One module in a bioagent detector currently under development involves a new two-heater, flow-through polymerase chain reaction (PCR) module which is being designed to save space and power and to reduce the amplification time. As in all PCR devices, thermal cycling requires three temperatures and residence times. These are 90–95°C for DNA denaturation, 50–65°C for hybridization and 72–77°C for replication with a time ratio of 4:9:4. The current design uses two heaters with heat conduction in the substrate providing the hybridization temperature. Typically, the flow and temperature fields in microfluidic devices have three-dimensional complexity, thus numerical simulations were performed to provide design guidelines in the development of the two-heater PCR device. The lattice Boltzmann (LB) method was used to perform low Reynolds number (typically Re = 0.10) simulations for two and three dimensional channel geometries having various wall temperature distributions. The momentum and thermal lattice Boltzmann equations were coupled via a body force term in the momentum equation. Initial computations using two- and three-heater configurations in two dimensions demonstrated excellent comparisons with published data provided that both the top and bottom walls were heated. If only one wall was heated, large vertical thermal gradients occurred resulting in non-uniform temperature fields. However, when the same conditions were applied to three dimensional channels, lower temperatures were observed in the center of the channel regardless of the wall temperatures or channel aspect ratio. Parametric studies were performed to evaluate the effects of thermal coupling, thermal diffusion coefficients, entrance temperatures, wall temperature configurations and channel geometry. If was found that moderate variation of the thermal diffusion coefficient produced only minor differences in the temperature field, and large changes in the thermal coupling magnitude demonstrated transition from natural to forced convection flows. The simulations also indicate that the largest effect on flow and temperature uniformity arises from the applied wall temperature distribution (various thickness channel walls). It was found, in 2D, that if the channel wall starts from ambient temperature, the applied heating, on the outer surfaces only, may not result in the desired wall or fluid temperatures. However, once the channel walls are heated to a uniform temperature, excellent temperature distributions are obtained for both thick and thin channel walls. These results indicate that the two-heater design has potential in providing a new flow-through PCR device. However, careful attention must be paid to the wall heater design to provide the required sample temperatures.
