Flow Patterning - Eric Rice and Wesley Thompson

Motivation

Nanoscale structures have been studied extensively over the past few years.¹ Nanoscale structure design has become an important area of interest because novel physical and chemical properties are exhibited relative to the bulk material. Over the past few decades, researchers have thouroughly progressed in research areas which relate size dependence to electrical, optical, and magnetic properties of materials such as metals and polymers.¹ A well known fact about nanoscale structures is that these structures exhibit a high surface area to volume ratio. The creation of smaller structures results in an increase in structure density , resulting in enhanced function. For example, this translates into more accurate chemical sensors or faster microprocessors.

Let's reflect on a specific example: microprocessors in computing. A microprocessor contains many small transistors, or semiconductors which allow the transmission of electronic signals. These transistors are patterned on a silicon chip using a microfabrication technique, such as electron beam lithography. Over the past three decades, Intel has led the charge in developing next generation microprocessors.²

Figure 3. Scanning electron micrograph of nanoscale transistors.³

They found that smaller transistors allow for a higher power density per chip, resulting in an increase in performance.

• 

  Figure 1. Transistor size decreases over the past 30 years.²

• 

  Figure 2. Power density increases with the development of higher transistor density chips, over the past 30 years.²

Similarly, by increasing the number of features on a transistor, performance has proved to increase. However, when designing smaller transistors, the length scale of feature sizes must also decrease resulting in a number of difficulties.²

Figure 4. Increasing the number of features on a transistor is difficult while simultaneously decreasing the size of transistors.²

Microprocessing is only one of many attributed areas of major nanotechnology development. Other technological advances in nano and micro patterning are finding their way into a number of integrated scientific fields such as materials chemistry, biology, and engineering.² Not surprisingly, much of the interfacial and biological phenomena occur at the micro or nanoscale. These scientific fields have high potential to be significantly impacted by developments in nanostructure fabrication.

Patterning of Nanoscale Structures

From the creation of transistors to the patterning of single molecules, lithographic methods have advanced alongside the developments of novel materials. Both the properties of substrate and patterning molecule of interest highly influence which patterning methods are chosen. Patterning can be accomplished through the direct deposition or removal of a functional material onto/from a previously conformed substrate.² Deposition of a single atomic layer, or monolayer, or multiple layers of functional material is often accomplished through self assembly or directed assembly. This means either the chemistry of the molecule can determine how the pattern assembles or another technique must be used to deposit a pattern of the material. The pattern design, the choice of material we wish to pattern onto our substrate, and the properties of the substrate will determine the ideal patterning method. A few methods are commonly used for patterning, each with their own advantages and drawbacks.

Techniques	Advantages	Drawbacks
Nanoimprint Lithography 2	Great resolution High throughput Re-usable stamp	Difficulty in producing useful/accurate stamp
Marangoni Effect5	Topographical 3D patterning capability No device required for surface directed flow Diversity of patterns Unique structures	Requires controlled temperature, air environment UV exposure must cause a resultant change in polymer surface tension to obtain patterns
Conformal Coating 2	Wide range of curing techniques available (brushing,spraying,dipping) Large range of materials which can be dissolved in solvents and subsequently cured Potential for high throughput	Requires pre-conformed surface
Dip-Pen Nanolithography 2	Ink need not be reacted With sol-gel ink, wide range of materials can be patterned Good control over tip writing	Feature size determined by tip/surface meniscus size Direct writing by diffusion is impossible for many materials Impractical for high melting point materials Expensive equipment required Not scalable
Etching 2	Commonly used for layered applications Good control over z-direction depth based on etch time and etch rate Good resolution	May require multiple solvents and masks Multiple step process
Laminar Flow Coating 4	Convenient, rapid, and inexpensive one step process Ease of manipulation of biological interactions with surface chemistry Side-by-side gradient capability 3D structures can be exposed to flows Wettability of substrate can drive pattern formation	Not high throughput Limited to flat substrates

Due to the suitability for biological applications, flow coating will be the focus of continued discussion.

Flow Coating in Microfluidic Systems

The conception of laminar flow coating is attributed to Dr. Whitesides, in 1996.⁴ This technique is well known, but the number of published articles are not extensive, despite its value in a research setting. Perhaps limitations on scalability make the technique unattractive for many researchers. Otherwise, due to the non-intrusive method of introducing flow into a previously molded device, flow coating is highly relevant for fundamental biological and materials interface research applications. Microfluidic channel systems, in combination with laminar flow, can be used to pattern material from solution onto a surface. A number of accomplishments achieved by flow coating have been step concentration gradients using controlled diffusive mixing⁶, preparation of hydrogel gradients for observation of cell interactions⁷, and 3D growth environments for cells⁴.

• 

  Figure 5. Two streams combine into two a single channel. The interface of the two fluids is clearly visible.⁴

• 

  Figure 6. Inlet streams of green and red are fed into a series of diffusive zig-zag mixers to create a color gradient.⁶

• 

  Figure 7. Growth medium is patterned on a substrate and bovine capillary endothelial cells observed in the middle square while human bladder cancer cells are observed on the inner and outer squares.⁴

• 

  Figure 8. A gradient of concentration of RGDS peptide was decreased from left to right, creating a gradient of cell attachment on a hydrogel.⁷

Other great accomplishments, such as patterning single nanoparticles⁸, exposing a single cell to two spatial environments⁴, and patterning bacteria along surfaces are significant examples of the power of flow coating in biological applications.⁹

• 

  Figure 9. The interface of the two fluids coat a nanoparticle with fluorescent Rhodamine 6G on one side of the nanoparticle.⁸

• 

  Figure 10. A single erythrocyte is exposed to trypsin over the left side of the cell.⁴

• 

  Figure 11. Escherichia coli is patterned onto a substrate.⁹

Benefits

• Convenient, rapid, and inexpensive one step process
• Ease of manipulation of biological interactions with surface chemistry
• Side-by-side gradient capability
• 3D structures can be exposed to flows
• Wettability of substrate can drive pattern formation

Drawbacks

Figure 12. Lateral diffusion can be significant. Increasing flow rate and decreasing flow distance will decrease this diffusion across the fluid-fluid interface. ¹⁰

• Lateral diffusion¹⁰
• Not scalable by industry standards
• Methods confined to a relatively narrow range of materials
• Deposition of coatings or bacteria may collect within channels
• Not applicable for large curved substrates
• May require excess use of flow material

Fabrication

Figure 13. Schematic of the soft lithography process.

For laminar flow conditions, low Reynolds number <<2000 requires a small value of velocity, small diameter of the microfluidic channels, and high values of the viscosity and density of the fluid.⁴ Due to the small size ~50 um microns for the microfluidic channels, the range of fluid properties and velocity parameter choices is large.

The microfluidic device is fabricated through the use of photolithographic methods, where UV light is transmitted through a photomask, used to selectively crosslink photoresist in order to produce a master mold for the device. Poly(dimethylsiloxane) (PDMS) is then cured on the master mold and peeled off in order to fabricate the usable devices from the mold.

At this point, the applications of the device is considered. For example, if the device is used for human capillary research, what buffer is appropriate for flow through the device?

Material Considerations

For a process engineer, material considerations are the most important. Not all materials can be patterned inherently; some require pre-modification. In addition, the resolution at which a material can be patterned depends highly on material properties. Processes are designed with process parameters for specific resolutions, and modification to the size of features can likely lead to resolution issues, and significant reconsideration of a process must be taken. A number of questions must be asked in order to designate which materials are best suit for the task.

• What are the physical, chemical, thermal, and electrical requirements for the coating material itself?
• How fast does the material dry/coat? Are there time constraints for the patterning process?
• What temperature range will the device or material encounter?
• Can the material be modified or reworked after patterned?
• What is the purpose of the pattern and what weaknesses will the pattern exhibit in an applied environment?
• Wetting properties of the material?
• How will structure geometry influence the material properties?

The purpose of a pattern is typically sought. The geometry of patterned nanostructures is very important, and geometries seen in nature have been used to inspire novel patterns. In order to create an effective coating. A pattern is effective if generated with a functional material, and size, shape and geometry can be controlled throughout the fabrication process.

A number of patterned features are important for the study of biological systems such as isolated features, overlapping features/stripes, dots, lines, and particles. These features can provide an array of benefits and drawbacks, so it is important to pattern features which provide optimal functionality.

A few examples of patterned materials are:¹

• Metals, organic polymers, organic/inorganic crystals
• Nanoparticles and nanowires/nanorods
• Metal and metal oxide nanotubes
• Deposition of clusters and nanocrystals on graphite or other semiconductive materials to obtain novel 3D nanosystems
• Single and multi-walled carbon nanotubes
• Nanobiological systems (DNA, antibodies, microorganisms, enzymes etc.)

Depending on the design parameters of interest, the choice of material we wish to pattern onto our substrate, and the properties of the substrate itself, there are a variety of patterning approaches which must be evaluated.

2D and 3D Geometries

Repeating 2D and 3D geometries at the micro-scale have a wide range of uses. Patterned 2-dimensional structures have already been shown to have countless applications in areas such as tissue engineering,¹⁶ biosensors,¹¹ and diagnostic assay systems.¹² If 3-dimensional structures of such complexity and resolution to the 2D ones were available, then huge improvements could be made in crucial areas of research including photonics,¹³ data structure and storage,¹⁴ and tissue/ organ engineering.^15,16 A number of techniques for the fabrication of 3D structures via flow coating in some way are currently being researched and implemented. On a basic note, 3D structures made from flow coating are made via either a bottom-up or a top-down approach.¹⁷ Bottom-up methods including polymer phase separation,¹⁸ molecular self-assembly,¹⁹ or colloidal assembly²⁰ are inexpensive and can cover a large surface area but face limitations when it comes to the type and geometry of motifs that can be made.¹⁷ Top-down approaches on the other hand are capable of much more precise size and shape control of the structures¹⁷. The issue is that the point-by-point or layer-by-layer processing necessary for these methods makes approaches such as gray-scale photolithography,²¹ direct 3D writing,²² or two-photon lithography²³ are quite time consuming.

One large advance in the field of 3D structure fabrication via microfluidic flow is the idea of interference lithography (IL). IL allows for the rapid design and generation of intricate and densely packed features in 1D, 2D, and 3D without the loss of focus on the structures. IL also allows for control over geometric parameters such as symmetry and volume fraction.¹⁷ However, these phase-mask interference lithography (PMIL) methods offer a lot of restrictions in terms of shape and material properties because of how they are made. PMIL is typically done by flood exposing a spin-coated layer of photoresist film through a phase mask¹⁷, which just by the nature of the processing limits the throughput as well as requiring a mask and thus only being applicable to arrays of features and not arbitrarily shaped structures. However, aside from purely mechanical limitations, PMIL does not allow for the creation of chemically anisotropic structures. This refers to the presence of multiple regions of segregated chemical functionality on the same structure (See Janus particle figure above). A structure with precisely controlled material properties and texturing across a number of length scales is important in a variety of applications such as tissue engineering,²⁵ self-assembly,²⁴ and particle diagnostics.²¹

In order to allow for much more precise features, laser interference lithography (LIL) was developed. With the advancements in the modern laser over the last 20 years, we now have the ability to create basically any structure geometry. The optical interference of multiple beams at the same time has the ability to create structures on the nanoscale with near unlimited complexity. Optical interference with lasers is unique in that it allows for control over the phase, intensity, polarization and direction of the interfering beams. IL is essentially the use of two or more interfering light sources to create periodic regions of low and high intensity light which are then recorded in a photosensitive material. Using more laser beams only gives you more control over the interference patterns.²⁵

Figure 5: A representation of the periodicities created with different beam interference patterns. With 2 beams you can get 1-dimensional periodicity, with 3 beam you can get 2D periodicity, and with 4 beams you can achieve 3D periodicity.²⁵

LIL is used in many applications relating to optical devices on the scale of wavelengths of light. Photonic crystals for example are a common structure created via LIL. Photonic crystals change the motion of photons due to the induced periodic changes in the dielectric constant. Implementing a photonic band gap allows for the restriction of specific wavelengths of light or a range of wavelengths. A photonic crystal with a band gap is similar to a semiconductor with a band gap, it only allows certain frequencies of light to propagate through it much like the semiconductor only allows the propagation of specific electronic frequencies. 1-dimensional photonic crystals are readily available and have been for awhile. They have known applications in thin film optics such as dielectric mirrors and filters. 2-dimensional photonic crystals allows for the light to be localized at a surface or plane.²⁵ 2D crystals are used in such things as low energy lasers,²⁶ sensors,²⁷ and optical fibers.²⁸ 3-dimensional photonic crystals are a much newer development because they are the most sophiscated in that they allow for complete control over the band gap such that a given frequency of light can not propagate in any direction in the crystal. Structures can be placed in the crystal to guide light in the desired direction, and there is no way that it can go any other way. Metallic 3D photonic crystals can be used in thermal photovoltaic power generation cells.^29,30 Yablonovitch has proposed making a 3D photonic crystal where there was an overlap of the electronic and photonic gap as a way to reduce spontaneous emission and enhance the performance of lasers, heterojunctions, bipolar transistors, and solar cells.³¹ Such ideas as zero threshold lasers³² and superprisms³³ have also been proposed as possible to fabricate via LIL with many laser beams.³⁴

In order to combat the chemistry limitations however, several new microfluidic techniques that combine your typical lithography and photopolymerization with the unique flow properties present in a device with micrometer length scales have emerged.^34,35 Doyle and co-workers have developed a simple, flow-through microfluidic process known as stop-flow lithography (SFL).²⁷ This new technique allows for photolithography to be performed on a flowing stream of oligomer, and thus enables the high throughput synthesis of a large number of micro-scale particles in any 2D shape by using a variety of polymer precursors.²⁷ This method also allows for careful tuning of the structures' chemical anisotropy, however, the use of a transparency mask has restricted this method to the formation of solid 2D shapes with large feature sizes.³⁵

The most novel of discoveries in this field is the introduction of stop-flow interference lithography (SFIL) by J. H. Yang and co-workers. SFIL integrates the best of IL and SFL into one single technique for the synthesis of 3D patterns. SFIL "enables the high-throughput synthesis of three-dimensionally patterned, transparency-mask-defined polymeric particles with sub-micrometer feature size using liquid, biocompatible, oligomeric precursors". The breakthrough advantage of the increased throughput arises from the ability to repeated form and flush arrays of patterned particles in less than one second. The speed is possible because of the oxygen-induced inhibition of free-radical polymerization reactions at PDMS surfaces.²⁸ Material advances on the other hand result from the fact a free flowing oligomeric liquid of low viscosity can be patterned in a continuous fashion without the need for a step-by-step spin-coating, exposing and developing method as previously used. The properties of the structures that are being made can be tweaked with high accuracy by varying the intensity of the light as well as inhibitor and photoinitiator concentrations. Lastly, chemical anisotropy can be introduced in a structure if desired by taking advantage of the laminar co-flow of liquids that occurs in microfluidic devices as a direct result of the dominance of diffusive over convective transport.³⁵

Figure 6: Experimental setup and results for an example SFIL trial³⁵. a) A schematic of the experimental setup showing the PDMS device and integrated PDMS phase mask with a cross sectional view shown inset right of (a). Shutter-regulated and transparency-mask-defined light from either a Hg lamp or a pulsed laser enters up through the phase mask and into the stationary polymer film that is currently inside of the microfluidic device. This then results in the formation of an array of structures, each of which now has a 3D pattern defined by the given phase mask. The patterned particles are then flushed out, and the flow is immediately stopped before another subsequent array of particles can form³⁶. b) Brightfield image of an array of patterned triangles of side length 60 μm formed in a channel that was 600 μm wide by 30 μm tall. c) DIC image of the triangles shown in (b) after they have been suspended in a solution of ethanol. The patterned grid-like structure formed via the phase mask is visible on the surface of the particles and even more noticeable on the inset SEM image.²⁶

Future Directions

Developments in microfabrication, automation, microelectronics, and biology have already led to the creation of novel technologies for biological research. A few examples of these developments are nanoelectronics, integrated optics, biological sensors, and drug delivery systems. Integration of research advancements between fields are likely to compromise many technological advancements in the near future.

References

[1] Woodson, M.; Liu, J. Functional Nanostructures from Surface Chemistry Patterning. 2007, No. July 2006, 207–225 DOI: 10.1039/b610651j.
[2] Danowitz, A. Kelley, K. Mao, J. Stevenson, J. P. Horowitz, M. Recording Microprocessor History. Communications of the ACM, 2012 Vol. 55 No. 4, Pages 55-63
[3] Wu, Wen-li. "COMPLETED: Dimensional Metrology for Nanoscale Patterns." NIST: Material Measurement Laboratory. NIST, n.d. Web. <http://www.nist.gov/mml/msed/functional_polymer/dimensional-metrology.cfm>.
[4] Whitesides, G. M.; Ostuni, E.; Jiang, X.; Ingber, D. E. Soft Lithography in Biology. 2001.
[5] Katzenstein, J. M.; Janes, D. W.; Cushen, J. D.; Hira, N. B.; Mcgu, D. L.; Prisco, N. A.; Ellison, C. J. Patterning by Photochemically Directing the Marangoni E ff ect. 2012.
[6] Dertinger, S. K. W.; Chiu, D. T.; Jeon, N. L.; Whitesides, G. M. Generation of Gradients Having Complex Shapes Using Microfluidic Networks. 2001, 73 (6), 1240–1246.
[7] Khademhosseini, A.; Langer, R.; Borenstein, J.; Vacanti, J. P. Microscale technologies for tissue engineering and biology. PNAS 2006, 103 (8).
[8] J. H. Yang, D. Dendukuri, T. A. Hatton, E. L. Thomas, P. S. Doyle. Angew. Chem. Int. Ed. 2007, 46, 9027 –9031.
[9] Takayama, S. Patterning cells and their environments using multiple laminar. PNAS 1999, 96 (May), 5545–5548.
[10] Stone, H. A.; Stroock, A. D.; Ajdari, A. Engineering Flows in Small Devices. Annu. Rev. Fluid Mech. 2004, 36 (1), 381–411 DOI: 10.1146/annurev.fluid.36.050802.122124.
[11] R. S. Kane, S. Takayama, E. Ostuni, D. E. Ingber, G. M. Whitesides, Biomaterials 1999, 20, 2363.
[12] D. C. Pregibon, M. Toner, P. S. Doyle, Science 2007, 315, 1393
[13] Y. Lu, Y. D. Yin, Y. N. Xia, Adv. Mater. 2001, 13, 415.
[14] D. A. Parthenopoulos, P. M. Rentzepis, Science 1989, 245, 843.
[15] C. J. Bettinger, E. J. Weinberg, K. M. Kulig, J. P. Vacanti, Y. D. Wang, J. T. Borenstein, R. Langer, Adv. Mater. 2006, 18, 165.
[16] F. Gelain, D. Bottai, A. Vescovi, S. Zhang, PLoSONE 2006, 1, e119.
[17] E. L. Thomas, D. B. Alward, D. J. Kinning, D. C. Martin, D. L. Handlin, L. J. Fetters, Macromolecules 1986, 19, 2197.
[18] S. Zhang, Proc. Natl. Acad. Sci. USA 2003, 21, 1171.
[19] A. D. Dinsmore, A. G. Yodh, D. J. Pine, Phys. Rev. E 1995, 52, 4045.
[20] J. C. Galas, B. Belier, A. Aassime, J. Palomo, D. Bouville, J. Aubert, J. Vac. Sci. Technol. B 2004, 22, 1160.
[21] G. M. Gratson, M. J. Xu, J. A. Lewis, Nature 2004, 428, 386.
[22] S. Kawata, H. B. Sun, T. Tanaka, K. Takada, Nature 2001, 412, 697.
[23] S. C. Glotzer, M. J. Solomon, N. A. Kotov, AIChE J. 2004, 50, 2978
[24] Juhl, A. T. PhD. Dissertation. University of Illinois at Urbana-Champaign. 2010.
[25] Painter, O.; Lee, R. K.; Scherer, A.; Yariv, A.; O'Brien, J. D.; Dapkus, P. D.; Kim, I. Laser Science 1999, 284 (5421), 1819 -1821
[26] Baker, S. E.; Pocha, M. D.; Chang, A. S. P.; Sirbuly, D. J.; Cabrini, S.; Dhuey, S. D.; Bond, T. C.; Létant, S. E. Applied Physics Letters 2010, 113701-04.
[27] Russell, P., Photonic Crystal Fibers Science 2003, 299 (5605), 358-362.
[28] Lin, S. Y.; Moreno, J.; Fleming, J. G., Applied Physics Letters 2003, 83 (2), 380-382.
[29] Fleming, J. G.; Lin, S. Y.; El-Kady, I.; Biswas, R.; Ho, K. M. Nature 2002, 417, 52-55. 106
[30] Yablonovitch, E., Inhibited Spontaneous Emission in Solid-State Physics and Electronics. Physical Review Letters 1987, 58 (20), 2059.
[31] Yablonovitch, E., Inhibited Spontaneous Emission in Solid-State Physics and Electronics. Physical Review Letters 1987, 58, 2059-2062.
[32] Kosaka, H.; Kawashima, T.; Tomita, A.; Notomi, M.; Tamamura, T.; Sato, T.; Kawakami, S. Applied Physics Letters 1999, 74, 1370-1372.
[33] D. Dendukuri, D. C. Pregibon, J. Collins, T. A. Hatton, P. S. Doyle, Nat. Mater. 2006, 5, 365.
[34] J. W. Kim, A. S. Utada, A. Fernandez-Nieves, Z. B. Hu, D. A. Weitz, Angew. Chem. 2007, 119, 1851; Angew. Chem. Int. Ed. 2007, 46, 1819.
[35] D. Dendukuri, S. S. Gu, D. C. Pregibon, T. A. Hatton, P. S. Doyle, Lab Chip 2007, 7, 818.
[36] C. Decker, A. D. Jenkins, Macromolecules 1985, 18, 1241