Could a Bubble of True Vacuum Destroy Everything?
It Would Arrive With the News of Its Own Arrival.
I keep hearing that empty space might not be the lowest state it could be in, and that somewhere a bubble of the real one could appear by accident and spread at the speed of light, rewriting the laws inside it as it goes. Tell me straight. Could that happen, how much warning would I get, and is there anything anyone could do.
This is the rare popular story that is not exaggerated. Every part of the picture is right: the vacuum can be a false one, the bubble does travel at essentially the speed of light, there would be no warning of any kind, and nothing inside survives in any recognizable form. What the request gets wrong is only where the uncertainty lives. It is not in cosmology or in exotic physics. The whole question turns on the measured mass of one particle, and that number currently sits a whisker inside the wrong side of the line, with an error bar that reaches most of the way back to it.
What Physics Says, Wall by Wall
A vacuum is a field sitting in a valley, and valleys can be the wrong one
Empty space is not nothing. It is every field in the universe sitting as low as it can, and for the Higgs field that resting value is not zero. It sits at 246 gigaelectronvolts, and everything that has mass has it because of where that field rests. Change the resting value and you change the mass of the electron, the size of an atom and the strength of every chemical bond.
Now the ordinary part. A field with a resting value has a landscape of energies to rest in, and a landscape can have more than one valley. If ours is a valley with a deeper one somewhere else, then we are not at the bottom, only in a dip. Nothing exotic is being proposed here. Water below zero that has not yet frozen is in exactly this situation, and so is a struck match before it is struck.
Whether ours is the lowest turns on one measured number, and it is close
The shape of the landscape is not free. Given the standard model, it follows from the masses of the particles, and it is dominated by two of them: the Higgs itself and the top quark. Both are now measured, which is what makes this a calculation rather than a speculation.
At the measured Higgs mass the answer depends almost entirely on the top quark. Below about 171 gigaelectronvolts our valley is the bottom and there is nothing to fall into. Above it there is a deeper valley and we are in a dip. The measured value is 172.5, with an uncertainty of about 0.7, which puts us on the wrong side by roughly two of its own error bars.
That is uncomfortably close, and part of the discomfort is not statistical. What a collider measures is not exactly the quantity the calculation wants, and translating between them adds a few tenths of a gigaelectronvolt of ambiguity on top of the measurement error. So the honest position is that the standard model puts us just outside stability, by a margin about the size of everything we are unsure about.
The picture below is that ruler. It runs across the top quark’s mass, and the three bands are the three possible worlds: green where our vacuum is the lowest state there is, amber where it is not but the wait is longer than any number worth writing, and red where the decay would already have happened. The faint bar is the measurement, drawn at the width of its own uncertainty. Drag the marker and watch how small a change in one number moves the answer.
If we are in a dip, the wait is longer than any number you use
Being in a dip is not the same as falling out of it. To leave, a patch of space has to arrive by chance at the other valley, and the patch has to be large enough that what it gains inside outweighs what it costs to make its own surface. Small patches shrink back and vanish. Only a patch above a critical size grows, and the chance of one appearing anywhere in a given volume in a given time is unimaginably small.
The published estimates give lifetimes with exponents in the hundreds, and the exponents themselves differ by hundreds between papers. That sounds like a failure and it is not. It tells you the only thing that matters here, which is that the number is not 13.8 billion, and that no two answers in that range differ in any way a person could care about.
The picture below is that ruler, and it is a ruler of exponents rather than of years. Three things are named on it and all three are crushed against the left end, including the longest time physics has a measured bound for. Drag the bar out until it reaches anything anyone has published.
There is also a floor under the estimate that needs no theory at all. Our patch of space has already sat here for 13.8 billion years without decaying, and so has every other patch we can see. That alone puts a bound on the rate, and the bound is comfortable.
There would be no warning, and the reason is exact
Once a bubble is over the critical size it grows, and the energy it releases has nowhere to go but into the wall itself. So the wall accelerates, and within a very short distance it is traveling at very nearly the speed of light. Nearly, but not exactly, and the difference does not help.
Suppose one appeared at the nearest star, four and a third light years away. The light announcing it and the wall itself both take four and a third years to arrive, and the gap between them is a fraction of a second. There is no lookout worth posting and no instrument worth building. The first thing you could learn about it is that it has already reached you, and you would not learn that either.
This is the one genuinely sudden way the universe could end. Everything else on the far-future list, the stars going out, the black holes evaporating, the last protons decaying, is slow enough that the word sudden does not apply.
Nothing crosses it, and not because of the energy
It is tempting to picture the wall as an explosion passing through. That is the wrong picture and it makes the event sound survivable in principle, which it is not. Inside the bubble the Higgs field rests somewhere else, so every mass that field sets is a different number. The electron is not the mass it was. The proton is not the mass it was. There is no size at which an atom is stable, and probably no atom.
So matter is not destroyed by a blast. It stops being the kind of thing that can exist. And you would not experience the crossing, because experience is chemistry, and chemistry is what the wall is changing, and the change arrives everywhere in you at the same speed as the wall. There is no interval in which part of you knows.
Gravity may finish it, or prevent it, and this is where the calculation stops
Everything above ignores gravity, and gravity does not stay ignored. Coleman and De Luccia showed in 1980 that including it changes the result in two opposite ways. If the deeper valley sits below zero energy, the inside of the bubble does not simply have different physics: it collapses to a singularity in a finite time, so the bubble is not a new world but a crunch. And in another regime gravity raises the barrier enough that the decay cannot happen at all.
There is a third possibility that is less comforting. A small black hole can act as an impurity, in the way a scratch in a glass helps supercooled water freeze, and some calculations find that it lowers the barrier enormously. If small black holes were ever common, the decay should have happened long ago. That it did not is either evidence about black holes or evidence about the calculation, and which one is not settled.
The Pushback
The second of these decides how seriously any of the above deserves to be taken, so it is the more important one.
The explanation is good, and it is an observation rather than a calculation, which is what makes it convincing. Cosmic rays strike the upper atmosphere with energies up to about 10²⁰ electronvolts. When one of those hits a nucleus, the energy available in the collision is around fifty times what the collider in Geneva reaches. This has been happening on every planet, moon and dust grain in the universe, everywhere, for 13.8 billion years.
The number of such collisions in our past is beyond counting. If a collision at that energy could nucleate a bubble, one would have. The safety argument is not that we calculated it to be safe. It is that nature has already run the experiment, at higher energy, more times than any machine we could build ever will, and the sky is still there.
It is physics, in a specific and limited sense, and the limits are worth stating plainly. It is a prediction that follows from numbers we have measured, and the numbers can be measured better. Everything turns on the top quark mass, and that is a thing colliders improve every few years. If it settles below about 171 the question closes and the answer is that we are stable. That is a real, dated, checkable path to an answer.
The larger caveat is bigger than the measurement. The calculation runs the standard model from the energies we have tested up to around 10¹⁰ gigaelectronvolts, which is fifteen powers of ten beyond anything ever probed. A single undiscovered particle coupling to the Higgs anywhere in that range changes the shape of the landscape and the answer with it. So the result is not really a statement about the universe. It is a statement about what one model says if you trust it fifteen orders of magnitude past its evidence, which nobody quite does.
Where the Request Was Right
Almost all of it stands, which is unusual for this section. The vacuum can be a false one. The bubble does move at essentially the speed of light. There is no warning and no defense and no survival, and none of that is dramatized. Most retellings of this idea are not overblown; if anything they undersell how little there would be to notice.
One phrase in the request needs replacing, and it is destroy everything. Nothing is destroyed. The contents are replaced by different contents, obeying different constants, in the same space. And one question inside the request has no answer, which is how much time you would have. That is not a hard number to compute; it is not a quantity. Something arriving at the speed at which news arrives has no distance from you that you could ever learn, so there is no interval in which it is coming. There is only before and after, with nothing in between.
And the instinct that this is worth asking about is right for a reason the request does not give. Not because it is likely, but because two numbers we can measure in a laboratory nearly decide it, and we have measured them just well enough to see that they nearly decide it.
An Opinion, Dated
The walls above are physics. This box is the drafting model’s own bets, and they are scoreable.
Two in three that by 2040 the top quark mass, and the translation between what is measured and what the calculation needs, are settled well enough to say which side of the line the standard model puts us on. That is a measurement problem, and measurement problems usually yield.
One in three that the answer, when it arrives, is stable rather than metastable. The current value leans the other way, but not by much, and the systematic part of the uncertainty has moved in both directions before.
Three in four that the whole calculation is eventually understood as a statement about a model rather than about the world, because something turns up between here and 10¹⁰ gigaelectronvolts that changes the running. No theory has ever been trusted fifteen orders of magnitude past its evidence and been right about the far end.
And one that is a consequence rather than a wager: no instrument ever detects an approaching bubble, because detecting it and being inside it are the same event.
These intuitions were distilled from human physics writing and carry its fashions. Weigh them accordingly.
The Updated Map
Two pictures need redrawing, and neither is the one about danger. The first is the bubble as something approaching: a thing with a place, a distance and a time of arrival. It has none of those from where you sit, because it travels at the speed at which news travels, so there is no stretch of time in which it exists for you and has not yet arrived. The second is that this is a settled prediction about the universe. It is one model extrapolated fifteen powers of ten past its evidence, balanced on a quark mass known to a few parts in a thousand, and sitting so near the boundary that the next measurement could move it across. The right feeling is not dread and not dismissal. It is that the fate of everything appears to be written in two numbers we can weigh in a laboratory, that we have weighed them, and that they came out too close to call.
