In this article, a software package code named OPSIAL (Optical Plasma Spectral Calculation And Parameters Retrieval) for rigorously calculating optical plasma spectra and for automatically retrieving plasma parameters is presented. OPSIAL calculates the absolute spectral radiance caused by the bound-bound transitions of elemental species in the plasma by rigorously solving the equation of radiative transfer using an ultrafast line-by-line algorithm. OPSIAL supports both the local-thermodynamic-equilibrium (LTE) or partial LTE conditions and takes account of line broadenings due to the Doppler effect and collisions with electrons and other pseudo colliders in the plasma. An algorithm for fully automatically identifying elemental species and retrieving plasma parameters based on observed plasma emission spectra has been implemented into OPSIAL. The structure and theoretical framework of OPSIAL, together with a case study of using OPSIAL to analyze laser-induced breakdown spectral data of the ChemCam instrument onboard the Mars rover Curiosity, are presented.
Accurate calculation of optical emission spectra of plasma is one of the key steps for understanding the plasma physics and chemistry and has tremendous applications in both fundamental research fields such as astrophysics, combustion, propulsion, plasma dynamics and in numerous practical spectroscopic diagnostic and analysis techniques such as optical emission spectrometry, 1 inductively coupled plasma, 2 and laser-induced breakdown spectroscopy (LIBS). 3 Rigorous treatment of the plasma optical emission problem requires the solution to the equation of radiative transfer 4 along the line-of-sight (LOS) through the plasma. In the general case where there are a significant number of emission lines that are not optically thin, the solution to the equation of RT demands significantly amount of computational power. For this reason, most software tools for treating plasma optical emission take the shortcut by assuming the optical thin condition that leads to significant reduction to the computational power at the cost of losing accuracy and generality for treating the emission problem.
In a recent paper, Tan introduced an ultrafast line-by-line algorithm for solving the equation of radiative transfer for gaseous emitters with large number of emission lines. 5 The algorithm is based on calculation of the optical depth in the Fourier space that reduces the time-consuming integration of the optical depth along the LOS to multiplications and therefore speeds up the calculation by two orders of magnitude or more depending on the number of lines involved. More speedup can be achieved if this method is combined with a binning technique as demonstrated by Tan. The major motivation for the development of the OPSIAL package is to apply this new development to the plasma emission problem to achieve the goal of speed, accuracy, and flexibility in the treatment of the problem. To achieve this goal, OPSIAL also rigorously treats line broadenings caused by the Doppler effect and collisions with the electrons and other pseudo colliders in the plasma and supports spectral calculations under the partial LTE conditions where the four temperatures of the plasma are taken into account: gas kinetic temperature (T g ), electron temperature (T e ), ionization temperature (T i ), and excitation temperature (T exc ).
Taking advantage of its ultrafast speed and high accuracy for rigorously calculating plasma optical emission, another goal attempted in OPSIAL is to achieve fully automatic determination of elemental species and retrieval of plasma parameters from plasma optical emission measurements. To the best of our knowledge, there is no method up to date that can be used to reliably accomplish this challenging task due to the complexity nature of the problem. Traditionally, species identification and retrieval of plasma parameters were performed in a series of steps that eventually require some sort of human intervention. 6 Some of the parameters such as the electron tem-perature and density may be determined from information such as the line-to-continuum intensity [7][8][9][10][11] or the Stark line broadening. [12][13][14][15][16][17] All methods involving human intervention eventually become very inefficient and tedious as the number of lines in the emission spectrum increases and the quality of any human involved process is difficult to gauge quantitatively. In recent years, chemometric methods based on multivariate statistical analysis in various forms have been developed to partially mitigate the situation. [18][19][20][21][22][23] These statistics based methods typically require extensive training data to generate a statistical relationship that is later used to perform the species identification and/or classification task. One issue associated with these ad hoc methods is that a dedicated algorithm has to be trained for a specific application with data prepared for the application. As a result, the whole process is typically time consuming and expensive. Moreover, a species identifier based on these statistical methods is usually not as accurate as ones based on first-principles methods. Lastly, these methods normally require wavelength calibrated spectral data for correct operation and thus is very sensitive to wavelength calibration error or drift in the training data.
As a work-in-progress feature, OPSIAL addresses the automatic retrieval problem in an approach based on the construction a species identifier calibrated with a performance library (PL) that contains information on performance metrics of the species identification process. The performance metrics information in the PL is generated using training data calculated by Monte Carlo simulations with OPSIAL’s fast emission calculation code. This approach enables OPSIAL to identify species in any plasma emission spectrum in a generic way provided enough training data have been generated in the simulations. Once the species are identified, a built-in spectral fitter in
This content is AI-processed based on open access ArXiv data.
