The STIMES ring trial on sulfur dioxide, carbon monoxide and benzene was carried out at the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen (LANUV NRW) from 12 to 14 September 2023. The trial brought together 19 participants for sulfur dioxide, 16 for carbon monoxide and 10 for benzene, representing a range of measurement techniques: 16 ultraviolet‑fluorescence instruments for SO₂, three nondispersive infrared analyzers for CO, one gas‑chromatography system for benzene and one absorption‑tube method for benzene. The goal was to compare results for gas mixtures that span regulatory limits and typical outdoor concentrations, to establish reference values with defined uncertainties, and to assess cross‑sensitivity to interfering species.
Five gas mixtures (PG 1–5) were prepared with SO₂ concentrations from 30 µg m⁻³ to 350 µg m⁻³ and CO concentrations from 1 mg m⁻³ to 10 mg m⁻³. Benzene was evaluated in a separate comparison using ORSA tubes supplied by Dräger; each participant received four coded tubes for BTEX analysis. Participants were asked to report their measurement uncertainties, which were then used to calculate En‑numbers according to DIN ISO 13528.
The statistical analysis of the participant data showed that the standard deviation of SO₂ measurements ranged from 0.6 to 5.1 µg m⁻³, corresponding to a mean relative standard deviation of 1.5 %. For CO the spread was 0.04 to 0.08 mg m⁻³, with a relative standard deviation between 0.9 % and 4.1 %. Benzene exhibited a standard deviation of 0.3 to 0.5 µg m⁻³, giving a relative standard deviation of 7 % to 27 %. Z‑score evaluation revealed that two of the 19 SO₂ methods marginally failed the acceptance criteria, while all 16 CO and all 10 benzene methods met the requirements.
En‑numbers for SO₂ and CO were largely within the acceptable range of –1 to +1, indicating satisfactory performance. Two benzene methods produced En‑numbers outside this interval, reflecting challenges in estimating measurement uncertainty for this compound. Cross‑sensitivity tests were performed for SO₂ against water vapour and nitric oxide, and for benzene against water vapour and ozone. For SO₂ at 100 µg m⁻³ and 40 % relative humidity, the mean cross‑sensitivity was –2.8 µg m⁻³, well below the maximum permissible value of 26.6 µg m⁻³. When exposed to 500 nmol mol⁻¹ NO, the mean positive cross‑sensitivity was about +8 µg m⁻³, again below the limit of 13.3 µg m⁻³. Benzene cross‑sensitivities were marginal and did not exceed regulatory thresholds.
The ORSA‑tube comparison for BTEX provided an independent check of benzene results. The four tubes per participant were coded and analysed by the same methods used in the ring trial, ensuring consistency across laboratories.
The collaboration involved state environmental agencies, measurement laboratories, and the manufacturer of the ORSA tubes. LANUV NRW coordinated the event, supplied the gas mixtures, and managed data collection. Dräger supplied the ORSA tubes and coded them for blind analysis. The trial was funded by the Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, reflecting a commitment to maintaining high‑quality ambient air monitoring across German states. The outcomes of the ring trial will inform future calibration protocols, improve uncertainty estimates, and support the continued reliability of air‑quality data used for regulatory and public‑health decisions.