Improvement of UN N.5 test method, in anticipation of potential use for the characterization of substances which in contact with water emit toxic gases
Résumé
The test method UN N.5 is intended to the characterization of hydro-reactive substances which in contact with water emit flammable gases. This test method and related classification criteria are defined in the UN Manual of Test and Criteria for the Transport of Dangerous Goods. In the context of CLP regulation, the UN test N.5 is also designated as the possible candidate to the newly implemented physico-chemical hazard class of "substances and mixtures which in contact with water emit toxic gases", for which for this moment there is no regulatory test method and for which no international standard exists. In practice, experience shows that the implementation of this test method leads to technical issues with respect to metrology, that generate significant uncertainties of the results potentially leading to wrong classification of the test sample. Based on an experimental study, this work targets the optimization and the evaluation of metrological performance of three suitable devices for the measurement of emitted gas volume. The tests consisted in a first step to implement the volumic measurement devices tested under controlled flow conditions. These tests were carried out with hydrogen and air, to mock up adequate variation spectrum of experimental conditions of substance classification tests that emit flammable gases or toxic gases. In a second step, the measuring devices were tested under experimental conditions for measuring the volume of hydrochloric acid or hydrogen generated by reactions between: i) water and aluminum chloride, ii) water and sodium borohydride. The results regarding measurements of volume and flow calculations are consistent with previus data published in 2014 on the same products by another laboratory. The results also reveal significant improvement in accuracy data of measurement devices selected and optimized, as compared to ordinary usual practice, whereas this is obtained without any chemical incompatibility issues with the toxic and/or corrosive gases potentially released. The results obtained show, for a gas flow rate of 1 mL/hr, an error less than 10%. This error can even be reduced to 5% for a 2 mL/hr flow rate. The standard deviation determined on the basis of three runs is lower than 10% for these two flow values. In conclusion, on the basis of a reduction factor of 10 between the classification thresholds of substances that emit flammable gases in contact with water and the hydro-toxic substances, the tested devices would be suitable for hydro-toxic substances classification tests.