The volumetric method is widely used to measure the gas adsorption isotherms which have the most basic importance for characterizing porous and nonporous colloid materials. However, this method contains a surprisingly large number of sources of experimental errors in contrast to its theoretical and technical simplicity.
Seventeen sources of errors are pointed out arising from the vacuum line, the vacuum techniques, the method of pressure measurement, the estimation of dead volumes, the pretreatment and the preparation of the samples to be measured. Among them, the sources of errors relating to the sample seem. to have the greatest contribution to the total error, since colloid materials are generally very sensitive to the physical, chemical and mechanical treatment such as preheating, weighing and evacuation.
A sample cell for gas adsorption is newly developed in this laboratory, which enables the precise measurement of the dead volumes and of the sample weight without exposing to air, and a very easy evacuation in the vacuum system before gas adsorption.
The reliability of the absolute values of the amount of gas adsorbed measured under these precautions is not always high, because almost no standard scales for these physical quantities are available. In this study, a search for reference materials are carried out as follows. Microporous silica gels of granular type were sufficiently purified, grinded and sieved in the range from 32 to 42 JIS mesh, mixed well using a special type mixer to attain a good homogenuity by the courtesy of Mr. Kozo Yamamoto and Mr. Kazunori Nobuhara of Fuji-Davison Co., Ltd. These were distributed to 14 laboratories in order to compare the values measured in each laboratory. The adsorption isotherms obtained and the BET surface areas, C values and the pore volumes computed from these data differ from each other much more largely than those which would be expected from the sample heterogenuity. This shows the strong necessity to set several materials as references, in order to improve the reliability of these values and the measuring techniques.
View full abstract