In Part I of this paper, general trends in the realisation of this Round Robin Test on chloride analysis in concrete (carried out as part of the work developed by the Technical Committee TC 178-TMC), as well as the results corresponding to total chlorides were given. The present paper reports the results corresponding to the extraction of water soluble chlorides. Results concerning the determination of the penetration front of chlorides by colourirnetric techniques will be given in Part III. A total of 30 different laboratories around the world have participated in this Round Robin Test. All of them have performed analyses of total chlorides, having performed 64 different determinations of total chlorides in triplicate specimens. Among them, 20 laboratories have analysed water soluble chlorides. The total number of different determination of free chlorides, in triplicate specimens, has been of 37. Concerning the colourimetric method, 7 laboratories have taken part in the determinations, being the total number of different determinations of the colourimetric front, of 10. Two different methods of extraction from the solid sample (B1 and B2) have been tested. It was also decided that other methods could also be used to the choice of each laboratory, as "other methods" (I33). (The complete recipes of these methods can be found in Annex 1. The statistical treatment of the data has been carried out according to the International Standard ISO 5725-1981 (F) for the determination of the reliability of testing methods: "Determination of the repeatability and reproducibility by Round Robin tests". Finally, it has been concluded that both methods, B1 and B2, are reliable enough to extract water soluble chlorides from concrete. However the method that each laboratory currently uses (other methods) is not recommendable.
Dans la partie I de cet article, les tendances g~n&ales de la r&lisation de l'essai interlaboratoire (effectu~ par la Commission Technique TC 178-TMC) lors de l'analysedu chlorure dans le b~ton, ainsi que les r&ultats du chlorure total ont (t( donn&. Cet article rapporte les r&ultats des chlorures libres. Le front de colorim&ie sera rapport6 dans la partie IIL Une trentaine de laboratoires ont participd a cet essai interlaboratoire, et 64 analyses diff&entes sur des &hantillons triple exemplaire ont dt6 effectu&s. Parmi ces laboratoires, 20 d'entre eux ont analysd les chlorures libres et 37 analyses diffdrentes ont dt( effectu&s en triple exemplaire. Deux m(thodes diff&entes d'extraction des chlorures libres a partir des solides (B1 et B2) ont (t( test&s. Les laboratoires pouvaient avoir le libre choix d'utiliser {, d'autres mdthodes,} (B3). (La recette compl~te de ces m~thodes est report& dans l'annexe i). Pour la ddtermination de la fiabilitd des m(thodes de test, le traitement statistique des donn&s a dt( re'ah's6 en accord avec la norme internationale ISO 5725-1981 (F) : {~D~termination de la rdpdtabitit~ et de la reproductibilit6 par essais interlaboratoires,}. Comme r&ultats finaux, les deux mdthodes, B ...