NEMS, smaller than MEMS, are nano-scaled systems having dimensions of 10-10 m (molecular size) that range upwards to 10-7 m or 0.1 to 100 nanometers. Studying nano-sized structures involves understanding the physics of building up molecules from atoms. A nano-sized process of all insects involves building up chitin, the major material of the exoskeleton, from chitin's components. NEMS employ novel materials such as carbon nano-tubes, quantum wires and quantum dots. Our 'nanotech revolution' involves problems of mass-producing nano-transistors and nano-diodes, nano-switches and nano-logic gates in order to construct nano-scale computers having terascale capabilities. Can we do it? Probably; but perhaps with a few changes in our thinking, insects might help.


Mechanical systems lack the similar modularity and topologi-cal simplicity of electronic circuits, so small mechanical devices are hard to manufacture. MEMS devices compared with larger machines, have fewer rigidly linked parts, and more are intrinsically compliant. As with chips, we can use planar lithographic processes to make masks for sequences in constructing MEMS.

Generally then, both NEMS and MEMS are micro-assemblies of parts having electronic and mechanical functions with NEMS being much smaller than MEMS. We house and integrate these latter assemblies on a single silicon chip.

A Generic MEMS System

To appreciate just how far we are from building bees, a current generic MEMS system might contain a micro-pump, a flow sensor, and an electronic control circuit. The pump delivers a variable flow into a micro-channel, and the circuit controls the rate of pumping. How do we build it?

First Simulate

To simulate a MEMS system on a computer, we first model each component of the system. Then we make a coupled liquid simulation, but disparities arise both at temporal and physical scales. Unlike an integrated circuit for which we have many programs that can test for errors in design, MEMS lack such verification tools. Having strong linkages between domains of energy in a device makes analysis difficult as computer models rapidly grow unwieldy and computationally prohibitive. So we simplify yet again.

Then Again and Again

We can simplify a model by mathematically lumping its functions together, by reducing its resolution or graininess, or by lowering a model's dimensions. We do almost anything to make the simulation work better; but even complicated mathematical descriptions quickly grow too simplistic, especially if the model is to describe a specific geometric domain or transitional region where our understanding of the underlying meso-physics is incomplete. Remember, bees live at dimensions where our knowledge of the physics a bee encounters daily is largely unknown. Again, one reason for our incomplete understanding is our reliance on overlapping scales in our descriptions.

The complex forms of macro-sized systems and MEMS devices may be incomparable. Macro-sized mechanical devices function in multiple energy domains and utilize many components. Some share topological boundaries, such as fluids bounded by moving parts. Design and manufacture of the more complicated macro-mechanical devices employ several techniques. We have difficulty integrating diverse techniques and tools into generalized sequences or programs because often no general relations or even correlations between a device's form and what it does exist. Example: an automobile production line.

VLSI Devices Obey Simple Laws

Very large-scale integration or VLSI is the current range of size for miniaturizing microchips. VLSI refers to microchips having hundreds of thousands of transistors on each chip. We can use structured design methods to build VLSI systems, because a computer chip functions only within a single domain of energy, and chips obey strict rules. Each of the many transistors on a chip maps directly, one to one, from function to topology, and each transistor obeys simple rules for interconnections. An example of such a 'rule' is Kirkoff's current law that states that the current entering a node equals the current leaving it. There is an analogous law for voltage. Such simple laws used to model circuits have far reaching consequences, as these rules are the starting points for analyzing any circuit. Unfortunately, we still must find similar basic current rules for flows of energy and information for both MEMS and biological systems.

0 0

Post a comment