what is the best internal standard substance for L-lactate

[lyh]

for the requirement of my work, i want to analyze l-lactate in phosphate buffer by HPLC. so my procedure is as follows:
first, add 1mL l-lactate in phosphate buffer solution of certain concentration, then add 15 uL 6N H2SO4. next, add 10mL ethyl acetate to extract and then separate and evaporate. after using pure water to dissolve the residue and filtering the sample, inject the sample into HPLC. (Chiralpak ma+ column, and mobile phase is 1mM Cupric sulfurate : methanol = 9:1, and wavelength is 250 nm)
if i follow the whole procedure without internal standard substance(iss), the reproducibility is not good enough. so i want to choose a proper iss.but i tried acetic acid, propanoic acid, butanoic acid, oxalic acid, and citric acid before, each time i couldn't get good result. i am exhausted.
so i willl really appreciate if someone can give me some suggestion about the internal standard substance of l-lactate or my procedure.
thanks if somebody can help me.

[Bill Tindall]

The first step is to identify the source of variability. Otherwise I don't know how to pick an internal standard that could improve the situation. For the internal standard to imporve the precision its performance must correlate with the compound of interest in the step that results in poor precision.

If the variability is in the extraction then you have a problem. The internal standard would need to extract in the same poor way as the compound of interest. It would be easier to improve the extraction than find something that correlates.

[tom jupille]

lyh, since you tried with various internal standards, go back and check the following:

was the precision (%RSD) of the areas better or worse than the precision of the area ratios (i.e., did the use of an internal standard make things better or worse).

If the precision was worse with the internal standard, that suggests that your dominant source of error is either in separation chemistry (e.g., incompletely resolved or tailing peaks) or in the integration.

If the precision was better with an internal standard (but still not good enough), that suggests that the dominant source of error is either in sample work-up (e.g., incomplete extraction as Bill suggested) or injection.