Compound-specific carbon isotope analysis (CSIA) has become an important tool in biological, archeological, and geological studies as well as in forensics, food sciences, and organic chemistry. If sensitivity could be enhanced, CSIA would further have an improved potential for environmental applications such as, for example, in situ remediation studies to assess contaminated environments, identification of pollutant degradation pathways and kinetics, distinction between degradation/formation mechanisms, or, verification of contaminant sources. With this goal in mind, we have developed methods to determine delta(13)C values of commonly reported groundwater contaminants in low-microgram per liter concentrations. Several injection and preconcentration techniques were evaluated for this purpose, i.e., on-column injection, split/splitless injection, solid-phase microextraction (SPME), and purge and trap (P&T) in combination with gas chromatography-isotope ratio mass spectrometry. The delta(13)C values of the target compounds were determined by liquid injections of the analytes dissolved in diethyl ether or, in the case of P&T and SPME, by extraction from water spiked with the analytes. P&T extraction was the most efficient preconcentration technique reaching method detection limits (MDLs) from 0.25 to 5.0 mug/L. These are the lowest MDLs reported so far for continuous-flow isotope ratio determinations, using a commercially available and fully automated system. Isotopic fractionation resulting from preconcentration and injection was investigated and quantified for the priority groundwater pollutants methyl tert-butyl ether (MTBE), chloroform, tetrachloromethane, chlorinated ethylenes, benzene, and toluene. The isotopic fractionations caused by the extraction techniques were small but highly reproducible and could therefore be corrected for. P&T was characterized by a higher reproducibility and smaller isotopic fractionations than SPME. Among the liquid injection techniques, cold on-column injection resulted in slightly better precision compared to split/splitless injection. However, the MDLs determined for liquid injections were 4-6 orders of magnitude higher (i.e., 9.5-2800 mg/L) than for P&T and SPME. Since both of the latter methods are solvent-less, a better chromatographic resolution was obtained than for the liquid injection techniques. The P&T and SPME methods described here are also applicable for CSIA of D/H ratios, which require 10-20 times higher analyte concentrations than C-13/C-12 analysis. Finally, the applicability of the described methods is demonstrated for pollutant concentrations of only 5-60 mug/L in environmental samples.