The Repository Absolute Magnitude Space

Betelgeuse's Companion Has Four Published Masses

October 2, 2026 · Runtime 13:02 · Watch on YouTube

In December 2024 the SPHERE camera on ESO's Very Large Telescope recorded a faint point 52.32 milliarcseconds from Betelgeuse, about 1,200 times fainter than the star. Since then that point, Siwarha, has been given a distance, an orbit, an age and four different masses. None of them came out of the camera.

Transcript

A point found after a century 0:00

This is Betelgeuse, as the SPHERE camera on the Very Large Telescope sees it. And that faint point below and to the left, inside the cross hairs, is something astronomers have been trying to find for about a century. It is the star's companion.

Everything published about it rests on one night. On the sixth of December twenty twenty four, a camera called SPHERE recorded a point of light fifty two thousandths of a second of arc from the star, and about twelve hundred times fainter.

Four masses, none from the camera 0:32

Since then, that point has been given a distance, an orbit, an age, a spectral type, a name, and a mass. Four different masses, in fact, across four papers, spanning a factor of five. None of those masses came out of the camera. This episode ends on the number that did.

A six-year cycle read as an orbit 0:53

Betelgeuse was announced as a double star several times in the last century, and none of those claims was ever confirmed. What changed in twenty twenty four is that two teams read the star's slowest rhythm, a six year cycle in its brightness, as the signature of an orbit.

And an orbit makes a prediction. If the companion was real, it would swing furthest from the star in December twenty twenty four. Four telescopes were pointed at it around then. The results came out in twenty twenty five and twenty twenty six, and the decisive check is still ahead.

The point already has a name. In September twenty twenty five the International Astronomical Union approved Siwarha, Arabic for her bracelet, circling the hand of the giant.

The problem is contrast 1:43

The difficulty is contrast. Betelgeuse is among the brightest stars in the sky, and the companion sits almost on top of it. The air smears every star into a shimmering blob, so SPHERE bends a deformable mirror more than a thousand times a second to undo it.

Seeing, diffraction, disc, separation 2:02

Put the angles on one scale. The blur from the air that night was about half a second of arc. The finest detail an eight metre mirror can resolve in this red light is about twenty thousandths. Betelgeuse's own disc is forty three. The companion sits at fifty two, just past the edge of the star.

And the star is not a clean point to subtract. Its surface boils, its outflow is full of dust, and every one of those structures is brighter than the thing being searched for.

Letting the sky turn: 5,520 frames 2:36

So the team let the sky turn. The telescope held its optics fixed while the field of view rotated, which means the instrument's own flaws stay put, and anything real on the sky swings around the star. Over the sequence, the field turned by about forty seven degrees.

The observing log is short. Four sequences of one thousand three hundred and eighty frames each, five thousand five hundred and twenty images in all, taken in about two hours through a narrow red filter at six hundred and forty five nanometres. The air was steady that night. That is the raw material. Nothing else went in.

SPHERE was built to hide bright stars behind a small mask. For Betelgeuse, the observing log says, no mask was used at all.

Two algorithms: 6.1 and 5.1 sigma 3:28

Then two algorithms went looking through the same frames. One, called PACO, builds a statistical model of the noise, patch by patch. The other, principal component analysis, removes the most common patterns. PACO found the point at six point one sigma. The second method found it at about five.

Gemini North, three days later 3:52

Three days later and a hemisphere away, the Gemini North telescope in Hawaii caught the same spot with a speckle camera, at only one point five sigma. Alone, that is barely a hint. Next to SPHERE, it is the same position, found by two instruments that share nothing.

The Gemini team had also looked in February twenty twenty, in the middle of the Great Dimming, and saw nothing beside the star. One empty epoch and one with a point in it is what they used to argue this is not a background star that happens to line up.

The position agrees 4:28

Here is the position as each of them measured it. Fifty two, plus or minus two, from Gemini. Fifty two point three two, plus or minus zero point one eight, from SPHERE. Fifty three point two from the second algorithm. The angle around the star agrees too, within a few degrees.

The brightness does not: 4.7 times 4:50

The brightness is where they stop agreeing. The PACO extraction puts the companion at eight point two four ten thousandths of Betelgeuse's light, about one part in twelve hundred. The principal component analysis, run on the very same frames, gets one point seven six. That is a very different star.

Same telescope, same night, same images, and a factor of four point seven between the two brightness values. The authors trust the first, because the second method eats part of any faint source while it removes the glare, and they say so. It is a reasonable choice. It is still a choice.

So there are two measured numbers: a position on the sky, and a ratio of brightness. Everything from here on is built on top of them.

The chain of conversions 5:44

The chain runs like this. The position becomes a separation in space once a distance is chosen. The separation becomes an orbit once a mass is chosen for Betelgeuse. The brightness becomes a mass for the companion once an age is chosen. Each arrow is an assumption.

The first arrow is the one from the last episode. The paper adopts one hundred and sixty eight parsecs, the distance from stellar modelling, and gets eight point eight astronomical units.

Kepler, and the distance that was chosen 6:17

The second arrow is Kepler. With sixteen and a half to nineteen suns for Betelgeuse, that separation means a period of at least five and a half to six years. And that matches the six year brightness cycle that predicted the companion in the first place.

Now run the same arithmetic at two hundred and twenty two parsecs, the radio distance from last time. The separation grows to eleven point six astronomical units, and the shortest possible orbit stretches past eight years. It no longer matches the six year cycle. The paper says so, in those terms.

Then it draws the conclusion. The closer distance, it says, better fits the hypothesis that the cycle is an orbit. That is stated openly. But notice what it means: the distance was chosen by the idea being tested.

Four published masses 7:14

The third arrow is the one that made the headlines: the mass. This is where the literature stops converging. The same object has now been weighed four times, and each weighing came out heavier than the one before, with a single exception.

Here they are in the order they appeared. One point one seven suns, as a minimum. Then zero point six. Then one point six. Then two point six to three point one. From the lightest to the heaviest is a factor of five, in two years, for one point of light.

The first two came from motion, not light. A century of velocity records shows Betelgeuse rocking toward and away from us by about one and a half kilometres per second on the six year cycle, and that rocking weighs whatever pulls it.

How a brightness becomes a mass 8:08

The last two came from light, and the SPHERE value is a chain of its own. Take the brightness ratio. Model Betelgeuse at thirty seven hundred kelvin, scaled to amateur photometry from that same week. Correct for dust. Put both stars at one hundred and sixty eight parsecs. Assume they were born together.

Only then read a mass off a stellar model. Every link is an assumption, and the authors flag the loosest one. The dust around Betelgeuse is patchy, it could not be measured in these data, and so the true range is, in their words, probably broader.

Follow that chain to the end and you get a portrait. A star of eleven to twelve thousand kelvin, about twice the Sun's radius, fifty to ninety times its luminosity, just settled onto the main sequence. The second filter, tuned to hydrogen, showed no glow, which fits a young star rather than one still forming.

Hubble and Chandra saw nothing 9:12

And two other telescopes disagree with the answer. At the same point in the orbit, the Hubble Space Telescope looked in far ultraviolet and saw nothing. Chandra looked in X rays and saw nothing. A hot young star of three suns is bright in the ultraviolet.

Lay the four instruments side by side. In visible red light, a point at six sigma. In blue light, a hint at one and a half. In far ultraviolet, nothing, and the team behind that search says its limit rules out a hot star above about two suns. In X rays, nothing either.

Reflected light, and the LL Pegasi wake 9:54

This August the ultraviolet team posted a way out, not yet peer reviewed. The light SPHERE caught may not be the companion's own. It may be Betelgeuse's light, reflected off the shock of gas and dust the companion plows up in the outflow. ALMA has already mapped a companion's wake around another giant, LL Pegasi.

So the same point supports two readings. If the light is the companion's own, it is a hot blue star of two and a half to three suns. If the light is reflected, its brightness says little about its mass, and the lighter estimates come back into play.

Notice which numbers survive either way. The masses from motion do not depend on where the light comes from. The masses from brightness do. And the brightness ratio itself still stands in both readings, because it is a measurement.

Can dust make the cycle? A tenth of it 10:51

The SPHERE paper also asks whether the companion could cause the six year cycle by clearing dust from our line of sight. It estimates that the companion evaporates dust out to about a third of an astronomical unit, and would open a clear window on about twelve per cent of the star's disc.

At best, that dims or brightens the star by about a tenth of a magnitude. The cycle it was meant to explain is half a magnitude. The authors propose that the cycle comes mainly from the motion itself, not from the dust.

What the instrument returned 11:25

Put the whole paper through one filter: what did the instrument return? A separation, an angle around the star, and a ratio of brightness. Everything else, the astronomical units, the period, the age, the mass, needed at least one number the camera never saw.

Still a candidate until 2027 11:45

Even the word companion is still provisional. The authors call it a candidate until it is seen on the opposite side of Betelgeuse, half an orbit later. If the six year orbit is real, the next widest swing falls around November twenty twenty seven. That observation settles more than any model can.

52.32 +/- 0.18 milliarcseconds 12:06

So the episode ends on the number the camera returned. Fifty two point three two thousandths of a second of arc, plus or minus zero point one eight. About a third of one per cent. That is the width of a coin seen from sixty miles away, measured to better than a tenth of a millimetre.

Next: the six-year cycle and Gaia 12:26

One problem is left, and it is bigger than this star. The companion was predicted from a six year cycle, and about a third of luminous red giants have a cycle like it. In April, a team that includes one of the companion's own predictors checked those stars with Gaia.

Most of them behaved like single stars. And the paper says Betelgeuse's cycle may not be an orbit either. Next time: the six year cycle, and why Gaia says it is not caused by a companion, even though one was found.

Description and sources

What the instrument returned is a separation (52.32 +/- 0.18 mas), a position angle (117.1 deg) and a brightness ratio ((8.24 +/- 1.04) x 10^-4), found at 6.1 sigma. The same frames run through a second algorithm give a brightness 4.7 times lower. Everything else is conversion: 8.8 AU needs an adopted distance of 168 pc; an orbit of at least 5.5 to 6 years needs a mass for Betelgeuse, and at 222 pc the orbit stops matching the six-year brightness cycle, so the paper adopts the distance that 'better fits the hypothesis'. The mass of 2.6 to 3.1 suns needs a stellar model, dust correction, the distance and an assumed common age. Across two years the published masses run 1.17 (minimum), 0.6, 1.6 and 2.6 to 3.1. Hubble saw nothing in far ultraviolet and Chandra nothing in X rays, and a preprint from August 2026 proposes that much of the light is Betelgeuse's own, reflected off the companion's bow shock. The object is still formally a candidate until it is seen on the far side of its orbit, around November 2027.

PRINT-READY, FROM THIS CHANNEL

The Conversion Chain, Vol. 01 — the method, and four numbers taken apart with it

https://therepository.gumroad.com/l/the-conversion-chain

PRIMARY SOURCES

Montarges, M. et al. (2026). VLT/SPHERE imaging of the candidate companion of Betelgeuse. A&A 711, L12; ESO release eso2611 (28 July 2026) -- SPHERE-ZIMPOL, 3 and 6 December 2024, no coronagraph, separation 52.32 +/- 0.18 mas, PA 117.12 +/- 0.60 deg, flux ratio (8.24 +/- 1.04) x 10^-4 at 644.9 nm, S/N 6.1 (PACO); PCA 53.23 mas, (1.76 +/- 0.29) x 10^-4, S/N ~5; 8.80 AU at an adopted 168 pc, 11.6 AU at 222 pc; period at least 5.5-5.9 yr (8.2-8.7 yr at 222 pc); 2.6-3.1 Msun, B8.5V-B9.5V, 'probably broader'.
Howell, S. B. et al. (2025). The Probable Direct-imaging Detection of the Stellar Companion to Betelgeuse. ApJ 988, L47 -- Gemini North / 'Alopeke, 9 December 2024, 52 +/- 2 mas, PA 115 +/- 5 deg, ~1.5 sigma; no detection on 17 February 2020; ~1.6 (1.4-2.0) Msun; next maximum elongation 26 November 2027.
Goldberg, J. A., Joyce, M. & Molnar, L. (2024). A Buddy for Betelgeuse. ApJ 977, 35 (arXiv:2408.09089) -- M sin i = 1.17 +/- 0.7 Msun from the long secondary period.
MacLeod, M. et al. (2025). Radial Velocity and Astrometric Evidence for a Close Companion to Betelgeuse. ApJ 978, 50 (arXiv:2409.11332) -- 0.6 +/- 0.14 Msun, K ~ 1.5 km/s, period 2,110 d.
Goldberg, J. A. et al. (2025), ApJ 994, 101 (HST far-UV non-detection); O'Grady, A. J. G. et al. (2025), ApJ 992, 107 (Chandra non-detection).
Goldberg, J. A. et al. (2026). Disco in the Dust: reflected light at the bow shock around Betelgeuse's companion. arXiv:2608.11672 -- preprint, not peer reviewed.
Shariat, C., El-Badry, K., MacLeod, M. & Leiner, E. (2026). Gaia astrometry disfavors a binary origin for long secondary periods. arXiv:2604.09767 (next episode).
ALMA (ESO/NAOJ/NRAO)/H. Kim et al. -- LL Pegasi 3D view (potw1710a). Name Siwarha approved by the IAU WGSN, 22 September 2025.
Arithmetic shown on screen, recomputed: 1/8.24e-4 = 1,214; 8.24/1.76 = 4.7; 52.32 mas x 168 pc = 8.79 AU, x 222 pc = 11.61 AU; 1.22 x 644.9 nm / 8.2 m = 19.8 mas; 11.22+12.31+12.20+11.03 = 46.8 deg of field rotation; 0.18/52.32 = 0.34%; 3.1/0.6 = 5.2.

ARCHIVE PLATES: all 21 are credited on screen in the video and listed in this project's manifest.

Absolute Magnitude takes one number the internet repeats and asks what an instrument actually returned. Every episode ends on a measured figure with its source on screen, and says plainly which part of it was converted afterwards.

No AI-generated imagery is used on this channel. Every frame of archive footage is real, publicly released material, credited on screen with its licence; every diagram is drawn in code from the numbers in the cited papers.

Corrections are pinned, not quietly edited.

#Betelgeuse #Siwarha #Astronomy #ESO #AbsoluteMagnitude