molecular geometry

[RichD]

How do chemists determine the precise locations
of individual atoms within a molecule?


For atomic physics, the only thing we can observe
is spectra. Bombard the sample with radiation,
modulate the temperature, read the spectra. But
we can't micro-photograph and see the locations
of the components.

For instance, fatty acids:
https://tinyurl.com/fatty-acid00


--
Rich

[Ian Gay]

x-ray diffraction in solids. For small molecules, microwave spectroscopy
in gas phase, probably with isotopic substitutions.
--
*********** To reply by e-mail, make w single in address **************

[mrou...@shaw.ca]

Ian is correct that x-ray diffraction is one of the most satisfying techniques in terms of getting structural information. But there are alternatives. Even a simple proton NMR tells you a lot about how the atoms in a molecular are connected. Add to that fancier NMR experiments that give you atom-to-atom distances like the nuclear Overhauser effect, or experiments that give direct connectivity information like INADEQUATE (NMR people are notorious for silly acronyms), and suddenly a whole world of structural characterization opens up with the need to crystallize a compound.
Then there is mass spectroscopy. Because typical mass spec experiments break up molecules into pieces, the masses of the pieces tell you something at least about local connectivity. Chemists used to get a lot to mileage out of that, although I'm not sure how much this kind of puzzle-solving activity still features in structure determination given advances in NMR and the ready availability of X-ray diffraction equipment in most places.
And on the topic of X-ray diffraction, there are also powder methods for molecules that don't crystallize well. Though they require a lot more effort for their interpretation, they do eventually yield structural information.
Then there is neutron diffraction, only available in a few places, but great for getting information about the locations of hydrogen atoms, which typically are invisible to X-ray methods.
There are a ton of techniques. In the final analysis, it's a question of what your sample is like, and how much effort you want to invest.

[RichD]

On October 3, mrou...@shaw.ca wrote:
>> How do chemists determine the precise locations
>> of individual atoms within a molecule?
>> For atomic physics, the only thing we can observe
>> is spectra. Bombard the sample with radiation,
>> modulate the temperature, read the spectra. But
>> we can't micro-photograph and see the locations
>> of the components.
>> For instance, fatty acids:
>> https://tinyurl.com/fatty-acid00
>

> Ian is correct that x-ray diffraction is one of the most satisfying techniques in terms of
> getting structural information.

That's a special case, restricted to to crystals.

> Even a simple proton NMR tells you a lot about how the atoms in a molecular are connected.
> Add to that fancier NMR experiments that give you atom-to-atom distances like the nuclear
> Overhauser effect, or experiments that give direct connectivity information like INADEQUATE

> and suddenly a whole world of structural characterization opens up with the need to crystallize a compound.

What is the output of the NMR, how does the chemist reconstruct the structure of the molecule?
Are they solving Schrodinger's equation?

> Then there is mass spectroscopy. Because typical mass spec experiments break up molecules
> into pieces, the masses of the pieces tell you something at least about local connectivity.

Again, do they have mathematical models, using quantum mechanics?

If you look at spectra, all you get are lines at various wavelengths.
Do they obtain phase information?

--
Rich

[mrou...@shaw.ca]

On Sunday, October 3, 2021 at 4:50:22 PM UTC-6, RichD wrote:

> On October 3, M. Roussel wrote:
> >> How do chemists determine the precise locations
> >> of individual atoms within a molecule?
> >> For atomic physics, the only thing we can observe
> >> is spectra. Bombard the sample with radiation,
> >> modulate the temperature, read the spectra. But
> >> we can't micro-photograph and see the locations
> >> of the components.
> >> For instance, fatty acids:
> >> https://tinyurl.com/fatty-acid00
> >
> > Ian is correct that x-ray diffraction is one of the most satisfying techniques in terms of
> > getting structural information.
> That's a special case, restricted to to crystals.

True, but a lot of things can be crystallized, and in many cases, the structural information is exquisitely precise.

> > Even a simple proton NMR tells you a lot about how the atoms in a molecular are connected.
> > Add to that fancier NMR experiments that give you atom-to-atom distances like the nuclear
> > Overhauser effect, or experiments that give direct connectivity information like INADEQUATE
> > and suddenly a whole world of structural characterization opens up with the need to crystallize a compound.
> What is the output of the NMR, how does the chemist reconstruct the structure of the molecule?
> Are they solving Schrodinger's equation?

No. The data you get from NMR can be more or less complicated. In a proton spectrum, you get resonant frequencies, with splitting of peaks that depends on what a given atom's neighbours are (roughly speaking). Then you go through the puzzle-solving activity of figuring out what the structure must be given the splitting pattern, aided also by the knowledge that certain functional groups reproducibly show up near certain frequencies. The nuclear Overhauser effect gives you atom-to-atom distances, so again, figuring out the structure is a puzzle-solving activity (often sorted out by a computer program, which reduces the puzzle to an optimization problem). Two-dimensional experiments like INADEQUATE sort the signals along two dimensions. INADEQUATE in particular gives you connectivity information in addition to the peak splitting. More puzzles to put it together...

> > Then there is mass spectroscopy. Because typical mass spec experiments break up molecules
> > into pieces, the masses of the pieces tell you something at least about local connectivity.
> Again, do they have mathematical models, using quantum mechanics?

Nope. More puzzles to solve. If you get a piece that has the mass of a methyl group, provided you know that nothing else in your system would have that mass (methyl = 15, so it could be confused with an NH fragment in theory, although most fragmentation methods wouldn't make that), then you know that somewhere in your molecule there is a methyl group. You would typically start out knowing the empirical formula of your compound (from elemental analysis). You also know something about how molecules break up under the particular ionization conditions that your mass spec uses. Mass spec doesn't usually give you quite enough to figure out the structure of a compound, but combined with other methods (typically spectroscopic methods), you can often puzzle it out.

X-ray crystallography is very direct. NMR is, if not direct, sufficiently powerful that you can really nail down the structure if you're willing to do enough experiments. Other methods (mass spec, IR spectroscopy, etc.) tend to give more indirect information, so you often have to put several of these techniques together to get a structure.