Ghodssi R., Lin P., MEMS Materials and Processes Handbook
Подождите немного. Документ загружается.
664 D.W. Burns
593. V. Lehmann: Porous silicon – A new material for MEMS, 1996 IEEE Micro Electro
Mechanical Systems, MEMS
96, pp. 1–6, (San Diego, CA, Feb. 11–15, 1996)
594. V. Lehmann: The physics of macropore formation in low doped n-type silicon, J.
Electrochem. Soc. 140(10), 2836–2843 (1993)
595. P. Steiner, W. Lang: Micromachining applications of porous silicon, Thin Solid Films 225,
52–58 (1995)
596. P.A. Kohl, F.W. Ostermayer Jr: Photoelectrochemical methods for III-V compound semicon-
ductor device processing, Ann. Rev. Mater. Sci. 19, 379–399 (1989)
597. C. Levy-Clement, A. Lagoubi, M. Tomkiewicz: Morphology of porous n-type silicon
obtained by photoelectrochemical etching I. Correleations with material and etching param-
eters, J. Electrochem. Soc. 141(4), 958–967 (1994)
598. R.L. Smith, S.D. Collins: Porous silicon formation mechanisms, J. Appl. Phys. 71(8),
R1–R22 (1992)
599. D.R. Turner: Electropolishing silicon in hydrofluoric acid solutions, J. Electrochem. Soc.
105(7), 402–408 (1958)
600. S.S. Tsao, T.R. Guilinger, M.J. Kelly, P.J. Clews: Multilevel porous silicon formation, J.
Electrochem. Soc. 136(2), 586–587 (1989)
601. H. Takai, T. Itoh: Isolation of silicon film grown on porous silicon layer, J. Electron. Mater.
12(6), 973–982 (1983)
602. A. Richter, P. Steiner, F. Kozlowski, W. L ang: Current-induced light emission from a porous
silicon device, IEEE Electron Device Lett. 12(12), 691–692 (1991)
603. V. Lehmann: The physics of macroporous silicon formation, Thin Solid Films 255, 1–4
(1995)
604. R.W. Fathauer, T. George, A. Ksendzov, R.P. Vasquez: Visible luminescence from silicon
wafers subjected to stain etches, Appl. Phys. Lett. 60(8), 995–997 (1992)
605. S. Liu, C. Palsule, S. Yi, S. Gangopadhyay: Characterization of stain-etched porous silicon,
Phys. Rev. B 49(15), 10318–10325 (1994)
606. K. Imai, H. Unno: FIPOS (Full Isolation by Porous Oxidized Silicon) technology and its
application to LSI’s, IEEE Trans. Electron Devices ED-31(3), 297–302 (1984)
607. F. Beck: Integrated Circuit Failure Analysis, A Guide to Preparation Techniques, Chapters
3, 6 (Wiley, Chichester, 1998)
608. C.A. Harper (Ed.): Handbook of Materials and Processes for Electronics, Chapter 7
(McGraw-Hill, New York, NY, 1970)
609. W.R. Runyan, T.J. Shaffner: Semiconductor Measurements and Instrumentation, 2nd edn,
Chapters 2, 9 (McGraw-Hill, New York, NY, 1998)
610. K.R. Williams: Silicon Preferential (Defect Delineation) Etch Rates, In R. Hull (Ed.),
Properties of Crystalline Silicon, Chapter 16.3 (INSPEC, London, 1999)
611. W.R. Runyan, T.J. Shaffner: Semiconductor Measurements and Instrumentation, 2nd edn.,
Chapter 2 (McGraw-Hill, New York, NY 1998)
612. S.A. Campbell, H.J. Lewerenz (Eds.): Semiconductor Micromachining, vol. 2, Chapter 10
(Wiley, Chichester, 1998)
613. W.R. Runyan, T.J. Shaffner: Semiconductor Measurements and Instrumentation, 2nd edn.,
Chapter 9 (McGraw-Hill, New York, NY, 1998)
614. T.L. Chu, J.R. Gavaler: Stacking faults in vapor grown silicon, J. Electrochem. Soc. 110(5),
388–393 (1963)
615. W.C. Dash: Copper precipitation on dislocations in silicon, J. Appl. Phys. 27(10), 1193–1195
(1956)
616. H. Robbins: Junction delineation in silicon, J. Electrochem. Soc. 109(1), 63–64 (1962)
617. C.S. Fuller, J.A. Ditzenberger: Diffusion of donor and acceptor elements in silicon, J. Appl.
Phys. 27(5), (1956)544–553
618. P.J. Whoriskey: Two chemical stains for marking p-n junctions in silicon, J. Appl. Phys.
29(5), 867–868 (1958)
8 MEMS Wet-Etch Processes and Procedures 665
619. C.P. Wu, E.C. Douglas, C.W. Mueller, R. Williams: Techniques for lapping and staining
ion-implanted layers, J. Electrochem. Soc. 126(11), 1982–1988 (1979)
620. T. Mills, E.W. Sponheimer: Precision VLSI Cross Sectioning and Staining, Proceedings of
20th Annual Reliability Physics Symposium, pp. 214–220 (San Diego, CA, March 30–31,
1982)
621. B.L. Sopori: A new defect etch for polycrystalline silicon, J. Electrochem. Soc. 131(3),
667–672 (1984)
622. D.G. Schimmel: Defect etch for <100> silicon evaluation, J. Electrochem. Soc. 126(3),
479–483 (1979)
623. V.D. Archer: Methods for defect evaluation of thin <100> oriented silicon epitaxial layers
using a wet chemical etch, J. Electrochem. Soc. 129(9), 2074–2076 (1982)
624. K.H. Yang: An etch for delineation of defects in silicon, J. Electrochem. Soc. 131(5),
1140–1145 (1984)
625. M. Wright Jenkins: A new preferential etch for defects in silicon crystals, J. Electrochem.
Soc. 124(5), (1977)757–762
626. Transene Company, Inc., Danvers, MA 01923: Wright Etchant, http://www.transene.com/
wright.html, accessed December 24, 2010
627. T.C. Chandler: MEMC Etch – A chromium trioxide-free replacement for delineating
dislocations and slip in silicon, J. Electrochem. Soc. 137(3), 944–948 (1990)
628. F. Secco d’ Aragona: Dislocation etch for (100) planes in silicon, J. Electrochem. Soc.
119(7), 948–951 (1972)
629. R.T. Howe, R.S. Muller: Polycrystalline and amorphous silicon micromechanical beams –
Annealing and mechanical properties, Sens. Actuators 4, 447–454 (1983)
630. W.A. Pliskin, R.P. Gnall: Evidence for oxidation growth at the oxide-silicon interface from
controlled etch studies, J. Electrochem. Soc. 111(7), 872–873 (1964)
631. H. Baltes, O. Brand, G.K. Fedder, C. Hierold, J. Korvink, O. Tabata (Eds.): Advanced Micro
and Nanosystems, vol. 2, CMOS-MEMS (Wiley-VCH, Weinheim, 2005)
This is Blank Page Integra xxxii
Chapter 9
MEMS Lithography and Micromachining
Techniques
Daniel R. Hines, Nathan P. Siwak, Lance A. Mosher, and Reza Ghodssi
Abstract Photolithography is a patterning process that uses light to transfer a
pattern from a mask to a photosensitive polymer layer. The resulting pattern can
either be etched into the underlying surface or used to define the patterning of a layer
deposited onto the masked surface. This is essentially a two-dimensional process
that can be repeated numerous times to fabricate various structures and devices. A
classic use of these techniques is the fabrication of transistors on a silicon substrate
as practiced in the semiconductor industry. Development over a number of years has
yielded optimization of processing conditions, equipment, and materials to achieve
ever smaller sized features and an increased density of integration. In the effort to
fabricate even smaller (submicron to nanometer) features, other fabrication meth-
ods have also been developed such as electron beam lithography, focused ion beam
lithography, and nanoimprint lithography. This chapter presents an introduction and
practical approach to lithography and micromachining techniques in the context of
MEMS device fabrication. The topics that are discussed here include: UV lithog-
raphy, grayscale lithography, e-beam lithography, X-ray lithography, direct-write
lithographies and imprint lithographies. A detailed description including highlighted
examples of each of these lithographic techniques is presented in this chapter. At
the end of the chapter a compilation of hands-on case studies is presented to assist
readers in implementing these techniques in their own laboratories and developing
custom fabrication capabilities that fulfill their own unique requirements.
9.1 Overview
Photolithography as it is practiced today in the semiconductor and MEMS indus-
tries employs a light source, an optical projection system (stepper), a mask, and a
photoresist film coating the surface of a substrate into which a desired pattern is to
D.R. Hines (B)
Laboratory for Physical Sciences, College Park, MD, USA
e-mail: hines@lps.umd.edu
667
R. Ghodssi, P. Lin (eds.), MEMS Materials and Processes Handbook,
MEMS Reference Shelf, DOI 10.1007/978-0-387-47318-5_9,
C
Springer Science+Business Media, LLC 2011
668 D.R. Hines et al.
be transferred. Typically the patterning of the photoresist layer is not the desired
end result but rather the means by which the entire surface of the desired sub-
strate can be patterned in a parallel fashion. The advancements in resolution of the
photolithographic process have been a primary driving force behind the increased
device density on IC chips that has been empirically predicted by Moore’s law [ 1].
The combination of smaller wavelength light sources, improvements in optical pro-
jection systems and highly engineered chemistry of photoresists [2–8]haveledto
rapid advances in chip design to the point where the modern era has seen entire
music collections replaced by a single iPod
TM
handheld device and the prevalence
of cell phones in modern culture. Today we can carry around laptop computers that
operate at 2 GHz, have hundreds of GBs of storage capacity, are approximately 1 in.
thick, and weigh only a few pounds!
In 1957 Jay Lathrop and James Nall of the U.S. Army’s Diamond Ordnance Fuse
Laboratory in Maryland patented a photolithographic technique used to deposit thin-
film metal strips approximately 200 µm wide to connect discrete transistors on a
ceramic substrate. In 1958, Jay Last and Robert Noyce at Fairchild Semiconductor
built one of the first optical projection systems and used it to make many identical
transistors on a single Si wafer. In 1954, Bell Labs developed and combined the
primary fabrication procedures associated with oxidation, photomasking, etching,
and diffusion processes that underlie IC production to this day. The first monolithic
integrated circuits developed in the late 1950s contained only a small (∼10) number
of electrical components. An example of a circa 1961 integrated circuit containing
four transistors and five resistors is shown in the left panel of Fig. 9.1. In contrast,
Fig. 9.1 Left, top; images of a circa 1961 integrated circuit from Fairchild Semiconductor contain-
ing four transistors and five resistors. Left, bottom; image of a typical packaged IC chip with the
top cut open. Right; images of an Intel Pentium4 processor containing approximately 55 million
transistors
9 MEMS Lithography and Micromachining Techniques 669
development of lithography techniques has led to microprocessors with approxi-
mately 55 million transistors, as in the Intel Pentium4 processor (2001) shown in
the right panel of Fig. 9.1.
The same basic fabrication processes developed by the semiconductor industry
for the fabrication of IC chips are also used in the MEMS industry, and although
fabrication of devices within the MEMS community draws directly from these litho-
graphic techniques, MEMS requirements demand the development of processing
capabilities that go beyond those employed for IC chip fabrication [9, 10]. For exam-
ple, in MEMS devices, it may be desirable to pattern polymeric (or elastomeric)
materials to create flow channels rather than passivation layers, introducing different
materials into the fabrication process, as well as additional material functionality.
Also employed are silicon-on-insulator wafers which can be used to create freely
moving structures (such as gears) or mechanical members that are only partially
attached to the underlying substrate (such as cantilevers and microbeams) rather
than monolithic Si transistors. Using similar materials to produce different compo-
nents often requires modified, if not dramatically different, processing steps. Beyond
conventional lithographic fabrication, there are emerging fields of soft lithography
[11–21] and related printing techniques that show great promise for MEMS fabri-
cation. Many of these techniques have been developed as procedures for patterning
and processing unconventional materials such as organic thin-films, polymeric sub-
strates and dielectric layers, 3-D structures produced by combining molding and
sintering processes, inkjet printing of active or structural materials, and so on. These
unconventional fabrication methods have the potential to drastically change the way
electronics and MEMS devices are fabricated and open new possibilities for device
applications that never existed before.
Photolithography, although not the only one, is by far the most developed pat-
terning method available for the fabrication of MEMS devices. The process, with
the main components illustrated in Fig. 9.2, is described in detail in Section 9.2.
Fig. 9.2 Common
components in a
photolithography system
670 D.R. Hines et al.
Briefly, UV light is used to create a pattern of open areas in a polymer film called
photoresist. Material is then either added to or removed from these open areas. The
desired patterns are first designed as opaque and transparent regions on a glass plate
called a photomask. The photomask is mounted into an optical projection system
where an image of the mask is focused onto a semiconductor wafer coated with
photoresist. After exposure, the photoresist layer is developed to reveal the pho-
tomask pattern transferred into the photoresist. Both the optical projection system
and photoresist chemistry are highly engineered to achieve high resolution and high
repeatability within the photolithographic process.
For standard photolithography, the photomasks and photoresist chemistry are
both designed to produce 2-D patterns within the photoresist film. However, the
photomask can be modified to transmit different intensities of UV light. This is
accomplished by using closely spaced pixels rather than fully opaque and/or trans-
parent areas, or by using variably transmissive mask media. The limitations on pixel
dimensions imposed by mask manufacturing and the resolution of optical projec-
tion systems, as well as the variable optical absorbance of mask media give rise to
specific transmission intensities referred to as gray levels. These gray levels can be
used to structure the thickness profile of a photoresist layer. Given appropriate etch
parameters, this process of grayscale lithography (Section 9.3) can be used to create
3-D structures on a substrate surface.
UV light is not the only energy source that can be used to change the solubility
of regions within a polymer film. For example, the solubility of PMMA can be
dramatically changed by exposure to electrons and X-rays [7]. Both electrons and
X-rays have shorter wavelengths than UV light and so can achieve higher resolution
patterns. The main difference between UV lithography and both e-beam and X-ray
lithographies is simply in the light source and method used for exposure.
For e-beam lithography (Section 9.5.1) the “light source” is essentially an SEM
that has been modified to switch the electron beam on and off as it is s canned. Using
a scanned electron beam, the resolution of e-beam lithography is greater than that
of UV lithography. However, a disadvantage is that the exposure now involves a
high vacuum system and is a serial rather than a parallel process, as in wide-area
exposure or projection. The electron beam has to be moved point-by-point across
the area to be exposed rather than all points exposed simultaneously, such a direct-
write approach will drastically increase the time to expose a pattern. For applications
that require massive parallelism and mass production, this can be problematic.
In the case of X-ray lithography (Section 9.4), the light source is typically a syn-
chrotron, and wafers are exposed similarly to a contact projection system analogous
with UV lithography. The advantage of X-ray lithography is the extremely high
aspect ratio and near vertical sidewalls that can be patterned into the photoresist.
Disadvantages include the need for exposures to be performed at a s ynchrotron,
cost and difficulty of X-ray mask production, and the low sensitivity of resists to
X-rays thus demanding long exposure times.
UV light, electrons, and X-rays can all be used to modify chemical bonds within
a polymer film but do not have enough energy to remove material directly from
a substrate surface. For these energy sources, a photoresist layer can be patterned
9 MEMS Lithography and Micromachining Techniques 671
by changing the chemistry of the film (i.e., solubility) but the substrate cannot be
patterned directly. By contrast, ion beams are energetic enough to remove atoms
directly from a substrate surface. Focused ion beam lithography (FIB; Section 9.5.2)
can be used to directly pattern a substrate surface without the need for a photoresist
layer. Here, substrate atoms are physically removed by a process called sputtering.
Both electron and ion beams can be used to modify the chemistry of a gas introduced
above the substrate such that the resulting gas molecules are either deposited onto
the substrate or chemically bonded with substrate atoms to form volatile compounds
that are easily removed from the substrate. In this way, both electron and ion beam
lithography can be performed directly on a substrate surface.
Of specific interest for MEMS applications is the ability to utilize direct-write
methods to fabricate 3-D mechanical structures onto a substrate. At the nanoscale,
material can be added to a substrate by coating an AFM tip with a specific molecule
and then dragging the tip across the substrate. Thus with dip-pen lithography
(Section 9.5.4) molecules get pulled off the tip to form a line on the substrate, much
like a nanoscale fountain pen. On a larger scale, laser beams can be used to transfer
materials from one substrate to a second substrate. This is performed by depositing
a desired material over a sacrificial layer on the first substrate. A laser beam is then
used to vaporize the sacrificial layer thus removing the desired materials from the
first substrate and transferring it over to the second substrate. Laser beams can also
be used to polymerize precursor polymer solutions.
Stereolithography ( Section 9.5.6) can be performed by submerging a substrate
into a precursor polymer solution and moving the laser beam and stage in a
predetermined pattern to produce a 3-D polymerized structure attached to the
substrate.
Printing techniques can also be used to fabricate MEMS components. Inkjet
printers (Section 9.6.1) were first designed to deposit high-contrast, passive inks
onto a substrate (paper) in a step-by-step fashion. Pressure pulses are used to force
ink through small holes in the printing head. New inks are being designed that allow
the printing of 3-D structures made from polymers, metal nanoparticles, organic
molecules, and the like [22–24].
Other forms of printing such as soft lithography and nanoimprint lithography are
being used to pattern films on substrate surfaces. Soft lithography (Section 9.6.2)
uses a PDMS stamp coated with a specific molecule. The stamp is brought into
contact with the substrate resulting in some or all of the film coating the PDMS
stamp to be inked onto the substrate. Nanoimprint lithography (NIL; Section 9.6.3)
creates a pattern in a resist layer by direct contact with a templated surface. The
resulting pattern in the resist layer is similar to that produced by UV lithography,
however, NIL can achieve pattern resolution in the tens of nm.
The last printing method to be discussed in this chapter is transfer printing
(Section 9.6.4). This method can be used to directly assemble dissimilar materi-
als onto a common substrate. The process r elies on differential adhesion to transfer
a patterned film from one substrate t o a second substrate. As long as the adhesion
is properly engineered, a wide variety of dissimilar materials can be sequentially
assembled onto the same substrate to produce both electronic and mechanical
672 D.R. Hines et al.
devices in ways that are difficult if not impossible to achieve using conventional
lithographic processes.
This chapter begins by presenting a detailed discussion of conventional pho-
tolithography methods with an emphasis on MEMS processing. Related lithographic
methods based on X-rays, electron beams, and ion beams are then presented. The
chapter also covers fabrication processes developed more specifically for MEMS
applications such as grayscale lithography and microstereolithography that are
meant for fabricating 3-D structures. The chapter concludes with new, emerging
fabrication techniques based on soft lithography and printing techniques that have
only recently been considered and developed for MEMS processing applications.
9.2 UV Lithography
The concepts underlying photolithography are quite simple. The process uses a
patterned mask comprised of transparent and opaque regions. Using an optical
projection system, this pattern is illuminated onto a photosensitive film called a pho-
toresist. When light strikes the photoresist, it causes a chemical change to take place
within the photoresist film which in turn changes the ability of the photoresist to
be dissolved by various solvents. This allows specific areas of the photoresist layer
to be removed, which uncovers (exposes) the underlying substrate surface for fur-
ther processing. The real advantage of the photolithographic process is that it can be
performed over large areas at a very high resolution. The photolithographic process
contains four main components (photomask, optical projection system, photoresist,
and substrate) and can be described by a five-step process.
I. Deposit photoresist
II. Expose photoresist
III. Develop photoresist
IV. Transfer pattern
V. Remove photoresist
These steps are illustrated in Figs. 9.2 and 9.3 with each component and process-
ing step labeled. Each of the four components and five processing steps is discussed
in appropriate detail below. Specific examples relevant to MEMS processing are
highlighted.
9.2.1 Photo Masks
Masks for photolithography are plates that contain transparent and opaque regions
laid out to form a desired pattern [25]. These patterns contain microscopic features
that form parts of electronic circuits or parts of MEMS devices. An example of a
photomask is shown in Fig. 9.4. Typically the plate is either glass or quartz that
has been designed for transparency in the UV and contains an opaque patterned
9 MEMS Lithography and Micromachining Techniques 673
Fig. 9.3 Photolithography processing steps
metal film such as a typical 800 Å thick chromium layer on the top surface (much
like a front surface mirror). Iron oxide is another common patterning material for
photomasks. It has the advantage of being transparent in the visible while opaque
in the UV, making it easier to align to previously patterned features in multilayer