Flow boiling in microchannels and microchannel arrays is an effective way of achieving the high heat transfer rates required in compact evaporator designs. One objective in designing compact evaporators is to generate dry vapor at a temperature not exceeding the saturation temperature more than necessary, at the lowest possible surface temperature. Water is used as a model substance in this study, in which different evaporator designs are compared primarily with the aid of high-speed videography in the visible wavelength range.

While the evaporation in single microchannels has been the subject of intensive research over the last years, relatively little effort has been expended on the situation in arrays of many (several dozen) microchannels. For a practical application, however, pressure loss over the device is often an important concern, and for a commercially relevant throughput range, there is often no alternative to designs involving such large numbers of channels. Microchannel arrays with liquids undergoing flow boiling are well suited to the removal of heat from devices such as semiconductor chips with high surface power densities. For such heat removal applications, complete dryout of the microchannels is undesirable and can be highly detrimental. For evaporator applications, however, the situation is quite the opposite since complete dryout of the microchannels towards their downstream ends is what a good design should provide. This goal can always be achieved by making the microchannels sufficiently long. The requirement of minimized pressure loss mentioned above, however, also puts restraints on the length for the individual microchannels in such arrays.

In this paper, several microchannel arrays with different channel lengths and layouts are compared. To allow for comparison between the different array designs, several parameters were held constant. The arrays always consisted of 68 channels, each 200 μm wide and 100 μm deep separated by 100 μm wide ridges, and were made by mechanical micromachining of steel or brass, respectively. Array lengths varied between 20 mm and 80 mm.

The experimental set-up consists of a steel frame with electrically powered cartridge heaters, an exchangeable microstructure foil with the microchannel arrays, and a window of glass or Makrolon^™. Water was supplied by an HPLC pump at a constant flow rate of 0.5 kg h^-1, corresponding to a mass flux of 61 kg m^-2s^-1.The two phase flow phenomena can be observed through the window and microscope optics by a high-speed video camera operating at up to 1825 frames per second at a resolution of 1024×256 pixels. Illumination can be provided alternatively by halogen lamps with gooseneck light guides, or by a stroboscope lamp with a flash duration of approximately 10 μs.

Microchannel array evaporators for industrial applications will have to consist of metal, such as stainless steel. When such a device is connected to the (inlet) tubing via metallic connectors, heat conduction against the direction of water flow can cause boiling to start already in the inlet plenum. Once this happens, relatively large bubbles are formed in the inlet plenum, periodically breaking up at the microchannel array entrance. The pulsation resulting from these events is primarily responsible for the observed periodic breakdown of complete evaporation. There are two possible ways of mitigating this effect:

• Minimizing the size of the inlet plenum, and hence the residence time there, should move the boiling front downstream and into the microchannel array proper.
• Alternative multistage evaporator designs should cope better with a feed already containing some vapor.

Extended Abstract Status: Not Uploaded