A few tools have become de-facto standard in EPR spectroscopy, such as the EasySpin toolbox for spectral simulations. However, routine protocols for reliable data acquisition, processing, and analysis are mostly missing. The quality of too many published papers speaks of its own. Having virtually every group develop their own very basic tools (mostly small scripts) does not necessarily improve the situation. Additionally, all too often EPR investigations will not yield results due to lack of thorough analysis and the ability to extract the necessary parameters from the data.
Reliability has several aspects: First of all, the hardware needs to reliably provide identical results for measuring the identical sample under identical conditions. Secondly, data acquisition needs to be performed accurately and appropriate for the task. And thirdly, data processing and analysis needs to be correct and reproducible.
The “research-grade” EPR spectrometers will usually provide the first aspect, i.e. repeated measurements of the same sample under identical conditions will provide identical results, both qualitatively and quantitatively. Own experience shows, however, that this is not always true for the “benchtop” series of EPR spectrometers, particularly with respect to signal intensities. As long as quantitative EPR measurements are not required, this is less of a problem, though still somewhat worrying.
Accurate data acquisition using the appropriate parameters for the sample and measurement conditions should be granted and is clearly achievable. However, due to the intrinsically complex measurements and many parameters, all too often already data acquisition cannot be guaranteed to be accurate. Hence, distorted signals with all impact on later data analysis cannot be ruled out. The simplest solution is proper training of the EPR operators and developing routine protocols for handling new samples with so far unknown signals. The next step is automating most of these aspects, freeing the scientists from spending time in the lab with routine tasks and allowing them, inter alia, to focus on the analysis part of the data.
Finally, the lack of generally available, robust, and proven tools for data processing and analysis often prevents extracting crucial parameters from the acquired data, eventually hampering EPR spectroscopy to contribute to the scientific questions at hand. Only further developing robust and proven tools for data processing and analysis using the existing knowledge and libraries from digital signal processing and optimisation and integrating these into a unified and modular package will help here.