This page summarizes all major changes and improvements to C96. In some cases the summary includes a link to the Hot Fixes page. Hot Fixes are minor changes to the coding to correct problems not yet fixed in the distributed source. These fixes will be incorporated in the next release of the code Here is a list of known problems with the current version of the code.
This version is strongly influenced by discussion held at the Lexington meeting on Astrophysics of Photoionized Plasmas. That meeting served to identify sources of differences in results among various codes and atomic data that have changed. The following section outlines changes that have been made since C94 was released in December of 1999.
new in next release
The default treatment of the Solomon process, the dominant H2 destruction mechanism in a PDR, has been changed from Tielens & Hollenbach 1985 to Bertoldi & Draine 1996. The resulting H2 column densities can be quite different (more than 1 dex in some cases). This has no impact on the predictions of the large H2 molecule, which uses self-consistent destruction rates rather than a simple approximation.
The TLUSTY model atmospheres may be used as part of the incident continuum. The data files are on the TLUSTY web site, and this command will read any one of their continua as a table. The syntax of the command is table tlusty "filename". The spectral energy distribution file "filename" will be used.
new in C96 beta 5, 2003 March 21
The cloudy executable now has a command line option.
Robin Williams (Cardiff) added an option so that, if you execute the code as
Tables generated by Starburst 99 can directly be used as part of the incident stellar continuum. Starburst 99 generates tables of emergent stellar continua as a function of burst age. The table starburst command has two arguments, a filename that must appear within a pair of double quotes, and the age of the burst, which will be interpolated from the continua in the file. The following command line would read data from the file "starburst99.dat" and interpolate to an age of 108 yr;
table starburst log age = 8, file = "starburst99.dat"
The normalization of the continuum to a luminosity or flux of photons could be inexact at the one or two percent level for older stellar atmospheres, and in the case of modern stellar atmospheres, with very strong spectral features, by as much as 6 - 7 percent. The logic was improved to be exact for all continuum shapes. Thanks to Torsten Elwert for finding this problem.
The chemical composition printed at the start of the calculation now has a separate block that shows the composition locked up in grains. This block of information is now printed after the details of the continuum (because the needed quantities have only been evaluated at that time).
The background command produces a cosmic background that is the sum of three parts - the thermal background, a component that peaks in the ultraviolet, and a power law. That power law component had been a pure unbroken power law that extended to both the high and low energy limits of the code. The extrapolation to very long wavelengths resulted in a large brightness temperature in that part of the continuum. This had no effect for low density gas but would result in significant free-free heating for dense gas (n > 1010 cm-3). The power law component is now approximated with that produced by the table power law command. Thanks to Christopher C. Stevenson for finding this problem.
The second generation (H-Ni) Rauch stellar continuum can now be used, in addition to the first generation continua.
Rate H2 forms on grains updated to Cazaux, S., & Tielens, A.G.G.M., 2002, ApJ, 575, L29
The entire chemistry network has been updated to the following references: Abel et al. (1997), Galli & Palla (1998), Hollenbach, Takahashi, & Tielens (1991), Maloney, Hollenbach, & Tielens (1998), Millar, Farquhar, & Willacy (1997); and Stancil, Lepp, & Dalgarno (1998).
The OI 1.31m, 1.13, 2.90m, and 4.60m lines had the wrong wavelength - they were 10,000 times too small. Thanks to Kevin Volk for finding this problem.
The high-n Lyman lines are not punched in the optical depths arrays. This will be fixed in the next release. Thanks to Adrian Turner for finding the problem.
Ionization by inner shell lines has been added. These include the 1s-2p transitions listed by Behar & Netzer (2002, ApJ, 570, 165) and the "UTA" lines in Behar, Sako, & Kahn (2001, ApJ, 563, 497). Many thanks to Ehud Behar for providing a complete list of the UTA transitions in the second paper.
Solar abundances have been updated to for C&O (Allende Prieto et al. 2002, 2001), while N, Ne, Mg, Si, and Fe are from Holweger (2001). The remainder of the first thirty elements comes from Grevesse & Sauval (1998). The solar oxygen abundance is dramatically lower than was previously assumed, and shows that emission lines in Orion had it right all along.
The column densities of more excited levels are now predicted. The set of excited levels now includes CI*, CI**, OI*, OI**, SiII*, CII* and CII**. These are printed when the print column density command is used to print all column densities. Access to these numbers is also possible with cdColm, and is described where cdColm is described, both in the header file cddrive.h and part III of Hazy.
The wavelength of the e- - e+ annihilation line has been set to zero. In the past the intensity was printed with a label of "e-e+" and "wavelength" of 511, indicating its energy in keV. The code now explicitly gives the units for the wavelengths and misidentities the lines as 511A. The wavelength is now zero to prevent misunderstandings. Thanks to Kirk Korista for pointing out this problem. 02 June 21.
Transition probabilities for S IV UV lines have been updated to Hibbert, A., Brage, T., & Fleming, J., 2002, MNRAS, 333, 885, the new values are significantly different from the previous ones.
new in beta 4, 2002 June 1
Names of several of the routines used to drive the code have changed, in an attempt to make the naming convention more uniform. The header file cddrive.h has a complete list of all public routines and a brief description of their function. Among others, cdGeTe is now cdGetTemp_last, CdEms is now cdEmis, and cdNoexec is now cdNoExec. Consult cddrive.h for an authoritative list of the current names.
The punch command no longer produces a title on the first line of output. I usually bring these files into a spreadsheet, and delete the title. But a title will be produced as the first line in the punch file if the keyword title appears somewhere on the punch command.
Electron scattering escape is now included in the escape probability for all lines. This does change the intensity of lines in dense gas at the 10 - 20% level. The process can be disabled with the no scattering escape command.
The fine structure splitting for the [O II] collision strengths has been reset to being proportional to the statistical weights of the levels. There has been a great deal of activity regarding the new collision strengths, both as a result of the Lexington meeting, where this was discussed extensively, and the recent paper by Copetti & Writzl, 2002, A&A, 382, 282.
Logic added to allow heavy-element ionization to increase dramatically with depth. The code saves time by not computing the ionization and populations of stages of ionization with trivial abundances. It never considered the case where the level of ionization of the gas increases dramatically, although this can happen with a rapidly accelerating wind. The code can now identify this case and allow the range of ionization considered to become larger in deeper regions.
HD cooling added using equation 6 of Puy et al. A&A, 1999, 345,
723. There is also a new
The 1910┼ line of 13C III is now predicted, using the transition rate given in Clegg, R.E.S., Storey, P.J., Walsh, J.R., & Neale, L. 1997, MNRAS, 284, 348. In the printout the line appears with the label 13C3 1910. The 12C/13C ratio is set with the new set 12C13C command. The code assumes that 12C+2 / 12C+2 is equal to 12C/13C.
new in beta 3, 2002, Feb 22
The algorithm for the quantum heating of grains has been significantly improved, and quantum heating of grains has become the standard treatment. This follows the treatment outlined by Guhathakurta, P., & Draine, B. T. 1989, ApJ, 345, 230. A Mie code for creating grain opacity functions for any size distribution and refractive index data has been included in the code.
The column density in CII* is now printed. Column densities will be printed if the print column density command is entered. The entry CII* is the column density in the excited level within the ground term.
The continuum energy mesh is now fully allocated at run time, rather than having static vectors for the information. This means that the resolution of the continuum mesh can be made as fine as you like without having to change vector sizes, and also makes it possible to change the continuum mesh resolution at run time. The continuum mesh is described in an external data file, continuum_mesh.dat, and it can be changed without recompiling the code.
Much of the development work on the He-like isoelectronic sequence has been completed, but some work, especially the atom's topoff and collisional effects, remain to be done.
The collision strength for [N I] has been set back to the 1977 value (see that update in beta 2). The discussion among the people at the Lexington meeting seemed to indicate that observations seem to rule out a collision strength as small as the new value.
print line sort command works again, now has option to limit the wavelength range that is printed. This command sorts the output spectrum by wavelength rather than by spectroscopic identification. The print line sort wavelength command will sort the lines by wavelengths. There is a range option to specify a lower and upper bounds to the wavelength. The command also has the form print line sort intensity to sort in terms of decreasing intensity.
Format of emission line wavelengths in output has been reorganized. In previous versions the wavelength was an integer, and the printed wavelength often had implicit wavelengths. For instance, the [NeII] 12.8m line was denoted as Ne 2 128. All wavelengths have been converted to floating numbers, and the number is followed by a letter to specify the units. Examples are He 1 3889A or C 2 157.6m. A indicates Angstroms and m microns. Commands that read in emission lines (for instance, the optimize lines command) need to specify all of the digits that are printed, as well as the "m" if the units are microns. As a result of this change some line wavelengths have changed.
Recombination contribution to [O II] 3727, 7325 lines updated to Liu et al., 2000, MNRAS, 312, 585. This can be quite important in some nebulae.
Atomic data for Fe19 7046, 1328.90A updated to Butler, K., & Zeippen, C.J., 2001, A&A, 372, 1083, data for Fe20 721A, 578A updated to Butler, K., & Zeippen, C.J., 2001, A&A, 372, 1078 and Merkelis, G., Martinson, I., Kisielius, R., & Vilkas, M.J., 1999, Physica Scripta, 59, 122
Recombination contribution to [N II] 5755 added using equation 1 from Liu et al., 2000, MNRAS, 312, 585. This can be quite important in some nebulae.
The slit and beam options, which had been on the sphere command, have been moved to the new aperture command. This simulates the effects of observing the computed structure with a long slit or small pencil beam.
The print flux and distance commands were introduced. The distance command sets the distance to the object from the Earth. If this distance is set, then the print flux command will tell the code to predict the observed fluxes at the Earth rather than emission line luminosities.
There are now print column densities and punch column densities, to output the predicted column densities.
Any number of Lyman lines for atoms and ions of the H-like and He-like iso-electronic sequences can now have their optical depths predicted with the punch optical depths command.
new in beta 2 of C96.00, Apr 2001
Collision strengths for [OII] were updated to McLaughlin, B.M., & Bell, K.L. 1998, J Phys B 31, 4317, line intensities changed at the 10 - 20 percent level.
Collision strengths for [NI] updated to Tayal, S.S., 2000, ADNDT, 76, 191 - these are 10x smaller for the 5200A lines!
The charge transfer network totally updated, including major changes in helium - heavy charge transfer. Detail of the new rates are on Phillip Stancil's web site. Hydrogen ionization reactions are more important than before. The biggest changes were in BLR clouds, where the partially ionized zone is more ionized as the result of updated chemistry. Balmer lines and continuum are 20-30 percent brighter as a result.
new in beta 1 of C96.00, April 2001
The grain physics has been updated as per Weingartner, Joseph C., & Draine, B. T. 2001, ApJS, 134, 263. The largest effects where that the work function for graphite has been changed from 0.5883 Ryd to 0.3235 Ryd and the sticking probabilities are much smaller. The photoelectric heating efficiencies are quite different as a result and the grain photoelectric heating rates have changed considerably.
The pgrains command makes it possible to resolve the grain size distribution function, solving for grain properties and emission as a function of their size. This is described in van Hoof et al. (2000).
The emission line output has been reordered, with comments to indicate the origin of the lines.
The wavelengths of lines between 1 and 10 microns are printed with one additional significant figure.
The wavelengths in the emission line output now have a character to indicate the units - these are either "A" for ┼ngstrom or "m" for microns. An integer to the right of this character indicates how many of the printed numbers are to the right of the decimal point. (This new letter and number should not be included when a line is identified to the program by wavelength - only the integer part should be specified, as it always has been.) The old format will be used instead if the command set oldwavelengths is entered.
The resolution of the continuum array has been increased across the sub-mm through UV range. This is to permit full resolution of the grain emission features now predicted.
CO is now a multi-level rigid rotor model, with any number of levels and full treatment of radiative transfer.
The boron iso-electronic and homologous sequences are now a better multi-level atom. As a result all five lines within the first intercombination multiplet are explicitly predicted. This includes the ions CII], NIII], OIV], SiII], and SIV].
The identifier in the printout for the summed intensity of the intercombination multiplet for the boron sequence has changed . The old and new labels are C 2 2326 => TOTL 2326, N 3 1750 => TOTL 1750,O 4 1402 => TOTL 1402, Si 2 2325 => TOTL 2325, and S 4 1406 => TOTL 1406. Each of these total intensities is now followed by the intensities of individual lines within the multiplet.
Intensities of the intercominbation multiplet in the boron sequence are unchanged at very low or very high densities, but can be nearly a factor of two stronger when the lines are optically thick, since now two or more lines are able to radiate while, in the equivalent two-level atom, only one line was visible. Intensities of the IR fine structure lines generally changed by less than 5%.
print faint command is now the print line faint command
The form of the optical depth scale used in the wind command has been changed to equation 3 of Caster, Abbott, & Klein, 1975. Previously it assumed constant velocity, and now uses the local value of the acceleration to establish the effective optical depth scale.
Hydrogen collision data have been updated to Anderson, H., Ballance, C.P., Badnell, N.R., & Summers, H.P., 2000, J Phys B, 33, 1255, and helium collision data to Bray, I., Burgess, A., Fursa, D.V., & Tully, J.A., 2000, A&AS, 146, 481-49.
punch line optical depths command added. The output file will list the line energy in Rydbergs, the log of the line optical depth, and a label giving the ionic species. The log of the smallest optical depth can be specified as an optional number on the line. The command accepts the units option to change the energy of the line to other units, such as Angstroms, microns, etc.
Charge transfer data base updated. Jon Slavin noticed that one of the He+ charge transfer reactions listed in Arnaud & Rothenflug (A&AS, 60, 425) had "disappeared" from the code. The entire CT database was revisited. The papers by Pequignot & Aldrovandi (1986; A&A, 161, 169) and Prasad and Huntress (ApJS ApJS, 43, 1) were used for several reactions that had no quantal calculations. This affected mainly atomic iron and silicon. Thanks to Jon Slavin for discovering this problem. 01 Jan 15.
The ISM oxygen abundance was changed to Meyer et al. (1998, ApJ, 493, 222).
background cosmic ray ionization rate changed to agree precisely with Tielens & Hollenbach 1985a (ApJ 291, 722) Table 10. A range of values within an order of magnitude of their value can be found in the literature. Exact agreement with their values makes PDR comparisons easier.
Hydrogenic bound-free cooling rate coefficients now from LaMothe & Ferland 2001. These are exact for all levels of hydrogenic species. The previous values for n<15 were valid, but no accurate rates for n>=15 were previously known.
The transition probability for SiIII 1892 has been updated to Callegari, F., & Trigueiros, A.G., ApJS, 119, 181. The line is typically about 50% stronger in BLR calculations.
All atomic data for the SIII ion have been updated to Tayal, S.S., and Gupta, G.P. 1999 ApJ 526, 544. Some collision strengths differ by large factors from Galavis, M.E., Mendoza, C., & Zeippen, C.J. 1995, A&AS, 111, 347, which was used before.
In the current version of the code the line wavelength is an integer. This is to allow a precise match with a wavelength specified during a search. An integer format has obvious limitations in specifying fractions. To overcome this, the format of the wavelength changes with wavelength, in an attempt to have several significant figures. For wavelengths longward of 100mm the wavelength is given in mm. For wavelengths longward of one micron the wavelength is given in tenths of microns. For wavelengths between 10┼ and 100┼ the wavelength is given in tenths of ┼, and in hundredths of ┼ below this. In previous versions of the code the wavelength was in tenths of an ┼ below 100┼. A new command, the set oldwavelength command, will cause the code to behave as it did in the past.