Atomic Spectroscopy at Imperial College

We are funded by the STFC of the UK to study atomic and molecular spectra of importance to astrophysics. The principal objective of our work on atomic spectrometry has been to upgrade the atomic database of wavelengths, energy levels and transition probabilities for astrophysically important atoms and ions to the accuracy required to exploit the high quality of the observations from current space- and ground-based telescopes. The age of the pre-FTS (Fourier Transform Spectroscopy) database is disguised by the fact that relatively modern compilations are actually based on spectra recorded many years earlier - frequently in the 1930s and in some cases still earlier. Improvements in the accuracy of wavelengths (and hence of atomic energy levels) of an order of magnitude are now needed for modern astrophysical applications, along with information on line broadening effects such as hyperfine structure and isotope shifts. This improvement can be readily achieved by the combination of high resolution, good signal-to-noise ratio, and accurately linear wavenumber scale offered by FTS.

Historical background

At the time that our first UV FT spectrometer became functional (the late 1980s), the FT spectrometer at NSO, Kitt Peak had already been used to record high quality laboratory spectra in the visible and infrared regions, and there was a clear need for the UV and VUV spectra, not only to complete the line lists but also to allow full analysis of the atomic energy levels. The importance of the analysis is not only to improve the accuracy of the known energy levels (which in turn feeds into semi-empirical calculations of atomic parameters) but also to identify new energy levels and classify many of the previously unidentified lines in both stellar and laboratory spectra. We have also measured experimental transition probabilities and oscillator strengths, important in elemental abundance determinations in astrophysical objects. We have studied line broadening effects such as hyperfine structure splitting and isotope shifts. Our policy has been to tackle the most important spectra systematically, while also making measurements on specific lines that are of high importance and particular interest to astronomers.

Our UV FT spectrometer was first applied to the spectrum of iron, the most abundant of the astrophysically important transition elements. The spectra of Fe I and Fe II were recorded at Imperial College from the visible to 200 nm and (later) into the VUV region, using as source an iron-neon hollow cathode lamp. These spectra were combined with infrared spectra from the National Solar Observatory to make a revised multiplet table for Fe I . The analysis of the Fe II data is nearing completion by former Group member Gillian Nave, now at NIST(USA). The spectra were also used to generate recommended iron wavelength standards in the visible region. This paper was followed by three others, giving precision Fe I and Fe II wavelengths in the ultraviolet and precision Fe I and Fe II wavelengths in the infrared, and, finally, precision vuv wavelengths of Fe II.

New Atomic Data for Astrophysics Applications.

Large scale analysis of atomic spectra

• Continuing with the iron-group elements, we have recorded the spectra of nickel, chromium, cobalt, vanadium, titanium, and manganese from the visible region to the VUV.


• The analysis of the Ni I spectrum , using Imperial College spectra, was carried out by U.Litzen (Lund University). Analysis of the Ni II spectra, containing most of the VUV lines, is a high priority for the Imperial College Group.


• Work is currently in progress at Imperial College on Mn II, and the final Mn I spectral analysis is in preparation for publication.


• We have carried out an extensive analysis of the cobalt spectra. The term analysis of Co I resulted in 60 new atomic energy levels and the classification of several hundred previously unidentified lines. The improvement from the term analysis of Co II was even more dramatic: over 200 new energy levels were found and the number of classified lines was doubled. The data was also used in theoretical calculations of eigenvalues and eigenvectors to predict energy levels and transition probabilities in excellent agreement with experimental values. In addition, the hyperfine structure in Co I was investigated: over 1000 line profiles were fitted to yield values of the magnetic dipole hyperfine interaction constant for nearly 300 levels, for two-thirds of which no previous values existed. This was the first large scale systematic study of hyperfine structure for a particular atomic species. ( Our current work on hyperfine structure includes Co II, Mn I, Ta I and Ta II, V I and V II.)


• We recently published the results of a large scale analysis of the V I spectrum, improving wavelength and energy level accuracy by over an order of magnitude. Our analysis of the V II spectrum is in preparation for publication.

Accurate transition probabilities

More accurate and more complete transition probabilities, or oscillator strengths are also much needed by astrophysicists. Applications of this data include the new stellar models (non LTE/3D), new regions of interest such as the IR, particularly for elemental abundance studies, and studies of Galactic Chemical Evolution.

Careful intensity calibration of high resolution laboratory spectra allows measurement of relative intensities of sets of lines from common upper levels, and, provided the sets are complete, these so-called branching ratios can be combined with measured or calculated level lifetimes to give absolute transition probabilities.

We have applied this method to:

• Transition probabilites in Ti II for about 700 lines from 89 levels.


• As part of the FERRUM project, we have measured sets of branching ratios in Fe II in the UV-VUV spectral region. We combined these with level lifetime measurements carried out at the Lund Laser Centre to yield Fe II oscillator strengths.


• Our project on log gfs for IR lines (H band) iron group elements is ongoing, driven by urgent needs of large scale stellar elemental abundance measurements for Galactic evolution studies (APOGEE).

Atomic data for cool stars

A current area of research is atomic spectroscopy for applications in studies of sub stellar objects.  Further details are given for this at: Cool Star project.

Atomic data for doubly ionised iron group elements

We have an ongoing series of projects studying doubly-ionized iron-group spectra, which are important in hot star spectra, using a Penning discharge source as a light source. The majority of the strong transitions in these spectra fall in the UV-VUV region, in the region uniquely accessible for Fourier transform spectroscopy by our VUV Fourier transform spectrometer. The visible-VUV spectra recorded at Imperial College are supplemented by IR spectra recorded at NIST (National Institute of Standards and Technology, USA). The analysis of spectra of Fe III is underway, and is leading to improvements in accuracy of over an order of magnitudein wavelengths and energy levels. Wavelengths standards of Cr III have been published, and our analysis of the Co III spectrum is in progress.

Atomic data for other, non-iron group elements

In a collaboration with the Graz Atomic physics group (Graz University, Austria) the large scale analysis of the spectra of Ta I and Ta II is being undertaken; finding new energy levels, and identifying many previously unknown transitions.

We completed measurements of the Ag I spectrum , resulting in at least order-of-magnitude improvements in accuracy of wavelengths and energy levels. In addition to the astrophysics needs, this Ag I study was driven at the time by atomic physics needs, where accurate wavelengths were required in projects looking into using a beam or atomic fountain of laser-cooled silver atoms as an optical frequency standard.

Selected examples of applications of our atomic data arising from astronomer requests for small scale atomic data for particular lines:

• Atomic data in the visible spectral regions for new non-LTE/3D stellar atmosphere models. Our data for particular lines in Co and Mn has been used to determine new abundances for these elements in the sun and other stars. The new solar abundances have far reaching consequences for the field, as all stars are compared to the sun, the solar abundances are effectively a "cosmic yardstick".


• Interpretation of the spectra from Hubble Space Telescope of the chemically peculiar star Chi Lupi. As part of the Chi Lupi pathfinder team our accurate atomic data has been used in analysis of this rich, high resolution spectrum. For example: our hyperfine structure analysis for cobalt led to the first accurate determination of cobalt abundance in this star, and analysis of the hyperfine structure and isotope shift in selected lines of Pt II led to the identification of isotopic anomalies in the platinum abundance.


• Time variation of fundamental physical constants may in fact have varied with time. A test of this hypothesis is to compare laboratory measurements of certain resonance lines with telescope observations of the same lines formed in very distant quasars with large red shifts. We have collaborated in this with J.Webb and his colleagues at the Australian National University by measuring absolute wavelengths with an uncertainty of 0.002 cm^-1, or about 0.008 pm, for resonance lines of Mg I and II, Cr, Zn and Ni, and Ti . Our atomic data is being used in the ongoing investigations looking for evidence for the time variation in fundamental constant, the fine structure constant alpha, at present results appear to be statistically significant but not conclusive.


• Thorium-neodymium clock, a method of determining the age of the Galaxy from the ratio of stellar abundances of thorium, having only one long-lived isotope, and neodymium, a stable reference element. The Th line chosen is blended with several lines, and we undertook two studies to unscramble these. The first involved weak lines of iron and nickel, and the second weak lines of cobalt and vanadium, including hyperfine structure in the former. Galaxy age estimations are still in progress using this new data.

Industrial Applications.

Glow Discharge Studies

The Imperial College Spectroscopy Team were members of the EU Marie Curie Research Training Network on Glow Discharge Sources, GLADNET. The Glow Discharge Source is used in industry for analytical purposes, and is capable of investigating the composition of very thin coatings and layers. The industrial sectors interested in GD are of crucial importance for Europe. They include life sciences (biocompatibility of medical implants etc.), nano-technology (composition of very thin layers etc.) and thin films, but also more traditional sectors such as car manufacturing. As applications of this technique increase, and accuracy improves, it has become vital to understand the physics of this technique in order to correctly interpret measurements made with it in industry.

As part of this network our PhD student, Sohail Mushtaq, (PhD awarded July 2011), has visited industrial partners for transfer of knowledge, and we have hosted extended training and knowledge transfer visits from GLADNET partners. Our high resolution VUV Fourier Transform Spectrometer was used to study the effects of trace gases, such as oxygen, on the Glow Discharge. The results are of importance in improving the accuracy and reliability of the Glow Discharge analytical technique used in industry in compositional analysis of materials. Our results have been published, and this work is now continuing in collaboration with London Metropolitan University. Imperial College final report on GLADNET project.