Skip to content

simLIBS provides Python class to simulate LIBS spectra with NIST LIBS Database interface.

License

Notifications You must be signed in to change notification settings

MKastek/simLIBS

Repository files navigation

Simulated LIBS

PyPI version License Tests ZENADO DOI

SimulatedLIBS provides Python class to simulate LIBS spectra with NIST LIBS Database interface.
SimulatedLIBS also allows the creation of simulated data sets that can be used to train ML models.
SimulatedLIBS was mentioned is FOSS For Spectroscopy by Prof. Bryan A. Hanson, DePauw University.

Project created for B.Eng. thesis:
Computer methods of the identification of the elements in optical spectra obtained by Laser Induced Breakdown Spectroscopy.

Thesis supervisor: Paweł Gąsior PhD
e-mail: [email protected]
Institute of Plasma Physics and Laser Microfusion - IPPLM

Installation

pip install simLIBS

Import

from simLIBS import SimulatedLIBS

Example

Parameters:

  • Te - electron temperature [eV]
  • Ne - electron density [cm^-3]
  • elements - list of elements
  • percentages - list of elements concentrations
  • resoulution
  • wavelength range: low_w, upper_w
  • maximal ion charge: max_ion_charge
  • websraping: 'static' or 'dynamic'

Static websraping

libs = SimulatedLIBS(Te=1.0,
                     Ne=10**17,
                     elements=['W','Fe','Mo'],
                     percentages=[50,25,25],
                     resolution=1000,
                     low_w=200,
                     upper_w=1000,
                     max_ion_charge=3,
                     webscraping='static')

Plot

libs.plot(color='blue', title='W Fe Mo composition')

Save to file

libs.save_to_csv('filename.csv')

Interpolated spectrum

SimulatedLIBS interpolates retrieved data from NIST with cubic splines.

libs.get_interpolated_spectrum()

Raw spectrum

Raw retrieved data from NIST

libs.get_raw_spectrum()

Dynamic webscraping

libs = SimulatedLIBS(Te=1.0,
                     Ne=10**17,
                     elements=['W','Fe','Mo'],
                     percentages=[50,25,25],
                     resolution=1000,
                     low_w=200,
                     upper_w=1000,
                     max_ion_charge=3,
                     webscraping='dynamic')

Plot

libs.plot(color='blue', title='W Fe Mo composition')

Ion spectra

After simulation with parameter websraping = dynamic, ion spectra are stored in ion_spectra (pd.DataFrame) and can be plotted.

libs.plot_ion_spectra()

Random dataset of samples

Based on .csv file with chemical composition of samples, one can generate dataset of simulated LIBS measurements with different Te and Ne values.

Example of input_composition_df pd.DataFrame:

W H He name
50 25 25 A
30 60 10 B
40 40 20 C
SimulatedLIBS.create_dataset(input_composition_df,
                             size=100,
                             Te_min=1.0,
                             Te_max=2.0,
                             Ne_min=10**17,
                             Ne_max=10**18)

Example of output .csv file:

200.0 200.1 200.2 200.3 200.4 ... H W Te Ne
0 0 0 0 0 0 ... 2 50 1.43 1.08e+17
1 0 0 0 0 0 ... 0 0 1.06 1.08e+17
2 0 0.1 0.1 0.1 0.1 ... 0 68 1.82 1.18e+17
3 0 54.5 56.7 54.4 48.4 ... 0 3 1.25 1.06e+17
4 0 121.7 143.1 140.5 115.3 ... 0 84 1.08 9.23e+17

Animations

SimulatedLIBS can be helpful in creating LIBS animations mostly for educational purpose.

Resolution animation

Changes in resolution in range: 500-10000.

Electron temperature animation

Changes in electron temperature Te in range: 0.5-5 eV.

Electron density animation

Changes in electron density Ne in range: 0.7-1.3 e+17 [cm^-3].

References

  • M. Kastek, et al., Analysis of hydrogen isotopes retention in thermonuclear reactors with LIBS supported by machine learning. Spectrochimica Acta Part B Atomic Spectroscopy 199: 106576. DOI: 10.1016/j.sab.2022.106576.

Used in Research

  • Chen Z, Chen Z, Jiang W, Guo L, Zhang Y. Line intensity calculation of laser-induced breakdown spectroscopy during plasma expansion in nonlocal thermodynamic equilibrium. Opt Lett. 2023 Jun 15;48(12):3227-3230. DOI: 10.1364/OL.488250.