Actual chemistry simulations are still computationally unfeasible. In the "ten thousand atoms is a lot" territory.

So you could certainly make a game with hardcoded reaction paths for it, but capturing the actual scope and breadth of real chemistry just isn't really going to happen.

On the other hand you do raise a really good point about something which bugged me a lot when I was learning lab chemistry: getting the mechanical sensibility for what processes you need to do is hard and frankly IMO it's not taught particularly well. There would, I think, be a lot of value in a decent fidelity VR simulation of some of these processes just so you could get a sense of what the mechanical arrangement of things you need to do will actually be.

i.e. testing out a particular glassware and fluid path apparatus would have a lot of utility in making experimental work more structured and developing the intuition for what things should look like.

we have blender, webGL, three.js, godot, blend2web, so much tech...and the pedagogy is...where? in the toilet ...idk? i really need to get smarter asap...lol. So many cool ideas to explore, particularly in STEM.

A simulator like this would require a lot of simplifications: a table of template chemical reactions and thermochemical mixing rules which are bound to be wrong in edge cases. It doesn't take many reagents before you hit a combinatorial wall where the combination has never been tried in practice, and while it could be estimated with quantum mechanics in theory, that gets expensive.

This is what popped into my head as well. In situations where you have multiple phases so that Phase A is required to reach a certain temp before adding Phase B, but Phase A must reach a specific temp then cool back down within a range before adding Phase B. Once Phase A and Phase B are combined, do you stir, blend, whisk, whip, etc? What happens when you need to use 0.14 grams, but add 0.15 grams?

Each little variance as minor as they may seem will affect the results. Making something one time is a definite achievement. Making the same thing multiple times, or successfully scaling the formula up/down is also noteworthy as well. (assuming doing things by hand as amateur from the title implies to me)

All of that to say, it would be MFing impressive if some sort of simulation could accurately demonstrate these kinds of variations in the results. Could be useful to try to recreate where something went wrong as a sort of reversing/debugging the result.

sorry i'm not sure whos comment to reply under - so i'll just stick it here - thanks for the explanations :) I am ignorant of maths, chem, comp science...so my question was posed purely for my own knowledge.

What i described could be likened to raytracing for light...but for chemicals...

If something like this is possible, would there be no use or other limitations to making a "descriptive" simulation. ie. libraries of animated reactions? Or is this what comicjk is refering to?

assuming the kinds of experiments listed in this article will turn out similar results if performed with similar materials?

It's more like, we can do raytracing, but what's really needed here is simulating actual photons scattering, complete with all the quantum effects that involves.

That's a good response. I was trying to figure out how to reply, but didn't like anything I came up with. I would honestly have no idea on how to program a computer to simulate the individual atoms and the activity of their electrons. How to tell the computer when two atoms can bond with the sharing of an electron, how to tell it when a covalent bond can be made, etc. That's probably a lot to do with I barely understand it myself.

Also, I'd assume the the sim would assume all equipment is 100% pollutant free, all ingredients are 100% pure so that there is no adulterants being introduced to the formula. Also a room of perfect humidity and temperature etc.

Maybe not the answer you expect, but a sufficiently ab-initio simulation is not something our current knowledge of chemistry permits! You could approach your problem with various degrees of approximation. Personally, I'm optimistic that we may make great progress in this area in the coming years, decades, and centuries. It's traditionally described as a lack of computing power, but I'm guessing there's a fair chunk of knowledge yet undiscovered that could help.

Some years back there was work on VR lab sims. I fuzzily recall niches of "virtual high-school lab", "intro to lab best practices", and "protocol training in life sciences"? No idea if any of it reached market.