Nemesis, the hypothetical "death star", is supposed to be a massive body that orbits the Sun at long distances and periodically sends comets from the Oort Cloud into the inner solar system. These comets impact the Earth and cause extinction events. It hasn't been found, and the theoretical case for it is not compelling anyway.

My question here is about this curious line from Richard Muller's page at LBL.

Fortunately, several all-sky surveys are underway that should find Nemesis in the next few years, if it is there, and rule out Nemesis if they don't. (Nemesis could hide if it were a black hole, but that is not very plausible.) These surveys include Pan-Starrs and the LSST.

How do we know it's not plausible for Nemesis to be a black hole? For that matter, how do we know it's not a neutron star?

  • $\begingroup$ See also astronomy.stackexchange.com/questions/10124/… $\endgroup$
    – ProfRob
    Sep 26, 2019 at 8:16
  • 2
    $\begingroup$ arxiv.org/abs/1909.11090v1 $\endgroup$
    – Keith
    Sep 26, 2019 at 23:21
  • $\begingroup$ Answering this question might require divining what the author means by "rule out" and "not very plausible". If you need a dark, quiet object of almost any mass, you can propose a primordial black hole and explore the consequences, as in the paper Keith links to above. But perhaps "not very plausible" is intended to mean that a stellar black hole isn't very plausible, and when Muller said "rule out Nemesis", he meant "rule out any Nemesis belonging to any class of object we actually know about, as opposed to classes of object we can speculate". $\endgroup$ Sep 27, 2019 at 15:56

2 Answers 2


If the Sun had been born in a relatively wide binary system with a star that was to become a black hole or neutron star via a supernova, then (a) it is quite likely that such a system would be disrupted by that supernova and we would not be in a binary system now; (b) there should be evidence of the supernova in the form of very high abundances of the daughters of certain short-lived radionuclides incorporated into solar system material. There is some evidence of the latter, but not I think enough for the Sun to have been in a binary system with such a star (though I might be checking this).

An alternative argument is that the Sun is captured in orbit by the stellar remnant at a later date. This avoids the supernova problems, but the capture process is inherently unlikely in our Galaxy once stars have left their birth environments, especially capture which is tuned rather precisely to yield just less than zero for the resultant system potential energy of a very wide binary. Capture by a "normal" star would in any case be much more likely than capture by a relatively rare compact object.

  • $\begingroup$ The sun is what would be captured not the other way around. $\endgroup$
    – Joshua
    Sep 26, 2019 at 16:35
  • $\begingroup$ Nemesis would be the more massive object; put the embryonic sun on a much more energetic orbit and have it be the captured body rather than the capturing body and it looks a little more sane. $\endgroup$
    – Joshua
    Sep 26, 2019 at 16:38
  • 7
    $\begingroup$ @Joshua "An alternative argument is that the Sun is captured..." I can't see which bit of that sentence implies that it is the Sun doing the capturing? $\endgroup$
    – ProfRob
    Sep 26, 2019 at 16:41

I think we can knock this down directly. The predicted semi-major axis is 1.5 light years. It was plausible at the time it was written that such an object could remain hidden. It is no longer plausible that a black hole with the required minimum mass to form would have escaped decades of automated asteroid search. It's too close, has too much parallax, should have too much transverse velocity, and would have shown up as gravitational lensing in the asteroid searches by now. They're looking for objects that aren't the same from frame to frame. This would trip the detectors.

The first-pass processing for asteroid search is simply take two plates of the same part of the sky and diff them by something that's little more than XOR. You now have bright spots where something moved. These are checked against a known table of foreground objects, and anything not matched gets looked at by a human. An unexpected gravitational lens will show up because it didn't cancel between the frames because the light magnitude was too different. If it's not dead-on (and this will be most of the time) it will also have moved.

The orbital mechanics involve in the capture would greatly distrupt the Sun's galactic orbit in the plane of the capture. We note that life appeared on the Earth's surface almost as soon as it was sufficiently cool, and was not fried by the galactic radiation. This in turn requires the sun to have the Z axis oscillation around the galaxy it has or smaller; which constrains Nemesis to be on or near the galactic plane.

But I've got a hole I can't close. If we ignore the normal Nemesis postulates enough we can end up with it being the Sun's companion for the entire time. The typical capture process should be disruptive to planetary formation around the embryonic Sun but the postulated orbit is more than far enough away to avoid this problem. This requires an exotic capture method, but we were most likely going to end up with one anyway.

  • 1
    $\begingroup$ Can you be clear what lensing you expect to see in what objects. Something 1.5 light years away (it depends on the mass of the black hole actually) has a parallax motion of about 2 arcseconds. I agree that this would have a lensing effect on the positions of background stars and something like this might show up in a careful analysis of the full Gaia catalogue when it emerges. The presence of the black hole would not affect the measured positions of foreground sources. $\endgroup$
    – ProfRob
    Sep 26, 2019 at 17:17
  • $\begingroup$ Oh, but you mean a "wobble" in the position of background stars would be seen in surveys for foreground asteroids. Well, yes perhaps but that would rely on it being in the foreground of suitably close stars with very carefully measured positions. I may have done such a claculation - I will check. $\endgroup$
    – ProfRob
    Sep 26, 2019 at 17:20
  • 3
    $\begingroup$ Check astronomy.stackexchange.com/questions/16578/… I think a 10 solar mass black hole at 1.5 ly would have to get within 0.04 arcsec of the position of a distant star to deflect the light by 20 microarcseconds. There is only a 0.4 in a million chance of this for a background star with $V<15$ and 20 microarcseconds is only detectable with Gaia. In a more typical asteroid survey you might detect deviations (if you were really looking) that were 1000 times bigger and a million times less likely. $\endgroup$
    – ProfRob
    Sep 26, 2019 at 17:31
  • $\begingroup$ Using the same formula, light rays passing within 0.25 degrees of the Sun's centre at a distance of 1 au will be deflected by ~2 arcsecs (as observed). $\endgroup$
    – ProfRob
    Sep 26, 2019 at 17:54
  • $\begingroup$ Joshua, good answer. To add, wouldn't the wobble from the barycenter of the Sun & Nemesis be apparent from Earth? We have been observing the Sun for millennia, and would have noticed if the background stars were periodically out of position. $\endgroup$
    – Jim
    Sep 27, 2019 at 15:58

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .