Stanford Univ. developing MEMS for Medicine

At Stanford Univ., academic and industry researchers are developing MEMS for a variety of medical applications that includes preventative medicine, implantable devices, biocompatible coatings, and microfluidics. Both the Stanford Nanofabrication Facility, a research lab within the NSF-funded National Nanofabrication Users Network (NNUN) that is chartered to provide nanofabrication R&D resources to researchers in academia and industry, and the Center for Integrated Systems (CIS), a partnership program between industry and academia, are involved in the research.

Professor Greg Kovacs is both an MD and PhD, and leads a variety of biological MEMS development efforts. Kovacs is most interested in preventative medicine, such as detecting biological warfare agents and other toxic substances. Sensors for these substances can be made by growing a thin layer of cells on top of chips containing detectors. "It is not that difficult to grow cells on silicon," Kovacs said. In a laboratory incubator, such devices can live for months, but the researchers are working on creating self-contained units that can provide a nutrient broth at body temperature, as well as supply carbon dioxide and oxygen. Kovacs and student Derek DeBusschere believe they can make a self-contained unit that will keep the sensor operational for two weeks.

MEMS could also help improve pacemakers and defibrillators. "The basic function of pacemakers hasn't changed a lot in 20 years," Kovacs said. All pacemakers include a computer, some memory, and a telemetry link; additional sensors might provide more information to the user's doctor, such as blood oxygen measurements, pH level, or other information. People with pacemakers tend to be subject to other diseases as well, and a group of implanted sensors could provide early warnings about the patient's health. "Maybe someday we might implant these in healthy people," Kovacs said, "maybe people at risk, as well."

The engineering issues for such devices are substantial. In addition to requiring FDA approval, the devices must be small, energy efficient, and remarkably robust. They must have a multiyear lifetime, as the devices will be in a situation where they cannot be recalibrated. When you submerge something in blood, it must avoid being rejected by the body and resistant to corrosion.

Another focus of work at CIS is in developing biocompatible coatings for chips. Kovacs says there are two major kinds of coatings: bulk coatings such as silicone, and those deposited in thin layers by gas plasmas such as silicon carbide or Teflon^®.

Yet another area of research involves microscale fluid handling. Tiny pumps and tubes the diameter of a hair can be used with tiny volumes of fluids. Part of the work at Stanford includes developing sensors to measure the amount of organic carbon in a stream of water. The group has developed sensors that can measure as little as 500 ppt of heavy metal ions in water for applications that need very pure water.

Research Scientist Mary Tang, the biotechnology liaison at NNUN's Stanford facility, said developing microfluidics for biological applications is very important. "Biological systems are inherently 'wet' so the general bioMEMS research focus on microfluidic handling is pretty much unavoidable," Tang said (DNA is an exception, in that it can be dried and rehydrated and manipulated).

There are a number of difficulties in working with fluids in such tiny quantities: accurate metering is enormously difficult and laminar flow conditions mean fluids can generally mix only by diffusion. "It's a different world at this scale," Tang said.

Silicon micromachined pumps and valves have been shown to be effective in controlling fluids at the nanoscale level, but at the moment they are fairly complicated and expensive. However, many biological applications of MEMS are focused on potentially disposable devices.

Nevertheless, MEMS is an enabling technology for biologists who are finding a wide range of tools useful in their research. "Some of their devices are almost purely electronic, while others are astonishingly simple," Tang said "It's amazing what some biologists can accomplish with just a microchannel defined in glass."

Tang said other work involves miniaturizing tools down to micron sizes. Advantages of reducing size include minimizing the amount of the sample, maximizing purity, minimizing the time a process takes, and miniaturization enables scientists to manipulate and detect processes at the cell level, or even within cells.

Minimizing the sample size may not seem like a big dealmost laboratory methods use only microliters now, but it can be helpful if one must run a battery of tests.

The time savings can be significant. Tang said sequencing a sample of DNA to about 600 bases using the classic slab gel method takes about 4 to 6 hours; using a capillary method provides results with the same accuracy and precision in about 2 or 3 hours. "In theory," Tang said, "on a chip you could expect to get the same accuracy and precision in 20 to 30 minutes." At the moment, slightly less precision can be obtained on glass chips in about 20 minutes. In practice, both the slab gel and capillary methods are multiplexed users load up to 96 samples at once and process them in parallel. Microchannels on a chip could also be multiplexed.

Another potential time savings would result from integrating different components at a chip level. "Reading out sequences doesn't take as long as preparing the sample," Tang said. Ideally, this process would be automated on a micron scale.

Micro-scale techniques also enable screening in parallel. Both drug discovery and clinical applications can take advantage of screening. For drug discovery, a company may make tens of thousands, or millions, of variations on a theme, and test them all. In a clinical application, one sample could be screened for many genetic defects.