Viscosity in HPLC SEC with Polysaccharides samples

[kurian]

Hi,

I've samples injected at 200 µg on TSK5000PWxl 7.8 x 300 mm at a flow 750 µl/min, elution is PO4 50mM NaCl 200mM pH 7
Samples are polysaccharides and is viscosity is high (i think so)

System is a Merck-Hitachi EZ-Chrom configuration.
We have profile not very equal between injections but reproductibility is ok (RSD about 3% for 2 injections for each 3 batch columns)

On another system HPLC (Waters maybe, but i don't know because it's another service), the reproductibility is very poor with 20% RSD between 2 injections on a same sample...
Profile show difference after the apex.
If we injected 100 µg, we have no problem, a good gaussian curve and good reproductibility.

I think it's the viscosity sample and difference between dead volume system the cause of difference between good reproductibility on Merck and poor on Waters (or Agilent...)

Someone have another idea ?

[sam.pedraglio]

I know, it's a very stupid question but, did you use the same column on the second instrument?
In my knowledge different dead volumes influence the separation, if you work in gradient mode, no experience regarding the viscosity.

[Uwe Neue]

With a highly concentrated sample, you may get precipitation/redissolution. For the assay of polysaccharides, you do not need a background electrolyte. As a matter of fact, it could be counterproductive. Any reason why you use the buffer?

[kurian]

It's the same column in the other instrument, yes.

Reason for buffer at saline concentration is use to stop polysaccharides adsorption.

[Uwe Neue]

Unless your polysaccharide carries charges, you do not need salt nor buffer. Unfortunately, people lump all kinds of things under the name "polysaccharides", so I can't tell what you are doing, or what you need to run your sample.

[kurian]

The sample is without elution buffer before injection.

But when the system inject, the sample is dilute with elution buffer before it happens to column head (i think).

The reason i think PS is more diluted by elution buffer on my system is lenght between injector and head of the column.
Thus the PS viscosity is less than PS viscosity on a system with a short dead volume.

Don't you think ?

[HW Mueller]

If your polysaccharide really needs a certain pH and ionic strength to prevent absorption than one would think that there is an adsorption difference between the two systems, due to different mixing. You already have one way to get around the problem, the other is changing the sample solvent.

[Uwe Neue]

If the polysaccharide is indeed not charged, the use of buffer and salt is counterproductive.

[kurian]

I have understand, thanks.
But if i use a buffer or Water, sample dilution before column head is the same.

[kurian]

My real question is :
Is it possible PS sample has a high viscosity and it's more diluted on my system with high dead volume than another system with poor dead volume ?

We have the solution : inject 100 µg instead of 200 µg
But i would like to understand why there is so much difference between my system and another...

[Uwe Neue]

Of course, the sample dilution will depend on the pre-column bandspreading of the system, which in turn depends on the precolumn volumes, such as the volume of the tubing.

[HW Mueller]

So it seems if viscosity played a role the bandspreading, that is peak broadening, etc., would be worse in the system with the higher dead volume.
Also I wonder whether the area of the peaks generated by the two systems is the same and also if the start of the peaks are at the same rt.

[danko]

Kurian,

Would you please check whether the injector loops are ≥ 200 µL on both systems?
Also, there could be some significant differences between the one system’s injection rate compared to the others ditto – it’s worth checking this setting.
Finally if you showed us chromatograms illustrating the case, it would be much easier for everybody to conceive and offer an explanation – potentially solution.

Best Regards

P.S. I don’t believe too much in differences in pre-column dilution, caused by differences in the system volume. Besides, system volume can be determined quite easily and if you find huge differences you could perhaps reduce them – just for the sake of having less trouble running gradient, if nothing else. I might be wrong, but I think the problem is to be found another place.

[kurian]

here are the samples in two same concentrations injections in different batch column..
Two injections at 2 mg/ml and 2 at 1 mg/ml

We see a perfect profile at 1 mg/ml

The two first profiles are Dextran at 3800 KDa

[danko]

Hi again,

It looks to me you’re overloading the column. I assume the shown data is generated on the system that didn’t perform as you wished.
Maybe you should show data from the system that performed according to your requirements (i.e. no shouldering).
I think we’ll need the values on the Y-axis as well.

Best Regards

P.S. Did you check the loop volumes on the two systems?