Help Please, Urgent

[wildfish]

Chaps, I got a problem here to separate four weak acid compounds (peaks 1,2,3,4). So far I got three of them well separated but two of the peaks (peak 3 and 4) sometimes partly co-elutes under my LC coditions. Can anyone give me some suggestions in the change of gradients. My current method is as below:

Column: Hypersil Gold aQ, 150*4.6mm, 5um
Tem: 30C
flow rate: 2ml/min
mobile phase A: 0.1% Phosphoric acid
mobile phase B: Acetonitrile

Gradient program:
Time (min) %B
0 min 18%
12min 20%
18min 30%
25min 30%

[zokitano]

Hi wildfish,

From the information that you gave us, can I say that you have retention time instability for the peaks 3 and 4, when all other parameters are the same?

If so, maybe using only phosphoric acid to maintain certain pH and ionic strength is not sufficient in this case. Note that during the gradient you have decrease of the phosphoric acid content, so you're changing both the pH and the ionic strength of the mobile phase.

Have you tried to run your analysis using phosphoric buffer (let's say H2PO4-/H3PO4 for low pH mobile phase) instead of only phosphoric acid?

Good luck

[wildfish]

Hi wildfish,

From the information that you gave us, can I say that you have retention time instability for the peaks 3 and 4, when all other parameters are the same?

If so, maybe using only phosphoric acid to maintain certain pH and ionic strength is not sufficient in this case. Note that during the gradient you have decrease of the phosphoric acid content, so you're changing both the pH and the ionic strength of the mobile phase.

Have you tried to run your analysis using phosphoric buffer (let's say H2PO4-/H3PO4 for low pH mobile phase) instead of only phosphoric acid?

Good luck

Thanks Zokitano. But the RT for the two peaks (3 and 4) are not that bad. the two peaks actually are not separated very well especially with injection volume above 10uL. The problem is that if I use injection volume of 10uL or less, Peak 3 will be very hard to be quantified (two small), if I use injection volume of above 10uL, peak 3 and 4 will interfer with each other (hard to integrate). So I am wondering whether there are other ways to separate the two peaks well then I can do a big injection volume to quantify all peaks, e.g. some optimization on the gradient curve??? but how??

[Klaus I.]

Thanks Zokitano. But the RT for the two peaks (3 and 4) are not that bad. the two peaks actually are not separated very well especially with injection volume above 10uL. The problem is that if I use injection volume of 10uL or less, Peak 3 will be very hard to be quantified (two small), if I use injection volume of above 10uL, peak 3 and 4 will interfer with each other (hard to integrate). So I am wondering whether there are other ways to separate the two peaks well then I can do a big injection volume to quantify all peaks, e.g. some optimization on the gradient curve??? but how??

At this time-range you use a iscratic step (30 %B), it is possible to receive a better resolution when raising the organic content moderatly in this time range and the peaks have a sharper profile.
Have you tested the influence of the column temperature?

[zokitano]

You should try to obtain a good book for HPLC, like: Introduction to modern liquid chromatography from Snyder, Kirkland and Dolan.

Anyway, you could start one scouting gradient let's say from 10% B to 100% B (if you use buffer be careful with the buffer solubility at high %B content) in let's say 15 min. Observe the retention time of the first and the last eluting peak under these conditions. If the time difference between the retention time of the last and first eluting peak is lower or equal to 0.25 of the gradient time (in our case 15 min), than probably isocratic elution is a way to go. If retention time difference is higher or equal to 0.40 of the gradient time (15 min) than you should consider gradient elution for your analytes. If it is between 0,25 and 0.4 of gradient time, than both isocratic or gradient could be applied (should be checked).

And, to find the right gradient which will separate all your analytes is the method development part. Record the retention times of all analytes under isocratic elution of predetermined % of B (acetonitrile), let's say once with 10, then 20, 30% B and so forth. Using equations or software for prediction you could find the first gradient conditions to start implementing them as LC experiments. With fine tuning you can find your best conditions.
This is the easier way. The harder way is trial and error way (time consuming)

Also the selectivity of the separation of your acidic analytes can be optimized (changed) using different pH of the mobile phase, different organic modifiers (than acetonitrile), different column chemistry, temperature etc.

You need to understand these basic principles in order to find your best separation.

Good luck!

[Klaus I.]

You should try to obtain a good book for HPLC, like: Introduction to modern liquid chromatography from Snyder, Kirkland and Dolan.

Yes, but he is also advised to buy "High-Perfomance Gradient Elution" from Snyder and ...; It is a more convinient way to understand real gradient hplc applications.

[unmgvar]

3 things that i would check
1. like zokitano said, go for a real buffer. as you prepare this 0.1% solution how precise are you really? this can be influencing the chromatography, especially if you are in-line mixing
2. try to use 2 premix solutions, containg the start and end ratio of your A+B solutions. this can also be a factor of stability
3. your flow rate is 2ml/min. why not decrease the flow rate and use more organics? using a gradient also can improve peak shape, because it reduces diffusion effects and then you can improve the separation between your peaks

[carls]

What is the composition of your injection solvent? 10-20uL is not an excessive volume for this column format, especially if the injection solvent is the same strength or weaker (greatly prefered) than the initial mobile phase.

[sepscientologist]

Plenty of things to try. I would start at 20% and leave it til the first peak comes out
then bump it to 28% and see if the next three resolve.