M E T H O D S I N M O L E C U L A R B I O L O G YTM
John M. Walker
School of Life Sciences
University of Hertfordshire
Hatfield, Hertfordshire, AL10 9AB, UK
For other titles published in this series, go to
Microengineering in Biotechnology
Michael P. Hughes and Kai F. Hoettges
University of Surrey, Guildford, UK
Michael P. Hughes
University of Surrey
Centre for Biomedical
Duke of Kent Building
United Kingdom GU2 7TE
Kai F. Hoettges
University of Surrey
Centre for Biomedical
Duke of Kent Building
United Kingdom GU2 7TE
Library of Congress Control Number: 2009933982
# Humana Press, a part of Springer ScienceþBusiness Media, LLC 2010
All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or
hereafter developed is forbidden.
The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as
such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.
While the advice and information in this book are believed to be true and accurate at the date of going to press, neither
the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be
made. The publisher makes no warranty, express or implied, with respect to the material contained herein.
Printed on acid-free paper
Just as the twentieth century witnessed developments both in electronic engineering
and molecular biology which have revolutionized the way we live, so the twenty-first
Since so many biochemical procedures occur at the molecular level, microelectronic
engineering offers the opportunity to reduce the way in which such procedures are
performed to the same level. The advantages of miniaturized analysis are manifold;
reducing the sample volume increases reaction speed and detector sensitivity whilst
reducing sample and reagent requirements and device cost. Some of the world’s largest
having potential to provide, for example, bench-top versions of large and expensive
equipment that could make analyses like flow cytometry as commonly available as gel
electrophoresis is now.
The aim of this book is to provide biochemists, molecular biologists, pharmacolo-
gists and others with a working understanding of the methods underlying microengi-
neering and the means by which such methods can be used for a range of analytical
techniques. It describes the methods by which microengineered devices can be built to
perform a number of applications and considers how the field may progress by examin-
ing some more complex lab on a chip devices which have great potential in the
advancement of the way in which molecular biology is performed. We also hope that
this book will be of use to microengineers, both as a reference guide for practical
microengineering techniques and as a route into the development of new devices for
biological applications. The union of molecular biology and microelectronics offers
huge promise but one which will be all the stronger wherein both sides understand the
needs of the other.
27 April 2009
Michael P. Hughes
Kai F. Hoettges
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1. Microfabrication Techniques for Biologists: A Primer on
Building Micromachines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2. The Application of Microfluidics in Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
David Holmes and Shady Gawad
3. Rapid Prototyping of Microstructures by Soft Lithography
for Biotechnology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Daniel B. Wolfe, Dong Qin, and George M. Whitesides
4. Chemical Synthesis in Microreactors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Paul Watts and Stephen J. Haswell
5. The Electroosmotic Flow (EOF) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gary W. Slater, Fre ´de ´ric Tessier, and Katerina Kopecka
6. Microengineered Neural Probes for In Vivo Recording . . . . . . . . . . . . . . . . . . . . .
Karla D. Bustamante Valles
7. Impedance Spectroscopy and Optical Analysis of Single Biological Cells
and Organisms in Microsystems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Shady Gawad, David Holmes, Giuseppe Benazzi, Philippe Renaud,
and Hywel Morgan
8. Dielectrophoresis as a Cell Characterisation Tool. . . . . . . . . . . . . . . . . . . . . . . . . .
Kai F. Hoettges
9. AC-Electrokinetic Applications in a Biological Setting . . . . . . . . . . . . . . . . . . . . . .
Fatima H. Labeed
10. Wireless Endoscopy: Technology and Design. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
David R. S. Cumming, Paul A. Hammond, and Lei Wang
Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
GIUSEPPE BENAZZI • School of Electronics and Computer Science, University of
Southampton, Highfield, Southampton, UK
KARLA D. BUSTAMANTE VALLES • Orthopaedic & Rehabilitation Engineering Center,
The Medical College of Wisconsin & Marquette University, Milwaukee, WI, USA
DOUGLAS CHINN • Sandia National Laboratories, Albuquerque, NM, USA
DAVID R. S. CUMMING • Department of Electronic and Electrical Engineering,
University of Glasgow, Glasgow, UK
SHADY GAWAD • Swiss Federal Institute of Technology, Lausanne, Switzerland
PAUL A. HAMMOND • Department of Electronic and Electrical Engineering, University
of Glasgow, Glasgow, UK
STEPHEN J. HASWELL • Department of Chemistry, University of Hull, Hull, UK
KAI F. HOETTGES • Centre for Biomedical Engineering, University of Surrey, Guildford,
DAVID HOLMES • School of Electronics and Computer Science, University of Southampton,
Highfield, Southampton, UK
KATERINA KOPECKA • Department of Physics, University of Ottawa, Ottawa, ON,
FATIMA H. LABEED • University of Surrey, Guildford, Surrey, UK
HYWEL MORGAN • School of Electronics and Computer Science, University of Southamp-
ton, Highfield, Southampton, UK
DONG QIN • University of Washington, Seattle, WA, USA
PHILIPPE RENAUD • Swiss Federal Institute of Technology, Lausanne, Switzerland
GARY W. SLATER • Department of Physics, University of Ottawa, Ottawa, ON, Canada
FRE´DE´RIC TESSIER • DepartmentofPhysics,UniversityofOttawa,Ottawa,ON,Canada
LEI WANG • Department of Electronic and Electrical Engineering, University of Glas-
gow, Glasgow, UK
PAUL WATTS • Department of Chemistry, University of Hull, Hull, UK
GEORGE M. WHITESIDES • Department of Chemistry and Chemical Biology, Harvard
University, Cambridge, MA, USA
DANIEL B. WOLFE • Department of Chemistry and Chemical Biology, Harvard
University, Cambridge, MA, USA
Microfabrication Techniques for Biologists: A Primer on
In this chapter we review the fundamental techniques and processes underlying the fabrication of devices
on the micron scale (referred to as ‘‘microfabrication’’). Principles laid down in the 1950s now form the
basis of the semiconductor manufacturing industry; these principles are easily adaptable to the production
of devices for biotechnological processing and analysis.
Key words: Fabrication, photolithography, photomask, etching, thin films.
Fabrication of electronic devices on a micron scale was developed
by the integrated circuit industry, beginning with the invention of
the integrated circuit (IC) simultaneously by Jack Kilby of Texas
Instruments and Robert Noyce of Fairchild in 1958. Microma-
chining began with the challenge issued by Professor Feynman in
1959 to build a tiny motor. In 1965, Gordon Moore of Intel
postulated what is now known as Moore’s Law, where the data
density of integrated circuits doubles every 18 months. Today,
2006, the largest companies are fabricating complex state-of-the-
art chips with over 40 photomask layers, including 10 metal layers,
and are approaching 109transistors in an area the size of a postage
stamp. Lateral dimensions are below 100 nm and shrinking all the
time. Vertical dimensions for the thinnest oxides are a few mono-
layers, with films of metal and dielectrics on the order of a few
nanometers (nm)to amicrometer (micron,mm).A modernsilicon
integrated circuit fabrication facility costs over $2 billion to build
M.P. Hughes, K.F. Hoettges (eds.), Microengineering in Biotechnology, Methods in Molecular Biology 583,
DOI 10.1007/978-1-60327-106-6_1, ª Humana Press, a part of Springer Science+Business Media, LLC 2010
and equip. The new MESA micromachine laboratory at Sandia
National Laboratories costs almost $500 million. Only large cor-
porations and governments can make this kind of capital invest-
ment. Chips must be made in huge quantities with exceptional
quality control to achieve an adequate return on investment.
Because chip-making processes advance so quickly, process equip-
ment is typically depreciated after 2 years.
An integrated circuit is fabricated using only four steps: film
growth and deposition, patterning, etching, and annealing. A
typical process begins by depositing a film, spinning photoresist,
patterning and developing the resist, etching the thin film, and
finally stripping the resist. By repeating these steps over and over,
the most complex devices can be made. This is a gross oversimpli-
fication but it demonstrates how a number of simple steps can be
combined to make very complex devices. In industry, complex
processing tools such as etchers and thin film deposition systems
are dedicated to a single process. Engineers in semiconductor and
micromachine factories devote entire careers to reducing variabil-
ity by characterizing processes and equipment and thus improving
device yield. Even so, scrap rates run up to 50% of all material
started in the line, depending on the maturity of the process.
Mature lines with well-characterized processes will skip testing in
wafer form and package every device made, rejecting the defective
parts at that stage, indicating that well over 99% of all wafers were
Most modern wafer processing machine tools are ‘‘cassette to
designedwaferholdersknownas cassettesor boats.Humansnever
handle wafers. Instead, cassettes are placed into ‘‘indexers’’ that
returning them to a cassette when the process is finished. Since
humans make mistakes and add defects, they must be removed as
far as possible from the wafers.
Here we must distinguish between the terms ‘‘fab’’ and ‘‘lab.’’
Fab refers to a large, commercial factory set up to turn a profit,
which produces a small variety of integrated circuit devices that
have similar processes. Lab refers to a smaller facility, usually at a
university or government laboratory, set up to produce a wide
variety of electronic, optical, and mechanical devices with many
is designed for high throughput (low cost) and must be designed
Uptime is a major consideration, and fabs will schedule routine
maintenance into the process. Although programmers who use
chips to create amazing programs get all the glory, the people who
design and build the machines that make the chips are the real
heroes because the semiconductor industry arguably has the high-
est technology machine tools in existence. Creating a 25-mm
square chip with dozens of layers, thousands of process steps,
geometries below 0.25 mm, and one billion transistors with no
defects is not an easy task!
The emerging micromachine industry has benefited from all
the advances made by the semiconductor industry, but is several
Moore’s law generations behind it. In many micromachining
applications, lateral dimensions are above 1 mm, and vertical
dimensions are above 100 nm. Universities and other small labs
do not need the expensive dedicated cassette to cassette machines
designed for high throughput and quality control. This chapter is
written for the small lab, with contrasts to large production facil-
ities. We hope to give rules of thumb that help get devices fabri-
cated without the reader having to ‘‘re-invent the wheel’’ while
learning to build microdevices.
Why is this chapter called ‘‘Microfabrication for Biologists’’
when it has absolutely no reference to biology in it? Because most
commercially successful micromachines have biological applica-
tions and most of those are microfluidic devices that sense some
driving micromachining away from silicon-based devices that
often have moving parts to devices that analyze fluids built on
glass and polymer substrates.
Many good books on the subject of microfabrication have
been written (see Suggested Reading List). These books list
some exotic and complex processes and the reader is referred to
these books and others for rigid analysis and mathematical detail.
first device needs practical information to get started on a project.
Here we attempt to supply the reader with a lot of practical knowl-
edge, gained from the author’s experience in building devices in
industry, university, and government laboratories. The approach
used in this chapter is to teach a novice engineer, scientist
or biologist how to design a device pattern using a CAD
(computer-aided design) program and how to develop a process
run sheet to get it fabricated with the help and assistance of
professionals who work in microdevice process labs every day.
We have avoided the use of trade names where possible. This
chapter is intended to be a practical guide to getting started in
the field, making no attempt to define the underlying physics and
chemistry. Since most biologically oriented micromachining
involves substrates that are not silicon or period II–VI or III–V
semiconductor compounds, this chapter is written with that in
Any device built by micromachine techniques has hundreds of
potential variables to control. Small changes in any one of these
can have a large impact on the final product. The keys to
successful microdevice fabrication are good process control in
Microfabrication Techniques for Biologists3
well-characterized machines, a well-defined fabrication process,
well-designed photomasks, and careful fabrication. To maximize
the possibility of success, it is highly recommended that the novice
microfabricator devote a great deal of time to designing the device
and process before beginning production. Properly designed
photomasks and a complete process run sheet, or traveler, can be
assembled by talking with people familiar with the process tools
their many variables in the proper order, and a simple example is
given below. In the world of microfabrication, the tiniest of details
of all the processing is necessary to repeat a successful process.
engineers building their first microdevice typically expect the first
attempt at fabrication to work. This is rarely the case, so one
should expect several scrap wafers before having success. The
reason is that in a research setting, well-characterized, reproduci-
ble, standardized processes are not often available, and as Mur-
phy’s law states, anything that can go wrong will. With hundreds
of variables to control, and only limited knowledge of any given
process step, fabricating a device correctly during the first attempt
is not highly probable. An engineer or biologist must fabricate the
device under the limitations of the laboratory equipment available
and the time available to characterize processes, but the resources
available to the researcher limit processing capabilities.
The first consideration for microfabrication is the substrate, or
wafer, the device will be built upon. Integrated circuits are built
on wafers, where micromachines can be built on a variety of sub-
strate materials, so the term substrate is used interchangeably with
wafer. Micromachining is often divided into two categories, bulk
and surface. In bulk micromachining, the substrate is formed into
the finished device, whereas in surface micromachining layers are
added on top (or on the bottom) of the substrate, and the added
layers are formed into the desired device. Many technologies
combine surface and bulk micromachining. It is highly recom-
mended that one obtain many more substrates than are required
for the project to use as dummy wafers, test wafers and to take into
account scrapped wafers caused by misprocessing. Plan on drop-
ping and breaking some substrates too. It is also recommended
that substrates be scribed at the beginning of the process to make
traceability easier. The variables affecting the specification of
substrates can be seen in Fig.1.1.
The selection of a substrate determines much of the
subsequent processing. Most microfabrication tools are designed
for silicon wafers with a flat. Theflat issimply an area ground off of
the edge of the round wafer. It is there so that the crystal orienta-
tion can be determined and is used by automated processing
equipment to orient the wafer. Square or other shaped substrates
can be used, but non-circular substrates are much harder to work
sizes that fit into most machines, so it is advantageous to use a
standard size. Two-inch (50 mm), three-inch (75 mm), four-inch
(100 mm), and six-inch (150 mm) substrates are common stan-
dards. However, the metric and English sizes are not necessarily
compatible in some machine wafer holders (1 inch=25.4 mm).
When ordering substrates,several variablesmustbe accounted
for. After diameter, and thickness, one must determine the flat-
ness, surface roughness, edge grind, the finish on the back of the
substrate, temperature limitations, coefficient of thermal expan-
sion, and transparency to a given wavelength of light and
mechanical toughness. Poorly ground edges can be a source of
particles when they chip, and can initiate cracks which will cause
the substrate to break. In a plasma etch process, note that the
conductivity of the substrate can have an impact on the way the
Silicon has the advantage that it is widely available, relatively
inexpensive and very well characterized. It is also the most perfect
material obtainable, being a slice of a single crystal. The crystal-
linity can be taken advantage of in directional etches and in
that can determine how successful the device fabrication process will be.
Microfabrication Techniques for Biologists5
applications where it is desired to etch through the wafer. The
most common types are (100), (111) and (110). These are the
Miller indices, indicating which crystal plane makes up the surface
of the wafer. Silicon can be ordered with either p (boron) or n
doping can be done with an ion implanter or high-temperature
furnace. If thin membranes are required in the device, silicon may
be the optimum material because some wet etchants, such as
KOH, can be highly selective between different crystal planes
and levels of doping. Custom silicon can be ordered from many
manufacturers, but off-the-shelf material is usually much cheaper.
Glass is being used more often in micromachining applica-
tions. It is amorphous, rather than crystalline like silicon. Dozens
of types of glass are available, and it can be ordered in custom
shapes and sizes. It is inert, transparent, and mechanically tough.
Some types of glass can be very inexpensive, such as microscope
slides, and some, such as fused silica and quartz, can be very
expensive. Many glasses are chemically impure, and because it is
amorphous, it can have a non-uniform microstructure. This can
affect etching, optical, and electrical properties.
Polymer substrates are now being used. It is very difficult to
generalize about polymers, since so many are available. Plastics are
inert to acids and can be formed by molding and embossing, but
plastics tend to have a non-uniform microstructure. Polymer
substrates can absorb water and solvents, changing their
Metal substrates have been used, but due to the reactivity of
metals they generally hold up very poorly during chemical treat-
ments. It is also difficult to get surfaces with low roughness in
metals. Ceramics are also used, but the polycrystalline nature of
ceramics can present severe processing difficulties
Wafer supply companies often have film deposition equip-
ment. If all of your substrates need a particular film, it may be
cheaper, faster, and cleaner to buy wafers with a film already
deposited on them by the supplier.
used to expose some kind of photoresist. The most common kind
is made of a thin film of chrome on a square of glass that is
transparent into the UV range. Many different kinds of glass are
on stray reflectance from the front surface. Most align tools use a
mask that is 1 inch/25 mm larger than the wafer size.
Contact the commercial mask fabrication shop at the begin-
ning of the design cycle so that your design is compatible with the
shop’s tooling. Many drawing programs are available, but line
types and shapes used by generic drawing programs cannot be
recognized by mask shop tools. Mask shops require the data to
be in specific formats, so an important factor is being able to
convert the finished CAD drawing into a format that the mask
capable of saving the data in an acceptable format.
Inexpensive photomasks can be designed using any drawing
program and printed out using standard office printers on Mylar
transparencies. Such masks are only effective for larger geometries
with loose tolerances, but can be made quickly and cheaply.
A reticle is a photomask that is scaled, typically used in various
types of steppers. The advantage of using a reticle over a photo-
mask is that small geometries are achieved optically by the end
user, rather than by trying to write small features directly on the
photomask. Each type of stepper requires its own reticle, where
photomasks can be used in most machines, if sized right. If your
device is small and several copies can fit on a single wafer, a shop
may generate a reticle and step it over and over to make a mask,
rather than write the same pattern over and over.
Sometimes a photomask has a pellicle attached to it, a thin
membrane that is set away from the chrome geometry by a spacer.
Many designers have been frustrated because an expensive mask
does not do what it was designed to do. As illustrated below, the
mask design is not independent of the process. Several decisions
must be made during the design stage.
1. Wafer or substrate and die size
2. Type and size of mask
3. Dark or light field
4. Spot size
5. Minimum dimension/pitch
6. Defect density
8. Process bias
9. Positive or negative resist
10. Alignment keys
12. Dicing the wafer into individual chips
13. Howthe device will bepackaged orhow the inputs andoutputs
will travel from the macroworld to the chip scale microworld.
Microfabrication Techniques for Biologists7
One of the firstdecisions to make isthe die, or chip, size. Theword
die comes from the IC industry, where a typical wafer will have
compound semiconductor) chip inside that could be anywhere
from a fraction of a mm2to ?30 mm2in size. These individual
dies are cut from a larger wafer. Many micromachine devices can
take up an entire wafer. If your device is smaller than the wafer it
may be beneficial to have many dies on the same wafer.
The type of lithography tool to be used and the substrate size
determine the size and type of the mask to be made. Contact
printers, steppers, and projection printers all have many specific
mask requirements, so decide on the alignment tool before doing
the CAD design. Masks can be ordered as masters or copies.
Copies are cheaper than masters, but can have lower resolution.
A master must be made to make a copy. If contact printing is to be
used, a mask can be damaged, so it may be economical to use
copies rather than the original master.
Dark field masks are mostly chrome with holes in the film;
light field masks appear to be mostly glass with chrome geometries
on them. The type of resist used and the desired pattern determine
whether dark or light field is called for. Masks are usually drawn
light field, and dark field is specified on the mask shop’s run sheet.
The way a mask exposes photoresist is a function of how much
geometry is present and how it is distributed. During lithography,
dense geometries in one area may expose differently than areas
are a function of geometry density and size.
A mask shop starts with a photoresist covered chrome film and
the writing is a major determinate of the price, with 0.25 mm being
write. Both line and space (pitch) minimum dimensions on the mask
can determine what type of equipment is used to make the mask:
larger dimensions are sometimes cheaper to make. The smallest geo-
the sum of the line space. Mask shops may define a minimum pitch.
size of defects allowed. If your device has small, dense geometries,
very few defects in the form of missing chrome or extra spots of
chrome can cause fatal defects in the device. If you have large, low
generalrule, a defect 10% ofthegeometrysize may causeproblems.
The tolerances in a given layer and the tolerances between layers
can also impact the price of a photomask. If all lines and spaces are
above 1 or 2 mm the mask can be fairly inexpensive. Mask shops
cannot alter their quality control for each mask. If a mask has tight
tolerances and low defect levels, a shop may have to make it several