[Enthalpy]

Hello you all!

I'm currently using ArgusLab to compute the enthalpy of formation of small molecules:
www.arguslab.com/downloads.htm
and I don't often get the 10kJ/mol accuracy I'd need to compare rocket fuel candidates, hence I ask for your help and explanations.

I understand AM1, PM3, MND0, UFF, Amber are sets of mathematical expressions, or maybe software functions, to evaluate the energy of a molecule in some given shape. ArgusLab, or other software around them, optimize the form of the molecule to minimize the energy accordingly.
(1) Is that correct?

These energy estimation functions are common to various molecule optimization and drawing-display software, which should give the same result if using the same function.
(2a) Is it right?
(2b) Would you suggest a better free software for Windows? (I know, free and Windows don't go together)

Such software computes the energy of the molecule alone in the Universe, without any rotation nor vibration - as if it were a dilute gas at zero Kelvin - as compared to the elements taken in their normal state but at zero Kelvin.
(3) Do you agree?

My main error source for a liquid at 298K is that I should have added the molecule's rotation and vibration heat at 298K, subtracted its vaporization heat, and subtracted the heat of the elements at 298K. This is bigger than the software's error at zero Kelvin.
(4) Your opinion?

Rules-based software isn't bad at estimating boiling points, so maybe I could take its estimates for enthalpy of vaporization and heat capacity to correct the computed heat of formation from a gas at zero Kelvin to a liquid at 298K, though heat capacity varies a lot in between.
(5) Could I?

Thank you!

[tamim83]

I can answer the first 2 for you. 

1.) All of these are methods for optimizing the geometry and calculating the energy of molecules.  AM1, PM3, and MNDO are semiempirical methods, meaning they use fundamental equations to get the energy, but substitute experimental parameters and/or ignore parameters to make the calculations go faster.  These are quantum baed methods.  UFF and AMBER are force fields used in molecular mechanics.  Molecular mechanics uses classical mechanics (like Newton's laws) to calculate the energy.  The force fields parameterize all of the interactions between nuclei in a molecule as if they are moving in an electron field.  Molecular mechanics is very useful for large molecules like proteins. 

2.) In theory, if I use the same level of theory to get the energy of the same molecule using two different software packages, I should get the same answer.  What may differ is the algorithms used in the programs to implement each theoretical method.  In other words, the time it takes to do the calculation may be different. 

As far as freeware goes, there is GAMESS.  You can use Avogadro to visualize your results.  Avogadro also lets you use molecular mechanics to optimize the geometry of your molecule. 

Hope this helps some. 

[Enthalpy]

Thank you so much, Gamim83! This really helps.

More contributions are welcome as well!