Major elements in rainwater

Five 60 L polypropylene containers to be used for the preparation and homogenisation were rinsed with ultra pure water, interconnected, provided with a drain tap at the bottom and filled with hydrogen peroxide (30 g kg ) to disinfect. After a contact period of 6 h the hydrogen peroxide solution was circulated in the tanks by pumping continuously for 16 h. The tanks were emptied and rinsed with ultra pure water. Samples of the water and from the walls of the tanks were bacteriologically examined (diluted broth agar medium, 10 days at 25°C). Samples of the rinsing water had typical microbial number concentrations (CFP) [3] of less than 1 cm. For the examination of the walls 

The drain tap of each tank was also cleaned with hydrogen peroxide (30 g kg ), leaving the tap filled for at least 2 h. The PTFE tip which was mounted to the drain tap for filling the ampoules was cleaned in the same way. Just before filling the PTFE tip was heat-sterilised. A test was carried out to estimate the losses by evaporation during the preparation of the reference material. The test consisted of filling three ampoules with ultra pure water. After weighing, the ampoules were left at room temperature for a period of 7 days. The relative loss by evaporation was less than 0.02% /day for both types of ampoules (diameter of 40 and 30 mm for respectively CRMs 408 and 409). 

To establish possible contamination from the quartz ampoules, the water that was in contact with the quartz for 7 days was analysed. The analyses resulted in concentrations of the determinands below the limits of detection of the methods applied. For each of the reference materials, 1200 ampoules were conditioned for at least 24 h with the solution which they would contain. During conditioning, the ampoules were loosely closed with a synthetic foil. Preliminary experiments had shown that one treatment to bring the walls in adsorption equilibrium with the solution was sufficient. 

The two reference materials were prepared from ultra pure water (0.05 pS m ), to which freshly prepared solutions of ammonium sulphate, sodium nitrate, magnesium nitrate hexa hydrate, sodium chloride, calcium chloride (all reagents pro analysis quality), hydrochloric acid and ultra pure nitric acid were added. 

To prevent air from the laboratory to enter into the containers a slight flow of pure nitrogen (produced from liquid N2 passed over a column of active coal) was passed over the liquid surface in each container. Homogenisation of the solution was performed in the polypropylene containers, which were connected in series with PTFE tubing and covered with a close fitting polyethylene lid. A centrifugal pump connected to the containers with PTFE piping ensured constant circulation of the solution. The pump had no metal parts in contact with the solution. 

---