Wednesday, February 11, 2015

Do black hole firewalls have observable consequences?

In a paper out last week Niayesh Afshordi and Yasaman Yazdi from Perimeter Institute study the possibility that black hole firewalls can be observed by their neutrino emission:
In their work, the authors assume that black holes are not black, but emit particles from some surface close by the black hole horizon. They look at the flux of particles that we should receive on Earth from this surface. They argue that highly energetic neutrinos, which have been detected recently but whose origin is presently unknown, might have originated at the black hole firewall.

The authors explain that from all possible particles that might be produced in such a black hole firewall, neutrinos are most likely to be measured on Earth, because the neutrinos interact only very weakly and have the best chances to escape from a hot and messy environment. In this sense their firewall has consequences different from what a hard surface would have, which is important because a hard surface rather than an event horizon has previously been ruled out. The authors make an ansatz for two different spectra of the neutrino emission, a power-law and a black-body law. They then use the distribution of black holes to arrive at an estimate for the neutrino flux on Earth.

Some experiments have recently measured neutrinos with very high energies. The origin of these highly energetic neutrinos is very difficult to explain from known astrophysical sources. It is presently not clear how they are produced. Afshordi and Yazdi suggest that these highly energetic neutrinos might come from the black hole firewalls if these have a suitable power-law spectrum.

It is a nice paper that tries to make contact between black hole thermodynamics and observation. This contact has so far only been made for primordial (light) black holes, but these have never been found. Afshordi and Yazdi instead look at massive and solar-mass black holes.

The idea of a black hole firewall goes back to a 2012 paper by Almheiri, Marolf, Polchinski, and Sully, hereafter referred to as AMPS. AMPS pointed out in their paper that what had become the most popular solution attempt to the black hole information loss problem is in fact internally inconsistent. They showed that four assumptions about black hole evaporation that were by many people believed to all be realized in nature could not simultaneously be true. These four assumptions are:
  1. Black hole evaporation does not destroy information.
  2. Black hole entropy counts microstates of the black hole.
  3. Quantum field theory is not significantly affected until close to the horizon.
  4. An infalling observer who crosses the horizon (when the black hole mass is still large) does not notice anything unusual.
The last assumption is basically the equivalence principle: An observer in a gravitational field cannot locally tell the difference to acceleration in flat space. A freely falling observer is not accelerated (by definition) and thus shouldn’t see anything. Assumption three is a quite weak version of knowing that quantum gravitational effects can’t just pop up anywhere in space that is almost flat. Assumption 3 and 4 are general expectations that don’t have much to do with the details of the attempted solution to the black hole information loss problem. The first two assumptions are what describe the specific scenario that supposedly solves the information loss problem. You may or may not buy into these.

These four assumptions are often assigned to black hole complementarity specifically, but that is mostly because the Susskind, Thorlacius, Uglum paper nicely happened to state these assumptions explicitly. The four assumptions are more importantly also believed to hold in string theory generally, which supposedly solves the black hole information loss problem via the gauge-gravity duality. Unfortunately, most articles in the popular science media have portrayed all these assumptions as known to be true, but this isn’t so. They are supported by string theory, so the contradiction is a conundrum primarily for everybody who believed that string theory solved the black hole information loss problem. If you do not, for example, commit to the strong interpretation of the Bekenstein-Hawking entropy (assumption 2), which is generally the case for all remnant scenarios, then you have no problem to begin with.

Now since AMPS have shown that the four assumptions are inconsistent, at least one of them has to be given up, and all the literature following their paper discussed back and forth which one should be given up. If you give up the equivalence principle, then you get the “firewall” that roasts the infalling observer. Clearly, that should be the last resort, since General Relativity is based on the equivalence principle. Giving up any of the other assumptions is more reasonable. Already for this reason, using the AMPS argument to draw the conclusion that black holes must be surrounded by a firewall is totally nutty.

As for my personal solution to the conundrum, I have simply shown that the four assumptions aren’t actually mutually incompatible. What makes them incompatible is a fifth assumption that isn’t explicitly stated in the above list, but enters later in the AMPS argument. Yes, I suffer from a severe case of chronic disagreeability, and I’m working hard on my delusions de grandeur, but this story shall be told another time.

The paper by Afshordi and Yazdi doesn’t make any use whatsoever of the AMPS calculation. They just assume that there is something emitting something close by the black hole horizon, and then they basically address the question what that something must do to explain the excess neutrino flux at high energies. This scenario has very little to do with the AMPS argument. In the AMPS paper the outgoing state is Hawking radiation, with suitable subtle entanglements so that it is pure and evolution unitary. The average energy of the emitted particles that is seen by the far away observer (us) is still zilch for large black holes. It is in fact typically below the background temperature of the cosmic microwave background, so the black holes won’t even evaporate until the universe has cooled some more (in some hundred billion years or so). It is only the infalling observer who notices something strange is going on if you drop assumption 4, which, I remind you, is already nutty.

So what is the merit of the paper?

Upon inquiry, Niayesh explained that the motivation for studying the emission of a possible black hole firewall in more general terms goes back to an earlier paper that he co-authored, arXiv:1212.4176. In this paper they argue (as so many before) that black holes are not the endstate of black hole collapse, but that instead spacetime ends already before the horizon (if you want to be nutty, be bold at least). This idea has some similarities with the fuzzball idea.

I am generally unenthusiastic about such solutions because I like black holes, and I don’t believe in anything that’s supposed to stabilize collapsing matter at small curvature. So I am not very convinced by their motivation. Their powerlaw ansatz is clearly constructed to explain a piece of data that currently wants explanation. To me their idea makes more sense when read backwards: Suppose the mysterious highly energetic neutrinos come from the vicinity of black holes. Given that we know the distribution of the black holes, what is the spectrum by which the neutrinos should have been emitted?

In summary: The paper doesn’t really have much to do with the black hole firewall, but it deals instead with an alternative end state of black hole collapse and asks for its observational consequences. The good thing about their paper is that they are making contact to observation, which is rare in an area plagued by lack of phenomenology, so I appreciate the intent. It would take more than a few neutrinos though to convince me that black holes don’t have a horizon.


Rastus Odinga Odinga said...

"Already for this reason, using the AMPS argument to draw the conclusion that black holes must be surrounded by a firewall is totally nutty. "

Thank you.

But there is a question of basic importance here: how is it possible that so many intelligent people believe that "drama" is a real option? I fear that it is due to a fundamental shortcoming in the way we are educating physicists.

By the way, it is not correct to say that this particular paper has not much to do with firewalls. It has in fact absolutely nothing to do with them.... people are getting desperate to persuade other people to download their papers. I am waiting for a paper titled "I AM SATAN INCARNATE!!!!" That ought to get some attention. Maybe even some cites.

The Universe said...

I root for relativity, but I’ll tell you this: you give up on the equivalence principle. See the quotes by Synge and Ray on page 20 in . It’s an “enabling” principle only. It applies to an infinitesimal region, a region of no extent, which means it doesn’t actually apply. As for being roasted, see An Apologia for Firewalls where reference 87 is to Friedwardt Winterberg’s Gamma Ray Bursters and Lorentzian Relativity. IMHO this is the real firewall. See Einstein talking about the speed of light being spatially variable. If you fall into a black hole, you fall faster and faster. But the “coordinate” speed of light is getting slower and slower. Falling bodies do not slow down, and nor can they fall faster than the local speed of light. Something has got to give, that something is you, and gamma ray bursters aren’t nutty. Nor is the idea that spacetime ends at the event horizon, or Oppenheimer’s original frozen-star black hole. What stabilizes collapsing matter is that the speed of light goes to zero. It can’t go lower than that, so there is no more gravitational gradient, and no more curvature. The paper by Afshordi and Yazdi doesn’t say this, but maybe they’re getting warm.

Uncle Al said...

"If you give up the equivalence principle, then you get the “firewall” that roasts the infalling observer. Clearly, that should be the last resort, since General Relativity is based on the equivalence principle." Photons and matter never fall identically within a gravitational field, by a factor of 2.

Einstein-Cartan-Kibble-Sciama gravitation contains chiral spacetime torsion EP violation. Eötvös balance, chemically and macroscopically identical single crystal enantiomorphic space group test masses, all observables except chirality exactly cancel. Theory alone cannot falsify a defective postulate - that boson photon vacuum symmetries are exactly true for fermion quarks (hadrons). LOOK.

kashyap vasavada said...

Do they define vicinity precisely? If neutrinos are emitted by firewalls then they should emit photons and other particles also.In that case, we will have to change the whole concept of black hole!

hush said...

Middle-aged couple searching for:

Any surface area of any event horizon with any mass.

Will except disjoint existence, Bell violation and collapse.

Bob and Alice.

Sabine Hossenfelder said...

kashyap: No, they don't. It's something like a Planck length. Yes, they do emit other stuff as well, please read post. They argue that neutrinos have best chances to reach us. Best,


Uncle Al said...

Phys. Rev. Lett. 114, 061301 (2015), arXiv:1410.7590, DOI: 10.1103/PhysRevLett.114.061301
"...massless particles like photons are no longer confined to traveling exactly on geodesics...bend differently depending on their spin."

Physical theory chokes on chirality. Do a Heimlich maneuver by looking. Other than looking, there is little quantum gravitation and SUSY will not pursue.

nicolas poupart said...

It is reminiscent of the EPR paradox : "watch how it's absurd, it's probably fake" becomes "watch how it's absurd, it's probably true".

Arun said...

As for my personal solution to the conundrum, I have simply shown that the four assumptions aren’t actually mutually incompatible. What makes them incompatible is a fifth assumption that isn’t explicitly stated in the above list, but enters later in the AMPS argument. Yes, I suffer from a severe case of chronic disagreeability, and I’m working hard on my delusions de grandeur, but this story shall be told another time.

I like! I really, really like!

hush said...

A holographic principle provides preservation (conservation?), not loss.

The principle needs a surface area too.

A vicinity where any observable favors the principle.

once was physicst said...

"but this story shall be told another time"

That would be great.

Amos said...

Uncle Al wrote: " Photons and matter never fall identically within a gravitational field, by a factor of 2." What do you mean by that? If, for example, we could measure the deflection of neutrinos (moving at nearly the speed of light) from a distant star as they graze the Sun, are you saying the deflection would differ from the deflection of light by a factor of 2? It's hard for me to imagine such a gross violation of the equivalence principle.

Munish Ratanpal said...

Hi Bee, wanted to know your thoughts about this recent paper ( which predicts no singularities by applying quantum corrections to the RayChaudhari equation.

Sabine Hossenfelder said...


As I have already explained on both G+ and facebook, it's a combination of several ideas each of which is already speculative, controversial and problematic. Adding several of these doesn't make the problems go away, it just makes the result even more controversial and problematic. I do not have the faintest clue why stuff like this gets hyped. I will make a prediction for you: you will never hear anything of this supposed breakthrough again. Best,


Uncle Al said...


Your objection's answer (plus proton cosmic rays) is GR. Don't validate, falsify. Two masses locally vacuum free fall. EP direction divergence is bad, speed divergence is acceptable. 1) Prior observation is not contradicted. 2) Theoretic curve fittings are sourced. 3) A never tested observable is active.

Bright corner streetlights leave a dark middle of the block where answers hide. Gravitation assumes something empirically defective. Observe outside its postulates, respecting (1), (2), (3).

Amos said...

Uncle Al: “Your objection's answer is GR.” I don’t think GR answers my objection, because my objection was consistent with GR. In fact, the equivalence principle alone implies that the gravitational deflection of neutrinos (at near light speed) and photons is the same. I object to the claim that matter falls differently than light by a factor of 2. The only answer GR would give to my objection is “Yes, Amos, you are right, as usual!”

Uncle Al: “Two masses locally vacuum free fall. EP direction divergence is bad, speed divergence is acceptable.” You seem to be saying we can rule out “direction divergence” for two free-falling particles, but not “speed divergence”. That seems like a non-starter to me, because if two particles are following the same spatial path (at different speeds) in one frame, they are following different spatial paths in another frame. There’s no meaningful distinction between “direction divergence” and “speed divergence” – barring some absolute frame.

Uncle Al: “Prior observation is not contradicted.” That’s a good principle, but it requires some judgment. Strictly speaking, prior observations can never be contradicted, because they are prior. We can never wade in the same river twice. Empiricism requires us to make judgments about what is “sufficiently the same” (did we ever conduct this experiment during a full moon?), and how far we can generalize our inductions (does it matter?). For example, some people would say we can already rule out chirality violations of the equivalence principle, based on prior observations that contradict an absolute frame. Others might disagree, but we can’t avoid making judgments. There can be no empiricism without rationalism.

Uncle Al said...


1) GR accurately predicts all observed trajectories (orbits). Observe "nicht einmal falsch" test masses.
2) "There’s no meaningful distinction between “direction divergence” and “speed divergence" The universe has no preferred direction; speed is arbitrary. Pawnbroker rotation: Three small niobium-plated balls in hard vacuum are Meissner levitated - enantiomorphic space groups single crystal alpha-quartz plus amorphous fused silica. Now apply Eötvös experiment divergences over 24 hours. Two balls oppositely rotate.
3) "prior observations can never be contradicted" Euclid and maps. Newton and GR, QM. SUSY-predicted proton decay and Super-Kamiokande.

Bee decries EP violation. Uncle Al offers it six different footnoted ways. Look.

Zephir said...

The emanation of neutrinos and gamma rays belongs into AWT predictions (actually it was already observed - just ignored like any other finding inconsistent with official belief). Most of black holes are dense neutrino stars in essence and their event horizons are equivalent with their physical surface (aka "firewall") - so that they emanate light-weight particles like any other stars - preferentially via their polar jets. This is the primary reason, why the black holes in mature galaxies are rather quiet unobtrusive objects - they do radiated most of their matter into outside already...

Patat Je said...

I have a question for you, writer of backreaction. Lately, I have been thinking a lot about black holes. I realized that the situation at the even horizon of a black hole is truly special. I think scientists are making a mistake when they believe that there is no drama involved at a black hole horizon. The think same mistake is made as in the past concerning the twin paradox. People thought that the situation for one half of the twin who is stationary is the same for the other half of the twin who is accelerating. This is not true because the first feels no acceleration, while the other does, so their situations are not equivalent.

When a person crosses the event horizon, feet first, it is possible for his head to still escape. And this person temporarily can't receive information from his/her feet. This seems to me quite dramatic. Outside a black hole this is no problem, you can always communicate with your feet, and your feet are not lost forever. The two situations are not equivalent.

But perhaps I'm totally wrong. I want to ask you writer of backreaction, because you seem to be the expert. You will probably say I am totally wrong. :)

Sabine Hossenfelder said...


Yes, you are totally wrong... The event horizon is a global construct. It is a) not a physical thing, just an imaginary boundary and b) it cannot be noticed locally. The event horizon would not prevent the information from your feet reaching your head because you cannot sit stretched across the event horizon. It's not possible to stay there, half in, half out, you inevitably fall in. And that allows the information to reach your head, just as usual. Best,


Patat Je said...

I can imagine two very small objects very close to each other, and having almost the same trajectories. One small object moves just past the event horizon and doesn't fall in, the other falls in. I could imagine something similar happening to people moving past a very large black hole without large tidal forces. They don't feel their bodies torn apart, but they discover parts of their body are suddenly gone!

Sabine Hossenfelder said...


No they don't. Take the equations, do the math. I'm telling you it's not the case. I don't understand why you insist on what your 'imagination' tells you. It's wrong.

Phillip Helbig said...

I think some people are confused because, for a black hole of high mass, the tidal effects at the event horizon can be very small. However, the escape velocity is still the speed of light, so anything which passes through the event horizon will start falling fast.

In Newtonian theory, assuming that light is a particles which travels at the speed of light, one also gets the speed of light as the escape velocity at the same radius as the Schwarzschild radius. However, in Newtonian theory, while one would have to (impossibly) move faster than the speed of light on an inertial trajectory to cross the "event horizon" to the outside, one could still escape, using a rocket, ladder, etc.

Patat Je said...

I reread what I wrote down. I said: "One small object moves just past the event horizon and doesn't fall in".

I think people could interpret this as me saying that the object goes into the black hole, and then goes out again. But that's not what I mean. I actually meant, the object moving alongside the edge of the event horizon, while still staying outside of it. I could have better used the word "alongside" instead of "past". I meant with the word "past" more like driving past a train in my car when the railroad and the highway run alongside and close to each other, and when initially the train is ahead of me, while I am faster in my car.

Imagine an object in orbit around the huge black hole, but with the perihelion almost nearing the event horizon, but just staying outside of it. Now also imagine a second much smaller object in orbit around the first object. I'm sure it is possible for the second object to cross the event horizon (if you place the perihelion close enough to the event horizon) ultimately being absorbed by the black hole, while the first object stays just outside of the event horizon, moving merrily along on its orbit around the black hole, not noticing anything unusual, except that his companion is gone. I think an astronaut in close orbit around a black hole could lose his feet this way!

I know this is an unusual situation and you have to prepare everything very precisely. But I think it's possible.

I think you could explain it to me what I'm doing wrong, Phillip Helbig.

And I have to add, that I do have some background in physics, albeit on a sophomore level. I dropped out in the fourth year. I have followed courses about special and general relativity. I have read about metrics and tortoise coordinates and such. I also have the book "Spacetime and Geometry: An Introduction to General Relativity " in my possession, written by Sean Carroll.

Sabine Hossenfelder said...


Both I and Phillip have now told you that you are wrong, and have told you why you are wrong. I do not understand why you continue to insist on your "imagination". There are no stable orbits close by the black hole horizon and everything that's inside will inevitably fall towards the singularitly. The astronaut you imagine will not 'lose his feet', he'll simply fall in, end of story. The other option is of course that he accelerates his head outwards so fast that it rips off, in which case he will indeed 'lose his feet', but that isn't in violation of the equivalence principle either.

If you want to do the math, then use the geodesic equation and calculate the trajectories of objects nearby the horizon. Best,


Patat Je said...

I think I understand now. You need an incredibly powerful jetpack to escape a black hole while losing your feet. The jetpack will be so powerful it will rip you apart anyway. And if you are in an orbit with the perihelion very close to the event horizon you will also be torn apart.

Sabine Hossenfelder said...


Yes, that's one way to think of it. Except that there are no stable orbits close to the event horizon. Look it up somewhere if you don't believe me. There's something called the innermost stable circular orbit which is at something like sqrt(27)/2 times the Schwarzschild radius, ie way out. (I might misremember the exact number, but it's something between 2 and 3). Best,


Patat Je said...

I meant elliptical orbits. I looked it up. Unstable circular orbits for a radius between 3GM and 6GM. Stable above 6GM. So below 3GM the geodesic always crosses the event horizon at 2GM and goes into the black hole! So you always need a jetpack if you want to escape from the space below 3GM.

But okay. I was wrong. I thought I had discovered something new!