented in this article are compared to a previous study that
characterized p-n junctions using Electric Force Microscopy (EFM) [4].

The properties sought out in this study were local capacitance characteristics,
small-scale variations of dielectric constant and resistivity, p-n junction delineation, and
oxide quality. To obtain these measurements, a scanning probe system was utilized
which can collect electrical data as well as capture the topographical features of the
surface of the sample. However, some difﬁculties arose due to uncertainty of the tip
shape of the scanning probe, interaction between the tip and the contact area of the
sample, and any other forces that may be present.

The samples used in this experiment were cut and lapped so that the p-n junction
could easily be accessed. The oxide layer had the conﬁguration depicted in Figure 1.1.
The measurements were carried out using a Nanonics SNOM/AFM 100, which is a
tapered optical ﬁber that is covered by a layer of gold. The tip of the ﬁber is curved
downward in order to correctly serve as an AFM tip. The SNOM/AFM 100 is mounted

on the stage of an optical microscope and connected to an AFM/STM control system.

The SNOM signal is obtained in illumination mode in near-ﬁeld by sending an

Ar+ laser down through the ﬁber and collecting the reﬂected signal in far-ﬁeld with a

Field oxide

 

Passivation layer

 

 

 

 

 

 

 

 

 

 

 

Figure 1.1 Schematic of the cross-section of the sample [4].

photo-diode. The probe is also kept in constant contact with the sample by laser

feedback, which acts on the topography of the sample. For the electrical contrast of the

sample, an oscillating voltage was applied to the circuit that was completed by the

contact between the probe and sample. The signal was collected with a lock-in ampliﬁer.
The result of the experiment gave the topography, the optical, and the electrical

constant image all simultaneously. The parameters used in the experiment were:

0 Scan area — 8 x 8 pm2

0 Vpn — 5V

0 V0SC — 50 mV

0 coyosc- 100 kHz

where Vpn was the bias voltage, and V0SC was the oscillating voltage applied to the tip of
the probe.

It was found that the surface of the sample was ﬂat, topographically. There was a
difference in color between the silicon and the passivation layer, and the junction was
clearly seen in all three images. The electrical contrast area gives the clearest image
because the depletion region acts as a capacitive area. Because an alternating current is

used, the capacitive area induces the enhancement of signal in this area.

1.2.2 Measurements of MEMS in Microprobe Station

The ability to take measurements on microelectromechanical systems (MEMS) is integral
to being able to predict their reliability in a service environment. An article out of the
Southwest Research Institute in San Antonio, Texas uses a micro probe station to test and
measure how MEMS will hold up to the vacuum of space [5]. MEMS provide
advantages as space sensors and are coveted because of their small mass and size, low
power requirements, and precise dimension control.

Some of the applications for the MEMS will allow them to be shielded from the
space environment, such as magnetometers and gravity gradient monitors. For
applications where the device is exposed to vacuum, special issues arise. Service
temperature can vary by 70° C, there is high radiation exposure, and even more of a
concern is vacuum welding between moving parts [5]. Because of these issues, high
cycle testing of MEMS under vacuum is necessary, and requires microprobes to carry out

the measurements.

The measurement system employed by this research group is similar to the one
shown in Figure 1.2. This system pumps to vacuum with an ion pump, to avoid

mechanical vibration, and an oil-free molecular drag pump [5]. There are four probes

 

Figure 1.2 Micro probe station used in space application measurements [6].

which can be translated in the xy, and 2 directions with a resolution of less than 1 pm.
This control allows for manual actuation of the MEMS by using the microprobes to
supply voltage to the various pads on the devices [5]. The sizes of several of the devices
measured in the microprobing station are listed in Table 1.1. Some of these devices are
only 26 pm in width. This requires probes with extremely ﬁne tapers to carry out the
measurement of the devices. The properties measured of the devices were such things as

required operating voltage and Q — factor, or resonant frequency.

 

Table 1.1 Dimensions of the MEMS devices measured in micro probe station [5].

 

 

 

 

 

 

 

 

End Mass Size (,um)
Device Type Size Layer Length (,um) Width (gm) Thickness (,um)
P18 Small Poly 1 34.0 26.0 — 30.0 2.0
PlL Large Poly 1 98.0 66.0 — 110.0 2.0
P281 Small Poly 2 34.0 26.0 — 30.0 1.5
P232 Small Poly 2 34.0 26.0 — 30.0 1.5
P2M Medium Poly 2 58.0 46.0 — 62.0 1.5
P2L Large Poly 2 98.0 66.0 — 110.0 1.5

 

 

 

 

 

 

 

1.2.3 Peltier-effect Module for Highly Localized Temperature Manipulations

There are certain instances that come up in biomedical research where localized, precise
manipulation of heating or cooling is necessary. An answer to this requirement is to use
an array of the probes that can be inserted or placed adjacent to the microsample or area
in need of temperature manipulation. An example of an application that could beneﬁt
from this type of accurate heating/cooling control is in the ﬁeld of metabolic research.
The transmission activity of a nerve can be slowed down or even stopped when in the
presence of a temperature drop, from 37° C to 5° C [7].

A restriction to this type of research is the availability of probes small enough to
perform these types of localized temperature manipulation. This research done at
Louisiana State University proposes constructing a device that can locally cool nerve

cells in laboratory rats on a microfabricated probe [7].

The ﬁeld of cryobiology could also potentially beneﬁt from cooling devices
assembled on microprobes. Cryomicroscopy is a technique used to monitor the response
of cells and cellular systems to freezing temperatures. In this technique, volumes of 4 ul
of cellular solution are normally used.

A problem that can arise conducting this research is that ice can form inside the
cells, rupturing the membranes. If the cooling can be precisely controlled, this
intracellular ice formation can be avoided [7]. A cooling microprobe can be used as a

seeding probe to ensure extracellular ice formation without intracellular ice formation.

microprobes

/\

. ‘ . ‘ ‘ ' Titanium substrate

A‘— Cold junction

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

——>
¢

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hot junction

Figure 1.3 Cooling device schematic with thermoelectric chip and metallic microprobes

[7]-

The thermoelectric effect is taken advantage of as the driving force for the

localized cooling. The cooling device is made up of two components: a cooling

thermoelectric chip and an array of metallic heat conducting microprobes attached to the
top surface of the chip.

A schematic of the device is shown in Figure 1.3. The thermoelectric chip serves
as the substrate, thus when an electrical current is applied, the top surface is cooled,
which will be conducted through the microprobes. The current could also be reversed,
and the top surface would be heated.

The microprobe array was fabricated using a technique known as the LIGA
process (a graphical outline of this technique is depicted in Figure 1.4), which involves a
combination of deep-etch x-ray lithography, electrodeposition, and molding to fabricate
high aspect ration microstructures [7]. The LIGA process is advantageous compared to
other microfabrication techniques because the ability to make microprobes from various
metals and alloys.

A thin layer of polyrnethylmethacrylate (PMMA) is glued onto a plate of
titanium. Then a mask with the design for the microstructure patterned on it is placed
over the PMMA layer. It is then exposed to x-rays, leaving only the masked pattern on
the metal ﬁlm after the exposed portion is dissolved away in a molecular—weight sensitive
solvent [7]. Then nickel is electroplated into the voids left from the x-ray lithographical
process.

What is left after this step is freestanding nickel structures approximately 1000
pm in height. Then the titanium plate, with the microprobes, is attached to the
thermoelectric chip via a high temperature and thermally conductive epoxy. The
microprobes, made of nickel, are then coated with a biocompatible metal such as gold or

platinum.

lO

 

W§ : Photoresist

Titanium plate

X-raysi I i i i I X—raymask

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

i Exposed resist

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Developed resist

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Electrodeposited Ni

 

 

 

 

 

 

 

 

 

 

Ni microprobes

 

Figure 1.4 Schematic diagram of the LIGA technique [7].

ll

1.2.4 Scanning Nanolithography Using a Material Filled Nanopipette

Hong, et a1, describes a procedure that utilizes nanopipettes to deposit photoresist onto a
glass substrate [8]. Conventional photolithography is extremely useful, but as electronics
become smaller, more precise methods to deposit photoresist must be developed, to make
way for nanolithographic techniques. The greatest limiting factor in is the ability to
create a mask with micrometer and even nanometer resolution. With current technology,
this is an extremely difﬁcult task.

Using a pulled nanopipette, in combination with scanning probe microscopy such
as scanning tunneling microscopy, atomic force microscopy, or near-ﬁeld optical
microscopy, the liquid photoresist can be deposited on the surface of the substrate

through a nanometer-sized aperture.

__ photoresist

 

 

 

 

nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn
ooooooooooooooooooooooooooooooooooooooooooooooo
oooooooooooooooooooooooooooooooooooooooooooooooo

Figure 1.5 Nanopipette depositing a dot of photoresist onto a gold-coated glass substrate

[8].

 

   

12

Liquid photoresist is deposited through the opening at the tapered tip of a quartz
micropipette where the mask is basically written on the substrate. The photoresist can be
deposited point-by-point, as depicted in Figure 1.5, or can be written on like a pen on
paper. The micropipettes have an inner aperture at the opening of the tapered end
ranging from 10 -— 100 nm. With this resolution and accuracy, the designs of photoresist
that could be created could be limitless.

This method eliminates the need for a mask, needs no spin on coating of
photoresist, and requires no ultraviolet light source that conventional photoresist
procedures use. The system developed for depositing the photoresist using the
nanopipette requires that the pipette-substrate separation distance remain only a few
nanometers apart. A shear force control method can be employed to regulate this
particular parameter.

Dots o f p hotoresist 300 nm in diameter were deposited on a gold-coated glass
substrate at a distance of 1 ,um apart. Typically, each dot was approximately 80 nm in

height.

13

2. THEORY

2.] Composite Material Mechanics

The glass capillary tube and metal wire, once the glass is pulled down (Fig. 2.1), can be
dealt with like a glass matrix-metal ﬁber composite. This is the approach taken in
previous studies where probes were fabricated with a micropipette puller resulting in

extremely small tip diameters [9,10]. A good mechanical bond between the glass and

124-:- ~..;:-. ”71/ =.;, 1 31-, 3 a
3' ,z; I’dldliIILIi {I 3'11"!th

m Glass capillary tube

 

 

 

 

Figure 2.1 Schematic of area of glass tube of diameter, dm, pulled down around the metal
wire of diameter df. This is the area where the glass and metal are considered act as a
composite material mechanically [10].

metal is essential, as the glass will be the only force pulling on the metal wire. Figure 2.1
is a close up schematic of the middle region of where the glass and metal wire touch for
the example of platinum wire inside a glass tube. Figure 2.2 shows the dimensional
changes of the glass matrix after the initial step is executed from a cross-sectional
viewpoint. The inner and outer diameters deform concentrically, and the thickness of the

tube remains constant.

0.30 mm

b).
0.32 mm

 

Diameter of wire
remains constant

 

 

A L
V r

1.2mm

Figure 2.2 Dimensional changes between glass tube and metal wire a). before the initial
pull and b). after the initial pulling step (not to scale). This is a cross-sectional view of
Fig. 2.1.
Assuming that there is good bonding between the glass and metal then some

aspects of the mechanical behavior of the system can be assumed. Namely, the glass and

metal should deform in the same manner under the right conditions (dimensions,

micropipette puller parameters, etc). Therefore, each component should experience the

same amount of strain, a (change in length/original length),
8:8 =8 =8], (2.1)

where EC is the strain on the composite, cm is the strain on the matrix, and 8f is the strain
experienced by the ﬁber. With that said, it can also be said that the total force on the
system is the summation of the force on the matrix and the force on the ﬁber. This is
depicted in Figure 2.3. The equation for the total force acting on the composite is
F, =Fm +Ff, (2.2)
where F,,, is the force acting on the matrix and F f is the force acting on the ﬁber. The
total force on the system is directly related to the dimensions of the components by the
equation for stress,
F = 0' . A , (2.3)
where 0' is the stress and A is the cross-sectional area. Figure 2.3 shows the manner in
which the force is exerted on the composite material.
Now that an expression for force has been established, it can be related to strain
through an expression known as Hooke’s Law [11]. The law links stress and strain by a

proportional value known as Young’s modulus or the modulus of elasticity
E = 3. (2.4)

This equation can be rewritten in terms of force and area

E=F/A, (2.5)

8

 

and the force can then be written as

F=E-e-A, (2.60)

16

and the entire force, Ftotal, acting on the composite can be written as

F(=(Em-£-Am)+(Ef-£-Af), (2.6b)

Ftotal Ftotal
(— H —)

\U/ decompose Force into

 

 

- F1 acted on glass

 

 

 

 

 

- F2 acted on wire
F1 1 1 F1
4— —>
I 1
Glass
F2 F2
£— —>
Wire

 

Ftotal = F1 + F2

 

 

 

 

 

 

Figure 2.3 Total force acting on composite [10].

Then, substituting Eq. (2.6b) into Eq. (2.4), the resulting equation can be reached to

determine the elasticity of the composite [12]

 

=(Em-Am+Ef-A,)g/A,

E

(2.7)

(‘

Because the strain appears on the top and bottom of the equation, it can be cancelled out,
which leaves an equation for the elasticity of the composite which is based only on the
elasticity of its constituents and their respective dimensions. Thus, if there is a known

elasticity that will optimize the probe fabrication process, it can be obtained through the

17

materials selected and their dimensions. However, this is only a starting point, as it only
accounts for behavior at room temperature. The probe fabrication procedure is assisted
by a laser, which changes the properties of both the glass matrix and metal reinforcement

ﬁber. It can be used as a reference to what has been accomplished in previous work.

2.1.] Mechanisms of the Deformation of F iber-Reinforced Composite in Tension

There are ﬁve mechanisms that contribute to deformation in ﬁber-reinforced composites
that are under an axial load. Each stage represents different degrees of stress that the
composite is under. The ﬁrst stage is the elastic deformation stage. Both the ﬁber and
the matrix are bearing the load at this point. Initial crack formation takes place during
this stage, but no crack growth [13].

In the second stage of deformation, matrix-ﬁber debonding begins to occur on a
localized level. Also, the matrix begins to experience signiﬁcant cracking. Matrix-ﬁber
debonding occurs on a wider scale in the third stage, and now the ﬁbers begin to bare the
majority of the stress. The entire load is transferred to the ﬁbers in the fourth stage and
failure is the ﬁfth stage [13].

The composite featured in this research is a glass matrix and a continuous, single
wire, which acts as the reinforcement ﬁber. This composite goes through each of these
stages, however, does so while being struck by a C02 laser. The laser is contributing a
signiﬁcant amount of energy to the composite, breaking bonds Si-O bonds of the glass
matrix. Eventually, the energy will propagate through the glass matrix and reach the

metal wire.

18

The metal will experience an increase in atomic vibrations when it absorbs the
energy of the laser [14]. This affects the plasticity of the metal. As temperature
increases, the yield and tensile strengths decrease, and ductility increases (Fig. 3.1).
More of this topic is discussed in the next chapter. Thus, the plasticity is improved for
both the glass matrix and the metallic reinforcement ﬁber, a necessary facet considering
the extensive amount of plastic deformation that must take place. A way to understand

how much the plasticity is affected is to investigate the viscosity of the composite.
2.2 Viscosity of Glass-Metal Composites

Because the matrix of the composite material used is glass, viscosity is a property that
must be given consideration. Viscosity, i], is the property that determines ﬂuid ﬂow [13].
More precisely, it can be thought of as the resistance of a substance to ﬂow in a ﬂuid-like
behavior. It represents the amount of internal friction present within the ﬂuid when
velocity gradients are present. If a ﬂuid is viscous, then it has a higher degree of internal
ﬁiction and will have a higher resistance to ﬂow. A less viscous ﬂuid will ﬂow easier.
When a micropipette puller draws a glass tube down, it heats the glass and then
applies force in tension to pull the glass into two separate pipettes. The heat decreases
the viscosity of the glass, so that it will deform in a desired manner. If the viscosity of
the glass is not decreased, and tensile force is applied to the glass tube, a brittle fracture
will result (Figure 2.4) with practically no strain observed before failure. A large strain
before failure is necessary to make a micropipette (Figure 2.5), and this is where the

viscous properties of heated glass are of importance.

l9

 

Figure 2.4 Brittle fracture of a borosilicate glass capillary tube that was not heated during
deformation. The force was applied axially in tension.

 

Figure 2.5 Image of glass tube pulled down into a micropipette with a platinum wire
inside.

20

If the metal wire deforms uniformly with the glass, then we can assume that the metal is
deforming at the same rate as the glass. Thus, the viscosity for the composite as a whole
can be determined. The viscosity for a composite material is deﬁned as
n = o/strain rate, (2.8)
and the expression for strain rate is
Lo
dL ’
dt

where L0 is the original length of the tube or rod and dL/dt is the change in length divided

 

(2.9)

by the change in time, or the deformation rate. Combining Eq. (2.8) with Eq. (2.9) gives

a formula for the viscosity of a composite

”22$. (2.10)

If) ’

dt

which is a ratio between the applied stress and the strain rate [16, 17]. The units of
viscosity are N-sec/mz, or Pas. This formula will be used to analyze the probe fabrication
technique to test how certain parameters of the micropipette puller affect the viscosity of

the composite and affect the outcome pull.

21

3. MATERIAL SELECTION AND EXPERIMENTAL METHOD

3.1 Material Selection

The microprobes (microelectrodes) that were created as a part of this research were
fabricated from a glass capillary tube and a thin metal wire. The central difﬁculty of the
microprobe fabrication was inventing a method where, upon application, a glass capillary
tube and a thin metal wire inside could be drawn down together simultaneously to form a
tapered micropipette that had a conducting metallic channel running to the very tip. This
method calls for the uniform deformation of two mechanically different materials, and
the microscopic dimensions involved only amplify the level of precision necessary for the
success of this method.

The selection of the materials to be used is of utmost importance to the fabrication
procedure. Based on research already done in this area, the metal selected should be soft,
annealed wire. Therefore we look to the precious metals, known for not only their
ductility, but also for their corrosion resistance and excellent electrical conductivity.
These properties make metals such as gold and platinum ideal candidates. Metals such as
copper and nickel were investigated as well. Table 3.] lists various materials involved in
this research and some relevant physical properties.

The types of glass selected for this research were borosilicate and silica. These
are glasses that the micropipette puller used to fabricate the probes was designed for. The
ease of fabrication of borosilicate glass makes it preferable to use over silica, however,
some of the applications in mind for the probes requires the high melting point and

strength of Si02. Therefore, both types of glass capillary tubes were experimented with

22

in conjunction with several different varieties of metal wire. The list of Table 3.1 is the

catalog of materials considered for this project, with particular focus on gold, platinum,

silica glass, and borosilicate glass.

This section will explain step-by-step the method employed to create micro

probes from glass capillary tubing and metal wire. The instrument used to make the

probes will be described in detail, along with the integration of the fabricated probes into

a scanning probe measurement system.

Table 3.1 List of materials used in research along with select properties.

 

Elastic Modulus

Electrical Resistivity

 

 

 

 

 

 

 

 

 

 

Matenal (GPa) at 25 °C Melting Temp. ( C) (Q-cm)
Gold‘s 77.2 1064 2.2 x 104’
Nickel” 270 1455 6.4 x 10‘6
Copper18 110 1083 1.7 x 10"
Constantan18 162 1300 52.0 x 10’6
Platinumlg’zo 164.6 1770 10.6 x 10'6
Platinum/Iridum'g‘zo 202.3 1800 25.0 x 10"
Platinum/Rhodium'8’20 212.6 1840 18.8 x 10°

Borosilicate18 64 800 --
Silica‘g‘z' 72.9 1410 1 x 1020

 

 

 

 

23

 

3. 1. 1 Mechanical Property Considerations

The mechanical properties of the metal wire and glass capillary tube dictated how
much force was needed in pulling the pipette into two separate pieces. Through our
research, we found that gold and platinum gave the most promising results as the core
material for the microprobe. These metals are relatively mechanically soft, with gold at a
hardness value of 25 Vickers and platinum at a hardness value of 40 Vickers [18]. These
values are obtained with a Vickers Hardness Test that uses a pyramidal shaped indenter,
which is made from diamond. Loads of various magnitudes are applied to the surface of
the material being tested, and the hardness of the material is based on the diagonal made
by the indenter. The Vickers hardness is determined by this equation

_ 1.854F

HV D2

 

, (3-1)

where F is the applied load and D is the diagonal of the impression made by the indenter
(mm) [15].

Intuitively, the softer the material is, the larger D will be, indicating how hard a
material is. Generally speaking, the softer a material is, the easier and more uniformly it
will deform. In order for a probe tip to be made that has a diameter less than a micron,
the wire (initially 50 — 80 pm in diameter) will need to undergo extensive plastic
deformation axially. However, the deformation is taking place as a laser is heating the
glass and metal. What this requires is that the glass and metal reach a certain viscosity
together, so they will draw as a cohesive unit. If the glass and metal are not deforming

together simultaneously, a uniform taper will not result.

24

3.1.2 Temperature Effects on the Mechanical Properties of Metals and Glass

The inherent softness of these precious m etals allows for m ore uniform plastic
deformation. Of course, the pulling procedure takes place at elevated temperatures
(metals start to lose their mechanical properties at approximately three-quarters of their
solidus temperature), softening whatever metal lies inside of the glass, along with the
glass itself. Still, it was found that the best probes (ﬁnest taper) were of gold and
platinum. Figure 3.1 shows the general relationship between mechanical behavior of
materials and increasing temperature.

The effect of temperature on metals is complex because not only is there a purely
physical effect (the thermal plasticity is related to the amplitude of thermal atomic

vibrations which increase when the metal is heated), there are physiochemical process

Ductility

Yield Strength

Mechanical Properties

Tensile Strength

 

 

>
Temperature

Figure 3.1 Effect of rising temperature vs. mechanical properties of metal [15].

25

that occur as well . An expression was developed for predicting the effect of temperature
on the mechanical properties of metals which is
M 2 = M. expt—af: — r. )1. (4.2)

where M1 is a given mechanical property at t1, M2 is the resulting mechanical property at
t2, and a is a temperature coefﬁcient of the property [14]. The temperature coefﬁcient
takes into account the manner in which the stress is applied, the extent of plastic
deformation, the strain rate, etc.

Table 3.2 shows the effect of increasing temperature on some of the mechanical
properties of nickel. This table illustrates the effect that temperature has on the properties
of metals. Of signiﬁcance are the tensile strength and the percent elongation. These are

important to this type of probe fabrication in that the metal is undergoing extreme plastic

Table 3.2 Temperature dependant mechanical properties of nickel [l9].

 

 

 

 

 

 

 

 

 

 

 

Temperature (K) Yielghfligngth “£233 Tensiﬁigfngth Elepiition
293 148 0.31 462 47
400 153 0.31 459 46
500 140 0.31 459 44
600 138 0.31 462 46
700 1 15 0.31 328 64
800 100 0.31 245 68
900 69 0.31 176 72
1000 59 0.31 121 82
1 100 45 0.31 83 95

 

 

 

 

 

 

26

deformation in tension. Figure 3.2 illustrates an example of the decreasing trend in yield

and tensile strength and the increasing trend in strain to failure versus increasing

Mechanical Properties vs Temperature for Ni

 

 

 

 

500
. 4504.._l-___l 2l__l ______
E , - - - - - - c -
. 1': 400
1 g 350 1 1 1 I - -- w A. ,,
a: 300 1 1 1 1 ~ 1 1 111 1111 1 IOYeildStrength
.1; 2501 1 111111.1111111 I(GPa_)
o I Tensrle Strength
E 200 I — _ _ _ _ I _ _ _ — I (GPa)
I 5 150 ’L 4” * o r 6 ' m * ' 7" .7 m r I Strain-to-failure' ’
g 10011 1 1 1 1 9 11.111 1 1 1 1._. 11» ._ 3%)2 2
i 50 11 a . _ L. L ._, _._. -- "L. _Q ‘0‘ _.
0 : 1 : :
200 400 600 800 1 000 1 200

Temperature (K)

Figure 3.2 Effect of temperature on mechanical properties of nickel [19].

temperature. However, in the actual pulling process the temperature is not known exactly
due to the fact that the metal wire is surrounded by glass, so the thermal energy that the
metal sees must be estimated.

The mechanical properties of the glass also come into affect, as the glass is the
component that dictates the shape of the probe tip. Table 3.3 lists the relationship
between mechanical properties and increasing temperature for silica glass. This data
shows that the mechanical properties of quartz behave differently from metal under the

inﬂuence of elevated temperature. The modulus of elasticity, E, increases as temperature

increases for silica.

27

Table 3.3 Temperature dependent mechanical properties of silica glass [22].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Temperature (°C) Youngéllsgpdulus Sheazcgzglulus Poisson's Ratio
25 72.9 31.3 0.165
100 74 31.6 0.171
200 75.1 32 0.173
400 77.2 32.8 0.177
600 78.7 33.3 0.182
800 80 33.7 0.187
1000 81.1 34 0.193
1100 81.4 34.1 0.194
1200 81.5 34.1 0.195
1250 81.4 34 0.197

Elastic Modulus vs Temperature for Si02
82 .
81 1 1 1 1 14 1’ 11’ 1
a€80 1 1 1 o 1-
} )7 79 11 1 111 1 11 111 1.1 - A» 1'1 1 11 ~11 — 1
I 3 78 I11 11 111 1~ — 111-1—— 111-1— 1111 111— 1 11
:1
,, , ,, Q. ,,,,, __, ,, _ ,_ , ,, ,,_
'2 77
I 2 76111 11 1 1111 ~11 11 1 1-1 1
3 3 75 -4 4 . -1. i. _ 1 -1
z a 7411 o1
73 __._ __ . .__
72 a . . . . .
O 200 400 600 800 1000 1200 1400 .
Temperature (C)

Figure 3.3 Plot of elastic modulus versus increasing temperature for silica glass.

28

 

Another property that is affected by temperature is viscosity. The values for viscosity of
silica and borosilicate glass are listed in Table 3.4. The values are also listed for
increasing temperature. Figure 3.3 plots elastic modulus versus increasing temperature
for silica glass and ﬁgure 3.4 plots increasing temperature versus viscosity for silica and

borosilicate.

Table 3.4 Viscosity versus temperature for silica and borosilicate [22].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Silica Borosilicate
Temperature (°C) Log viscosity (P) Log viscosity (P)

500 - 15

600 - 12.15
700 - 9.72
800 - 7.88
900 - 6.97
1000 16.8 5.48

1 100 14.67 4.74
1200 12.82 4.17
1300 11.21 3.7
1400 9.8 3.4
1600 8.12 -

1800 6.74 -
2000 5.6 -
2200 4.64 -
2400 3.82 -

 

 

 

 

 

29

 

Temperature (C) vs Viscosity i + Silica

+ Borosilicate

 

2250 11

1750

12501 11

Temperature (C)

750 -

 

 

 

250

 

0 2 4 6 8 10 1 2 14 1 6 18
Log Viscosity (P)

Figure 3.4 Plot of log viscosity versus temperature [22].

From composite material mechanics, it is known that there is a critical volume
fraction, V”, where the plasticity of the composite material becomes equal to the
plasticity of the matrix [9]. This helps in selecting the dimensions of the metal wire and
the dimensions of the glass capillary tube. The ratio A = Dw/Dg, deﬁnes the volume
fraction V of the reinforcement ﬁber, where D... is the diameter of the metal wire and Dg,
is the outside diameter of the glass tube [9]. Therefore, by holding the diameter of the
wire constant, the outer diameter of the glass can be varied to reach a point where V z V6,
[9]. Nanonics [2, 9] reported the best results for volume fractions between 0.2 and 0.4 for
platinum wire and borosilicate glass. An analysis of the mechanics of the platinum and

borosilicate pulling procedure can be seen in Chapter 2.

30

Another factor to consider is how the metal and glass properties compare to one
another. Referring back to Table 3.1, gold and borosilicate glass share similar properties
mechanically and thermally. Of course, the plasticity of gold and borosilicate differ
greatly, borosilicate glass being very brittle and gold being ductile. However, the melting

points of the two are quite similar and actually were very useful to us for this study.

 

Figure 3.5 Wire melted inside pulled micropipette.

An example of a poor match of glass and metal can be seen in ﬁgure 3.5. Using
gold with silica does not work well because the melting point of gold is far less than even
the softening point of silica. Therefore, the gold will melt inside the tube before the silica

begins to pull, which does not give gold the chance to pull simultaneously with the silica.

31

3.2 Probe Fabrication

The probe fabrication technique developed for this research was based on previous works
mentioned in the introduction [1,2,3]. They devised a method in which a metal wire
(gold, platinum, silver, etc.) was inserted into a capillary tube of borosilicate glass. This

step is depicted in Figure 3.6.

Heat Applied
Pull Jr Pull

 

Figure 3.6 Schematic illustration of wire threaded through glass tube [10].

Then using a micropipette puller tapered the glass tube, along with the wire inside, down
into two separate pieces. Figure 3.7 depicts the probes created by the micropipette puller,
noting the great deal of tapering near the tips.

The puller works by heating the center of the tube and pulling on each end of the
tube to taper the glass. One could pull hundreds of tubes down into micropipettes a day,
it is a very simple task given a capable puller. However, these pullers are not designed to
create micropipettes from the glass tube including the wire, which is why a very speciﬁc

technique must be developed through experimental methods.

32

 

Figure 3.7 Resulting probe tip shape from the fabrication procedure [10].

The micropipette is a useful instrument and, as mentioned in previous sections,
ﬁnds many applications in biological research and now more commonly in other
branches of scientiﬁc research. The unique properties of glass allow for such a simple
process to produce a powerful instrument. The glass, in the case of a micropipette probe
acts as the matrix for the metal wire if the combination is thought of as a composite
material.

A metal wire could not be fashioned this sharp and precise by any other physical
material reduction method while at the same time creating an insulated covering for the
metallic core. The glass forms as it would and induces the metal wire to form along with
it. This requires the glass and the metal to have a certain degree of viscosity. The wire is
undergoing extensive plastic deformation, as there have been metal tips reported with
diameters less than 50 nm from a wire that was originally 50 ,um, a 1000% reduction in
diameter [9]. This is no trivial matter, as a 50 ,um wire is very sensitive to breaking and

melting. A small tear or separation of the wire near the tip can signiﬁcantly affect the

33

ability of the probe to conduct electricity. Measures can be taken, however, to aid the

deformation procedure.

3. 2. 1 Sutter P-2000 Micropipette Puller

The Sutter P-2000 (Figure 3.8) uses several parameters to deliver the proper amount of

heat, force, and timing to fashion a micropipette, or in this case, a probe with 3 micron

 

Figure 3.8 Sutter P-2000 micropipette puller [23].

scale tip diameter. A C0; laser supplies the heat and was selected for this micropipette
puller for several reasons. The wavelength of this type of laser matches the resonant

frequency of the silica lattice, allowing the fabrication of pipettes out of silica. There is

34

no metal residue on the pipette because of clean heat delivered by the laser. Finally, laser
heat can be turned off instantly, leaving no residual ﬁlament heat.

This is advantageous in that many pipettes can be fabricated in a relatively short

amount of time with little or no waste produced. Once a program to create the micro
probes is established, a large quantity of them can be made shortly. However, the
reproducibility of each pull is an issue. A pulling program may work well one time, then
fail to reproduce those results for several attempts. There are many factors that affect the
reproducibility of each pull, for example, the way in which the wire is placed in the glass
tube, whether or not the mirror is clean, whether or not the glass tube is of correct
dimension, etc.
The ﬁrst parameter is HEAT, which dictates the output power of the laser. The energy of
the laser will be absorbed by the glass, which will soften once adequately heated. The
range of the HEAT parameter is 0—999, with 500-600 able to soﬁen thin wall silica
tubing. HEAT values between 300 and 400 are able to soften borosilicate glass [24].
Laser power is 10 Watts maximum at 999, 0 Watt at 100 and can be calculated as
Watts=(HEAT-100)/ 90. Temperature will be a complicated function of the HEAT value,
time, the glass size and other factors.

The next parameter is the FILAMENT. The FILAMENT selects the scanning
pattern of the laser beam. The range of this parameter is 0-15, which deﬁnes the length
of the scanning pattern in millimeters; a value of 1 would produce a 1 mm seaming
length, a value of 2 would give a 2 mm scanning length, etc. Figure 3.9 shows an

example of the display where the parameters are entered [24].

35

 

 

HEAT FILAMENT VELOCITY DELAY PULL
600 4 5 5 l 28 70

 

 

 

 

 

 

 

Figure 3.9 Sample of the display where the program is entered into the P-2000.

The VELOCITY speciﬁes a certain glass temperature to trigger the hard pull.
The rate at which the clamps move apart is measured as the glass softens and begins to
stretch. This increasing velocity is related to the viscosity of the glass, which is a
function of the glass temperature. The range of this parameter is 0-255 with useful values
falling between 10 and 100, with units of millivolts of transducer output. If the
VELOCITY value is too high, the glass will separate before the pull [24].

The PULL is the parameter that controls the force of the hard pull on the glass.
The total pull force acting on the glass comes from a slight gravity pull and the force
programmed into the micropipette puller. Thus, the total pull force is made up of these
two parts.

The glass carrier bars (Figure 3.10) are attached to cables that are brought by
pulleys to a common point of attachment on the plunger of a solenoid, which is the
source of the weak gravitational force mentioned above. When the glass tube is loaded
into the carrier bars, the bars are brought close together, lifting the plunger. The weight
of this plunger along with all of the components is approximately 0.171 kg [24].

The programmable force is produced by passing current through the coil of the
solenoid. The current is turned on at after the velocity value has been reached. The
current that is supplied to the solenoid is based on the PULL value; a value of 255 sends

2.5 amps. For a pull value of 221, the load applied is 53 lb.

36

The DELAY parameter dictates the amount of time between the shutting off of
the heat (the C02 laser), and the initiation of the hard pull. This allows for time to allow
the glass to cool a bit. Using a high delay value will generally result in a shorter taper
length and a larger tip diameter. The range for this parameter is 0—255 and the units are
milliseconds. The default setting for this parameter is set at 128. At this value, the hard

pull will activate at the same time as the deactivation of the laser beam. If a value greater

 

Figure 3.10 Image of the carriage system of the P-2000.

than 128 is used, the hard pull will be initiated after the laser turns off. Similarly, if a

value less than 128 is chosen, the hard pull will begin before the deactivation of the laser

[24].

37

Much of the early experimentation with the P-2000 involved testing how each
parameter affected the size and shape of the tip. This testing began on pipettes without a
metal wire inside. Four of the parameters would be held constant while the ﬁfth was
varied throughout its range of values. After obtaining a rough feel for each parameter,
several different metals were selected to use as the electrical conducting core for the
probe. T he metals were chosen based on their physical properties and what had been

experimented with in previous studies.

Trip Point

Velocity

 

1
Heat Heat Value$

 

 

 

 

Pull // +11 sec—«1 . Pull Value

 

 

Delay Value (128 msec)

 

 

Figure 3.11 Pull cycle when a DELAY greater than 128 is selected [24].

The VELOCITY and DELAY parameters dictate the timing of the pull as
illustrated in Figures 3.11 and 3.12. The VELOCITY determines the point at which the
laser turns off. The DELAY is the time between the laser deactivation and the PULL

activation. Therefore, if no DELAY is used in a given program, the pull will initiate as

38

soon as a certain viscosity is reached. Likewise, if a DELAY of less than 128 is used, the

pull will begin before the laser is deactivated.

Trip Point

Velocity

 

1 Delay Value (128 msec)
Heat Value /

Heat 1 _‘
Pull 1 "Bee ->|

 

 

 

 

 

 

 

Figure 3.12 Pull cycle when a DELAY less than 128 is selected [24].

3.3 Micro Probe Fabrication Method

Before extensive testing with the metal wire began however, a manipulation of
the glass tube was needed. As seen in the sketch, ﬁgure 3.13, a large dimensional gap
exists between the inner wall of the glass tube and the metal wire. The micropipette
puller is optimized for pulling glass tubing that has an outer diameter of 1.0 mm.
Through the preliminary experimentation, capillary tubing with 1.0 mm outer diameter
and 0.5 mm inside diameter was settled on, although silica tubes with 1.0 mm outer
diameter and 0.3 mm inner diameter were tested as well, with varying degrees of success.

Inserting a 50 pm wire inside of a tube with 0.5 mm inner diameter leaves a signiﬁcant

39

gap between the wire and glass.
A problem arises for several reasons because of this dimensional mismatch. First,
air is an excellent insulator of thermal energy. If the air is absorbing most of the energy

from the laser, the wire will not receive enough energy necessary to soften and alter its

\\\\\\\\\\\\\ \

4

\\
‘ \\\\\\\\\\\

///

Glass tube .
Metal w1re

 

Air gap

/ W
//
’wﬂ

Figure 3.13 Sketch of cross-sectional view of the dimensional mismatch between the
inner wall of glass capillary tube and the metal wire.

mechanical properties. When the force of the pull is exerted on the wire without
sufﬁcient heating, the wire will snap and result in two ﬂat ends rather than a gradual taper
to a ﬁne point. The ﬁrst step would be to reduce the dimensional mismatch between the
inner wall of the glass tube and the metal wire. Figure 3.14 shows one of the several
results that can occur from bypassing the ﬁrst step.

In the article by Fish, et a1. [9], they recommend breaking the pulling procedure

down into two or three steps. The ﬁrst step is to pull the glass down far enough that the

40

inner diameter of the tube is the same as the diameter of the wire (Figure 3.15). A

speciﬁc program was designed to perform this task.

end of i '
metal wire ,

 

Figure 3.14 Example of microprobe that was pulled with the wire breaking
prematurely. As seen, the tapered glass extends far beyond the point where the wire
ended.

One program, designed for 1.0 mm outside diameter, 0.5 inside diameter
borosilicate glass and 0.05 mm gold wire, called for
0 Heat — 400
o Filament — 4
0 Velocity — 50
0 Delay — 200

0 Pull — 20

41

This program was designed for 1.0 mm outside diameter, 0.5 mm inside diameter
borosilicate glass and 0.05 mm gold wire. The DELAY value of 200 gives the glass 72
msec to cool before the hard pull begins. This means the glass won’t ﬂow so easily and
will regain some of its mechanical strength, leaving it in tact. Also, a smaller PULL
force allows the inner diameter of the tube to reach the wire and then stop.

It must be noted however, that instances such as the one depicted in ﬁgure 3.14
can occur even if the ﬁrst step is carried out. Executing the ﬁrst pulling step decreases
the frequency of such events to a certain extent and the only probes deemed “successful”
were a result of including the ﬁrst pulling step. Figure 3.16 depicts a cross sectional view
of the middle portion of ﬁgure 3.15 where the glass is collapsed around the wire.

The second step is simply a heating the glass/metal combination for several
seconds to provide better connection between the glass and metal. It is important that the
glass and metal move together when they deform. That way, the amount of tapering of
the metal wire is maximized. However, through signiﬁcant experimentation, it was

concluded that this step is not always necessary.

 

42

Figure 3.15 Glass tube collapsed around the metal wire as the ﬁrst pulling step in the
probe fabrication procedure [10].

lass
g metal

 

Figure 3.16 Cross-sectional view of middle portion of Figure 3.15. The glass is pulled
down so that the air gap that existed between the glass and the metal is either minimal or
eliminated completely.

The third step in the procedure is the ﬁnal pulling step, which is the most difﬁcult
to program. The desired result will look similar to the sketch of Figure 3.17. After the
ﬁnal step, if the pull appeared successful (the metal wire was continuous to the very end
of the tip), the very tip of the glass was etched with hydroﬂuoric acid (HF).

The preliminary observation of the probe was done under an optical microscope,
making it extremely difﬁcult to see the tip of the probe, even more difﬁcult to determine
if there was a thin coating of glass covering the metallic tip. Even the thinnest coating of
glass could render the probe useless, particularly if the use of the probe requires that it
conduct electricity. Therefore, the tips of the pipette were given an HF bath to remove

any glass that may have covered the metal.

43

 

Figure 3.17 Illustration of fabricated micro probe as a result of the pulling procedure
executed by the P-2000 [10].

The HF bath consisted of one part 50% HF and six parts H20. This gave a
solution with a glass-etching rate of approximately 120 nm/min. A bath of about two
minutes was adequate for most of the pulled probes, yet sometimes the probes were
exposed to the HF for longer periods of time to ensure an appropriate amount of etching
would occur. Only the tip was dipped into the HF bath.

Unfortunately, there was no direct way to observe if all of the glass had been
removed from the tip of the probe. The only way to determine this was by viewing the
sample in a scanning electron microscope (SEM). If there was no charging at the tip, it

was assumed that the glass had been removed.

3.4 Probe Measurement Technique

Once fabricated, the probes were viewed under a SEM to determine if indeed the metal

wire was continuous to the very tip of the pipette and all of the glass had been removed

44

during the HF etching procedure. If so, the probe was integrated into a four-point probe
arrangement.

In a traditional four-point probe set-up, there are two leads for current and two
electrodes for voltage drop measurement. This is demonstrated in Fig 3.18. In this
conﬁguration, the probes that measure the voltage drop are held at a constant spacing, I,
from each other. This spacing allows for the calculation of electrical conductivity in the
sample from the expression

_11

01—, 3.2
VA ( )

where, 0' is the electrical conductivity, 1 is the amount of current passed through the

sample, V is the voltage, and A is the cross-sectional area of the sample.

V- V,

metal contact V \ /

 

 

  

Figure 3.18 Schematic of traditional four-point probe arrangement.

What we altered as part of our project was moving V- to the electrode that I- is
using. Then, we mounted V+ (our fabricated microelectrode) onto a xyz translational

stage so that it is free to scan the entire surface of the sample. The current leads were

45

connected to the sample through metal contacts on the sample. These contacts were put
on the sample through sputter coating, metal evaporation, or electrochemical deposition.
This creates a contact resistance at these points. Because our probe is free to translate,
the contact potential at the metal contact can be measured, and then added to the
measured voltage drop. This is advantageous over the traditional four-point probe
because of this ability to measure the contact resistance.

Figure 3.19 shows a schematic of the scanning one-probe setup. The current
leads are connected to a Keithley 2400 source meter, which was set to deliver 10 mA
across the sample. The voltage probes were run to the Keithley 2182 nanovoltmeter to

record the voltage drop across the sample [10].

  

metal contact

7 1+

 

 

Figure 3.19 Illustration of one-point scanning probe setup. The micro probe is mounted
on a xyz transnational stage and acts as the positive voltage probe. This probe is free to
scan over the entire surface of the sample.

The basic methodology of the experiment using this arrangement is to ﬁrst

measure the resistance one end of the sample where the contact has been deposited.

Then, the scanning probe is moved toward the opposite end of the sample and the voltage

46

drop changes in accordance with the increased probe spacing between V. and V +. This

procedure will be discussed further in the Results portion of the write up.

47

4. RESULTS

4.1 Probe Fabrication

Microprobes were fabricated from glass capillary tubes and metal wires. The goal of
the fabrication process was to create a probe that had as small of a tip as possible.
Desirable results of the tip diameter were on the order of 1 pm and below. Also, it was
imperative that the wire be continuous to the very end of the tip, for purposes of
conducting electricity. The results for the fabrication of micro probes from gold and
platinum showed promise. Some of the other metals, however, proved more difﬁcult to
create the probe tips that were of the desired tip diameter.

The design of the pulling procedures will be presented in this chapter, along
with results of the probe fabrication and the conductivity measurements done with
platinum microelectrodes. The probes were integrated into a variation of the four-point
probe conﬁguration, where the micro probe acted as a scanning electrode. Conductivity
measurements were conducted across the surface of the sample to test the effectiveness
of the fabricated micro probes as electrodes.

Images of the probes will be presented along with the pulling procedure for
each. Each pulling procedure consists of three steps and produces two micro probes.
The ﬁrst step, after the wire is inserted into the glass tube, is to collapse the glass down
around the wire (Figure 3.15). The second step is a heating only step, to ensure a good
mechanical bond between the glass and metal. The third step is the ﬁnal step, where the

heat and pull are such that the glass and metal are pulled apart (Fig. 3.7).

48

4.1.1 Critical Volume Fraction

Will the glass matrix and metal wire behave and deform together as a composite
material? These are two materials that have distinctly different mechanical behavior,
yet in the proper volume fractions, the desired mechanical behavior for the composite is
possible. There is a critical volume, V”, where the plasticity of the composite material
becomes equal to the plasticity of the matrix [9]. The volume fraction, A, of the wire

can be deﬁned as
A = 1’1, (4.1)

where df is the diameter of the wire and dm is the outer diameter of the glass tube.

It is actually the thickness of the glass tube that is of signiﬁcance. The article
from The Hebrew University of Jerusalem [9] describes a particular glass-metal
combination they used to make microprobes. Using 0.08 mm platinum wire, annealed,
and 1.2 mm outer diameter, 0.3 mm inner diameter, they executed the step to collapse
the glass tube around the metal wire. Therefore, the inner diameter of the glass
becomes 0.08 mm and the outer diameter of the tube becomes 0.32 mm, as can be seen
in Fig. 2.2.

For this particular glass-metal combination, the authors reported pulling tips as
small as 50 nm in diameter. They state that this is possible for tube and wire and glass
dimensions where 0.2 < A < 0.4. For the dimensions mentioned above, A = 0.25. Of
course, there is a speciﬁc pulling program that must be designed for this glass tube and

metal wire composite. This same critical volume value would not necessarily work the

49

same if the glass used was silica rather than borosilicate, or if the metal was not
platinum, but silver instead.

This critical volume fraction ratio simply gives a starting point, helping to
determine the dimensions of the glass and metal that can work well together in the
fabrication procedure. At the critical volume fraction of the ﬁber, as stated above, the

plasticity of the matrix represents the plasticity of the entire composite [9].

4.1.2 Fabrication of Microprobes from a Borosilicate Glass Tube and Pt Wire

This combination of glass and metal had a very large mismatch in physical properties.
The melting point of platinum is above 1400° C whereas the melting point of
borosilicate glass is closer to 1100° C approximately. This made the ﬁrst step very
simple; the borosilicate glass was easily pulled down around the platinum wire without
the wire weakening at all. In the second step however, because the glass is already
weakened, the high heat needed to melt the platinum wire inside was too much for the
glass to handle.

Using 1.0 mm outer diameter, 0.5 mm inner diameter along with 0.05 mm
platinum wire, the glass was pulled down around the wire. Then several different
programs were experimented with, all of them resulting in poor results. Heat values
ranging from 600 — 900 and each value of ﬁlament was tested. The pull values were
leﬁ small, as well as the velocity. The delay values were kept high, to hold intact the
integrity of the glass as long as possible.

What was found in every case besides those where the heat was set to a very

high value (above 800) is that the glass would separate but the wire would not. In the

50

cases of high heat, the combination of glass and metal would pull, but it would not form
a small tapered tip that could be useful as a submicron probe. Instead it would usually
form a ﬂat, broad tip.

Further experimentation with this combination could lead to more promising
results, as Nanonics and the Hebrew University of Jerusalem reported fabricating
similar probes with these two speciﬁc materials and the Sutter P2000 [9]. However,
Nanonics used speciﬁc dimensions where the volume fraction of the composite was
between 0.2 and 0.4. Under these conditions, they pulled probes with tip diameters of

50 nm.

4.1.3 Fabrication of Microprobes from a Silica Tube and Pt Wire

The melting points of platinum and silica are signiﬁcantly closer than that of platinum
and borosilicate. Therefore, this combination was explored a bit more comprehensively
than the glass/metal combinations that have been discussed already, even though it
seems that the relative dimensions of the matrix and ﬁber are more important than their
respective physical properties. Some of the potential applications for the probes call for
service temperatures above the soﬁening point of borosilicate glass. For certain
applications, then, silica would be required as the matrix material.

The ﬁrst trial was carried out with 1.0 mm outer diameter, 0.3 mm inner
diameter silica glass and 0.08 mm diameter annealed Pt wire. The 0.2 < A < 0.4
guideline [ultrafast] was not purposely followed in this case because we were no longer
dealing with borosilicate glass. However, for these dimensions, the volume fraction of

glass to metal was 0.3, and promising results were obtained.

51

For this pulling procedure, ﬁrst the platinum wire was inserted into the quarts

tube. Then, the following program was used to pull the tube down around the wire:

0 Heat — 720

o Filament — 5

0 Velocity 1 50

0 Delay ~— 250

0 Pull ~— 145
This program used a high heat value to soften the silica, and a high delay value (122
msec before the pull was initiated) to give the glass a moment to recover and not
separate completely, but instead only stretch. A high pull value was required as well to
stretch the silica. The result of this step is seen in ﬁgure 4.1.

Because of the geometry of the tube and wire, this step only had to be run one
time (the tube did not need to be stretched very extensively because the inner diameter
of the tube was 300 microns and the diameter of the wire was 80 microns, therefore
only about a 200 micron gap had to be closed between the glass and the metal).

This was enough to collapse the glass around the Pt wire and eliminate most of
the gap that existed between the inner diameter of the tube and the Pt wire. Then, the
program was executed to heat — but not pull — the glass and metal. This step was
executed in an attempt to achieve a stronger bond at the interface of the two materials.
After this step was completed, the ﬁnal pulling step was initiated. This step called for
the following program:

0 Heat — 650

o Filament — 5

52

0 Velocity — 50
0 Delay— 100

0 Pull — 100

 

Figure 4.1 First pulling step of 0.3 mm inner diameter silica tube and 0.08 mm Pt wire.

A high heat value was employed for this step to soften the silica and the Pt so
they would taper ﬁnely during the hard pull. The pull will activate for 28 msec while
the laser is still on, then the laser will turn off as the pull continues. Therefore, part of
the pull is taking place while the glass and metal are at their hottest points. This should
be the best time for the pull to take place because both the glass and the metal should be

in their most readily deformable states.

53

This program yielded the result depicted in Figure 4.2. The diameter of the tip
was less than 20 microns, relatively speaking not a desirable outcome. However, the
outcome of this pull exhibited a very nice taper, and encouraging sign for developing a
more sophisticated pulling procedure.

The silica matrix will not only allow the probe to be utilized in a high
temperature environment, it will also be more robust, as silica is a stronger glass than
borosilicate. This probe would have a tremendous advantage over a similar probe that
used borosilicate glass. Nowhere in the literature were probes made from silica and Pt
found. However, a better taper is required in order for this probe to be of signiﬁcant

value.

 

Figure 4.2 Result of pull with 0.3 mm inner diameter silica tube and 0.08 mm diameter
Pt wire. The diameter of the tip is approximately 0.02 mm.

54

Several variations of this program were investigated. Another tip that was
produced was fabricated using the same initial step to collapse the glass around the
wire. The second step was altered however, using a higher heat, lower ﬁlament, same
velocity, higher delay, and smaller pulling force. This was done to give the metal wire
a slightly longer time to soften so the tip would have a smaller diameter. The program
goes as follows:

0 Heat — 850
0 Filament 1 4
0 Velocity — 50

0 Delay — 135

0 Pull — 80

 

Figure 4.3 Image of probes made of 0.3 mm inner diameter silica glass and 50 um Pt
wire.

55

This program yielded another good result as can be seen in Figure 4.3. Many programs
leading up to this image produced poor results. Each program was tweaked and tuned
until a good result was obtained. A good result is one where the taper is ﬁne and the
wire is pulled to the very end, as seen in Figure 4.4. A 50 um Pt wire was used for the
probes seen in Figures 4.3 and 4.4 along with 1.0 mm outer diameter, 0.3 mm inner

diameter silica tubing. The critical volume fraction then became 0.3.

 

Figure 4.4 Zoomed in view of one of the probes depicted in Figure 4.3. The diameter
of the tip is less than 5 pm.

This result for Pt wire in silica glass was very promising and with more variation
of the program, a smaller, more ﬁnely tapered tip could result. This was an

improvement over the result obtained in Figure 4.2

56

4.1.4 Fabrication of Microprobes from a Borosilicate Glass Tube and Au Wire

This glass and metal combination worked the best for creating micro probes. The
dimensions of the materials were 1.0 mm outer diameter, 0.5 mm inner diameter for the
borosilicate tube, and 0.05 mm diameter gold wire. The pulling procedure, results of
the pulling procedure, and the measurements taken using the probe will be presented in
this section. The volume fraction for this glass-metal composite is 0.5.

The fabrication procedure for gold wire and borosilicate glass followed the same
initial step as the procedure described previously. The wire was inserted into the not yet
pulled borosilicate glass tube, then the tube with wire inside were placed into the P2000
micropipette puller. The following program was put into the puller to achieve the
results seen in Figure 4.5:

0 Heat —- 400

0 Filament 1 4
0 Velocity — 50
0 Delay — 200
0 Pull — 20

This program was executed two times to stretch the tube so that the inner
diameter of the tube was the same as the diameter of the gold wire. Again, a high delay
was used to allow for recovery of the softened glass for a brief time period so that only
stretching of the tube occurs, and not separation. A small pull value was employed as

well so the glass will not pull apart.

57

gold wire

 

Figure 4.5 Initial pull that shows a borosilicate tube pulled down around a gold wire.

At this point a different step was executed that had not been used in any of the
previous pulls. One side of the tube was unclamped and the entire pipette is shifted
over a couple of millimeters and then clamped again. Figure 4.6 illustrates where the
laser would hit before the pipette was shifted and approximately where the laser beam
strikes the pipette after the shift. Figures 4.7a and 4.7b show the difference of the
results of the ﬁnal pull with and without the shift.

The shift was done right before the ﬁnal pulling step and our best results were
obtained when this step in the pulling procedure was employed. The ﬁnal pull program
was:

0 Heat 1 370

o Filament — 4

58

0 Velocity — 40

0 Delay — 0

0 Pull — 65
A smaller heat value than what was used in the initial pull was used because the gold
wire touched the borosilicate glass, allowing more heat to transfer directly to the wire.
A delay value of zero was used so the pull will initiate as soon as the laser activated. A

larger pull value was used so the probe tips tapered down extensively. It is not known

Laser beam
before shift

 

Laser beam
after shift

Figure 4.6 This ﬁgure illustrates where the laser beam would strike the glass normally
and where it strikes the tube after it has been shifted over.

exactly why shifting the glass over a few millimeters makes such a dramatic difference

in the outcome of pulling procedure. Most likely, the gold wire is well gripped by the

59

glass to the right of where the laser beam strikes, and the result is a very ﬁnely tapered

tip with gold continuous to the very end.

 

Figure 4.7 These images show the difference in the outcome of the pull when a). the
glass was shifted alter the ﬁrst step and b) the glass was not shifted after the ﬁrst step.

A magniﬁed look at the result of this pulling procedure was done using scanning
electron microscopy (SEM). The tips were etched with HF to expose the metallic tip
before they were viewed in the SEM. Several different images will show the variety of
results that each fabrication procedure can yield (each using the fabrication procedure
described above, with the shift after the ﬁrst step). Figure 4.8 shows a relatively ﬂat
and straight tip. There is not, however, extensive tapering at the end of this probe, as
there is in Figure 4.9. Interestingly, the tip in 4.8 is ﬂat, the tip in 4.9 is sharp, and the
tip in 4.10 is rounded.

Each probe in Figures 4.8, 4.9, and 4.10 was pulled with the same pulling
program that included the glass shift. It is clearly seen how each pull varies, especially
at the micrometer level. Each tip has a different outer diameter, and shape. There are
several possibilities as to why each pulled tip has such variation, including how the

glass and metal were mounted in the micropipette puller (wire was kinked inside glass

60

tube), contaminants on the surface of the glass during the pull, minute differences in the
pull force, etc. In an extremely sensitive fabrication process such as this, the smallest
variations can lead to dramatic changes in the result.

The sizes of each pulled probe vary as well. Figure 6.13 shows an outer

diameter

 

Figure 4.8 SEM image of gold probe pulled using Sutter P2000 micropipette puller.

61

of roughly 3-4 microns. Figure 4.9, on the other hand, clearly shows more tapering at

the tip of the probe. From the scale on the micrograph, it is estimated that the tip of this

 

Figure 4.9 SEM image of gold and borosilicate probe.

particular probe has a diameter of approximately 500 nm. It is not clear in this image,
however, if the Au is exposed at the tip. The bottom half of the probe, the brighter

portion of the image, is borosilicate glass. Some results clearly show that the glass was

62

removed from the tip during the hydroﬂuoric etches. Figure 4.10 is an instance where
the acid etch was not even needed. The gold wire protruded from the glass as a result of
the pull. The diameter of the tip of this probe was approximately 450 nm. These

images represent the best of the micro probes fabricated with the micropipette puller.

 

Figure 4.10 SEM image of gold micro probe.

It is quite interesting to see in this image that the ﬁnal fracture of the glass is
ﬂat, although it tapered quite extensively. This type of ﬁacture is advantageous because

no hydroﬂuoric etch is required.

63

4.2 Analysis of Probe Fabrication

The elastic modulus can be calculated for the composites experimented with and
compared to composites used in the literature that produced successful results.
However, the elastic modulus doesn’t reveal much information about the
therrnomechanical properties of the composite. Therefore, the viscosity of the
borosilicate-gold composite will be investigated to examine how certain parameters will
affect the viscosity of the composite during the pull. A low viscosity is desired, as a

lower viscosity during the pull will mean the composite will ﬂow with greater ease.

4.2.1 Calculations of Modulus of Elasticity

The elastic properties of the glass-metal composites is of interest in determining what
the dimensions of the glass tube and metal wire should be to optimize the pull to
achieve a sharp taper and small tip diameter. The calculations are made for the stage
when the glass is pulled down to the wire diameter, as seen in Figure 4.1. Comparisons
of the dimensions used and resulting modulus of elasticity values of the Nanonics
probes [9] will be compared to the modulus of elasticity values of the probes fabricated
in this research effort.

First, the elasticity for the borosilicate glass and platinum wire composite
described in Fish et a1. [9] will be calculated. The dimensions for the components after
the glass capillary has been pulled down around the metal wire are as follows:

0 Borosilicate — 0.32 mm outer diameter and 0.08 mm inner

diameter

0 Platinum — 0.08 mm diameter

64

0 Platinum 1- 0.08 mm diameter
The modulus of elasticity for the borosilicate and platinum are 64 GPa and 171 GPa

respectively. The modulus of elasticity for the composite is then (Eq. 2.7)

__ [(64.0GPa 9.6 -10‘8 m) + (171.0GPa - 6.4- w" m)] :9

E, _7
1024-10 772

 

EC = 60.0 GPa + 10.687 GPa = 70.687 GPa
For the platinum and quartz composite designed for this research, the dimensions are:
0 Silica — 0.267 mm outer diameter, 0.08 mm inner diameter
0 Platinum — 0.08 mm diameter
The modulus of elasticity for the silica and platinum are 72.9 GPa and 171 GPa
respectively. These values can be put into the equation to determine the modulus of

elasticity for the composite.

_ [(72.9GPa -6.488 110‘8 m) + (171.0GPa - 6.4 110'9 m)] :>

EC
7.1281104»:

 

EC = 66.354 GPa + 15.352 GPa = 81.706 GPa
For the gold and borosilicate composite designed for this research, the dimensions are:
0 Borosilicate — 0.1 mm outer diameter, 0.05 mm inner diameter
0 Gold — 0.05 mm diameter
The modulus of elasticity for gold is 77.2 GPa. These values can similarly be
calculated using the Rule of Mixtures to ﬁnd the overall modulus of elasticity for the

composite.

E _ [(64.0GPa - 7.5 ~10’9m)+(77.ZGPa-2.5-10'9m)] 3
C 1.0-10’8m

 

EC = 48.0 GPa + 19.3 GPa = 67.3 GPa

65

These are the values for the modulus of elasticity for the three glass-metal composites

analyzed.

4. 2.2 Viscosity

The viscosity can be calculated for a glass-metal composite using Eq. (2.10). I f the
viscosity can be determined, it could help in predicting what values of parameters to
input to obtain optimal results. Theoretically, the lower the viscosity for the composite,
the better it should ﬂow and the better the taper should be. Certain parameters can be
varied, while the others are held constant, to see how each parameter affects the
viscosity of the pipette pull.

To calculate the viscosity for the composite, the applied stress, the original
length, the strain rate is needed. The initial length of the pulled down portion of the
glass and metal is measured after the ﬁrst pulling step, and the change in length is
measured after the ﬁnal pull. Also, the P-2000 displays how long it took to execute the
pulls; therefore the strain rate can be calculated. The diameter of the glass and metal is
known after the glass is collapsed down around the metal wire, and the force is known
from the PULL parameter. Thus, the viscosity for each pull can be calculated from Eq.
(2.10). Table 4. 1 lists several values of viscosity where the HEAT value on the
micropipette puller is varied while the other parameters are held constant.

Figure 4.11 is a plot of these data points and it can be clearly seen that HEAT
values between 365 and 375 result in the lowest viscosity for the composite. The other

parameters were

0 Filament — 4

66

0 Delay — 0
0 Pull - 70

This is nearly the same program that was used to fabricate the probes of Figures 4.8 1—

4.10.

Table 4.1 Table of values where viscosity is calculated for different HEAT values.
Each of the other parameters was held constant.

 

 

 

 

 

 

 

 

 

 

 

 

HEAT AL (mm) Time (sec) (niArfrisAec) Viscosity (Pas)
315 0.00943 1.82 0.00518 6.29E+09
335 0.00863 1.41 0.00612 5.33E+09
355 0.00817 1.18 0.00692 4.71E+09
365 0.01358 1.12 0.01213 2.69E+09
375 0.0138 1.14 0.01211 2.69E+09
395 0.00846 1.04 0.00813 4.01 E+09

 

 

Viscosity was determined for pulls where the VELOCITY parameter was
increased as the other parameters where held constant. The VELOCITY parameter
dictates at what level of glass viscosity should the hard pull initiate. Table 4.2 presents
the calculated viscosity that is a result of changing VELOCITY values.

The lowest viscosity value was obtained at a VELOCITY value of 40. The
other parameters were set to:

0 Heat — 365
o Filament—4
o Delay—0

0 Pull — 7O

67

Viscosity vs. HEAT

‘ + Viscosity
| _ _ _,

___ ,__1‘, 7

 

Viscosity (*10‘9) (Pas)

 

 

 

 

300 320 340 360 380 400

Figure 4.11 Plot of the data in Table 4.1. Values of HEAT between 365 and 375 result
in the lowest viscosity of the composite.

Table 4.2 Values of viscosity for different values of VELOCITY.

 

 

 

 

 

 

 

VELOCITY AL (mm) Time (sec) (135:6) Viscosity (Pas)
20 0.01028 1.31 0.00785 4.1513+09
40 0.014 1.21 0.01157 2.82E+09
60 0.0081 1.28 0.00633 5,155+09
80 0.01017 1.36 0.00748 4.36E+09
100 0.01035 1.23 0.00841 3.87E+09

 

 

 

 

 

68

 

Figure 4.17 shows the trend of this data, where it can be seen that a VELOCITY value

of 40 results in the lowest viscosity and, consequently, the best pulling result.

~ -—)|(— Viscosity

 

 

Viscosity(*10"9) (Pas)

 

 

 

O 20 4O 60 80 1 00 1 20
Velocity

Figure 4.12 Plot of viscosity versus VELOCITY for Au/Borosilicate microprobe
fabrication.

4.3 Conductivity Measurements

Conductivity measurements on different samples were carried out using the probes
fabricated with the micropipette puller. The probe was mounted on a xyz translation
stage and was used as one of the voltage probes. A sample of bismuth telluride
(BCzTC3), a thermoelectric material, was scanned across its surface to obtain the results

of its electrical conductivity. This sample was measured using a platinum/silica

69

microprobe as seen in Figures 4.3 and 4.4. The results of the measurement will yield

the contact resistance and the electrical conductivity of the sample.

4. 3.1 The One-Point Scanning Probe Method

The one-point scanning probe conﬁguration is depicted in Fig 4.13. This illustration
how the micropipette-based probe was integrated into the altered four-point probe set up
as a seaming probe. The micro probe was mounted on a xyz translation stage so that
the conductivity could be scanned across the surface of the material being tested.
Contacts were made on the sample by electrochemical deposition of nickel on the sides.
The current was supplied through these contacts and the low voltage was measured at

that point as well. The scanning probe was used as the high voltage probe.

 

 

 

 

 

 

 

V' I.
1' metal
contact
/
/ é
/ /
/ <— —> /
/ Sample /
WV 1

Figure 4.13 Scanning one-point probe conﬁguration for conductivity measurement on a
given sample.

70

An advantage of the method used, consider it a one-point scanning probe, is that
the probe spacing is not held constant. This allows for the conductivity measurement of
the sample along with the measurement of the contact resistance between the sample
and the deposited metal electrodes.

Therefore, the scanning one-point method is simply a variation on the four-point
probe technique, which allows for the measurement of the contact resistance by

measuring the voltage drop at the contact.
4. 3.2 Electrical Conductivity Measurements Using Pt Microprobes

Platinum micro probes were integrated into the one-point scanning probe conﬁguration
(Fig 4.13). Figure 4.14 is a plot of voltage difference versus change in probe spacing
measured on a sample of bismuth telluride (BizTe3), a thermoelectric material. The
measurement was done using one of the Pt electrodes similar to the one seen in Figure
4.4 to demonstrate the effectiveness of the probes. The value for the conductivity for
the sample can be extracted from this data using the formula

[Al
0 =—,
AVA

(4.2)

where I is the applied current, Al is the distance between the probes, AV is the measured
voltage difference, and A is the cross-sectional area of the sample. Below (Table 4.3) is
data obtained from an AC measurement on BizTe3 using a Pt microelectrode. The
contacts were made by electro deposition of nickel.

A lock-in ampliﬁer was used to supply the oscillating sinusoidal signal with a 1

Q resistor connected in parallel to the sample

71

Table 4.3 Collection of electrical conductivity data of BizTe3 using the scanning probe

conductivity technique [10].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Probe Spacing (cm) A V (mV)

0 0.05
0.025 0.051
0.05 0.052
0.075 0.055
0.1 0.057
0.125 0.059
0.15 0.061
0.175 0.063
0.2 0.064
0.225 0.065
0.25 0.069
0.265 0.071
0.3 0.075
0.325 0.075
0.35 0.079
0.375 0.082
0.4 0.084
0.425 0.086
0.45 0.099

 

 

This data is plotted in Figure 4.14. A large jump in voltage difference is seen
between the last two data points. The contact resistance between the nickel-plated
contact and the sample can be identiﬁed as it accounts for the large jump in AV, from
0.086 mV to 0.099 mV, when all previous values for AV were approximately 4 mV.
Thus, simply taking the difference between these last two data points and dividing it by
the applied current can determine the contact resistance for this situation.

_ 0.099(m V) — 0.086(m V)

Rconmct _
9.76(mA)

=1.33m§2

 

72

 

 

 

—O— Voltage difference (mV)
----- slope (dV/dl)

 

0.1

0.09
S‘

E 0.08
8
C
o

a“; 0.07
.35
‘O
o
a)
.9

£5 0.06

0.05

0.04

 

812Te3 AC conductivity measurement using
Pt microelectrode

 

i

ITWITT

IlTrrfT

l

 

I T T I I I l’ I Y I I l I I T I I T 1 Y‘ﬁ I i T 1,.) i V

----- y = 0.048099 + 0.086852x R= 0.99517

 

I
I

 

1 J J l l l L i 1 l L A l l l I l L 1 l l l L l l l l l l

 

-0.1

0 0.1 0.2 0.3 0.4 0.5

probe spacing (cm)

 

 

Figure 4.14 Plot of AC conductivity measurement on BizTe3 sample using Pt micro

probe [ 10].

The slope of the plot in Figure 4.14 represents the change in voltage difference

over the change in probe spacing and Eq. (4.2) can be written as

1

(AK) . A
A1

0' = (4.3)

by dividing the top and bottom by Al. The current used for this conductivity

measurement was 9.76 mA and the cross-sectional area of the sample was 0.103 cmz.

73

Therefore, the electrical conductivity can be obtained by placing these values into Eq.

(4.3).

0 _ 9.76(mA)
0.08968(mV/cm) - 0.103(cm2)

 

=1058.72(S/cm)

The accepted value for the conductivity for BlzTCj, lies between 800 and 1200 S/cm

[10].

74

 

5. CONCLUSION
5.1 Submicron Scanning Probe Measurement System

Submicron probes were created from glass capillary tubes and gold and platinum metal
wires using a CO; laser-based micropipette puller. Probes with tip diameters of less
than 500 nm were fabricated using methods developed during this research, and
integrated into a scanning one-probe measurement system. This system features the
ability to measure the electrical conductance of a sample along with the contact
resistance between the sample and the current lead. Another feature of the system is the
potential to measure low-scale dimension samples. The probe system was successfully
tested on a 0.231 cm2 BizTe3 sample using a Silica/Pt probe as the scanning probe.
However, with probe tips less than 500 run, much smaller samples could be measured.
The viscosity of the glass-metal composite can be determined for any of the
settings of the P2000. This property was investigated, and for the borosilicate-gold
probes, the program developed for the fabrication yielded the lowest viscosity, and

consequently the most uniform tapers and smallest tip diameters.

5.2 Suggestion for Future Work

5. 2. I Mount Probes in Microprobe Station

The measurements done for this research were carried out under a low
magniﬁcation optical microscope and a simple translational stage. This led to several
broken probes, as they are fragile and deform under relatively small levels of pressure.

Our research group recently acquired a microprobe station similar to the one depicted in

75

Figure 1.2. The station features digital micrometers for translation of the probe, a
vibration-free working surface, a high magniﬁcation optical microscope, a
turbomolecular vacuum pump, and a temperature controller.

The submicron probes could be mounted into this probing station and be used as
a scanning electrode, similar to the conﬁguration used in this research, with much better
manipulation control and visual capability. Much of the glass would have to be etched
away to mount the probes into the station, however. The probes are fed through a small
hole and clamped into place with a setscrew. This setscrew is also an electrical contact
for the probe. If the probe is covered in glass, it will be electrically insulated from the
rest of the system and there will also be difﬁculty clamping it into position. Therefore,
the glass should be etched away approximately 1 em up the probe length. This will

allow for contact with the bare wire both electrically and mechanically.

5. 2.2 Continued Improvement on Probe Fabrication Procedure

Further research should be built on the research that has been accomplished on the
fabrication of submicron probes. Supplementary research into gold and platinum
probes will eventually lead to smaller and smaller probe tips, ideally down to 50 nm.
The programs designed for the micropipette puller can be further manipulated in
attempt to achieve smaller probe tips. Another area of interest lies in developing
programs that will pull silica glass and gold wire together. There is interest in growing
silicon nanowires by pulsed laser deposition on the tips of metal-ﬁlled micropipettes.
This would yield a nanowire that is wieldable and is grown directly onto an electrical

contact, as depicted in Figure 6.1.

76

 

Si Nanowire

gold

/

 

 

 

 

Figure 5.1 Illustration of Si nanowire grown off tip of silica-gold submicron electrode.

77

APPENDIX A

Below are tables of temperature dependant properties of platinum and platinum

alloys.

Table A.1 Temperature dependant properties of pure Pt [20].

 

 

 

 

 

 

 

 

 

Temp. (6C) Elastic Modulus Mod. Of Rigidity Poisson's

(GPa) (GPa) Ratio
25 164.6 54.2 0.518
200 159.3 52.9 0.506
400 153.3 51.1 05
500 149.1 50 0.491
600 145.6 48.9 0.489
700 141.9 47.7 0.487
800 137.8 46.6 0.479
900 132.7

 

 

 

 

 

 

Table A.2 Temperature dependant properties of Pt 10%Rh [20].

 

 

 

 

 

 

 

 

 

 

 

Elastic Modulus Mod. Of Ri 'dit Poisson's
Temp' (°C) (GPa) (GPa)g1 y Ratio
25 212.6 78 0.363
200 206.3 75.4 0.368
400 197.9 72.1 0.372
500 193.3 70.5 0.371
600 188.7 68.7 0.373
700 183.9 66.9 0.374
800 179.2 65.2 0.38
900 175 63.4
1000 169.7
1100 164.9

 

 

 

 

 

 

78

Table A.3 Temperature dependant properties of Pt 10%Ir [20].

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 Elastic Modulus Mod. Of Rigidity Poisson's
Temp' ( C) (GPa) (GPa) Ratio
25 202.3 73.4 0.378
200 196.6 71.1 0.382
400 188.3 68.1 0.382
500 183.9 66.4 0.385
600 178.8 64.8 0.381
700 173.6 62.8 0.381
800 170.7 58.1
900 166.4
1000 162.2
1100 157.1
1200 150.8
Elastic Modulus vs Temperature for Pt Alloys
220 I
Taztolera~44~ ——#Ai
gszooawrk. sea—«iii -
1:1907“ a ~ 9- " . —-— ~ .. v - i-rﬁﬂ
‘.3180~s* a a A-—'- _ we ##‘ﬁ’Pt .
131704.“ ,. a a ~ ~ I i a - lPt10%Rh;,
2160_.___._ ‘— — — — —-—-— — i mom
.3 150 ~— 4 a l r .w— — ~-~—~‘**
3 140 “W ~ - + WW1 T ~ ~ 4
“‘1 130 T A a 4- AAAAAA 9i- 9 a
120 4 , . .
0 200 400 600 800 1000 1200 1400

Figure A.1 Comparison of elastic modulus vs. temperature between pure Pt, Pt 10%Rh,
and Pt 10%Ir [20].

79

10.

ll.

12.

7. REFERENCES

. Robert M. Dowben and Jerzy E. Rose, “A Metal-Filled Microelectrode”,

Science, Vol. 118, pp. 2224,1953.

Bradford D. Pendley and Hector D. Abruna, “Construction of Submicrometer
Voltammetric Electrodes”, Analytical Chemistry, Vol. 62, No. 7, pp. 782-784,
1990.

www.nagonics.co.il

G. Tallarida, “Imaging P-N Junctions by Scanning Near Field Optical
Microscopy, Atomic Force and Electrical Contrast Microscopy”, Materiali e
Dispositivi per la Microelettronica, pp. 1-2.

DJ. McComas and GP. Miller, “Space applications of microelectromechanical
systems: Southwest Research Institute® vacuum microprobe facility and initial
vacuum test results”, Review of Scientific Instruments, Vol. 74, No. 8, pp 3874-
3878, May 2003.

www.janis.com

Lei Huang, Wanjun Wang, and Michael C. Murphy, “Peltier-effect module for
highly localized temperature m anipulations”, R eview of S cientiﬁc Instruments,
Vol. 70, No. 11, pp. 4398-4403, August 1999.

Mun-Heon Hong, Ki Hyun Kim, “Scanning nanolithography using a material
ﬁlled nanopipette”, Applied Physics Letters, Vol. 77, No. 16, p. 2605, 2000.

G. Fish, O. Bouevich, S. Kokotov, K. Lieberman, D. Palanker, I. Turovets, and
Aaron Lewis, “Ultrafast response micropipette-based submicrometer
thermocouple”, Review of Scientific Instruments, Vol. 66, No. 5, pp. 3300-3306,
May 1995.

A. Kubota, “Scanning Probe Electrical Conductivity Measurements Using
Submicro-Electrodes Fabricated by Nonlithographic Methods”, Masters Thesis,
Michigan State University, 2003.

Callister, William D., Materials Science and Engineering: An Introduction, 4th
Edition, New York, John Wiley & Sons, 1997.

Gibson, Ronald F, Principles of Composite Material Mechanics, McGraw-Hill
Inc., 1994

80

l3.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Z. Li, J. Lambros, “Dynamic thermomechanical behavior of ﬁber reinforced
composites”, Composites Part A, Vol. 31, pp 537-547, 2000.

Savitsky E.M., The Inﬂuence of Temperature on the Mechanical Properties of
Metal and Alloys, Stanford University Press, 1957.

Groover, Mikell P., Fundamentals of Modern Manufacturing, Prentice Hall, Inc.,
1996.

G. Banuprakash, V. Katyal V.S.R. Murthy, and GS. Murty, “Mechanical
behaviour of borosilicate glass-copper composites”, Composites Part A, Vol.
28A, pp. 861-867, May 1997.

Robert H. Doremus, “Viscosity of silica”, Journal of Applied Physics, Vol. 92,
No. 12, pp. 7619-7629, August 2002.

wwwinatlsxom

R. L. Williamson, B.H. Rabin, and J.T. Drake, “Finite element analysis of
thermal residual stresses at graded ceramic-metal interfaces. Part 1. Model
description and geometrical effects”, Journal of Applied Physics, Vol. 74, No. 2,
pp 1310-1320, July 1993.

Jurgen Merker and David Lupton, “High Temperature Mechanical Properties of
the Platinum Group Metals”, Platinum Metals Rev., Vol. 45, No. 2, pp. 74-82,
2001.

GM. Bartenev, The Structure and Properties of Inorganic Glasses, Wolters-
Noordhoff Publishing, 1970.

Narottam P. Bansal and RH. Doremus, Handbook of glass properties, Orlando:
Academic Press, 1986.

www.surtter.com

Sutter Instrument Company, P-2000 micropipette puller operation manual.

81

 

l11111111111111»111111111

 

 

‘h-

(1'4" ‘