Hydrogen bonding in HPLC

[Mike H.]

If you had a compound that was bonding strongly to a poroshell C-18 column, how would this affect the peak shape? What additives could be used to weaken this?

Background: I'm working on a 6 compound separation and have met my goals ie. tailing, resolution etc, but one peak is broader than the rest. The peak width for it is 1 min where as the peak width of the other peaks is ~0.15min. The only substituent groups of importance are 2 ketones. What would be a good next step to reduce the broad peak? thx

[HPLC chemist]

Usually, a broad peak can be sharpened (reduce column interaction) by substituting a 'high carbon' C18 column and is more hydrophobic.

If this doesn't work you can try adding an 'ion pairing' agent like TEA (which will mask this interaction) to your mobile phase. However, an 'ion pairing' agent will permanently change the selectivity of your column (you can't wash it off) which means the column is thus dedicated.

[mattmullaney]

Hi Mike H.,

Is the column you're using a Poroshell SB-C18? This phase is nonendcapped and low in Carbon Load (8%). Going to a stationary phase with a higher carbon load and/or using an ion-pairing agent may help as per HPLC Chemist's suggestions.

There are some other possible causes for this type of behavior. The analyte may not be completely soluble in the mobile phase, the mass injected on the column of that analyte may be too great, the analyte may have multiple conformers (like captopril), or maybe this peak could be carried over from the previous infections made on the LC instrument. Also, is the mobile phase buffered? Is this method isocratic or gradient? Does the problem peak elute in the middle of the chromatogram or at the beginning or end of the running time?

Have any of these possibilities been looked into? This list is not likely to be exhaustive.

[Jahroll42]

Thanks for formular Hydrogen bond.
sbobet
maxbet m8bet
แทงบอล sbobet
ทางเข้า maxbet หวยออนไลน์ royal1688

[Mike H.]

I'm using an Poroshell 120 EC-C18 2.7µm, 150mm x 4.6mm. It's double endcapped with a 10% carbon load.

Current screening MP is 200 mmol HN4Fm pH 3.3/AC gradient. 95:5 ramp to 20:80 in 30 minutes. Peak is second in elution order at ~rt 7min.

One other thing of possible interest is that this molecule can be existing in keto and enol forms. I tried a run at column temp 60° to see if we could force one form, but it didn't seem to affect this peak.

I can't use TFA due to the pH min for this column is 2.0. Would 0.1% H3PO4 suppress hydrogen bonding, if that is the cause? thx

[mattmullaney]

Hi Mike H.,

Okay, this stationary phase is based on Eclipse Plus...with a solid core. This phase is of intermediate density as you say above, and endcapped. The analyte of particular interest is a di-ketone, just to make sure.

I am not sure that you are prohibited from using TFA altogether...I'm not saying either that TFA is "the solution" to what is going on, at this moment. This is worth reading, anyway:

http://blog.sepscience.com/separationsc ... reparation

You can always add TFA to water to afford a final pH of the eluent that is > 2 if you want. Probably less than 0.1% (v/v) level.

200mM ammonium formate/ACN in a gradient? 200mM seems to me like a high concentration--is this 20 mM and a typo?

Anyway, to me, this isn't a problem with lack of an ion-pairing agent given the info we have. If this is a di-ketone, then the presence of Lewis bases (Formate ion, for example) will encourage the formation of the keto form (tautomerism) of the di-ketone via internal hydrogen bonding, as you note above.

This reference may help a lot:

http://www.sciencedirect.com/science/ar ... 7310000890

Journal of Chromatography A, Volume 1217, Issue 12, 19 March 2010, Pages 1912-1915.

"β-Diketones: Peak shape challenges and the use of mixed-mode high-performance liquid chromatography methodology to obtain fast, high-resolution chromatography", by M.J.Rance and S.Wilson

Abstract
Historically, indirect methods have been used for the HPLC analysis of β-diketone compounds because of the very poor peak shapes and resolution obtained on conventional HPLC stationary phases. In this paper we demonstrate that it is possible to obtain good peak shapes for underivatised β-diketone compounds, in a simulated reaction mixture, using only conventional mobile phases with mixed-mode stationary phase HPLC columns. Optimum conditions were obtained using the mixed-mode reversed-phase strong anion exchange column Primesep B, supplied by SIELC Technologies, with a 0.1% aq. TFA/MeOH gradient method and a column temperature of 55 °C.

[Mike H.]

200 mmol wasn't a typo. It's a mobile phase for another compound where I had to really up the concentration to control tailing. It wasn't targeted for this specific. It was just what I had on the instrument. I probably don't need 200mmol for this separation. I was just getting a baseline.

I don't have any of the mixed mode columns in the lab, but I've been telling my boss we need to get a couple for MD.

The separation is pretty decent except for a couple minor issue.
1. The peak width of the di-ketone peak, while still broader than the other peaks, is symmetrical T=1.3. This may be as good as I can get it based on the new info you shared. ty btw.

2. The separation has 5 compounds, 3 of which have different UV maxs. I'm using a DAD for development so this isn't a problem. The problem is that one of the compounds has it's UV max at 240nm. Any gradient using NH4Fm buffer has severe UV drifting at wavelengths <250nm. So If I use 240nm as on of my wavelengths, I'm dealing with a chromatogram starting at 0mAu and ending at -200mAu. I'd really like to avoid this.

What are my options? How does the FDA view methods that would drift this much? Can I collect at a higher wavelength, along the slope of the UV curve for the peak, to get a flatter baseline?

[HPLCaddict]

If you don't need that high salt concentration in the mobile phase, I'd try to premix Eluent A as 200 mM NH4Fm:Water:ACN 20:75:5 and Eluent B as 200 mM NH4Fm:ACN 20:80. Run the gradient from 100%A to 100%B. This way, the Formate concentration is held constant at 20mM over your gradient and won't cause that severe baseline drift.
Alternative if you're using a quaternary system: Use three eluents, A: Water; B: ACN; C: 200 mM NH4Fm. Run the gradient from 85:5:10 A/B/C to 10:80:10 A/B/C (channel C held constant, so there's a constant 20mM salt over the gradient run).

[mattmullaney]

Hi Mike H.,

Wow, 200 mM ammonium acetate in water, pH 3.3. Okay...

1. I'm afraid you're correct if the stationary phase cannot be altered.

2. You're going to have baseline drift with these conditions from refractive index effects alone. Also, acetate begins to absorb strongly below 220 nm (data @ 10 mM, too!). Premixing a little of the aqueous eluent into the ACN may help with drift, but long-term it'd be better, I think, to explore a different, lesser acetate concentration.

Last I checked, you're not "locked" into using any lambda max--given the spectra of all analytes, can any wavelengths be used in compromise?

Does your LC allow for a reference wavelength? If applied with care, this can help with baseline drift. Agilent has an accurate procedure for setting this up correctly.

Which are the lambda max wavelengths of the components of interest? You'll have fewer problems if they're all > 220 nm.

I'm not sure of FDA's opinion of baseline drift, that said, they do prefer the ease of quantification...potentially drift harms this facet of things.

I'd try to re-develop the method, lowering the acetate concentration first...and require the spectra of all five peaks...and see if there's a good wavelength to use as a compromise for all of the peaks.

Let's see how others weigh in, and thank you.

[mattmullaney]

HPLCaddict is a faster typist than I am...

I concur with his suggestions.

[Mike H.]

I scrapped the 200mmol. I used it as a screen since the system was already set up for it. Changed to 20 mmol PPM pH 3.2. Elution order is the same tailing is improved ~1.0-1.3.

Under these conditions a pair of regio-isomers partially resolved. I'm at a crossroads. Previously I was unable to resolve these via hydrophobic mode reverse phase so I used a biphenyl column isocratic 0.1%TFA/MeOH condition.

It is not necessary to resolve this pair because the undesired regioisomer will be tested for 2 steps back in the synthesis, but might be useful to resolve with this method at some point. The isomer are different only in the position of a CN group on a benzene ring. Both have a COOH group. The nitrile group for one isomer is 1 position closer to the COOH group.

I know this is limited information, but does anyone have any best tries on how to target this difference for a hydrophobic separation?

[mattmullaney]

Hi Mike H.,

Please, let's see if I understand (posted chromatograms may also help). The di-ketone peak(s) (tautomers) appear more or less the same now with 20 mM NH4OAc as with 200 mM NH4OAc, both pH 3.3, with ACN in gradient separations.

Now, is it also that two regioisomers as you describe above are only partially resolved with 20 mM NH4OAc where previously they were resolved with 200 mM NH4OAc?

My guess is that to resolve these the pH of the separation will have to be explored...the pKa values of the regioisomers will be different. Likely that another method will be needed at some point, or is the biphenyl-based method useful for quantification of all five peaks? (maybe not).

Will there be a need for Quantification and Resolution to the baseline of all five peaks?

[Mike H.]

I'd submit a chromatogram if I could upload directly. Too confusing going through 3rd party picture sites.

For those interested, we have a method this will be acceptable for this stage. 0.1% H3PO4/ACN gradient. 95:5>20:80 in 30 minutes. 50°, EC-C18 poroshell column. I get good separation of 5 of the 6 peaks, K'>2 and T of 1.0 for all peaks except the co-eluting acid regio-isomers which is 1.5. For these I have a isocratic method with 0.1% TFA and MeOH on a Biphenyl column that separates the pair at R 2.5.

Thx everyone for helping me reach a solution.

[mattmullaney]

Sounds Good--Best Wishes!