SED library

The library of model spectral energy distributions (SEDs) is available at this link.

If you have any questions, drop me a line @gmail: mstalevski.astro

Below are some basic information on SED file format and parameters of the model. For more details, please refer to Stalevski et al. (2012a), and Stalevski et al. (2016). For any further questions and clarifications, feel free to contact me.

If you make use of the models, any feedback would be very welcome. If your work requires models with parameter values not currently included, let us know and we will do our best to provide those. If you cannot get good fits to the observations for some sources, seeing those examples would help us understand which part of parameter space needs to be explored further or sampled better, or if an additional component is required.

Please visit https://skirtor.streamlit.app/ for a very handy way to quickly visualise the model SEDs and explore how different parameters affect their shape (thanks to Devang Liya!)

Please read through this page for some important information.

SED files naming convention and format

File name example: t5_p1_q0_oa50_R20_Mcl0.97_i30_sed.dat

t: tau9.7, average edge-on optical depth at 9.7 micron; the actual one along the line of sight may vary depending on the clumps distribution.

p: power-law exponent that sets radial gradient of dust density

q: index that sets dust density gradient with polar angle

oa: angle measured between the equatorial plan and edge of the torus. Half-opening angle of the dust-free cone is 90-oa.

R: ratio of outer to inner radius, R_out/R_in

Mcl: fraction of total dust mass inside clumps. 0.97 means 97% of total mass is inside the clumps and 3% in the interclump dust.

i: inclination, i.e. viewing angle, i.e. position of the instrument w.r.t. the AGN axis. i=0: face-on, type 1 view; i=90: edge-on, type 2 view.

Flux in the SED files is calculated for a source at a distance of 10 Mpc and luminosity L_AGN=10^11 L_sol, where L_sol = 3.839e26 W (= 3.839e33 erg/s) (without solar neutrino radiation).

It is trivial to rescale the flux so that model SEDs can be applied to any given source luminosity and distance, just keeping in mind the scaling of dust emission, size and mass (see Ivezic & Elitzur 1997; equation 14 in Fritz et al. 2006; Honig & Kishimoto 2010).

Included in the SED library is the file with total dust masses for all the models; again, keep in mind these values (as well as size of the torus) need to be rescaled when applied to the sources with different L_AGN (see references above).

Primary source of radiation, the accretion disk, is represented by a point-like central source with anisotropic emission pattern, as suggested by Netzer (1987).

The adopted R_in is 0.5 pc (note that in Stalevski et al. 2016 the different value is reported; that is a typo). However, the actual R_in value depends on the polar angle, following anisotropic emission pattern of the accretion disk.

Header of the SED files contains the required description. A couple of clarifications:

# column 3: direct stellar flux:

# column 4: scattered stellar flux

"stellar" is a remnant from the original version of the code which provided only stars as primary source of radiation. In AGN models presented here, it stands for accretion disk emission.

"direct" means what reached the instrument at the given viewing angle, after being absorbed or scattered away by the dust. If there is no dust along the line of sight for the given viewing angle it will be identical to the "transparent" flux.

# column 7: transparent flux:

"transparent" stands for the original primary source flux (accretion disk), unaffected by the dust.

"transparent" flux is the same for all the models, but different for each inclination, as anisotropic emission for primary source is assumed.

-------------------------------------------------------------------------------------------------------------------------------------------------------------------

# column 1: lambda (micron)

# column 2: total flux; lambda*F_lambda (W/m2)

# column 3: direct stellar flux; lambda*F_lambda (W/m2)

# column 4: scattered stellar flux; lambda*F_lambda (W/m2)

# column 5: total dust emission flux; lambda*F_lambda (W/m2)

# column 6: dust emission scattered flux; lambda*F_lambda (W/m2)

# column 7: transparent flux; lambda*F_lambda (W/m2)

1.00000000e-03 1.12460249e-14 1.09750614e-14 2.70963495e-16 0.00000000e+00 0.00000000e+00 5.25636674e-13

1.14815362e-03 3.79601070e-15 3.73912587e-15 5.68848218e-17 0.00000000e+00 0.00000000e+00 6.20419813e-13

1.31825674e-03 1.15030177e-15 1.10217035e-15 4.81314199e-17 0.00000000e+00 0.00000000e+00 7.32294308e

...

...

...

7.58577575e+02 2.00305690e-17 3.37207694e-19 0.00000000e+00 1.96933613e-17 0.00000000e+00 3.37373427e-19

8.70963590e+02 1.01973387e-17 2.22817149e-19 0.00000000e+00 9.97452155e-18 0.00000000e+00 2.22900413e-19

1.00000000e+03 5.04538628e-18 0.00000000e+00 0.00000000e+00 5.04538628e-18 0.00000000e+00 0.00000000e+00

-------------------------------------------------------------------------------------------------------------------------------------------------------------------

SED example showing different flux components.

Primary source: accretion disk

Model SEDs are computed assuming the following composition of power-laws for the primary source, i.e., the accretion disk:

However, the actual shape of the accretion disk SED has very little influence on the resulting dust emission: what matters is the total amount of emission coming out from the disk. We demonstrate this in the plot below, where total and dust flux SEDs are compared for two models obtained with different accretion disk SEDs: one with the above power-law composition, and the other one with the modified version (as assumed by Schartmann et al. 2005 and Feltre et al. 2012). Looking at total flux in face-on view (top left), there appears to be a significant difference. However, this is caused just by the different normalization of the input SED in the optical/IR range. If we compare the dust fluxes (top right), they are almost the same, with some marginal difference in the silicate feature. In the bottom raw, the same for edge-on case is presented: again, the difference is marginal, even in total flux.

IMPORTANT: If AGNs of your interest have different SED shape as a function of wavelength (i.e. different slopes in the UV/optical range), you can subtract the power-law composition we assume for the model, and add the one that fits better your needs. You just need make sure the new disk SED is normalized to the same value as the original. This is a straightforward procedure, as outlined below:

Convert lambda*F_lambda to F_lambda (because SED files are given in lambda*F_lambda):

(1) Take columns # 4,5,7 (primary scattered, total dust, transparent), and divide it by column # 1: lambda

    ($4/$1; $5/$1; $7/$1, where $n stands for column number)

Normalize and add the new accretion disk SED:

(2) Integrate the primary accretion disk flux from SKIRTOR model ($7/$1)

(3) Normalize the new accretion disk SED to have the same value that you get in step #2

(4) Add new normalized accretion disk flux to the sum of primary scattered and total dust flux:

    $4/$1 + $5/$1 + newADnorm

Keep in mind that we assume anisotropic emission pattern for the accretion disk in the model, so the accretion disk flux in models with different viewing angle (inclination, parameter "i" in the file names) will have different normalization.

* * * This procedure is to be applied to the type 1 viewing angles, those with dust-free line of sight, i.e., i<90-oa, where 'i' is inclination and 'oa' half opening angle (i and oa in *sed.dat file names). Type 2 viewing angles i>=90-oa can be used unchanged: the accretion disk flux is completely absorbed, so it makes no difference. * * *

Alternatively, you can follow the procedure described in Yang et al. (2020):

Repeat the steps 1-2-3 as above to convert lambda*F_lambda in columns # 3, 4, 5, 7 to F_lambda and normalize the new accretion disk SED.

(4) Calculate the flux ratio of the original and new normalized accretion disk SEDs at each wavelength: 

k_lambda = origAD_lambda / newADnorm_lambda

(5) The new components of the direct and scattered accretion disk flux (corresponding to the columns 3 & 4) can be obtained by multiplying the original ones with the factor k_lambda. The total dust flux remains unchanged.

(6) Add the new direct and scattered flux components to the total dust flux.

This method has the advantage that it works for both type 1 and type 2 AGNs, including the scattered component and thus, strictly keeps the energy balance.