How would one define a peak in chromatography and how is it different than an anomalie?

Answer 1

Yet again, I wish people would give more information in their question: what do you mean by "an anomalie" which is I assume is misspelled "anomaly"?

I will assume you mean either 1) how can you differentiate between the peak you want versus something you don't want (such as a side-product in a reaction), OR 2) how can you differentiate between a real peak due to a real compound versus problems--things stuck on the column that are slowly eluting out, signals due to background buffers, etc.

For 1) If you want to know whether a given peak is something that you want or not, the only way to know for sure is to use a different technique--NMR, MS, etc. This is why many chromatography-type setups are connected with mass spectrometers: GC-MS, LC-MS, CE-MS, etc. If you don't have access to something like that, you can compare the elution time to a known standard (for example if you are purifying a mixture of compounds but already have some of the purified material, you can run it under the same conditions, and if the peak in the mixture comes out at the same time it's probably the same thing).

2) Many problems can occur with chromatography; salts/buffers in solution can absorb at the wavelength you're using to monitor, there may be stuff adsorbed on the column from previous users that will come off when you use it, and depending on the system you may have a setup where electrical spikes can result in things that look like peaks.

Some solutions: basically the peak shape is your big clue.
* If buffers in the solution you injected are absorbing, you will often see a huge peak right in the beginning of your run that will rapidly go away, as the liquid in the system contains less and less of your injected sample and more of the eluent.
* If there is stuff adsorbed on the column, it will typically appear as a really broad peak. The way to ensure that this is not a problem is to do a blank run before you do your actual run. Setup the exact routine you plan to use, and run it after injecting nothing but solvent. If you see any peaks, you know they're junk coming off the column, in which case you need to clean it.
* If you work in an old building that does not have independent electrical circuits, sometimes a big electricity-using machine coming (a fridge, a laser, a spectrometer) on can give you a spike in your chromatogram--this peak will be very abnormally narrow in time, much much narrower than any real peak. If this is a problem try working after 5pm or so when a lot of people have left the building and big machines are less likely to be used.

Lastly, if you're trying to figure out a signal-to-noise problem, a) a "real" peak is usually 2 or 3 times the noise level, although people who've stared at this stuff long enough can often pick out real peaks at less than that level, again the "shape" of the peak is the giveaway.

If you want more information please clarify your question. Good luck in solving the problem.

Answer 2

Generally a peak is defined as a rise in signal that is two times greater than the signal noise.