Opened 10 months ago

Last modified 10 months ago

#379 new defect - wrong answer

ioniation discontinuity introduced by continuum lowering

Reported by: Gary J. Ferland Owned by: nobody
Priority: major Milestone: C19_branch
Component: etc Version: trunk
Keywords: Cc:

Description

with continuum lowering included there is an He ionization discontinuity at ~1e14 cm-3. Continuum lowering disabled for final plot

toggle the database continuum lowering line to turn glitch on/off

Attachments (3)

clower.in (377 bytes) - added by Gary J. Ferland 10 months ago.
input script, toggle database continuum lowering to turn glitch on/off
clower.pdf (45.7 KB) - added by Gary J. Ferland 10 months ago.
ionization with / without continuum lowering
HeIonFrac.pdf (43.4 KB) - added by Gary J. Ferland 10 months ago.
He+ and He2+ ionization fractions

Download all attachments as: .zip

Change History (5)

Changed 10 months ago by Gary J. Ferland

Attachment: clower.in added

input script, toggle database continuum lowering to turn glitch on/off

Changed 10 months ago by Gary J. Ferland

Attachment: clower.pdf added

ionization with / without continuum lowering

Changed 10 months ago by Gary J. Ferland

Attachment: HeIonFrac.pdf added

He+ and He2+ ionization fractions

comment:1 Changed 10 months ago by Gary J. Ferland

The culprit is the collisional ionization boost of the top level, in iso_collide.cpp:

92 sp->fb[sp->numLevels_max-1].ColIoniz? *= 100.;

A comparison of runs with the collisional ionization boost commented out or not, and with lowering on or off is attached. The red and purple curves is what Gary sent the other day. The other two are similar, save we do not including the collisional ionization enhancement. With lowering and without the collisional ionization boost, we recover the tail in the plot. The collisional ionization boost makes a big difference when lowering is not included.

Looks like the collisional boost does not help in this test.

comment:2 Changed 10 months ago by Gary J. Ferland

The physics that is happening is nicely shown in Fig 1 of this paper by Mike Seaton. Levels come into LTE with the continuum because of very fast collisional ionization / 3-body recombination with very high-n levels. With our default very limited number of levels, only going up to n~20, these collisions are slow and the levels do not come into LTE with the continuum. Hence the need for the boost (a common trick in stellar atmosphere codes).

This particular model has 50 collapsed levels so goes pretty high and the ionization rates are becoming very large. The boost is in terms of the highest n collisional ionization rate so becomes very large. Then it totally goes away then the continuum starts to lower, since the density should be high enough to have brought the levels into LTE. Hence the bump.

Note: See TracTickets for help on using tickets.