spectro-aras.com

Been having a difficult time processing LISA Infrared spectra in Isis 5.93. I've tried to process using the LISA (argon IR) and/or the filemode config (under the General tab) using the following dispersion and Ar lines:

2
3.6299
6965.431
7067.218
7272.936
7383.98
7723.761
7948.176
8266.077

Every-time I try to use the calibration assistant to get my calibrate just prior to get a response curve, the "X coordinate of line at wavelength 6965.431 changes to the normal Ne wavelength of 5944.82 and I get a large RMS value in the calibration assistant box. I've also tried to manually place these line values (with their corresponding x-positions) in the dispersion button under the Profile tab, but I get the same error; I wonder if the "Spectral domain for profile scaling" under settings needs to be changed to another value (using 6400 and 6750 for L1 and L2).

Is there some-kind of LISA IR processing guide that Christian put out? All I can find is the normal visual guide or his IR spectra response.

One more question, is there a spectra database we can access with ranges below 7400A? Most of the miles and Mybase end at this wavelength, but I need stellar spectra to go to at least 9500A.

James

Hi James,
Did you entered the correct dispersion parameter for your spectrograph : 3.6299 ?
The 8266.077 Neon line is faint, I tested successfull the 8264.52 Argon line.
I join my NIR LST calibration files (for a lhires III with 600 l/mm gratting).
Michel

Hello James,

Don't know if it helps, here is my calibration file for LHires 600/mm visual (R~2700):

CALIB\g600_IR_CaIImid.lst:

2
0.72806580645161290322580645161292
8103.693
8115.311
8264.522
8408.21
8424.648
8521.442
8667.944

Many of these lines should be resolved by LISA (some little contamination at lower res more and more likely to happen, so you either include many of these, or include very few but very sure ones).

ISIS log:
Wavelength fitting error
Line #1 x = 12.981 lambda = 8103.693 dlambda = -0.017
Line #2 x = 28.803 lambda = 8115.311 dlambda = 0.021
Line #3 x = 232.993 lambda = 8264.522 dlambda = -0.006
Line #4 x = 430.084 lambda = 8408.210 dlambda = -0.001
Line #5 x = 452.658 lambda = 8424.648 dlambda = 0.003
Line #6 x = 585.749 lambda = 8521.442 dlambda = 0.002
Line #7 x = 787.622 lambda = 8667.944 dlambda = -0.001
--------------------------------------------------------------------------
Wavelength file : C:\csill\428exm\20181031_LH600IR\34_v694Mon_m12C\calib\g600_IR_CaIImid.lst
Coefficient a4 : 0.000000E00
Coefficient a3 : 0.000000E00
Coefficient a2 : -4.629648E-06
Coefficient a1 : 0.73210
Coefficient a0 : 8093.476
--------------------------------------------------------------------------
RMS : 0.013956
--------------------------------------------------------------------------

Calibration image: If you chose manual calibration via file, first I recommend you to calibrate the lamp image with itself (copying and renaming appropriately). Don't forget checking 2.General -> Sky not removed. Then load the first few attempts (processing with order=1 and few lines for the very first attempts) in VSpec and see how the lines fit visually: play with Tools -> Elements / Lineident dropdown list. You can get extra lines upon your wish, and finetune the calibration (I know, this is not easy but can be done once).

- Peter

PS.: I'd also appreciate to know, which library to use in order to get fluxes - as for real use (= extracting profile fluxes), I highly doubt these spectra are for serious use (continuum may change rather unreliably, and I filter is very rare).

I am speculating to use this one:
http://www.iac.es/proyecto/miles/pages/ ... ibrary.php

For a wider wavelength range, I could imagine scaling with a matching Pickles star to extend such a region.

- Peter

Thanks for all the feedback. Yes my computed dispersion came out to be 3.6299 A/px using Bass project 1.9.8; no trouble getting a calibrated spectra there!
Just to clarify, here are screen shots of the Isis processing pages I have just before running the calibration assistant and having Isis reset to visual neon lines with
the "X coordinate line at wavelength =" turning back to 5944.82:

General Tab setting:

Setting Tab:

Finally, I tried to use the manual setting in dispersion based on the Ar lines seen and their X-positions:

Still no go....the "X coordinate line at wavelength =" reverts back to 5944.82A and the RMS error in the calibration assistant box becomes 13 or more.

James

Hi James,

I have never used the assistant, though I only have the ALPY, not the LISA. Doesn't it combine the lamp and Balmer lines in the star spectrum to improve the calibration, which would not be needed for IR spectra ?

What happens if you just use the file mode without using the calibration assistant? ie like method 2 described here for the ALPY but using the file mode instead of the predefined line option and entering the X coordinate and wavelength of one of the lines
http://www.astrosurf.com/buil/isis/guid ... ration.htm

Robin

RE:"What happens if you just use the file mode without using the calibration assistant?"

Excellent suggestion! I thought about it and decide to make it even simpler and use the LHIRESIII mode in the Settings tab, and the "File Mode" under
the General tab, with my the Dispersion in manual mode with the same Ar lines indicated in my LISA-IR.lst file, ran and got a good calibrated spectra (see below):
HD007351 (S3+/2) star (no Image Response correction):


I still need to find a IR .fits spectra source MILES or other stars from 7400-9500A; the one mentioned by Peter is only good for spectra btw 8350 and 9000A and they might be "Normalized" spectra around the value of (y-axis) 1: NOTE: ps below.......

James

p.s. I found this site which has great IR coverage (3000-25000A!) that helps with generating IR and UV low resolution Image response files for spectral calibration:
https://www.eso.org/sci/observing/tools ... brary.html

Thanks to the help here I was able to process my IR LISA data for TU And and combine it with my Alpy 600 data for TU And; see:

[NOTE: All identified lines other than the Balmer lines are speculative]
I used a M5III star's IR spectra (ukm3iii.fits) from here: https://www.eso.org/sci/observing/tools ... brary.html to get my IR curve (on HD007351)
for TU And from 6100-9600A. More spectra to follow!

James

James Foster wrote: I used a M5III star's IR spectra (ukm3iii.fits) from here: https://www.eso.org/sci/observing/tools ... brary.html to get my IR curve (on HD007351)
for TU And from 6100-9600A.

I think these are just the Pickles spectra that are already in VSpec and ISIS (to 10000A at least)
How are people working in this region handling the Telluric bands ?

Robin

Robin Leadbeater wrote:How are people working in this region handling the Telluric bands ?

Robin

It depends what kind of data you want and what resolution / coverage.

At medium resolution, I am trying to cut out the tellurics manually (ISIS -> Continuum -> double click 2 times over a region, doing Edit the 1st few or last few pixels if necessary, chosing representative value...).
Then do "Automatic 1".
This method is not reliable, but should be good enough for EW assessment at medium resolution.
For a real use, I think wide slit + absolute flux was necessary - at my narrow range. (For LISA wide coverage, maybe EW ratios enough - professional astronomer needed to judge).

Continuum shape distortions matter more and more when the red component is cool type (M), and you want to decompose it, allowing a more accurate EW estimate (this is the final product what we want) and perhaps bury out other details (Paschen, CaII triplet).

Something for you don't need much care, would be some weak and narrow Fe I and Ti I absorptions, but that requires medium resolution and very high SNR, such that unusual for amateur instruments. ( http://adsabs.harvard.edu/full/1994A%26AS..106..413H )

Another usage with LISA IR could be, supernovae (CaII triplet should be in huge emission).

A good method can be to reduce the tellurics by software (VSpec??), but I never did that - except nearby H-alpha tellurics using ISIS (only for fixing the response curve). Hints welcome...

- Peter