The position of droplets within such microvoids is governed by the surface tension of the droplet fluid, the wettability of the fluid with respect to microvoid walls con-tacted during di
Trang 248 A.D Johnson, and V.V Martynov, Proc 2nd Int Conf.
Shape Memory Superelastic Technol., Pacific Grove, CA, 1997,
pp 149–154.
49 M Kohl, K.D Strobanek, and S Miyasaki, Sensors and
Actua-tors A72: 243–250 (1999).
50 Y Bellouard, T Lehnert, T Sidler, R Gotthardt, and R Clavel,
Mater Res Soc Mater Smart Syst III 604: 177–182 (2000).
51 Y Bellouard, T Lehnert, J.-E Bidaux, T Sidler, R Clavel,
and R Gotthardt, Mater Sci Eng A273–275: 733–737
(1999).
52 K Kuribayashi, S Shimizu, M Yoshitake, and S Ogawa, Proc.
6th Int Symp Micro Mach Hum Sci Piscataway, NJ, 1995,
pp 103–110.
53 S.T Smith and D.G Chetwynd, Gordon and Breach, 1994.
54 J.M Paros and L Weisborg, Mach Design 27: 151–156
(1965).
55 Y Bellouard, Ph.D Thesis, Lausanne, EPFL, n ◦2308 (2000).
56 W Nix, Scripta Materialia, 39(4/5): 545–554 (1998).
57 R.D James, Int J Solids Struct 37: 239–250 (2000).
58 R Gorbet, Ph D Thesis, University of Waterloo, 1997.
Microtubes are very small diameter tubes (in the
nanome-ter and micron range) that have very high aspect ratios
and can be made from practically any material in any
combination of cross-sectional and axial shape desired In
smart structures, these microscopic tubes can function as
sensors and actuators, as well as components of fluidic
logic systems In many technological fields, including smart
structures, microtube technology enables fabricating
com-ponents and devices that have, to date, been impossible to
produce, offers a lower cost route for fabricating some
cur-rent products, and provides the opportunity to miniaturize
numerous components and devices that are currently in
existence
In recent years, there has been tremendous interest in
miniaturization due to the high payoff involved The most
graphic example that can be cited occurred in the
electron-ics industry, which only 50 years ago relied exclusively on
the vacuum tube for numerous functions The advent of
the transistor in 1947 and its gradual replacement of the
vacuum tube started a revolution in miniaturization that
was inconceivable at the time of its invention and is not
fully recognized even many years later
Miniaturization resulted in the possibility for billions of
transistors to occupy the volume of a vacuum tube or the
first transistor, and it was not the only consequence The
subsequent spin-off developments in allied areas, such asintegrated circuits and the microprocessor, have spawnedentirely new fields of technology It is quite likely that otherareas are now poised for revolutionary developments thatparallel those that have occurred in the electronics indus-try since the advent of the first transistor
These areas include microelectromechanical systems(MEMS) and closely related fields, such as microfluidicsand micro-optical systems Currently, these technologiesinvolve micromachining on a silicon chip to produce nu-merous types of devices, such as sensors, detectors, gears,engines, actuators, valves, pumps, motors, and mirrors on
a micron scale The first commercial product to arise fromMEMS was the accelerometer that was manufactured as asensor for air-bag actuation On the market today are alsomicrofluidic devices, mechanical resonators, biosensors forglucose, and disposable blood pressure sensors that are in-serted into the body
The vast majority of microsystems are made almost clusively on planar surfaces using technology developed tofabricate electronic integrated circuits The fabrication ofthese devices takes place on a silicon wafer, and the de-vice is formed layer-by-layer using standard clean-roomtechniques that include electron beams or photolithogra-phy, thin-film deposition, and wet or dry etching (bothisotropic and anisotropic) Three variations of this conven-tional electronic chip technology can be used, for example,
ex-to make three-dimensional structures that have high pect ratios and suspended beams These include the LIGA(lithographie, galvanoformung, abformung) process (1,2),the Hexsil process (3), and the SCREAM (single-crystalreactive etching and metallization) process (4) The tech-nique most employed, the LIGA process, which was de-veloped specifically for MEMS-type applications, can con-struct and metallize high-aspect-ratio microfeatures This
as-is done by applying and exposing a very thick X-ray sitive photoresist layer to synchrotron radiation Features
sen-up to 600 microns high that have aspect ratios of 300 to
1 can be fabricated by this technique to make truly dimensional objects The Hexsil process uses a mold thathas a sacrificial layer of silicon dioxide to form polysili-con structures that are released by removing the silicondioxide film A third approach is the SCREAM bulk mi-cromachining process that can fabricate high-aspect-ratiosingle-crystal silicon suspended microstructures from a sil-icon wafer using anisotropic reactive ion etching Note,however, that like the conventional technique used to makeelectronic circuits, all of these variations use a layered ap-proach that starts on a flat surface
three-In addition, there are some disadvantages of the tional electronic chip fabrication technique and its modifi-cations, even though there have been numerous and veryinnovative successes using these silicon wafer-based tech-nologies This is due to the fact that these technologiesrequire building up many layers of different materials aswell as surface and bulk micromachining which leads tosome very difficult material science problems that have
conven-to be solved These include differential etching and layingdown one material without damaging any previous layer
In addition, there are the problems of interconnecting ers in a chip that have different functions An example ofthis is a microfluidic device in which there are both fluidic
lay-Next Page
Trang 348 A.D Johnson, and V.V Martynov, Proc 2nd Int Conf.
Shape Memory Superelastic Technol., Pacific Grove, CA, 1997,
pp 149–154.
49 M Kohl, K.D Strobanek, and S Miyasaki, Sensors and
Actua-tors A72: 243–250 (1999).
50 Y Bellouard, T Lehnert, T Sidler, R Gotthardt, and R Clavel,
Mater Res Soc Mater Smart Syst III 604: 177–182 (2000).
51 Y Bellouard, T Lehnert, J.-E Bidaux, T Sidler, R Clavel,
and R Gotthardt, Mater Sci Eng A273–275: 733–737
(1999).
52 K Kuribayashi, S Shimizu, M Yoshitake, and S Ogawa, Proc.
6th Int Symp Micro Mach Hum Sci Piscataway, NJ, 1995,
pp 103–110.
53 S.T Smith and D.G Chetwynd, Gordon and Breach, 1994.
54 J.M Paros and L Weisborg, Mach Design 27: 151–156
(1965).
55 Y Bellouard, Ph.D Thesis, Lausanne, EPFL, n ◦2308 (2000).
56 W Nix, Scripta Materialia, 39(4/5): 545–554 (1998).
57 R.D James, Int J Solids Struct 37: 239–250 (2000).
58 R Gorbet, Ph D Thesis, University of Waterloo, 1997.
Microtubes are very small diameter tubes (in the
nanome-ter and micron range) that have very high aspect ratios
and can be made from practically any material in any
combination of cross-sectional and axial shape desired In
smart structures, these microscopic tubes can function as
sensors and actuators, as well as components of fluidic
logic systems In many technological fields, including smart
structures, microtube technology enables fabricating
com-ponents and devices that have, to date, been impossible to
produce, offers a lower cost route for fabricating some
cur-rent products, and provides the opportunity to miniaturize
numerous components and devices that are currently in
existence
In recent years, there has been tremendous interest in
miniaturization due to the high payoff involved The most
graphic example that can be cited occurred in the
electron-ics industry, which only 50 years ago relied exclusively on
the vacuum tube for numerous functions The advent of
the transistor in 1947 and its gradual replacement of the
vacuum tube started a revolution in miniaturization that
was inconceivable at the time of its invention and is not
fully recognized even many years later
Miniaturization resulted in the possibility for billions of
transistors to occupy the volume of a vacuum tube or the
first transistor, and it was not the only consequence The
subsequent spin-off developments in allied areas, such asintegrated circuits and the microprocessor, have spawnedentirely new fields of technology It is quite likely that otherareas are now poised for revolutionary developments thatparallel those that have occurred in the electronics indus-try since the advent of the first transistor
These areas include microelectromechanical systems(MEMS) and closely related fields, such as microfluidicsand micro-optical systems Currently, these technologiesinvolve micromachining on a silicon chip to produce nu-merous types of devices, such as sensors, detectors, gears,engines, actuators, valves, pumps, motors, and mirrors on
a micron scale The first commercial product to arise fromMEMS was the accelerometer that was manufactured as asensor for air-bag actuation On the market today are alsomicrofluidic devices, mechanical resonators, biosensors forglucose, and disposable blood pressure sensors that are in-serted into the body
The vast majority of microsystems are made almost clusively on planar surfaces using technology developed tofabricate electronic integrated circuits The fabrication ofthese devices takes place on a silicon wafer, and the de-vice is formed layer-by-layer using standard clean-roomtechniques that include electron beams or photolithogra-phy, thin-film deposition, and wet or dry etching (bothisotropic and anisotropic) Three variations of this conven-tional electronic chip technology can be used, for example,
ex-to make three-dimensional structures that have high pect ratios and suspended beams These include the LIGA(lithographie, galvanoformung, abformung) process (1,2),the Hexsil process (3), and the SCREAM (single-crystalreactive etching and metallization) process (4) The tech-nique most employed, the LIGA process, which was de-veloped specifically for MEMS-type applications, can con-struct and metallize high-aspect-ratio microfeatures This
as-is done by applying and exposing a very thick X-ray sitive photoresist layer to synchrotron radiation Features
sen-up to 600 microns high that have aspect ratios of 300 to
1 can be fabricated by this technique to make truly dimensional objects The Hexsil process uses a mold thathas a sacrificial layer of silicon dioxide to form polysili-con structures that are released by removing the silicondioxide film A third approach is the SCREAM bulk mi-cromachining process that can fabricate high-aspect-ratiosingle-crystal silicon suspended microstructures from a sil-icon wafer using anisotropic reactive ion etching Note,however, that like the conventional technique used to makeelectronic circuits, all of these variations use a layered ap-proach that starts on a flat surface
three-In addition, there are some disadvantages of the tional electronic chip fabrication technique and its modifi-cations, even though there have been numerous and veryinnovative successes using these silicon wafer-based tech-nologies This is due to the fact that these technologiesrequire building up many layers of different materials aswell as surface and bulk micromachining which leads tosome very difficult material science problems that have
conven-to be solved These include differential etching and layingdown one material without damaging any previous layer
In addition, there are the problems of interconnecting ers in a chip that have different functions An example ofthis is a microfluidic device in which there are both fluidic
Trang 4lay-and electronic functions Clearly, there are numerous
ma-terials issues central to this technology
Other technologies are available that, like conventional
lithography, can construct or replicate microscopic features
on a flat surface These approaches include imprint
lith-ography that involves compression molding (5), lasers (6–
8), ion beams (9) and electron beam (10) micro-machining,
soft lithography (11), writing features into the surface
us-ing an atomic force microscope (12,13), and very limited
application of deposition using a scanning tunneling
mi-croscope (14,15) The majority of these technologies are not
discussed in detail because there is not a close link to
mi-crotube technology
In addition to the processing problems mentioned
be-fore, there are other limitations inherent in conventional
lithographic techniques that are based on planar silicon
For example, in some applications such as those that
in-volve surface tension in fluidics, it is important to have a
circular cross section However, it is impossible to make
a perfectly round tube or channel on a chip by
conven-tional technology Instead, channels on the wafer surface
are made by etching a trench and then covering the trench
by using a plate (16,17) This process can produce only
an-gled channels such as those that have a square,
rectan-gular, or triangular cross section Because of the
limita-tions already mentioned, we heartily agree with Wise and
Najafi in their review of microfabrication technology (18)
when they stated, “The planar nature of silicon technology
is a major limitation for many future systems, including
microvalves and pumps.”
In the literature, there are at least two technologies in
addition to microtubes that remove microfabrication from
the flatland of the wafer One uses “soft lithography,” and
the other uses laser-assisted chemical vapor deposition
(LCVD) “Soft lithography,” conceived and developed by
Whitesides’ outstanding group at Harvard, encompasses a
series of very novel related technologies that include
micro-contact printing, micromolding, and micromolding in
cap-illaries (11) These technologies can fabricate structures
from several different materials on flat and curved
sur-faces By example, structures can be fabricated using
mi-crocontact printing by first making a stamp that contains
the desired features This stamp, which is usually made
from poly(dimethylsiloxane) (PDMS) has raised features
placed on the surface by photolithographic techniques The
raised features are “inked” with an alkanethiol and then
brought into contact with a gold-coated surface, for
ex-ample, by rolling the curved surface over the stamp The
gold is then etched where there is no self-assembled
mono-layer of alkanethiolate Features as small as 200 nm can
be formed by this technique However, the
microstruc-tures produced by this technique are the same as those
produced by standard techniques, except that the
start-ing surface need not be flat By usstart-ing these techniques,
submicron features can be fabricated on flat or curved
substrates made of materials, such as metals (19),
poly-mers (20), and carbon (21) In addition, these technologies
can be used to make truly three-dimensional free-standing
objects (22,23)
Another step away from the standard planar silicon
technology is the LCVD process (24,25) which can “write
in space” to produce three-dimensional microsystems In
this process, two intersecting laser beams are focused in avery small volume in a low-pressure chamber The surface
of the substrate on which deposition is to occur is brought
to the focal point of the lasers The power to the lasers
is adjusted so that deposition from the gas phase occursonly at the intersection of the beams As deposition occurs
on the substrate surface, it is pulled away from the focalpoint Under computer control, the substrate can be mani-pulated so that complex, free-standing, three-dimensionalmicrostructures can be fabricated
In addition to LCVD and soft lithography, only crotube technology offers the possibility of truly three-dimensional nonplanar microsystems However, in con-trast to these two technologies, microtube technology alsooffers the ability to make microdevices from practically anymaterial because the technology is not limited by electrode-position or the availability of CVD precursor materials Inaddition, in contrast to these other technologies, microtubetechnology provides the opportunity to make tubing andalso to make it in a variety of cross-sectional and axialshapes that can be used to miniaturize systems, connectcomponents, and fabricate components or systems that arenot currently possible to produce
mi-Microscopic and Nanoscopic Tubes and Tubules
Commercially, tubing is extruded, drawn, pultruded, orrolled and welded which limits the types of materials thatcan be used for ultrasmall tubes as well as their ultimateinternal diameters In addition, it is not currently possi-ble to control the wall thickness, internal diameter, or thesurface roughness of the inner wall of these tubes to a frac-tion of a micron by these techniques Using conventionaltechniques, ceramic tubes are currently available only assmall as 1 mm i.d Copper tubing can be obtained as small
as 0.05 mm i.d., polyimide tubing is fabricated as small as
50µm i.d., and quartz tubing is drawn down as small as
2µm i.d This means that quartz is the only tubing
com-mercially available that is less than 10µm i.d This quartz
tubing is used principally for chromatography
There are, however, other sources of small tubing thatare presently at various stages of research and develop-ment For some time, several groups have been using lipids
as templates (26–28) to fabricate submicron diameter ing These tubes are made by using electroless deposi-tion to metallize a tubular lipid structure formed from aLangmuir–Blodgett film Lipid templated tubes are veryuniform in diameter, which is fixed at∼0.5 µm by the lipid
tub-structure Lengths to 100µm have been obtained by this
technique which is extremely expensive due to the cost ofthe raw materials
Other groups are making submicron diameter tubulesusing a membrane-based synthetic approach This methodinvolves depositing the desired tubule material within thecylindrical pores of a nanoporous membrane Commercial
“track-etch” polymeric membranes and anodic aluminumoxide films have been used as the porous substrate.Aluminum oxide, which is electrochemically etched, hasbeen the preferred substrate because pores of uniformdiameter can be made from 5–1000 nm Martin (29–31)polymerized electrically conductive polymers from the liq-uid phase and electrochemically deposited metal in the
Trang 5pore structure of the membrane Kyotani et al (32,33)
de-posited pyrolytic carbon inside the pores of the same type
of alumina substrate In each case, after the inside walls of
the porous membrane are covered to the desired thickness,
the porous membrane is dissolved leaving the tubules A
variation of this technique, used by Hoyer (34,35) to form
semiconductor (CdS, TiO2, and WO3) nanotubes, includes
an additional step Instead of coating the pore wall directly
to form the tubule, he fills the pore with a sacrificial
mate-rial, solvates the membrane, and then coats the sacrificial
material with the material for the nanotube wall The
sac-rificial material is finally removed to form the nanotube As
in the lipid process, all of the tubules formed by this process
in a single membrane are uniform in diameter, length, and
thickness But in contrast to the lipid process, the diameter
of the tubules can be varied by the extent of oxidation of
the aluminum substrate Although diameters can be
var-ied in this process, it should be clear that these tubules are
limited in length to the thickness of the porous membrane
In addition, the wall thickness is also limited in that the
sum of the inside tubule diameter and two times the wall
thickness is equal to the starting pore diameter
Using a sol-gel method, tubules can be made in about
the same diameter range as in the membrane approach
By hydrolyzing tetraethlyorthosilicate at room
tempera-ture in a mixtempera-ture of ethanol, ammonia, water and tartaric
acid, Nakamura and Matsui (36) made silica tubes that had
both square and round interiors The tubules produced by
this technique were up to 300µm long, and the i.d of the
tubes ranged from 0.02 to 0.8µm By introducing minute
bubbles into the sol, hollow TiO2fibers that have internal
diameters up to 100µm have also been made (37) by using
the sol-gel approach
On an even smaller scale, nanotubules are fabricated
using a number of very different techniques The most
well-known tube in this category is the carbon “buckytube” that
is a cousin of the C60buckyball (38–42) Since carbon
nano-tubes were first observed as a by-product in C60production,
the method of C60formation using an arc-discharge plasma
was modified to enhance nanotube production The process
produces tubules whose i.d is in the range of 1–30 nm
These tubules are also limited in length to about 20
mi-crons Similar nanotubes of BN (43), B3C, and BC2N (44)
have been made by a very similar arc-discharge process
In addition, nanotubes of other compositions (45,46) have
been prepared using carbon nanotubes as a substrate for
conversion or deposition
An alternative technique for manufacturing carbon
tubes that have nanometer diameters has been known to
the carbon community for decades from the work of Bacon,
Baker, and others (47–50) The process produces a
hol-low catalytic carbon fiber by pyrolyzing a hydrocarbon gas
over a catalyst particle The fibers, which vary in
diame-ter from 1 nm to 0.1µm have lengths up to centimeters,
can be grown either hollow or has an amorphous center
that can be removed by catalytic oxidation after a fiber is
formed
Other nanoscale tubules whose diameters are slightly
larger and smaller than buckytubes have been made from
bacteria and components of cytoskeletons and by direct
chemical syntheses Chow and others (51) isolated and
purified nanoscale protein tubules called rhapidosomes
from the bacterium Aquaspirillum itersonii After the
rhapidosomes are metallized by electroless deposition andthe bacteria are removed, metal tubules approximately
17 nm in diameter and 400 nm long are produced ing a similar metallization technique, metal tubes havebeen fabricated (52) whose inner diameters are 25 nm byusing biological microtubules as templates These micro-tubules, which are protein filaments of 25 nm o.d andwhose lengths are measured in microns, are components ofthe cytoskeletons of eukaryotic cells In contrast to tubulesproduced from biological templates, the tubules produced
Us-by direct chemical synthesis involve using the technique ofmolecular self-assembly Some of the nanotubules that fallinto this category are made from cyclic peptides (53), cy-clodextrins (54), and bolaamphiphiles (55) Cyclic peptidenanotubules have an 0.8 nm i.d: and can be made severalmicrons in length Other self-assembled nanotubules thatrange from 0.45 to 0.85 nm i.d have been synthesized fromcyclodextrins (54,56) in lengths in the tens of nanometers.Although it is clear that individual nanotubules are cur-rently useful for certain applications, such as encapsula-tion, reinforcement, or as scanning probe microscope tips(57), it is not obvious how individual nanotubules can beobserved and economically manipulated for use in devicesother than by using a scanning probe microscope (58) Untilthis problem is solved, the future of individual nanotubes
in devices is uncertain However, this problem can be cumvented if the nanotubules are part of a larger body such
cir-as in an array
If oriented groups or arrays of submicron to micron meter tubes or channels perpendicular to the surface of thewafer or device are desired, there are at least four meansavailable to make them Using the technique described be-fore for making anodic porous alumina, a two step repli-cation process (59) can be used to fabricate a highly or-dered honeycomb nanohole array from gold or platinum.The metal hole array is from 1–3 micron thick and has holes
dia-70 nanometers in diameter For smaller tubes or channels,
a technique (60) has recently been developed to draw downbundles of quartz tubes to form an array This process pro-duces a hexagonal array of glass tubes each as small as
33 nm in diameter This translates to a density of 3×1010channels per square centimeter Even smaller regular ar-rays of channels can be synthesized by a liquid crystaltemplate mechanism (61,62) In this process, aluminumsilicate gels are calcined in the presence of surfactants toproduce channels 2–10 nm in diameters Finally, channels
of∼4 nm in cross section can be produced (63) larly to the surface of an amorphous silica film by forminghematite crystals in a Fe–Si–O film and then etching awaythe hematite crystals
perpendicu-Finally, several technologies exist to make channels orlayers of channels of desired orientation in solid objects.These technologies are another spin-off of the photolitho-graphic process used for integrated circuits On a two-dimensional plane, channels that range in size from tens
to hundreds of microns in width and depth have been ricated (16,17) on the surface of silicon wafers by stan-dard microphotolithographic techniques Forming of mi-croscopic channels and holes in other materials originated
Trang 6fab-in the rocket propulsion community fab-in 1964 Work at
Aerojet Inc (64) produced metallic injectors and cooling
channels in metallic parts using a process that included
photolithographic etching of thin metallic platelets and
stacking the platelets followed by diffusion bonding of the
platelets to form a solid metallic object that has
micron-sized channels The group at Aerojet has recently modified
its technique to use silicon nitride Variations on this
tech-nique include electrochemical micromachining and sheet
architecture technology
Electrochemical micromachining (65,66) avoids
gener-ating toxic waste from acid etching by making the thin
metal part covered with exposed photoresist the anode in
an electrochemical cell where a nontoxic salt solution is the
electrolyte Sheet architecture technology (67) developed
at Pacific Northwest National laboratory is used to
fabri-cate numerous microscopic chemical and thermal systems,
such as reactors, heat pumps, heat exchangers, and heat
absorbers These devices may consist of a single
photolitho-graphically etched or laser-machined laminate that has a
cover bonded to seal the channels, as described before, or
may consist of multiple layers of plastic or metal laminates
bonded together
It is quite apparent from this brief and incomplete
review, that a number of very novel and innovative
ap-proaches have been used to make microsystems as well
as tubes and channels whose diameters are in the range
of nanometers to microns In the next section, the basics
of microtube technology which complements these other
technologies are discussed
AFRL MICROTUBE TECHNOLOGY
Properties and Production of Microtubes
Except for self-assembled tubules, the microtube
tech-nology developed at the Propulsion Directorate of the Air
Force Research Laboratory (AFRL) can produce tubes in
the size range of those made by all of the other techniques
cited In contrast to tubing currently on the market and
the submicron laboratory scale tubing mentioned before,
microtubes can be made from practically any material
(in-cluding smart materials) and will have precisely controlled
composition, diameter, and wall thickness in a great range
of lengths In addition, this technology can produce tubes
in a great diversity of axial and cross-sectional geometries
For most materials, there is no upper diameter limit, and
for practically any material, internal diameters greater
than 5 µm are possible In addition, for materials that
can survive temperatures higher than 400◦C, tubes can be
made as small as 5 nanometers by using the same process
To date, tubes have been made from metals (copper,
nickel, aluminum, gold, platinum, and silver), ceramics
(silicon carbide, carbon, silicon nitride, alumina, zirconia,
and sapphire), glasses (silica), polymers (Teflon), alloys
(stainless steel), and layered combinations (carbon /nickel
and silver/sapphire) in sizes from 0.5–410µm Like many
of the techniques described before, microtube technology
employs a fugitive process that uses a sacrificial
man-drel, which in this case is a fiber High-quality coating
techniques very faithfully replicate the surface of the fiber
on the inner wall of the coating after the fiber is removed
By a proper choice of fiber, coating, deposition method, andmandrel removal method, tubes of practically any compo-sition can be fabricated Obviously, a great deal of materialscience is involved in making precision tubes of high qual-ity Some scanning electron microscope (SEM) micrographs
of a group of tubes are shown in Fig 1
Cross-sectional shapes and wall thickness can be veryaccurately controlled to a fraction of a micron, which isnot possible by using any of the approaches cited before.Numerous cross-sectional shapes have already been made,and some of them are shown in Fig 2 These micrographsshould be sufficient to demonstrate that practically anycross-sectional shape imagined can be fabricated As seen
in Fig 2, the wall thickness of the tubes can be held veryuniform around the tube It is also possible to control thewall thickness accurately along the length of the individ-ual tubes and among the tubes in a batch or a continuousprocess It can be seen in Fig 2 that the walls can be madenonporous It will be shown later that the microstructure ofthe walls and extent of porosity that the walls contain canalso be controlled In addition to the possibility of cross-sectional tube shapes, using a fugitive process also allowsfabricating tubes that have practically any axial geometry,
as is shown later
The maximum length in which these tubes can be madehas yet to be determined because it depends on many vari-ables, such as the type of tube material, the composition
of the sacrificial tube-forming material, and the degree ofporosity in the wall It is possible that there is no limitation
in length for a tube that has a porous wall For nonporouswall tubing, the maximum length would probably be in themeter range because there is a direct relationship betweenthe tube i.d and the maximum possible length However,for most applications conceived to date, the length needonly be of the order of a few centimeters Based on a quickcalculation, it is apparent that even “short” tubes have atremendous aspect ratio For instance, a 10-µm i.d tube
25 cm long has an aspect ratio of 2500
Using microtube technology, there is no upper limitation
in wall thickness for most materials To date, free-standingtubes have been made whose wall thickness range from0.01–800 µm (Fig 3a) Most of the microtubes tested to
date have demonstrated surprising mechanical strength
In fact, preliminary studies of both copper and silver tubeswhose wall thickness is in the micron range have shownthat microtubes can have up to two times the tensilestrength of an annealed wire of the same material of thesame cross-sectional area Besides precise control of thetube wall thickness and composition, the interior surface ofthese tube walls can have practically any desired texture ordegree of roughness In addition, the walls can range fromnonporous to extremely porous, as seen in Fig 4, and theinterior or exterior surfaces of these tubes can be coated byone or more layers of other materials (Fig 5),
In addition to free-standing microtubes, solid monolithicstructures that have microchannels can be fabricated bymaking the tube walls so thick that the spaces between thetubes are filled (Fig 6) The microchannels can be randomlyoriented, or they can have a predetermined orientation
Trang 7(a) (b)
(c)
(d)
Figure 1 Examples of microtubes: (a) 10-µm silicon carbide tubes; (b) 410-µm nickel tubes;
(c) 26-µm silicon nitride tube; and (d) 0.6-µm quartz tube.
Any desired orientation or configuration of microtubes can
be obtained by a fixturing process Alternatively,
compos-ite materials can be made by using a material different
from the tube wall as a “matrix” that fills in the space
among the tubes The microtubes imbedded in these
mono-lithic structures form oriented microchannels that, like
free-standing tubes, can contain solids, liquids, and gases,
and as act as waveguides for all types of electromagnetic
energy
Microtube Applications
Discrete thinner walled microtubes are useful in areas
as diverse as spill cleanup, encapsulation of medicine
or explosives, insulation that is usable across a very
wide range of temperature, and as lightweight structural
reinforcement similar to that found in bone or wood The
cross-sectional shape of these reinforcing tubes can be
tail-ored to optimize mechanical or other properties In
addi-tion, thinner walled tubes are useful as bending or
ex-tension actuators when fabricated from smart materials
Thicker walled tubes (Fig 3b: nickel and SS) that are just
as easily fabricated are needed in other applications, such
as calibrated leaks and applications that involve internal
or external pressure on the tube wall
The ability to coat the interior or exterior surface ofthese tubes with a layer or numerous layers of other ma-terials enlarges the uses of the microtubes and also allowsfabricating certain devices For example, applying oxida-tion or corrosion protection layers on a structural or spe-cialty tube material will greatly enlarge its uses A catalystcan be coated on the inner and /or outer tube surface to en-hance chemical reactions The catalytic activity of the tubecan also be enhanced by increasing the porosity in the wall,
as shown before in Fig 4 Multiple alternating conductiveand insulating layers on a tube can provide a multiple-pathmicrocoaxial conductor or a high-density microcapacitor
As stated before, the interior surface of these tube wallscan have practically any desired texture or degree of rough-ness This control is highly advantageous and allows usingmicrotubes in many diverse applications For example, op-tical waveguides require very smooth walls, whereas cat-alytic reactors would benefit from rough walls (Because ofthe fabrication technique, the roughness of the tube wallinterior can be quantified to a fraction of a micron by usingscanning probe microscopy techniques on the mandrel.)
Trang 8(a) (b)
(c)
(d)
Figure 2 Tubes larger than 1µm i.d can be made in any cross-sectional shape such as (a) 17-µm
star, (b) 9× 34-µm oval, (c) 59-µm smile, and (d) a 45-µm trilobal shape.
Figure 3 Tubes can be structurally sound and have (a) very thin walls or (b) thick walls.
Trang 9Figure 4 Microtube that has a porous tube wall.
Microtubes can be made straight or curved (Fig 7),
or they can be coiled (Fig 8) Coiled tubes whose coils
are as small as 20µm can be used, for example, as
flex-ible connectors or solenoid coils For the latter
applica-tion, the coils could be of metal or of a high temperature
superconductor where liquid nitrogen flows through the
(a)
(b)
Figure 5 (a) Sapphire tube that has a silver liner (b) Nickel tube
that has a silver liner.
Figure 6 Solid nickel structure that has oriented microchannels.
tube Another application for coils is for force or sure measurement No longer are we limited to quartz mi-crosprings Using microtube technology, the diameter andwall thickness of the tube, the diameter of the coil, thetube material, and the coil spacing can be very precisely
pres-(a)
(b)
Figure 7 Examples of curved silver tubes: (a) single tube;
(b) multiple tubes.
Trang 10(a) (b)
Figure 8 (a) Section of “large” coiled tube (b) Open end of coiled tube.
controlled to give whatever spring constant is needed for
the specific application In addition, these microcoils can
be made from a variety of smart materials and used as
actuators or sensors For example, the length of a spring
made from Nitinol® can easily be changed by applying
heat It is also possible to wrap one or more coiled spring
tubes around a core tube (Fig 9) Applications for this kind
of device range from a counterflow heat exchanger to a
screwdrive for micromachines (For the screw application,
the wrapped coil cross section could be made rectangular.)
Like coiled spring tubes, bellows can be used as
microin-terconnects, sensors, and actuators and can be made in
practically any shape imaginable Figure 10a shows a
bel-lows that has a circular cross section, and the belbel-lows in
Fig 10b has a square cross section and aligned bellows
segments The bellows in Fig 10c is square and has a
twist A slightly more complex bellows shown in Fig 10d
is a tapered-square camera bellows that has a sunshade
to demonstrate the unique capability of this technology It
demonstrates the ability to control cross section and
ax-ial shape and to decrease and increase the cross-sectional
Figure 9 A coiled tube wrapped around a tube or fiber that can
be used as a heat exchanger or as a microscopic screwdrive.
dimension in the same device Bellows fabricated by crotube technology can have a variety of shaped ends forconnections to systems for use, for example, as finned heatexchangers, hydraulic couplings for gas and liquid, or staticmixers for multiple fluids The bellows in Fig 10e has athicker transitional region and a dovetail on the end forconnection to a device machined on a silicon wafer The fe-male dovetail to mate with this bellows is a commerciallyavailable trench design (68) on a silicon wafer that pro-vides a way to attach the bellows to the wafer, which can bepressurized by using proper sealing (No other technologyavailable can join a fluidic coupling to a wafer for pressur-ization to relatively high pressures.)
mi-If one end of the bellows is sealed, an entirely new group
of applications becomes possible For example, if a bellowsend is sealed, the bellows can be extended hydraulically
or pneumatically In this configuration, a bellows could beused as a positive displacement pump, a valve actuator, orfor micromanipulation As a manipulator, a single bellowscould be used for linear motion, three bellows could be or-thogonally placed for 3-D motion, or three bellows could
be attached at several places externally along their axes(Fig 11) and differentially pressurized to produce a bend-ing motion This bending motion would produce a microfin-ger, and several of these fingers would make up a hand Thelarge forces and displacements possible by using this tech-nique far surpass those currently possible by electrostatic
or piezoelectric means and fulfill the need expressed
by Wise and Najafi (18) when they stated that “In thearea of micro-actuators, we badly need drive mechanismscapable of producing high force and high displacementsimultaneously.”
For most applications, it is necessary to interface crotubes and the macroworld This is possible in a num-ber of ways For example, a tapering process can be used
mi-in which the diameter is gradually decreased to micron mensions Alternatively, the tubes and the macroworld can
di-be interfaced by telescoping or numerous types of ing schemes (Fig 12) An example of a thin-walled 5-µm i.d.tube telescoped to a 250-µm o.d tube is shown in Fig 13
manifold-A tube of this type could be used as a micropitot tube and, ofcourse, could be made more robust by thickening the walls
Trang 11(a) (b)
(e)
Figure 10 (a) A conventional round bellows (b) A straight bellows that has a square cross section
(c) A square bellows that has a twist (d) A tapered square camera bellows that has a sun shade to demonstrate the versatility of the technique (e) A round bellows that has a dovetail connector.
Although microtube technology has unique capabilities,
it should be obvious that no single technology can fill all of
the requirements imposed by diverse applications Thus,
microtube technology cannot easily compete with other
technologies in certain applications One of these involves
gas and liquid separation such as in chromatography For
example, quartz tubing that can be extruded and drawn
in very long lengths is inexpensive and available in cron dimensions However, note that even in areas such
mi-as separation, there are niches for microtubes that volve the composition of the tube material, the cross-sectional shape, or the inner wall coating For example,Fig 14 shows microtubes manifolded to a tubular framefor a specific gas separation that requires microtubes
Trang 12Figure 11 Microtube bellows finger: (a) unpressurized; (b) pressurized.
of a specific composition, precise diameter, and wall
thickness
Currently, these tubes have been made by a batch
pro-cess in the laboratory, but the technique is equally suited
to a continuous process which would be more efficient and
also much easier in some cases Obviously, a continuous
process would reduce costs For most materials, costs are
already rather low because, unlike some other processes,
expensive tooling is not required For many materials such
as quartz, aluminum, and copper, the anticipated cost is
(a)
(b)
(c)
(d)
Figure 12 Different ways of transitioning microtubes to the real world: (a) taper, (b) telescope,
(c) bundle, and (d) manifold.
∼$0.01/cm for thin-walled tubes For precious metals such
as gold or platinum, the cost would be significantly higherdue to the cost of raw materials
Microtubes have almost universal application in eas as diverse as optics, electronics, medical technology,and microelectromechanical devices Specific applicationsfor microtubes are as diverse as chromatography, encap-sulation, cross- and counterflow heat exchange, injectors,micropipettes, dies, composite reinforcement, detectors,micropore filters, hollow insulation, displays, sensors,
Trang 13ar-(a) (b)
Figure 13 (a) A thin-walled 5-µm i.d tube telescoped to a 250-µm o.d tube (b) View of the small
open end of the telescope.
optical waveguides, flow control, pinpoint lubrication,
crosponges, heat pipes, microprobes, and plumbing for
mi-cromotors and refrigerators The technology works equally
well for high- and low-temperature materials and appears
feasible for all applications that have been conceived to
date As can be seen, there are numerous types of devices
that have become possible as a result of microtube
tech-nology One category of devices that is highlighted is that
based on surface tension and wettability
MICROTUBE DEVICES BASED ON SURFACE TENSION
AND WETTABILITY
Now, there is great interest in developing microfluidic
sys-tems to decrease the size of current devices, increase their
speed and efficiency, and decrease their cost because
mi-crofluidic systems have the potential, for example, for
dras-tically decreasing the cost of certain health tests, allowing
implantable drug delivery systems, and very significantly
reducing the time needed to complete the Humane Genome
Project Microtube technology based on surface tension and
wettability is unique in its capabilities and is truly an
en-abling technology in the microfluidic field
Figure 14 Microtubes are manifolded to a tubular frame for gas
separation.
As miniaturization of mechanical, electrical, and fluidicsystems occurs, the role of physical and chemical effectsand parameters has to be reappraised Some effects, such
as those due to gravity or ambient atmospheric pressure,are relegated to minor roles or can even be disregardedentirely as miniaturization progresses Meanwhile, othereffects become elevated in importance or, in some cases,actually become the dominating variables This “downsiz-ing reappraisal” is vital to successful miniaturization In
a very real manner of speaking, new worlds are enteredinto in which design considerations and forces that arenormally negligible in real-world applications become es-sential to successful use and application of miniaturizedtechnology
Surface tension and wettability are closely related nomena that are greatly elevated in importance as minia-turization proceeds Surface tension involves only thestrength of attraction of droplet molecules for one an-other (cohesive forces), but wettability also includes thestrength of attraction of droplet molecules to molecules
phe-of the wall material (adhesive forces) It is important
to realize that surface tension and wettability are ally not comparable in effect to normal physical forces atmacroscopic levels For example, surface tension is usuallyignored when determining fluid flow through a pump ortube Its effect is many orders of magnitude smaller thanpressure drop caused by viscosity because the difference
usu-in pressure P between the inside of a droplet and the
outside is given by the Young and Laplace equation ofcapillary pressure (69,70):
in the case of balloons, however, confining forces in surfacetension are caused by the affinity of molecules of dropletmaterial for one another Because molecules are missing a
Trang 14binding partner looking outward on the surface of a drop,
they pull on their nearest neighbors
Normally, droplet dimensions in most macroscopic
ap-plications are measured in thousands of microns
There-fore, pressure differences due to surface tension are
incon-sequential and typically measure far less than atmospheric
pressure For comparison, pressure drops resulting from
viscous flow are typically of the order of magnitude of tens
of atmospheres When r is of the order of microns, however,
pressure differences due to surface tension become
enor-mous and frequently surpass tens of atmospheres This is
precisely the reason that fine aerosol droplets are so
diffi-cult to form However, the formation of tiny droplets is not
specifically the focus of discussion here, but rather their
behavior in miniature voids, such as cavities, capillaries,
and channels that are shaped so that they partially confine
the droplet The position of droplets within such microvoids
is governed by the surface tension of the droplet fluid, the
wettability of the fluid with respect to microvoid walls
con-tacted during displacement, the geometric configuration of
the walls that confine the fluid droplets, and any pressure
external to the droplet Microdevices fabricated from these
microvoids can be made to operate when wettabilities are
greater than or less than 90◦, but not exactly 90◦ They
can operate using either nonwetting or wetting fluids The
difference between wetting and nonwetting fluids in
capil-laries can be explained by using Fig 15
In Fig 15a, a nonwetting fluid droplet is forced into a
single microtube An insertion pressure has to push the
nonwetting droplet inside the microtube because of the
re-pulsion between the droplet and the walls Once it is inside,
however, no further pressure is necessary In fact, any
pres-sure simply moves the nonwetting droplet along the
micro-tube at a velocity determined by the applied pressure and
the frictional forces between the droplet and the microtube
wall Note that the nonwetting droplet becomes elongated
when it is constrained in the capillary and has a
convex-shaped interface along the axis of the capillary In addition,
it can be seen that the radius of the nonwetting droplet is
now greater than the radius of the microdevice tube and
that the contact angleθ with the capillary surface is
be-tween 90◦and 180◦, which is the contact angle for a totally
nonwetting droplet In contrast, the situation is very
differ-ent if the fluid totally wets the microtube surface, as seen
Dropletradius
Non-wettingdroplet
Wettingdroplet
Press
(b)
Figure 15 Behavior of fluid droplets in capillaries: (a)
nonwet-ting droplet; ( b) wetnonwet-ting droplet.
in Fig 15b In this case, the fluid is sucked into the tube, and fluid flow is governed only by frictional forces.This is the situation in normal macroscopic applications.For wetting fluids, the ends of the droplets are concave be-cause the walls of the microdevice are wet by the dropletand attract the droplet molecules The contact angle forwetting fluids is between 0 and 90◦, 0◦ indicates a totallywetting fluid In this article, the term nonwetting refers
micro-to a contact angle greater than 90◦, and the term wettingmeans a contact angle less than 90◦
The behavior of a microtube device that employs wetting droplets is easily understood if one compares it tothe mercury intrusion method (71–73) of measuring thepore-size distribution within porous solids This technique
non-is based on the understanding that the pressure needed toforce a nonwetting fluid into a capillary or a pore in a solid
is given by the relationship proposed by Washburn (71):
whereθ is the contact angle of the fluid with the material under test, P is the external pressure applied to the non- wetting fluid, and r is the radius of the capillary or pore
which, act is assumed for simplicity, is spherical and has
a constant diameter This equation is valid for any fluid incontact with a capillary or porous solid whose contact an-gle is greater than 90◦ Once the external applied pressureexceeds that needed to insert the nonwetting fluid into aconstant-diameter capillary or pore, the nonwetting fluidflows into that particular diameter capillary or pore until
it fills it Then, the volume of the intruded fluid is a directmeasure of that particular capillary’s or pore’s void volume
If a smaller capillary or pore branches off the larger meter void, it remains unfilled until the insertion pressure
dia-is radia-ised sufficiently high that Eq (2) dia-is again satdia-isfied,and the process repeats itself
In contrast to the mercury intrusion method of mining pore volume, instead of determining the pore vol-ume, the emphasis in devices based on microtube tech-nology is placed on the movement and the position of thedroplet in the confining voids These droplets can be wet-ting or nonwetting As will be apparent later, a myriad ofsmart microdevices are based on surface tension and wet-tability
deter-Because these microdevices have no moving cal parts, they are very reliable, can be used in both staticand dynamic applications, and are very rugged They canexperience pressures or forces far beyond their normal op-erating range and still return to their original accuracy andprecision In addition, unlike technology built up on a sili-con wafer, these microdevices can be made from practicallyany material Thus, high-temperature microdevices can
mechani-be fabricated by properly choosing the device and dropletmaterial
Devices That Use the Interaction of Nonwetting Droplets and Gases and Wetting Fluids
This group of devices uses the surface properties of als, primarily surface tension and wettability, as the prin-cipal means of actuating and controlling motion by and
Trang 15Figure 16 Nonwetting droplet inserted into a microtube under
pressure (a) Constant diameter tube (b) Tube that has a
transi-tion to a smaller diameter.
within microtube devices These devices, which have no
moving mechanical parts, can perform mechanical tasks
whose scale of motion is measured in microns
These devices, similar to other microdevices based on
surface tension and wettability, are composed of various
sizes of nonwetting droplets inserted into microscopic voids
of various shapes and sizes These voids can be in the form
of cavities, capillaries, and channels that are shaped so
that they partially confine the droplet Gas or wetting fluid
is placed in the microcavities along with the nonwetting
fluid During operation, the nonwetting droplets move in
response to fluid or gas pressure or vice versa Specifically,
these nonwetting droplets may translate within a void of
the microtube device that is filled with the gas or wetting
fluid, translate from one void space to another, or rotate
in a fixed position Microtube devices of this type can stop
fluid flow or act as a check valve, a flow restricter, a flow
regulator, or a gate, for example The minimum dimension
of the voids in these devices typically ranges from about
20 nm to about 1000µm.
In Fig 16a, a non-wetting fluid droplet is forced through
a single microtube An initial insertion pressure has to
push the nonwetting droplet inside the microtube If the
diameter of the microtube in Fig 16a decreases at a
cer-tain point to form a telescoping microtube (Fig 16b), a
considerably higher pressure must be applied by a gas
or wetting fluid to squeeze the nonwetting drop into the
smaller section of the microtube In contrast, if a wetting
fluid is employed instead of a gas, as before, it is also sucked
into the smaller diameter section, completely filling all the
available space in the microcavity By inserting an
appro-priately sized nonwetting droplet into a tapered microtube
or a microtube that has a transition to a smaller
dimen-sion that is filled by a second fluid that wets the tube walls,
all flow of the wetting fluid can be stopped by applying a
pressure that forces the nonwetting droplet to block the
Non-wetting
tube
Bypasstube
(a)
Non-wettingdroplet
Bypasstube
Bypasstube
(b)
Figure 17 Microtube check valve: (a) flow possible through
by-pass tubes; (b) flow is blocked.
entrance to the smaller section of the cavity This is thesituation in Fig 16b where the nonwetting droplet hasbeen forced to the intersection of the larger and smallermicrotube sections by the flowing gas or wetting fluid.Figure 17a,b illustrates an extension of this concept Byadding additional small-diameter bypass-flow paths to oneend of a doubly constricted tube, flow is possible only inthe direction of the end that has the added flow paths at-tached to the cavity Of course, these bypass tubes must beproperly sized to prevent nonwetting droplets from squeez-ing into them This microtube device in Fig 17 acts as acheck valve and has no solid moving parts This cannot
be achieved at the macroscopic level because forces thatarise from surface tensions of fluids are too small due tothe much larger geometries employed
Figures 18 and 19 are further extensions of this sameconcept In Fig 18, bypass tubes are left off the microtubecheck valve and convert it to either a microtube flow limiter(Fig 18) or a microtube flow restricter (Fig 19) In Fig 18,because the nonwetting droplet and the larger tube wallform a seal, the only wetting fluid flow that can occur ineither direction when the nonwetting droplet travels backand forth is equal to the volume of the larger tube sectionminus the volume of the nonwetting droplet In Fig 19, thediameter of the nonwetting droplet is now smaller than thediameter of the larger microtube section but larger than
Non-wettingdroplet
Figure 18 Microtube flow limiter.
Trang 16Non-wetting droplet
Figure 19 Microtube flow restricter.
the diameter of the smaller microtube Thus, flow can take
place around the nonwetting droplet However, fluid flow
is not merely restricted, but is entirely stopped if there is
enough flow to push the drop to one end that blocks the
smaller tube
Figure 20 illustrates a microtube pressure/flow
regu-lator In this device, bypass tubes have openings or are
open along their entire lengths to a conically shaped
tran-sitional region placed between the larger and smaller
dia-meter tubes Furthermore, the lengths of the joined bypass
tubes (now better described as bypass channels) up to the
conical transitional region can be varied Increased
pres-sure or flow forces the nonwetting droplet farther into the
conical transitional region and exposes more flow channel
openings to wetting fluid The result is increased flow of the
gas or wetting fluid as a function of pressure By suitably
sizing the nonwetting droplet, properly orienting the
de-vice, correctly shaping the transitional cone, and precisely
positioning bypass channels, this device can also function
as a microtube pressure-relief valve; no flow occurs
un-til some predetermined pressure is exceeded Then, flow
takes place as long as pressure is maintained Note that
only two bypass flow channels are shown in Fig 20 This
was done to simplify the drawing Any convenient
num-ber, one or more, of channels can be employed Finally, by
making bypass-flow channels vary in cross-sectional area,
uniformly increasing or decreasing flow can be produced
as a function of pressure
In addition to a check valve, it is possible to use
nonwet-ting fluids to make a positive closure valve that has zero
dead space to control a gas or wetting fluid Figure 21a,b
Wettingfluid
Non-wetting
droplet
Conicaltransition
Bypasstube
Bypasstube
Figure 20 Microtube flow or pressure regulator.
Non-wettingfluid
Inletduct
Filltube
HeaterEnd
bulb
Outletduct
Micro-channel
(a)
Inlet fluid
HeaterNon-wetting
fluid
Micro-channel
Inletduct
(b)
Figure 21 Positive closure microtube valve that has zero dead
space: (a) top view; (b) side view.
illustrates a microvalve composed of a fill tube joined to
an end bulb, where two microchannels are attached to thefill tube In this example, the nonwetting droplet controlsthe flow of a wetting fluid or gas through a microchannelwhose thickness is less than that of the fill tube In this mi-crovalve, an inlet fluid flows through an inlet duct and theninto one of the microchannels If the fill tube is not blocked
by the nonwetting droplet, the inlet fluid traverses the blocked fill tube at the point where both microchannels at-tach to it Then, the fluid exits the microvalve through anoutlet duct as outlet fluid In Figure 21a,b, the nonwetting
Trang 17un-A B OUT 0 1 0 1
0 1 0 1 1
0 0 1
Non-wettingdroplet 1
Non-wettingdroplet 2
Figure 22 Microfluidic logic circuit that acts as comparator:
(a) output; (b) no output.
droplet is activated by a heater and only partially fills the
fill tube Obviously, many other forms of activation are
pos-sible, and it is possible to assemble these microvalves in
parallel to control large flows of liquids or gases
A final category of microfluidic devices in which one fluid
controls another can be understood by the more complex
examples given in Fig 22 These figures present microtube
devices that use surface tension and wettability in fluidic
logic circuits that are fully digital, not analog The device
in Fig 22 functions as a comparator; there is an output
only if the inputs are equal Thus, if pressure is applied
to either branch A or B, the gate (nonwetting droplet 2)
closes, as in Fig 22b, and no flow occurs (and no pressure
is transmitted) between branches C and D If equal
pres-sure is applied to A and B or no prespres-sure is applied to A
and B, the gate remains open as in Fig 22a, and flow occurs
(and pressure is transmitted) between C and D
Nonwet-ting droplet 1 is returned to the center position
when-ever pressure is removed because surface tension always
minimizes droplet surface area and a sphere has the
low-est surface area per unit volume of any object Only at the
center position can it be a sphere, and unless placed under
unbalanced force by pressure from A or B, it remains at
the center Numerous other types of logic circuits, such as
OR, NOR, AND, and NAND gates, can also be fabricated inthis manner By combining a number of these logic compo-nents in a suitable arrangement, digital operations can beperformed identically to those of electrical devices Instead
of electricity being on or off in a circuit, pressure is applied
or not applied, and fluid flow does or does not occur
Microdevices Based on the Positions and Shapes of Nonwetting Droplets
In this group of microdevices, the basic principle of eration is the movement or shape change of nonwettingdroplets in tubes, channels, or voids that have at least onemicroscopic dimension This movement on shape changeresults from external or internal stimuli The change indroplet shape depends on the cavity shape and alwaysminimizes the surface free energy of the droplet The cavitythat constrains the droplet in these devices can be sealed
op-or can have one op-or mop-ore openings The shape of this ity determines the reaction of the droplet to a stimulus,
cav-as well cav-as the use of the microdevice, and the output thatcan be obtained from it Uses for these microdevices are asdiverse as sensors, detectors, shutters, and valves
As just stated, microtube sensors based on surface sion and wettability are one type of device in this group.Some of these sensors respond to one or more external stim-uli such as pressure, temperature, and gravity or acceler-ation by changes in the displacement or shape of liquidinterfaces contained within microtubes and/or microchan-nels that have either fixed or variable axial geometries andcircular or noncircular cross-sectional profiles Other sen-sors respond to internal stimuli, such as a change in surfacetension of the liquid droplet or a change in the wettability
ten-of the microdevice’s internal walls Some ten-of these sensorscan quantify the displacement or change in shape of theconstrained droplet that is results from external or inter-nal forces acting on it
An example of one of the simplest microdevices in thisgroup of devices is a microtube pressure sensor (Fig 23)that uses a nonwetting fluid in the form of a droplet Fig-ure 23a illustrates the position of the droplet when the
entrance pressure Pent is equal to the device pressure Pdev.Figure 23b illustrates the position of the droplet when the
entrance pressure is greater than the device pressure Pdev,
and Fig 23c illustrates the position of the droplet whenthere is a much higher entrance pressure This sensordemonstrates the reaction of such a device to an outsidestimulus which in this case is an increase in externally ap-
plied pressure Pent As can easily be seen, the shape of thenonwetting droplet changes in reaction to increases in the
applied external pressure Pent More precisely, increasing the external pressure Pent, that acts through an entrance
microtube squeezes the droplet into ever smaller diameterlocations within a microcavity, which results in displacingthe nonwetting interface toward the smaller diameter end
of the device For this type of sensor, this microcavity may
be tear shaped, circular, or have practically any shape, aslong as there is a change in at least one dimension andthis dimension is from 0.003–1000µm For simplicity, the pressure on the smaller side of the microdevice Pdev, which
Trang 18opposes the external pressure Pent, is set at zero in this
fig-ure Pdevis most easily thought of as a residual gas pressure
left over inside the device from the actual fabrication
pro-cess It does not have to be zero, as shown later The only
requirement for this type of sensor is that Pdevbe less than
Pent and that both pressures be smaller than the
burst-ing strength of the walls of the microtube pressure sensor
It should be apparent from Fig 23 that an overpressure of
the device will push the droplet further into the device tube
than designed for, but when the pressure is released, the
droplet will return to its equilibrium position As long as
the walls of the device have not been damaged and
main-tain their original shape, the sensitivity and accuracy of
this device and the others that are described following will
be unaffected by overpressure
The reaction of the droplet to external pressure is
eas-ily calculable from surface tension theory (the change in
radius of the smaller end of the droplet is inversely
pro-portional to applied the external pressure), but the
ac-tual decrease in radius and resulting displacement of the
nonwetting interface can be understood only intuitively or
observed visually by microscope, as presented in Fig 23
Figure 24 illustrates a modification of this microtube
pres-sure sensor based on surface tension/wettability which
enables nonvisual determination of the displaced
inter-face This nonvisual response to the reaction (movement
or shape change) of the droplet interface caused by a
stim-ulus can take many forms One of these is a change in
electrical resistance A center contact, which has a
mea-surable electrical resistance, is inserted through the
mi-crotube device and establishes electrical contact with the
nonwetting droplet This center contact can be a wire, tube,
Meter
Resistance wireDroplet
a very high resistance per unit length compared to the sidecontact, the actual droplet itself, and compared to the re-mainder of the circuit that connects both contacts to theresistance measuring apparatus
As stated previously, the opposing pressure Pdev neednot be zero It has been set at zero thus far for simplicity.For this type of sensor, it merely needs to be less than the
externally applied pressure Pent; otherwise, the nonwetting
droplet could be expelled from the microtube pressuresensor
Note here that the devices shown schematically canmeasure a variety of external or internal stimuli The pres-sure sensor in Fig 24, for example, could also measure ac-celeration and oscillation along the device axis as well asrotation and temperature, which affect both the thermalexpansion and the surface tension of the droplet If anothercenter contact is also placed in the device on the end op-posite the present center contact, the device can measureacceleration in two directions In addition, it should be ap-parent that to measure parameters such as temperature,rotation, acceleration, or oscillation, it is not even neces-sary for the entrance tube and the device tube to be open
to the atmosphere Thus, to measure these external uli or some internal stimulus, a totally sealed cavity wouldfunction as well as the open pressure sensor in Fig 24.For simplicity, only pressure sensors are shown sche-matically, and it should be understood that the deviceswork equally well in reaction to many other stimuli Apartial list that includes vibration, acceleration, rotation,temperature, electromagnetic fields, and ionizing radiationdemonstrates the broad scope of this sensor technology.When a wetting droplet is employed in place of a non-
stim-wetting droplet, instead of needing Pentto reach some valuegiven by Eq (2) to force the droplet into the microtube, itgoes in automatically This behavior is often referred to as
“wicking.” In contrast to the nonwetting droplet, no sure is needed to get the drop into the tube The fact thatwetting droplets behave similarly to nonwetting droplets in
Trang 19Figure 25 Microtube pressure sensor that has a “Yes-No”
straight port.
some respects means that microdevice sensors can employ
either wetting or nonwetting fluids as the droplet
mate-rial and still serve the same function For microtube
pres-sure sensors, this would require switching the side of the
droplet that actually “feels” the applied pressure
Microde-vices that sense other stimuli would also need similar kinds
of modification These would be specific for the actual
sens-ing application
Figure 25a,b illustrates the relationship between the
radius of curvature of the small end of the drop and the
various pressures involved for a microtube pressure sensor
that has a digital type of response; this will hereafter be
referred to as a “Yes-No” response In Fig 25a, the applied
pressure Pent is not sufficient, compared to Pdev, to force
the nonwetting droplet into the straight port Therefore,
the continuity apparatus registers an open circuit, or “No”
response In Fig 25b, the applied pressure Pentis sufficient
to force entry of the nonwetting droplet into the straight
port Once the droplet has entered the straight port, it
completely fills it, and the continuity apparatus registers a
closed circuit, or “Yes” response As mentioned previously,
this same kind of “Yes-No” response can be duplicated
by using wetting fluids However, because a wetting fluid
would spontaneously wick into the smaller diameter tube,
the only difference would be that now an applied pressure
of sufficient magnitude would need to be directed to Pdev
to expel the wetting droplet from that same straight tube
Therefore, for a wetting fluid, the continuity apparatus
would work in reverse to the “Yes-No” response for a
non-wetting droplet In this case, Pdevmust be greater than Pent,
and therefore, an open circuit signifies “Yes,” and a closed
circuit signifies “No.” However, because a wetting droplet
adheres to the microdevice walls, including the straight
port, fluid remaining on these surfaces might compromise
the accuracy of the continuity apparatus Therefore, it
is preferable to use nonwetting droplets in this kind of
sensor This same logic applies to most microdevices based
on surface tension and wettability, and so in the discussion
that follows, only nonwetting behavior is illustrated
There are obviously many other means for
measur-ing displacement of a nonwettmeasur-ing droplet Other basic
electrical parameters that can be employed are capacitance
and inductance Note here that for all of the tioned techniques for measuring displacement of a nonwet-ting droplet by using electrical means, the electrical proper-ties of the nonwetting droplet must, of course, be suitablefor the measurement technique employed For some ap-plications, the resistance of the nonwetting droplet must
aforemen-be sufficiently low to permit measuring the resistivity ofthe center contact accurately enough for the application,
at hand For other applications, the conductivity must behigh enough to enable measuring capacitance accurately.For certain applications, permeability must be sufficientlydifferent between the nonwetting droplet and its surround-ing medium in the microdevice to allow measuring induc-tance accurately enough to satisfy the demands of the de-sired application These electrical property requirementsare most likely to be different, depending on the measuringtechnique employed and the particular application beingdeveloped Note that it is also possible to combine two ormore readout techniques in a single device
For either multirange or redundancy-driven tions, a great deal of variation is possible These varia-tions are in the form of identical or different devices, cavity,and/or channel or tube configurations, as well as identical
applica-or different types of readouts Many different types of vice channel or tube configurations are possible that willgive either linear or nonlinear responses, as well as analog
de-or digital responses to the stimuli being sensed Fde-or ample, a gradual taper would produce a linear response,whereas a very rapid taper would give a nonlinear re-sponse In another example, a device such as that shown inFig 25 could be modified with a tapered section to followthe constant dimension tube This would result in a dig-ital response followed by an analog response In addition
ex-to these differences in individual sensors, multiple sensorscould all be used together simultaneously or switched on
or off as needed
As mentioned previously, the presence or absence ofnonwetting material in a straight tube (Fig 25) enablesthe pressure sensor or other microdevice that derives itscapabilities from surface properties of materials to func-tion in a digital or a “Yes-No” mode of response Figure 26illustrates another very simple kind of “Yes-No” readout re-sponse for a pressure sensor that does not have a straighttube Now, the center contact in Fig 24 has been trun-cated, so that it does not make contact with the non-wetting droplet for low values of the pressure difference
between Pentand Pdev This lack of contact, or gap, is shown
in Fig 26 The truncated center contact makes contactonly with the nonwetting droplet once a predetermined
Trang 20Figure 27 Microtube pressure sensor that has a variable
“Yes-No” central contact.
pressure difference exists (Pent greater than Pdev) Once
this occurs, the continuity meter signals that contact has
been made, and the desired “Yes-No” readout response is
provided Once sufficient pressure difference has been
es-tablished, the truncated center contact can then function
as a center resistance contact, center contact, some other
kind of readout implement, or some combination thereof
The truncated center contact should not move relative to
the microdevice walls in this simple form of “Yes-No”
read-out microdevice It should be apparent that devices of this
type that have a truncated center contact can also serve as
an electrical switch, based on surface tension and
wettabi-lity, that can be made to operate independently of gravity,
can be impervious to radiation, and can be activated by
numerous stimuli
A more sophisticated “Yes-No” readout microdevice
pressure sensor is illustrated in Fig 27 Now the truncated
center contact is attached to a positioner, which, in turn, is
attached to a positioner holder, which is itself held firmly
in place relative to the actual microdevice walls In Fig 27,
the positioner holder is shown attached to the microdevice
walls The positioner is any type of device that can move the
truncated center contact relative to the microdevice walls
in a predetermined fashion Examples of such positioning
devices are numerous They can be the type where the
op-erator sets the gap and thereby controls the device’s
sensi-tivity, such as a pressurized microbellows and a
piezoelec-tric crystal Alternatively, the positioning device can be the
type that is influenced by its environment, such as those
made from photostrictive, chemostrictive, electrostrictive,
or magnetostrictive materials, which change length due to
light, a chemical environment, or an electric or magnetic
field In these types of “smart”materials, the positioner
can be controlled in real time by its environment, and
thus the device can respond to two stimuli simultaneously
Moreover, using any such positioning devices, the gap can
be changed by a feedback circuit By altering the size of the
gap either before or during actual operation of the
microde-vice, the amount of pressure difference needed between Pent
and Pdev to establish contact and thereby evoke the
“Yes-No” readout response or continuity, as measured by the
continuity apparatus, can be changed Thus, the sensitivity
of this device can be changed by an operator, by its
environ-ment, or by a feedback circuit As before, once continuity
has been established for the simple “Yes-No” readout
re-sponse microdevice of Fig 27, the truncated center contact
exam-as the resistance meexam-asuring apparatus shown in Fig 24
If this is done, both digital and analog readout responsescan be garnered from the same sensor More than one cen-ter contact of different lengths and/or more than one sidecontact can also be employed in Fig 27, thereby providingmultiple digital responses from one device
Note that the sensing techniques mentioned thus farhave all been relatively simple and have employed prin-ciples of physics that are intuitively easy to understand:changes in resistance, capacitance, or inductance Anothersimple technique for detecting the position of a droplet in-terface is using an electromagnetic beam impinging on adetector that is blocked by the advancing surface of thedroplet This type of arrangement can basically give only
a “Yes”-“No” response Two other techniques that can also
be employed to monitor displacement of the nonwettingdroplet interface are optical interference and electron tun-neling These techniques are capable of much higher lev-els of resolution of the nonwetting droplets’ displacement,which results in greater levels of sensitivity
Figure 28 illustrates the readout technique that ploys optical interference The only additional require-ment that must be imposed to use this technique is thatthe nonwetting droplet must reflect at least some of theelectromagnetic radiation input through the fiber-opticsinput/output cable back through the same cable If theseconditions are met, an interference pattern can then begenerated between the incoming and outgoing rays of ra-diation that can be detected by a suitable apparatus located
em-at the opposite end of the fiber-optics input/output cable.This interference pattern will be highly dependent on theposition of the internal interface of the nonwetting droplet,
as well as on the wavelength of radiation employed fore, it is an extremely accurate technique for monitoringany displacement of that interface
There-Figure 29 illustrates the readout technique for tron tunneling There are two primary differences betweenthis readout technique and the previous readout tech-nique that employs a truncated center contact, as illus-trated in Fig 27 In this apparatus, the truncated centercontact is replaced by a very sharp needle-shaped elec-trode In addition, the continuity apparatus is replaced by
elec-a much more sensitive electron tunneling current tor that can measure the tiny electrical currents gener-ated when the needle-shaped electrode moves very close
detec-to the internal interface and creates gaps of the order ofatomic dimensions As in optical interference, tunnelingcurrent measurements are many times more sensitive to
Trang 21Droplet interfaceTunneling tip
(Tunneling current)
Figure 29 Microtube pressure sensor that has a tunneling
cur-rent readout.
displacements of the internal interface than simpler
read-out techniques discussed initially Obviously, any other
de-tecting technique used in scanning probe microscopy, such
as atomic force, magnetic force, or capacitance, can be used
in place of the needle-shaped electrode and the
current-detection circuit In addition, the sensitivity of these
de-vices, as in the device in Fig 27, can be changed by varying
the gap between the tip and the droplet
At this point, it is well worth mentioning again that the
movement of a droplet in a microscopic tube, channel, or
void and the process of remote measurement of
displace-ments of the internal interface itself is of critical
signifi-cance to the devices shown, not the actual kind of remote
measurement technique employed Whatever technique is
employed affects only the accuracy of the remote readout
Thus, regardless of the measurement technique,
displace-ments of the internal interface in reaction to external
stim-ulus will be the same and will depend only on the surface
tension and wettability of the sensor components and on
sensor geometry
Until now, it has been tacitly assumed that motion of
the internal interface during any remote readout of its
displacement is negligible This is not necessarily so Any
measurement of the position of the internal interface will
take some finite amount of time If there is motion of the
in-ternal interface during this finite measurement time, the
position of the internal interface will be some sort of
av-erage readout If this is acceptable to the designer of the
microdevice, all is well If it is not, either the method of
remote readout or the level of precision of the analytical
instruments employed must be changed to increase the
speed of readout to the degree required Once this has
been done, microdevices based on surface tension and
wet-tability that have remote readout capabilities can
func-tion either as static or dynamic analytical detectors or
sensors
Until now, it has also been assumed that the shaped
or tapered microtubes or microchannels within which
droplets move or flow under the influence of surface tension
and wettability and some external forcing agent such as
pressure or acceleration, had circular cross sections This
does not have to be so Figure 30a, b illustrates flow of
an elongated nonwetting mercury droplet constrained on
result of setting r equal to zero in the relationship given
in Eq (2), to fill in all corners completely This can never
be true for two reasons First, there is no such entity asinfinitely high pressure Second, in Fig 30a,b, bypass flow
of externally applied pressure Pent will occur through allopen corners, thereby reducing the actual pressure applied
to the nonwetting droplet An analogous situation occurswhen a child shoots an irregularly shaped pea through acircular straw Even though gaps equivalent to the opencorners in Fig 30a,b exist around the pea, the child canstill expel the pea from the straw simply by blowing hardenough, thereby producing a sufficiently effective pres-sure on the pea to accomplish the purpose This is exactlythe situation that exists for flow of nonwetting droplets
in noncircular microtubes or microchannels Therefore, allprevious arguments for remote sensing of droplet inter-faces in circular cross-sectioned microtubes or microchan-nels apply equally well to remote sensing of droplet inter-faces in microtubes, microchannels, or voids that have anytype of noncircular profile Moreover, noncircular micro-tubes or microchannels can certainly be used in conjunc-tion with circular microtubes or microchannels in the samemicrodevice In fact, there is very good reason to do so.Noncircular microtubes or microchannels can be fabricatedrelatively easily by using techniques such as photolithog-raphy and LIGA on a surface This is currently done onsilicon wafers by a sequence of deposition and/or etchingtechniques in a number of different ways, two of which will
be given A noncircular channel can be formed, for ple, by etching the channel in the surface and then coveringthe channel by sealing a glass plate over it Alternatively,for example, the noncircular channel can be formed byetching a channel in the surface and then filling it with
Trang 22exam-a sexam-acrificiexam-al mexam-ateriexam-al Another mexam-ateriexam-al is deposited over
the filled channel and then the sacrificial material is
re-moved to leave a microchannel However, no matter how
the noncircular channel is formed in the surface, a bypass
flow of gases occurs through its open corners, as illustrated
in Fig 30b, if the liquid has a high contact angle As just
mentioned, this makes it difficult to apply pressure
accu-rately and reproducibly to some nonwetting droplets
con-tained within such microtubes or microchannels This is
not true for circular microtubes or microchannels Thus,
the presence of small circular microtubes or
microchan-nels at appropriate positions in any device fabricated from
noncircular microtubes or microchannels will allow either
gases or wetting fluids to apply hydrostatic pressure to
microdevices that contain nonwetting droplets at 100%
ef-ficiency The reverse is also true A circular cross section in
the device will also allow the nonwetting droplet to apply
force to the gas or wetting fluid at 100% efficiency This also
means that it is possible to have wetting and nonwetting
fluids in the same microdevice Finally, regardless of the
cross-sectional shape of the microtubes, microchannels, or
voids, all wetting fluids will have 100% efficiency
It is extremely important to realize that the previous
discussion also illustrates that an elongated non-wetting
droplet confined within a microtube or microchannel that
has, for the sake of illustration, square walls can serve
purposes other than remote sensing For example, it can be
used to act as a shutter in optical applications, where the
presence or absence of the droplet controls whether or not
light or other electromagnetic radiation is allowed to pass
through the square microchannel walls In this instance,
the nonwetting droplets function in much the same fashion
as a window blind by controlling whether or not light is let
through a window depending on whether or not the blind is
up or down It could also control particle beams in a similar
manner
Figure 31 illustrates a much more familiar looking
shutter mechanism that could very easily function
iden-tically to traditional mechanical shutters An end bulb is
connected to a fill tube, and both are filled by a nonwetting
liquid, which is called the working fluid and is opaque
for the particular application A rectangular void is also
connected to the fill tube, but its thickness is less than
the diameter of the fill tube (The thickness of the void
in this figure is exaggerated for clarity.) The shutter that
has constant void thickness is illustrated in the open
configuration in Fig 31a,b, where the incident radiation
or particle beam passes through the shutter, and is closed
in Fig 31c,d where the incident beam or radiation is
blocked by the shutter The void width and void length
can be much greater than the void thickness and only
one void dimension has to be macroscopic to carry out a
shutter’s function This illustrates an extremely important
point: although all of the dimensions of a device can be
microscopic, only one dimension of a device must be in
the range where surface tension and wettability become
dominant factors in the device’s reaction to internal or
external stimuli to consider the device a microdevice The
rectangular shutter of constant void thickness illustrated
in Figure 31 must be considered a microdevice because of
the microscopic dimensions of its thickness, even though
Light
LightEnd bulb
WorkingfluidFill tube
ThicknessVoid
Fluid interface
(a)
Non-wetting interfaceHeater
Length
WidthEnd bulb
Fill tube
Void
Workingfluid
(b)
Light
Working fluidinterface
(c)
Working fluidinterface
(d)
Figure 31 Macroscopic microtube rectangular shutter: (a) side
view of open shutter; (b) top view of open shutter; (c) side view of closed shutter; (d) top view of closed shutter.
it can have very macroscopic dimensions for one or more
of its other features This is true for all microdevices based
on surface tension and wettability In this example, themethod of actuation of the rectangular shutter shown inFig 31 is derived by an externally generated electrical cur-rent input through a heater contained within the workingfluid As the working fluid expands due to this heat input,
Trang 23any gas bubbles or other gas-filled voids contained within
the working fluid become compressed, thereby raising
the internal pressure Pint within the working fluid This
thrusts the internal interface farther and farther into the
rectangular void At some point, the radius will decrease
sufficiently, so that the internal interface will shoot across
the void length, providing that the volume of compressed
gas-filled voids is much larger than the volume of the
shutter This will close the shutter in the same fashion
as the “Yes-No” devices described earlier in Fig 25 If gas
bubbles or other gas-filled voids are not present in the
working fluid, then the heat input will not cause a “Yes-No”
type of reaction, but rather will enable the shutter to close
more gradually In this way, a partially closed shutter can
be maintained by controlling the heat input appropriately
The gradual closing capability can be obtained for a
pressure-activated shutter by employing a void thickness
that has a decreasing taper A different tapered end is
shown at the end of the rectangular void where the
work-ing fluid stops in the closed position This is to minimize
the water hammer effect and does not have to be present
for the shutter to work Of course, external pressure or
some other external stimuli as well as an internal stimuli
could also be used in place of the heater, and the shutter
would still function If pressure were employed, it would
then be a pressure sensor that has some macroscopic
dimensions that would be very easy to observe Filling of
the void would signify that a certain pressure had been
reached Obviously, there are numerous other applications
of this technology but only two others will be mentioned
One involves using a reflective nonwetting fluid, so that a
mirror results when the void space is filled, and the second
application encompasses a much larger microscopic void
area If the void space is made as large in area as a window
pane, solar energy acting on the reservoir could be used
to force liquid into a void of microscopic dimensions in the
window pane and thus block sunlight from going through
the window if an opaque nonwetting liquid is employed
The void shown in Fig 31 has constant thickness and
is rectangular Neither parameter is necessary Figure 32
illustrates a circular shutter, which has a straight top face
and a curved bottom face that make up the void Now,
de-pending on the amount of expansion of the working fluid
caused by electric power supplied to the heater, the shutter
can be completely open when all of the working fluid is
con-tained within the outside bulb, completely shut and have
no circular gap in the center at all, or anywhere in between,
as Fig 32 illustrates Certainly, both the top face and
bot-tom face can be curved or straight, and virtually any
shut-ter geometry can be employed Likewise, actuating
tech-niques other than heat input to the working fluid by an
internal heater can be used External heat input by
radia-tion or conducradia-tion or changes in the internal pressure of the
working fluid by any other means can be used to achieve
the same resulting shutter behavior
As mentioned earlier, surface tension and wettability
govern the position of droplets within microdevices Thus,
in addition to the external stimuli already mentioned, any
external stimuli that changes either the surface tension
of the droplet or the wettability of the surface can be
de-tected by a suitably designed microdevice sensor Some, but
(a)
Heater
Outsidebulb
Workingfluidinterface
CirculargapLight
Topface
BottomfaceLight
Outside bulb
Heater
Interface
Workingfluid
(b)
Figure 32 Circular shutter or iris: (a) side view; (b) top view.
not all, such stimuli include the following: temperature,magnetic field, electrical field, rotation, radiation, andbeams of particles
Until now, all of the microdevice sensors illustrated havebeen designed to respond to external stimuli This is not theonly mechanism for displacing microdevice droplets Anycompositional change that occurs within droplets them-selves or on the walls of microdevices can also change sur-face tension or wettability These changes can be eitherreversible or irreversible and can be caused by a gas or wet-ting fluid in the device along with the nonwetting droplet
In addition, the surface tension of the droplet increases
as both the temperature and rotation increase If these
or any other internally induced change in a microdevice’ssurface tension or wettability occurs, it can be detected andmonitored remotely using any of the techniques describedpreviously Obviously, these internally induced changes in
a microdevice’s surface tension or wettability can also beused to move the droplet(s) to perform work
In addition, no actual dimensions of either microtubes
or microchannels have yet been discussed Assuming anonwetting fluid such as mercury, which has a surface ten-sion at room temperature of approximately 470 dynes/cmand a contact angle on glass microdevice walls of roughly
140◦, one can calculate the following droplet radii for theindicated internal pressures using Laplace’s equation mod-ified to include the effect of wettability (Table 1):
In this section, we have shown that the flow of dropletswithin microtubes and microchannels that is controlled bysurface tension and wettability can be used to sense, quali-tatively and quantitatively, any environmental factor thatacts on a droplet or affects either its surface tension or
Trang 24Table 1 Calculation of Droplet Radii
rdev(µm) Pint(lb/in 2 ) Pint(kPa)
wettability, or both This sensing can be performed
remotely by a variety of techniques It has also been
demon-strated that wetting droplets can sometimes be used with
only minor device modifications Static as well as dynamic
remote sensing can be performed, and microtubes or
mi-crochannels that have circular or noncircular cross
sec-tions and variable axial geometries can be employed either
individually or together The reaction of these devices to
any stimuli can be tailored by the device geometry and the
method of sensing and result in linear and nonlinear
ana-log output, as well as digital output The use of this
tech-nique in nonsensing applications that perform
mechani-cal functions was also demonstrated Finally, whenever at
least one microdevice dimension lies between 1000µm and
0.003µm, it has been shown that it can perform all of the
various tasks that have been discussed
Nonwetting Droplets That Perform Work
The preceding discussion has indicated that the movement
of nonwetting droplets within microtubes or
microchan-nels can be used for sensing and controlling fluids In
ad-dition, microscopic nonwetting droplets can be used for
mechanical control and manipulation within microdevices,
including tasks such as position control, moving objects,
deforming objects, pumping fluids, circulating fluids, and
controlling their flow Obviously, complex machines and
en-gines can be produced by properly joining actuator and
pumping elements
An application of a microscopic nonwetting droplet for
low friction position control is a microtube liquid bearing
shown in Fig 33 Referring to Fig 33a, for example, the
bearing assembly is a microtube that has one or more
cir-cular channels on its circumference that actually join the
microtube’s interior void space in a narrow ring-shaped
opening A center rod only slightly smaller in diameter
than the bearing assembly is supported by nonwetting fluid
that fills the circular channels This fluid cannot leak out
around the center rod if the gap is small enough because
too much pressure (Table 1) is required to form the droplet
of smaller radius that would be able to leak Therefore, the
center rod is free to either rotate or translate axially within
the bearing assembly It is called an external bearing
be-cause of this outside configuration The only restraining
forces involved are frictional forces between the center rod
and the non-wetting fluid
Figure 33b illustrates a reciprocal situation called a
microtube internal bearing A straight walled microtube
is used A central rod has at least one groove about the
circumference, and the nonwetting fluid fills this groove,
which allows both rotational and translational motion
(a)
Center rodCircular channel
Non-wettingfluid
Circular channel
Non-wettingfluid
Figure 33 Shaft supported by nonwetting fluidic bearings:
(a) rotating and translating outer bearing; (b) rotating and lating inner bearing; (c) inner/outer bearing that rotates but does not translate.
trans-Figure 33c is a mixed combination of internal and externalmicrotube liquid-bearing locations In this configuration,however, only rotational motion is easily achieved Fortranslation to occur, shearing of a wetting droplet musttake place Although this is not as difficult as forming asmall-radius annular droplet, it still involves generatingnew droplet surface area and therefore requires more force
to produce translation than for either the purely internal
or purely external bearings
Figure 34 illustrates a microtube liquid bearing thatwill not allow significant translational motion It is a
Non-wettingbearings
Figure 34 Thrust bearing that incorporates four nonwetting
fluid bearings.
Trang 25PistonGap
Non-wetting
droplet
Figure 35 Piston actuated by nonwetting fluid.
thrust bearing that uses four separate microtube liquid
bearings in an external configuration As before, an
in-ternal or mixed configuration is also possible, and
ad-ditional microtube liquid bearings that use surface
ten-sion/wettability effects can be employed
Figure 35 illustrates how a nonwetting droplet can be
employed as an actuator; its motion down a microchannel
or microtube is used to move a loosely fitting piston It is
important to note that as long as the droplet does not wet
the piston and walls, none of the droplet will squeeze by
the piston if the clearance is less than the radius of the
nonwetting droplet The nonwetting droplet is deliberately
shown only in part in Fig 35 to demonstrate that the
actual mechanism, external stimuli, or internal stimuli,
or wetting fluid, that causes it to push on the piston
is unimportant Its ability to function to transmit force
mechanically and thereby perform work is all that matters
CONCLUSIONS
Microtubes appear to have almost universal application in
areas as diverse as optics, electronics, medical technology,
and microelectromechanical devices Specific applications
for microtubes are as wide ranging as chromatography,
en-capsulation, cross- and counterflow heat exchange,
injec-tors, micropipettes, dies, composite reinforcement,
detec-tors, micropore filters, hollow insulation, displays, sensors,
optical waveguides, flow control, pinpoint lubrication,
crosponges, heat pipes, microprobes, and plumbing for
mi-cromotors and refrigerators The technology works equally
well for high- and low-temperature materials and appears
feasible for all applications that have been conceived to
date
The advantage of microtube technology is that tubes
can be fabricated inexpensively from practically any
mate-rial in a variety of cross-sectional and axial shapes in very
precise diameters, compositions, and wall thicknesses of
orders of magnitude smaller than is now possible In
con-trast to the other micro- and nanotube technologies
cur-rently being developed, microtubes can be made from a
greater range of materials in a greater range of lengths and
diameters and far greater control over the cross-sectional
shape These tubes will provide the opportunity to
minia-turize (even to nanoscale dimensions) numerous products
and devices that currently exist, as well as allowing the
fabrication of innovative new products that have to date
been impossible to produce
Space only allowed presenting one application of tube technology to new innovative products in greater de-tail This one application comprised those devices that arebased on surface tension and wettability The few devicesthat were shown as examples demonstrated the breadth ofthis technology only in one field The application of micro-tube technology to other fields is considered equally richand limited only by a designer’s imagination
micro-ACKNOWLEDGMENTS
The invaluable help provided by Hong Phan in tion, Marietta Fernandez in microscopy, and Tom Duffey
fabrica-in artwork is greatly appreciated Ffabrica-inancial support from
Dr Alex Pechenik of the Chemistry and Materials ScienceDirectorate, Air Force Office of Scientific Research was re-sponsible for much of this work
BIBLIOGRAPHY
1 G Stix, Sci Am 267: 106–117 (1992).
2 W Menz, W Bacher, M Harmening, and A Michel, 1990 Int Workshop on Micro Electro Mechanical Syst (MEMS 90), 1990,
6 S Weiss, Photonics Spectra 108 (Oct 1995).
7 M Mullenborn, H Dirac, and J.W Peterson, Appl Phys Lett.
10 B.D Terris, H.J Mamin, M.E Best, J.A Logan, D Rugar, and
S.A Rishton, Appl Phys Lett 69: 4262–4264 (1996).
11 Y Xia and G.M Whiteside, Angew Chem Int Ed 37: 550–575
(1998).
12 M Ishibashi, S Heike, H Kajayama, Y Wada, and
T Hashizume, J Surf Anal 4: 324–327 (1998).
13 H.J Mamin and D Rugar, Appl Phys Lett 61: 1003–1005
16 J.C Harley, M.S Thesis, University of Pennsylvania, 1991.
17 D.J Harrison, K Fluri, K Seiler, Z Fan, C.S Effenhauser, and
A Manz, Science 261: 895–897 (1993).
18 K.D Wise and K Najafi, Science 254:1335–1342 (1991).
19 P.C Hidber, W Helbig, E Kim, and G.M Whitesides,
Trang 2622 R.J Jackman, S.T Brittain, A Adams, H Wu, M.G Prentiss,
S Whitesides, and G.M Whitesides, Langmuir 15: 826–836
34 P Hoyer, Adv Mater 8: 857–859 (1996).
35 P Hoyer, N Baba, and H Masuda, Appl Phys Lett 66: 2700–
40 T.W Ebbesen and P.M Ajayan, Nature 358: 220–223 (1992).
41 S Iijima, T Ichihashi, and Y Ando, Nature 356: 776–779
(1992).
42 T.W Ebbesen, Phys Today 49(6): 26–32 (1996).
43 N.C Chopra, R.J Luyken, K Cherrey, V.H Crespi, M.L.
Cohen, S.G Louie, and A Zettl, Science 269: 966–967
(1995).
44 Z Weng-Sieh, K Cherrey, N.G Chopra, X Blase, Y Miyamoto,
A Rubio, M.L Cohen, S.G Louie, A Zettl, and R Gronsky,
47 R Bacon, J Appl Phys 31: 283–290 (1959).
48 R.T.K Baker, M.A Barber, P.S Harris, F.S Feates, and R.J.
52 R Kirsch, M Mertig, W Pompe, R Wahl, G Sadowski,
K.J Boehm, and E Unger, Thin Solid Films 305: 248–253
(1997).
53 M.R Ghadiri, J.R Granja, R.A Milligan, D.E McRee, and
N Khazanovich, Nature 366: 324–327 (1993).
54 A Harada, J Li, and M Kamachi, Nature 364: 516–518 (1993).
55 J.-H Fuhrhop, D Spiroski, and C Boettcher, J Am Chem Soc.
115: 1600–1601 (1993).
56 G Li and L.B McGown, Science 264: 249–251 (1994).
57 H Dai, J.H Hafner, A.G Rinzler, D.T Colbert, and R.E.
Smalley, Nature 384: 147–150 (1996).
58 L Langer, L Stockman, J.P Heremans, V Bayot, C.H Olk,
C Van Haesendonck, Y Bruynseraede, and J-P Issi, J Mater.
64 H.H Mueggenburg, J.W Hidahl, E.L Kessler, and D.C.
Rousar, AIAA/SAE/ASME/ASEE Joint Propulsion Conf Exhibit Proc Nashville, TN 1992, #AIAA 92-3127.
65 M Datta, J Electrochem Soc 142: 3801–3805 (1995).
66 M Datta, Electrochem Soc Interface, 32–35 (Summer 1995).
67 Microcomponent Chemical Process Sheet Architecture, U.S Pat 5,811,062, Sept 22,1998, R.S Weneng, M.K Drost, C.J Call, J.G Birmingham, C.E McDonald, D.K Kurath, and
M Friedrich.
68 C.Gonzalez, S.D Collins, Transducers 97, Chicago IL, 1997,
pp 273–274.
69 T Young, Trans R Soc London 95: 65 (1805).
70 P S de Laplace, Mechanique Celeste Supplement to Book 10,
1806.
71 E.W Washburn, Proc Nat Acad Sci USA 7: 115 (1921).
72 H.L Ritter and L.C Drake, Ind Eng Chem Anal Ed 17: 782
INTRODUCTION
Molecular imprinting is a process for synthesizing als that contain highly specific recognition sites for smallmolecules (scheme 1) The imprinting process consists
materi-of a template molecule (T ) that preorganizes functional, polymerizable monomers (M ) This template assembly (T A) is copolymerized with a large excess of cross-linking
monomers to produce an insoluble network material.Extraction of the template molecule leaves behind a regionthat is complementary in size, shape, and functional grouporientation to the template molecule The resulting poly-meric materials are referred to as molecularly imprintedpolymers (MIPs):
Trang 27Extract
Rebind
Cross-linkingmonomerM
Scheme 1
Intermolecular forces between the template, functional
monomers and polymer matrix determines MIP
struc-ture Pre-organization of the functional monomers has
been achieved with covalent (1–4), electrostatic (5,6), and
hydrogen bonding interactions (7–9) The most common
method of polymer formation employs free radical
initia-tion These materials are tough, easy to handle, and have
been shown to retain their recognition properties for over
five years without loss of selectivity or capacity (10,11)
As a result, they have advantages over more fragile
bio-logical systems for molecular recognition and have been
referred to as “plastic antibodies” (12) Under the proper
conditions, MIPs have shown the ability to function as
se-lective binding materials (13–15), microreactors (3,16,17),
facilitated transport membranes (18,19), and catalysts
(20–23) Their utility, coupled with their longevity and
robustness, make MIPs candidates for applications in
separation science, as industrial catalysis, and for
medi-cal diagnostics (24,25)
Molecular imprinting began as an attempt to develop
“synthetic antibodies.” Initially referred to as “specific
adsorption,” a technique developed by Dickey utilized
silica polymers to create synthetic receptors for the dye
molecules methyl orange and ethyl orange (26,27) Dickey
found that a material prepared using methyl orange was
able to re-adsorb that molecule 1.4 times better than ethyl
orange, and that the selectivity could be reversed by
us-ing ethyl orange durus-ing the polymerization These
re-sults were followed by the first imprinted chiral
station-ary phase (CSP), which was prepared in 1952 by Curti
and Columbo Dickey’s method was used to imprint silicate
polymers withD-camphorsulfonic acid and enantiomers of
mandelic acid (28) Curti and Columbo were able to achieve
chromatographic separation of camphorsulfonic acid
enantiomers and mandelic acid enantiomers using this
process
The first report of an imprinted organic polymer was
made by Wulff in 1972 (29) A D-glyceric acid template
was covalently bonded to amino styrene and 2,3-O-
p-vinylphenylboronic ester and was then incorporated into
a divinylbenzene polymer (29,30) This important
develop-ment increased the specificity of interaction between
tem-plate and functional groups within the polymer and
al-lowed for more controlled design of functionality within
the imprinted polymer Another major contribution to
the field of molecular imprinting was made by Mosbach
and co-workers in 1984 when they introduced a new
method of forming the pre-polymerization complex based
on noncovalent interactions (31) Instead of covalently printing target molecules, a template assembly was formedutilizing electrostatic and hydrogen bonding interactions
im-to specifically position functional monomers Since thistime, there has been a tremendous amount of work in thearea of molecular imprinting, much of which has been re-viewed in the literature (11,32–34,35–38)
POLYMER CHEMISTRY Organic Materials
Molecularly imprinted polymers are almost exclusivelycomposed of highly cross-linked organic polymers Thesepolymers may be prepared using a variety of monomersand polymerization methods By far, the most commonmethod has been the free radical polymerization of acry-late functionalized monomers A detailed description of thechemistry of free radical polymerization is available in theliterature (39)
Typical formulations consist of equal parts of monomer/template assembly and an inert deluent, referred to as aporogen The choice of porogen is important since it mustsolubilize the monomers and, in the event of noncovalentimprinting, must not suppress interactions between theimprinting molecule and the functional monomers Theporogen also serves as an important determinant in the for-mation of the final polymer morphology
The resulting network polymers are macroporous in ture Their macroporosity arises from the phase separation
na-of solid polymer from solvent and unreacted monomer Thetiming of the phase separation is dependent on the extent
of polymerization and the polymer solubility in the gen The phase separated polymer particles aggregate andcreate a high internal surface area (40,41) This is par-ticularly important since the highly cross-linked networkspresent a barrier to facile transport of solvent and reagents
poro-to the functional sites The high internal surface area (200–
400 m2/g) exposes a large fraction of polymer mass to theexternal solvent and reagents
The high percentage of cross-linking monomer (80 to
>90 mol %) that is usually employed results in a
poly-mer with a rigid three-dimensional structure In addition
to providing a rigid structure, the cross-linking provides
a scaffold for positioning functional groups within voids
Trang 28Table 1 Representative Examples of Templates, Monomers, and Their Associated Complexes Used
for Covalent Imprinting
O
OH
OH B
O
O
O O
N
H N
N H
Cu OH
OH2
N
H N
N H
Cu O O
enabling recognition of the imprinted molecules The
inter-action between the template molecule and the functional
monomers can be covalent or noncovalent in nature The
former can consist of any number of interactions
includ-ing electrostatic, hydrogen bondinclud-ing, hydrophobic, andπ–π
interactions
Covalent Imprinting
The covalent method of imprinting requires the
synthe-sis of a polymerizable derivative of the template molecule
Covalent bonds are chosen so as to permit their cleavage
following the polymerization After the cleavage step, the
monomers should present functionality that can interact
with the imprint molecule either by reforming the covalent
linkage or by noncovalent interactions Some common valent bonds used for this purpose are carboxylic (42) andboronic esters (43–45), ketals (13,46), and Schiff bases (2).Also included in this area are metal complexes that bindand orient the template (47–49) Representative examples
co-of covalent template assemblies are given in Table 1.The covalent method of pre-organization has several ad-vantages over noncovalent methods Perhaps the most im-portant is that there is no requirement for excess functionalmonomer This minimizes the amount of nonspecific bond-ing that can occur between the analyte and the polymermatrix A nonspecific interaction in this case is any bondingthat occurs other than between the analyte and the definedbinding site In addition, covalent imprinting may pro-duce more “well-defined” binding sites, thus reducing the
Trang 29Table 2 Representative Examples of Templates, Monomers, and Their Associated Complexes Used
for Noncovalent Imprinting
(51)
Hydrogen bond
N
N N N
NH2
O OH
N
N N N N
O O H
O O
O N
H2N
O OH
O N H
N H
H H
O H O
distribution of binding sites with varying affinities for the
template molecule Potential disadvantages of this
tech-nique include the additional steps required to synthesize
the template-monomer complex and for the chemical
cleav-age of the template from the polymer If rebinding is to
oc-cur by reformation of covalent bonds, the rate of rebinding
may be slow and not suitable for certain applications such
as separation media for chromatography
Noncovalent Imprinting
Non-covalent imprinting employs weaker interactions
be-tween the template, functional monomers and
cross-linking monomers to position the functional groups and
de-fine the binding cavity (31,50) The interactions that have
been most commonly utilized are hydrogen bonds,
electro-static,π–π, and hydrophobic interactions Some
represen-tative examples of noncovalent template assemblies are
given in Table 2
The noncovalent method has the benefit of being
rela-tively simple to carry out It requires fewer steps than
the covalent approach and is more compatible with
au-tomated or combinatorial methods In addition, the types
and number of potential templates is greater than what
can be achieved using the covalent approach This method
is limited, however, to templates that can interact
some-what strongly with functional monomers In addition, an
excess of functional monomers is usually required in the
polymerization mixture to ensure optimal results In part
because of this requirement, polymers prepared in this
manner contain large numbers of functional groups, which
can contribute to nonspecific bonding and in general mayprovide lower affinity sites than can be achieved with co-valently imprinted polymers
Polymer Networks
In most imprinting systems the bulk of the polymer is made
up of difunctional cross-linking monomers A wide variety
of these have been developed, with the majority being late or styrene based (Table 3) Cross-linking monomersplay an important role in determining the bulk proper-ties of the materials In addition, compatibility betweenthe cross-linking monomers and all other monomers in theformulation must be established
acry-Preparation and Processing The initial step in the
im-printing protocol is formation of a prepolymerization plex between the molecule to be imprinted and the func-tional monomers which will provide recognition (Fig 1).This complex can be formed through either covalent or non-covalent interactions In both cases, the formation of thecomplex must be reversible to allow extraction/rebinding
com-of the template molecule This complex is combined withthe cross-linking monomer(s), initiator, and the porogen.Oxygen, which can interfere with free-radical polymeriza-tions, is removed from the formulation either by freeze–pump–thaw methods or by displacing it from the solution
by sparging with an inert gas
Free-radical polymerization of the monomer/porogenmixture is initiated either photochemically or thermally
Trang 30Table 3 Representative Polymer Networks Used in Molecular Imprinting
Styrene/divinylbenzene Nonpolar/
Ethyleneglycol Dimethacrylate (EGDMA)
Polar aprotic trifunctionalized
Cross-linkingmonomer(s)Initiatorporogen
30 - 200 micron < 30 micron
UV Lightorheat
Figure 1 Schematic representation of the molecular imprinting process.
Trang 311 2 3
H OH
NH
N SN
N
O
H OH
NH
Due to the large number of initiators available, the range
of conditions in terms of radiation intensity, temperature,
and time that is required is large The choice of conditions
and method of initiation is dictated by the sensitivity of the
template to temperature and/or photochemical radiation
The polymer that is produced is often a monolithic
solid For noncovalent imprinting, releasing the template
molecule from the polymer is often achieved by extracting
the polymer using a solvent such as methanol In some
cases, it is beneficial to use a small quantity of acid to aid
in the removal of the template For covalent imprinting,
the conditions for removal of the template are dictated by
the requirements for chemical cleavage of the template–
monomer bond
The next step in the processing sequence involves
crush-ing the polymer followed by sizcrush-ing of the resultcrush-ing particles
Many of the chromatographic applications of imprinted
polymers require the polymer to be ground into small
par-ticles (25–150µm) and sized to a uniform range.
APPLICATIONS
To date, the applications of MIPs have fallen into four
areas: (1) separation technology, (2) catalysis (enzyme
mimics), (3) recognition elements for sensors, and (4)
anti-body and receptor site mimics
Separation Technology
The use of MIPs as stationary phases for chromatography
is the most studied application of this technology The list of
molecules that have been examined includes biomolecules
such as amino acids, small peptides, proteins,
carbo-hydrates, nucleotides, nucleotide bases, steroids, and a
number of small drug molecules The majority of
separa-tion studies have focused on the preparasepara-tion of stasepara-tion-
station-ary phases for high performance liquid chromatography
(HPLC) (59) MIPs have also been used in thin layer
chro-matography (TLC) (60), capillary electrophosresis (61,62),
membranes (18,19) and solid-phase extraction (SPE)
(63–65)
Mosbach and coworkers demonstrated direct
enan-tioseparation of β-blocker (-)-S-timolol (template 1;
scheme 2) using a chiral stationary phase (CSP) prepared
by molecular imprinting (66)
Macroporous copolymers of EGDMA and MAA,
pre-pared with template 1 as the imprinting molecule, were
employed as a chiral stationary phase Following
extrac-tion of the template with acetic acid, the MIP staextrac-tion-
station-ary phase was used for separation of racemic (template 1)
by HPLC The imprinted polymer stationary phases
gave baseline separation of racemic timolol with a tion factor (α) of 2.9 The separation factor (α) is a measure
separa-of the separation separa-of two peaks in a chromatogram and isgiven byα = k2/k1, where k is equal to the capacity factor
of each peak The capacity factor is a measure of binding
strength and is defined as k = (t − t0)/t0, where t is the tention time of the solute, and t0is the void time The struc-turally related racemic derivatives atenolol (template 2)
re-and propanolol (template 3) were not separated into their
enantiomers, nor were they retained on the column to anysubstantial degree
Shea et al explored the ability of MIPs to separate tiomers of a series of benzodiazepine molecules (9) Enan-tiomerically pure benzodiazepine derivatives (molecules4–8, Scheme 3) were used to imprint EGDMA polymersusing MAA as the functional monomer The resulting ma-terials were ground into particles and analyzed by HPLC
enan-Cl
R' N
Scheme 3
Regardless of the benzodiazepine used for the ing, enantioselectivity could be achieved for the entirefamily of benzodiazepines In each case, the imprintedenantiomer was retained over the nonimprinted isomer(α > 1; Table 4) Additionally, the greatest separation foreach polymer corresponded to the benzodiazepine whichhad been imprinted (bold entries) The ability of ben-zodiazepine MIPs to produce separation for a family ofmolecules in favor of the absolute configuration of the tem-plate is an important result of this study
imprint-Synthetic polymer complements to biologically tant nucleosides have been prepared using 9-ethyladenine(9-EA; polymer 9, scheme 4) as the imprinting molecule(15) The molecular basis for the imprinting arises fromthe fact that adenine and its derivatives are known toform complexes with carboxylic acids 9-EA was employed
impor-as the template; methacrylic acid (MAA; polymer 10) wimpor-asemployed as the functional monomer (scheme 4) The for-mulation included EGDMA as the cross-linking monomer
in chloroform solution Polymerization was initiated tochemically at 12◦C in the presence of AIBN
Trang 32pho-Table 4 Benzodiazepine α Values for Retention on MIP Stationary Phases
Benzodiazepine Derivative Injected MIP
Template
4-Hydroxy benzyl (7)
Benzyl (6)
Indolyl (8)
Isopropyl (5)
Methyl (4)
N
NN
H
OOHO
O
H
OOH
EGDMA AIBN hυ
N
NN
+
OHO
Ka 78,000 M−1 CHCl3
9-EA (CHCl3)
OO
N
NNHH
H
OOH
Scheme 4
When used as the stationary phase for HPLC, the 9-EA
imprinted column packing was found to have a strong
affi-nity for adenine and its derivatives The remaining purine
and pyrimidine bases elute close to the void volume
NH 2
N
N H O
HN
N H O
H2N
Scheme 5
An illustration of the selectivity is given in Fig 2
An HPLC trace of a sample mixture of adenine (A;template 11), cytosine (C; 12), guanine (G; 13), andthymine (T; 14) reveals C, G, and T elute close tothe void volume (∼2 min at 1 ml/min flow) while ade-nine elutes in 23 minutes In contrast, polymers made
by replacing 9-EA with benzylamine show no ity for the nucleoside bases: all elute between 1.6 and2.0 minutes
affin-Most studies with imprinted polymer have been carriedout using polymer particles either in the batch or chro-matographic mode There have been several reports how-ever of the preparation of imprinted polymers fabricated
Trang 33CGAT
C, G, T
Figure 2 HPLC race of a 4.6 mm× 10 cm column packed with
polymer templated with 9-ethyladenine (9-EA) A mixture of
ade-nine (A), thymine (T), guaade-nine (G), and cytosine (C) was injected.
Adenine elutes at 23 minutes: all other nucleosides elute close to
the void volume (2 min) The mobile phase is MeCN:AcOH:H2O
in the ratio 92.5:5:2.5.
into membranes for the selective transport and
separa-tion of molecules One example involves the preparasepara-tion
of MIPs with 9-ethyladenine (template 9) as the template
and methacrylic acid (10) as the polymerizable functional
monomer (scheme 4) (18)
The free-standing films of imprinted polymers were
ob-tained by the polymerization of films of a DMF solution of
ethylene glycol dimethacrylate and methacrylic acid with
9-ethyladenine (template 9) as the template
Transport studies were carried out with an H-shaped
two-compartment cell The concentration of substrate in
the receiving phase as a function of time was quantified
by HPLC Under steady-state conditions, a linear
correla-tion between amount of substrate transported and time
was observed The transport rates of adenine, thymine,
Figure 3 Plot of the facilitated transport
of adenosine from an equimolar methanol/
chloroform (6:94 v/v) solution (adenosine =
guanosine= 26 µM) The membrane was
im-printed with 9-ethyladenine The selectivity
O
NH2H
0.02.55.07.510
at a comparable rate to those of cytosine and thymine Areference polymer membrane of the same monomer com-position prepared without template 9, on the other hand,transports adenine at rates comparable to those of thymineand cytosine
Sensors
MIPs have been used as the recognition element in cal sensing devices Biosensors utilize a recognition ele-ment, such as an antibody or enzyme, in conjunction with
chemi-a trchemi-ansducer A chemicchemi-al signchemi-al results from the binding ofthe analyte to the receptor that is then transformed into anelectrical or optical signal that can be monitored In manycases, the development of recognition elements lags far be-hind the transduction methods available, such as optical,resistive, surface acoustic wave (SAW) or capacitance mea-surements There is considerable potential for new chem-ical sensing devices if suitable recognition elements wereavailable
There are several potential advantages that may begained by utilizing MIPs as the recognition element inplace of biological receptors Since MIPs are syntheticreceptors, they have a virtually unlimited pool of analy-ates In addition, MIPs are stable to harsh conditions thatmay not be compatible with biologically based sensors Fur-thermore, the ability to incorporate a signaling element,such as a fluorescent probe, in the vicinity of the binding
Trang 34Table 5 Summary of Sensor Applications Using Molecularly Imprinted Polymers
Fluorimetry Dansyl- L -phenylalanine MAA, 2VPy 0–30µg/mL (68)
Luminescence PMP Eu(III) + DVMB 0.125–150000µg/L (49,72)
site could be utilized in sensor applications It may also
be possible to prepare sensors with an array of polymers
imprinted for any number of analyte molecules for
fabri-cation of a single sensor capable of detecting many
sub-stances Some examples of sensor applications using MIPs
and their detection ranges are shown in Table 5
The use of luminescence spectroscopy combined with
fiber optics can provide systems for sensor applications
Utilizing molecularly imprinted polymers in combination
with these systems can add chemical selectivity to these
types of sensors This technology has been applied to the
detection of many types of compounds including nerve
agents (49), herbicides (72,73), drugs molecules (70), and
biomolecules (68,74)
Mosbach et al demonstrated the use of a MIP based
fiber-optic sensing device for a fluorescence-labeled amino
acid (dansyl-L-phenylalnine; template 17, scheme 6) (67)
These materials were prepared using MAA and 2-vinyl
pyridine as the functional monomers and EGDMA as the
cross-linking monomer A system was devised in which
polymer particles were held against the fiber tip using a
nylon net (Fig 4)
In both systems the polymers showed an increased
affi-nity for the imprinted enantiomer This was demonstrated
as an increase in the measured fluorescence in the
fiber-optic system In addition, the ability to follow increases in
concentration by monitoring the increases in fluorescence
was demonstrated Binding of the analyte to the polymer
OH O OH
O
N
OH O
N
17
EGDMA MAA 2-Vinyl pyridine
Dissociation Association
Scheme 6
O-ringsPolymer particlesNet
Window
Optic-fiber from light source
Fiber bundle to light detector
Figure 4 Fiber-optic sensing device configuration A layer
con-sisting of 2 mg of polymer particles was held below the quartz glass window The layer was kept in place by a nylon net (20µm
pores) and an O-ring An optical fiber was connected through the quartz window onto the polymer surface Part of the total light emitted from the polymer was collected and guided through the fiber bundle to the detector (visible light-sensitive photodiode).
could be followed in real time as the time required to reach
a steady-state response was 4 hours This rebinding timecorrelates with the equilibration time of this imprintedsystem
Trang 35Powel et al prepared imprinted polymers for the
detection of cAMP (polymer 18) (68) These polymers
are unique in that they contain a fluorescent dye,
trans-4-[ p-(N,N-dimethylamino)styryl]-N-vinylbenzylpyridinium
chloride (19, scheme 8), as an integral part of the binding
site The polymer serves as both the recognition
ele-ment and the transducer for the detection of cAMP The
polymer was prepared by polymerization of trimethylol
trimethacrylate (TRIM), 2-hydroxyethyl methacrylate
(HEMA; 20), and the fluorescent functional
mono-mer, template 19, in the presence of cAMP The
polymeriza-tion was initiated photochemically at room temperature
R
H2N
OH
OH
OO
The fluorescence of the imprinted polymer was shed when in the presence of aqueous cAMP The quench-ing was found to reach a maximal level between 10 and
dimini-100µM cAMP In addition, no quenching was observed in
a control polymer prepared without template Addition of
cGMP (21) had little effect on the fluorescence of both the
imprinted and control polymer
O O O
OH
N N
N N
NH2O
OH
OH O
O OH
O O O
OH
N N
N N
NH2O
OH
OH
O OH
O OH
O
O P O
O O O
OH
N N
N N
NH2
18
TRIM,MeOH AIBN