Better would be citizen R&D but citizen science is a commonly known name so we'll go with that.
Not much can be done yet.
- A lot of knowledge is necessary to get going.
- Unlike e.g. in 3D printing one does not (yet) get anything macroscopic physical and useful in nature to showcase as resulting product.
- theoretical investigations and modeling is difficult too, but still much easier than experimentation
While astonishingly it's possible to resolve atoms in ones living-room for a very small budget, (one can e.g. resolve atoms of graphite single crystals HOPG or atoms on gold surfaces, both very resilient against oxidation in air) at this point (2018) doing actual physical R&D for APM at home is unrealistic.
Interesting advanced machanosynthesis reactions are not possible since:
- Testing more advanced reactions require a very good vaccum (UHV)
- (Lack of cooling is actually not that much of a problem. Weak easily diffusing bonds aren't of much interest anyway)
- Testing less advanced in solution biomineralization mechanosynthesis still requires advanced preparation facilities
Best bet is foldamer technology.
Especially structural DNA technology where one can already order the strands over the web. Short DNA strands (oglionucleotides to be pricese) are not really cheap yet, but also not exorbitantly expensive anymore. DIY STMs would likely have sufficient resolution to resolve these structures, but these structures are non-conductive, so AFMs are needed, which are more difficult to bild for equivalent resolution. Wheter a DIY AFM with sufficient resolution to resolve (and manipulate) structural DNA structures can be built with a very tight DIY budget is a good and open question.
Interesting to observes is the "DIYification" of microbiology (aka biohacking) especially 3D printed DIY microfluidics might rise from this field and allow for efficient usage for many assembly experiments of the still precious oglionucleotide base material.
- Microscopic microscopes for the masses ...
Once one manages to install the old Nanoenginner-1 software (especially difficult to do on Linux) then it's quite easy to get started playing around designing some new types of crystolecules. The Nanoenginner-1 software is old has some issue. It urgently needs some overhaul (rebuilt from scratch on more modern software fundaments).
First off: With all of the few crystolecules that have been designed and simulated by now (2018):
- The goal was not to establish big library of random parts.
- The goal was to establish a rough proof of the general existence and functionality of this class of physical objects.
- The goal was to establish some far apart points in design space that span a convex hull in design space that encompasses as much "design space volume" as possible. When these spanning post designs do not work (well or at all) as is (e.g. because being a bit too hight or a bit too loose in fit) it is not a big issue. In later implemented actual physical design slight tweaks can be made to make it work.
One could argue that if only enough designs are made then some designs are bound to work. But isn't that a giant waste of effort? Well if it really is a fun game then it at least has the value of the entertainment it provides to users.
To build up big libraries (where some designs are likely to hit the sweet spot) various (semi)automated design series might be a better approach. Setting up those is too complex to be gamified though. It really seems not possible to make that into a proper game.
Another danger: Designs that are right away (that is possibly prematurely) built up into very complex and big assemblies the upcoming necessity to quite drastically tweak the base parts may well leas to a the whole thing needing even bigger redesign (aka design from scratch). The typical card-house effect.
Design of crystolecules requires some knowledge about what works and what does not, since atoms do not always behave like a construction toy (actually its more of an exception than the rule - but in the spirit of engineering an exception that is thought after). That is especially important when the design software is unaware and nonreminding/nonenforcing of some of these limitations (like it is the case in Nanoengineer-1 for example). So a gamified design tool needs to have extensive empiric knowledge incorporated, so to keep users from making poisonous bombs instead of proper nanomachinery.