EP Impurity Method for Timolol maleate is not robust

[max_planck]

Dear Colleagues,

I am currently doing a robustness study of pharmacopendial methods and noticed that the EP impurity method for Timolol maleate is not robust regarding the retention time of the API peak Timolol maleate.

Under nominal conditions my peak comes at approximately 7 minutes but when I vary the amount of methanol by +/- 10% in eluent A, the retention times shift to 30 and 2 minutes respectively. This I find a bit to extrem and is not robust enough for daily routine testing.

The method details are listed in the current EP and are as follows:

Column: 125 x 4 mm, 5 um Octadecylsilicate. The column I use is a Nucleosil C18 von Machery & Nagel with the same dimensions.
Buffer: A sodium octane sufonate solution adjusted to pH 3.0 with 99% acetic acid. I apply a 4.32 g/L sodium octane sulfonate solution adjusted to pH 3.0 with glacial acetic acid.
Mobile Phase A: 50:50, Buffer: Methanol (v/v)
Mobile Phase B: Methanol

Gradient:

0 -10(min) 97.5%A 2.5%B
10 - 11(min) 97.5%A to 70%B ; 2.5% A to 30%B
11 - 14.5(min) 70%A 30%B

Flow rate: 2.0 ml/min
Detection: 295 nm
Injection volume: 20 ul

I varied the pH by +/- 10% and it had no significant effect on the retention time of the Timolol maleate peak. My question therefore is, could this extrem shift (instability) in retention time be due to my Nucleosil C18 column, methanol as eluent/solvent and/or the Mobile phase ratio of 2.5% to 97.5% during the first 10 minutes?

Many thanks in advance!

Max P.

[HPLCaddict]

If I remember correctly, there is a rule of thumb that for (isocratic!) methods an increase of 5% for the organic modifier yields a 2-fold decrease of the capacity facor. So you can imagine that your change of 10% will for sure change the capacity factor and thus retention time significantly. Although of course this is a gradient method which makes predicitons a lot tougher.

In this present case, what do you mean by 10% change? Did you change the gradient steps by 10% or did you change the composition of mobile phase A by 10%?

[max_planck]

Did you change the gradient steps by 10% or did you change the composition of mobile phase A by 10%?

I prepared Mobile phase A twice, one with 40% methanol (400 ml methanol and 600 ml buffer) and the other with 60% methanol (600 ml methanol and 400 ml buffer).

[max_planck]

Although of course this is a gradient method which makes predicitons a lot tougher.

The Timolol maleate peak comes at approximately 7 minutes and is therefore within the isocratic step of the first 10 minutes. The gradient can therefore be dismissed as it starts 3 to 4 minutes after the Timolol maleate peak.

[HPLCaddict]

Let's do some quick and dirty calculations:
You're using a 125x4mm column, that one should have a dead volume of roughly 1 mL. At a flow rate of 2 mL/min that gives a dead time of ~0.5 minutes. Your peak normally elutes at 7 minutes. The capacity factor therefore should be about (7-0.5)/0,5 ~ 13. Increasing Methanol by 10% should decrease the capacity factor roughly 4-fold, that is 3.25. Transforming back to retention time gives 3.25x0.5+0,5 ~2.1 minutes. Wow! You had 2 minutes retention time? These rules of thumb sometimes are fantastic .

You see, the retention time changes are perfectly reasonable. The column is innocent.
Personally, when doing robustness testing, I'd like to include practical considerations concerning the parameters to be changed. E. g. in your case: Your mobile phase A is buffer/methanol 50:50. How do you prepare 1L of mobile phase? Measuring 500 mL buffer, measuring 500 mL MeOH, mixing together. How do you measure those 500 mL? Using a 500 mL or perhaps 1000 mL measuring cylinder I suppose? Is it really probable (in practice), to be off 10%, that is 50 mL??? Of course, you should challenge your method, but some challenges are simply too big .

[max_planck]

Let's do some quick and dirty calculations:
You're using a 125x4mm column, that one should have a dead volume of roughly 1 mL. At a flow rate of 2 mL/min that gives a dead time of ~0.5 minutes. Your peak normally elutes at 7 minutes. The capacity factor therefore should be about (7-0.5)/0,5 ~ 13. Increasing Methanol by 10% should decrease the capacity factor roughly 4-fold, that is 3.25. Transforming back to retention time gives 3.25x0.5+0,5 ~2.1 minutes. Wow! You had 2 minutes retention time? These rules of thumb sometimes are fantastic .

You see, the retention time changes are perfectly reasonable. The column is innocent.
Personally, when doing robustness testing, I'd like to include practical considerations concerning the parameters to be changed. E. g. in your case: Your mobile phase A is buffer/methanol 50:50. How do you prepare 1L of mobile phase? Measuring 500 mL buffer, measuring 500 mL MeOH, mixing together. How do you measure those 500 mL? Using a 500 mL or perhaps 1000 mL measuring cylinder I suppose? Is it really probable (in practice), to be off 10%, that is 50 mL??? Of course, you should challenge your method, but some challenges are simply too big .

Hello HPLCaddict,

many thanks for your help! Those rules of thumb are often amazingly accurate. The mobile phase is prepared by combining equal volumes of methanol and buffer (i.e. 500 mL + 500mL) using a 1000 mL volumetric cylinder with an error of +/- 5mL. In addition to this, I wonder if its necessary to have 97.5% A and 2.5%B during the first 10 minutes. What if I just turn off channel B during the first 10 minutes and correct the ratio of methanol : buffer in mobile phase A i.e. 500 mL buffer mixed with 510 mL methanol? Just to reduce a further error introduced by the pump. Regarding the 10% challenge you are right, +/- 50 mL from 500 mL is to much but even 5% has a significant impact as the API peak still could run into one of the impurity peaks.

Best regards

Max

[JI2002]

I thought that the 10% change in the test means a relative change of 10%, not absolute change of 10%. For example, if the original method requires 40% MEOH/60% water, you can change it to 44% MEOH/56% Water or 36& MEOH/64% water.