Issue link: https://www.e-digitaleditions.com/i/1118782
26 Pharmaceutical Technology BIOLOGICS AND STERILE DRUG MANUFACTURING 2019 P h a r mTe c h . c o m Analytics virtually the same and were independent of the viscos- ity (e.g., the average molecular weight and molecular- weight distribution). Therefore, a silicone oil standard of one molecular weight and molecular weight distri- bution can be used for the quantitation of leachable silicone oil of different average molecular weight and molecular weight distribution. The data also show that response factors for the vola- tile silicone oil oligomers were lower than the silicone oil standard. This indicates that a portion of the vola- tile cyclic siloxanes escaped prior to atomization be- cause of their volatility and were not detected. There- fore, volatile cyclic siloxanes will not be accurately quantitated by ICP–OES (e.g., their amounts will be under-estimated). Method sensitivity. The ICP–OES method did not have a response distinguishable from the background noise when silicone oil concentration was below 0.1 ppm. When increasing the silicone oil concentration above 0.1 ppm, the response gradually became more distinguishable from the noise. The noise level varied significantly after adequate buildup of carbon within the instrument detector during analysis, affecting in- strument sensitivity and precision. For the purposes of this study, any response with a reading below 0.1 ppm was considered noise. Silicone oil at a concentration of 0.5 ppm can be mea- sured with good precision. Six measurements of 0.5 ppm silicone oil solution in xylene yielded responses as follows: 0.5205, 0.5176, 0.5283, 0.5293, 0.5240, and 0.5289. The percent relative standard deviation of the six measurements was 1.0%. Method non-interference. Eleven of the 15 formulations were stored in silicone oil-free glass containers at 5 °C, 25 °C, and 40 °C for 30 days and were then analyzed by ICP–OES, with the data summarized in Table IV. The data indicate that all the formulations stored in silicone oil-free glass containers after 30 days had ICP– OES responses below 0.1 ppm, the noise level of the ICP–OES method. This indicated there was no inter- ference for the detection and quantitation of leachable silicone oil from the formulations. Linearity. Silicone oil solutions prepared in xylene solution at different concentrations (0.5 ppm to 25 ppm) were analyzed by ICP–OES, and the responses were plotted against the concentrations seen in Figure 1. The data showed a linear correlation of the ICP–OES responses with the silicone oil concentration. The cor- relation coefficient was 0.995. Method recovery. The silicone oil was extracted into the organic solvent xylene prior to ICP–OES analysis to avoid possible interference from inorganic silica. A liquid-liquid extraction was used for all formulations with no surfactant to transfer the leachable silicone Table III. Relative response factors of silicone oil of different molecular weight. Silicone oil viscosity(cSt) Average molecular weight * Relative response factor Plasma view 3800 0.99 Analyst 5970 0.97 350 cSt 13,700 0.99 500 cSt 17,300 0.99 1000 cSt 28,000 0.99 D3 (hexamethylcyclotrisiloxane) 222 0.72 D4 (octamethylcyclotetrasiloxane) 296 0.42 D5 (demethylcyclopentasiloxane) 370 0.36 * The average molecular weight data are from Viscosity Correlation to Molecular Weight for Clearco PSF Fluids (6). The exact molecular weights of the silicone oil used in this study may be slightly different; the molecular weights are included for information purposes. Table IV. Non-interference results. Formulations 5 °C 25 °C 40 °C 1 0.035 0.031 0.026 2 0.022 0.009 0.007 3 0.009 -0.004 -0.003 4 0.028 0.024 0.018 5 0.011 0.016 0.012 6 0.004 -0.012 -0.006 7 -0.006 -0.005 -0.008 8 0.015 0.022 0.017 9 -0.025 -0.062 -0.089 14 -0.007 -0.004 -0.004 15 0.015 0.026 0.025