Rosie Braunholtz
Data overload is a primary challenge current astronomers face. With the next generation of telescopes, there will soon be an explosion of data and processing data will no longer be feasible to do ’manually’. I created a program to aid the process of identification of a large database of emission lines. These emission lines come from a catalog of X-ray spectra taken from the telescope XMM-Newton. These observations include stars, X-ray binaries, supernovae remnants, active galactic nuclei and clusters of galaxies.
Identification of emission lines was made possible by use of AtomDB which is a large database of ions used by astronomers for spectral modeling. A small subset of ions known to appear prominently in astronomic spectra were selected from the enormous AtomDB database. This was done by “manually” going through 2D intensity plots of a few astronomical objects (binary star, supernova and galaxy cluster) and selecting the most prominent emission lines from each and matching them with closest and strongest ions from AtomDB.
The program I created matches each emission line from the catalog with one or more ions from the small subset.
A match between the catalog and ions in the AtomDB database is based on a difference in wavelength within 1%. Nearly half of all lines in the database matched closely (within 1%) to one or more of the lines in small subset. Expanding the subset of AtomDB lines used in matching, for instance adding lines found in different types of astronomic objects, would be expected to radically increase the number of close matches found. Overall this algorithm was a small step towards automating spectroscopy in Astronomy.
High resolution X-ray spectroscopy provides astronomers with information about the hot and energetic Universe. Two major X-ray space-observatory facilities were launched in 1999 - Chandra and XMM-Newton. The spectra from these X-ray telescopes have resulted in major scientific breakthroughs.
The ATHENA X-ray observatory is to be launched in 2028. It will carry the most sensitive X-ray telescope yet launched. Below is a comparison of the Perseus cluster seen by Hitomi grating spectrometer SXS (X-ray telescope launched in 2016) and a simulation of ATHENA grating spectrometer X-IFU. This illustrates the improvement of sensitivity due to effective area increase between the two instruments.
How to process this overload of data? Development of technology and scientific instruments such as ATHENA is great news to astronomers. However the progression of this technology is advancing so fast that it takes a long time to process all the data. Creating algorithms and software that enable computers use to process data automatically (and very quickly) is essential
Second slide:
Three celestial objects’ X-ray spectra were analyzed manually. These were AB Doradus, a binary star system; DEM L71 a supernova remnant and lastly MCXC J1023.6+0411 a galaxy cluster taken from a mega catalogue of clusters MCXC.
Spectral lines of Hydrogen-like and Helium-like ions are some of the most well known features in X-ray spectra. . (Hydrogen meaning only one electron, helium 2 electrons)
He-like ions often produce a series of triplet lines. The emission line O VII (which come as a triplet) have enormous diagnostic power.
Ions in DEM L71:Hydrogen - like ions: Fe XXVI, O VIII, C VIHelium - like ions: Fe XXV, O VII, Ne IX.
The spectrum of both AB Doradus and DEM L71 were very similar. The spectrum for MCXC J1023.6+0411 looked very different as it contained a redshift.
The strongest line in the MCXC J1023.6+0411 spectrum was at a wavelength of 24.5 A°and assumed to be O VIII, in which case it’s actual wavelength is 18.97 A° . From this the redshift was able to be calculated in the spectrum using redshift equation.
Third slide : CIELO-RGS is a Catalog of Ionised Emission Lines Observed by the Reflection Grating Spectrometer (CIELO-RGS) on board the XMM-Newton space observatory.
The aims of the catalogue is to make easier use of emission features in the archive of public RGS spectra.
Catalog includs over 12000 emission lines from X-ray binaries, active galactic nuclei, supernovae remnants, stars and clusters of galaxies.
AtomDB is an atomic database designed for X-ray plasma spectral modeling.
A small subset of ions were selected from the enormous AtomDB database.
This was done by selecting the most prominent emission lines from 2D intensity plots of the 3 astronomical objects (binary star, supernova and galaxy cluster), and then matching these prominent lines with the closest and strongest ions from AtomDB.
Fourth slide:
An algorithm was made to match identified ions from AtomDB with the CIELO-RGS catalog. The program matches each (astronomic) line from CIELO-RGS with one or more ionic emission lines from the table of identified ions from AtomDB. A match between the catalog and ions in the AtomDB database is based on a difference in wavelength within 1%.
Histogram for the number of matches in the 12000 observation lines in the catalog.
As can be seen from the graph most of the observations have no ”suitable” match ie. no match within 1% tolerance.
The number of observations with zero matches is 6071, which is over half the catalog.
Conclusion
The final algorithm successfully takes data directly downloaded from CIELO-RGS catalog and can find the nearest matches within a percentage and also the match with the minimum percentage difference for each observation.
This algorithm would not be replicable on a bigger scale. To be expanded to more spectra, a bigger database than 40 ions would have to be created.
For future work to make it more accurate a probability could be assigned to each line. With the relative probability being, for example, whether this line was likely to be seen in a supernovae or not.
Overall this algorithm was a small step towards automating spectroscopy in Astronomy.