Influence of the material of weighing bottles on loss-on-drying reproducibility

Abstract

One of the factors influencing the uncertainty of residual moisture measurements in biological medicinal products is the accumulation of electrostatic charge on the surfaces of weighing bottles and laboratory balances, which results in poor weighing reproducibility. The authors believe that the simplest and most economical solution to this problem is to use weighing bottles made of a conductive material, e.g. metal. The aim of the work was to evaluate the influence of the material of weighing bottles on the reproducibility of loss-on-drying (LOD) methods. Materials and methods: Model samples for the study were prepared from a sucrose-gelatin medium by lyophilisation and subsequent moisture sorption to achieve a certain residual moisture content. The authors assessed the samples’ mass uniformity using Shewhart’s X-charts, and analysed their residual moisture content using a loss-on-drying procedure with glass and metal weighing bottles. Statistical processing of the results was carried out by calculating the main statistical indicators: Student’s t-test and Fisher’s F-test. Results: Four batches of model samples were prepared and standardised in terms of average mass using Shewhart’s charts. The effect of weighing bottle materials was most pronounced at low residual moisture contents (less than 0.5%), with the relative standard deviation (RSD) values for the results obtained with glass and metal weighing bottles reaching 76% and 35%, respectively. For the samples with a higher residual moisture content (2–5%), the minimum RSDs with glass and metal weighing bottles were 15% and 6%, respectively. Conclusions: The study allowed for evaluating the influence of the material of weighing bottles on the results of LOD measurements and demonstrated a higher reproducibility with metal weighing bottles. This confirms the possibility of using metal weighing bottles in quality assessment of biological medicinal products for human use with LOD methods.