A Matlab algorithm developed for automatic layout generation of thin film heaters

(A major challenge in this technology was to make a heater
Th precise spatial control of temperature is a major challenge in this technology. To address this … I proposed…)

Spatial thermal control of localized fluid volumes is a major challenge in the lab-on-chip (LOC) field. Thin film heaters are commonly used in LOC devices to carry out many key functions that require high temperature uniformity, including the polymerase chain reaction (PCR) and melting curve analysis (MCA). Uniform heating, particularly, is difficult to achieve due to complex patterns of heat loss to a variable external environment. Moreover, the conventional heater design approach is to iterate from an initial best guess towards an optimal design, but without prior knowledge of the exact power distribution that is required to produce the desired temperature profile. Such an iterative process enormously increases the time and cost of the design process.

Recently I developed a method that enables rapid-turnaround design of system-customized thin film heaters for precise spatial temperature control in planar structures of any size and shape. Only a single layer of conductive material is necessary to produce the desired temperature profile. The method allows for maximizing energy efficiency and minimizing the heater footprint. Owing ot its non-iterative nature, the method is capable of generating an optimal heater design in few minutes automatically.

At the heart the method is an algorithm written in Matlab that is inspired in the finite element method. The algorithm divides the pre-determined power density field of the heater into small segments and for each segment computes the width of the track piece that meets the required power. Since the width is computed for each segment rather that for the entire geometry, the size and shape of the heater can be arbitrary. Since the segment size is free-to-choose, very high resolution power density fields can be achieved, only limited by fabrication constraints.

The method has been fully implemented in a monolithic polymer chip for genetic analysis and the produced heater layouts have been tested by simulation and experiment. The method applied in four different heater designs has been demonstrated to produce the expected results repeatably. Fast, highly selective and uniform heating within 1 ºC in >94% of the chamber volume was clearly demonstrated in the four designs at 95 ºC.


Fabricated and simulated heaters:


Close up of fab. chip next to simulation.