PROFILE

Microfabrication at Micralyne: evolution of MEMS and microfluidics from exploration to commercialisation

Microfabrication at Micralyne: evolution of MEMS and microfluidics from exploration to commercialisation

---

Company background

Micralyne is a growing, profitable, private company. Using semiconductor microfabrication technology to design and manufacture miniaturised devices, it provides industry-leading solutions that enable the biochemistry and telecommunications revolutions.

Founded in 1982 as a unique collaboration between industry, government and university and originally called the Alberta Microelectronic Centre (AMC), the organisation was a not-for-profit institute located at the University of Alberta. After having enjoyed steady growth of commercial revenue throughout the 1990s, Micralyne became a private corporation in 1998. Since that time the company has grown its revenues steadily and has been consistently profitable, having becoming a rare example of successful privatisation of a not-for-profit research institute.

In 2000, Micralyne moved into a new 40 000 square foot, state of-the-art manufacturing/research facility in Edmonton, Alberta. This move has allowed for the space, equipment and personnel growth required to respond to the ever-increasing demands of customers for cutting-edge miniaturised products manufactured at production-level volumes. Now employing over 80 employees, Micralyne has amassed broad experience in moving microfabricated products from early development into manufacturing, and works extensively in strategic relationships with market-leading companies like Applied Biosystems, MDS Sciex, JDS Uniphase, Signature Bioscience and Lynx Therapeutics. It is Micralyne’s goal to become the premier manufacturing partner for those companies and institutions in the life-sciences marketplace that desire to make microfluidics and MEMS technology a vital component of their process. Micralyne’s researchers and engineers have become highly skilled at quickly drawing out marketable products from loosely defined R&D concepts. Prospective customers are drawn to Micralyne because it is one of the only companies in North America that provides R&D, prototyping and volume manufacturing of microfabricated parts under one roof. Many of its largest clients use Micralyne as their sole supplier.

Technical background

The leading edge of MEMS and microfluidics has progressed from a multitude of diverse, exploratory ideas being prototyped in the early 1990s to a smaller number of proven commercialised devices and an even greater number of exploratory devices to seed the next wave of commercial products. To a large degree, Micralyne experienced a similar progression in its focus: beginning a decade ago as an R&D centre fabricating small quantities of proof-of-principle devices for academic and government clients, and more recently operating as a privatised manufacturer of tens of thousands of OEM devices for industrial clients.

Of Micralyne’s ∼85 employees, some 30 are directly involved in microdevice R&D. Over the last 10 years, senior research staff such as Graham H. McKinnon (VP, R&D), Yuebin B. Ning, Glen A. Fitzpatrick, James N. Broughton, R. Niall Tait, Christopher J. Backhouse and H. John Crabtree have been crucial to internal and client projects, and have published in a variety of journals relating to MEMS and microfluidics.

Since the inception of MEMS and microfluidics, Micralyne has fabricated and often co-developed devices with the pioneers of these fields. In more recent years, an emphasis on larger-scale manufacturability has come to the fore. This experience has helped develop both volume microdevice manufacturing and applications expertise.

Microfabrication procedures for glass, quartz and silicon substrates that are tailored to large-scale manufacturing are essential to bring functional prototype devices forward as cost-effective, commercialisable products. Such procedures would include substrate cleaning, metal and semiconductor thin-film deposition (physical vapour deposition methods using sputtering, resistive heating or an electron beam; or chemical vapour deposition methods (CVD) such as plasma-enhanced CVD or PECVD), photolithography, thin film and bulk substrate etching (isotropic and anisotropic wet chemical etching, reactive ion etching or RIE), metal deposition, substrate bonding (fusion, anodic), hole drilling (ultrasonic milling, laser drilling), dicing, and packaging. Fig. 1 shows a substrate being loaded into one of the sputtering system in the clean-room. An extensive review of such procedures has been written by Micralyne researchers Ning and Fitzpatrick.¹ In addition, non-technical expertise and discipline such as proper manufacturing QA/QC are crucial to ensure that all manufactured devices in a lot meet their requirements. To this end, Micralyne adopted the ISO 9001 system, becoming certified early in 2001.

Fig. 1 Processing in the clean-room. A large substrate is loaded into a sputtering system.

Examples of MEMS and microfluidics devices and research

Some of the breadth of manufacturing and applications expertise gained in a decade of microfabrication can be represented by surveying a few of the devices that we have developed internally, co-developed, or fabricated for academic or corporate researchers.

An early MEMS device pioneered by Yan Loke and McKinnon (Micralyne) in collaboration with Mike Brett at the University of Alberta was a high-g accelerometer.² The design incorporated an array of silica cantilever beams ranging from 120 to 400 μm in length, some of which are shown in Fig. 2. Upon sufficient acceleration, a given beam deflected across the 2.5 μm gap and closed the electric circuit; the threshold acceleration could be determined by the number of beams that deflected. Electrostatic and centrifugal testing was performed on the devices and showed accurate acceleration measurement within a range of 400–20 000 g.

Fig. 2 High-g threshold accelerometer. (A) Silicon plate with SiO₂ cantilever beams, each 80 × 2.2 μm, and ranging in length in the image from 260 to 400 μm. Taken from ref. 2 with permission from Elsevier Science. (B) The cantilever beams in the lower Si plate in (A) deflect upwards to close the circuit with the metal contact in the upper Pyrex cover plate, given sufficient acceleration. Devices were tested from 400 to 20 000 g.

A silicon micromachined capacitive microphone³ was conceived by Ning, Mitchell and Tait (Micralyne). PECVD-deposited Si₃N₄ was used as the diaphragm and backplate, and amorphous Si was used as the sacrificial layer, freeing the diaphragm; a photo and schematic are shown in Fig. 3. The microphone was operated at a bias voltage of 6 V, and showed a flat frequency response from 100 Hz to 10 kHz, with a sensitivity of 7 mV Pa⁻¹.

Fig. 3 Micromachined silicon capacitive microphone. (A) View from above the microphone die. (B) Cross-section of the microphone showing several aluminium and silicon nitride layers.

Ferrofluidic valves⁴ have been developed by Herb Hartshorne, Ning and Backhouse (Micralyne) in collaboration with William Lee of Defence Research Establishment Suffield (Canada). In these valves, 9 mm, 5.5 kG permanent magnets are used to move ferrofluids into a channel to block it, or into a side reservoir, leaving the channel open to fluid flow. The valves were tested to 0.4 atm above atmospheric pressure via water backpressure. An illustration of the valving mechanism is shown in Fig. 4.

Fig. 4 A ferrofluidic valve. (A) Open valve. Horizontal channel has triangular reservoir containing darker ferrofluid. Magnet pulls the ferrofluid down into the reservoir allowing aqueous fluid flow through the channel. (B) Closed valve. Magnet has drawn the ferrofluid up from the reservoir to block the thin horizontal fluidic channel.

Research pioneered by Bing He and Fred Regnier of Purdue University in collaboration with Tait (Micralyne) saw revolutionary column supports microfabricated inside the column channels,⁵ as shown in Fig. 5. The columns and supports were made from a silica substrate with reactive ion etching to a depth of 10 μm. The micromachined column support structure circumvents bead-packing irregularities and allows for uniform convective flow.

Fig. 5 Microfabricated chromatographic supports and columns. Smallest silica support pillars are 5 × 5 × 10 μm, with 1.5 μm channels in between. Taken from ref. 5 with permission from the American Chemical Society.

Another example of chip-based chromatography⁶ comes from Richard Oleschuk and D. Jed Harrison at the University of Alberta, in collaboration with Ning (Micralyne). In this instance, weirs were microfabricated on the glass chip to trap beads to form the separation bed. It was found that packings could be inserted, stabilised and removed at will. The bead chamber and filling procedure are illustrated in Fig. 6. The device was used for 500× sample preconcentration and CEC separations.

Fig. 6 Beads trapped by weirs to form a separation bed. (A) SEM photo of the empty bead chamber. (B) Photo of chamber being filled with beads. Taken from ref. 6with permission from the American Chemical Society.

Micralyne has introduced a new product combination to expedite the researcher’s entry into the field of microfluidics research: Standard Microfluidic Chips and the Microfluidic Tool Kit, or μTK, an instrument support platform, both illustrated in Fig. 7. The μTK consists of high voltage power supplies and laser-induced fluorescence detection in a user-configurable system, controlled by single graphical user interface. The system can be expanded up to eight independent 6 kV power supplies and has two fluorescence excitation options, red and green.

Fig. 7 Microfluidic chips and instrumentation. (A) Glass chip with two channels, 8 and 85 mm long, intersecting midway along the short channel, 5 mm from one end of the long channel. Channels are 20 μm deep and 50 μm wide. (B) Photo of Microfluidic Tool Kit, consisting of 6 kV power supplies, LIF detection, data acquisition hardware and control software.

In collaboration with the Backhouse group at the University of Alberta, Micralyne has performed preliminary research using Standard Chips and the μTK. The first study, headed by Crabtree (Micralyne), was aimed at elucidating the source of separation irregularities observed while performing routine CE experiments on the chips.⁷ The particular irregularities were baseline and migration time drift observed after multiple repeat injections. It was found that the source of the irregularities was pressure-driven backflow from certain wells on the chip, and that the pressure-driven flow arose from the well menisci (e.g. Laplace pressure); siphoning was found to be a relatively insignificant factor, under the conditions studied.

The second study, headed by Footz in the Backhouse lab, focussed on a new method to inject/isolate the product of a PCR reaction from its template.⁸ In essence, both DNA fragments are electrophoretically-driven across the short channel in a gel-filled simple cross Standard Chip (forward-loading), the potential is reversed, and the trailing larger product can be isolated during the reverse-loading at the channel intersection. Serial plug injections from the intersection down the long channel of the chip through the course of the forward- and reverse-loading stages conclusively showed that primer, primer + product, and finally product, were present at the intersection as predicted. These results are illustrated in Fig. 8.

Fig. 8 Molecular Gating. (a) Successive injections at the channel intersection taken at 10 s intervals during forward-loading of the short channel (LIF detection 13 mm downstream of the intersection in the long channel). After 10 s, neither primer nor product had reached the intersection; after 20 s, primer had reached the intersection; and after 30 s both were present in the injected plug. (b) Successive injections at the channel intersection taken at 3 s intervals during reverse-loading of the short channel (detection as in A). After 9 s, the injected plug sees a dramatic reduction in the faster (shorter) primer, and after 15 s of driving the sample back to the sample well only the desired PCR product is present in the injected plug. The inset graph portrays this electropherogram clearly. (c) A schematic to describe the events of A and B: the left-most injection corresponds to 20 s into forward-loading, the middle injection corresponds to 30 s into forward-loading, and the right-most injection corresponds to 15 s into reverse-loading. Taken from ref. 8 with permission from Wiley–VCH.

In its relationships with corporate clients, Micralyne strives to aid the client with the development of their microfabricated product, from earliest prototypes through to manufacture. Participating at the conceptual stages ensures that such a product is designed to meet its performance requirements with manufacturability and cost-effectiveness in mind.

Applied Biosystems, a leading manufacturer of DNA sequencers and other genetic analysis equipment and reagents, has a development programme with Micralyne to develop next-generation microfluidic components for applications in genetic analysis and high-throughput screening. Some early research in this area by Sue Bay and co-workers at Applied Biosystems and Backhouse (at Micralyne) focussed on DNA sequencing in long channels.⁹ 50 cm channel glass plates with 48–188 microchannels per plate were developed for this purpose. Using POP-6^™ polymer and BigDye^™ sequencing standard, read lengths of 640 bases were achieved with 98% accuracy on this device, equal to the performance in a fused-silica capillary of similar cross-sectional area.

Sydney Brenner and co-workers at Lynx Therapeutics have developed a technique¹⁰ termed Massively Parallel Signature Sequencing^™, which uses batch processing of a microfabricated flow-cell that houses one million beads to determine the sequence of the DNA attached to each bead. The 5 μm microbeads are arranged in an array that is quasi-random, but that is fixed (by the microstructures of the flow cell) and imaged repeatedly during the course of the fluidic processing. No gel-based fragment sizing is required to elucidate the sequence. Fig. 9 illustrates a flow cell and bead array contained therein.

Fig. 9 Flow cell for Massively Parallel Signature Sequencing^™. (A) Longitudinal cross-section shows the dam against which the beads are trapped, and reagent inlet and outlet ports. (B) Close-up top view of bead chambers. (C) Lateral cross-section of bead chambers. (D) Photo of quasi-random array of beads. Taken from ref. 10 with permission from Nature Publishing Group.

Challenges on the horizon

Recent completion of the human genome project and the advent of combinatorial chemistry have continued to place ever-increasing demands on analytical chemistry methodologies and increased sample throughput. Given the advantages of micromachined microfluidic systems, motivation has focused on smaller devices, capable of handling numerous samples at one time. Primary motivation is being focused on ‘higher throughput—lower cost’ in an effort to make such devices more common-place.

The long-term benefits of microfabricated analytical devices are numerous. For example, the drug discovery process is a long, labour- and cost-intensive process. Combinatorial chemistry has resulted in an explosion in the number of potential drug candidates being discovered. Multi-million compound libraries have been generated, and synthesis is estimated to be proceeding at the rate of over one million new compounds per year.¹¹ Hence, existing analytical techniques, such as high performance liquid chromatography, are quickly becoming the bottleneck in the identification and pursuit of potential lead compounds. One can imagine the consequences and frustrations associated with the realisation that a potential ‘wonder drug’ may be buried within a compound library, awaiting discovery. Ultra-high throughput drug screening has the potential to greatly accelerate the drug screening process, and minimise the time associated with the identification of potential leads. This would aid in the identification of potential drug candidates and facilitate the process of bringing them through screening, to clinical trials, and eventual FDA approval, such that they become available to patients on a shorter time-scale.

Several drug discovery and disease identification efforts are dependent upon on genetic analysis. Hence, microfluidic devices that are capable of increasing the speed and efficiency of genetic sequencing would also be of enormous benefit. An increase in the speed with which an illness can be identified and understood translates into a reduction in the time a patient must wait for the development of specific therapies, and a potential cure. Micralyne is poised to make significant contributions to the development of a wide variety of these devices, designed for analytical purposes anywhere in the analytical spectrum from fundamental research through drug discovery to bedside diagnostics. Moreover, these contributions can guide the realisation of such devices in terms of cost-effective fabrication, making them more amenable to widespread use.

However, some important challenges exist in the development and realisation of these devices. One such challenge is the creation of devices that are capable of performing multiple, complex fluid handling steps in series on a fully integrated device. Development of a device that is capable of taking samples though a series of sample preparation, analysis and detection steps would greatly increase the speed and efficiency of analysis. Reduced sample usage and elimination of numerous time-consuming, off-chip sample handling procedures are the promise of microfluidic devices, and may make certain medical assay protocols that use expensive reagents more widely available and affordable. Stacked against this advantage is the challenge of manipulating smaller sample sizes—being able to interface nanolitre-consuming chips with microlitre real-world samples.

Yet another challenge to face is the incorporation of several of these integrated devices on a single substrate, to achieve highly parallel sample throughput on a single device. Implicit is the need for automated systems that are capable of loading and running these devices, and for complex, multi-channel detection systems for data collection. While these challenges seem difficult, the μTAS community is well on its way to realising these goals.

Over the next five years we are likely to witness the development of devices that are capable of performing several complex fluid handling procedures, such as sample preparation and analysis. These devices will increase the speed and capacity for sample handling, and decrease the time associated with critical analysis such as drug screening and genetic analysis efforts.

Acknowledgments

The authors would like to thank Graham McKinnon, Glen Fitzpatrick, Yuebin Ning and Sandra Katz (Micralyne) for helpful discussions, and Kevin Corcoran (Lynx Therapeutics), Richard Oleschuk (Queen’s University), D. Jed Harrison (University of Alberta), Fred Regnier (Purdue University) and Chris Backhouse (University of Alberta) for kindly allowing us to show their work.

References

1. Y. Ning and G. Fitzpatrick, in Biochip Technology, ed. J. Cheng and L. Kricka, Harwood Academic Publishers, Philadelphia, 2001, pp. 17–38.
2. Y. Loke, G. H. McKinnon and M. J. Brett, Sensors Actuators A, 1991, 29, 235–240.
3. Y. B. Ning, A. W. Mitchell and R. N. Tait, Sensors Actuators A, 1996, 53, 237–242.
4. H. Hartshorne, Y. Ning, W. E. Lee and C. Backhouse, Development of Microfabricated Valves for uTAS, uTAS ′98, Banff, Canada, Kluwer Academic Publishers, 1998, 379–381.
5. B. He, N. Tait and F. Regnier, Anal. Chem., 1998, 70, 3790–3797.
6. R. D. Oleschuk, L. L. Shultz-Lockyear, Y. Ning and D. J. Harrison, Anal. Chem., 2000, 72, 585–590.
7. H. J. Crabtree, E. C. S. Cheong, D. A. Tilroe and C. J. Backhouse, Anal. Chem., 2001, 73, 4079–4086.
8. T. Footz, S. Wunsam, S. Kulak, H. J. Crabtree, D. M. Glerum and C. J. Backhouse, Electrophoresis, 2001, 22, 3868–3875.
9. C. Backhouse, M. Caamano, F. Oaks, E. Nordman, A. Carrillo, B. Johnson and S. Bay, Electrophoresis, 2000, 21, 150–156.
10. S. Brenner, M. Johnson, J. Bridgham, G. Golda, D. H. Lloyd, D. Johnson, S. Luo, S. McCurdy, M. Foy, M. Ewan, R. Roth, D. George, S. Eletr, G. Albrecht, E. Vermaas, S. R. Williams, K. Moon, T. Burcham, M. Pallas, R. B. DuBridge, J. Kirchner, K. Fearon, J. Mao and K. Corcoran, Nature Biotechnol., 2000, 18, 630–634 . Also, visit the Lynx website for more information on MPSS: http://www.lynxgen.com/wt/tert.php3?page_name=mpss.
11. C. Robbins-Roth, From Alchemy to IPO; the Business of Biotechnology, Perseus Publishing, Cambridge, MA, 2000.

H. John Crabtree *, Michael Finot , Jennifer J. Lukomskyj and Vicky Walker
Micralyne Inc., 1911–94th St., Edmonton, Alberta, Canada T6N 1E6.
E-mail: www.micralyne.com

---