Chromatography Forum

Optimizing to me means you vary parameters first in relatively large steps then in small ones for fine tuning, checking peak shape and separation, sometimes sensitivity. If I don´t know too much about the compound I might start with pH 2, check the behavior at pH 8, then zero in on the best pH. If one has an optimized standard an adjustment might become necessary when the final matrix is analyzed. By that time I usually knew enough about the compound´s behavior that this was not very difficult.
The reason I asked is that I like to get a picture about how column preferences, parameter preferences, are established.

HW Mueller,

On the past mostly I used C8 column with acetonitrile and phosphate buffer as mobile phase, running gradient elution since isocratic wasn't effective and efficient to separate two analytes.

I tried to use pH 3.0. Then I used pH in the mid region of two pKas (pH = 6.6). Actually I've already got optimum method on both pH condition. Unfortunately there was a serious problem. On each condition (blank injection = 0 uL) there was a ghost peak appeared exactly at the retention time of acid compound. I've changed the composition of mobile phase and even tried to change the column with C18 and C12 but that ghost peak still appeared. I think this wasn't a "carry over" since the ghost peak also appeared even at the first blank injection.

How could it be?

Finally I prefer to use isocratic and I think CN column could give more effective and efficient separation for those two analytes.

How about using MeOH instead of ACN?

We always try acid, neutral and basic pH on 1 HPLC column, with ACN, running 2 gradients. If we already got our separation in these runs, we optimise the separation, otherwise, we change to MeOH.

If we really don't get any separation, we might consider a column change, but even then a CN is not our first choise.

My point: if you can avoid CN-columns: avoid them as much as possible.

Go find the source of the ghost peaks instead of using a (temporary) solution where you don't see these peaks

Kind regards

Ace

OK, so you did not optimize pH, but what did you optimize at the two pH you mentioned? Ionic strength, type of organic modifyer, amount of org. modifyer, flow rate, temp. gradient? I assume that the "ghost" is due to a gradient.
That brings me to another question: The gradient separated better than isocratic?
If I read you correctly than you were stuck on the CN column also, until someone suggested to use another pH (was it 4.5 instead of 3 or 6.6?).
In other words, I still wonder how you landed on a CN column.

Ace,

I think ACN was a better choice than MeOH because MeOH has lower solvent strength than ACN whereas the analyte (especially acid compound) was retained quite strongly in the column. Besides high percentage of MeOH would result high backpressure, so I didn't consider to use MeOH.

What pH do you usually use on each acid, neutral and basic condition? And I wonder why you don't prefer to use CN column? Is it because CN is more susceptible to get loss of its bonded stationary phase or are there any reasons?

HW Mueller,

I think you would suggest that we have to choose a proper pH first before determine other parameters. What I've done were I choosed one pH, then optimized mobile phase composition, flow rate, and temperature. When this wasn't optimum yet, I started to use another pH. That was why I've already tried two pH conditions.

The analytes have different polarity in certain pH condition hence gradient was better than isocratic. Furthermore I think CN has quite different polarity than other RP columns (C18 and C8) so I assumed that it would give a better separation for two analytes those have great variation on pKa (4.4 and 9.5). When using CN, I thought pH 6.6 was a proper one before it showed unstable retention time (decrease). Finally I prefer to use pH 5.4 (with acetate buffer) as suggested by Liu.

I'd like to hear other consideration from all of you to obtain a better method

Thanks for all

Basic HPLC method development technique:

Start at low pH = 2.5 - should minimize silanols, increase retention of acidic compounds and improve peak shape of basic compounds.

Should give you the best robust method and preserve column lifetime.

Basic HPLC method development technique:

Start at low pH = 2.5 - should minimize silanols, increase retention of acidic compounds and improve peak shape of basic compounds.

Should give you the best robust method and preserve column lifetime.

Tina choose the pH based on the pKas of the analytes. The analytes are basic and acid (see starter post). I remember that Dr Neue suggested to use pH in the mid of the pKas for mobile phase in this forum (but, I do not remember the title of the topic). I think this is the reason why she uses pH 6.6 in her method.

Ace,

I think ACN was a better choice than MeOH because MeOH has lower solvent strength than ACN whereas the analyte (especially acid compound) was retained quite strongly in the column. Besides high percentage of MeOH would result high backpressure, so I didn't consider to use MeOH.

What pH do you usually use on each acid, neutral and basic condition? And I wonder why you don't prefer to use CN column? Is it because CN is more susceptible to get loss of its bonded stationary phase or are there any reasons?

Oké, so you really considered using MetOH.
pH acid = 0.1% phosphoric acid.
pH neutral = 0.8g/l ammonium acetate
pH basic: 0.8g/l ammonium bicarbonate

The reason why I don't like CN columns? They are fairly unstable, require longer equilibration times, are less robust, ...

Did you try with a C18 column with pH 6.6?
How about these results what was wrong with these?

gl

Ace

I've ever tried to use C18, running with gradient elution. The problem was there was a ghost peak appeared exactly at the retention time of acid compound even at the first blank injection (I ran baseline before running standard solution). For more details, please check on my last posts in this forum (Topic: Baseline Absorption in Gradient Elution).

Thanks to syx for those kind explanation. The topic in this forum that you've referred might be "pKa and Buffering Capacity".

Without knowing much besides ionizability I would start with pH ~ 2 to work on a first approximation about the org. modifyer. If I knew the pKa to be that what you say I would have started with a pH similar to that suggested by Liu, but then optimized it.
Actually, my real concern is that column comparisons, etc., will appear in public without having given each type of compared column, etc., a fair chance.