An improved method is described for the routine analysis of methyl tert-butyl ether (MTBE), tert-amyl methyl ether (TAME), and tert-butyl alcohol (TBA) in petroleum fuelcontaminated groundwater samples using solid-phase microextraction (SPME) and deuterated internal standards combined with gas chromatography mass spectrometry (GC-MS). Factors affecting method performance (SPME fiber selection, headspace or liquid extraction, extraction time, calibration conditions, salt addition, method sensitivity, and matrix effects) are evaluated using groundwater samples from a chalk aquifer contaminated with petroleum fuel containing MTBE, TAME, and TBA. The detection sensitivity and analytical efficiency of the method was optimized for these compounds using a PDMS-Carboxen fiber, sample NaCl content of 25% (w/v), and extraction time of 30 min. Internal calibration standards (deuterated MTBE and TBA) are necessary to control extraction errors during analysis. SPME extraction efficiency and detection sensitivity for the oxygenates and TBA decreased as the background matrix concentration of benzene, toluene, ethylbenzene, and xylenes (BTEX) increased up to 300 mg/L total BTEX. However, reliable measurement of MTBE, TAME, and TBA was possible in this BTEX matrix using deuterated internal standards. The precision, accuracy, and reliability of the method were verified by analysis of certified standards. Analytical accuracy, determined from replicate (n ) 10) analysis of spiked laboratory standards and groundwater samples, was 97%, with a precision of 1.6-2.9% for MTBE, 3.1-5.8% for TAME, and 1.6-1.7% for TBA. Method detection limits under the conditions described are 2 µg/L for TBA (by liquid sampling) and 1 µg/L for MTBE and TAME (by headspace sampling). This sensitivity can be increased for MTBE by further refinement of the method.