> Hello,
> I'm an amateur scientist in Denver, CO, looking to expand my lab a
> bit. I would like to be able identify different compounds in a
> solution of (whatever), and have a computer identify the ingredients.
>
> I *assume* I need:
> 1) a spectrophotometer with a computer interface and software
> 2) a way to seperate the individual compounds using chromatography
> (HPLC?)
> 3) a database to identify the individual spectra
> 4) I'm not sure what the most useful spectra are - IR / Vis / UV?
> 5) Are there any free computerized databases out there?
>
> This is brand new territory for me. What used system, which I might
> find on Ebay would be the best 'bang for the buck'?
>
> Thanks for your advice!
> Best regards,
> -Mark


I solve this problem using thin-layer chromatography combined with an old
grating IR instrument.

I can think of these advantages in this approach:

1) The chromatography is cheap and easy requiring very little equipment:

- Jars
- Solvents (certainly much less than HPLC)
- UV lamp
- Visualizing agents

2) You can find references for Rf values, developing systems, and
visualizing agents.
(It certainly helps if you know what your looking for.) TLC plates
show up on
Ebay all the time.

3) It's possible to make TLC semi-quantitative using some sort of
densitometry
....for example a computer scanner and freeware software such as Scion
Image.
(In the process, you'll learn a bit of image analysis.)

It's also possible to pick up instruments designed for TLC scanning on Ebay.
e.g.

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=26230&item=7503575157&rd=\
1&ssPageName=WDVW

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=26230&item=7503599058&rd=\
1

4) You can do preprative TLC or column chromatography to isolate compounds.
From these isolations, you can then collect other important characterization
data such
as melting points etc.

You can do centrifugal TLC for preprative work using an instrument such as
this:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=48720&item=7503488286&rd=\
1&ssPageName=WDVW

Be careful when buying any IR or UV spectrophotometer. I would certainly
try and
accertain whether the source works in any IR or UV instrument or UV HPLC
detector as
the replacement of them can be very costly.....if you can find the lamp.

You can get single wavelength detectors for HPLC pretty cheap but again,
make sure the UV source works.

A simple HPLC pump and injector can certainly be put together. If you run
in isocratic
mode you could recycle the mobile phase easily. The chromatography will be
less versatile (powerful).

Software for matching UV/VIS spectra exists but there's some significant
limitation in what you can identify from a UV/VIS
spectrum. Any molecule with a particular chromophore(s) is going to give
essentially the same UV/VIS spectrum.

Outside mass spec, I tend to think IR spectroscopy gives you the largest
information content concerning unknown
identification. It's possible to build your own IR libraries and code the
software to do library searches, but it would be
a big project. It's probably better just to use your head and 'interprete'
the IR spectra. Of course getting a good
IR spectrum depends on your success in isolation of compounds via TLC etc.

What you're wishing for in terms of having an instrument identify all the
unknowns in a mixture is probably best met by a
GC/MS, but what are the chances of actually coming up with such an
instrument via Ebay cheaply? I think slim.

My advice is to consider TLC and a good dose of "old school" organic
chemistry thinking. Do some research
into the techniques of the organic experiementalists before the advent of
modern computerized instrumentation.

Take to heart these words by Albert Hoffman:

"For structure determination, which today can be conducted rapidly and
elegantly with the help of spectroscopic
methods (UV, IR, NMR) and X-ray crystallography, we had to rely, in the
first fundamental ergot studies,
entirely on the old laborious methods of chemical degradation and
derivatization."

LSD, My Problem Child p.10