I recently procured an Alpy 600 and am gaining knowledge in acquiring and processing astronomical spectra. I have learned much but have a specific question regarding instrument response.
On a recent night I acquired spectra of bet Gem (AAVSO Standard Star) and nearby type A0V reference star HD64648 in addition to a few other targets and reference stars. I processed HD64648 and generated from it and a Pickles A0V library spectrum an instrument response curve. I applied this instrument response to two other A0V stars, HD41695 and HD61887 and found a good, but not great match. Next I applied the instrument response to bet Gem and got a result that compared poorly with a Pickles K0III star in terms of the shape of the continuum.
Doing some research I have come to the understanding that instrument response depends not only on the characteristics of the telescope and imaging camera, but also on atmosphere and the spectroscope itself (i.e the slit). Note that my instrument response curve was based on a reference star at approximately the same airmass (~1.02 for bet Gem and ~1.07 for HD64648), so atmospheric dispersion effects should have been minimal when applied to bet Gem. So I'm wondering if the positions of the target and reference star on the slit makes a difference. My specific question then is this: When acquiring spectra of a target and a reference star, should I make sure both are at the same location on the slit? Put another way, does a difference in slit position between target and reference star affect the instrument response? If yes, why? If no, then what else can contribute to the problem I described?
Also, if it makes a difference, I have a focal reducer forward of the spectroscope and measured my F ratio at 5.7
Scott
Hi Scott,
Yes the position of the star on the slit (and focus) can affect the shape of the spectrum through either atmospheric dispersion (ie the light from the star is dispersed vertically relative to the horizon according to wavelength) or through chromatic aberrations in the telescope optics which means the mix of wavelengths varies depending on focus and the distance from the centre of the star image. Achromatic refractors and focal reducers are particularly problematic and are best avoided if possible. Christian Buil, has some analysis of the combined effects here
http://www.astrosurf.com/buil/dispersion/atmo.htm
As well as placing the centrally star on the slit, running with the slit vertical relative to the horizon (ie at the parallactic angle) helps with atmospheric dispersion and keeping the focus constant is critical if your telescope optics are achromatic. The trick is to try keep everything as far as possible the same between measuring the reference and target and if you think things are changing, take a reference before and after. A relative flux calibration accuracy of say +-5% above 4000A is reasonable target with care but this gets much harder as you move further into the UV.
The other thing to watch is the quality of your reference. It is preferable to use stars with known measured spectra if possible and if relying on published classifications for stars check the reliability of this eg using Brian Skiff's catalogue and and chose references with low interstellar extinction.
http://vizier.u-strasbg.fr/viz-bin/VizieR?-source=B/mk
You might also find this page on my website of interest which has more hints and tips on relative flux calibration and some examples
http://www.threehillsobservatory.co.uk/astro/spectroscopy_21.htm
Cheers
Robin
The other thing to watch out for with bright stars like Beta Gem is the effects of scintillation which can dramatically vary the shape of the spectrum between short exposures necessary to avoid saturation with the ALPY 600 on such a bright star. It is important to combine sufficient individual exposures to give a total exposure long enough to average out this effect
Cheers
Robin
Robin,
Thank you for your explanations, advice, and references. Although you confirmed (I think) that a difference in slit position between reference and target can affect flux calibration, I was left to wonder why. I know that dispersion is a function of slit width, so any difference in slit width (e.g. due to manufacturing tolerances) along the slit could affect the flux calibration. So If I place the reference and target at the same position on the slit I'm assured the dispersion for both due to the slit itself will be the same. Is that correct? And if I place them both centrally on the slit then the spectrum is centered on the flat image, which I presume is desirable (but don't know why).
Scott
Hi Scott,
By position on the slit I mean the position across the slit, not along it (ie such that the centre of the star image is offset relative to the centre of the slit so a different slice of the star image is sampled) The diagrams in the reference to Buil's website show what I mean.
The position along the slit should not in theory affect the shape of the spectrum provided a good flat correction has been made. It is however good practise to place the reference star at the same location along the slit in case for example the flat does not completely correct for vignetting in the spectrograph, the effect of which would be different at different points along the slit
Cheers
Robin
Thanks Robin. All makes perfect sense. I'm taking this all in as I increase my understanding and apply it to developing my techniques. The goal, of course, is to consistently produce quality spectra free from defect to the extent possible. For me it is important to understand the reasons/physics behind a particular recommended technique, whether for acquisition or processing and your explanations and references has provided me this insight. Next clear night I plan on implementing these lessons-learned and see how much improved my results are. Thanks again.
Scott
Found this reference to be useful;
https://ui.adsabs.harvard.edu/abs/1982PASP...94..715F/abstract
Yes ideally one should run with the slit running in the parallactic direction to minimise the effect of atmospheric dispersion (This is where alt az mounted telescopes have an advantage) but in practise most amateurs do not bother and just take extra care when working close to the horizon to chose reference stars close to the target in both altitude and azimuth, combined with placing the target and reference at the same XY coordinates so the effects cancel. For me it is focus shifts coupled with the achromatic focal reducer which gives the biggest problem (A C11 run with a focal reducer when using the ALPY. If starting from scratch again I would chose a fully reflective telescope for spectroscopy.) I mentor in spectroscopy for the BAA (perhaps the AAVSO have a similar system ?) and other common issues I see are in the choice of binning and sky background zones. (Make sure all the spectrum is included in the binning zone and none of the spectrum is included in the sky zones) and over smoothing the response curve, giving a poor fit, particularly at the blue end where the sensitivity is low and dropping fast. If you still are having problems I could take a quick look at your data. Do you have anywhere on line you could put a set of images?
Cheers
Robin
When the weather clears here I will take a set of test images and post them to my Google shared drive. When I have them I will send you the link.
Scott
Robin,
I sent you a link to my Google shared drive with images from a recent session.