It is not justified to downscale a tool, because small items are appealing, or because  one has access to micro fabrication facility. While micro fabrication   proved to  be immensely  useful in electronics and other fields, its success  in microfluidics has been limited, partly due to misidentification  of the principal scaling parameters, and  due to an ill defined purpose. Let us consider  downscaling of a simple flow system , that will meter and mix sample with a reagent , will allow the mixture to incubate for defined period of time, and will facilitate monitoring of a resulting  product that will represent   the content of the target species. 

Miniaturization of Flow Analysis
microTAS & Lab-on-chip

Since the purpose of miniaturization is to reduce sample and reagent consumption, then the parameters  to be considered are 1) volume of the flow channel, 2)` the way in which sample will be mixed with reagent 3) incubation (residence) time available for chemical reactions to proceed.. The concept of lab on chip, also called microTAS,  is based on assumption that mixing by  radial mass transfer can be efficiently accomplished by diffusion (Section 0.2.6.). Therefore the  maximum internal diameter of the flow channel must be 0.1mm or less. To provide incubation time of 50 seconds at a flow velocity  of 1mm/second, the channel  should be at least 5cm long, requiring a pressure of 15 bars to propel the solution at a constant, continuous flow rate. Combined sample and regent volumes  will be in the range  0.5 to 1 microliter (microTAS). While fabrication of such a device is feasible, its use for  real life  assays faces numerous  obstacles. The main issue is that the ratio of surface (of the conduit walls) to volume of the transported liquid, is far higher than in conventional flow systems, which has an unwanted effect of enhanced adsorption of species from processed samples. Another challenge is devising pumps, that would generate the low flow at a constant flow rate, and detectors suitable for monitoring target species in sub microliter volumes, moving trough 100 micron channel.   While monitoring fluorescence or electrochemical responses is feasible,  the  sensitivity of UV-VIS spectrophotometry that  is used in about 60% of all reagent based assays, is severely restricted  due to shortness of the light path. While elecrophoretic drive and electrokinewtic sample injection ( above) solves mechanical problems of pumping and injection, it severely limits the range of potential applications. In this context it is revealing to read critical and honest evaluation of microTAS technology by one of its inventors. (Mukhopadhyay 2009)

R. Mukhopadhyay Anal. Chem. 81, 4169, 2009.

Small is beautiful.
E.F. Schumacher.