Dereddening flux corrected spectra with ISIS

[David Boyd]

When I deredden a flux corrected spectrum with ISIS using the Extinction tab in the MISC tab, the y-axis of the output spectral profile is in relative intensity (ie scaled to 1 at the wavelengths specified on the Settings tab) rather than in erg/cm2/s/A as for the input spectrum. The output spectrum is the same as if I had reddened the original relative intensity spectrum before applying flux correction.

Is this a bug or is there a reason why it does this?

Thanks,
David

[David Boyd]

Christian, can you make any comment on this?

Thanks,
David

[Christian Buil]

ISIS does not calculate the absolute extinction of a star because the only parameter supplied in the function is the color excess, that is to say:

E(B-V) = (B-V)observed - (B-V)intrinsic

The correction applied to the input spectrum at this stage is only a relative correction of the profile shape. So the values in ergs are lost in the actual computation in the output product, and given in relative intensity scale. If you have the absolute magnitude extinction A(v) you are of course free to apply a multiplier to the result to the actual result for recover the absolute intensity.

Christian

[David Boyd]

Christian, thank you for your explanation. I understand that this is a relative correction of the profile shape and not a correction for absolute extinction.

I’ve been looking into extinction correction in papers by Cardelli and Fitzpatrick & Massa. If I understand it correctly, extinction can be accounted for by multiplying the calibrated flux by 10^[0.4*A(V)] and reddening by multiplying the flux by 10^[0.4*E(λ-V)]. If we adopt an average value of R(V) = 3.1 for interstellar extinction in the galaxy, then A(V) can be calculated as R(V)*E(B-V). E(λ-V) = A(λ)-A(V) and a calculation for A(λ)/A(V) can be found, for example, in Cardelli’s paper.

It looks as though the dereddening correction you apply in ISIS is equivalent to the second of the above corrections. If this is correct, I am still not clear why is it not possible to scale the original flux calibration by this relative correction and retain the result in flux units, but perhaps I am misunderstanding something?

I see ISIS applies the scaling relative to the value 1.0 at 6650A. Is it because the relative intensity in ISIS is usually normalised to 1.0 around that wavelength?

David

[Francois Teyssier]

Hello David,

Three possibilities :

1. VOSpec compute the correction for extinction
Use Cardelli-O'Connell law with Rv = 3.1 (Rv = 5 in Orion region)
Unfortunatly, VOSpec save only in .dat files
http://www.sciops.esa.int/index.php?pro ... age=vospec

2. In ISIS
- build a "unit" spectrum (all intensities = 1)
- with the extinction tool, compute the correction for your unit spectrum
- normalize the corrected unit spectrum at lambda = 5550 angströms
- multiply the last one by Av (Magnitude extinction for V band : Av = Rv x E(B-V) and save it : this is your fonction for correction of extinction
- finally, you multiply your flux calibrated spectrum by the function for correction

3. Mix
In VOSpec, you open your unit spectrum, compute the extinction and save it
Then you open your function in ISIS and multiply your calibrated spectrum by he function

1 = 3
2 gives a correct result at 10%

François

[David Boyd]

Thank you Francois. The method I am now using is similar to your option 2.

I have created a spreadsheet which, given the colour excess E(B-V), computes the extinction correction to the flux as 10^[0.4*A(V)] and also a spectral profile for interstellar reddening as 10^[0.4*E(λ-V)] with E(λ-V) calculated using Cardelli's formula with the value 1.0 at 5500A. I then use ISIS Arithmetic to multiply the flux calibrated spectrum by these two factors. For stars with significant interstellar reddening, this gives spectra which match the catalogued types well.

David