WR 104: Technical Questions

This page is for a more in-depth technical discussion of the risk posed by WR 104. There is jargon and techincal stuff ahead, and I won't try to explain all the terms. Ask your local friendly Astronomer if in doubt.

This discussion focusses on the risk to earth from a potential future gamma ray burst (GRB) in WR 104. Note that this was not a big part of my scientific article published in Astrophysical Journal. The article only had a paragraph or two about this scenario. But for understandable reasons, the press and public have picked up on this aspect of the work. It is maybe a little more widely relevant than, say, "radiative braking in colliding-wind systems" (which is discussed at length in the paper).

Furthermore let me add that any discussion of the future behavior of this complicated star system by necessity covers quite a lot of difficult astronomical terrain. I am in no way an expert in all the aspects one needs to cover to see the argument through. So I invite anyone reading this who sees a hole in my logic, or can add insight or detail to the discussion to get in touch and let me know what I have mistaken or missed.

The chain of logic from the observations through to a plausible credible future threat to Earth's biosphere has a lot of steps, and I will try to go through each of them with a realistic discussion of what we (or rather I) know and don't know about each step.


Will WR 104 explode?

OK - firstly remember that WR 104 is a binary star system with two very massive stars. So the answer is a double-yes. Both of these stars will eventually explode as a supernova. In fact, in the very distant future, in the unlikely event that the system is not disrupted by these two core-collapse supernovae, there is the possibility of a third inspiral and merger event of the two remnants. It is plausible that a short GRB might come of this, but this very-distant future is not the concern here and I won't discuss it further.

When will WR 104 explode?

The WC spectrum Wolf-Rayet component of the binary should explode sometime within the next few hundred thousand years. To an astronomer, this is the last known stable phase in the life of these massive stars, and it could go anytime. I know of no way to predict this more accurately. The other component, the supergiant O/B star, likely has a much longer fuse than this, and for the rest of this discussion I focus mainly on the WR component.

What causes GRBs?

There are two basic types of GRB, short and long (referring to the observed duration of the burst). Wikipedia or google will tell you more than you ever wanted to know. In brief, a popular model for the short ones is a coalescence of two massive bodies (like 2 neutron stars). Here we are concerned with the long bursts. There is pretty good observational evidence linking these with type Ic supernovae, which are in turn believed to be exploding Wolf-Rayet stars. So WR stars are the prime progenitor object for long GRBs.

So will WR 104 generate a GRB?

This is probably the biggest uncertainty in the story. Remember that the whole field of gamma ray burst astronomy is quite new, and ideas are changing and evolving all the time. The key property a WR supernova needs to turn it from an ordinary spherical explosion (harmless to Earth) into a ferocious, directed gamma-ray burst is spin. The star has to be spinning quite fast, and it is along this spin axis that the gamma ray beam will emerge.

This is where things get more complicated and murky. Could WR 104 be spinning fast enough? There is some evidence that Wolf-Rayet stars in our own galaxy would likely be slowed down too much to make a GRB. The prime observational characteristic of a WR is its strong wind - this is really what defines the whole class of star. But this wind can gradually, over time, carry away the angular momentum of the star, slowing it down below the critical threshold and defusing the potential GRB.

The efficiency of this mechanism depends on the strength of the wind, which in turn depends on the metallicity of the star. Lots of metals gives strong winds resulting in strong braking. So the argument runs that in our present-day galaxy, the metallicity is high enough to defuse most or all of the long GRBs. As we look into the distant cosmos and back in time, we see lower metallicity, and start to see GRBs going off at cosmological distances.

This should all be very reassuring - not only are we safe from WR 104, but we are safe from *all* long GRBs in our galaxy. Except that there are some niggling doubts. New research like this is controversial - nobody is really sure if it is quite right yet. Astronomers have not yet managed to treat all the complicated physics (such as magnetic fields) correctly in the models. If Adrian Melott is right in speculating that the Ordovician mass extinction at 440MYrs ago was caused by a GRB, then this is close enough in time to the present epoch (similar enough metallicity to today) to be a concern.

WR 104, however, has one more ace up its sleeve which may trump the arguments about the wind braking. WR 104 is a binary star, and in its earlier life, the two stars almost certainly interacted with one another in a roche-lobe overflow. There is strong evidence for this in my work, with a nearly exactly-circular orbit betraying a prehistory of tidal circularization. In fact, the existence of the WR star in this system may even rely on this turbulent past of Roche-lobe overflow and envelope stripping of the WR progenitor.

So, what does all this mean for the stellar spin? In such dynamical processes, all sorts of outcomes are possible, including both an increase or a braking in the spin of the stars in the binary system. For some binary star systems, little tugs from the binary orbit can compound over time spinning the stars up to high rates. Could this happen for WR 104? Unlikely, as the 8-month period means that the stars are actually quite far apart and effects like this will be weak.

So, back to our question of a whether to expect GRB in WR 104? Is it spinning quickly enough? Highly unlikley. In fact, I have had to revise the information presented here to downgrade the risk several times after talking to experts in the field. The bottom line is that only very rare supernovae go on to generate a GRB, so something quite special has to happen - and the discussion above about spin only begins to touch upon the complexity of the situation. This area of research is just too new, and there are too many unknowns.

But perhaps there is one exotic scenario which should give us a tiny pause for thought. If the OB companion swells to become a supergiant before the advent of the supernova in the WR component, then we might be in trouble. A common-envelope phase would be rapidly followed by an inspiral of the cores, and then all sorts of high-spin, supercritical core scenarios become possible again... I should hasten to add that this could not happen overnight, and the WR 104 system as it presently is structured could not do this without changing its form completely (giving us many milennia of warning).

If there is a GRB in WR 104, is it pointed our way?

Approximately, yes, but the detail of this question throws up another set of quite curly issues which need to be discussed further.

With my data, I can perform a fit to the inclination of the spiral plume being ejected by the binary system. Under the assumption that this should in reality be a pure, planar, Archimedian spiral (likely a good assumption), then it is easy to get a computer model to fit the inclination angle to the line-of-sight. The formal best-fit is an inclination of 12 degrees, but this is a hard number to extract. The difficulty lies in that a small tilt in the structure only gives a very small change in the aspect ratio. So values anywhere in the range from exactly face-on to 16 degrees are probably OK. In fact, maybe even a wider margin than that needs to be included, because we are looking through a partly optically-thin strucutre. Try tilting or tipping a wineglass or fishbowl - you don't get much idea how tilted it is from its appearance.

It is important to point out that what we have measured here is the axis of rotation of the Binary Star system. NOT the axis of spin of the WR star. The crosshairs of any future GRB will point along the spin axis. So, can these two axes be related?

The answer here is probably yes. In the event of a prehistory of Roche Lobe Overflow and tidal circularization, I think most models would predict an alignment of the spin and orbital axes of this binary star system (if, indeed, they were not already roughly aligned since the binary star was presumably formed from a common collapsing and rotating cloud). Still, this is certainly another link in the chain of logic that could to be firmed up more.

The final piece of information needed to answer the question is to know how wide is the beam from the burst? How big is the danger zone - a laser beam or a shotgun blast? Many different values for the opening angle of GRBs can be found in the literature. I have seen numbers from 2 to 20 degrees, and of course it is likely that there is a distribution of angles for different GRBs. As a rule-of-thumb for a typical GRB, I have adopted a value of 12 degrees.

In summary: maybe we *are* a little close for comfort to the danger zone, but there is still plenty of room for Earth to duck aside. Spectroscopic observations of the binary pair can certainly firm up the inclination of this system removing one source of doubt discussed in this subsection.

Conflicting data from Spectroscopy


Grant Hill of Keck observatory announced results from studies with a spectrograph on this system at the 2009 AAS meeting. He found significant variations in radial velocity consistent with a binary orbit. However, the orbital inclinations (hence plausibly the GRB beam direction) significantly larger than 12 degrees. No numbers were published, but vaules in the range 30-50 degrees were mentioned. Picking up the story at this time, press stories generally say "disaster averted".

This disagreement can't be reconciled with some clever model. One or other of these two inclinations is right; the other is wrong. Back in 2009, the smart money would probaly have been that my imaging (bleeding edge new technology) was wrong and that classical spectroscopy is likely right. However, Grant never actually published the data or orbital model in any peer-reviewed paper, so it was difficult to judge.

Unfortunately for the risk-to-Earth, I am now convinced that Grant's inclination is incorrect. Imaging from French researchers in 2018 using new Adaptive Optics technology completely vindicated my spiral model, including the inclination. Furthermore, I have tried reducing Grant's original Keck data again, and find that it is extremely difficult to extract anything reliable about the stellar orbits from it. Like most WR systems, the powerful lines from the fast wind which vary with time completely dominate everything, and no clean signal from the stellar photospheres could be found. In short, I think this was a case of a very difficult spectroscopic dataset to extract the specific information Grant was after.
This is the way Science goes, and on a personal note, Grant is a fantastic scientist still doing awesome work at Keck. I make a point to hang out every time I visit. Scientists must be able to flatly disagree with a colleague, gracefully.

Give some Rough Odds?

In nearby galaxies, the observed GRB rate is about one in every 100 million years. If we suppose a beaming of about 500:1, then the real rate is 500 times higher (the other 499 events are not pointed our way). So we have an approximate true rate of say one every 200 000 years. If there are 1000 Wolf-Rayet stars in our galaxy, and each has a full lifetime of a million years, then we have an approximate Wolf-Rayet supernova rate of one every 1000 years. So we can ballpark the approximate chances of any given Wolf-Rayet supernova producing a gamma-ray burst: it takes about 200 WR supernovae before you see one GRB. The odds, therefore, are at around the percent level. Given the massive uncertainties and ballpark numbers pulled in for this argument, I do not trust it all that much, but the odds are strongly against (but not overwhelming) a GRB. However, these statistical games are not an excuse for avoiding doing the physics properly, and ultimately more study is needed to figure out whether the system is dangerous.

In the worst-case scenario of an aligned GRB, what then?

Consequences are mainly related to global impacts on the biosphere and climate-change triggered by the large dose of radiation. This is all discussed in the work of Melott (and others), and it is also worth looking at the "Bad Astronomy" link on the previous web page. This is all the work of others, and I won't repeat it all here.

The good news is that we are not all *that* close to WR 104. For a fully-fledged GRB, we may be within the dangerous range but it is by no means a point-blank shot. If SN/GRBs form a continuum of events ranging from highly directed gamma beams through to slightly egg-shaped supernovae, then this means that we are safe from all but the more extreme focussed beam events. To carry a lot of clout over larger distances, a smaller cone angle is needed, tilting the odds and making it increasingly less likely that Earth is in the beam.

Bottom Line?

There are a lot of missing or incompletely understood links in the chain of logic connecting WR 104 to a real threat to Earth. But conversely, I don't know of any showstopper that categorically rules it out either. I think it is an obligation for scientists to keep an open and impartial mind to any event that could have consequences to a large number of people, and this would appear to be one of them, even if the odds appear quite long at the present time.

Back to Peter's Home Page

School of Physics
Sydney University
NSW 2006, Australia