Chromatography Forum

I'm having issues trying to obtain a 10:1 signal:noise ratio (S/N) for my LOQ solution, which is 0.1% of the working sample concentration. The current conditions are 1.5mL/min, Run time-20minutes, and 10mcL injection volume, and 50:50 ACN:Buffer. Increasing the injection volume didn't improve the S/N ratio, so what is the best reccomendation? Options available are increasing the working sample concentration, changing the mobile phase composition, or increasing the flow rate. Where do I go from here??

Are you running on HPLC, OR LC-MS.
lc-ms technique
Try to change the compostion to mobile phase 60:40 ( ACN: Buffer).
what is your ionization mode? ESI/heated Nebulizer?
In ESI if you are running on ESI check mobile phase mentioned above. and increase collision energy/reduce depend on the compound to increase ionization thus sensitivity as a result the signal to noise ratio
will increase.

It seems very strange, that if you injected more you are not seeing an increase in peak area? Might be worth just re-checking this, you might not exactly see a proportional increase in peak response from increased injection, but you should at least see some!

Other possible solutions:

Adjust detector sensitivity
Reduce sample volume (i.e. increase sample concentration)
Try to increase overall sample size (i.e. increase sample concentration)
Try gradient method to get sharper peaks (should increase s/n)
Alternative wavelengths?
Fluorescent possibilities?

Pass on a bit more detail of you overall method and the forum should be able to offer more guidance...

What is the wavelength you're running at (I'm assuming UV here)?
If it's low, you may need a fresh lamp to minimize BL noise.

Also, check the data collection rate and the time constant (or risetime or filter or whatever your detector calls it). If those are wrong, you can degrade the S/N. For Dionex UV detectors, optimum data collection rate is ~15/W (Hz) where W is the peak width in seconds at 50% for the narrowest peak of interest. The best time constant is ~W/10 (sec).

Increasing the injection volume didn't improve the S/N ratio,

Can you clarify that statement a bit? Did both the peak height and the noise stay the same, or did both increase in proportion?

Lower flow-rate should improve the sensitivity. Try 1 mL/min.

BR Danko

Lower flow-rate should improve the sensitivity. Try 1 mL/min.

BR Danko

I doubt it. LC Van Deemter plots strongly resemble "L"s.

Something is strange...

If you increase the inj vol, the signal should increase proportionately but the noise should stay the same.

Dear sdo7575:

First thing first, you should suppress the noise! Use an online degasser (or degass the mobile phase!); use a pulse-dampener; and connect a small-ID capillary to the outlet of the detector.

And what is your detector? Is your problem solved?

Alfred.

I’ll have to disagree with you DR, on this one. Sorry!
I assumed that sdo7575 used a kind of a standard analytical column e.g. 4.6 mm inner diameter, packed with 5 μm particles. The flow optimum would be then between 0.8 and 1.0 mL/min.
In fact I can’t think of a standard analytical column that will display efficiency optimum at 1.5 mL/min flow rate. I forgot one ting though: The retention time should be adjusted to the original one by adding some more ACN in the mobile phase.
Another thing is, as Albany-12303 points out, it makes no sense, if larger injection volume hence more analyte loaded, doesn’t help. It makes me think, that the noise might be measured around the injection noise and this would be a reasonable explanation. The noise should be determined around the same place as the analyte elutes, but on injection of a blank sample. Maybe there is some carry-over and then I would understand the problem. But anyway, I bet the sensitivity will improve if the flow rate is set to 1 mL/min and the retention time is adjusted to the “normalâ€

Agreed, that sensitivity will improve if flow rate is optimized, but the improvement will be, at best, very small.

I had assumed that noise was being measured appropriately - maybe it's time for sdo7575 to elaborate on the S/N measurement technique being employed.

(I'm not addressing the "inject more sample" argument for the benefit of those who may not have that option - eg: some stability indicating methods).