Fifth attempt at solar spectroscopy -- high resolution monchrome images

[thesmiths]

We keep making incremental improvements to our setup. This time, we took the apparatus from our third attempt and replaced the 200mm Canon camera lens with a Baader 250mm f4.1 finder scope. The Canon 40D was replaced with a Imaging Source DMK41 monochrome camera. The 10 micron slit was replaced with a 5 micron one (again from Thorlabs). The equipment setup now looks like this:



In the foreground is the makeup mirror used to reflect and focus sunlight from the window onto the slit. The 1x5 inch collimator projects the light onto the 50mmx50mm 1800 l/mm grating. The Baader Vario-Finder is mounted on a Stronghold Tangent Assembly so as to adjust for tilt. There is a helical focuser at the back and then the DMK41 camera. The front optics gets draped in a dark towel to block out stray light.

Since the finder scope has 25% more focal length than the camera lens, there is higher resolution just from the optics. The narrower slit did not make much difference, I think. The DMK41 has 4.65 micron pixels vs the Canon 40D with 5.7 microns so that's another 23%. Arguably since it's monochrome vs Bayer matrix, there's an even further increase in resolution.

Here is a selection of images taken with IC Capture (set to automatically take BMP files at 2 sec intervals). There has been a little Photoshop done (levels and unsharp mask) but that's basically it. First, the area around 517nm with the "Magnesium triplet":



Then the area around 598nm with the "Sodium doublet":



The area around 656nm with "Hydrogen alpha":



The area around 687nm with the "Oxygen B band":



Finally, the area around 434nm with "Hydrogen gamma" and lots of other lines:



As expected, the resolution with the DMK41 is better. But not only that, with the Canon 40D -- even though it was modded with removal of the standard filter -- it was not easy to resolve the hydrogen alpha line very clearly and the oxygen B band in the IR was not visible. So all in all, a good improvement on our last attempt.

[thakursam]

Great results, thanks for sharing details of setup.

[swisswalter]

Hi Douglas

great resolution with that setup :

[Derek Klepp]

A great improvement thanks for the update.

[Montana]

Beautiful lines but no pretty colours

Alexandra

[Merlin66]

Very nice results there!
Well done.
The 1800 l/mm grating and the 250mm fl imaging lens coupled with the DMK41 pixel size is giving very good results. The optimum slit gap would be around 5 micron (4.65 x 2 x 125/250)
At this resolution, you're picking up the thermal "broadening" of the various lines.
Due to this, the Ha line always appears wider...look at the comparison between the two sodium lines -one is much tighter than the other.
You should be able to pick up the CaH&K lines with the DMK....

(Slit size corrected!)

[thesmiths]

To Ken: I'm still a little uncertain what corresponds to "optimum" slit size. Does it mean that all lines are effectively resolved as "lines" rather than as the broadened structures that they really are? So if to increase the slit to 20 microns, all the lines would appear the same width? I think capturing thermal, Doppler and Zeeman broadening/splitting is very interesting. So it's nice to be able to resolve not all lines being equal.

In our next iteration, I am hoping to push the resolution to the max by looking at the 2nd order lines in the blue and see if we are able to detect Doppler shift due to the Sun's rotation. My initial calculation, using the Ken Harrison spreadsheet, indicates it should be possible.

[Merlin66]

Firstly, I screwed up my arithmetic!!
The slit, as you have it is ideal!

The optimum slit width won't give lines of the same width, but will allow maximum resolution of the lines and the recording of detail in and around them.
There's plenty of light for solar spectra, so the usual issues of getting enough signal for a good SNR can be skipped.
It comes down the the Nyquist sampling of the finest detail recorded. If we work on a minimum sampling of 2, then the smallest detail recorded will be almost 10micron across. For your optics, there's a magnification ratio between the actual slit gap and the "projected" slit gap onto the CCD = focal length of the imaging lens/ focal length of the collimating lens = 2. This means that a 5 micron slit will give an image (taking into account the anamorphic factor) and a total angle of say 38 degrees, of 26micron - a oversampling rate of (26/5=5)!!

At these settings you'll achieve a creditable resolution of 0.56A (R=10000)
However, in the 2nd order you can't get much beyond 4500A before the angle of the grating becomes excessive...you'd need a Littrow type arrangement. But in this range the possible resolution kicks up to 0.15A (R=29000)
(Sorry for the confusion - I'll re-edit the original post to correct.)