Page 1 of 1 [ 7 posts ] 

IrishJew
Blue Jay
Blue Jay

User avatar

Joined: 22 Apr 2013
Age: 43
Gender: Male
Posts: 82

08 Jun 2013, 2:11 pm

Is it possible to compute a substance's boiling point and melting point by simply knowing the molecular structure of that substance?

For example, let's say I "invent" a substance. Say: 3,4-dichlorophenyl calcium 3-fluoroglutarate. Could I then, using some equation, derive its boiling point, melting point, and then maybe some other characteristics?



Stargazer43
Veteran
Veteran

User avatar

Joined: 6 Nov 2011
Age: 34
Gender: Male
Posts: 2,604

08 Jun 2013, 3:01 pm

To the best of my knowledge, there is no easy way to do that. There are several correlations that do exactly that (such as the Joback method), but I don't think that any have proven to be viable enough to be used as an all-inclusive formula with high accuracy. Depending on the specific compound in question though, you can probably find an apporpriate correlation to use that will give a reasonable degree of accuracy, it just may take some searching to find an appropriate one. However, if you do obtain the boiling point and melting point at a given temperature and pressure (as well as a small amount of additional data, all pretty easy to measure), then you can interrelate that to other conditions using pretty simple calculations. They're really simple tests to perform so it's usually one of the easier physical properties to measure. I'm assuming you're referring simply to pure components and not mixtures also.



The_Hemulen
Blue Jay
Blue Jay

User avatar

Joined: 19 Apr 2013
Age: 30
Gender: Female
Posts: 83
Location: UK

08 Jun 2013, 3:44 pm

The most generally applicable approaches would involve molecular simulation, either running molecular dynamics simulations and observing behaviour directly or calculating free energies and using these to work out the melting point (more accurate). *Lots* of different methods have been developed. The most accurate of these will do pretty well, though the values won't be exact.

(I am doing postgraduate research in molecular simulation - I can answer in more detail if you so wish!)



IrishJew
Blue Jay
Blue Jay

User avatar

Joined: 22 Apr 2013
Age: 43
Gender: Male
Posts: 82

11 Jun 2013, 12:38 pm

Thanks, guys! Perhaps a complete solution to this problem is up for the taking? I imagine it would involve the concept of tesselation. For example, 1,2 dichlorobenzene might have a higher boiling point than 1,4-dichlorobenzene because the molecules of the former tesselate better than the molecules of the latter. What do you think?



Stargazer43
Veteran
Veteran

User avatar

Joined: 6 Nov 2011
Age: 34
Gender: Male
Posts: 2,604

11 Jun 2013, 12:43 pm

IrishJew wrote:
Thanks, guys! Perhaps a complete solution to this problem is up for the taking? I imagine it would involve the concept of tesselation. For example, 1,2 dichlorobenzene might have a higher boiling point than 1,4-dichlorobenzene because the molecules of the former tesselate better than the molecules of the latter. What do you think?


Believe it or not, I am extremely familiar with both of those chemicals lol, I have a ton of data on them (more than I ever wanted to have), I should have time to pull it up later tonight if you want. What's your interest? BTW, the boiling point difference is pretty small, about 5 degrees Celsius if my memory serves correctly, I'll look up the exact values later. They can be separated by distillation but crystallization is typically the better option, since their melting points are significantly different.



The_Hemulen
Blue Jay
Blue Jay

User avatar

Joined: 19 Apr 2013
Age: 30
Gender: Female
Posts: 83
Location: UK

11 Jun 2013, 1:39 pm

IrishJew wrote:
Thanks, guys! Perhaps a complete solution to this problem is up for the taking? I imagine it would involve the concept of tesselation. For example, 1,2 dichlorobenzene might have a higher boiling point than 1,4-dichlorobenzene because the molecules of the former tesselate better than the molecules of the latter. What do you think?


A solution involving tesselation only takes into account entropic contributions (resulting from how ordered the system is). This might work well for classes of similar compounds, e.g. the chlorobenzenes you mentioned, but wouldn't be a general approach for predicting melting points of any compound. A general solution would also have to take enthalpic contributions to the free energy of the system into account, i.e. the bond strengths between molecules are important.



IrishJew
Blue Jay
Blue Jay

User avatar

Joined: 22 Apr 2013
Age: 43
Gender: Male
Posts: 82

12 Jun 2013, 2:56 pm

Stargazer43 wrote:
What's your interest?


Haha. I'd rather not say. It wouldn't be in my best "interest". Let's just say in certain endeavours, I ran into some difficulties I had to overcome.