With Levitation one can bear very high speeds in a very small space when load isn't excessive. It offers friction levels even lower than superlubrication. If even less friction is needed sufficient emty space surrounding the movement trajectory is needed to do throw and catch maneuvers through vacuum.
Dynamic electrostatic control
Some form of electrostatic sensor is needed.
A lot should be known from MEMS here.
Electrostatic lagrange points
Mechanically forcing two equal e.g. positive charges with sufficient difference in charge [factor 24.65 = 25/2+sqrt(621)/2] to circle around each other around their hypothetical barycenter with their hypothetical natural rotation period gives two stable Lagrange points in L3 L4 for small charges of the opposite sign - just like in celestial mechanics. The ability to deviate from the natural movement might allow for further optimisation of the stable points. [To inverstigate: is there a better and if is there an optimal configutation] The usefulness of this isn't quite obvious.
Note that any (sufficiently isolated) zero dimensional nano sized object is subject to notable quantum mechanical wave dispersion and tunneling (meaning in colorful words that it kind of "dissolves" an reappears somewhere else by chance).
Beside thermally induced dipoles there's also a not yet well understood but certainly practical relevant component of the Van der Waals force that originate from the suppression of non harmonic modes of heat radiation ( ... virtual particles ... ) in nano to micro scaled gaps. Certain geometries like an elongated ellipsoid over a circular hole in a plate) lead to static levitation. [Todo: find out why this shape]
Negative compression bearings
When bushing and axle of a DMME-bearing form an increasingly big but not too big gap the force can switch from inward from all directions to pulling outward in all directions at the same time but still provide a stable center for the axle. This lowers the waviness and coupling of the bearing and makes it more levitation like.
This happens whe the shafts surface lies between the minimum and the inflection point of the Lennard Jones potential of the bushing atoms. Going beyond the inflection point (unloaded bearing) makes the axle stick to one side of the bushing. Going for the minimum of the potential leds to zero local stiffness (like in tensegrity structures).
Magnetism does not scale well with shrinking size it becomes very weak at the nanoscale thus it's mainly useful in levitating macroscopic parts. The manipulation of magnetic properties of diamondoid materials falls unther the non mechanical technology path
There is also the method of optical tweezers (and smaller plasmonic tweezers) which could be counted as "levitation methods". (Note that these methods are likely insufficient in stiffness and force for single atom placement especially in advanced mechanosysnthesis.)
Also there is sonic "levitation" for bigger things immersed in gases. The contact to a gas comes with much higher friction though.
- nanoscale turbomelecular pumps (but the may not be needed since positive displacement pumps do just fine)
- carriage particle accelerators
- moving surface medium movers
- Bearings can be stable despite attractive interactions between their surfaces (at K. Eric Drexlers website)
- Wikipedia:Q factor