Determination of population size of glass-like carbon sampling used in weibull analysis

Abstract

Weibull analysis is a statistical tool that can be used to support the processing optimization of fragile materials. In this context, this work presents a study about the probability of failure of a carbon material named glass-like carbon (GC) based on this tool. GC is a fragile material that has presented increasing importance in different fields, for example, optical, medical and electrochemical areas. Thus, GC samples based on polyfurfuryl alcohol heat treated under nitrogen atmosphere were prepared and afterwards characterized by flexural strength tests. The Weibull modulus of different sampling size (population of 5, 10, 15, 20, 25 and 30 specimens) of polymerized resin and GC samples were evaluated in order to determine the minimum number of specimens necessary to guarantee the Weibull analysis efficiency. The results show that the minimum sample population equals 20 for both GC and polymerized resin specimens assures the method's trustworthiness. Larger sample population does not present significant variation in the Weibull moduli.