Chromatography Forum

Thanks EVERYONE for your input into my issues ~ as you all know I'm a rookie at this ~

I'm using a Shimadzu HPLC system and it is relatively old. Therefore, I have to manually degas the mobile phases and manually inject my samples (I need to turn a knob). The ELSD detector was bought a few months ago just before I started using it.

So now I should really be making my standards in atleast 50:50 ACN/DI right. And I can try increasing this to 90:10 ACN/DI if it shows no improvement? Carl, you say buffers are highly recommended for HILIC seperations, how can I input this into my method?

Also thank Labdog, zokitano, lmh, and hollow.

Hi everyone

I have done some of the things you guys mentioned but it hasn't improved much ~

At first, I made up a solution of glucose and galactose with 50/50 ACN/Water. I injected this into the column which was running at 80/20 ACN/Water and it showed no peaks at all.
So I tried increasing the ACN in the mobile phase to 95/5 ACN/Water and I ran that for 30 minutes. Injected the same samples and again showed no peaks at all.

Now, I tried adding MeOH into the mobile phase as you mentioned, running at 95/3/2 ACN/MeOH/Water and I left that running for another 30 minutes. I injected a sample of galactose and a peak came out at 5.5 min (running at 1mL/min). I said to myself "good, atleast theres a peak!". After that I injected a sample of glucose and, guess what, a peak came out at 5.5 min! The peak shape looks nearly identical to that of galactose.

I then had time to change the column to a shorter amino column and see how that goes, but the retention time for both sugars where the same, around 1.5 mins.


I don't know what else I can do now ~ any thoughts?

Thanks a million
Wilfred

did you verify that your system is working well, in means of flow rate, leaks etc?

if 5.5 min is your void time with 1ml/min with this column, I guess there is something else wrong. (I would expect a t0 of somewhere between 2.5-3.0 min to be fair)

- Verify the flow by using a volumetric flask and stop watch. (simple take the effluent after the RI)

- is your system working properly in reversed phase mode?
e.g. normaly I use a test mix of aceto and butyrophenon and a void marker such as thiourea or uracil. Works +/- on any C18 with MeOH 60°/Water40%, Runtime about 7xt0.

- are your samples of big enough concentrations? try 10-20 g/L or more.

- is your RI working well? do you purge your refernce cell well prior to the injection?

- try sugars which differs more than glucose and galactose, e.g. use glucose and sucrose in this step of the troubleshooting
acc. to the Phenomenex note, they should elute somewhere around 10 and 16 min, on your column, with a flow of 1 ml/min, 40°C, ACN/H2O 80/20.

Hi everyone

At first, I made up a solution of glucose and galactose with 50/50 ACN/Water. I injected this into the column which was running at 80/20 ACN/Water and it showed no peaks at all.
So I tried increasing the ACN in the mobile phase to 95/5 ACN/Water and I ran that for 30 minutes. Injected the same samples and again showed no peaks at all.

Now, I tried adding MeOH into the mobile phase as you mentioned, running at 95/3/2 ACN/MeOH/Water and I left that running for another 30 minutes. I injected a sample of galactose and a peak came out at 5.5 min (running at 1mL/min). I said to myself "good, atleast theres a peak!". After that I injected a sample of glucose and, guess what, a peak came out at 5.5 min! The peak shape looks nearly identical to that of galactose.

I then had time to change the column to a shorter amino column and see how that goes, but the retention time for both sugars where the same, around 1.5 mins.


I don't know what else I can do now ~ any thoughts?

Thanks a million
Wilfred

Did you use 1mL/min for all of the above analyses? Remember the void volume of this column is ~2.9mL so run times will be long at 1mL/min (k = 3 at RT = 11.6min).

Since you did not see peaks at 80/20 the analytes were retained too strongly at this mobile phase composition. In this case you need make the mobile phase stronger which in HILIC mode means you need more water.

Try the separation again at 75/25 ACN/water with a run time of 30 minutes if using 1mL/min. You can reduce the run time to 15 min at 2mL/min.

Are you running the column at 40C?

For this experiment, make separate standards of glucose and sucrose dissolved in 75/25 ACN/H2O.

Regarding your question about buffer, it is generally not needed in sugar separations since sugars are non-ionizable.

lack of selectivity is one thing, but no peak at all another...

- maybe you have to forgive me this question, but when you don't see any peaks at all, is the RI in the correct mode or possibly still purging the reference cell? I don't know your system, but if you have to do manual injections, your RI is still of a generation which needs to be switched between ref- and analytical-cell by hand?

- obey Carls advice and always pre-mix your mobile phase and run 100% on one channel when working with RI. Otherwise you can get a really noisy baseline that let your peaks disappear in the noise.

- maybe Carl is also right with this, and your analytes are too strongly retained with 80/20 (but you have seen peaks once, and fru or glc should elute within k=3). So I'm not that confident about this.
But you can try stronger eluents as well ("don't exclude too many things before you've tried them...").
So, for a systematic approach you could start with e.g. 60/40 and decrease the % of water by 10 or 5 % steps. Always equilibrate the column and the RI-reference cell well (use at least 30-40 ml for this). At some conditions, you'll hopefully see the retention starts to increase significantely.

- if you're using high acetonitrile content with your "samples", make sure that your sugars keeps dissolved and aren't precipitated (and filtered off prior to injection...). Go as high as possible with the organic, but not more
Maybe it's better to inject less of a more concentrated sample.

As I've written in my last post, I strongly recommend to have your system checked in means of the flowrate, injector (rotor and loop), and detector. Try to reproduce some other, simple application, that is/was normaly run on this system.

Just that I'm writing this:
Did you use on-line mixing in all experiments so far? Then I could also think of a faulty proportioning valve!
(BTW: do you have a high- or low-pressure gradient system?)
->> pre-mix eluent and use only one channel.

(Verify the gradient accuracy, before you start gradients with ELSD.)

Thanks guys for your feedback

Hollow: I will try measure and verify the flow rate tomorrow and see if there is anything wrong with the flow rate. And yes the system is in revered phase mode and the suagr stds I am using now is 1g diluted to 100mL with 50/50 ANC/Water. I'm new with the RI and there isn't a 'purge' button on the RI but I do purge the pump whenever I change mobile phasesto get rid of all the bubbles etc. Theres a 'balance' button on the RI which I have to press occasionally as sometimes the RI goes off the scale for some reason and I have to balance it or zero it.

Carls: Yes I used 1ml/min for all the analyses I did. I will try using 75/25 tomorrow as well and make glucose and sucrose stds with the same conditions. I did try incubating the column at 40C but it didn't really show any effect so I turned it off. Maybe I should keep it at 40C? I will let you know how this goes tomorrow ~ FIngers X!

Hollow: There is a 'reference' button that I can press on the RI. The technicians don't even know how to work these machines, neither does my supervisor really. This is how it runs here at this University. We figure out everything by ourselves which is way I am asking you all for help. And I do mix my mobile phases by hand and use one channel. I will give 60/40 a go, then increase it like you mentioned. And maybe I need to try a more concentrated sample. I did use an on-line mixing in most of the experiments but for the 80/20 they were mixed by hand and fed in one channel.

Thanks and will keep you all posted!

There is a 'reference' button that I can press on the RI. The technicians don't even know how to work these machines, neither does my supervisor really. This is how it runs here at this University. We figure out everything by ourselves which is way I am asking you all for help

hm, ok.
do you have or can you get the manual of your RI? maybe someone of this forum can provide it, if you post the manufacture and model? Or just google for it and/or ask manufacturer.

Basically, a RI detector consists of two cells, one reference and the analytical one. The signal will then be measured as the difference in refraction index of the solvents in these two cells.
Therefore the reference cell has to be filled with your eluent before injecting a sample. During the run, the reference cell is "closed" (excluded from the flow path), that's why RI doesn't work with gradients.
So to fill the reference cell, one has to switch between these two cells either manually or by software (I guess you have to do it manually).

Question: When you push the "Reference Button", is there any inidicator (Display or LED) which indicates the actual selection? hopefully there is... otherwise you have to figure it out how to switch between the two cells yourself.

Without knowing any details of your detector, I would proceed like followed (every time you set up a new mobile phase or start a new sequence) :
- set up your mobile phase, premixed, one channel 100% when using RI, degas it well (if possible by Helium sparging)
- purge pump as usual, then connect the column
- while equilibrating your column with the eluent (at least 30-40 ml with HILIC on your 4.6x250 column), select the reference cell (to purge and fill it with the mobile phase).
- after that, switch to the measuring cell, let it also equilibrate (maybe 5-10 minutes) until the baseline is stable.
- reset the baseline to "zero" with the "Balance bttn". (balance before each injection (if there is no "autozero" function with the software)
- make your injection(s)

Beside this, RI is also sensitve to temperature differences, so it is advisable to keep your column at the same temperature as is your RI-detector. If you can thermostate your RI (maybe there is also a "temp bttn"?), I would use something of 35 or 40°C for both, RI and column. Otherwise I would not heat the column.
Power on your RI detector at least half a day (better overnight) before starting your experiments, so every part is well temperature equilibrated. Leave it powered on if you plan to continue your work the other days or so.

A conc of 10g/L should be sufficent to see some peaks.

Nevertheless, still do the check for flow the rate, so you're sure that there is no other problem hidden.

...and finaly, if you manage to bring your machine to work, write it down in a short SOP. Any of your successors (as well as your supervisor) will be deeply grateful for it

for an RI you must "balance" between the reference and sample cell
so you must purge the reference cell with your mobile phase before you start analysing

you said in the beginning that you have a RI coupled with an ELSD
i advise you to use either one of them because otherwise you are limited not using gradients with the RI
but alone you can use the ELSD and do gradients for better separation.
for ELSD you need to have the nebuliser gas generally N2, extra cost for better separation and sensitivity

you should give the entire models names and parts of your system.
so that those that know them can give you more details on how to set them up
what is the software as well?