Chromatography Forum

Hi all.

My problem is that I have negative peaks in my chromatograms.

I am using a Shodex Pb2+ column with a deashing precolumn and RI detector for wood sugars analysis. Mobile phase is water and the column is operated at 85 C and 0.6 ml/min flow rate as suggested in ASTM protocol. yesterday I injected H2O into the column to see if it is clean and discovered that I have negative peaks. These negative peaks were present in my sample-standard chromatograms but the elution times were very close to those of the sugars so it was very difficult to see them unless at very low sugar concentrations, so I didn't spot them before. As they interfere with my analysis I really need to solve this problem.

In order to find "where" these peaks come from I have changed water source, cleaned injection loop, sample syringe and containers but no improvement. I have removed the precolumn and made a H2O injection to see if the precolumn is responsible but negative peaks are there only at shorter elution times. I am almost certain these peaks are not chemicals, impurities etc. since they appear in all samples, standards, and H2O same as the mobile phase. My last thought is temperature difference between column (85 C) and detector cell (35 C, this is fixed and cannot be changed), or sample at ambient environment. ASTM states the column-detector temps should be as close as possible. Will try to check this by injecting water in column at 35-40 C or warming the water before injecting.

Any suggestions will be most welcome.
Thanks in advance and apologies for writing so much.

Cheers

My suspicion would either a temperature effect, as you suggest, or that somehow a small amount of your eluent is flowing through the reference side of your RI detector. Depends on the model, but I'd completely isolate the reference side.

Certainly the disparity in temperature between column and detector is a worry, and you could also be seeing an effect from that. One test would be to choose one compound and inject at varying concentrations and observe any change in the negative peak size, and also whether the positive peak response in linear. Then reduce the column temperature to 50C and see if the responses change, adjusting flowrate to obtain same retention time.

Please keep having fun,

Bruce Hamilton

OK Bruce thanks a lot for your reply and advice.

This is what I have done since:

I have operated the column at 35 C -same temperature as detector- and the negative peaks remain. So a temperature effect should be ruled out.

The positive peaks response is linear. In fact i get a calibration curve of R squared 0.999-0.996 depending on the compound I am analysing, which I think is quite satisfacory. *the lower r squared values of course for the analytes that elute close to the negative peaks.

The negative peak response is not dependent on analytes concentration.

Two things I noticed that surprised me though and I would like an opinion:

1. When I switch the valve to "inject" without injecting a sample into the loop, the negative peaks dissapeared. I repeated and got the same result. When I injected a wated "sample" they reapeered.

2. When I flush the detector reference cell with mobile phase I see on the baseline the same "patern" of negative peaks. It looks as if a "signature" of the reference cell being flushed with fresh mobile phase and then the base line equilibrates again. By the way the Detector is a Shimadzu RID-6A, and by looking at the manual diagram it is a Fresnel type refractometer.

Does this give any clue as to what might be the problem?

(P.S. Currentrly I am not working cause I found a microbial infection in my water-mobile phase. This water is suposedly ultrapure passed through 0.45 um membrane, active carbon, ion exchange resin, and uv lamp, and still a microbial infection occurs. From now on I will pass the water through a 0.2 um nitrocellulose membrane to remove bacteria. So if anyone reads this, take extra care of your water or aqueous solution mobile phase.

SO now I am washing the containers, changing the filters, flushing the system with metanol to disinfect, and next I will connect the lead column and pass water/ethanol through to clean it up -I think the column is clean cause I dont see any pressure build up, and I leave it in 10% ethanol in water overnight to prevent any growth. I will monitor this).

I forgot to mention that:

1. I ruled out the posibility of dissolved oxygene and CO2 causing negative peaks by degassing a water sample and injecting it.

2. I did a literature search and I have found a couple of papers with chromatograms of sugar analysis with the same method and there are negative peaks there as well...so it could be due to the method (for whatever reason )and cannot be helped. Not a very promising thought though.

Thanks again