THE VARIABILITY OF DRY-MASS CORRECTION FOR CERTIFIED REFERENCE MATERIALS

A realistic or an acceptable level of variability that can be expected between laboratories when reporting dry-mass corrections for candidat e reference materials is not known. The proper use of solid certified reference materials (CRMs) necessitates a dry mass correction so the e lemental composition can be reported as a mass of analyte per unit dry mass of sample. The water content of seven, biological and environmen tal, certified reference materials (Measurements and Testing Programme , formerly the BCR) was studied as a function of time, each in seven m icro-environments ranging from 15 to 94-percent relative humidity. For the determination of water in these different relative humidity envir onments, one CRM was observed to have a reproducibility as high as 23- percent (alpha = 0.05). This degree of variability suggests reproducib ility consistently less than 10-percent (alpha = 0.05), as observed in interlaboratory campaigns to certify candidate reference materials, r epresents acceptable between laboratory variability.