SF propagation (Python)

[phockett]

Add scale-factor propagation to Python code for correct X-sections.

[phockett]

SF prop now incorporated.

Added SF as non-dimensional coord to matrix element Xarray.
Added to scaling calcs. in mfblm().

Might be an issue in some cases if SF is dependent on multiple dims, if only Eke dependent then all is OK.

[phockett]

Issue closed.

[phockett]

Actually, may still be an issue generally - trying to iron out again with AF calculation tests. Issue might be due to complex-valued SFs. (See GDoc PAD notes 2018 for more.)

[phockett]

Now mostly fixed (for geometric calc version) in 1aad3f9

This gives correct AF betas (tested only for N2 LF case so far), and XS from ePS definition (matches GetCro results).

Still need to confirm absolute XS Mb values however.

[Planetenbesitzer]

Sorry, but what is this scale-factor exactly? I don't know when and how to use SF to correct my calculation.

[phockett]

Good question, and apologies this is not documented elsewhere already!

Short answer: you probably don't need to use the scale factor (SF).

---

Longer answer for reference:

• The SF is output by ePolyScat DumpIdy segments for matrix elements, which look something like:

----------------------------------------------------------------------
DumpIdy - dump out phase corrected and properly normalized dynamical coefs
----------------------------------------------------------------------

Use photoelectron kinetic energy =       0.810000 eV
Corresponding photon energy =      14.402000 eV
Scale Factor to square root of Mbarn =  0.20663304E+01  0.39041597E+01


Sym Cont =A2     Targ =A2     Total =A1   
Sym Cont =    2  Targ =    2  Total =    1
NGroup =    4  Energy =  0.81000000E+00
LMax =   12  nrdimTarg =    1  NDipoleMatrixTypes = -100
   m   l mu ip it Value
  -2   2  0  1  1 -0.11977756E+00 -0.14663879E-01
   2   2  0  1  1  0.11977756E+00  0.14663879E-01
...

(This is from the NO2 demo at

ePSproc/data/photoionization/no2_demo_ePS.out

Line 2128 in 1c0b8fd

DumpIdy - dump out phase corrected and properly normalized dynamical coefs

)

• In ePSproc these Scale Factor to square root of Mbarn values are read, and stored in the Xarray metadata for each set of matrix elements.
• In testing, I found that these are already incorporated in the matrix elements, so you don't need to worry about them in general for computing X-sections or AF/LF results (see https://epsproc.readthedocs.io/en/latest/methods/LF_AF_verification_tests_060720_tidy_100920.html which compares results from ePolyScat getCro segments for the LF, with results computed using the ePSproc routines from the matrix elements for the LF case - this is a good sanity test of the numerics), although I did leave in the option to multiply by the SF in the current dev code.
• For changing the normalisation of the matrix elements you might want to use the SF, but I haven't explored this.

---

Out of interest, what are you working on?

[phockett]

Also, there is a new verifyAFiso() function you can use for the type of comparison mentioned above, see the docs and source for details. (Note this is currently in the dev branch only, and not yet written up in the main documentation.)

[Planetenbesitzer]

Thank you so much for your answer!! My task is to calculate beta4 coefficient for sequential ionization of CH3I, using dipole matrix elements. Wish you happy life Shutao Zhang 祝您工作生活顺利. 张舒涛 发件人: Paul Hockett 发送时间: 2023年5月12日 8:34 收件人: phockett/ePSproc 抄送: 张舒涛; Comment 主题: Re: [phockett/ePSproc] SF propagation (Python) (#6) Good question, and apologies this is not documented elsewhere already! Short answer: you probably don't need to use the scale factor (SF). Longer answer for reference: • The SF is output by ePolyScat DumpIdy segments for matrix elements, which look something like: ---------------------------------------------------------------------- DumpIdy - dump out phase corrected and properly normalized dynamical coefs ---------------------------------------------------------------------- Use photoelectron kinetic energy = 0.810000 eV Corresponding photon energy = 14.402000 eV Scale Factor to square root of Mbarn = 0.20663304E+01 0.39041597E+01 Sym Cont =A2 Targ =A2 Total =A1 Sym Cont = 2 Targ = 2 Total = 1 NGroup = 4 Energy = 0.81000000E+00 LMax = 12 nrdimTarg = 1 NDipoleMatrixTypes = -100 m l mu ip it Value -2 2 0 1 1 -0.11977756E+00 -0.14663879E-01 2 2 0 1 1 0.11977756E+00 0.14663879E-01 ... (This is from the NO2 demo at https://github.com/phockett/ePSproc/blob/1c0b8fd409648f07c85f4f20628b5ea7627e0c4e/data/photoionization/no2_demo_ePS.out#L2128) • In ePSproc these Scale Factor to square root of Mbarn values are read, and stored in the Xarray metadata for each set of matrix elements. • In testing, I found that these are already incorporated in the matrix elements, so you don't need to worry about them in general for computing X-sections or AF/LF results (see https://epsproc.readthedocs.io/en/latest/methods/LF_AF_verification_tests_060720_tidy_100920.html which compares results from ePolyScat getCro segments for the LF, with results computed using the ePSproc routines from the matrix elements for the LF case - this is a good sanity test of the numerics), although I did leave in the option to multiply by the SF in the current dev code. • For changing the normalisation of the matrix elements you might want to use the SF, but I haven't explored this. Out of interest, what are you working on? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>