MEMS (MicroElectroMechanical Systems) have been used by the medical industry since the early 1980s and have continued to add applications and revenue throughout the past decade. According to a forecast by the Nexus Task Force, the market for bio-medical MEMS will reach $18 billion in 2005.

The two dominant MEMS biomedical technology are micromachined pressure sensors and accelerometers. But highly-accurate chemical sensors and custom microstructure applications are being widely adopted.

Millions of micromachined pressure sensors wind up in medical equipment each year in applications to monitor blood pressure, vital signs, blood analysis, and diffusion pump drug delivery systems.

They are also used in a growing number of newer applications. Surgeons will soon be using forceps, retractors and drills with embedded pressure sensors that provide real-time information about the medical procedure. One of the cutting edge technologies in development are so-called "data knives," that can be used for procedures in which precision and safety are paramount such as fetal surgery or brain surgery.

Pressure sensors embedded at the tip of the smart knife would, for example, make brain operations safer by telling the surgeon when the knife edge is near a blood vessel. In addition to the pressure sensors, data knives might also integrate a cauterizer, sensing/stimulating electrodes, ultrasonic cutting elements and strain sensors along the scapel's length. Verimetra, a Pittsburgh, PA, based company filed a patent for the devices in 2003.

Most pressure sensors are created using a micro-machined Wheatstone bridge piezo-resistive technology. These devices measure a wide range for pressures down to less than 0.1 psi. The pressure sensing element integrates an etched diaphragm structure with piezo-resistors. The diaphragm's thickness determines the range of pressures for the application. The near perfect elasticity of silicon is the critical reason for the high accuracy of these sensors.

As the diaphragm moves under pressure, stress is concentrated in specific areas of the silicon. The piezo resistors that have been ion implanted in these areas change value as they are stretched or compressed and measurements of the resistive values are directly related to the pressure being measured. The MEMS designer uses the location and orientation of the resistors to define a very precise relationship between the change in resistive value and the pressure that causes deformation of the diaphragm.

Since precise blood pressure monitoring is required in intensive care units, ICUs are a perfect venue of this technology. In one version, the sensor is part of a device that includes a bag of saline solution and its associated tubing and needle. Fluid passes from the bag into the patient. With every beat of the patient's heart, a pressure wave moves along the fluid path and is detected by the sensor.

Pressure sensors are also used during eye surgery as part of a system that removes debris inside the eye. During the retinal procedure, fluid is continuously removed from the eye and replaced after being cleaned. MEMS-based pressure sensors are used to control the vacuum that removes fluid from the eye. Another sensor in the vacuum control loop measures barometric pressure and feeds it back to the pump's electronics.

A third distinct application type for pressure sensors is infusion pumps and drug delivery systems. When a blockage prevents fluid from moving down the tube, the sensing element detects a pressure spike and sounds an alarm. These drug delivery systems can vary in their underlying technology.

One approach is to place diaphragm in the fluid path and interface this diaphragm with a pressure sensor that has a matching diaphragm. Pressure spikes in the plastic tube that delivers the medication increase the pressure across the diaphragm and the sensor records the change. Another method employs a strain gauge connected to a flexible beam. The plastic tube rests on the beam and moves as pressure increases.

These approaches to drug delivery have been in use for quite some time. A newer technology delivers medication by relying on diffusion through silicon nanomembranes using a process that can produce pores that are 20 to 100 nanometers wide. These devices are typically implanted under the skin and incorporate sensors as well to maintain a therapeutic concentration of the drug.

The complete system is frequently called a smart pill and incorporates a sensor array—usually a pressure sensor—electronics, a battery and a drug reservoir as its major components. These devices are extremely small—7 mm—and can be configured as two capsules, one inside the other.

By rotating the inner capsule either the pressure sensor array that diagnoses the patient is exposed, or, the portion of the device that releases the drug. They are used extensively for insulin infusion, administering morphine from inside the body to control extreme pain, and cancer chemotherapy, where they are injected into the bloodstream, travel to tumor sites and deliver a therapeutic dose of radiation.

Accelerometers measure frequency and amplitude of vibrations. They can also determine title angle in steady-state conditions and the amplitude of a single shock or pulse. When accelerometers are implemented as MEMS, the have uses indicators of a patient's activity while sleeping—important for treatment of sleep apnea—as well as for heart pacemakers.

Just like pressure sensors, accelerometers take advantage of piezoresistive effects and the elastic characteristics of silicon. A sensing mass is supported by beams in which the piezoresistors are implanted. When a vibration or rotation occurs, the mass displaces which activates a change in the resistors.

The sensor provides constant output that can be directly related to the position of the sensing mass. The operating range can extend from below 0.1 g to over 500 g and the devices offer exceptional response for position and modal analysis.

The value of accelerometer to a person with a heart pacemaker is that it can adjust the pulse of the pacemaker to physical activity by reporting movement to a microprocessor. For patients suffering from sleep apnea—the heart stopping while asleep—between one and three accelerometers are placed on the patient and an alarm sounds if he does not move within a specified period of time.

An emerging MEMS technology that can sense chemical and biological properties on a platform the size of a chip typically combines silicon fabrication technology with microfluidics. Not too surprisingly, microfluidics owes much of its early progress to the development of ink-jet printer technology. A simple—but useful—definition of microfluidics is the device has one or more channels with at least one dimension less than 1 mm. Lab-on-a-chip products can evaluate any number of chemicals but the most talked about topic is DNA analysis.

Labs-on-a-chip technology has several advantages: fluid volume inside the channels is very small (usually on the order of nanoliters); the amount of reagents and analytes used is quite small; and fabrication techniques are relatively inexpensive and adaptable to mass production.

In a manner similar to that for microelectronics, microfluidic technologies enable the fabrication of highly integrated devices for performing several different functions on the same substrate chip. One of the long term goals in the field of microfluidics is to create integrated, portable clinical diagnostic devices for home and bedside use, thereby eliminating time consuming laboratory analysis procedures.

There are two common techniques to force fluid through microchannels. In pressure driven flow, in which the fluid is pumped through the device via positive displacement pumps, such as syringe pumps. Electrokinetic flow is more involved and depends on a phenomenon called electroφsmotic pumping.

If a microchannel's walls have an electric charge—as they generally do—then an electric double layer of counter ions will form at the walls. When an electric field is applied across the channel, the ions in the double layer move towards the electrode of opposite polarity. This causes the fluid near the walls to start moving and this transfers into convective motion of the bulk fluid. If the channel is open at the electrodes, the velocity profile is uniform across the entire width of the channel. However, if the electric field is applied across a closed channel (or a backpressure exists that just counters that produced by the pump), a recirculation pattern forms in which fluid along the center of the channel moves in a direction opposite to that at the walls. By modifying the electric field, the motion of the fluid can be controlled.

Although there are several ways to separate the elements of the sample, one that is illustrative in principle is the H filter, which was developed at the University of Washington in the mid 1990s. The H-filter (Figure 1) allows continuous extraction of molecular analytes from fluids containing interfering particles—blood cells, bacteria, microorganisms, dust, and viruses—for example, without a membrane filter or some other component that requires cleaning or replacement.

Figure 1:  The H-filter allows continuous extraction of molecular analytes from fluids containing interfering particles.

This device is designed to process the small fluid volumes (nanoliters to microliters) analyzed by microfluidic instruments, although it can be scaled up to process fluids at arbitrarily high flow rates. Products based on H-filter separatation tend to be inexpensive enough to be disposable. Biomedical applications include:

• Upstream in-line blood plasma separator (replaces centrifuges)
• Separator, reactor and extractor for bio-pharmaceutical and drug-discovery
• Artificial kidney
• Genomic analysis systems, sequencers and mass spectrometry

After the molecules are isolated by the filter, the must be analyzed. The T sensor is one option that, once again, is illustrative of the general process.

In a typical T-sensor assay one input stream contains an analyte of interest, such as a protein or a drug. The other fluid stream contains a receptor molecule such as a fluorescent indicator, an antibody, a pH indicator, an enzyme, or some other reactive species. The flow of the two streams is kept completely laminar and no convective mixing occurs.

The only way molecules in opposite streams can mix is by molecular diffusion across the interface of the two fluid streams. The chemical binding or other reaction events than occur along this centerline produce a measurable signal, usually fluorescence, which can be used to calculate a parameter of interest for the analyte, such as concentration or diffusion coefficient. A diagram of a T sensor is shown in Figure 2.

Figure 2:  T sensors use parallel microfluidic channels to identify molecules

T sensors and H filters have complementary attributes for product design with the most important being that they are relatively inexpensive to manufacture and therefore can be used in disposable products. T sensors have been used in the following applications.

• Clinical chemistry analyzer
• Drug and hormone analyzer
• Enzyme and protein analyzer
• Chemistry/coagulation system

A variety of materials can be used to make microfluidic devices but the process itself is very similar to chip making. A photoresist (positive or negative) is spun onto and silicon substrate. The photoresist is exposed to UV light through a high-resolution mask with the desired device patterns.

After washing off the excess photoresist, the silicon wafer is placed in a wet chemical etching bath that etches the silicon in locations not protected by photoresist. The result is a silicon wafer in which microchannels are etched. Often, a glass coverslip is used to fully enclose the channels and holes are drilled in the glass to allow fluidic access.

At least one very large semiconductor company has jumped into the bioMEMS space in a big way. ST Microelectronics has marketing MEMS products for automobiles for years. Now it has leveraged its research and world-class silicon manufacturing facilities in a new line of bioMEMS.

Its first product is a lab-on-a-chip specifically designed for DNA research. DNA analysis chips are used to diagnose genetic diseases, perform drug discovery, test livestock and monitor water supplies. The InCheck™ platform (see Figure 3) integrates two fundamental steps of genetic analysis: amplification and detection.

Figure 3:  The InCheck Lab-on-a-Chip speeds DNA analysis

Since blood samples don't have enough DNA for analysis, scientists "amplify," or copy the DNA target many times using a process called polymerase chain reaction (PCR).

Using InCheck allows scientists to perform PCR in 15 minutes instead of other techniques that might take a day or two. The bioMEMS approach also shrinks the equipment in size because external circuitry driving the chip can fit in a 20 x 20 x 20 cm box. This makes suitable for point-of-care applications.

A prepared DNA sample flows into the channels in the chip and is repeatedly cycled through three temperatures. The quantity of DNA is doubled in each cycle. The amplified sample then flows into a detection area on the same chip where gold electrodes have been preloaded with DNA fragments. Fragments attach to matching fragments on the electrodes and are detected optically.

Beginning with relatively simple pressure sensors and accelerometers about 20 years ago, MEMS technology has consistently added new complementary technologies such as micorfludics to address wider and wider markets. In the future, we can expect to see semiconductor manufacturing technology adopted more and more for bio-medical purposes that far exceed its present information technology uses.