A new sealed double tube combustion method was developed at Lawrence Livermore National Laboratory (LLNL) to extract carbon from modern steels and ancient iron artifacts. Iron samples were chemically pretreated with 10% nitric acid, vacuum sealed in 6 mm quartz tubes with CuO, vacuum sealed again inside 9 mm quartz tubes, and combusted at 1000 °C for a minimum of 10 hr. The resulting CO2 was graphitized routinely using hydrogen reduction (Vogel et al. 1989). After the initial phase of development, carbon yields of 100% were consistently obtained. The activities of two modern high carbon steels (treated as process blanks, manufactured using only coal as the carbon source) were determined to be 0.0077 ± 0.0009 (n = 12, ± 1 σ) for a 1.3% C steel and 0.0090 ± 0.0038 (n = 12, ± 1 σ) for a 1.9% C steel, indicating that very little contamination is introduced during the sample preparation process. Since the Iron Age began less than 5000 years ago, these background uncertainties should introduce errors of no more than ±30 years to the radiocarbon ages of actual artifacts. Two ancient iron artifacts of known date were analyzed and demonstrate that the new methodology can be used to obtain the correct date of manufacture for iron objects, provided that they are made exclusively using charcoal that was contemporaneous with the manufacture of the artifact. Since only 1 mg of carbon is required for accelerator mass spectrometry (AMS), very small iron samples can now be analyzed (50 mg of a 2.0% C iron or 1 g of a 0.1% C iron). We anticipate that this methodology will be particularly useful to archeologists who currently have to rely on context to date iron artifacts.