Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Apr 20.
Published in final edited form as: Acc Chem Res. 2010 Apr 20;43(4):518–528. doi: 10.1021/ar900178k

Co-Fabrication: A Strategy for Building Multi-Component Microsystems

Adam C Siegel †,§, Sindy K Y Tang †,§, Christian A Nijhuis †, Michinao Hashimoto †, Scott T Phillips †, Michael D Dickey †, George M Whitesides †,*
PMCID: PMC2857577  NIHMSID: NIHMS172183  PMID: 20088528

Conspectus

Whitesides Conspectus ar-2009-00178k.R2 edited and approved

In this Account, we describe a strategy for fabricating multicomponent microsystems in which the structures of essentially all of the components are formed in a single step of micromolding. This strategy—which we call “co-fabrication”—is an alternative to multilayer microfabrication, in which multiple layers of components are sequentially aligned (“registered”) and deposited on a substrate by photolithography.

Co-fabrication has several characteristics that make it a particularly useful approach for building multicomponent microsystems. It rapidly and inexpensively generates correctly aligned components (for example, wires, heaters, magnetic field generators, optical waveguides, and microfluidic channels)—and over very large surface areas. By avoiding registration, the technique does not impose the size limitations of common registration tools, such as steppers and contact aligners, on substrates. We have demonstrated multicomponent microsystems with surface areas exceeding 100 cm2, but in principle device size is only limited by the requirements of generating the original master.

Co-fabrication can also serve as a low-cost and minimal-equipment strategy for building microsystems. The technique is amenable to a variety of laboratory settings and uses fabrication tools that are less expensive than those used for multistep microfabrication. The process also requires only small amounts of solvent and photoresist—a costly chemical required for photolithography. In co-fabrication, photoresist is applied and developed only once to produce a master, which is then used to produce multiple copies of molds containing the microfluidic channels.

Co-fabrication represents a new processing paradigm in which the exterior (or shell) of the desired structures are produced before the interior (or core). This approach—generating the insulation or packaging structure first, and injecting materials that provide function in channels in liquid phase—makes it possible to design and build microsystems with component materials that cannot be easily manipulated conventionally (such as solid materials with low melting points, liquid metals, liquid crystals, fused salts, foams, emulsions, gases, polymers, biomaterials, and fragile organics). Moreover, materials can be altered, removed, or replaced after the manufacturing stage. For example, co-fabrication allows one to build devices in which a liquid flows through the device during use (or is replaced before use). Metal wires can be melted and re-set by heating (in principle, repairing a break). This method leads to certain kinds of structures—such as integrated metallic wires with large cross-sectional areas, or optical waveguides aligned in the same plane as microfluidic channels—that would be difficult or impossible to make with techniques such as sputter deposition or evaporation.

This Account outlines the strategy of co-fabrication, describing several co-fabricated microsystems that combine microfluidics with (i) electrical wires for microheaters, electromagnets, and organic electrodes; (ii) fluidic optical components, such as optical waveguides, lenses, and light sources; (iii) gels for biological cell cultures; and (iv) droplets for compartmentalized chemical reactions, such as protein crystallization.

I. Fabricating Microfluidic Channels

Co-fabrication starts with the design and fabrication of a single layer (or plane) of microfluidic channels. Microchannels in PDMS can be readily produced using soft lithography.[1] Using this approach, a master is first created by patterning photoresist using contact photolithography. To reduce costs for features having relatively large (>10 μm) dimensions, a transparency with the desired patterned printed with a high resolution printer can be used as the photomask. Subsequent casting of the PDMS pre-polymer against the master and curing yields the polymer replica with a network of channels. Bringing the polymer replica in conformal contact with a flat slab of PDMS (or other substrates including glass, oxidized silicon, or oxidized polystyrene) after oxidization of both surfaces irreversibly seals and completes the microchannels. The procedure for fabricating the channels takes less than 24 h from computer design to physical system.

For some applications, it is desirable to build the layer of microfluidic channels using a material other than PDMS.[2,3] For example, bulk and surface micromachining[4] rely on dry- or wet-etching cavities in silicon or glass substrates, followed by a bonding step with another surface. Micro-molding, micro-imprinting/embossing, and laser ablation are commonly used to produce microfluidic channels in plastics and other polymers.[5,6] Scanning beam lithography techniques, such as focused ion-beam, e-beam, and pulsed laser deposition can fabricate microchannels as small as 10 nm in width.[7] Patterning paper into hydrophobic and hydrophilic regions also provides a new approach to the fabrication of microfluidic channels.[8]

The microfluidic channels used in co-fabrication have some similarities with established injection molding strategies used for manufacturing polymeric, metallic, and ceramic parts.[9] These techniques utilize molds that are injected with a precursor (typically a powder or polymer melt) that can be processed further to form solid structures. Injection molding of miniature components (so-called “microcasting”) can form microscale parts (e.g. posts, gears) with features as small as 10 μm.[10] Co-fabrication also involves injecting materials into molds, but unlike injection molding, the microfluidic channels are integral parts of the final structure (whereas in injection molding the mold is typically separated after injection). For academic research environments, injection molding is also limited by the cost of the equipment and the high pressures required (e.g., 3–5 MPa).[10]

Ultimately, the choice of material used for microfluidic channels (e.g., the bas-relief material of a co-fabricated system) depends on the characteristics required of that material in the final application: temperature stability, elasticity, transparency, resistance to solvent, permeability, wettability, biocompatibility, scalability, index of refraction, etc. Table 1 summarizes materials often used for generating microfluidic channels.

Table 1.

Properties of materials used to form microfluidic channels in co-fabricated systems.

Material for Forming Channels Method for Forming Channels Advantages Disadvantages
PDMS SL fast prototyping, biocompat. well studied, transparent, flexible/unbreakable gas perm., melts at high temp, hydrophobic, absorbs small molec., swells with solvents
Polycarbonate μM, IM, HE, e-B, LL non-gas perm. struct. strong, transparent hydrophobic, susceptible to organic acids/bases, inflexible
Polymethylmethacrylate μM, IM, HE, e-B, LL non-gas perm. struct. strong transparent, cheap hydrophobic, susceptible to organic acids/bases, inflexible
Cycloolefin Copolymer μM, IM, HE, e-B non-gas perm., cheap, resists solvents and acids/bases hydrophobic
Perfluoropolyethers μM, IM, HE biocompat. resists solvents and acids/bases, liq. at R.T. gas perm., not commercially available, requires UV to cure
Polyurethanes SL non-gas perm., easy to make hydrophilic, cheap flexible/unbreakable more complex fab. than PDMS, difficult to bond layers or perform surface modifications
Silicon μM, e-B, LL well-studied, non gas-perm. resists solvents and most acids/bases breaks easily, not electrically insulating, expensive, opaque
Glass μM, e-B, LL cheap, resists solvents and most acids/bases, non gas-perm., transparent breaks easily, difficult to etch (req. HF) and seal to itself (req. high temp)
Thiolene SL fast prototyping, resists, cheap, solvents, transparent, low gas perm., biocompat. requires UV to cure
Open in a new tab

SL = Soft Lithography, μM = Bulk/Surface Micromachining, IM = Injection Molding, HE = Hot Embossing, e-B = E-Beam Lithography, LL = Laser Lithography

II. Filling Microfluidic Channels with Materials

Microfluidic channels are typically filled with liquids using either capillary action, or active injection by pump/vacuum. Capillarity depends on the relative free energies of the interfaces in a channel to “wick” aqueous fluids in microfluidic channels. It is also possible to fill channels by applying an external vacuum or pressure to force a material in the channel; this approach is especially useful for replacing fluids in previously-filled channels, or for applications in which the fluid must flow through the channel to provide function (e.g., optofluidics).

Microfluidic channels can be filled with solid materials by injecting the material in the channel in liquid form, and then allowing the material to solidify inside the channel. We have found it particularly useful to be able to inject molten metals and alloys (e.g., In, Sn, Ga, In/Sn, In/Ag, Ga/In) into microfluidic channels by heating the metals and channels using a hotplate and/or heat gun, drawing the metal in the channel using house vacuum, and allowing the metal to solidify by cooling; PDMS is able to withstand temperatures up to 400ºC for limited times.[11] It is also possible to inject liquid materials into microfluidic channels, and to solidify these liquids into solids by curing or chemical cross-linking. These materials include sol-gels,[12] thermally curable epoxies,[13] and photocurable pre-polymers.[14] Crystallization in microfluidic channels has also been demonstrated.[15] (Table 2)

Table 2.

Relationship between the function of a co-fabricated microsystem and the material injected in its microfluidic channels.

Function Class Function Phase Material in Microfluidic Channel
Fluidic Control and handling of liquid chemicals Liquid Aqueous-based solvents, buffers [16,17,22,23]
Control and handling of biological samples for diagnostics Liquid Blood serum, saliva, other bodily fluids [8,24]
Supply of nutrients to cell cultures Liquid, Gel Liquid-based media, Agar [19]
Gas Oxygen, CO2 or other bio-active gases
Supply of nutrients to multi-cell organisms Liquid Liquid-based media [20]
Application of pressure for valves/pumps Gas Air or other inert gases
Gel pH responsive hydrogels [38]
Colorimetric molecular sensing Gel Polydiacetylene-based hydrogel [39]
Compartmentalization of reactions Emulsion Aqueous detergents suspended in fluorinated carrier fluids [52]
Molecular crystallization Emulsion Aqueous detergents suspended in fluorinated carrier fluids [15,53]
Electronic Conduction of electrical current (wires, circuits, antennas) Solid Low-melting-temperature metals [11]
Generation of heat/control of temp. Solid Low-melting-temperature metals [11]
Generation of electrostatic fields (dielectrophoresis, cell lysis) Solid Low-melting-temperature metals
Generation of magnetic fields Solid Low-melting-temperature metals [27,28]
Molecular electronics/testing of SAMs Liquid Non-Newtonian fluidic metals (EGaIn) [26,35,36]
Electro-mechanical fluid actuation Sol Ferrofluids
Electrochemical Storage of electrical energy (batteries) Liquid Electrolytes or conductive liquids
Electroplating/electrochemical reactions Liquid Electrolytes or conductive liquids
Electrophoretic separation Gel Polyacrylamide or agarose gels [37]
Electrochromatographic separation Sol-Gel Porous photopolymerized sol-gel
Optical Guiding and switching light (waveguides) Liquid Liquids with high/low index of refraction [29,30]
Focusing light (lenses) Liquid Liquids with high/low index of refraction [31]
Generation of light (lasers, optical sources) Liquid Fluorescent or phosphorescent dyes [32,33]
Diffraction of light (diffraction gratings) Foam Gas-liquid foams [51]
Structural Micromechanical support (scaffolding) Solid Epoxies, cements, UV/heat curable polymers, or low melting-point metals [11,12,13]
Solid support for bead packing and chromatography Solid, Gel Porous polymer and gel-based monoliths [12]
Support for plant growth (hydrostats/artificial soils) Gel Sodium/potassium/ammonium polyacrylates
Tissue engineering Gel Collagen, gels, other biocompat. solids [14,40–44]
Open in a new tab

For some materials, it is necessary to modify the interior surface of the microfluidic channels chemically to improve the wettability of the injected material. For example, prior to injecting low-melting point metal (e.g., In) or metal alloy (e.g., eutectic In/Sn) into a PDMS channel to form a long, continuous wire, it is first necessary to oxidize the surface of the channel, and then to modify it with a silane (3-mercaptopropyltrimethoxysilane) so that the metal wets it.[11] When injected into untreated PDMS channels, most metals in liquid form do not wet the walls of the channels. As a result, they tend to fracture upon cooling and solidification, and to lose their electrical connectivity/conductivity.

III. Devices Produced Using Co-Fabrication

This section describes several devices that demonstrate the design, operation, and characterization of co-fabricated microsystems.

Liquid-Filled Channels (Microfluidics)

Over the past decade, the study of the flow of liquids in microfluidic channels and the applications of liquid-filled microchannels has emerged as an area of great scientific and technological interest.[16,17] A number of groups have investigated and modeled the physics of microfluidic laminar flow in microsystems.[18] Microfluidic channels filled with media are useful for the study of cells[19] or whole microorganisms.[20] Liquid-filled channels have been applied for the development of a variety of practical “lab-on-a-chip” devices including devices for high-throughput analysis,[21] generation of chemical gradients,[22] fluidic mixing,[23], and detection and diagnosis of disease.[24]

Metal-Filled Channels (Microsolidics)

Agrowing interest in automated lab-on-a-chip systems has fueled the development of microfluidic systems in which metal wires provide electronic function to the system.[25] Using the principles of co-fabrication, we have produced current-carrying wires by injecting liquid metal in microfluidic channels and allowing the metal to cool and solidify;[11] for very low melting temperature metals, we also used the metal directly in liquid form.[26] This method of fabrication—which we call microsolidics—ultimately generates solid metal structures in two or three dimensions that can be fabricated in close alignment with microfluidic channels (Figure 2). The wires can be fabricated with high aspect ratios and cross sectional areas up to 10 × 106 μm2; this large cross-sectional area is difficult to achieve in a small footprint using alternative techniques such as evaporation or sputter deposition, and makes it possible to use the wires for applications requiring high electrical current (e.g., for generation of magnetic fields). The general procedure consists of five steps: (i) fabrication of microfluidic channels in PDMS using soft lithography and rapid prototyping; (ii) plasma oxidation and silanization of the inside surfaces of the microchannels with 3-mercaptopropyltrimethoxysilane (0.1 M solution in acetonitrile) to make them wettable to the metals we use; (iii) introduction of molten metal into the channels by applying a vacuum to draw metal into the channels—walls of silanized channels are rapidly wet, for example, by liquid indium, eutectic indium-tin, or eutectic gallium-indium, walls of unmodified channels are not; (iv) cooling the channels to form solid metal microstructures; and (v) (if desired) forming the solder-filled system of channels into non-planar topologies by bending, twisting, rolling, or deforming. This last step—modifying the shape—works particularly well when the metal is encapsulated in thin (50–200 μm) layers of PDMS. Using this technique, it is possible to fabricate metallic structures with lateral dimensions as small as 50 μm, and thicknesses of 10 μm – 1 mm. Because the structures are embedded in polymer insulation, breaks or fractures in the metal can often be repaired by heating the metal above its melting point and cooling (perhaps with brief sonication).

Figure 2.

Figure 2

Open in a new tab

Filling microchannels with low-melting temperature solders (microsolidics). a) Schematic diagram describing the injection of low-melting point solders in microchannels. b) A metallic wire (100% In) embedded in a PDMS substrate; the wire can be twisted while still maintaining electrical conductivity (adapted from [11]). c) Mesh structures formed by injecting metal in PDMS microfluidic chambers. d-e) A system of three microfluidic channels, each filled with a different low-melting temperature material. In d), the outer channels are filled with 100% In metal, and the central channel is filled with a Bi56/Pb44 solder. In e), the outer channels are filled with 100% In metal, and the central channel is filled with 100% Se (a semiconducting non-metal); the three channels are connected together electrically, and form a metal-semiconductor-metal junction.

Microheaters

Microsolidic wires can be used to build precision-integrated heaters for microfluidic channels.[11] Indium metal (100%) was injected in microfluidic channels and cooled to form solid, resistive metal wires in microchannels 100 μm × 100 μm in cross-sectional area. Passing electrical current through the wires increased the temperature of the metal by Joule heating, which, in turn, increased the temperature of adjacent microfluidic channels. Microsolidic heaters with cross-sectional areas of 104 μm2 supported currents up to 2 A (2 × 108 A·m−2) without breaking down, and have been shown to heat static or flowing aqueous fluids in adjacent microfluidic channels to >100ºC.

In-plane electromagnets

Microsolidic wires have also been used to build integrated electromagnets for separations (Figure 3).[27,28] Siegel et al. co-fabricated metal wires in close proximity (distance ~10 μm) to liquid-filled microfluidic channels. Passing electrical current through the wires generated magnetic fields up to 2.8 mT and magnetic field gradients up to 40 T·m−1, orthogonal to the direction of current flow. These gradients captured/released and sorted superparamagnetic beads suspended in the adjacent liquid-filled microfluidic channels with forces up to 3.3 pN.

Figure 3.

Figure 3

Open in a new tab

Co-fabricated electromagnets. a) Photographs of three microfluidic channels as viewed from above at low magnification (left), high magnification (upper right) and the cross-section of the channels (lower right); all channels are 40 μm in height (adapted from [27]). The two outer channels, which form the electromagnets, are filled with solder, and the central channel is filled with solution. b) Passing electrical current through the outer metal wires generates a magnetic field, which is subsequently used to move superparamagnetic beads suspended in solution in the central channel. c-e) the capture/release of superparamagnetic beads in a microfluidic channel in proximity to two electromagnets (in black); the arrows represent the direction of electrical current passing through each electromagnet. The current generated a magnetic field oriented into the plane of the page with greatest magnitude at the edge of the channel closest to the electromagnet. The process was repeated over 1000 times with similar results.

Liquid-Liquid Optical Systems

The principles of co-fabrication can also be applied to fabricate reconfigurable optical components.[29,30] One category of these devices, which we call liquid-liquid (L2) systems, typically comprise two or more streams of liquid flowing in microfluidic channels, each possessing different optical properties (such as refractive indices) (Figure 4a). At low and moderate Reynolds numbers, flow of these streams is laminar;[17] this property creates optically smooth liquid-liquid interfaces which exhibit low optical losses from scattering.

Figure 4.

Figure 4

Open in a new tab

a) Schematic representation of the typical design of a L2 waveguide. The index of refraction of liquid 2 is higher than that of liquids 1 and 3 for the guiding of light. b) Optical micrographs of the cross section of the outlet of the microfluidic channel (outlined by the dashed box) as viewed through the transparent window. The rates of flow (μL/min) (and the residence time in seconds) of the core and the cladding are listed (adapted from [29]). c) Optical switch. (i)-(iii) Optical micrographs of the top view of the microfluidic channels. (iv)-(vi) Optical micrographs of the cross section of the end of the channel showing light exiting the L2 waveguides. The white arrows and lines represent the location of the ends of the branches of the microfluidic channel.

Liquids can be replaced and/or replenished continuously in liquid-liquid systems. This capability for replacement allows injection of liquids with different properties (e.g., index of refraction, absorption, fluorescence) to tune the optical output of a microsystem in real time. The ability to replenish liquids minimizes the impact of photobleaching and related phenomena, because bleached elements are replaced continuously. In addition, the liquid-liquid interface is deformable: by manipulating the flow conditions, it is possible to change the position or the shape of the liquid-liquid interface, and therefore, the path of light inside the optofluidic devices.

Optofluidic waveguides

Optical waveguides can be formed in microfluidic channels in which the laminar streams, themselves, control the direction of the flow of light.[29] These devices consist of two streams of liquids with lower refractive index (the cladding) sandwiching a stream of liquid with higher refractive index (the core) in a single microchannel; light is supplied from an optical fiber inserted in PDMS through a fiber port fabricated at the end of the channel (Figure 4b). By controlling the relative rates of flow of the core and cladding liquids, we changed the width of the core stream, and therefore, the width of the optical waveguide. Decreasing the ratio of flow rates of the core to the cladding streams decreased the core size from >100 μm to <10 μm, and switched the guiding from multi- to single-mode.

Optofluidic switches and couplers

Co-fabricated liquid-liquid components can be used to build complex devices that control the direction of propagation of light in microfluidic channels.[29] We built an optical switch by branching an liquid-liquid waveguide into three separate outlet channels (Figure 4c). Changing the relative rates of flow of the cladding liquids changed the path of the core liquid, and therefore, the path of the light. It was also possible to fabricate an evanescent coupler comprising two liquid-liquid waveguides sharing an inner cladding stream with a width < 5 μm. The efficiency of coupling of evanescent fields between the two waveguides was modified by changing the rate of flow of the liquids, which in turn, adjusted the width of the inner cladding stream

Optofluidic lenses

We used the principles of co-fabrication to build a liquid-liquid lens, where the curvature of the interface could be modified dynamically (Figure 5a).[31] The lens was formed by laminar flow of three streams of fluids in a microchannel; the index of refraction of the core stream was higher than the index of the cladding streams. The streams entered an “expansion chamber”, or a region of the microchannel with increased width. For appropriate rates of flow, the shape of the interface between the core and cladding streams in this region was biconvex; this structure focused light propagating in the plane of the expansion chamber perpendicular to the direction of flow of the liquids. Changing the relative rates of flow of the three streams modified the curvature of the interface—and thus the focal distance of the lens—in real time.

Figure 5.

Figure 5

Open in a new tab

a) Bright-field image an L2 lens co-fabricated in a single layer of lithography. (adapted from [31]). b) Top-view scheme for the array of L2 fluorescent light sources, consisting of parallel L2 waveguides in a single PDMS microchannel. Inset: Optical micrograph of the T-junction. c) Spectral output (solid line) from an array of L2 fluorescent light sources containing 0.5 mM solutions of perylene, fluorescein, and sulforhodamine B in DMSO/EG (1:1), with various cladding liquids (adapted from [32]).

Optofluidic light sources

We developed various on-chip fluidic light sources for optical detection and spectroscopic analysis in integrated microanalytical systems (μTAS).[32,33] In these systems, solutions containing fluorescent dye flowed through microchannels and were excited optically by incident light from an external halogen bulb or a pump laser (Figures 5b, c). Microchannel-based liquid-liquid fluorescent light sources circumvent the need to align and secure optical fibers to microfluidic channels, and can be produced with a range of cross-sectional diameters.

Electronic Interfacing with Surfaces

Sometimes it is possible to gain additional function in a co-fabricated system by placing the layer of microfluidic channels—roughly and with minimal precision—perpendicular to a chemically patterned surface. We co-fabricated an array of seven molecular-scale tunnel junctions[34] each comprising an electrically conductive self-assembled monolayer (SAM) connected to two metallic electrodes via a microfluidic channel (Figure 6a).[35] We aligned a PDMS microchannel perpendicularly to Ag electrodes on a glass substrate and sealed it to the surface; the Ag electrodes were patterned with SC11Fc (Fc = ferrocene) SAMs. Injecting liquid metal (EGaIn = eutectic Ga-In)[36] in the microchannel formed electrical contact between the SAM-Ag electrode pair and completed the molecular junction. An optical micrograph of the devices is shown in Figure 6b. The resulting Ag-SC11Fc//EGaIn junctions (‘−’ indicates a covalent bond and ‘//’ indicates a van der Waals contact) rectified currents with rectification ratios R (R = |J(−V)|/|J(+V)|, J = current density (A/cm2), V = the voltage (V)) of 102 at V = ± 1.0 V. The electrodes and microfluidic channel were arranged in a 7 × 1 crossbar scheme (comprising seven Ag electrodes and one EGaIn-filled microfluidic channel).

Figure 6.

Figure 6

Open in a new tab

Co-fabrication of molecular electronic devices in microfluidic channels. a) Schematic of the fabrication of the metal–SAM//EGaIn junctions. b) Top-down micrograph of two TS Ag-electrodes (of seven total) with EGaIn in the perpendicularly aligned PDMS channel. c) The average values of the current density (J) as a function of voltage (V) applied across the junction. The shape of the J/V curve is asymmetric: the difference in average slope of J for negative values of V and positive values of V illustrates the rectification of electronic signals through the device (adapted from [35]).

The co-fabricated tunnel junctions gave working devices in ~80% yield, and reliably rectified AC electronic signals.[35] Figure 6c shows the average current density as a function of the potential of the tunnel junctions. These results represent a first step in applying co-fabrication to produce functional molecular electronic devices in microfluidic channels.

In-channel Gels for Electrophoresis and Cellular Growth

Co-fabrication of gels and other multi-phase materials in a single layer of microfluidic channels produces microdevices with useful function in biology. Typically, gels are injected in the channels in liquid form, and then cured by heating, cooling, or UV irradiation. A variety of groups have injected polyacrylamide gels in microfluidic channels or networks of channels for applications in 1D and 2D electrophoretic separation.[37] Other types of gels, including collagen- and acrylate-based hydrogels have been injected in microfluidic channels for microfluidic valving and flow control,[38] colorimetric molecular sensing,[39] tissue engineering,[40,41] and for constructing cellular and biomolecular microenvironments.[42–44]

Bubble and Droplet Generation in Microchannels

There has been growing interest in producing emulsions such as foams in microfluidic devices.[45,46] We and others have used co-fabrication to generate monodisperse emulsions of bubbles,[47] droplets,[48] and polymer particles[49] in microchannels that exhibit remarkable regularity. Flow-focusing[47,49] and T-junction[48] geometries allow the user to tune the size and the fraction of volume that the bubbles and droplets occupy in the outlet channel. Garstecki et al. demonstrated the self-organization of bubbles in a variety of ordered gas-liquid and liquid-liquid lattices in microfluidic channels.[47] The form and structure of these lattices was adjusted by changing the pressures applied to the two immiscible fluids. These microfluidic lattices have four practical benefits. They are: i) self-organizing, (ii) self-healing, (iii) dynamically addressable, and (iv) made from components that are constantly replenished for real-time control and modification of the lattice.

Complex, periodic emulsions

Co-fabricating multiple flow-focusing generators in single layer of microfluidic channels makes it possible to produce complex, periodic structures of bubbles and droplets (Figure 7a).[50] The generators—when placed in close proximity to each other—allow fluctuations in the local fluidic resistance in a single channel, or in a network of channels, and thus make possible the synchronization of timing in formation of droplets and bubbles. This synchronization can produce stable, composite structures comprising different emulsions (droplets or bubbles) in the same continuous phase.[47] In one application of this approach, Hashimoto et al. used the self-organization of bubbles to produce microfluidic diffraction gratings.[51] Changing the external flow parameters (i.e., the rate of flow of water, and the applied pressure of nitrogen) adjusted the diameter of the bubbles that were generated from the flow-focusing geometry (and thus, the lattice constant of the grating structure), and made it possible to tune the angle of diffraction of light from the device dynamically and in real time.

Figure 7.

Figure 7

Open in a new tab

Co-fabrication of emulsions in microfluidic channels. a) (left) A schematic representation of three flow-focusing generators connected to a single outlet channel. The abbreviations “cp” and “dp” denote inlets with continuous phase and dispersed phase, respectively, and arrows indicate the direction of flow. (right) Optical micrographs of composite lattices comprising nitrogen bubbles and hexadecane droplets suspended in water (adapted from [50]). b) Using droplets as microreactors. Aqueous droplets containing two reagents were formed and encapsulated in a fluorinated carrier fluid to facilitate transport through a buffer (adapted from [52]). c) Crystallization in droplets. (top) A schematic illustration of the droplet forming process: decreasing the flow rate of the NaCl stream and increasing the flow rate of the buffer increased the concentration of NaCl in each droplet. (bottom) A polarized microphotograph showing crystallization of lysozyme inside 7.5-nL aqueous droplets containing PEG, salt, lysozyme, and acetate buffer with variable composition (adapted from [53]).

Chemical reactions in droplets

Droplets formed by flowing immiscible fluids in co-fabricated microfluidic channels can be used as vehicles for performing chemical reactions (Figure 7b).[52] Compartmentalization in droplets allows for the rapid mixing of reagents, provides precise control of the timing of reactions, and makes it possible to synthesize and transport solid products. Ismagilov and co-workers applied this strategy to the crystallization of proteins in microfluidic channels (Figures 7c).[53] The crystallization trials were set up inside 7.5-nL aqueous droplets, each containing solutions of protein, precipitants, and additives in variable ratios. This method allowed researchers to set up hundreds of crystallization trials under computer control at rates of several trials per second.

Integrated droplet generation and dielectrophoretic sorting

One advantage of co-fabrication is the ability to combine multiple materials—and thereby multiple functions—in a single device. An example is fluorescence-activated droplet sorting.[54] In this approach, Baret et al. compartmentalized single Escherichia coli in droplets produced from a co-fabricated microfluidic droplet generator, and then sorted the cells dielectrophoretically into separate chambers using metallic electrodes co-fabricated in the same device. This method combines the advantages of microtitre-plate screening and traditional fluorescence-activated cell sorting, and demonstrates how two strategies of co-fabrication (here metal electrodes and droplet generation) can be integrated in practical applications.

IV. Expanding Capabilities Through Co-Fabrication: New Methods and Materials

Co-fabrication is already a useful technique for a variety of applications in science and technology, but the approach faces several hurdles on its path to becoming a reliable technique for mass production. For example, the technique is currently limited to materials with low melting points (generally < 300°C) and with affinity for the surface of the channel walls. In addition, while materials such as aqueous fluids are carried into channels via capillary action, others must be injected manually using vacuum or pressure (typically via a syringe), making it difficult to produce devices rapidly en masse.

We believe that further research in co-fabrication will help overcome many of these limitations. Advances in automation, such as robotic-based fluid injection, will help facilitate production in high volume. In addition, because co-fabrication does not depend on the precise alignment of layers, it may be possible to produce the bas-relief for the channels using techniques not commonly used for microfabrication, such as plastic injection molding or laser ablation (which can be outsourced at low cost using a commercial vendor). We believe many of the practical advancements of co-fabrication will be motivated by efficiency and cost-control, and may emerge as the technique is adapted to commercial environments.

We also believe that new applications will evolve when new materials (and combinations of materials) are injected into microchannels. Table 2 suggests several possibilities, some of which have been demonstrated, and some of which have not. In co-fabrication, structural and functional elements are added to channels post-formation, and hence, the only factor limiting the choice of materials is the ability to fill channels with them. In the future, we envision systems comprising high-melting-temperature materials such as glasses, structural metals, or temperature-resistant polymers, and systems based on novel materials that are formed by combining two or more substances in a single microchannel (e.g., combination of PDMS and Fe3O4 to produce functional ferromagnetic ceramics).

V. Conclusions

Co-fabrication represents a new model for designing and building multi-component microsystems. The approach is especially applicable to technologies that require the integration and registration of multiple different materials in one or more layers with micron-precision, but where simplicity and economy of fabrication are key characteristics of the design process. We anticipate that advances in co-fabrication will lead to new, practical applications in energy storage (e.g., batteries), lab-on-a-chip systems, flexible electronics (e.g., for radio-frequency identification – RFID), systems biology, thermal management, biochemical sensors and diagnostics, self-standing 3-D structures (in which the surrounding insulation is removed after fabrication), and compound multilayer microsystems produced by combining several co-fabricated layers in a single device.

Figure 1.

Figure 1

Open in a new tab

Co-fabrication involves three steps: (i) designing the microsystem in such a way that a single layer of separate sub-systems of microfluidic channels provides the structural basis for all of the components and functions required in the final device, (ii) fabricating these channels by micromolding, and (iii) generating function by filling the separate sub-systems of microfluidic channels with appropriate functional materials in fluid form. In this example, a simple step of micromolding ultimately generates two independent, aligned metal wires, a microfluidic channel, and an optical structure.

Acknowledgments

The authors thank Katherine Mirica and Zhihong Nie for helpful discussions. This work was funded by the Bill & Melinda Gates Foundation under award # 51308 “Zero-Cost Diagnostics”, DARPA through the Caltech optofluidics grant, and the MF3 Micro/Nano Fluidics Fundamentals Focus Center. A.C.S. acknowledges financial support from the Howard Hughes Medical Institute and the Harvard-MIT Division of Health Science and Technology. S.T.P acknowledges financial support from NIH. C.A.N. acknowledges NWO for financial support.

Biographies

Adam C. Siegel received his Ph.D. in 2009 from Harvard University under the direction of George Whitesides in the Department of Chemistry. His doctoral work focused on the design and fabrication of microsystems based upon unconventional materials such as silicones, papers, and plastics, with applications in electronics and biotechnology.

Sindy K.Y. Tang is a Ph.D. student at Harvard University working under the direction of George Whitesides. Her research focuses on developing reconfigurable optical devices using liquid-liquid interfaces in microfluidic systems.

Christian A. Nijhuis received his Ph.D. at the University of Twente under the direction of David N. Reinhoudt, where he focused on the electrochemically controlled self-assembly of supramolecular complexes at surfaces and their application in nanofabrication. He currently holds a postdoctoral position in the George Whitesides group at Harvard University.

Michinao Hashimoto received his Ph.D. degree in Chemical Physics in 2009 from Harvard University, under the direction of George Whitesides. His doctoral work focused on development of methods for controlled generation of emulsions in microfluidic systems.

Scott T. Phillips is the Martarano Assistant Professor in the Department of Chemistry at The Pennsylvania State University. His research interests involve the design and synthesis of molecules with unique function, and include areas such as unconventional reaction methodology, environmental chemistry, and bioanalytical chemistry.

Michael Dickey received his Ph.D. in 2006 from the University of Texas at Austin under the direction of Grant Willson, where he focused on new patterning techniques such as imprint lithography. He recently completed a post-doctoral fellowship at Harvard University under the guidance of George Whitesides and is currently an assistant professor at NC State University.

George M. Whitesides was born August 3, 1939 in Louisville, KY. He received an A.B. degree from Harvard University in 1960 and a Ph.D. from the California Institute of Technology (with J.D. Roberts) in 1964. He was a member of the faculty of the Massachusetts Institute of Technology from 1963 to 1982, and joined the Department of Chemistry of Harvard University 1982, serving as department chairman from 1986-89. He is now the Woodford L. and Ann A. Flowers University Professor.

References

  • 1.McDonald JC, Whitesides GM. Poly(dimethylsiloxane) as a Material for Fabricating Microfluidic Devices. Acc Chem Res. 2002;35:491–499. doi: 10.1021/ar010110q. [DOI] [PubMed] [Google Scholar]
  • 2.Mukhopadhyay R. When PDMS Isn’t the Best. Anal Chem. 2007;79:3248–3253. doi: 10.1021/ac071903e. [DOI] [PubMed] [Google Scholar]
  • 3.Hung LH, Lina R, Lee AP. Rapid Microfabrication of Solvent-Resistant Biocompatible Microfluidic Devices. Lab Chip. 2008;8:983–987. doi: 10.1039/b717710k. [DOI] [PubMed] [Google Scholar]
  • 4.Bustillo JM, Howe RT, Muller RS. Surface Micromachining for Microelectromechanical Systems. Proc IEEE. 1998;86:1552–1574. [Google Scholar]
  • 5.Qin D, Xia Y, Rogers JA, Jackman RJ, Zhao XM, Whitesides GM. Microfabrication, Microstructures and Microsystems. Top Current Chem. 1998;194:1–20. [Google Scholar]
  • 6.Becker H, Gartner C. Polymer Microfabrication Methods for Microfluidic Analytical Applications. Electrophoresis. 2000;21:12–26. doi: 10.1002/(SICI)1522-2683(20000101)21:1<12::AID-ELPS12>3.0.CO;2-7. [DOI] [PubMed] [Google Scholar]
  • 7.Mijatovic D, Eijkel JCT, van den Berg A. Technologies for Nanofluidic Systems: Top-Down vs. Bottom-Up – A Review. Lab Chip. 2005;5:492–500. doi: 10.1039/b416951d. [DOI] [PubMed] [Google Scholar]
  • 8.Martinez AW, Phillips ST, Butte M, Whitesides GM. Patterned Paper as a Platform for Inexpensive, Low Volume, and Portable Bioassays. Angew Chem Int Ed. 2007;46:1318–1320. doi: 10.1002/anie.200603817. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Oswald TA, Turng LS, Gramann P. Injection Molding Handbook. 2. Hanser Gardner Publications; Cincinnati USA: 2007. [Google Scholar]
  • 10.Piotter V, Benzler T, Gietzelt T, Ruprecht R, Hausselt J. Adv Eng Mat. 2000;2:639. [Google Scholar]
  • 11.Siegel AC, Bruzewicz DA, Weibel DB, Whitesides GM. Microsolidics: Fabrication of Three-Dimensional Metallic Structures using Microfluidics. Adv Mater. 2007;19:727–733. [Google Scholar]
  • 12.Peterson DS. Solid Supports for Micro Analytical Systems. Lab Chip. 2005;5:132–139. doi: 10.1039/b405311g. [DOI] [PubMed] [Google Scholar]
  • 13.Kim E, Xia Y, Whitesides GM. Polymer Microstructures Formed by Molding in Capillaries. Nature. 1995;376:581–584. [Google Scholar]
  • 14.Cheung YK, Gillette BM, Zhong M, Ramcharan S, Sia SK. Direct Patterning of Composite Biocompatible Microstructures Using Microfluidics. Lab Chip. 2007;7:574–579. doi: 10.1039/b700869d. [DOI] [PubMed] [Google Scholar]
  • 15.Leng J, Salmon JB. Microfluidic Crystallization. Lab Chip. 2009;9:24–34. doi: 10.1039/b807653g. [DOI] [PubMed] [Google Scholar]
  • 16.Whitesides GM, Stroock AD. Flexible Methods for Microfluidics. Physics Today. 2001;54:42–48. [Google Scholar]
  • 17.Squires TM, Quake SR. Microfluidics: Fluid Physics at the Nanoliter Scale. Rev Mod Phys. 2005;77:977–1026. [Google Scholar]
  • 18.Atencia J, Beebe DJ. Controlled Microfluidic Interfaces. Nature. 2005;437:648–655. doi: 10.1038/nature04163. [DOI] [PubMed] [Google Scholar]
  • 19.Whitesides GM, Ostuni ES, Takayama S, Jiang X, Ingber DE. Soft Lithography in Biology and Biochemistry. Ann Rev Biomed Eng. 2001;3:335–373. doi: 10.1146/annurev.bioeng.3.1.335. [DOI] [PubMed] [Google Scholar]
  • 20.Hulme SE, Shevkoplyas SS, Samuel A. Microfluidics: Streamlining Discovery in Worm Biology. Nat Methods. 2008;5:589–590. doi: 10.1038/nmeth0708-589. [DOI] [PubMed] [Google Scholar]
  • 21.Thorsen T, Maerkl SJ, Quake SR. Microfluidic Large-Scale Integration. Science. 2002;298:580–584. doi: 10.1126/science.1076996. [DOI] [PubMed] [Google Scholar]
  • 22.Jeon NL, Dertinger SK, Chiu DT, Choi IS, Stroock AD, Whitesides GM. Generation of Solution and Surface Gradients Using Microfluidic Systems. Langmuir. 2000;16:8311–8316. [Google Scholar]
  • 23.Stone HA, Stroock AD, Ajdari A. Engineering Flows in Small Devices. Ann Rev Fluid Mech. 2004;36:381–411. [Google Scholar]
  • 24.Chin CD, Linder V, Sia SK. Lab-on-a-Chip Devices for Global Health. Lab on a Chip. 2007;7:41–57. doi: 10.1039/b611455e. [DOI] [PubMed] [Google Scholar]
  • 25.Verpoorte E, De Rooij NF. Microfluidics Meets MEMS. Proc IEEE. 2003;91:930–953. [Google Scholar]
  • 26.Dickey MD, Chiechi RC, Larsen RJ, Weiss EA, Weitz DA, Whitesides GM. Eutectic Gallium-Indium (EGaIn): A Liquid Metal Alloy for the Formation of Stable Structures in Microchannels at Room Temperature. Adv Func Mat. 2008;18:1097–1104. [Google Scholar]
  • 27.Siegel AC, Shevkoplyas SS, Weibel DB, Bruzewicz DA, Martinez AW, Whitesides GM. Co-Fabrication of Electromagnets and Microfluidic Systems. in Poly(dimethylsiloxane) Angew Chem Int Ed. 2006;45:6877–6882. doi: 10.1002/anie.200602273. [DOI] [PubMed] [Google Scholar]
  • 28.Shevkoplyas SS, Siegel AC, Westervelt RM, Prentiss MG, Whitesides GM. The Force Acting on a Superparamagnetic Bead Due to an Applied Magnetic Field. Lab Chip. 2007;7:1294–1302. doi: 10.1039/b705045c. [DOI] [PubMed] [Google Scholar]
  • 29.Wolfe DB, Conroy RS, Garstecki P, Mayers BT, Fischbach MA, Paul KE, Prentiss M, Whitesides GM. Dynamic Control of Liquid-Core/Liquid-Cladding Optical Waveguides. Proc Natl Acad Sci USA. 2004;101:12434–12438. doi: 10.1073/pnas.0404423101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Psaltis D, Quake SR, Yang CH. Developing Optofluidic Technology Through the Fusion of Microfluidics and Optics. Nature. 2006;442:381–386. doi: 10.1038/nature05060. [DOI] [PubMed] [Google Scholar]
  • 31.Tang SKY, Stan CA, Whitesides GM. Dynamically Reconfigurable Liquid-Core Liquid-Cladding Lens in a Microfluidic Channel. Lab Chip. 2008;8:395–401. doi: 10.1039/b717037h. [DOI] [PubMed] [Google Scholar]
  • 32.Mayers BT, Vezenov DV, Vullev VI, Whitesides GM. Arrays and Cascades of Fluorescent Liquid-Liquid Waveguides: Broadband Light Sources for Spectroscopy in Microchannels. Anal Chem. 2005;77:1310–1316. doi: 10.1021/ac048692n. [DOI] [PubMed] [Google Scholar]
  • 33.Tang SKY, Li Z, Abate AR, Agresti JJ, Weitz DA, Psaltis D, Whitesides GM. A multi-color fast-switching microfluidic droplet dye laser. Lab Chip. 2009;9:2767–2771. doi: 10.1039/b914066b. [DOI] [PubMed] [Google Scholar]
  • 34.Akkerman HB, De Boer B. Electrical Conduction Through Single Molecules and Self-Assembled Monolayers. J Phys Condens Matter. 2008;20:013001. [Google Scholar]
  • 35.Nijhuis CA, Dickey M, Whitesides GM. Manuscript in preparation. [Google Scholar]
  • 36.Chiechi RC, Weiss EA, Dickey MD, Whitesides GM. Eutectic Gallium-Indium (EGaIn): A Moldable Liquid Metal for the Electrical Characterization of Self-Assembled Monolayers (SAMs) Angew Chem Intl Ed. 2007;47:142. doi: 10.1002/anie.200703642. [DOI] [PubMed] [Google Scholar]
  • 37.Wu D, Qin J, Lin B. Electrophoretic Separations on Microfluidic Chips. J Chromat A. 2008;1184:542–559. doi: 10.1016/j.chroma.2007.11.119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Beebe DJ, Moore JS, Bauer JM, Yu Q, Liu RH, Devadoss C, Jo BH. Functional Hydrogel Structures for Autonomous Flow Control Inside Microfluidic Channels. Nature. 2000;404:588–590. doi: 10.1038/35007047. [DOI] [PubMed] [Google Scholar]
  • 39.Lee NY, Jung YK, Park HG. On-chip Colorimetric Biosensor Based On Polydiacetylene (PDA) Embedded in Photopolymerized Poly(ethylene glycol) Diacrylate (PEG-DA) Hydrogel. Biochem Eng J. 2006;29:103–108. [Google Scholar]
  • 40.Gillette BM, Jensen JA, Tang B, Yang GJ, Bazargan-Lari A, Zhong M, Sia SK. In Situ Collagen Assembly for Integrating Microfabricated Three-Dimensional Cell-Seeded Matrices. Nat Mater. 2008;7:636–640. doi: 10.1038/nmat2203. [DOI] [PubMed] [Google Scholar]
  • 41.Choi NW, Cabodi M, Held B, Gleghorn JP, Bonassar LJ, Stroock AD. Microfluidic Scaffolds for Tissue Engineering. Nat Mater. 2007;6:908–915. doi: 10.1038/nmat2022. [DOI] [PubMed] [Google Scholar]
  • 42.Wong AP, Perez-Castillejos R, Love JC, Whitesides GM. Partitioning Microfluidic Channels with Hydrogel to Construct Tunable 3-D Cellular Microenvironments. Biomaterials. 2008;29:1853–1861. doi: 10.1016/j.biomaterials.2007.12.044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Ling Y, Rubin J, Deng Y, Huang C, Demirci U, Karp J, Khademhosseini A. A Cell-Laden Microfluidic Hydrogel. Lab Chip. 2007;7:756–762. doi: 10.1039/b615486g. [DOI] [PubMed] [Google Scholar]
  • 44.Bruzewicz DA, McGuigan AP, Whitesides GM. Fabrication of a Modular Tissue Construct in a Microfluidic Chip. Lab Chip. 2008;8:663–671. doi: 10.1039/b719806j. [DOI] [PubMed] [Google Scholar]
  • 45.Shaha RK, Shuma HC, Rowata AC, Leea D, Agrestia JJ, Utada AS, Chua LY, Kima JW, Fernandez-Nievesa A, Martineza CJ, Weitz DA. Designer Emulsions Using Microfluidics. Mater Today. 2008;11:18–27. [Google Scholar]
  • 46.Garstecki P, Ganan-Calvo AM, Whitesides GM. Formation of Bubbles and Droplets in Microfluidic Systems. Bull Polish Acad Sci - Tech Sci. 2005;53:361–372. [Google Scholar]
  • 47.Garstecki P, Gitlin I, DiLuzio W, Whitesides GM, Kumacheva E, Stone HA. Formation of Monodisperse Bubbles in a Microfluidic Flow-Focusing Device. App Phys Lett. 2004;85:2649–2651. [Google Scholar]
  • 48.Thorsen T, Roberts RW, Arnold FH, Quake SR. Dynamic Pattern Formation in a Vesicle-Generating Microfluidic Device. Phys Rev Lett. 2001;86:4163–4166. doi: 10.1103/PhysRevLett.86.4163. [DOI] [PubMed] [Google Scholar]
  • 49.Xu SQ, Nie ZH, Seo M, Lewis P, Kumacheva E, Stone HA, Garstecki P, Weibel DB, Gitlin I, Whitesides GM. Generation of Monodisperse Particles by Using Microfluidics: Control over Size, Shape, and Composition. Angew Chem Intl Ed. 2005;44:724–728. doi: 10.1002/anie.200462226. [DOI] [PubMed] [Google Scholar]
  • 50.Hashimoto M, Garstecki P, Whitesides GM. Synthesis of Composite Emulsions and Complex Foams with the use of Microfluidic Flow-Focusing Devices. Small. 2007;3:1792–1802. doi: 10.1002/smll.200700238. [DOI] [PubMed] [Google Scholar]
  • 51.Hashimoto M, Mayers B, Garstecki P, Whitesides GM. Flowing Lattices of Bubbles as Tunable, Self-Assembled Diffraction Gratings. Small. 2006;2:1292–1298. doi: 10.1002/smll.200600211. [DOI] [PubMed] [Google Scholar]
  • 52.Song H, Chen DL, Ismagilov RF. Reactions in Droplets in Microfluidic Channels. Angew Chem Int Ed. 2006;45:7336–7356. doi: 10.1002/anie.200601554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Zheng B, Roach LS, Ismagilov RF. Screening of Protein Crystallization Conditions on a Microfluidic Chip Using Nanoliter-Size Droplets. J Am Chem Soc. 2003;125:11170–11171. doi: 10.1021/ja037166v. [DOI] [PubMed] [Google Scholar]
  • 54.Baret JC, Miller OJ, Taly V, Ryckelynck M, El-Harrak A, Frenz L, Rick C, Samuels ML, Hutchison JB, Agresti JJ, Link DR, Weitz DA, Griffiths AD. Fluorescence-activated droplet sorting (FADS): efficient microfluidic cell sorting based on enzymatic activity. Lab Chip. 2009;9:1850. doi: 10.1039/b902504a. [DOI] [PubMed] [Google Scholar]

RESOURCES