Microfluidics is basically tiny plumbing. Like so much of life, molecular biology is all about moving fluids from place to place; microfluidics lets one do this on a very small scale.

Microfluidic "chips" are slabs of PDMS (a transparent polymer similar to bathroom sealant) or glass, with microscopic grooves and indentations in them. These are bonded to a glass base, so that the grooves and indentations become pipes and chambers.

Typically, the channels in microfluidic chips are between a few microns and a few hundred microns across (a human hair, for comparison, is about a hundred microns across). At these small scales, liquids behave strangely. There is no turbulence, which means that two streams of liquid can flow alongside eachother in a channel without mixing, except by diffusion.

I use microfluidic devices to conduct rapid reactions (by adding reagents successively to a channel), and to observe the reaction products (for instance, using a microscope to detect fluorescence as molecules move past a given point in a channel). My main reason for getting into microfluidics is my single-molecule sequencing programme.













My group uses a method known as "soft photolithography" (mouse-over the blue text to see the images). The chip's pattern of channels is designed and printed as a high resolution transparency known as the mask.

Then, we take a silicon wafer, and pour on a few ml of photosensitive epoxy before spinning it at a precise speed in the spin coater which causes the epoxy to spread out to a uniform thickness. The epoxy is baked to harden it, and then exposed to UV light, through the mask using a mask aligner (we can also build up multiple layers if we want).

After another baking step, the wafer is rinsed with developer, which removes all of the epoxy except in those areas exposed to UV light - the clear areas of the mask. This leaves the pattern of channels as raised areas of epoxy.

Next, PDMS (a silicone rubber) is poured over the wafer. After about an hour at 60°C, the PDMS sets into a flexible solid, which can be peeled off and diced into individual chips. - the raised areas on the wafer are now shallow grooves in the PDMS.

Access ports are made by punching holes through the PDMS. Finally, the chip is bonded to a glass or quartz slide. This is done by exposing the PDMS to an oxygen plasma, to create reactive groups on the surface: the PDMS then bonds irreversibly to glass on contact. The finished chip is shown here from the bottom (glass) side, with blue dye to make the channels visible.












I set up the LMB's microfluidic facility in 2008, to enable the production of a wide range of microfluidic devices, primarily in PDMS.

Most fabrication facilities are in large cleanrooms under positive pressure, with scrupulously clean air and an airlock for entering and leaving. The LMB's facility differs from these substantially.

The shoebox-size room is kept under vaguely positive pressure by means of a high-volume HEPA-filtered fan, but has no airlock or staging area. We do not even wear the disposable "bunny suits" which are de rigeur in most fabrication facilities. Critical work is performed in a large laminar flow hood, with the air filtered through another HEPA filter and a charcoal filter to remove solvent vapour.

The advantage of this charmingly lax set-up is that the room is accessible without a 15-minute changing/cleaning routine. Most microfluidic fabrication does not, in fact, require the high levels of precaution used in most fab facilities.

Equipment includes a precision hot-block (modified thermocycler), Laurell WS400 Spin-coater, Neutronix Quintel mask aligner, GaLa PlasmaPrep2 plasma oven, and a cuckoo clock.