-5

Sorry my chemical knowledge is not very deep so some of what you assume is common knowledge will probably be new to me.

For a game i am trying to make i wrote a simple atom / compound simulator where i try to build compounds with simple covalent and ionic bonds but in a very basic sense without taking anything into account beside the number of electrons

so far my approach was to implement rules and every time i the simulator spills out a new compound i google if really does exist (did not find a comprehensive list of existing compounds, please link if you have one) and if it does not i dig deeper to understand the rules better

so after implementing dativ bonds i got a compound of C, Li and F where Li forms an ionic bond with C and C forms a covalent bond with F and F helps C out with a dativ bond to reach the 8/8

so far i only found CF5Li which would suggest that this compound can exist in some way but i am not sure if the dativ bond between C and F is "real" or my implementation is beyond flawed because its much more complicated (like for example that F does have "free" electrons but there is no overlap between the missing electrons orbits of C and the available ones of F)

braunbaer
  • 3
  • 2
  • 2
    The problem is that rules based on valency can only get you so far. You can generate lots of compounds that formally satisfy all atoms' 'desired' valency (or octet), but don't make sense IRL because they would immediately decompose to a more stable arrangement. I'm guessing you are a programmer - this is like trying to generate working programmes just using the syntax rules of a language: you can generate valid programmes, but the vast majority of them wouldn't make any sense. – orthocresol Jan 02 '23 at 16:43
  • 1
    You might think CF3Li should exist but it is unstable with respect to LiF and difluorocarbene – Waylander Jan 02 '23 at 16:45
  • @Waylander indeed, i was already thinking about some kind of "bonding priority" because i am also not able to create O3 or HNO3 , with your comment you gave me one more reason to dig into that – braunbaer Jan 02 '23 at 16:49
  • @orthocresol i am indeed a programmer, do you think its possible to define whats stable and what is not with enough knowledge ? – braunbaer Jan 02 '23 at 16:51
  • In principle, the answer is yes, you can - however, the amount of 'knowledge' you'd have to encode is huge, because determining stability (computationally) essentially involves solving physics equations (e.g. Schrödinger) from the ground up. Naturally, the less knowledge you provide it, the more errors you will run into. One sensible compromise might be to restricting the elements to a set of well-behaved, organic-ish elements, namely C, H, O, N; it's often possible to use lower-level heuristics (i.e. based on bonding patterns, not quantum mechanics) to make a guess as to their stability. – orthocresol Jan 02 '23 at 17:00
  • Unfortunately it is not yet possible to get an algorithm which is able to say if a given structure does exist. Let's take for example the molecule H-N=O. It should exist, by the octet rule, but it does not exist. Why ? – Maurice Jan 02 '23 at 17:01
  • @Maurice Nitroxyl does not exist? – braunbaer Jan 02 '23 at 17:03
  • @orthocresol i see, thank you very much for your input. restricting the elements will not be an option for me because the whole goal is to have a "good enough" representation of the "real rules" and than create fictional elements / atoms that not exist but follow the same rules for bonding but I appreciate your suggestion non the less – braunbaer Jan 02 '23 at 17:07
  • 2
    Generalising chemical rules across all possible chemistry is extremely ambitious. The vastly simpler problem of enumerating all possible simple hydrocarbons (just hydrogen and carbon with single bonds and no loops) is already very hard despite being nearly adressible using graph theory. See this question for some of the detail of that problem. – matt_black Jan 02 '23 at 17:37
  • @Braunbaer. HNO does not exist. If somebody did succeed in synthesizing it, it gets immediately transformed into $\ce{N2O + H2O } $ – Maurice Jan 02 '23 at 17:50
  • 1
    @Maurice Not in gas phase https://en.wikipedia.org/wiki/Nitroxyl and it hardly is on topic under this question. – Mithoron Jan 02 '23 at 17:52
  • @Mithoron. You are right. HNO is not on topic under this question. It was just here to illustrate that it is not always possible to foresee the stability of a structure by counting its electrons. – Maurice Jan 02 '23 at 21:03

1 Answers1

3

You want to blend $\ce{C}$, $\ce{Li}$ and $\ce{F}$ into one molecule and state "found" $\ce{CF5Li}$ as a solution. From this I take, there is no constraint on the numbers of atoms as long each element is present by one atom, or multiple.

With 3,3,3-trifluoropropyne there is an entry on PubChem beyond of living in silico only, with a CAS registry number 661-54-1 (reference). The compound is literature known since 1950,(1) described with as "a gas of boiling point c $\pu{-48^\circ{}}$ with an odour reminiscent of that of acetylene". Though rather on the expensive side for such a small molecule ($\ce{C3HF3}$, $M = \pu{94.04 g/mol}$), it is commercially available; for \$US 1,010.00 MilliporeSigma sells you $\pu{10 g}$ in a cylinder.

Attempting chemical overkill on paper, one could deprotonate this ethyne with butyl lithium (BuLi), formally creating $\ce{C3F3Li}$:

enter image description here

From a practical perspective, however:

  • the reaction has to be carried out at a temperature of $\pu{-48 ^\circ{}C}$, or below, to retain 3,3,3-trifluoropropyne in the liquid phase. Likely, a reaction temperature considerably lower will better dissipate the heat of reaction.

  • butyllithium is a base way stronger, than typically required to abstract a protons attached to an ethyne. With the presence of a conjugated, electron withdrawing trifluoromethyl group ($\ce{-CF3}$), sodium hydride often used for such a reaction equally would be multiple orders of magnitude stronger, than necessary.

    The byproduct of butyl lithium's reaction as a base is butane. Wikipedia's corresponding property box lists a boiling point of pure $n$-butane at atmospheric pressure in the range of $\pu{-1 \ldots +1 ^\circ{}C}$, so this doesn't leave the reaction mixture. This possibly could hamper attempts to isolate the envisioned product, $\ce{CF3C#CLi}$.

  • Haszeldine already described the preparation of compounds where hydrogen was replaced by a metal, $\ce{CF3C#CCu}$, $\ce{CF3C#CAg}$, and $\ce{(CF3C#C)2Hg}$ with copper, silver and mercury where "the first acetylides decompose quitely on gentle heating and violently when heated rapidly." Maybe $\ce{CF3C#CLi}$ is similar, i.e. unstable. Given the formation of e.g., tetramers for butylithium, the notation of $\ce{CF3C#CLi}$ as an isolated molecule might not be the correct representation of the structure of the reaction product, either.

This doesn't sound like an easy synthesis.

Because you mentioned your background leans more toward computation, than chemistry: the systematic generation of molecules in silico stays a field of ongoing research. Challenges include the identification of efficient algorithms (and their implementation) to shuffle the atoms to build chains, and rings. Then you need quality checks to retain molecules which are chemically plausible e.g., by count of valences (how many bonds an atom typically engages), or ring strain. Last but not least, you equally want to screen these structures for useful chemical and physical properties to identify interesting candidates for drug design, or material science -- still before performing any chemical synthesis in the lab.

Thus, examples like MAYGEN and OMG, or Reymond's massive compilations GBD-11, GDB-13, and GDB-17 can only provide a glimpse.

addition past acceptance of the answer: User @Mithoron suggests in a comment pentafluorobenzene as an alternative, i.e. implies a reaction like

enter image description here

This would be much more affordable ($\ce{C6HF5}$, $M = \pu{168.06 g/mol}$) per gram, as well per mole as the same source as above currently sells $\pu{5 g}$ for \$US 40.00. In addition, the analogue phenyl lithium (lacking the fluorine atoms) is readily available and used a chemical reagent solution. Thus, the synthesis departing from pentafluorobenzene likely is less troublesome than the one on 3,3,3-trifluoropropyne, too.

(1) Haszeldine, R. N. Synthesis of 1:1:1 Trifluoropropyne. Nature 1950, 165, 152–153; doi 10.1038/165152b0

(2) Peironcely, J.E., Rojas-Chertó, M., Fichera, D. et al. OMG: Open Molecule Generator. J Cheminform 2012, 4, 21; doi 10.1186/1758-2946-4-21 (open access); http://sourceforge.net/p/openmg

(3) Yirik, M.A., Sorokina, M. & Steinbeck, C. MAYGEN: an open-source chemical structure generator for constitutional isomers based on the orderly generation principle. J Cheminform 2021, 13, 48; doi 10.1186/s13321-021-00529-9 (open access); https://github.com/MehmetAzizYirik/MAYGEN

(4) https://www.gdb.unibe.ch/downloads/

(5) What are the GDB-13 criteria for "synthetically accessible organic molecules"?

Buttonwood
  • 29,590
  • 2
  • 45
  • 108