Above: A bucket blocks the Sun at the center of a 22° ice crystal halo.
.
There was a 22° halo around the Sun yesterday, a result of refraction by ice crystals high above the summer-like Vermont landscape. The inner edge of the halo was red, but the other colors were not well separated. As Galen and I admired the sight, I was wondering out loud whether there was an invisible smaller diameter halo formed by infrared light, and whether an infrared camera could capture it. I was bemoaning the problem of precisely comparing the diameter of halos in two photographs, when Galen said one word, Infragram. An Infragram camera makes an infrared image in one channel, and a visible light image in another, and both images are precisely aligned. I ran for the cameras.

The three channels of a normal color photo of the halo. In Fiji, I made a transect one pixel wide from that sharp stick near the sun directly to the right through the halo. Fiji produced a data file of the brightness for each pixel along the same transect in each of the three channel images.
.

The brightness of each color channel in a normal color photo of the halo. The transect from the little stick through the halo shows the hump of the halo is at a different place for each color. The radius of the red halo is slightly smaller than that of the green or blue halos.
.
Results for a normal color photo suggest that this approach might work. The halo captured in the red channel of the photo has a slightly smaller radius than the halos of other colors.
.

Three color channels (red, green, blue, left to right) in an Infragram photo taken with an A590 with a Wratten 25A filter replacing the IR block filter. The yellow lines are the transects from the sharp stick on the left through the halo. The transect I quantified was 10 pixels wide. It was easy to start the transect at the exact same spot on the sharp stick.
.

Detail of the 2000 pixel long transects near the hump of the halo. The blue channel in this Infragram camera records only near infrared light, and the bright hump of the invisible infrared halo is indeed to the left of (on the inside of) the brightest part of the halos in the red or green channels. The individual dots are averages of 10 pixels ("vertical columns") at each distance along the transects. The lines are 30 pixel moving averages.
.
So it would appear that there is indeed a smaller diameter invisible halo of infrared light inside the common 22° ice crystal halo. One fewer thing to wonder about, thanks to Public Lab tools, and Galen.