This invention pertains to polymeric micro-reactors, particularly a device and method of fabricating a complete polymeric micro-reactor for manufacturing chemicals such as nano-materials.
Chemical manufactures currently use a technique referred to as “scale-up” to massively produce chemicals using large-size batch reactors. These batch reactors often require large volumes of raw materials and products, which increase complications associated with large-scale transport and storage, and safety and health issues related to potential explosions, and toxin and flammable solvent leakages.
Microfluidic reactors for process scale-up, based on the concept of parallel processing, are increasingly showing potential for controlling the synthetic aspects of the final product to produce chemicals having higher yield and purity. Micro-reactors minimize some of the health and safety risks associated with traditional chemical scale-up processes, by increasing compound reaction efficiency and the controllability of compound reactions, and by reducing the amounts of raw materials and products needed to induce a compound reaction. Micro-reactors also have higher mass and heat transfer efficiency than traditional chemical processes, and may be used, for example, to perform wet chemical synthesis of nanoparticles. See S. J. Haswell, et al., “Micro-chemical Reactors: The Key to Controlling Chemistry,” Royal Society of Chemistry, vol. 250, pp. 25-33 (2000); and P. Watts et al., “Electrochemical Effects Related To Synthesis In Micro-reactors Operating Under Electrokinetics Flow,” Chemical Engineering Journal, vol. 101 (1-3), pp. 237-240 (2004).
Some obstacles associated with process scale-up using microfluidic reactors include, for example, the high costs associated with fabricating microfluidic reactors using existing rapid prototyping techniques, and the incompatibility of materials used in the fabrication process with chemicals produced by microfluidic reactors. Commercial manufacturers of micro-reactors have traditionally used stainless steel, silicon or borosilicate glass to replicate microfluidic reactors, which often involves lengthy and expensive photolithographic processes. These processes are incapable of achieving deep reactive ion etching (“DRIE”) chemistry, and thus reduce the fabrication efficiency of high aspect ratio channels.
There are several obstacles to the successful rapid proto-typing of microfluidic reactors using polymers. One major obstacle involves sealing of microfluidic channels. Another obstacle involves obtaining a strong bond between microfluidic patterns and substrates. Yet another obstacle involves connecting microfluidic reactors with other instruments, such as pumps, collectors, and detectors. These obstacles limit the ability to fabricate simple, low-cost microfluidic reactors using processes such as LIGA, embossing, casting, injection molding, and imprinting. (“LIGA” is a German acronym for “lithography, electrodeposition, and molding.”) See R. J. Jackman et al., “Microfluidic Systems with On-line UV Detection Fabricated in Photodefinable Epoxy,” J. Micromech. Microeng., vol. 11, pp. 263-269 (2001).
In the last few years, research has been very active on low-cost, mass production microfabrication techniques for manufacturing SU-8-based microfluidic reactors due to the superior chemical and mechanical properties of SU-8, in addition to its ease of fabrication using X-ray or UV-based LIGA processes. Complex and multilayered structures are generally produced with relative ease using SU-8 and other materials, such as polymethyl methacrylate (PMMA), polycarbonate (PC), and polydimethylsiloxane (PDMS), that are compatible with standard silicon processing conditions. As compared to other materials currently used to fabricate micro-reactors, such as PDMS and PMMA, SU-8 appears to be more suitable, especially for fabricating reactors having fluidic channels with large depths (up to 500 μm). However, there are several complications to fabricating microfluidic reactors with SU-8. First, sealing the microfluidic channels fabricated in SU-8 without clogging or blockage is not currently possible. Second, the surface tension of a liquid at the edge of the microfluidic pattern during spin-coating prevents the fabrication of a uniform surface pattern. See C. Lin et al., “A New Fabrication Process For Ultra-Thick Microfluidic Microstructures Utilizing SU-8 Photoresist,” J. Micromech. Microeng. vol. 12, pp. 590-597 (2002).
Until recently, there were no methods for sealing SU-8 microfluidic channels without clogging or blockage, nor were there any methods for fabrication of a uniform surface pattern during spin-coating. These problems were addressed in R. J. Jackman et al., “Microfluidic systems with on-line UV detection fabricated in photodefinable epoxy,” J. Micromech. Microeng., vol. 11, pp. 263-269 (2001), which discloses a process for sealing microfluidic channels using SU-8 without any cross-linking, and C. Lin et al., 2002, which discloses the use of a “constant-volume-injection” method to achieve a flat surface for overcoming edge-bead effects. However, these methods require additional process steps, the use of a thin film laminate of SU-8, the precise control of bonding temperatures, and the fabrication of uniform surfaces, which increase microfluidic fabrication costs.
U.S. Pat. No. 6,686,184 describes microfluidic networks and methods for fabricating microfluidic networks having one or more levels of microfluidic channels. In one embodiment, the microfluidic network comprises a polymeric structure having at least first and second non-fluidically interconnected fluid flow paths, wherein at least the first flow path comprises a series of interconnected channels within the polymeric structure.
U.S. Pat. Pub. No. 2003/0150555 and U.S. Pat. No. 6,123,798 describe methods for fabricating polymeric microfluidic devices that incorporate microscale fluidic structures without substantially distorting or deforming the structures. In one embodiment, the microfluidic device comprises a first polymeric substrate having at least a first planar surface with a plurality of channels disposed therein and a second polymeric substrate layer having at least a first planar surface bonded to the first planar surface of the first substrate, wherein the first surface of the second substrate has a lower glass transition temperature than the first surface of the first substrate. In another embodiment, the first planar surface of the second substrate is non-solvent bonded to the first planar surface of the first substrate, wherein the first surface of the second substrate does not substantially project into the plurality of channels.
U.S. Pat. No. 6,645,432 describes microfluidic systems and methods for fabricating complex, discontinuous patterns onto surfaces that can also incorporate or deposit multiple materials onto the surface. In one embodiment, a microfluidic system comprises a polymeric structure having at least first and second non-fluidically interconnected fluid flow paths, wherein at least the first flow path comprises a series of interconnected channels within the polymeric structure. In another embodiment, the microfluidic system comprises a polymeric membrane having a first surface with at least one channel disposed therein, and a polymeric region intermediate the first surface and the second surface. The intermediate region includes at least one connecting channel there-through which fluidically interconnects the channel disposed in the first surface with the channel disposed in the second surface of the membrane.
U.S. Pat. Pub. No. 2002/0108860 describes microfluidic devices and a process for fabricating microfluidic devices comprising emitting microdroplets of a polymeric material from a nozzle onto a substrate, and forming a pattern of microfluidic device features on the substrate using the polymeric material.
An unfilled need exists for a fast and inexpensive microfabrication technique for creating completely polymeric microfluidic reactors for synthesis of chemicals, such as nanoparticles.
We have discovered an inexpensive apparatus and method for microfabrication of complete polymeric (e.g., SU-8, PMMA, and PEEK) microfluidic reactors suitable for the synthesis of chemicals, particularly nanoparticles ranging in size between about 1 nm and about 2000 nm (e.g., mono, bi, tri, alloy, core-shell, polymeric, and metal-polymer nano-particles). This method is a precise process which uses polymeric microfluidic patterning techniques and a new microfluidic sealing technique, referred to as “flexible semi-solid transfer,” to fabricate high aspect ratio polymeric micro-reactors. The novel method provides an improved means for controlling the micro-reactor fabrication process.
In one embodiment, high quality microfluidic channels (e.g., 4-way mixers, multi-pole mixers, and multi-reaction channels) are patterned using SU-8 on a polymeric substrate, such as a PMMA or PEEK substrate. The microfluidic structure is sealed using a thin (about 40-100 μm) SU-8 film coated on a sacrificial substrate, and then exposed to a small dosage (less than about 480 mJ/cm2) of ultraviolet light. After some of the parts of the microfluidic structure are exposed to UV-light through a mask, the unexposed parts of the microfluidic structure are developed away and the structure bonded with PMMA or PEEK to produce a micro-reactor. Embedded structures may be fabricated between the substrates and inlet and outlet channels of the micro-reactor. The micro-reactor may also be combined with an integrated micro heat exchanger or other external or internal components such as micro pumps, valves, and micro separators. Optionally, to further strengthen the completely polymeric micro-reactor, the micro-reactor may be bonded to metallic substrates (e.g., stainless steel, copper, and gold substrates) or ceramic substrates (e.g., alumina and glass substrates).
A general purpose of this invention is to provide an apparatus and method for rapid production of completely polymeric microfluidic reactors for chemical synthesis, chemical process development, and process scale-up. More specifically, a purpose of this invention is to provide an inexpensive method for rapid fabrication of completely polymeric microfluidic structures suitable for the synthesis of nanoparticles (e.g., mono, bi, tri, alloy, core-shell, polymeric, and metal-polymer). There are several essentials for facilitating the replication of high quality polymeric micro-reactors. First, the fabrication process should be capable of sealing the microfluidic channels, while avoiding complications associated with clogging and blocking, in addition to avoiding the formation of a non-uniform surface pattern caused by liquid surface tension buildup at the edge of the pattern when using techniques such as spin-coating. Second, the fabrication process should be capable of providing a strong bond between the support substrate and the polymeric microfluidic structures. Finally, the fabrication process should be capable of providing suitable connectors between the microfluidic reactor and the external components such as reactant reservoirs, pumps, and inlets for inert gas, and to ensure leak-proof operations.
High chemical compatibility between materials used to construct the microfluidic structure is preferred. The microfluidic structure should be compatible with various solvents and harsh chemicals such as tetrahydrofuran, toluene, acetone, acids (e.g, HCl), bases (e.g., NaOH) used by commercial chemical manufacturers during synthesis, and it should be stable at temperatures up to 200° C. A preferred microfluidic patterning material is SU-8 (MicroChem Corporation, Newton, Mass.). SU-8 is preferred because it is suitable for fabricating reactors having fluidic channels with large depths (up to 500 μm), and it has superior chemical and mechanical properties in addition to its ease of fabrication using X-ray or UV-based LIGA. SU-8 has a high glass transition temperature range (between about 150° C. and about 220° C.), a high shear modulus (between about 6.26 MPa and about 7.49 MPa), Young's modulus from 2396-2605 MPa at R.T. and 653-1017 MPa at 150° C. The max operation pressure could be as high as 2.1 MPa for this prototype. It also has a low loss tangent (tan δ<<0.001).
Selection of Materials for Constructing the Microfluidic Structure
In determining an effective material for rapidly fabricating inexpensive prototype micro-reactors for chemical synthesis, several polymers (e.g., PMMA, PDMS and SU-8) were tested for chemical stability and compatibility. To determine the chemical stability of each polymer, samples were incubated for three days in test solutions such as THF, Lithium hydrotriethylborate, metal salt (i.e., PdCl2), THF solution, 50% (v/v) HNO3, and 50% (v/v) HCl. Afterwards, each polymer sample was inspected for degradation using infrared spectroscopy. THF (99.90% pure packaged under nitrogen), PdCl2 (99%), lithium hydrotriethylborate (LiBH(C2H5)3) as 1 M solution in THF, 3-(N,N-dimethyldodecylammonium)-propanesulfonate (SB-12), and acetone (reagent anhydrous, water<0.5%, 99.9+%) were purchased from Aldrich Chem. Corp., Milwaukee, Wis. No significant changes were observed for SU-8 in the THF, metal chloride or reducing agent solutions, including lithium hydrotriethylborate (LiBH(C2H5)3) as 1M solution in THF. SU-8 also remained stable for 12 hr in the 50% HCl or HNO3 solutions. In comparison, PDMS dissolved in less than 4 hr and PMMA swelled within the first two days.
To confirm that SU-8 had a high level of thermal stability, SU-8 films were prepared under different curing conditions and evaluated using Thermo Gravimetric Analysis (TGA) (TA Instrument Inc., New Castle, Del.). The SU-8 films were stable at temperatures up to about 200° C., with an observed weight loss of about two percent. Previous studies have shown that the Glass Transition Temperature (Tg) of SU-8 films varies depending on curing conditions, and that SU-8 prepared under standard conditions has a Tg of 150° C. Thus, the maximum operating temperature of SU-8-based microfluidic reactors is normally about 150° C. See K. Lian et al. (2003).
Selection of a Suitable Support Substrate for SU-8 Microfluidic Structure
Stainless steel, poly(methyl methacrylate) or PMMA, polyetheretherketone (PEEK), and silicon wafers were considered as a support substrate for the SU-8 microfluidic structure. Silicon wafer was found to be not suitable as a support substrate because it concaved to the side when loaded with an SU-8 multilayer microfluidic pattern due to shrinking stress loads resulting from the cross-linking of SU-8. When a thin layer (4.5 mm) of PMMA was used, the final microfluidic reactor concaved to the side of PMMA. However, when the thickness of the PMMA substrate increased to 12.5 mm, stable structures were obtained. (For PEEK and stainless steel substrates, a 4.5 mm thickness was sufficient.)
While PMMA, PEEK, and stainless steel have good mechanical and machining properties, other important factors for selection of the substrate, include thermal stability of the substrate, thermal expansion coefficient with respect to SU-8, and compatibility with various solvents and harsh chemicals (e.g., tetrahydrofuran, toluene, acetone, acid, and base). Typical thermal expansion coefficients (α), solubility parameters (δ) and Tg for SU-8, stainless steel, PMMA, PEEK, tretrahydrofuran (THF), and acetone are listed in Table 1. As shown in Table 1, stainless steel has good organic solvent compatibility and mechanical properties, but is not compatible with cross-linked SU-8, as shown by the differences in their thermal expansion coefficients. (Experiments have shown that even a small defect in an orifice connecting an SU-8 microfluidic structure body with a stainless steel substrate can lead to debonding and cracking along the defect, resulting in leakage.) In contrast, PMMA has good compatibility with cross-linked SU-8 (as shown by the similar δ and α values shown in Table 1), but may be damaged by solvents such as THF and acetone. By comparison, PEEK seems suitable as a support substrate due to its compatibility with SU-8 (as shown by the similar 8 and a values), its ability to maintain a high level of stability with organic solvents (as shown by the differences in 6 between PEEK, THF, and acetone), and its suitability for machining of orifices of different diameters and types.
Selection of a Suitable Sacrificial Substrate
Tests were conducted to determine if a semi-solid SU-8 film on a flexible sacrificial substrate could be successfully transferred to an SU-8 micro fluidic pattern, and the sacrificial substrate removed after curing the SU-8 film to seal the pattern. Several film substrates such as polyethylene, polytetrafluoroethylene, polycarbonate or printing film, and polyimide film were examined for their utility as a sacrificial substrate. Polyimide was found to be the most suitable film substrate, because liquid SU-8 100 adheres to the polyimide film uniformly, and could be peeled away after the SU-8 cured to a solid state without damaging the microfluidic pattern.
Optimization of the Exposure Dosage and Thickness of SU-8
To obtain a flexible semi-solid SU-8, the time required for solidification of SU-8 at different exposure dosages and its relation to the thickness of SU-8 coating were investigated.
Increasing the Thermal Stability of SU-8
To increase the thermal stability of the SU-8 microfluidic reactor, the SU-8 layer used in the sealing process was further exposed to UV-light at 300 mJ/cm2. A thermo gravimetric analysis of several 100 μm thick SU-8 films, prepared under varying conditions such as (i) exposure of the SU-8 at a standard UV-light dose of 480 mJ/cm2 after pre-baking at 65° C. for 20 min and 95° C. for 120 min, and (ii) exposure of the SU-8 at 300 mJ/cm2 (without any pre-baking), followed by post-baking at 65° C. for 20 min and 95° C. for 20 min, and exposure at 300 mJ/cm2, showed thermal stability beyond 205° C.
The micro-reactor was fabricated using PMMA (cast-type) and polyetheretherketone (PEEK) support substrates purchased from McMaster-Carr Supply Co., Atlanta, Ga. Other examples of polymeric support substrates which could be used to fabricate the micro-reactor include, SU-8, polypropylene, polyvinyl chloride, polycarbonate, and polyethylene. The patterning of microfluidic channels of SU-8, spin-coated on PMMA, PEEK and stainless steel substrates was carried out using a UV-light (220-450 nm, Model # 85110; Oriel Corporation, Stratford, Conn.). The UV-light dosage required to pattern the microfluidic channels varied up to about 1680 mJ/cm2 for a 500 um thick, soft-baked SU-8.
The multilayer, embedded SU-8 structure was then sealed using the flexible semi-solid transfer process. To achieve this, a 100 μm thick SU-8 film was coated on a sacrificial substrate 32 such as polyimide film (KAPTON®, Lanmar Inc., Northbrook, Ill.) and pre-exposed to a 300 mJ/cm2 dosage of UV-light (standard dosage is 480 mJ/cm2) without soft-baking to form a flexible, semi-solid layer 34 of SU-8 film, as shown in
Screws (not shown) were then threaded into the orifices to tightly position ferrules next to the SU-8 layer on the bottom of the orifices to prevent any leakage. Finally, multi-layer embedded SU-8 structure 36 was bonded to the bottom of the orifices in substrate 24 to prevent the microfluidic structure from cracking or debonding.
The inlet channels, as shown in
Table 2 shows the pressure drop of the micro-reactor estimated at different points in the reaction channels using the following equation:
See I. Simpson et al., Microfluidics: Applications in Chemical Processing and Analytical Science Proc. Imeche Micro and Nanotechnology (The Thermofluids Dimension, London, 1995). Here Q is flow rate, m3/s; Cfr is the friction coefficient for the rectangular cross section where the width w is bigger than the depth d; A is cross section area, m2; Dh is the hydraulic diameter calculated by the equivalent diameter of the same cross section area, m; L is length of the channel, m; and μ is viscosity of the feed, Pa·s. For a diluted THF solution, the viscosity and the diffusion coefficient may be treated as the pure THF solvent. The viscosity is 4.856×10−3 Pa·s at 25° C. At a combined flow rate of 760 μL/min, the micro-reactor had a pressure drop of 0.028 MPa, and a maximum pressure drop of 2.1 Mpa without any leakage.
aCapillary pressure drop for a rectangular cross section was used to estimate pressure. The diffusion coefficient of THF was 5.0 × 10−9 m2 · s−1. The viscosity of THF was 4.856 × 10−3 Pa · s at 25° C.
bRetention time was calculated based on the average flow rate.
To determine the long-term stability of the micro-reactor shown in
To demonstrate the effectiveness of the micro-reactor to synthesize nanoparticles, comparative tests were conducted using both a conventional batch process and a continuous flow polymeric micro-reactor to synthesize palladium nanoparticles. Palladium nanoparticles were first synthesized with the conventional batch process by reducing PdCl2 in THF (99.9% pure packaged under nitrogen) using lithium hydrotriethyl borate (LiBH(C2H5)3) as a reducing agent in the presence of 3-(N,N-dimethyldodecylammonium)-propanesulfonate (SB12) by modifying the wet chemical process in the reaction shown below. See H. Bonnemann, et al., “Nanoscale colloidal metals and alloys stabilized by solvents and surfactants: Preparation and use as catalyst precursors,” J. Org. Metal. Chem., vol. 520, pp. 143-162 (1996).
The reaction was conducted under inert atmospheric conditions using the Schlenk technique. The Schlenk technique is used to perform reactions under inert atmospheric conditions. PdCl2 (0.354 g; 2 mmol) was removed using a 250 mL triple-neck R.B. flask equipped with a flow control inlet adapter, with the flask evacuated completely. Next, the flask was filled with nitrogen and evacuated three times to remove oxygen. Next, 50 mL of THF was added to the reaction flask under nitrogen and the contents stirred magnetically. In a similar fashion, SB12 (0.67 g, 2 mmol) was dissolved under sonication in a 50 mL THF solution containing 4 mmol of lithium hydrotriethyl borate, and then added to a PdCl2 THF solution drop-wise. Afterwards, the reactants were stirred for an additional 30 min to complete reaction, and then 5 mL of acetone (reagent anhydrous, water<0.5%, 99.9+%) added to destroy reducing agent excess. Next, a 100 mL solution of ethanol (reagent anhydrous, water<0.003%) was added to the reactants, and the Pd nanoparticles were allowed to settle down. Next, supernatant was removed and the particles washed three times using a solution of 50 mL 1:1 volume ratio of THF:ethanol mixture to remove surfactant and other impurities, such as lithium salts. The nanoparticles were then dried using N2 to obtain a fine black powdery substance. (All of the above-mentioned chemicals were purchased from the Aldrich Chemical Company, Milwaukee, Wis., and used without further purification.)
Palladium nanoparticles were then synthesized with the polymeric micro-reactor from Example 1. Reactant reservoirs 40 (PdCl2 in THF) and 42 (LiBEt3H in THF) and nanoparticle solution collector 44, as shown in
Without wishing to be bound by this theory, it appears that the molar ratio of PdCl2/SB12 surfactant also affected the size of the palladium nanoparticles prepared within the microfluidic channels, as shown in
adn: the most probability diameter; n: the number mean diameter;
bFWHM01, the full-width at half maximum intensity for the most intensive peak;
c{overscore (a)}i: the mean lattice parameter calculated from SAED patterns.
An electron diffraction image analysis showed that Pd nanoparticles obtained from the micro-reactor were face-centered cubic (fcc) crystals with a lattice parameter of 4.217 Å, similar to those obtained from the conventional batch process. Compared with the bulk Pd foil, the lattice constants in sulfobetaine-stabilized Pd nanoparticles increased due to the nano size effect. The x-ray diffraction (XRD) pattern of Pd nanoparticles, as shown in
The peak widths for Pd nanoparticles from the micro-reactor resulted from the crystal size and micro strain effect caused by hydrodynamic forces occurring along the flow orientation in the micro-reactor. These results indicate that there was a slow growth of nanoparticles after a sudden formation of clusters from the saturation solution in the micro-reactor followed by the prevention of Ostwald ripening (i.e., the growth of large crystals from those of smaller size) through decomposition of the excess reducing agent which lead to nearly mono disperse nanoparticles.
The complete disclosures of all references cited in this specification are hereby incorporated by reference. Also incorporated by reference is the following publication of the inventors' own work: Y. Song et al., “Synthesis of Palladium Nanoparticles Using a Continuous Flow Polymeric Micro-reactor,” in Proceedings of the ICCE-10 Symposium on Composites/Nano Engineering, pp. 687-688, held in New Orleans, La. on Jul. 20, 2003; Y. Song et al., “Fabrication of a SU-8 Based Micro Fluidic Reactor on a PEEK Substrate Sealed by a ‘Flexible Semi-solid Transfer’(FST) Process,” J. Micromech. Microeng., vol. 14, pp. 932-940 (2004), and Y. Song et al., “Synthesis of Palladium Nanoparticles Using a Continuous Flow Polymeric Micro Reactor,” Nanosci. and Nanotech., vol. 4, No. 7, pp. 1-6 (2004). In the event of an otherwise irreconcilable conflict, however, the present specification shall control.
The development of this invention was partially funded by the Government under grant no. NSF/LEQSF (2001-04) RII-03 from the United States National Science Foundation. The Government has certain rights in this invention.