This handbook is designed to provide a selection of the most important and frequently used atomic spectroscopic data in an easily accessible format. The compilation includes data for the neutral and singly-ionized atoms of all elements hydrogen through einsteinium (Z = 1-99). The wavelengths, intensities, and spectrum assignments are given in a table for each element, and the data for the approximately 12,000 lines of all elements are also collected into a single table, sorted by wavelength (a "finding list").
More complete data for a smaller number of the most persistent lines of each spectrum are given in additional tables for each element. In addition to the wavelengths and intensities, the energy levels and transition probabilities (where available) are listed for a total of about 2,400 lines in these tables. We also give a separate table of energy level data for each spectrum which, although incomplete, includes levels additional to those involved in the persistent-line transitions.
More complete data than those selected for this Handbook can usually be found in references given with the tables for particular spectra. The data from most of the NIST compilations we have used are available online from the Atomic Spectra Database (ASD; see [MFKM99]). In addition to more extensive data for many of the spectra in this Handbook, the ASD has data for higher ionization stages of many elements and includes the references.
Although we have made heavy use of previous compilations, our tables for the great majority of elements include at least some data compiled by us from more recent original literature and, in some cases, from unpublished material. Our most extensive use of data from the original literature has been for the heavier elements. Although the data are incomplete, our wavelength and energy level tables for these elements, especially, comprise a supplement to the ASD. For example, the current version of the ASD includes energy-level data for only two spectra (Mo I,II) of the 72 spectra of the neutral and singly-ionized atoms of the elements Rb to Ba (Z = 37-56) and Hf to Ra (Z = 72-88). No complete and critical compilations of energy levels have appeared for most of these spectra since vols. 2 and 3 of Atomic Energy Levels [M52, M58]. For the actinide elements Ac-Es ( Z =89-99), we were able to rely almost entirely on the very complete compilation by Blaise and Wyart [BW92b].
For some spectra, especially some of the lighter elements, we have taken data from existing compilations that have been superseded by data reported in more recent literature.
A small selection of non-spectroscopic atomic data has been provided for each element. Included are the atomic number and weight and a list of naturally occurring isotopes, including the isotopic mass, the relative abundance, the nuclear spin (in units of h/2π), and the magnetic moment (in units of nuclear magnetons). For elements with no naturally occurring isotopes, the most commonly observed isotopes are listed. These data are taken from J. Emsley [E95].
For each of the elements a list of the strongest lines in the spectra of the neutral and singly-ionized atoms has been compiled. This list includes the wavelength, the ionization stage, the reference for the wavelength measurement, and an intensity. Unless otherwise noted, the spectroscopic data in this Handbook pertain to the naturally occurring isotopic mix for each element.
The wavelengths for many spectra have been taken from Reader et al. [RCWM80]. Wavelengths given to three decimal places in [RCWM80] have a stated uncertainty of less than 0.001 Å, and many of the two-place wavelengths in [RCWM80] are rounded off from three-place values in the original literature. For all transitions with wave numbers greater than 50,000 cm-1 the wavelengths listed are vacuum wavelengths; for those less than 50,000 cm-1 air wavelengths are given.
Laboratory observations carried out since the publication of [RCWM80] have yielded improved wavelength data for many spectra, including more accurate wavelengths, increased range of wavelength coverage, and more reliable assignments of observed lines to particular spectra. We have used more recent data for many of the spectra. However, it is important to note that our retention of wavelengths from [RCWM80] for any particular spectrum does not imply that more accurate data do not now exist.
In compiling data from the original literature, we have in many cases departed from the practice of [RCWM80] by quoting wavelengths without rounding off, especially in cases where the literature values were given to three or fewer decimal places. The original references should be consulted for uncertainty estimates.
Unlike the other tabulated data, the relative intensities are not basic data and must be used with caution. The relative intensities of the spectral lines observed for any element depend upon the light source and excitation conditions. Thus, even if the relative intensities observed in a particular experiment are adjusted to correct for the wavelength dependence of the sensitivity of the spectrometer and detector, the intensities will in general be different from relative intensities given by a previous observer or tabulated in a compilation such as this one. With a caveat that users should keep these considerations in mind, we list a relative intensity for each line. For some lines the wavelengths are so close to another that it was impossible to make two separate intensity measurements. For those lines the intensity of the blended line is given for each and both intensities are marked with an asterisk.
For uniformity we have assigned an intensity of 1000 to the strongest line(s) of each spectrum. In most cases the chosen line (or lines) can reasonably be regarded as the ultimate line (Sec IV). The relative intensities for most spectra here are based on values from [RCWM80]. We have attempted to give improved intensities for some spectra by using more recent and apparently more accurate data than those available to the compilers of [RCWM80].
It should be noted that the intensities in [RCWM80] for lines of neutral and singly-ionized atoms of about half the elements (mainly nd- and nf-shell metals) were taken at least in part from [MCS75]. These intensities were obtained from observations of 10-A, 220-V direct-current arc discharges between copper electrodes having 0.1% of the element under investigation. The relative intensities were put on a linear scale by the use of standardized lamps.
For several spectra we have altered some of the intensities found in the literature to give smoother transitions between wavelength regions covered by different observers. We have also adjusted reported intensities given in a single reference in some cases where the reported values were clearly affected by strong self-absorption and/or by large wavelength-dependent non-linearities. Such adjustments were necessary in order to assign the largest intensities to the inherently strongest persistent lines (usually the ultimate lines).
Lines we have selected as persistent (Sec IV) are indicated by the letter "P" following the intensity. For some spectra, other descriptive codes have been included to characterize the line shape or give other related information. They have the following meanings:
b - band head
In general, the character of a line depends on the spectroscopic source used, the resolution of the spectrometer, etc. Most of the line characterizations in our tables are quoted from [RCWM80], so that the characters given for lines of metallic elements usually pertain to the arc source used by Meggers et al. [MCS75] (see above). In some cases we have given character notations from [RCWM80] for lines for which the tabulated wavelengths were obtained with a very different (low-pressure) source.
In spectroscopic observations made with low concentrations of a particular element relative to other substances in the source, the number of observable lines of the element is found to decrease with decreasing concentration until only the most "persistent" or "sensitive" lines remain. Some authors refer to the last such line(s) as the raie(s) ultime, i.e., the ultimate line(s). Although the ultimate lines depend in principle on the source, the spectrometer, and other features of the experiment, a relatively small group of lines can be specified for each element that will include the ultimate lines as observed over a broad range of experimental conditions. We designate our selection of these lines "persistent lines."
The strongest persistent lines usually include one or more resonance lines, i.e., transitions to the ground level or term. We include at least one of the resonance lines in our persistent line table for each spectrum. The most sensitive or ultimate lines for many spectra lie in the vacuum-ultraviolet region (wavelength < 2000 Å). In such cases we have tried to include some lines above 2000 Å in the persistent lines list. We have also tried to make these tables more generally useful for many spectra by covering broader wavelength ranges than most tables of this sort. For all transitions with wave numbers greater than 50,000 cm-1 the wavelengths listed are vacuum wavelengths; for those less than 50,000 cm-1 air wavelengths are given.
In addition to the information given in the strong lines table, the list of persistent lines includes the energy levels involved in the transition, complete with configuration, term designations, and J values. Where available, the transition probability is also given, along with the reference from which it is taken.
The values are listed as Aki in units of 108 s-1. These Aki values can easily be converted to oscillator strengths, fik, gifik, or log(gifik), or line strength, S, by using the following formula:
where i refers to the lower energy level, k refers to the upper level, λ is the wavelength in Ångstroms, and g = 2J+1 for a given level.
The transition probability data are taken primarily from three compilations and from references cited therein. The NIST compilation [FW96] contains transition probabilities for about 9000 lines, covering most elements. Major recent compilations by Morton for elements from H to Ga [M91, M03] and from Ge to Bi (plus Th and U) [M00] have data for lines longward of the Lyman limit (911.754 Å) and include a number of useful new references.
The NIST transition-probability compilations include values with uncertainties ranging from 1% to larger than 50%, with an approximate uncertainty range being indicated for each line by an assigned letter. We have not included or assigned such letters here, but this information is given in the NIST publications and in the ASD for all transition-probability values taken from these sources. For data taken from the original literature or from non-NIST compilations, the user can consult the cited references for accuracy estimates.
Data pertaining to the two levels classifying each line are given with data for the lower level above that for the upper level. Included are the level values, configurations, term names, and J values for the levels classifying the line. The energy-level classifications for a few persistent lines are not known, as indicated by the absence of level values. The accuracies and spectroscopic designations of the levels are discussed in the next section.
The tabulated energy levels represent a selection of the available data for each spectrum, including all levels involved in the persistent-line classifications. At least the lower levels of the ground configuration and other low-lying configurations are given. The levels of some of the simpler spectra are given complete up through the highest tabulated level, but most of the known energy level structures of the more complex spectra are omitted here. The ionization energies are included except for a few spectra for which no reliable values are available. The reference for each level represents the source of the level value.
Estimated uncertainties for the levels can usually be found in the references. In our tables, the uncertainties are only roughly indicated by the number of decimal places or significant figures in the values. The uncertainty corresponding to a particular number of decimals may easily vary by an order of magnitude, however, even within the data for one spectrum. In most cases, the uncertainty is probably between 1 and ~30 units in the last decimal place or significant figure, but still larger errors can occur. The uncertainty in the relative position of two levels having different uncertainties is at least as large as the greater of the two level uncertainties. In this regard, it is important to notice the number of decimals given for the ground level; the uncertainties of the absolute values of the excited levels and ionization energy are at least as large as the indicated ground-level uncertainty.
The configuration and term notations are standard for NIST compilations. Explanations of the notations for the different coupling schemes and of the arrangement of the data can be found in [MZH78], [MW04] and online as a part of the "Help" section of ASD [MFKM99]. Some levels of complex spectra are given without term names, either because the level has not been interpreted theoretically or because the calculated eigenvector for the level yields no meaningful unique name. Some of these levels have been assigned simple numerical designations under "Term". The parity of levels lacking designations is indicated by a degree symbol in the "Term" column for odd-parity levels.
This table gives the wavelength, intensity, spectrum, and reference for each line in this compilation, listed in order of increasing wavelength. Although this list has fewer lines than the finding list of [RCWM80] or [MCS75], it includes some lines not given in the earlier publications.
We wish to acknowledge the generous help of several NIST colleagues. Arlene Robey gave expert assistance with data-handling and bibliographic aspects of the work. The cooperation and helpfulness of Jonathan Baker, Svetlana Kotochigova, Peter Mohr, Victor Kaufman, Gillian Nave, Joseph Reader, Craig Sansonetti, and Jack Sugar in supplying unpublished data are greatly appreciated. We thank Jeffrey Fuhr and Wolfgang Wiese for guidance in finding and assessing transition-probability data. We are also grateful for the assistance of Robert Dragoset and Gloria Wiersma in polishing the appearance of this website.
A number of colleagues from other laboratories have kindly provided unpublished data or electronically readable files of published data. We are very grateful to Vladimir Azarov, Jean Blaise, James Brault, Gordon Drake, Sveneric Johansson,Gabriele Kalus, Alexander Kramida, Ulf Litzen, Donald Morton, Byron Palmer, Juliet Pickering, Alexander Ryabtsev, Toshizo Shirai, and Ward Whaling for this help.