[azmanam]

I downloaded Avogadro a few weeks ago and have been playing around with it.  It's not entirely intuitive, but I'm able to get most of what I need.  (Except for electron density maps... every time I try to calculate that surface the program crashes...)

My biggest hurdle to using this more effectively is that I haven't taken a pchem class since undergrad, and I'm pretty sure we didn't do any molecular modelling (we did derive a LOT of equations by hand, though).

Let's ask these systematically.  I'm looking at the Setup Force Field menu.  My force field options are: Gaff, Ghemical, MMFF94, MMFF94s, and UFF.  I don't need the entire background on the creation of these force fields, but a brief description of what it is, what systems it's good for, and why these 5 were chosen for this list would be nice

In various menus, I have either a "Theory" or "With" or "Method" dropdown which all seem to mean the same thing.  The list includes: AM1, MNDO, MNDO-d, PM3, PM6, RHF, RM1, B3LYP, and MP2, among others.  Same questions.  How do I know which to pick, and what are the nuances of each, so I can make an informed decision other times.  Any reason one dropdown maybe contains 5 and another contains 10?

I also sometimes get a "Basis" dropdown containing: STO-3G, 3-21G, 6-31(d), 6-31(d,p), LANL2Dz, and cc-pVDZ, among others.  Same questions.

My "Calculation" dropdown lets me choose between Single Point Energy, Geometry Optimization, or Frequencies.  Same questions.

Final question: why all the cryptic abbreviations?  It really makes it unapproachable to the non computational chemist just picking up the software.

[arit]

My level of involvement with computational chemistry would also be best described as dabbling, but here goes...

Anything having something to do with "force fields" relates to molecular mechanics or dynamics. Force fields are instructions
on how to account for the interactions involved in making atoms stick together (bond lengths, e-e repulsion, etc..)
Since force fields are at least partially experimental, some force fields are better parametrized for different applications.
(CHARMM & AMBER for biomolecules, for example.)

I don't know the specifics of what goes on with the force fields provided as defaults with Avogadro, and their wiki isn't
too helpful on the subject, just stating that,

MMFF94(s): organic chemistry and drug-like molecules.
UFF: This force field can be used across the entire periodic table.
Ghemical: simple organic molecules

As you probably have noticed, even though UFF is said to be parametrized for the entire periodic table, that's bit
of a stretch.

For the next question, let's divide the abbreviations into two groups: (1) AM1, MNDO, MNDO-d, PM3, RHF, RM1, MP2 and (2) B3LYP
The first group is wave-function based "Hartree-Fock" methods. Most of them either contain more elaborate ideas than "just" RHF, or
are semi-empirical. How to choose which method to use depends on what question you are trying to answer, and the size of your
molecular system. I'd recommend googling around for more detailed descriptions about each one.

The second group is for the sole representative of density functional theory. DFT differs fundamentally from HF methods. The B3LYP
functional is probably one of the more commonly used for modeling small organic molecules. The appeal of DFT is that it tends to be less
computationally demanding than comparably accurate HF-methods.

The "Basis" dropdown refers to the basis set to be used in the (either HF or DFT) calculation. In addition to the level of theory, the chosen
basis set has a huge impact on both the accuracy and time consumption of the calculation. I'd suggest referring to a book on computational
chemistry for specifics on the basis sets. In short: bigger is better, but makes for lots of more calculations.

"Calculation" dropdown lets you choose what question you are asking about the system. For a new structure, the first thing you probably
want to do, is optimize the structure at some level of theory, and then go on about calculating the frequencies, or some other properties.
You can also mix & match, by optimizing the geometry with molecular mechanics and doing the frequency calculation using DFT or HF
methods.

The abbreviations are unfortunately cryptic (I have no idea what some of them mean), but at the same time they contain a lot of
detailed information to the specialist.

Sorry for the roundabout response, books are written about these things for a good reason

Basically, if you are just playing around on your desktop, you probably want keep mostly on the molecular mechanics side of things
since it's the least demanding method of the three. Both HF and DFT start getting quite heavy really fast (from a desktop perspective).

[jazzyjanelle]

I have 2 initial suggestions for you. Get any computational chemistry textbook from the library. They all do a decent job explaining the difference between the methods, basis sets, etc.

Second, find a computational paper that lists their method and try to reproduce the numbers they obtained. It is a good exercise that helps you learn what a method is calculating, what kind of properties you can calculate, etc.

Your first set of questions are molecular mechanics, which I am the least familiar with. The calculations treat the molecular bonds like springs. These are low-level calculations often used for large systems or when only a small amount of computational power is available.

The basis lets you choose how you want to model the atomic orbitals.
Some basis sets are better for some systems than other. For instance the LanL2DZ is often used for transition metals (maybe even f-block metals) while the others you have listed work well for organic molecules. I am not sure about cc-pVDZ.

The lingo does appear cryptic at first, but contains extra info. Some abbreviations stand for the last names of 4 people who developed a technique.

Good luck!

[azmanam]

Thanks for your responses.  I don't know if I know any more than when I started, but I should probably get a textbook if I want to really get into it.

Is it a safe assumption to assume that entries near the top of the dropdown are lower in computational power (thus lower in computer resource usage) and entries near the bottom are more involved - thus more accurate?

I'm mostly interested in viewing molecular orbitals and electron density maps of small organic molecules.  I don't envision needing really high level calculations or calculations on whole proteins or anything.  I'd like to plot these orbital and density surfaces, view them, then show them off to my undergrad ochem class so they can see these in 3d.  Can I stay to the top/middle of the dropdown lists and still get accurate results?

Thanks for taking the time to explain to an ochem poser like me

[Mm04302011]

Essentials of Computational Chemistry: Theories and Models by Christopher J. Cramer

This book is a very excellent text on basic principle of modelling.  It does not do much deriving, but does a good job at providing the reader with pragmatic skills.  It outlines everything you asked about in a very straightforward way.  It is relatively cheap (~$50) on amazon.  I have a copy, and although I am somewhere between a "dabbler" and an "expert", I reference/read this book at least a few times per month.  Hope this helps!

[azmanam]

Thanks!  I submitted an interlibrary loan for the book through my university   I'll let you know how I do!