Dissipation sharing
Advanced gemstone based atomically precise manufacturing systems will be able to work very efficiently. There are several reasons for this high efficiency. (among them the possibility to work in deeply cooled environment)
One reason that may allows squeezing out the last little bits of efficiency towards 100% is the idea of "dissipation sharing". With the technique of dissipation sharing efficiency is likely even to far surpass the efficiencies seen in biological solution based systems.
For a system to run predominantly in one direction (for it to have an arrow of time) the thermodynamic potential for the system must decrease. For solution phase chemistry the appropriate thermodynamic potential is the Gibbs free energy. One needs to devaluate/thermalize free energy.
For a single reaction the energy pumped into the thermal motions must significantly exceed kBT. Beside the obvious thing to lower the temperature (which has practical limits) there's another opportunity.
In solution based systems reactions are mostly isolated from each other every reactant molecule has its own degrees of freedom. Stiff machine-pase-systems allows one to link many reactions together. More concretely speaking: Many (indirectly all) of the molecular mills are linked together in the background stiffly by common rotating shafts. This effectively reduces many individual atom depositions to only a single one. Only for this one one needs to still expend en energy >> kBT to ensure forward motion.
How many reactions can be coupled together (the spacial and temporal reach of the "virtual reaction") is highly nontrivial question. Elasticity modes in the background mechanics introduce new unwanted degrees of freedom that need to be fed. Introducing mechanical advantage as soon as possible to gain high virtual stiffness in the background may be a feasible strategy.
Occasionally running backwards a slight bit may be bad for a nanofactory.
If the reactions are not fully reversible one might end up in an unknown failure state (atom at wrong location).
When progressing with open loop control (manipulators working blindfoled to a good part will be normal in advanced nanosystems) the resulting errors effects may range from irrelevant over performance degrading to fatal for the local subsystem. Depending whether on the location of the error ant the degree of consequential errors.
---
- More fine grained .. partially electronic ion concentration potential.
- ATP ... Energy in quanta (not quantum quanta here!)
- Having no pockets for the "exchange money" energy.
