We know that the universe is getting  bigger.

And we know that the speed that   the universe is getting bigger is also  getting bigger.

The standard assumption   is that the cosmic acceleration rate is itself is at least constant, which will surely result in   ultimate heat death of the universe.

But a recent survey of primordial sound waves frozen into the way   galaxies are sprinkled through the universe  reveals that this fate is now in question.

Back in 1998 we discovered that the expansion  of the universe is accelerating.

This rates as   one of the most important findings in the  past several decades in physics.

That’s   because dark energy—our placeholder name for  whatever causes this acceleration—is utterly   mysterious and may be the clue we need to move  beyond our current understanding of physics.

To be clear, we do understand the math of  how dark energy works—just add a constant   energy density to the fabric of space and  you get accelerating expansion, represented   by the cosmological constant in Einstein’s  general theory of relativity.

These days,   almost all of cosmology assumes something called  the Lambda-CDM model—CDM for cold dark matter,   not the topic of this video, and  Lambda for the cosmological constant.

We even have a prime candidate for this  sort of unchanging dark energy.

We expect   fluctuations in the quantum fields to  produce such a constant energy density,   resulting in an “energy of the vacuum” that  behaves exactly like a cosmological constant.

Problem is, with our current understanding  of the quantum fields in our universe,   we would expect those quantum fluctuations to most naturally do one of two things: to cancel out to exactly zero,   or pile up to an enormous, universe-destroying  value.

It’s very, very difficult to see   how those fluctuations should produce  such a tiny but non-zero vacuum energy   without extreme fine-tuning  of the field parameters.

There are proposals for how this  fine-tuning may have happened—for example,   we could be an extremely rare universe with a  habitable value for the vacuum energy among a   multitude of desolate ones.

But if you  don’t like these anthropic arguments,   we could also reject the idea that dark energy  is vacuum energy at all.

For example it might   be an undiscovered quantum field.

One thing  that a new quantum field might be expected   to do is to change in strength over time, which  would lead to a changing rate of acceleration.

So how would we look for a changing acceleration  rate?

Detecting that the expansion was   accelerating at all required us to map the  changing size of the universe back in time.

This was done by determining the distances  to a type of supernova—type 1a supernovae,   or exploding white dwarf stars.

This approach  gave us the expansion history over the most   recent half of the age of the universe.

That’s  enough to show the expansion rate is changing,   but not enough to see whether the  change in that rate is itself changing.

We also have a data point at the beginning  of time with the cosmic microwave background   radiation.

If you compare the expansion  rates predicted by these two methods   assuming a constant acceleration, you  get a mismatch.

The CMB measure gives an   expansion rate smaller than the supernova  measure.

This is the now-famous crisis in   cosmology.

Many believe that there’s a  systematic error in these measurements   somewhere.

But another possibility is that the  assumption of constant acceleration is wrong.

It’s starting to sound like it would be  a good idea to properly map the expansion   history back through as much of cosmic time  as possible with a single method.

Then the   acceleration history should be clear, and  with it the constancy or non-constancy of   dark energy.

Unfortunately we can’t see  supernovae at large enough distances,   and the CMB is only at the  furthest distance by definition.

But we do have another option: Baryon Acoustic  Oscillations—BAOs.

BAOs are like standard rulers   that stretch all the way back to the beginning  of the universe.

We talked about them before,   but let me give you a refresher.

In the few  hundred thousand years right after the big bang,   the entire universe was filled with hot plasma  and no stars or even atoms had formed.

This plasma   flowed into the regions of higher gravity that  would later collapse into clusters of galaxies.

As these massive inflows crashed together,  vast sound waves reverberated back out,   creating expanding shells of higher density  material.

These sound waves could travel   because matter was locked together with the  intense radiation that filled the universe.

But as the universe also expanded and cooled,  and at some point the first atoms formed; matter   and light were no longer locked together and so  our sound waves stopped moving.

The pattern of   those waves became frozen into the distribution of  matter at that moment a few hundred thousand years   after the big bang.

Later, galaxies and clusters  would form from this matter.

Most galaxies would   form near the centers of large overdensities—but  there should be much weaker shells of   overdensities surrounding each of these where  the sound waves were frozen.

We see these baryon   acoustic oscillations in the cosmic microwave  background—the light carried to us from the moment   of freezing.

But we also see that there are rings  of galaxies distantly surrounding galaxy clusters.

This is an incredibly faint signal—observable  only as a slightly higher probability of finding   galaxies separated by the diameter that these  rings have today.

That diameter is called the   sound horizon of the BAOs—the maximum distance  sound travelled before freezing—and it’s around   500 million light years after having  expanded with the rest of the universe.

This number is our standard ruler.

We know how big  the sound horizon should be at any point in the   history of the expanding universe.

By measuring  the angular size of this horizon on the sky, we   can find the true physical distance corresponding  to the galaxies on these BAO rings.

With one more   ingredient, that translates to an expansion  history of the universe.

That final ingredient   is redshift, and this is exactly what the  Dark Energy Spectroscopic Instrument measures.

DESI is a pretty amazing instrument and  survey.

It measures redshifts—the amount   that light is stretched as it travels  to us through an expanding universe.

We’ve been able to measure redshifts since the  20’s—that’s how we discovered the expansion of   the universe in the first place.

But DESI is able  to measure redshifts of 5000 galaxies or quasars at a time.

For every patch of the sky, 5000 robots place  as many optical fibers at precise positions on   a plate, corresponding to the precise locations  of galaxies in that patch.

Light from those galaxies   travels the optical fiber down to a spectrograph,  where it’s split into component wavelengths.

Characteristic spikes in emission due to known  elements in each galaxy then reveal the redshifts.

In this way, DESI has been able to measure  redshifts for 26 million galaxies and quasars   so far, and the plan is to get to 40 million by  the end of the survey.

This will ultimately result   in a giant three dimensional map of the nearby  universe over the past 11 billion years.

Combined   with the size of the BAO rings at each redshift,  this also gives us a map of the expansion history.

Back in April of this year, the DESI collaboration  published a series of papers with the results from   the first year of the survey—from those 6 million  redshifts.

One of those papers covered the baryon   acoustic oscillations, in which they determine  constraints on a few cosmological parameters.

One   of these is something called omega, which gives  the so-called “equation of state” of dark energy.

It tells us how much pressure is being produced  for a given energy density.

In Lambda CDM,   we expect omega to equal exactly -1 for all  of time, and the DESI results are in very good   agreement with omega = -1 and constant  dark energy, just as we have always assumed.

That sounds anticlimactic, but then something  odd turned up.

The DESI team combined the sound   horizon measurements from their redshift survey  with sound horizon measurements from the two   other types of survey: from the Planck satellite  measurements of the cosmic microwave background   at the beginning of time, and with a few  supernova surveys of the more modern universe.

Adding these other surveys led to an unexpected  result.

The combined data were better fit with a   model with omega -1 and in which dark energy  is getting weaker over time.

And that’s the   case whether the CMB or supernova results  were added to DESI separately or together.

The statistical significance of this result  depends on which combination of datasets you   look at.

It’s 2.6-sigma for DESI + CMB, and  up to nearly 4-sigma with a particular one   of the supernova samples included.

2.6-sigma  is considered not statistically significant,   although still worthy of curiosity and more  observations.

4-sigma is getting exciting,   even if it’s short of the normal 5-sigma  threshold for a very likely positive detection.

But we should also be quite wary of that 4-sigma  result because it depends on inferring distances   from supernovae.

The BAO sound horizons measured  this way are probably more subject to systematic   errors than are the CMB results.

That’s because  supernova distances depend on a longer chain of   distance measurements, and so there are more  places that some error could have crept in.

OK, let’s allow ourselves to get excited  just for a minute and ask what this means   if it’s true.

One of the more general models for  non-constant dark energy is called “quintessence”,   which has an equation of state that can change  over time.

The DESI paper focuses on a version   of this stuff called thawing quintessence, in  which an initially constant dark energy sort   of defrosts, becoming variable over time.

And  Desi finds this model is more consistent with   the various combined surveys than constant dark  energy.

If dark energy behaves this way then   it’s less likely that it will reverse course  and start to get stronger in the future.

That   may preclude the Big Rip scenario, in which dark  energy gets so powerful that it shreds the fabric   of space time.

So, the universe hopefully  doesn’t end as we described in this episode.

The opposite extreme to the big rip is the big  crunch, in which the expansion eventually slows,   then turns around and the universe collapses  into a reverse big bang—potentially even   kickstarting a new universe.

We don’t  expect that to happen even with no dark   energy whatsoever.

The current expansion rate is  enough to keep the universe expanding forever,   even ignoring dark energy and including the  slow-down from the inward gravitational pull   of all galaxies.

If the density of dark energy is  diminishing and eventually hits zero, that won’t   stop eternal expansion—although it will hold off  the accompanying heat death for a while longer.

To get a big crunch, dark energy would have to  change in a more fundamental way—it would have   to shift to a positive gravitational force  rather than a negative one.

And that’s not   impossible.

If the equation of state parameter  omega also changes, and in particular rises from   the current -1 to -⅓ or higher then dark energy  stops behaving like dark energy at all and could   work to recollapse the universe.

Then we might get  a proper crunchy end to the universe after all.

But perhaps the most exciting outcome of  discovering a non-constant dark energy is   that it may constrain other physics.

We’ve  been a bit stuck this past few decades,   unable to move beyond general relativity and the  standard model of particle physics.

Things like   string theory seemed promising, but have trouble  generating testable predictions.

String theory   is especially challenging because it predicts a  myriad of possible universes called the string   landscape and it seems impossible to determine  how to identify our universe in this space.

Well,   in 2018 a team of theoretical physicists showed  that only a subset of the string landscape should   lead to stable universes, and that a requirement  of that stability of that subset is that dark energy should diminish   over time.

So perhaps we have our first tested  prediction of string theory and string theorists are going to be   pretty excited if the DESI result pans out.

OK, one final word.

I mentioned the crisis in   cosmology—the tension between the CMB versus  supernova measurements of the expansion rate   as measured by the Hubble constant.

This  Hubble tension could be resolved with a   changing acceleration rate.

Unfortunately,  the acceleration rate would have to increase to resolve the tension,   not decrease as DESI is suggesting.

In fact the DESI results do   also infer an expansion rate the hubble constant—and they actually support the Planck CMB result and not the   supernova result.

So it seems we have to be tense  about the Hubble constant for a while longer.

And once again, please remember the this  result has not yet reached the threshold   of statistical significance that we need for  a confident declaration of a discovery.

We’ll   learn more soon though.

DESI has actually  completed its second year of observations,   and its results will come out  faster than the first year results,   so we’ll see if the finding firms up  or evaporates.

And in a few years,   when DESI has observed all 40 million galaxies  and quasars, and we have results from other   current or future programs such as the Dark Energy  Survey and the Vera Rubin observatory, we will   gradually be able to hone in on a more precise  value of the equation of state of dark energy.

And then we should know one way or  another—do we have a cosmological constant   or a cosmological variable.

And perhaps we’ll  finally understand what dark energy actually is,   and have evidence for the underlying  theory of everything that supports this,   and even know the final fate  of our expanding space time.