RESUMEN
Before the measurement of stable carbon isotope signatures (δ13C), plant samples should be well homogenized. Using a ball-mill fitted with poly tetra fluoroethylene (PTFE) plastic tubes is one of the most efficient and convenient methods. However, sample-tube plastic might contaminate plant samples during milling. In this study, a two-factor experiment was carried out using four growth chambers, with different relative humidity of the air (50% and 80%) and δ13C of the air (13C depleted and enriched). Leaf samples of Cleistogenes squarrosa (C4) were milled and homogenized using a ball-mill fitted with PTFE tubes and measured for δ13C, and the results were compared with that of leaf samples milled using metal tubes. Due to plastic contamination, 13C discrimination (Δ13C) of the two groups of plants, which were grown in CO2 with different δ13C and could be assumed as replicates, were significantly different (with an offset of 4.8 on average). The contamination led to errors in δ13C of individual leaves up to 8. Given the lower Δ13C value of C4 plants (normally about 1-8) compared with that of C3 plants, such an error caused by plastic contamination far exceeded the acceptable error range. By using a two-member mixing model that was similar to the 'Keeling plot', such errors could be effectively eliminated, and the δ13C of plant samples and pollutants were accurately estimated. Our results showed that the widely applied method of using PTFE tubes for ball-mill homogenization is not suitable for examining Δ13C of C4 plants, as it might lead to large errors. For studies with high precision requirements, samples should be milled in stainless-steel tubes.