Inter-laboratory comparison of NO2 and SO2 generated by dynamic dilution system under laboratory conditions: a technical discussion.

A workshop on analytical quality control (AQC) of ambient air quality measurement methods for nitrogen dioxide (NO(2)) and sulphur dioxide (SO(2)) was conducted by Central Pollution Control Board (CPCB) for officials involved in National Ambient Air Quality Monitoring (NAAQM) in India. Concentrations of NO(2) and SO(2) were generated by dynamic dilution system under laboratory conditions at low and high levels and measured using static dilution system and wet chemical methods laid down by CPCB under section 16(2)(h) of the air act 1981. CPCB provided the measured values as reference values for comparing the means obtained by the officials participated from thirteen organizations. A tolerance limit of +/-15% of the reference values was specified to accept the results. Generated concentrations, which were unknown to the participants, were measured using gaseous sampling assembly (Envirotech APM 411, New Delhi, India), and wet chemical methods laid down by CPCB i.e. the same methodology which is used by the organizations to generate the data of NO(2) and SO(2) in ambient air. Simultaneously, concentrations were checked by CPCB using automatic analyzers as a check on reference concentration. It is observed that results of automatic analyzers for NO(2) and SO(2) were within a tolerance of +/-5% with %RSD below 3. On the other hand, results of most of the participants showed variability in the measurements with %RSD ranging between +/-0.8 and +/-88.6 and exceedences of means from the tolerance limit with bias ranging between 1.4 and -59%. To check the cause of high variability in the measurements obtained under identical conditions, duplicate sampling was performed by one of the participants for SO(2) at low concentration level. In this study, results of wet chemical methods, automatic analyzers and results of duplicate sampling are analysed statistically to assess the cause of high variability in the measurements. Analysis of t-test and analysis of variance (ANOVA) showed highly significant results for NO(2) and SO(2) at high concentration levels (alpha 0.05) and for SO(2) at both the levels (alpha 0.01) respectively indicating some bias is existing either in the sampling or in analytical technique. Duplicate sampling performed to check precision in parallel measurements showed high %RSD indicating the presence of systematic error in sampling technique as the same calibration factor (CF) was used to measure the concentration of duplicate samples. Statistical analysis of flow rates of duplicate sampling showed that the sampling assembly could not maintain the constant flow rate within the +/-10% with that measured at the start of the sampling. This resulted in high %RSD and deviation from the reference values for the results of most of the participants, even after accepting +/-15% tolerance limit. There is a need to improve and evaluate this gaseous sample collection device under laboratory conditions to generate reliable database of NO(2) and SO(2) in ambient air.