Experiments in stellar spectroscopy

For those interested in imaging and image processing with DSLR, CCD, webcam or film.
Message
Author
User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Experiments in stellar spectroscopy

#1 Post by rwilkinson » Wed Jun 22, 2011 7:21 am

I've now mounted a 500 lines/mm grating filter on the front of my 50mm lens:
staranalyser500.jpg
A cheap diffraction grating fitted on the front of my 50mm lens
staranalyser500.jpg (218.05 KiB) Viewed 7490 times
This will split the spectrum of point sources of light (i.e. stars).
Not to be put off by clouds (or daylight), I've made myself a synthetic star for testing:
cal_lamp2.jpg
Testing with a home-made "synthetic star"
cal_lamp2.jpg (2.74 KiB) Viewed 7855 times
which produces this spectrum:
cal_lamp2_lines.jpg
Spectral lines from my synthetic star
cal_lamp2_lines.jpg (1.56 KiB) Viewed 7855 times
And from the spectral analysis:
we can deduce that it's actually a compact fluorescent lamp, viewed through a pin-hole! :-)
cal_lamp_id.gif
Spectral analysis of my synthetic star
cal_lamp_id.gif (7.95 KiB) Viewed 7490 times
Identification of the characteristic Mercury emission lines (at 4046 and 4358 Angstroms) in its spectrum allows me to use it as a calibration source for my spectrographs.

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#2 Post by rwilkinson » Fri Jun 24, 2011 12:15 pm

Rapid progress this week - I've now used a "proper" diffraction grating on a real star!

The Star Analyser 100 spectroscopy filter is designed to be fitted in front of the CCD chip, but it may also be used across the front of a lens:
1004xspectrograph.jpg
Star Analyser 100 fitted across the front of a 50mm lens on my 1004x-JG camera.
And yes, it is possible to get spectra from bright stars without using a driven mount!
1004xspectrograph.jpg (160.94 KiB) Viewed 7489 times
With a little IRIS processing, I extracted the spectrum:
vega10x5stack.jpg
Spectrum of Vega from 10x 5-sec exposures at f/22 in a twilight sky.
vega10x5stack.jpg (1.51 KiB) Viewed 7844 times
and then fed it into the VisualSpec program to do the analysis:
http://www.boltonastro.org.uk/Gallery/d ... fullsize=1

The scale across the bottom of this graph is the wavelength in Angstroms (1A = 10nm), and the resolution is principally limited by my camera's 640x480 pixels.
With this set-up I get around 15 Angstroms per pixel, but I could improve this by a factor of four using my DSLR with a 200mm lens.

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#3 Post by rwilkinson » Mon Jun 27, 2011 7:37 am

Well, I had another go at Vega last night, but this time with the Star Analyser grating on the front of a 200mm f/8 lens on my DSLR.
Although the extracted spectrum is now around than 4 Angstroms per pixel, the actual resolution of the spectral lines is around the same:
vega2.gif
Spectrum of Vega with DSLR
vega2.gif (6.95 KiB) Viewed 7827 times
and you can see that the response of this colour DSLR sensor is nowhere near as good as the monochrome one (Sony Ex-View) in my modified CCTV camera, particularly at the red end of the spectrum!

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#4 Post by rwilkinson » Wed Jun 29, 2011 7:26 am

Last night I took some images of Vega with the £2 grating on the front of my DSLR with 50mm lens.
Even though the sky was not very dark, I was still able to see the "Balmer" hydrogen lines in the spectrum very clearly:
vega500_gauss2.gif
Spectrum obtained with camera tracking
vega500_gauss2.gif (6.62 KiB) Viewed 7819 times
And then I turned off the drive, to see what I could manage with a fixed camera:
vega500_fix_guass2.gif
Spectrum obtained with camera on fixed tripod
vega500_fix_guass2.gif (5.59 KiB) Viewed 7819 times
and yes, they are still visible! 8-)

For this test, I took care to align the grating so that the spectrum was at right-angles to the star-trails.
I'm sure that the noise could be reduced considerably with a dark sky, taking more longer exposures, and using dark-frame calibration.

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#5 Post by rwilkinson » Fri Jul 01, 2011 1:35 pm

I'm back to working with the StarAnalyser100 filter now - this time with my 8" SCT (the filter is fitted just in front of my 1004x-JG camera). Since the field-of-view is very small, I've also got a flip-mirror in-line to help finding the target. But its CCD is sensitive enough that I can capture the starlight direct from the video stream (rather than needing its long-exposure facility).

My first target was Albireo, the famous yellow & blue double star in Cygnus. I used a single exposure setting for my images: so the blue component is fainter (and so its spectrum is more subject to noise).
albireo.png
Image of Albireo and the pair of derived spectra, showing some hydrogen absorbtion lines.
albireo.png (23.73 KiB) Viewed 7489 times
albireo.GIF
Analysis of the spectra using Visual Spec. The vertical axis is intensity and horizontal is wavelength in Angstroms (with blue on the left and red on the right).
albireo.GIF (19.07 KiB) Viewed 7806 times

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#6 Post by rwilkinson » Fri Jul 01, 2011 4:20 pm

I've just found another interesting function in the Visual Spec program: the ability to synthesize a colour spectrum based on a calibrated monochrome intensity profile.
Here are the results for the two components of Albireo:
albireo1.jpg
Synthesized spectrum of the yellow component of Albireo.
albireo1.jpg (5.47 KiB) Viewed 7805 times
albireo2.jpg
Synthesized spectrum of the bluecomponent of Albireo.
albireo2.jpg (5.21 KiB) Viewed 7805 times
so now I can see the actual colours with my black & white camera!

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#7 Post by rwilkinson » Sat Jul 02, 2011 12:17 pm

Having seen some absorbtion lines in a distant star, I thought I'd try to see some emission lines next, and with the famous planetary nebula the Ring (M57) high in the sky at the moment, it was too good an opportunity to miss:
m57_sa.jpg
M57 viewed through Star Analyser 100.
25x 15-sec with 1004x-JG camera on 8" SCT at f/10.
m57_sa.jpg (7.86 KiB) Viewed 7797 times
Notice how bright the first-order spectra are, compared with the "zeroth-order" (straight-through) image. But the difficulty is that this is an extended object rather than a point-source (like a star) and the spectra from different parts of the object overlap (I'd need to put a narrow slit in front of the grating to get a sharp spectrum). So I tried to select a narrow band around the brightest part of the nebula:
m57_sa2.jpg
Selection of part of the image for spectrum analysis.
m57_sa2.jpg (8.87 KiB) Viewed 7797 times
This produced the following spectrum profile:
m57_spec4.gif
Diffuse spectrum of the Ring Nebula
m57_spec4.gif (4.92 KiB) Viewed 7797 times
which is too diffuse to calibrate properly, but we can see the main peaks around 4861-5007 (Hydrogen-beta and Oxygen III) and 6548-6584 (Hydrogen-alpha and Nitrogen II).

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#8 Post by rwilkinson » Sun Jan 15, 2012 1:45 pm

I've been doing some more experiments with our StarAnalyser100 filter, this time mounted between my 1004x-JG camera and a 200mm lens.
This spreads the light spectrum from each star across the CCD at around 2nm per pixel - here's an example using part of the Pleiades cluster:
m45spectra.jpg
Part of the Pleiades cluster imaged with a StarAnalyser filter and 200mm lens
m45spectra.jpg (8.43 KiB) Viewed 7445 times
The Visual Spec program can convert this spectrum into a frequency plot, which can then be calibrated (taking the Telluric absorbtion band around 759nm caused by oxygen in the Earth's atmosphere). Here's an example with Sirius:
siriusgraph.png
VSpec graph of Sirius spectrum, calibrated for wavelength in Angstroms
siriusgraph.png (17.34 KiB) Viewed 7445 times
Now this observation is filtered by the frequency response of the CCD detector, but if we know what this is (from the Sony datasheet) then VSpec can correct for it by division:
siriuscal.png
Sirius spectrum before (cyan) and after (red) calibration for the CCD response.
siriuscal.png (5.59 KiB) Viewed 7445 times
And finally, the original colour spectrum can be reconstructed by VSpec:
sirius-col.jpg
Colour spectrum of Sirius, synthesized by VSpec from data from a monochrome camera and diffraction filter.
sirius-col.jpg (14.5 KiB) Viewed 7445 times

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy

#9 Post by rwilkinson » Sun Jan 15, 2012 2:01 pm

It's interesting to compare the spectra of three of the bright stars in the Southern sky - they are of different spectral types and so appear different colours.

Rigel, type B8:
rigel-col.jpg
Rigel spectrum from VSpec
rigel-col.jpg (14.44 KiB) Viewed 7445 times
Sirius, type A1:
sirius-col.jpg
Sirius spectrum from VSpec
sirius-col.jpg (15.22 KiB) Viewed 7445 times
and Betelgeuse, type M1:
betelgeuse-col.jpg
Betelgeuse spectrum from VSpec
betelgeuse-col.jpg (13.71 KiB) Viewed 7445 times
In each case the dark absorbtion band is caused by Oxygen in Earth's atmosphere (around 759nm).

User avatar
rwilkinson
Member
Posts: 1284
Joined: Mon Oct 12, 2009 6:47 pm
Location: Bolton
Contact:

Re: Experiments in stellar spectroscopy - M42 emission lines

#10 Post by rwilkinson » Mon Jan 16, 2012 3:24 pm

Whilst I had my spectrograph system pointed i the Orion region, I thought that it would be interesting to look for the emission spectrum from M42. As this is a diffuse object, I can't produce a proper spectrum without a slit, but we can still see the dominant emission lines with the diffraction filter:
m42spec.jpg
M42 viewed through the diffraction filter
m42spec.jpg (7.21 KiB) Viewed 7438 times
Isolating a narrow portion of the nebula (shown in the yellow box), I managed to extract the following profile:
m42graph.png
Profile of M42 spectrum from VSpec
m42graph.png (22.14 KiB) Viewed 7438 times
We can certainly see the main peaks around 4861-5007 (Hydrogen-beta and Oxygen III) and 6548-6584 (Hydrogen-alpha and Nitrogen II).
The re-synthesized colour spectrum looks like this:
m42-col.jpg
Synthesized colour spectrum of M42
m42-col.jpg (15.8 KiB) Viewed 7438 times

Post Reply