Micro- and nanosystems, scale effects and fabrication limits

Separate geometry from material behaviour

A microelectromechanical system combines small structures with functions such as sensing or actuation. Nanosystems involve dimensions where surface and size-dependent phenomena can become important. Micro and nano are useful scale descriptions, but neither label is a universal explanation of how every device behaves.

Geometric similarity preserves shape. Length scales with L, a corresponding area with L² and volume with L³. Consequently surface area per volume scales as 1/L. A process acting through an interface can therefore become more important relative to an inventory in the volume, even before any material-property change is considered.

State which physical model still applies

Surface adhesion, transport paths and electrical or thermal behaviour can compete differently as dimensions change. Quantum effects can alter some nanoscale properties, depending on material, dimension and conditions. They do not dominate every quantity merely because a device has a small feature. Choose the mechanism needed for the intended prediction.

The figure uses two cubes much larger than a nanometre to isolate geometry. It does not demonstrate a nanoscale material effect. This distinction matters: an area-to-volume calculation can be correct while a prediction of reactivity, friction or heat transfer remains unsupported because surface condition and transport were not specified.

Likewise, reducing one characteristic length does not establish complete similarity of a fabricated system. Leads, supports, coatings, packages and measurement connections may retain different dimensions. The small active structure can be strongly influenced by those surrounding interfaces.

Fabrication and metrology define the realised feature

Top-down patterning and material removal create structures from a larger starting object; bottom-up approaches assemble smaller constituents. Both require control of relevant dimensions and variability. A nominal design feature is not evidence of the realised feature or its distribution across production.

Measurement must define the feature edge, method and uncertainty. Inspection of a few favourable examples cannot establish uniform yield. Compare the actual requirement with process variation and metrology, then examine packaging, contamination and lifecycle changes. Small size alone establishes neither precision nor suitability.

Compare surface exposure and relative tolerance

For an original cube of side L, area is 6 L² and volume L³. With L=2 mm=.002 m, area/volume is 3000 m⁻¹. Reducing side length tenfold to .2 mm=.0002 m raises the ratio to 30000 m⁻¹. Area becomes one hundredth and volume one thousandth, giving ten times more surface per unit volume.

Now supply the same absolute dimensional deviation magnitude .01 mm in each case. Relative to 2 mm it is .5%; relative to .2 mm it is 5%. This is a ratio comparison, not a measured manufacturing capability, probability of defects or permissible tolerance.

The cube examples use millimetre and submillimetre dimensions. Neither is evidence of quantum behaviour. To predict an actual small device, the relevant material, interfaces, geometry, process and measurement evidence must still be supplied.

Try the reasoning

1. What happens to area/volume if similar cube side length halves?

   Show the answer and reasoning

   It doubles, because 6/L varies inversely with side length.

2. Does a nominal nanometre feature guarantee nanometre measurement uncertainty?

   Show the answer and reasoning

   No. Fabrication and measurement have separate variation and uncertainty.

3. Can area/volume alone predict reaction rate?

   Show the answer and reasoning

   No. Surface state, transport, chemistry and the appropriate rate relationship also matter.

What to carry forward

Use scaling to identify questions, then test the mechanism and realised feature.