Higher bearing surface area of smaller machinery

A scaling law.
When scaling things down the ratio between surface area and volume changes. Specifically halving the size of an object doubles its surface to volume ratio. This can be easily seen by cutting a cube into eight sub-cubes and calculating the ratio between the surface to volume ratios before and after the cutting. (wiki-TODO: add illustrative image)

As concern in regards to:
"macroscale style machinery at the nanoscale"

The potential consequences of rising surface area are one of the major concerns when it comes to the assessment of the feasibility of macroscale style machinery at the nanoscale.

Potential issues include:

• first and foremost: rising friction power losses
• second: rising corrosion rates
• third: dirt and lubricants clogging machinery

Up: Macroscale style machinery at the nanoscale

Concern: Friction-power losses

There are no less than three factors that work against the growing surface area effect when it comes to increasing friction power losses.

• first and most importantly: the rising throughput per volume scaling law
• second: the superlubricity effect
• and third: infinitesimal bearings (an "invention" of this wikis author)

It should be possible to keep power losses low enough for a practical functioning nanofactory (with large safety margins). Even systems more efficient than biological diffusion based systems may be possible. For details why see here: (wiki-TODO: add link).

Counter-factor: throughput per volume scaling law

(Main article: "Higher productivity of smaller machinery")

This law is less much less known than the scaling law for surface area per volume, but it plays a major role in compensating the rising friction effect of it.

• there is rising throughput-per-volume @ constant operation speeds
• or equivalently constant throughput-per-volume @ falling operation speeds
• this causes falling friction power losses -- quadratically falling :) since it is dynamic friction

Counter-factor: superlubicity

This is not exactly a scaling law but an effect that is only available at the nanoscale in atomically precise systems in dry sleeve bearings with non meshing atomic bumps. The effect is experimentally proven. For more details see the main article: Superlubrication.

Friction power losses can be lowered from three to five orders of magnitude compared to motion in a liquid.

Counter-factor: infinitesimal bearing

These machine elements distribute speed differences equally over multiple coplanar surfaces. Due to friction power falling quadratically with speed this kind of allows one to "cheat" scaling laws a bit.

Concern: Surface oxidation

Conter-factor: Non-oxidizing materials

This is one of the reasons why the choice for building materials falls mostly towards already oxidized materials -- flawless ceramics (aka gemstones) instead of metals.

Note: This is very much unlike macroscale machinery where unoxidized metals are usually the material of choice.

Conter-factor: Perfect sealings

Gas sealings are possible that are FAPP perfectly tight.

The interior space of gem-gum manufacturing devices (and their products) is exceptionally well sealable against outside gasses -- (wiki-TODO: add reference). So inside even oxidation sensitive materials can be used. (given they are not thermally sensitive that is they don't show surface diffusion).

Counter-factor: Compact products

Almost all of the machinery is not exposed the atmosphere.

In case of bulk products (most products) by far most nanomachinery surface is not located on the outside products surface, but in tightly sealed inside chambers. For the minute outside macro-product surfaces especially materials that are highly stable against oxidative (or other) chemical attack can be chosen.

Concern: Nanomachinery getting clogged

Lubricants (or solvents including water) may seem like gravel at the atomic scale. Small molecules are:

• very slippery (since there are very vew available DOFs for energy being dissipated into heat) and
• very strongly jostled by thermal motion (much faster than the machine motions)

thus they are unlikely to act like wrenches in gears(**). Nonetheless lubricants won't be used in nanofactories because with superlubrication one can achieve much lower friction levels (as already noted above in a previous section).

• Counter-factor: no lubricants present
• Counter-factor: no dirt present at inside machinery (FAPP perfectly sealed)

Dirt is somewhat of an issue at the outside of products and in the context of recycling where things may need to be pulled back in again.

(**) Off topic side note: Trapping solvent molecules on purpose should be possible despite the large speed differences (e.g. by tightly sealing big chambers).

Related

For quantitative calculations please consult Nanosystems (or its freely available predecessor paper).

Related pages:

• Friction
• Superlubricity
• Infinitesimal bearing

Effects on transport

Transport in general:

The further one wants to transport stuff and
the smaller the stuff one wants to transport
the more one wants to first bunch and link this stuff together via some "pre-transport". This minimizes the shear slide motion surface area between the parts-to-transtort to the static environment for the majority of the transportation distance. Naturally the distance of the "pre-transport" (and the possible "post-tranport") needs to be minimized.

(TODO: Work out the math here.)

If the main transport distance gets down near to the same order of magnitude of the pre- and post-transport distance then at some point one falls below a threshold below which it does not energetically pay of to do this "pre packaging for transport" any more.

For a general transport of parts that are not necessarily designed to fit together there needs to be a solution of transport containers (a bit over the size of the parts to transport) that do fit together (providing a shape adapter function).

Short range transport in Nanofactories

In a layered thin film chip like nanofactory design (as the current concept foresees) this bunching together with a minimal pre-transport path length is ensured. Well, it is better to call it inter-assembly-layer transport here, since there is no main- and post-transport. The main transport would coincide with the pre-transort of the next assembly layer. Well, if one stretches it, one maybe could maybe also see the path the final product takes when in usage (e.g. in human hands) as the a main transport and a recycling disassembly process that potentially happens at some other place else as the post-transport.

Back to topic: In convergent assembly the shortest possible inter-assembly-layer-transport is ensured by well designed convergent assembly. (between assembly layers interdigitating routing layers) (Convergent assembly that needs to heed equivalent layer stacking requirements due friction reducing slowdown at the bottommost levels).

In the case of convergent assembly in nanofactories there is usually no need for transport containers providing an adapter function since the parts are made to fit together and are put together to bigger fragments of the product right away.

At higher assembly levels and consequently bigger size scales sliding surface areas become very low. Also there now is plenty of space available that can be filled with infinitesimal bearings to reduce friction levels even further.

One may want to now trade this extremely low friction to just low friction for something else. One may want to trade it for a certain principle that can beef up assembly speed. Namely part streaming designs.

Part streaming designs are remotely similar to todays 3D printers where there is a continuous feed of plastic into a moving manipulator, but here with discrete parts that are themselves already pre-assembled from smaller discrete parts.

The idea is to stream the previous level pre-assembled current level unassembled parts through the manipulators arms interior (or on the manipulators arms surface) directly towards the end effector. This way the manipulator does not need to constantly go back and forth to pick up new parts while still retaining all of its general purpose capabilities. (Unlike lowest assembly level mechanosynthetic assembly where there too is streaming but no general purpose capabilities).

At the macro scale assembly levels (e.g. playing dice size to room size) one could even imagine highly dexterous tentacle like manipulators. Educational models may forgo on the usage of UV and further light absorbing nanomachinery protecting functionality and operate at a slowed down pace such that one can follow the parts wandering through with ones bare eyes.

Streaming designs may be sensible starting relatively early in the assembly level stack. That is as early as microcomponent assembly to product fragments (). At this size levels streaming would look much more mechanical though. There's simply not enough space yet for the aforementioned organic looking tentacle like designs. No space for thick swaths of infinitesimal bearings that are combined with advanced mechanical metamaterials that emulate elasticity.

The "productive nanosystems" concept video features streaming designs at the microcomponent to product fragment assembly level. There is no need to putting them at the end of the assembly level stack (as shown) though.

A tentacle streaming designs at the very topmost macroscale assembly is a quite conspicuous deviation from the classic more basic nanofactory thin film design. It's important to realize that despite the change in looks, proper design principles for nanofactories the we already have learned are not being abandoned.

Some speculations

Streaming designs can be combined with:

• (1) The classical pick and place design.
  Just make the tentacle manipulator pick up a part from the sub assembly cell it comes out from instead of letting it draw from its streaming supply
• (2) A parallel extrusion design.
  As if it where the last and uppermost assembly layer.

This may not make much sense in the context of first virgin assembly. Just as it is the 3D printing processes of today, with a cracked open cross sectional plane one can have access to just about everywhere. It may make more sense in case of product reconfiguration, where putting just a few parts from one hard to reach spot to another hard to reach spot by some complex manipulator paths may be quite a bit quicker than taking the whole thing fully apart. Completely down to its rather small base parts.

Long range transport for recycling

The necessity of long range transport also crops up in the context of recycling. Since what person A does not need anymore may be needed by person B that sits somewhere quite far away.

Again to reduce friction losses, for long range distance transports one may want to bunch parts (of all size scales) together to even bigger "parcels" before "shipment".

Microcomponents are the most versatile parts in the assembly level stack. They are simultaneously reusable and still rather fundamental (not as fundamental as crystolecules, but fused crystolecules lack in recyclability). Due to their versatility microcomponents may be the most desirable to send/ship to other far away places over long range distances where it's from a friction minimization standpoint better to have parcel sizes that are quite a bit bigger than microcomponents. Bigger parcels require and equate to bigger diameter of the transportation lines.

Taking a wild guess for diameters one could maybe think of:

• millimeter scale lines for many cross city scale (intercity) transport lines
• centimeter diameter lines for many cross country scale (international) lines
• decimeter diameter lines for many cross globe scale (intercontinental) lines

One may imagine those lines (superlube tubes) not entirely unlike conventional tube mail. Just that there is neither air nor vacuum in there. It would be a solid state stream of one very very long flexible "parcel". A "parcel" made up out of many very small transport containers. Containers that link together with enough elasticity emulating metamaterial capabilities, such that they can go around the necessary curves in the lines. Also the "parcel" (flex-pack-stream?) would be lubed in a quite thick shell of infinitesimal bearing metamaterial. (Getting infinitesimal bearing metamaterial to emulate elasticity at the same time might be quite difficult to design.)

To get microcomponents out of nanofactories and suitably packaged up in that "solid state stream" for transportation, one needs to make the nanofactories (possibly specialized ones?) assemble the solid state micro-parcel stream just like any other product. There is no need to somehow tap nanofactories "sidewards" between the assembly levels, as one may think. That won't work.

Excess material will need to be managed in caching depots. Not entirely unlike digital memory caches in computer architecture just much bigger (storehouse size) and with physical immutable contents instead of mutable bits and bytes.

If the requested quantities of some type of microcomponent are too high to being fulfilled by caches nearby caches farther away need to be used too. Old stuff never used by anyone anymore needs to be disposed of in a safe way with zero release of waste (e.g. burning, dissolving, ...)

Beside transport of microcomponents there is no reason for not having something even more similar to conventional tube mail. Still superlubricated but this time with real macroscale capsules where one can put in non gem-gum products too like e.g. bananas to give a completely arbitrary example.