Stanford Research Communication Program
  Home   Researchers Professionals  About
Archive by Major Area


Social Science

Natural Science

Archive by Year

Fall 1999 - Spring 2000

Fall 2000 - Summer 2001

Fall 2001 - Spring 2002

Fall 2002 - Summer 2003




Improving Microchannel Designs for Miniaturized Biological Analysis Instruments

Josh Molho
Mechanical Engineering
Stanford University
May 2001

During the past decade, the Human Genome Project, pharmaceutical research, and changes in the legal and healthcare systems have created a need for rapid and portable biological analyses. My research focuses on miniaturizing instruments that are used to identify biological molecules such as proteins and DNA. Typically, the bench-top sized versions of such devices employ a 50-100 centimeter (20-40 inch) long capillary tube as a "separation column." My research aims to design a more compact and efficient alternative to this long tube; an alternative that performs the same functions in a space the size of a credit card.

Starting in the early 1990s, the Human Genome Project sought to identify the thousands of genes comprising the "genetic blueprint" that describes the human species. During the same period, pharmaceutical companies performed around-the-clock testing of therapeutic drugs, the legal system became more accepting of DNA evidence, and new health care systems demanded cheaper clinical testing. These demands resulted in a need for faster and more portable biological analysis instruments. Surprisingly, the solution to these demands came from micro-fabrication techniques developed by computer chip manufacturers. By using these micro-fabrication techniques (e.g. photolithography and chemical etching), researchers were able to build tiny channels for the transport of liquids and testing of biological samples. These new devices are known as "labs-on-a-chip," or micro-Total-Analysis Systems (mu-TAS). Such systems integrate many analysis steps into one glass or plastic chip the size of a credit card. Previously, these steps were performed separately in large, bench-top sized instruments with human or robotic operators transporting samples between individual instruments.

Additional benefits of miniaturization include portability, a decrease in analysis times, and a reduction in the use of expensive chemicals. A typical mu-TAS is comprised of networks of channels that are 100 microns wide, 20 microns deep and a few centimeters long (a human hair is about 80 microns in diameter). A mu-TAS can perform many tasks, including sample preparation (e.g. removing desired molecules from a volume of blood), separation of the sample into its individual components (e.g. separating a mixture of five different proteins), and detection of the individual components. My research focuses on miniaturizing a common separation technique known as electrophoresis.

Electrophoresis can separate electrically charged molecules based on how fast they move in an applied electric field. An electrophoretic separation is often performed by first introducing a small amount of the unknown molecules into a liquid-filled, long channel or capillary called the "separation column." An electric field is then established along the length of the column by applying high voltage (usually a few thousand volts) to one end of the column while grounding the other end. When the electric field is applied, charged molecules in the sample begin to move; the speed at which a molecule moves depends on its charge and size. Since different molecules will move at different speeds, the initially mixed sample plug will separate into bands, each containing only one type of molecule. An optical, electrical or chemical sensor detects the arrival of these bands at the end of the separation column, and a molecule can then be identified by comparing its arrival time to that of a known standard.

In many cases of interest, the sensitivity of an electrophoretic separation increases as the length of the separation channel increases. Therefore, when miniaturizing electrophoresis, researchers have tried to use serpentine channels to fit a long separation length into a compact area. Unfortunately, turns in the separation channel reduce the effectiveness of the separation because of what is known as the "race-track effect." Imagine a turn that is a semi-circular arc. Molecules moving along the inside of the turn have a shorter distance to travel than molecules moving along the outside; just as the inside lanes on a circular jogging track are shorter than those on the outside. Thus, the molecules on the inside of the turn race ahead of the molecules on the outside, so that a turn tends to distort an initially straight band of molecules. To make matters worse, the electric field that is causing the motion of the molecules intensifies along the inside of the turn. Therefore, the molecules on the inside of the turn move faster in addition to having a shorter distance to travel.

Working with other researchers, I have designed new microchannel turns that allow all molecules entering the turn together to exit the turn at the same time. Our approach has been to mathematically model the turn dispersion problem, then design new turn geometries using computer simulation tools, and finally to build the new, "compensating" turns and test them experimentally in the laboratory.

Our first step was to mathematically model the race-track effect, including the effects of the different distances, field strength variations and molecular diffusion in the turn. Starting with a two-dimensional model, we mathematically averaged across the width of the channel, permitting us to derive an equation that quantifies the amount of dispersion created by circular-arc turns. This one-dimensional, mathematical model allows us to predict under what circumstances the dispersion caused by the turns will significantly reduce the effectiveness of the electrophoretic separation.

After studying the race-track effect for circular-arc turns, we used computer simulation tools to study more complex turn designs. We discovered that by varying the width of the channel in the turn, we were able to reduce the amount of sample dispersion. After some initial trial-and-error design, we employed computer-aided shape optimization to find the best turn design. The most successful design was predicted to create less than 1 percent of the dispersion that a circular-arc turn produces.

To test the turn designs experimentally, we used a 125-micron diameter cutting tool to machine channels into the surface of a credit card-sized plastic substrate. We sealed the channels by thermally bonding another piece of plastic on top of the substrate containing the machined channels. To measure the turn dispersion, we filled the channels with a special fluorescent water-dye mixture that allows us to optically mark the liquid with a laser beam. We tested each turn design by writing an initially straight line in the channel with a laser, just prior to the turn. After applying an electric field along the channel, we used a microscope and a digital camera to track the shape of the line as it entered and exited the turn. We then measured the stretching or dispersion of the line by examining the collected digital images. In general, we found that the computer simulations accurately predicted the dispersion that we measured in the experiments.

After successfully designing and testing a first set of compensating turns, we are now creating the next generation designs based on newer simulation results. These turn designs are an important building-block for the creation of effective miniaturized biological analysis instruments.