Contributed by Frédéric Breussin, Yole Développement
Driven by impressive recent progress in automating the identification of genes and proteins, the microfluidics market is approaching ~$1.5 billion and is on a steady ~20% annual growth path going forward. But the highly diverse range of products in the sector means highly fragmented demand for processes and materials.
Major segments driving growth require products with entirely different requirements for everything down to the substrate materials. Point-of-care clinical diagnostics devices demand low cost, disposable cartridges in automated testing systems to make the tests affordable. The R&D market, in contrast, looks for very precise, very complex chips to quickly test one sample against thousands of targets at once, to save researchers’ time and replace large complex equipment to bring its big savings.
But the choice is not as simple as just low cost plastic vs. more precisely patterned glass. Production volume, application, type of patterning, and optical properties all impact material choice, for both cost and performance.
Market size is one consideration. Polymer costs a fraction of a cent per square centimeter, while glass costs $.02, but glass is actually the lower cost choice for all but high volume production. A device on glass can be prototyped and then directly scaled up to volume production on the same equipment, so manufacture of low to mid volumes is cheaper than with polymer. Injection molding of polymer devices requires first making a costly mold, so costs come down only when that can be amortized over high volumes. Those volume requirements also mean that injection molding is not practical for making prototypes, so that’s usually done using PDMS or some other cheap and convenient material, so some redesign will often be required to port the process to injection molding for production.
The desired features are another issue. Very fine features can be precisely etched on glass, for higher resolution than molded plastic. Glass is also a good insulator, making it easy to provide an electric potential across the chip if needed. On the other hand, polymer provides greater freedom of design for large reservoirs and wells, sometimes needed in great numbers across plates for testing thousands of targets, and they can be expensive to etch or sandblast in glass. Polymer is also good for integrating valves and pumps, as silicone film can be easily added between polymer layers for flexible features easily actuated by pressure, for clever solutions like those from Rheonix and Fluidigm.
Optical properties are still another consideration. Since most microfluidic devices depend on optical readouts, they need optical access, for which the transparency of glass is an advantage. Not only are many traditional polymers not transparent, but they may fluoresce at the same wavelengths as used for detection, so more expensive and hardto-process alternatives like COC may have to be used instead.
Silicon is most used by companies with semiconductor production experience. And it makes most sense to use the higher cost material when silicon’s active properties are needed to integrate the processing on the chip.
That means that almost half the current microfluidics market, by value, opts for polymer, as plastic is the preferred choice for low cost diagnostics and point of care devices. Polymer is also used for the majority of microfluidics for pharmaceutical research, which often require arrays of wells that are easier to make in polymer. Many of these devices were developed for diagnostics markets needing low cost disposables, but are currently still used mostly in research.
Silicon and glass each account for nearly equal parts of the remaining half of the market. Glass is used for the majority of analytic devices, particularly for sample preparation for mass spectrometry and high pressure liquid chromatography. Silicon dominates the drug delivery applications, which consist mostly of inhaler nozzles for asthma medication and micro insulin pumps. It’s also widely used in the small but developing market for sophisticated and reusable industrial and environmental sensors.
Others are developing other approaches as well. Nano imprint technologies may have potential, like Sony’s stamping/embossing replication technique based on that used to make CDs, though feature depth is limited. Yet another approach, taken by Sophion Bioscience, is to combine glass, silicon and polymer in one device, making the reservoirs in polymer, the channels in glass, and the controls in silicon, then assembling the separate parts into a complex multiplexing device for drug research.
Production technology on all these substrates is maturing, with more professional, automated, quality production facilities now becoming readily available. Typically microfluidics foundries, such as Dolomite, Micralyne and Microfluidic ChipShop, have worked mainly with R&D groups to do relatively small production runs. But now we see companies leveraging their equipment from volume production of MEMS to offer mass production of microfluidics as well, particularly on glass and silicon.
With the impressive progress coming out of biomedical research labs in identifying biomarkers—and in starting to make progress on the equally challenging issue of figuring out practical ways to apply this new knowledge to improve people’s lives—there looks like plenty of opportunity for both companies that target low cost, high volume markets, and those that target the high margin niche markets.