科研之友
微信
新浪微博
Facebook
分享链接
摘要
This work describes and characterizes an algorithm for the nonambiguous reduction of multiexponentially decaying luminescence signals to scalar values of corresponding calibration temperatures. Previous evaluation schemes in phosphor thermometry make use of an intermediate step, where data reduction is achieved by fitting a model equation to phosphorescence decays in order to translate one or more fitting parameters into temperature. However, every slight mismatch between model equation and experimental data may lead to substantial errors in connection to noise-related inaccuracies during the retrieval of adequate fitting windows. Additionally, there is a need for fitting windows, capable of automatically adapting to largely varying signal time scales. In this context, the authors propose to set the fitting window length according to the time where the signal falls below a given percentage of the initial intensity. In comparison to fitting windows, defined by multiple decay times, modeling results suggest substantial precision benefits for as long as signal-to-noise ratios stay above 4. Nevertheless, by comparing signal shapes of measured curves directly with a library of temperature-calibrated decay signals, all necessary assumptions on the mathematical description of measured signals become redundant and evaluation errors connected to uncertain fitting windows are largely circumvented. Resulting capabilities of the proposed signal shape recognition method (SSR) in terms of temperature precision and accuracy were compared to a conventional least-squares fitting approach, using a set of temperature-calibrated phosphorescence decay signals from CdWO4. Accordingly, the SSR algorithm was found to reduce statistical temperature errors by at least 9 %.
- 出版日期2014-1
