Opened 22 months ago

Last modified 3 weeks ago

#358 new defect - wrong answer

PDR sims affected by induced processes

Reported by: Gary J. Ferland Owned by: Gary J. Ferland
Priority: blocker Milestone: C19_branch
Component: radiative transfer Version: trunk
Keywords: Cc:


tests reveal unexplained behavior in PDR sims with no ionizing radiation. Modify pdf_leiden_f4 to stop in zone 2. In its default state we get this hydrogen ionization for these two zones:

1.00E+00	5.18E-08
1.00E+00	3.51E-05 

the ionization goes up by ~3 dex.

Add the command "no induced processes" and we get

1.00E+00	5.18E-08
1.00E+00	5.19E-08

the answer does not change and is equal to the first zone, with induced processes. It is counterintuitive that induced processes would make the ionization higher in the second zone, which should be more shielded.

Attachments (2)

optimize-species-read.patch (3.7 KB) - added by mchatzikos 22 months ago.
patch for conditional loading of atomic data (1.5 KB) - added by mchatzikos 3 weeks ago.
Minimal script, in lieu of

Download all attachments as: .zip

Change History (6)

Changed 22 months ago by mchatzikos

Attachment: optimize-species-read.patch added

patch for conditional loading of atomic data

comment:1 Changed 22 months ago by mchatzikos

The patch that triggered the discovery of the unexplained behavior is attached (optimize-species-read.patch). It does not load atomic data for species that have been disabled.

comment:2 Changed 22 months ago by Gary J. Ferland

tried removing NO INDUCED and replacing with CASE B NO PHOTOIONIZATION. The latter disables photoionization of 2s. The large change in ionization still occurred, showing that photoionization of 2s is not responsible.

So we know: turning induced processes off and on disables / enables the anomalous ionization

it is not due to 2s metastables.

comment:3 Changed 3 weeks ago by mchatzikos

A minimal input deck to reproduce the results is attached. Relative to the original deck and the pdr_leiden.ini initialization file included by, the present script shuts off cosmic rays, allows for H-like Lyman pumping (removed the database H-like Lyman pumping off command, as it doesn't make a difference), and allows only for H, He, and N in the sim, among other things.

It is found that the rise of H+ in the second zone depends critically on the presence of N in the sim, as well as on charge transfer, and molecular physics. When nitrogen is disabled with element nitrogen off, the H+ ionization fractions in both zones become 0. The same is true when the no molecules command is used. When the no charge transfer command is used, the H0 and H+ ionization fractions become roughly equal to each other, about 1e-37, in both zones.

With this it becomes evident that the rise of H+ is due to the N molecular network, and its coupling to atomic and ionized H. What exactly causes the rise is not yet understood.

Note that it is still unclear why the no induced processes command has any effect on the H ionization balance.

Changed 3 weeks ago by mchatzikos

Attachment: added

Minimal script, in lieu of

comment:4 Changed 3 weeks ago by mchatzikos

Milestone: C17_branchC19_branch
Note: See TracTickets for help on using tickets.