First attempt at solar spectroscopy

[thesmiths]

I recently bought a Paton Hawksley spectroscope. It's actually a very nice simple instrument. I had intended to use it for simple chemistry demonstrations but then found if I pointed it at a bright cloud, I could also clearly see the Fraunhoffer lines in the sunlight. I decided to try to take some images with the spectroscope. This turned out to be a bit challenging since it's really designed for visual use. I happened to have a small CCTV lens that fits on my Imaging Source cameras. This fitted very nicely to the eyepiece of the spectroscope and would just about come to focus. Here's how the setup looked:



The first test is of course always the bright orange/yellow line of sodium. For this I put a wet wooden chopstick into some salt and held it over the gas stove. The resolution of this spectroscope is not high enough to split the sodium doublet (even visually).



The second image that is always fun (and easy) is a compact flourescent light bulb:



For these colour images I used my DBK21. In the next post, I'll put up the solar spectra.

[thesmiths]

Next, I pointed the spectroscope at a bright part of the sky (not directly at the sun). This is the spectra taken with the DBK21. It more or less shows what is seen visually. Actually, when viewed visually, the lines are much more distinct and sharp. The eye focuses much better with this instrument.



The DBK21 goes far into the infrared (which the eye does not) so there are lots of infrared lines in the image. The blue filter of the DBK21 seems to become very transparent in the infrared. There are also some shadows which have something to do with the internal construction of the spectroscope (these are out of focus visually). At this stage, I decided to switch to monochrome and used my DMK41. This gave the best results:



Lots more lines in the infrared and near UV are visible. In the next post, I will compare these lines to the known solar spectrum.

[swisswalter]

Hi Douglas

great results, thanks for sharing

[thesmiths]

Here is what the solar spectrum is supposed to look like:



And here are the assignment of the lines to particular elements (and to the oxygen bands in the atmosphere). The A band is the very dark band on the far right of the spectra image. The K and H lines are the last two very dark lines visible on the left side.



It turns out the hydrogen alpha absorption line is not very strong but the calcium K and H lines are very strong. Actually, I could not see the calcium lines visually (although my son could). This is a well known issue with calcium imaging.

I tried to calibrate the spectroscope but this turned out to be a bit hopeless since the optics have too much distortion. There is no linear relationship between diffraction distance and wavelength. Nevertheless, quite a fun experiment and also good test for the Imaging Source cameras.

Attached files

[robert]

Interesting report, thanks
Robert

[marktownley]

Way cool! Thanks for that!

[Merlin66]

A good start!
At least it would allow a "quick and dirt" check of filter transmission curves.
You can use the detailed solar spectrum from Liege etc to both idenify your lines and then use this to do a wavelength calibration.
The highly detailed solar spectrum by Lunette Jean Rösch is highly recommended.
http://ljr.bagn.obs-mip.fr/observing/spectrum/

Valerie Desnoux's VSpec (freeware) is a great tool - well supported.
http://www.astrosurf.com/vdesnoux/
onwards and upwards

[rsfoto]

Next, I pointed the spectroscope at a bright part of the sky (not directly at the sun). This is the spectra taken with the DBK21. It more or less shows what is seen visually. Actually, when viewed visually, the lines are much more distinct and sharp. The eye focuses much better with this instrument.



The DBK21 goes far into the infrared (which the eye does not) so there are lots of infrared lines in the image. The blue filter of the DBK21 seems to become very transparent in the infrared. There are also some shadows which have something to do with the internal construction of the spectroscope (these are out of focus visually). At this stage, I decided to switch to monochrome and used my DMK41. This gave the best results:



Lots more lines in the infrared and near UV are visible. In the next post, I will compare these lines to the known solar spectrum.

Hi,

Tried yesterday to calibrate the spectrum but looks like it is not very linear as I did not get far :S

Looks like there is an overlaping of the 2nd order (blue over red) on the right side of the spectrum over the infrared ¿?

BTW, the picture of my signature is a Solar spectrum :woohoo:

[papamidnight]

Nice Work!

[thesmiths]

Reply to RSFOTO: I did mention in my later post "I tried to calibrate the spectroscope but this turned out to be a bit hopeless since the optics have too much distortion. There is no linear relationship between diffraction distance and wavelength." I used RSpec to attempt a calibration. I assume the optics in the spectroscope itself plus the optics in the CCD camera lens has a lot of spherical aberation. Here is my attempt to manually align the spectrum to get a qualitative idea of which line is which:



It very hard to get them to line up all the way across the field. But I was actually pretty amazed to see so much at all. For those interested in "solar outreach", I would say having a simple hand-held spectroscope around to point towards the sky is a very striking visual demonstration of what hydrogen alpha and calcium K corresponds to.

With regards to what you called 2nd order overlapping, I mentioned that the "blue filter of the DBK21 seems to become very transparent in the infrared." The DBK (in contrast to the DFK) cameras have no IR cut filter. Many cheaper blue filters become quite transparent in the IR and so really these kinds of cameras should always be used with an IR cut filter to keep the correct colour balance.

[Merlin66]

Not having a IR cut off at around 7000A is a bonus for us in spectroscopy!
(The 2nd order can be easily suppressed using a red filter)
VSpec will certainly allow you to calibrate non-linear spectra.
Note: the UV-IR cut filters in "std" DSLR block below 4000A and above 7000A.
I've removed both the colour balance and the anti-alias filter from my 1000D for spectroscopy.

[rsfoto]

Not having a IR cut off at around 7000A is a bonus for us in spectroscopy!
(The 2nd order can be easily suppressed using a red filter)
VSpec will certainly allow you to calibrate non-linear spectra.
Note: the UV-IR cut filters in "std" DSLR block below 4000A and above 7000A.
I've removed both the colour balance and the anti-alias filter from my 1000D for spectroscopy.

Hi Ken,

I do not want to start a war here about spectroscopy software. :woohoo:

The problem itself is that the lines are a bit washy and having possibly the 2nd order overlaid on the red part makes it a bit more difficult.

BTW Rspec can also calibrate non linear spectra. Just an aside note :silly:

[colinbm]

Very nice work Douglas.The slit in this is nice & clean too
Col

[rsfoto]

Sorry, did not see that.

On the other side at one of your posts compraing your B&W spectrum with the coloured spectrum there seem to be many matches.

Again, sorry for missing your writing about calibration attempt

[rsfoto]

Hi,

Being stubborn as I am, I retried a calibration of your spectrum :whistle:

Also the grating looks quite linear :woohoo:

Calibrated spectrum



Calibration window