After installing a C-GC-IRMS-System (Elemental analyzer EA 1108 CHN + IRMS MAT 252) some problems arose with respect to the external precision of the δ¹⁵N values (standard deviation approx. 1‰ at < 50 μg N) after changeover from ¹³C to ¹⁵N measurements.

Attempts were made to determine the reasons for these effects and to eliminate them by changes of the Carlo Erba on-line combustion system (EA) and the split.

Non-quantifiable contributions, e.g., air nitrogen and/or N₂ in carrier gas helium were reduced by a blank correction.

Using Cyclohexanone-2, 4-dinitrophenylhydrazone it was shown that under the same conditions and after a common blank correction with small and large amounts of samples (60-650 μg, 20% N) an external precision of 0.2‰ (standard deviation, 10 cycles) can be reached.