Kindly note that we perform experiments for fructose dehydration into 5-HMF over heterogeneous catalysts.
For the product analysis, we used HPLC equipped with a refractive index unit (RIU)) detector and Aminex HPX-87H ion-exclusion column.
Expected products of fructose dehydration are as follow:
1. 5-HMF ( detected by HPLC)
2. Formic acid ( detected by HPLC)
3. Levulinic acid ( detected by HPLC)
4. Humins (humins have a wide range of condensed fructose or dehydrated HMF )
We have noticed from some reaction spectra the area under the curve of some peaks are very high ( which results in very high yield, sometimes higher than conversion which is not acceptable), and I couldn’t figure out the reason knowing that we have diluted the products before analysis to be in the range of calibration curve of the expected main product (5-HMF) and other side products (FA and LA) ( Attached you can find reaction network, HPLC spectra and RT of each product)
What do you mean exactly? I.e. is it in comparison to a standard curve of 5-HMF (though with RI I would not expect much differences in weight response between compounds)? Did you check all possible co-elutions (run a reaction medium blanck?
In my works, I have used the same column for the same analytes and not encountered such problems. However, what can happen easily is that the retention times shift. In this case the wrong peak is assigned to the analyte. This is especially problematic for fructose, as from this signal you calculate the conversion.
Another possibility is, as already mentioned, the co-elution. Don't only check the results table, but also how the chromatogram looks.
By the way, I have never had the need to dilute my samples, which of course eliminates an experimentator related error source.
Another point I can think of is your exoerimental procedure. Can you exclude that you "loose" analytes due to adsorption on the catalyst that you recover those analytes in a subsequent experiment without knowing?
Dear @Paul Korner
Many thanks for the hints and help.
What I meant is when I calculate fructose conversion and yield according to thier areas I got yield higher than conversion which is impossible.
And I figured out I only encounter this problem when u expect to have high 5-HMF yield. I thought about RT time shift. I will also check the spent catalyst via FTIR as you said it may adsorb analytes.
Two most likely causes: (1) This sounds like co-elution may be occurring, so the method is invalid. The cause could be something simple such as column fouling (which occurs over time, esp with higher sample loads and failure to properly wash the column after each use). (2) Insufficient sample preparation may result in other material being observed which is not well resolved by your method.
Basic troubleshooting of the HPLC system and column should be carried out to insure the pump flow rate is accurate; mobile phase is clean, made freshly each day and is unchanged; the temperature is stable and a brand new column should be substituted to see if the chromatography improves. Have someone experienced in HPLC and your specific hardware/software assist you to troubleshoot the system and method to find the cause.
As the authors stated co-elution is the first parameter that needs to be excluded and for this purpose, you may try to perform positive identification at your LC method for being sure about your HMF's peak purity.
You should spike the analytical purity HMF standard to your sample (at concentration that you encountered the problem) at least 3 different concentrations and observe that your peak is increasing linearly and calculated results were linear too. If there is co-eluted peak, non-linear peak increasing at HMF is inevitable...At the end of this test you should get appropriate recovery unless something doesn't shadow your peak.
Since you told that you are experiencing this problem at high 5-HMF yield (concentrated main product) you should be sure about your column capacity and resulting symmetrical peak shapes. At the attachment you showed, fronting peak shape caught my attention and could be the signal of your column loading capacity at its upper limit. When you exceed column capacity, asymmetrical peak shapes (typically fronting) may give unexpected (non-linear) results at higher concentrations. But as I said this peak shape could also be the sign of an impurity that co-eluted or an old column.
Emir...
- Badges
- Science method