HPLC methods Development for basic compounds

[dcobice]

hi, Im trying to development a new short gradient platform to cuantitate basics compunds.The old platform was with water and methanol + 0.1% of TFA and the gradient began with 70% of water to 80% of methanol and I used a ODS-3 3um HPLC columns.The total run was 36 minutes with a 10 minutes for reequilibration.Is someone have experience whit monolithics columns?

Thanks in advance!!!

[tom jupille]

If you're dealing with "small" molecules, that gradient seems like it might be quite shallow. What column dimensions and flow rate are you running?

One way of characterizing gradient separations is by the average k' value of peaks during the gradient (k*). This can be used as a surrogate for isocratic k' in predicting resolution, which means that resolution is proportional to [k*/(1+k*)]. That, in turn, suggests that, in general, there is little to be gained by running at k* values above 5 or so (if selectivity changes, then all bets are off!).

k* can't be measured directly from a single chromatogram, but you can estimate it from the separation conditions:

k* ≈ (tG/ΔΦ)(F/Vm)(1/S)

where:

• tG is the gradient time
  ΔΦ is the change in strong solvent composition, expressed as a decimal
  F is the flow rate
  Vm is the internal volume of the column
  S is the slope of the isocratic log(k') vs Φ plot for your analyte. For small molecules, S = 5 is a reasonably safe assumption.

You can turn that equation around and solve for gradient time:

tG ≈ k* S ΔΦ Vm / F

I'll assume a 100 X 4.6-mm column (for which Vm would be about 1 mL) and a flow of 2 mL/min. That suggests a gradient time of:

5 * 5 * 0.5 * 1 / 2 ≈ 6 minutes.

Even at a flow rate of 1 mL/min, you could get away with a significantly shorter (12 minute) gradient time that you are now using.

[dcobice]

hi, thank for you reply.... sorry I forgot to told you about the column dimension.The column was 150mm x 4.6mm x 3um and the flow rate was 1.0 ml/min.
The issue was that the first peak was at about 3 minutes,and the last peak show at 30 minutes retention time at the old platform.
I dont Known if it is posible to run a short gradient profile whit another column for example ( monolithics columns) perhaps I will reduce the runing time!!!

Thank very much for your help!!!!

Kinds regards!!

[tom jupille]

You could cut the gradient time in half (to about 18 minutes) on your present column and still have reasonable k*.

[dcobice]

thank you very much for your reply I trying to cut the gradient time at 18.00 minutes and I will compare the profiles in order to view changes in selectivity or relutions between peaks!!!

Thanks again!

[tom jupille]

If you see selectivity changes (relative positions of peaks changing) when you change the gradient time, then there may well be an "optimum" time. If so, you can figure out what it is either by successive approximations or by using appropriate modeling softeware.

[dcobice]

hi, thanks again for your reply.Do you know about software of HPLC simulation.... because I havent any in Argentina... Do you recomend me any???
Another topic. I whish I could begin my thesis in monolithic columns .. Do you know about papers or literature of these???
My enfasis is to aply these columns to optimize the time of the large gradients profiles in phamaceuticals products.

Thanks again in advance!!!

regards!!

[tom jupille]

The longest-established chromatography modeling program is called DryLab. It is developed and supported by Rheodyne:

http://www.rheodyne.com/products/chroma ... /index.asp

I am, by definition, biased because I spent a decade and a half supporting and helping to develop the program, but it really works quite well.

I just did a Google search on the keywords HPLC and monolith which returned over 900 hits. Two innovators in this area are Frantisec Svec and Nabuo Tanaka. Look for papers with either as a coauthor.

[dcobice]

hi, Tom thanks very much for your reply!!! i try to buy these software to improve my changes in the chromatography profile.
Then I will find something about monolithics columns in the net!!!

Thank again!!

[Einar Ponten]

This is a typically problem at pharma companies.

When it becomes too tricky we know that several users are very happy about (ZIC®)-HILIC for the analytes that elutes close to the void in any RP setup.

[pipettemonkey]

[quote]
S is the slope of the isocratic log(k') plot for your analyte. For small molecules, S = 5 is a reasonably safe assumption . [ /quote]


For ionizable compounds, will the slope S differ much depending on the ionization of the compound?

[tom jupille]

Yes, it can vary quite a bit; off-the-cuff estimates for S are probably accurate within a factor of 2 (i.e., saying "a typical S value is 5" actually means "we're pretty sure it's between 3 and 10).

For the present purpose, though, the object is to run under conditions that give "reasonable" k* values. Since gradient k* is the analog of isocratic k', "reasonable" means the same thing: somewhere between 2 and 10 as a starting point.

If different analytes have different S values, then you will see changes in peak spacing as a function of gradient steepness, and there may well be some optimum steepness.

[dcobice]

S is the slope of the isocratic log(k') plot for your analyte. For small molecules, S = 5 is a reasonably safe assumption . [ /quote]


For ionizable compounds, will the slope S differ much depending on the ionization of the compound?

Yes Its OK I understand I have 10 years of experience in chromatography , but Its a quite complicated because when I Started the gradient profile the first peak was at 3.25 minutes and the last peak was at 40 minutes I try to do all the changes to resolved my 15 peaks in 20 minutes but its imposible.I changes columns stationary phases with a lot of vendors, MS columns and the total run of the pfrofiles did not significants values.I started whit 80% of water and finished with 80% of methanol with 0.1% of TFA.I used a 5 slope in my gradient profile because a gretear values increase the slope and its imposible to cuantitate them!!
Thanks!

[Uwe Neue]

I looked at your separation from a different angle, i.e. the peak capacity that you have. In principle, it is fairly well optimized, and if you want to cut the run time in half, you will also reduce the peak capacity, i.e. the overall resolving power of your gradient. The fact that you have played with several columns and did not get anywhere indicates that this judgement might be correct.

However, there may be another way to skin the cat. You are using a methanol gradient. The viscosity maximum is exactly in the middle of this range. If you were to use an acetonitrile gradient instead of methanol, you will gain a good 50% in peak width, which may alow you to reduce the gradient time by a factor of 2. The general problem with this approach is that the elution order might get scrambled, but you did not appear to be bothered by this since you used different column chemistries.

I would try to start the gradient a bit earlier with acetonitrile, maybe around 20%, and go to 70% as a start. Since you are shooting about half the run time, I would use a 20 minute or 15 minute run time. Then run a gradient at 1 mL/min over the run time, and then increase the flow rate, keeping the run time constant. You will get selectivity changes from the switch to acetonitrile and from the changes in the gradient profile due to flow rate changes, but you may find a reasonable separation somewhere in between.

Anyway, this is a different thought process than what you have tried until now.

[PavelP]

Did you try to involve a second organic modifier into your gradient? Ethanol or IPA at low concentrations reduces quite significantly the retention times and, if you're lucky ( ), even without drastic changes in resolution.