Multiconvolutional approach for uncertainty estimation of the thermal diffusivity measured by flash method

The paper addresses the exactness of the results of the thermal diffusivity (a) of a Sn-Ag-Cu soldering alloy, mainly a new approach for measurement uncertainty (MU) estimation of the value assigned to the measurand. The method used for α measurement is the flash one implemented on an Anter Flash Line 3000. The weak point of this method is the exactness of the value assigned to the measurand due to the large uncertainty budget of the measurement process. Thus, uncertainty budget contains a great number of contributor factors to the MU assigned to the result as specimen factors (homogeneity, flatness etc), equipment factors (drift, calibration etc) operational factors (parameter settings, specimen coating etc). The authors purpose a multiconvolutional approach based on uniform probability distributions function (PDF) of the individual result. The fundamentals of the multiconvolutional approach are given in the paper together with the motivation for intentional use of this approach. This approach fulfills the requirements of the EN ISO/CEI 17025 and EN Guide 98-3 standards that underpin any effort for an accurate estimation of the MU. The paper brings a new view about MU estimation for α of Sn-Ag-Cu which could be extended for the interconnecting structures and not only.