The direct derivatisation of acetic acid with n-hexyl chloroformate and with benzyl bromide in water was evaluated. With n-hexyl chloroformate, acetic acid did not give the n-hexyl acetate derivative, but the reaction of acetic acid with benzyl bromide in aqueous solution resulted in the formation of benzyl acetate. The derivatisation of acetic acid with benzyl bromide and the headspace solid-phase microextraction (SPME) of benzyl acetate were optimised. Under optimum conditions, the limit of detection for acetic acid was 260 nM, and the relative standard deviation of the overall procedure at 1.10(-4) M acetic acid was 15.6% (n = 10). A linear response was obtained in the 1 x 10(-4) to 5 x 10(-6) M concentration range (R2 = 0.993, n = 6). Although Carbowax-divinylbenzene (CW-DVB)-coated fibres exhibited a higher extraction capacity for benzyl acetate, polyacrylate (PA) was selected, because its mechanical stability was better than that of CW-DVB fibres. Moreover, the relative standard deviation of the SPME was better with PA (1.5%, n = 10 at 1 x 10(-5) M) than with CW-DVB-coated fibres (8.0%, n = 10 at 1 x 10(-5) M). Thus, a new analytical method for the quantitative determination of micromolar concentrations of acetic acid in the aqueous phase was developed. This method is based on water-phase derivatisation with benzyl bromide, headspace SPME with PA fibres and GC-FID. It was observed experimentally that benzyl alcohol formed by hydrolysis of the reagent affected the fibre-gas phase partitioning of benzyl acetate.