ATOMDB Version 3.1.3 Release Notes (2025-March-20)

Minor bugfix: there is an issue with the argon continuum in kT=8keV plasmas in version 3.1.2. This is due to an error compressing the pseudocontinuum which caused an energy bin of zero width.

This did not affect PyAtomDB, but in XSPEC caused an infinitely large/small bin flux. If you did not have issues with NaNs appearing in your fit statistics, you are unaffected. 

In addition, the "units" column of the continuum and pseudocontinuum emissivities had been reset to keV, instead of photons cm^3 s^-1 keV^-1. This is purely a labelling issue, the numbers remain unchanged in the new release. 

There are no other changes to this release. Our thanks to Annie Heinrich for spotting the argon issue.

-- Adam Foster


ATOMDB Version 3.1.2 Release Notes (2025-January-31)

Error in the A-values for inner shell (n=2 ->1) transitions of Fe VII to Fe XVIII (excluding Fe XVII) have been discovered leading to some double counting. These have now been rectified. Thanks to Tyler Holland-Ashford for spotting this issue.

-- Adam Foster


ATOMDB Version 3.1.1 Release Notes (2025-January-16)

Minor bugfix: There is an error in the continuum of neutral ions at high temperatures (T>10^8K). This affects only the non-equilibrium plasma, with T>10^8K, and a very short timescale (ne*t < 1e9 cm^3 s^-1).

This was due to an error calculating bremsstrahlung emission. Thanks to Taras Kuzyo for finding this. As the error causes XSPEC and pyatomdb to not return a spectrum, if your model returned a spectrum then you are unaffected and no analysis needs to be redone.

Equilibrium file and ionization balance files are unaffected.

Also, AtomDB 3.1.0 was packaged with incorrect release notes, which have now been updated.

-- Adam Foster


ATOMDB Version 3.1.0 Release Notes (2025-January-02)

This is a major update to a large portion of the database, affecting nearly every ion. Here is a summary of the points:

(1) New ionization rates from Urdampilleta 2016 have been included.
(2) Multiple ionization rates from Hahn, Mueller and Savin 2016 have been included.

These new ionization rates generally increase the ionization state (so ions are more highly ionized at lower temperatures). 
The APEC (equilibrium) models will incorporate these changes automatically.
For all non-equilibrium models (nei, rnei, pshock, etc), the new eigenvector files will have to be initiated each time (see AtomDB)

Bugfix:
(3) Bugfix to the electron-electron bremsstrahlung emissivity (thank you Keith Arnaud)
(4) Fix to the fluorescence yield of n>=3 states (previously this was too low due to an error; n=2->1 lines were ok)

Other data updates:
(5) Addition of inner shell excitation rate up to n=15 for Be-like and higher charged ions of all elements
(6) Addition of DR satellite lines up to n=15 , n'=4 for Be-like and higher charged ions of all elements
These collectively update the previous holdings, which where n'=2, up to n=5 and only for H- and He-like ions.
(7) Addition of high n excitation rates for Lyman series and He-like high n transitions
(8) Update to spontaneous emissiion probabilities (a-values) of Li-like inner shell transitions

The physics of these updates will be covered in detail in an upcoming publication.

-- Adam Foster


ATOMDB Version 3.0.9 Release Notes (2017-August-07)

Bugfix: 
When incorporating inner shell fluorescence lines from Fe XVIII-XXIV, we have double counted the rates, leading to the inner shell lines being twice as strong as anticipated. We now fix this. This bug was introduced in v3.0.4. Our thanks to Liyi Gu for spotting this issue.

Updates: 
Fixed wavelengths of Na-like ions to NIST values.

Incorporated DR lines for H and He-like ions of all minor elements (i.e. F, Na, P, Cl, K, Sc, Ti, V, Cr, Mn) from Safronova. This is the same source as used fro all other elements already.

Updated DR line files to have an additional HDU with information describing the levels, so no more mystery "10024" levels with no obvious meaning.

-- Adam Foster

ATOMDB Version 3.0.8 Release Notes (2017-February-17)

Data from 2005 Whiteford was replaced by data from 2001 for Fe XXV collision strengths.

Data for H-like excitation was replaced with H-like data from Li et al 2015, but with high energy extrappolation for the final points

Wavelength and energy levels for Li-like valence shell transitions was updated to NIST values

Nickel XVII had one transition with an anolmalously large collision strength, this has been fixed.

-- Adam Foster

ATOMDB Version 3.0.7 Release Notes (2016-December-05)

When updating the energy levels to match NIST values, this resulted in some very weak transitions of He-like ions flipping to a negative wavelength. This mistake has now been fixed.

This caused no significant problem for spectra generated with PyAtomDB, but interacted in XSPEC in a strange manner resulting in "turning off" the O VII lines at ~ 0.1keV.

Thanks to Shinya Nakashima and Keith Arnaud for spotting and helping fix these issues.

-- Adam Foster

ATOMDB Version 3.0.6 Release Notes (2016-November-09)

A bug was discovered in creating the apec python code which led to double counting of low-lying recombination driven lines in Li, He and H-like ions. This has been corrected and new emissivity files have been created.

-- Adam Foster

---
ATOMDB Version 3.0.5 Release Notes (2016-October-28)

During packaging of 3.0.4, several elements were omitted and also there issues with the emissivities of some of the lowest energy levels.

This issues have been fixed. As some people had downloaded the files before the error was observed, we have incremented the version number to avoid confusion.

-- Adam Foster

---


ATOMDB Version 3.0.4 Release Notes (2016-October-04)

Large energy levels and wavelength updates for H through B-like ions, to match
NIST wavelengths.

Change to ionization balance calculation. The data from Bryans has been updated.
This reflects a change from interpolated grids to fit formulae, as opposed to an
actual change in the atomic data.

-- Adam Foster

---
ATOMDB Version 3.0.3 Supplementary Release Notes (2016-July-18)

Minor fix for interaction with XSPEC
The line data for the equilibrium models should have been sorted by the
element, ion and then reverse wavelength before release, to take advantage
of XSPEC speed improvements. This was not done, leading to spurious lines in the
equilibrium models (APEC, BAPEC, and all "v" letters thereof). 

This has been fixed. If your data has been fixed, the md5 checksum (done by 
either "md5sum apec_v3.0.3_line.fits" or "md5 apec_v3.0.3_line.fits", 
depending on your operating system) should be:

c66905cb963f076b6ff18697b433ca77  apec_v3.0.3_line.fits

The old data has md5 checksum:

f034f0c8ca110fdd113adddb872e242d  apec_v3.0.3_line.fits

Note that there have been no other changes to the files, this is simply
reordering the existing lines.

-- Adam Foster

---

ATOMDB Version 3.0.3 Release Notes (2016-May-19)

One bugfix and two updates from v3.0.2:
(1) Update: We have updated the wavelengths for fluorescence lines to those
from literature values, taken from a series of papers by Palmeri, Mendoza and
others (2008ApJS..177..408P, 2011A&A...525A..59P, 2008ApJS..179..542P, 
2012A&A...543A..44P, 2003A&A...410..359P, 2004A&A...414..377M, 
2003A&A...403.1175P)

(2) Update: We have corrected the fluorescence yields for lower Z ions, in 
particular Fe IX and below. These were previously overestimated as we were
missing an autoionization channel. In addition, the fluorescence yields have
been made to match the literature values, taken from the same papers as above.

(3) Bug Fix: The excitation rate for the Fe XXIV shells has been updated to the
data of 2011A&A...528A..69L

-- Adam Foster

---
ATOMDB Version 3.0.2 Release Notes (2015-Jun-10)

One bug and two updates have been fixed in AtomDB:
(1) Bugfix: fro unidentified reasons, the import of L-shell valence electron
electron collision excitation rates into the AtomDB database failed. This has
now been fixed. It did not affect "common" ions, e.g. iron L-shell, but did
affect minor elements, e.g. chromium.

(2) Improvement: Inner shell excitation & ionization rates, and hence 
fluorescence lines, for Fe I-VII are now included.

(3) Revamp: In order to keep file sizes manageable, the minimum emissivity for
lines to be included in the _nei_line.fits file has been changed as a function
of the ion's population in an equilibrium plasma:

Ion fraction (%)    Min emissivity (ph cm^3 s^-1)
    f > 1                     10^{-20}
    0.01 < f <= 1             10^{-19}
    f <= 0.01                 10^{-18}
    
Lines with emissivity less than this value are added to the pseudocontinuum
instead (calculated on 1eV energy bins)

-- Adam Foster

---
ATOMDB Version 3.0.1 Release Notes (2015-Mar-25)

Two bugs have been corrected in AtomDB:
(1) An error in the apec_python.py code in interpolating the rate coefficients
was fixed. This was caused by a numpy issue allowing a float to be
simultaneously >= a limit and < limit, presumably due to precision. This was
fixed by checking for when the numbers are identical independently of the
interpolation routine.

(2) An error in the Mn V data was discovered and fixed where rate coefficients
were tabulated as collision strengths.

-- Adam Foster

---
ATOMDB Version 3.0.0 Release Notes (2015-Mar-01)

AtomDB version 3.0.0 includes major updates for NEI plasmas. Massive inclusion
of inner shell excitation and ionization data, complete with autoionization
rates for all ions. The XSTAR photoionization database has been incorporated
too. New data formats for autoionization have been created (type
"AI") and for XSTAR photoionization data (type "PI").

A fuller writeup is beyond the scope of these notes. But it's big. See
publications in preparation. Every single file has had changes.

In addition, due to the increasing size of AtomDB, we have removed all the old
data files from tarball. We used to package previous versions of each file (e.g.
the old ionization and recombination rates). These are now available only in the
older tarballs (e.g. v2.0.2).

-- Adam Foster

---
ATOMDB Version 2.0.2 Release Notes (2012-Jan-17)

2 bugs have been fixed. One caused double counting of the
excitation-autoionization process in the ionization balance
calculation in several ions. Whilst nearly all Li and Na like ions are
affected, the ions most noticeably effect is for the
iron M shell ions, e.g. Fe+8, Fe+9. 

The other bug lead to a small number of lines of lithium like Al, Ar
and Ca having negative wavelengths due to an error importing the CHIANTI
database. These lines were then excluded from the _line.fits
files. THey ar enow included properly.

In addition, some further minor corrections were made:

fe_10_LV_v2_0_1_a.fits -> fe_10_LV_v2_0_1_b.fits:
   ELEC_CONFIG were all incorrect
   Error due to incorrect import of Chianti data.
fe_12_LV_v2_0_1_a.fits -> fe_12_LV_v2_0_1_b.fits:
   ELEC_CONFIG were all incorrect
   Error due to incorrect import of Chianti data.
ni_14_LV_v2_0_1_a.fits -> ni_14_LV_v2_0_1_b.fits:
   S_QUAN of level 121 fixed
   ELEC_CONFIG of level 143 fixed.
   Errors due to original Chianti mistake.
ni_16_LV_v2_0_1_a.fits -> ni_16_LV_v2_0_1_b.fits:
   LEV_DEG of level 32 fixed.
   Error due to original Chianti mistake.

Finally, the IONPOT keyword was added to the header of the IR fits
files, using values from NIST.

-- Adam Foster

---
ATOMDB Version 2.0.1 Release Notes (2011-Jun-23)

A bug was found in the APEC code affecting the calculation of radiative
recombination continua, causing an overestimate at certain temperatures. This 
did not affect v1.3.1, it was introduced in v2.0.0 when correcting quantum 
numbers in the database. This has been rectified.

While investigating this, several changes were also made to other data files:

 - For all non-hydrogenic ions, the ground state RRC calculation uses the Verner
  (1986ADNDT..34..415C) data. In some cases this had been set to zero since
   v1.3.1. It has now been restored for all ions.
   This has neccessitated changing the storage of PHOT_PAR parameters for the
   Verner type data. All updated LV files have been given the 2_0_1_a suffix.
   To provide continued access to legacy files, all LV files which are no longer
   used in the APEC runs have been incremented to their "b" versions after 
   changing their PHOT_PAR to be readable by the current version of APEC.
 
 - For hydrogenic ions, the RRC for many odd-Z ions had been turned off since
   v1.3.1. This has also been fixed for all H-like ions (updated versions given
   2_0_1_a suffix).
   
 - The TR_INDEX in the new IR files (ionization and recombination data) had
   become non-sequential. This had no effect on the output of the APEC code, but
   has been fixed (files given 2_0_1_a suffix).





-- Adam Foster

---

ATOMDB Version 2.0.0 Release Notes

Version 2.0.0 Improvements (2011-Jan-06)

The database structure for version 2.0.0 has not changed from earlier versions,
however nearly all the atomic data has been replaced with improved calculations.
All He-like collisional data has been replaced with R-Matrix calculations.
H-like data has been replaced with a combination of R-Matrix and Distorted Wave
data. Iron L-shell data has been included from R-Matrix calculations.

Quantum numbers identifying the various levels have been updated and/or
corrected, many were missing.

State Selective DR & RR rates have been included for all astrophysically
relevant ions, which can affect excited state populations significantly.

Total DR & RR rates are now those from Bryans+ 2009, with an adjustment for
updated DR rates into N-like ions, reflecting a further, more recent update to
the data which Bryans+2009 is based on.

Data taken from Chianti (usually for less-ionized ions) has been updated from
Chianti 2.0 to Chianti 6.0.1.

A new website has been established, www.atomdb.org. This will detail more
specifically (i.e. with examples and links to papers) the changes made here
and provides a place for feedback.

-- Adam Foster

---
Version 1.3.1 Improvements

Version 1.3.1 uses the same underlying atomic database (APED) as V1.3.0.
The only change is that the continuum emission is now calculated out to
50 keV, rather than stopping at 10 keV.  This means that ATOMDB can now
be used with redshifted observations, where previously the cutoff at 10 keV
became awkward.  The discerning user will notice the line emissivities are
very slightly changed, especially at high temperatures.  This is due to 
increased radiative recombination cascades from high-energy electrons; APEC
considers only electrons that contribute to the continuum, and so increasing
the continuum range from 10 keV to 50 keV increased the amount of RRC and
thus the cascades.  The change is less than 1% for all lines, however, which
is far smaller than our expected accuracy.

For completeness, we include the Version 1.3.0 Improvements and Caveats below.

Version 1.3.0 Improvements

1. This release contains dielectronic recombination lines for K
   shell transitions in Fe XVIII - Fe XXIV.  We have not yet included
   collisional excitation rates for these lines, although we expect
   to in the next release.  Dielectronic recombination, however, is the
   dominant source of 6.5-6.8 keV emission lines.  The data for these
   lines was kindly provided by Dr. Verne Jacobs, based on his paper
   Jacobs et al. 1989 Phys Rev A, 39, 2411.  

2. V1.3.0 also corrects an error in the Li-like Ni XXVI line emissivities
   at 10^8 K and above.  Incorrect values of the excitation rate coefficient
   were used leading to large overestimates of the Ni XXVI line emission for
   these temperatures.  The problem did not affect results below this 
   temperature. Thanks to Dr. Eugene Churazov for bringing it to our attention.

3. This release also corrects some transition probabilities in H-like ions.
   This error affects the line emissivity calculations for a low-density
   collisional plasma at the 1-2% level, but would be important if the 
   radiative rates found in APED are used for other purposes.  Thanks to Yair 
   Krongold-Herrera for discovering the problem.

Version 1.3.0 Caveats

1. We still have very little data on the delta n >= 1 transitions from
   L-shell ions of Ne, Mg, Al, Si, S, Ar, or Ca, with the exception of
   Lithium-like ions.  X-rays from these lines are largely below 0.25
   keV, and as a result, our spectral calculations in this range are very
   deficient.  We are working on this problem.

2. Only for some of the strongest lines do we include lines from
   levels above n=5.

3. Despite the addition of the Fe XVIII - Fe XXIV dielectronic recombination
   lines, we still do not include the so-called "cold" Iron line at 6.4 keV,
   which arises out of Fe I - Fe XVII ions.  We are also missing some 
   inner shell excitation transitions, for example from Fe XVI and O VI.
   We hope to include these data in the next release.

4. These models are derived from theoretical atomic data for which
   few experimental measurements have been made. Most emissivities are
   thought to be good to better than 30 percent; however, larger
   discrepancies are being found in some caes, likely the result of
   incomplete atomic physics. Such cases are under investigation as part
   of the Emission Line Project.

5. We note that the line ratio of Fe XVII 15.014 to Fe XVII 15.261
   is about 15 to 20 percent larger than the measured ratio (see Brown et
   al. 1998, ApJ, 502, 1015). While the discrepancy is within our
   expected uncertainties, it is quite likely that this discrepancy
   will be noticeable even at moderate (CCD) resolution. 
