This collated spectrum, showing the light of the Sun between ~360-920 nm, marks that, from two years of work, I have amassed 122k solar disks, into the atlas. This spectrum, made continuous and colorized, smoothed out gradients etc, are a sort of a byproduct of the individual cubes of the atlas. Still, nice to look at.
O I 7772 ~ continuum, dark smudge at center-top-left
On 2026-04-05, I imaged the Sun across multiple wavelengths, and found that areas that are dark smudges in oxygen are:
The images presented below are from scans recorded across several hours, each wavelength is the result of several scans stacked together. During acquisition, no major flares were reported. More about oxygen smudges here, here and here.
Using the familiar setup, I recorded four frames and stacked them together. The smudge appears somewhat embering-glowing in SDO AIA 1600, but as a bright spot in AIA/SUVI 304.
The smudge predictor is an attempt to estimate where a smudge could be expected:
stretch(someMath(AIA 1600) – someMath(AIA 1700)),
flares and smudges might appear bright in the image.
As I explained here, I keep seeing >>smudges<< in oxygen O I 7772. Smudges, ie regions that are darker than expected based solely on the location of flares and plages, yet seem to be related to whatever is happening in extreme UV. Here’s this day’s smudge.
2026-03-21-1244 solar disk, oxygen divided by continuum (and some other math), showing a smudge
2026-03-21-1244 smudge marked
WIP/DRAFT
The first light of my spectroheliograph was in the summer of 2024. It was a stock Sol’Ex, with a modified setup, using full aperture bandpass filters as an energy rejection method, instead of the recommended full spectrum ND filter or Herschel prism, flooding the instrument with the right kind of wavelength, while also managing the thermal load and keeping the zero order and any internal spectral fog at bay. I quickly left the comfort of the Fraunhofer lines, and stumbled into prominences in „unknown” wavelengths[1], a frustrating experience that lead me towards literature, and also propelled me to create the atlas of spectroheliograms[2].
In the meanwhile, almost as a side project, dedicating telescope time to this in an on and off manner, I started to hunt down prominences in various spectral regions and lines, some known and expected, some not so much. Thus today I know that the ultraviolet „unknown” wavelengths showing prominences were most likely the magnesium triplet around 383nm, that Paschen hydrogens are dwarfed by not just Balmer hydrogens but also by the helium at 1083nm, which shows coronal holes, just like the yellow and the deep red helium do. I’ve seen the hydrogen ladder almost up till the Balmer break, and seen that neutral calcium is something worthy to at least check for prominences, that the infrared oxygen triplet can do very funny stuff, with obvious prominences, and that ionized metals show dark plages and often times prominences that are easier to see than some helium lines.
During my campaign to observe the neutral oxygen triplet in the infrared, through full disk spectroheliograms, I noticed a few things:
Smudges are areas that stand out as darker than expected, after taking into account spots, flares, and active regions. After contrasting the oxygen against a nearby continuum, either by subtraction or by division, smudges are revealed to be darker in oxygen, when compared to nearby active regions (which are themselves darkened) and thus a bit surprising. Smudges are stable across several minutes/hours, or even a day or two, while the flares I recorded are often times darker and short lived, on the order of minutes mostly. These smudges seem related to active regions, but don’t stand out in GONG Hα, where they can be a mere brighter spot like many others that don’t show up as smudges, nor AIA 1700 where the underlying area usually looks like any other active area that doesn’t show a smudge in oxygen. Better correlation can be seen with say SUVI 0304 or SUVI 0094 as small bright features in EUV, but it isn’t one-to-one either. Smudges stand out when they are somewhat towards the disk center, inside say 2/3 solar radii.
Below I compiled some scans with a nearby SUVI and AIA disk for comparison. I included smudges (marked with S), small and bright flares (marked with F), for context.
Instruments: 62/400 doublet, EO 775/25 OD4 narrow band filter, OD 0.6 neutral density filter, modified Sol’Ex, with 1800 ln/mm grating, modified/cooled ASI 678MM, JSol’Ex. Variable to good skies.
O I 7772 (~) continuum, showing excess light on spots and shadows on plages
I observed the Sun in two oxygen lines, the O I 7772 and O I 8446, with the modified Sol’Ex (1800 gr/mm grating, band pass filters etc, 42/400 refractor). Unlike 7772 (the triplet), the 8446 (close doublet) doesn’t light up at the limb, as chromosphere. Regardless, I observed and processed the images. For λ7772, I have a full aperture etalon (CWL 775nm, FWHM 25NM, OD4 blocking), but only relatively wide bandpass filter(s) for λ8446, so parasite light is somewhat of an issue. Still, there is an interesting dark feature present in both observations, at around 4 o’clock, close to the edge.
O I 7772 divided by a nearby continuum, stretched. Showing prominences, excess light on sunspots, darkening on active regions, and a dark shadow where the flare occurred at around 2026-02-03 07:49Z
Active region 4366 shows quite the flares. I observed the Sun at around local noon, at the infrared oxygen triplet at 777nm, focusing on the O I 7772 line. Below I present data from two scans, UT 2026-02-03-0727 with little indication of flaring, and UT 2026-02-03-0749 with clear signs of something going on.
I see what seems to be a clear indication of flaring: regions in oxygen go darker when other wavelengths shine, somewhat like He I D3 does — albeit I’ve seen bright helium flares in both D3 and even in He I 5016 before. Instrument: the modified Sol’Ex spectroheliograph, with the 1800 ln/mm grating, and a front mounted full aperture Edmund Optics filter, CWL 775nm FWHM 25nm OD4. With the refractor thus stopped down to about 46/400, I recorded a time series. General observing conditions were poor, with passing clouds and bad seeing.
Processing: reconstructing the disks of the wavelength and a chosen continuum, dividing and stretching the two, and various other scripts for other kinds of output.
before, and during the flare. Hα for reference
Did I just record a solar flare, its imprint in some form, in the infrared oxygen triplet’s O I 7772 line? I think there is something in the data, I just don’t know what exactly.
Using a modified* Sol’Ex, I recorded the Sun at the infrared oxygen triplet, together with the neutral iron Fe I 7781. Passing clouds and some haze made the session less than ideal. The recordings have been reconstructed in JSolEx, and further processed with JSolEx’s ImageMath module.
The oxygen triplet lights up in an obvious manner near and beyond the disk edge, and prominences are visible after applying curves or subtracting a continuum from the disk at the wavelength. This is what I expected, based on what I read in the literature.
Between around 10:30Z-11:00Z, local time 12:30-13:00, there was a flare on the Sun, and for the second half of it, I recorded, still between the clouds.
blink compare: bad seeing and the clouds didn’t help
Paschen Gamma (enhanced) Solar Disk
Using the purpose built spectroheliograph, with stacked filters of 1075nm/50nm and 1100nm/50nm, I observed the Paschen gamma line (λ 10938.1Å), hoping to extract Hydrogen-resembling disk details. Neither the seeing, nor the low winter Sun helped the observation, worsening the already challenging endeavor. For this observation, I used the IMX 585MM camera, instead of the 678MM, hoping to get more photons per pixel, sacrificing some spectral resolution — mixed results.
Still, the image published here, with the signal enhanced, could well be the very first Paschen Gamma Solar Disk, bringing the details to life.
The line itself is much weaker than the He I D3, it is a nebulous blur on the spectrum on the disk center, and is dwarfed by He I 10830 in the flash spectrum. In this respect, it resembles the situation of the (Balmer) epsilon line, which is overwhelmed by the calcium valley, or that of the He I D3 which is somewhat subtly visible while scanning, and is fully revealed only through continuum subtraction.
