Well it will never get a perfect method, but thats how to approach it: 1) extract e.g. liquid liquid extractions of e.g. 1 L water by 10 ml hexane. phase separate concentrate, perfome a silica micro column clean up (1g silica) condense to ~1ml inject. if FID, GC is good enough I would expect LOQ of ~1 µg/L You can fine tune by decreasing the final volume, increasing the injection volume. In both cases you will need to take more efford for the clean up. Determining recovery rates for the process is crucial. I would anyway recommend checking standard textbooks before posing question in here ;-).