generally, excursions from 'baseline' (whether negative or positive excursions) on a refractive index detector indicate the elution of a material. I have no sense of the column or conditions, so I can only make guesses. (Among the many guesses is that a chiral separation might be taking place.)

Generally speaking, the detector is reporting an elution event (or events),  and your desire to remove it (or them) may be an erroneous approach to interpreting the data.

I bought a new column (Zorbax carbohydrate analysis) a week ago. When I had removed the column, negative peaks were not present (like other positive peaks, except t0). Standards were dissolved in water.

But when standards were dissolved in a mobile phase and the column was placed back, there were no negative peaks after t0, but negative peaks after carbohydrate peaks remained the same.

When blank(mobile phase) was injected, there were no negative peaks at all.

How it is possible that a new column material dissolves, when the chromatographic parameters are the same as in a column certificate?

The negative peaks at one column volume come from compositional differences in the instrument solvent and the sample solvent. In addition to retention effects, size exclusion effects can be sometimes seen and used to chromatographic advantage, depending on the pore size and the molecular weight of the sample. In those cases, material elutes before one column volume has been pumped through the column.

As Scot Abott says, if you inject any solution different from mobile phase in a system with RI  detector you will have a peak, unless if this solution has exactly the same RI index of the mobile phase. If it is negative or positive depend on the Ri index of the injected solution in relation to the RI index of the mobile phase. If samples and standards are dissolved in mobile phase prior to injection, these additional ghost peaks will probably disappear, since there will be no difference between solutions other than the substances present in samples and standards and their true peaks. 

hello,

Have you a correct baseline ?

this can be a technical problem with the inlet valve that ensures redirection flow to the cell messure or reference .

a little leak  will let pass your compound to the reference cell, and cause negative  peak after peak..

You don't specify the retention time of the negative peak : is it just after the positive peak ? Indeed, i already have this kind of problem, and the problem came from the reference cell that was not closed during the run. Additionnaly, i don't know which kind of RI detector you have. 

> pay attention that the reference cell is well closed (though the dedicated button)

> maybe there is a pb in the reference cell that is not well closed (leak of the eluent in it)

Thank you all for your answers. I've sent you some chromatograms (standard solution and blank).

Dr. Abbott and Dr. Junior, as you can see, when I inject blank (water), there is no negative peaks, except t0. When I inject standart solution (carbohydrates are dissolved in mobile phase), t0 negative peak is gone (not completely but still), but the other peaks remain the same as if the standard were dissolved in water)

Dr Merlier, Ive found a leak. After purging, I've noticed that some part of MP is pushed through the ref. cell.

Dr. Richard, as you can see, the negative peak (after carbohydrate peak) goes after 1,5 minutes (or 1 column volume).  The detector is: RID Waters 2410.

there is a big time lag between a sugar peak and the negative peak

indded, if there is a small leak towards the reference cell, the time lag will be important since the flow rate in the ref cell is low

But i don't know how to correct this problem ...

May be there is a problem with equilibration of a column? It could explain one column volume lags and negative peaks (ACN n(20)/D = 1.344, but H2O n(20)/D = 1.333).

Also there were no negative peaks, when I'd removed the column(mobile phase 25:75 (H2O/ACN), flow rate 1,4ml/min and standards were dissolved in water)

and when I used H2O instead of Mobile phase (flow rate 0,5 ml/min)

A long shot: is everything at the same temperature? I mean mobile phase, samples and standards?

Dr Junior,

Only a column and RID are thermostated, but samples and standards are not (room temp). i work with manual injector, and injection volume is 5 ul.. But still baseline was stable when I had injected blank.  

Thank you all for your answers.

I've noticed that a leak in the pressure relief valve was the major problem. When the flow was split and the flow rate (that comes to detector) was decreased twofold, there was no leak and no negative peak. There is no leak when flow rate smaller than 0,7 ml/min.

Thank you so much for these clarification

am asking that this is the first time happened to me and I check there is no any leak and day before I tried with no negative but today I find it , am wondering there is some problem with column or other factor like mobile ?

How is the system plumbed? Where is there a pressure relief valve? Refractive index detectors are flow sensitive, so there should be no flow branching from pump to detector.