Currently, there are two main preparation techniques used in the production of NO₂ diffusion tubes used in the UK NO₂ Network. The diffusion tubes may be prepared using:

Triethanolamine (TEA) in a 50:50 v/v solution with acetone, or

Triethanolamine (TEA) in a 50:50 v/v solution with water and a proprietary surfactant (Brij).

From the data obtained in the 1998 field intercomparison trial, it was identified that the two preparation techniques may produce diffusion tubes with different performance characteristics and a requirement was placed to further investigate the effect of preparation technique upon diffusion tube performance.

Figure 1 presented the average bias in the data for each participating laboratory relative to the chemiluminescent analyser. Also indicated is the preparation technique for the diffusion tubes used by each laboratory; light columns indicate diffusion tubes prepared using 50% v/v TEA in water with surfactant, dark columns indicate diffusion tubes prepared using 50% v/v TEA in acetone. The majority of the diffusion tubes (71%) exhibiting a negative bias were prepared using TEA in water, while the majority of those exhibiting a positive bias (73%) were prepared using TEA and acetone. This observation further exemplifies the potential effect of diffusion tube preparation on sampling performance. Indeed, the evidence would be fairly categorical if it were not for the results from a small number of laboratories, which show atypical diffusion tube performance for the preparation technique used.

In order to investigate further the potential effect of preparation technique upon diffusion tube performance, an additional nine tubes of each preparation technique were exposed as part of an independent test. Three diffusion tubes of each type were retained as travel blanks. After exposure, both sets of samplers were analysed by the AEA Technology laboratory. The results are shown in Table 2.

Table 2 clearly shows that the diffusion tubes prepared using acetone as the solvent for TEA exhibited a positive bias relative to the chemiluminescent analyser, (on average +11%), while those prepared using water and surfactant as the solvent, exhibited negative bias, ( on average -23%). The standard deviation of the results from diffusion tubes prepared using the TEA in water method is approximately 2.5 times that of the results from diffusion tubes prepared using TEA in acetone.

The results presented in Table 2 are broadly consistent with the results obtained by the supplying laboratory in the main 1999 intercomparison trial (presented in Table 3 below). However, in the main trial both supplying laboratories achieved similar standard deviations.