Latest revision as of 12:09, 21 May 2022

Chemical synthesis is ...

• ... one way to make atomically precise structures and is ...
• ... the oldest way to do so discovered/invented by humans.

The size of of atomically precise structures that can be made by chemical synthesis is limited to to small (to medium sized) nanoscale molecules.
This is because:

• (1) Mixing stuff in big flasks is slow and
• (2) It is bound to an exponential drop in yield
• (3) parallelization in synthesis for the individual product molecules is not possible due to limitations in specificity.
  Not to confuse with the parallelism of synthesis of many identical small molecules. This one is trivial and massive.

Making atomically precise products of macroscopic size via chemical synthesis is impossible.

Nature circumvents this by thermally driven self assembly which goes (per definition) beyond chemical synthesis.
Also nature only features topological atomic precision at the macroscale not positional atomic precision.

Making stuff from scratch

(1) Speed limit due to size of chemical reaction vessels

This can (and is beginning to) be greatly improved on by microfluidics.
Microfluidics does not need atomically precise technology with macroscale product size and is thus doable today (2021).
Related: Non atomically precise nanomanufacturing methods#Microscale

Once stuff needs to be separated more or less ingenious tricks need to be employed
(distillation, re-crystallization, liquid-liquid extractions, out-gassing, ...)
These each come with their own challenges in miniaturazisability.

Important "construction sites" of technology

• There are efforts in automation of chemical synthesis
• There are efforts in improvement on microfluidics

• Both eventually need to be combined.
• Both seem quiet relevant for getting towards advanced productive nanosystems faster.

(2) Exponential drop in yield

See main article: exponential drop in yield

Without any form of mechanosynthesis available yet
(aside from borrowing weakly mechanosynthetic biological nanomachinery)
This cannot be improved on in the case of linear molecules in solution floating around freely.
If there is a fatal error it's game over. Synthesis can no longer proceed.

As an analogy it's like a severe car crash on a narrow road that hopelessly blocks that road.
With every car passing having a chance to crash of 50% (bad drivers, horrific road) and only ten drives passing
the chance of the road not being blocked is 1/2¹⁰ which is one in a thousand.

(3) Averting the drop by going to sheets and volumes

Well. this at least can be improved on for structures with higher dimensionality like 2D sheets and 3D volumes.
Where the chemical synthesis (aka product assembly) can redundantly move around single failed spots.

That leads right into the next problem though.
The advantage of many open ends to continue the synthesis (assembly) process on is simultaneously a disadvantage.
If there are many spots to continue on then many of them need to be exactly the same as others.
This is because reactions for strong covalent bond formation (in some sense defining chemical synthesis)
gives only very very limited options for specificity. E.g. compared to thermally driven self assembly. A note on that later.

Going to weaker bonds (including e.g. Van der Waals bonds) there can be more specificity
and a larger number of differing assembly spots. Plus kinetic traps can be averted.
Also then this is no longer called chemical synthesis but thermally driven self assembly. So let's talk about this later.

Examples:

One prototypical case of chemical synthesis in 3D (stretching the term a bit) is crystal growth (often done with pure metals).
One gets more or less defined crystal borders from crystal growth.
In the case of more defined crystal growth the slowest growing crystal faces prevail.
This may be a bit counter intuitive but an animation can make that intuitively obvious.
One gets nanoparticles with quite high symmetry. BUT ... There are incredibly severe limitations:

• Shape is only minutely adjustable by growth condition/parameters – all shapes are convex polyhedral only (right?)
• Size is somewhat statistical since there is no defined termination point
• It's limited to one monolithic material and even switching that monolithic material for another is typically not a straightforward option

Overall this seems very mich not promising as a part of a fast targeted path toward advanced productive nanosystem like gemstone metamaterial on-chip factories.
Of course one cannot exclude that it may help in some obscure ways. But it's definitely not the technology to focus on.

Another case that is thinkable is "branching and fusing chemical synthesis".
This seems like rather exotic chemistry. (TODO: Is there existing work on this?)
Specificity for more than a few (two, three) active sites to continue forming covalent bonds on is hard and reduces the success rate for each of the spots.
As an analogy imagine a slightly braided river, blocking one path a ship can still redundantly pass trough another.
Each individual path has a a higher chance of being blocked though.

Borrowing the nanomachinery of nature

Nature has in some sense overcome the limitation to only nanoscale atomically precise products
as the existence of macroscale organisms impressively shows.
Then again over macroscopic scales

• it is only topological atomic precision (what links to what)
• it is not positional atomic precision like in a perfect single crystal of quartz. (where is that specific atom, same isotope atoms distinguishable by position)

As far as positional atomic precision over macroscale distances would be theoretically possible given thermal expansion, sound waves, and such. Related: Why identical copying is unnecessary for foodsynthesis

Limitations

Limitation ins product classes:

• peptides, proteins
• DNA, RNA
• Sugars, lipids

There are a lot of fancy tricks to make variations on this basic natural product types.
E.g. unnatural side chains on peptides.
But variations typically decrease yield and increase production costs.
For some variations one even needs to

• go back to in vivo synthesis without employing the nanomachinery of nature or
• take an intermediate approach (this exists too)

Peptides and proteins

A major focus lies on peptides and proteins since these make up the major chunk of soft nanomachinery in molecular biology. Other stuff usually just modifies proteins. (With the notable exception of the ribosome, the soft nanomachine that makes proteins. Not a coincidence.)

Callenges:
Expressing proteins in cells:

• sometimes the proteins don't fold properly because of interactions with other parts of the cell
• sometimes the proteins cannot successfully extracted and this is hard to predict.

Advantages: A major advantage of employing the nanomachinery of nature is its integrated error correction mechanisms
Also in the case of foldamers the subsequent folding process is helped along by so called chaperone proteins.
This goes beyond chemical synthesis though.

Interesting trivia:
Cells can produce astounding amount of artificial proteins without dying.
That looks quite odd in electron microscopy.

Self assembly

This goes beyond basic chemical synthesis.
See main article: Thermally driven self assembly

Related

• Spectrum of means of assembly – Mechanosynthesis
• Exponential drop in yield
• Atomically precise nanostructure
• Types of chemical reactions for synthesis

External links

• Liquid–liquid extraction - Extraction (chemistry)