Abstract: In this study, we characterized the amounts and isotopic compositions of the blank CO₂ remaining in vials after specific flushing times and examined how it influences the measured δ¹³C and δ¹⁸O values of carbonates. Results show that the blank CO₂ remaining in the vials due to either incomplete removal of air caused by short flushing time or an input from an inherent source would significantly influence the measured C and O isotopic compositions for small-sized carbonates (e.g. <100 g). The extent of this influence depends on the amount of blank CO₂ and the isotopic differences between the blank CO₂ and the carbonates. Therefore, we performed blank corrections to the measured δ¹³C and δ¹⁸O values, especially for samples of small size. The corrected δ¹³C and δ¹⁸O values when sample weight is bigger than 20 μg have an overall precision of <0.12‰ and <0.13‰, respectively. This testifies the practicability of the described blank correction strategy. Although the variation for measurements of the amount, C and O isotope of blank CO₂ could give rise to ~0.2‰ uncertainties for the corrected δ¹³C and δ¹⁸O values of small-sized carbonate samples, the uncertainties will be greatly reduced if we perform routine precise measurement on C and O isotope of lab air and use them for blank correction.