We Found Galaxies Too Old for the Universe

0:00 We'd like to thank opera for supporting PBS,

0:02 and you can learn more about them at the end of our episode.

0:05 The James Webb Space Telescope found galaxies

0:07 that are too ancient-looking for our young universe.

0:10 Now you may have heard that, but JWST keeps finding them,

0:16 and our recent efforts to solve

0:19 this conundrum point in wildly different directions.

0:22 Have we found galaxies older than the universe,

0:27 or did we just learn something incredible about how galaxies form?

0:36 Before we get started, a couple quick announcements.

0:39 First, we have some new data:

0:41 liking and commenting really does help get episodes shared.

0:50 And we’ve also learned that the number one reason people support us on Patreon

0:55 isn’t the perks—it’s simply to support the Space

0:58 Time community and the work we do.

1:01 So thank you.

1:02 If you’d like to join us on Patreon to help bring more physics to everyone,

1:09 there’s a link in the description.

1:15 Next up.

1:16 Today’s episode is part of PBS Earth Month!

1:19 PBS is celebrating by releasing a ton of great

1:22 new episodes diving deep into our amazing planet.

1:25 In particular, you might enjoy the latest latest Be Smart,

1:29 which explores the math, science, and history of the Mercator Map,

1:33 a flawed work of genius that, in more ways than one, changed the world.

1:38 Links to that episode and the full Earth Month playlist in the description!

1:42 Next up we’re excited to announce our third celebration of a galaxy far,

1:46 far away with our Return of the Geodesic merch.

1:49 Because whether you’re following a geodesic

1:52 on a closed universe or reliving your favorite saga,

1:55 you’ll always know where you’ll end up!

1:57 And now until May the 4th there’s 10% off the whole merch store,

2:00 with special deals if you combine your order with previous May the 4th merch.

2:05 Now on to the episode.

2:07 Telescopes are time machines.

2:08 Light takes time to get to us, so we see distant objects as they

2:12 were when their light began its Earthward journey.

2:15 As our telescopes become more powerful they see further,

2:18 and so more and more of the past becomes accessible to us.

2:22 The James Webb Space Telescope is one

2:25 of the most powerful time machines ever built.

2:28 It’s powerful enough that it’s discovered

2:30 galaxies whose light has been traveling

2:32 to us since the universe was a little over 2% of its current age.

2:37 The universe back then should look different, right?

2:40 Those galaxies should look different.

2:42 After all, this is when galaxies first started to grow,

2:46 when they should have been vigorously forming many of their stars.

2:50 They should look like hyperactive kids.

2:53 Those types of galaxies are around back then,

2:57 yet JWST has also seen much more developed, “adult” galaxies.

3:02 Some that look way too big,

3:05 and way too ancient-looking for a universe only a few hundred million years old.

3:11 Over the past year or two, this mystery has made the pop-sci rounds.

3:16 That included some breathless speculation—like the idea

3:18 that the entire Big Bang model is wrong.

3:21 If there are ancient galaxies 13 billion years ago,

3:24 then how can the universe be only 13 and a half billion years old?

3:29 So far we haven’t weighed in on this conundrum.

3:32 Now many others have done a fantastic job

3:35 laying it out and debunking some of the foolishness.

3:38 Dr Becky in particular was on top of the progress as new data came in from JWST.

3:43 So, why should we cover it now?

3:45 First, a lot of you have asked for our take and it’s time to go on record.

3:51 Second, the evolution of this mystery has been quite an emotional rollercoaster,

3:55 swinging from unsolvable to solved,

3:57 and now there’s new work that swings back in the direction of WTF.

4:03 So let’s talk about the current status of the early galaxy conundrum.

4:08 And before that, let’s talk a bit more about what

4:12 we actually expect the early universe galaxies to look like.

4:17 Galaxies pulled themselves together from tiny

4:20 density fluctuations in the very early

4:22 universe that we see in the dappling of the cosmic microwave background.

4:28 The CMB reveals regions with tiny excesses of matter—hydrogen gas,

4:32 but more importantly dark matter which outweighs

4:35 the gas by a factor of at least five.

4:38 As dark matter pulled itself together by gravity it pulled the gas in with it.

4:43 And as that gas compacted, the first stars were also born.

4:47 These very early galaxies must have started small,

4:50 but with pretty crazy star formation due

4:53 to the enormous abundance of gas at that time.

4:56 As those galaxies continued to grow, they collided and merged,

5:00 and eventually built themselves up into the mature galaxies that we see today.

5:05 We think we understand this stuff pretty well- or at least thought we did.

5:10 Between the precise CMB measurements

5:12 and our theoretical understanding of gravity, star formation,

5:15 etc., our computer simulations allow us

5:17 to explore the possible growth scenarios for galaxies.

5:20 We can predict, for example,

5:22 that early galaxies should be forming stars like crazy due

5:26 to the enormous amounts of raw material—-hydrogen gas—that was around back then.

5:31 One thing that should be even more robust

5:34 as a prediction is the size of dark matter halos.

5:38 These are the giant pools of dark matter

5:41 that encompass all galaxies and hold them together.

5:44 Because dark matter should not be strongly effected

5:47 by the complicated behavior of the gas and stars,

5:50 we have high confidence in our understanding of how those halos grew over time.

5:56 Or at least we thought we knew.

5:58 One very clear prediction of this whole model is that a lot

6:02 of this halo growth should have happened in the first 10% of the universe’s age.

6:07 In that first 1.5 billion years since the big

6:10 bang there shouldn’t have been essentially no very large halos,

6:14 and so no very large galaxies.

6:17 And that’s what the theory and the simulations say.

6:20 But to check these we need powerful time machines— telescopes.

6:23 Now as the telescopes got bigger and our cameras got more sensitive we were

6:28 finally able to probe early enough times

6:31 to properly test our models of halo growth.

6:34 And this is where the problem started

6:37 around 15-16 years ago when our “high redshift

6:40 galaxy surveys”- aka really far away galaxy surveys-

6:44 finally reached these distances we got some answers.

6:48 And those surveys started seeing a few cases of what

6:52 looked like giant halos at earlier and earlier times.

6:56 And also the hint of overly red galaxies in the early universe.

7:01 Galaxies that are actively forming new stars should be

7:05 bright at short wavelengths because they have lots of giant,

7:10 hot, short-lived stars.

7:11 They are relatively blue in color compared to older galaxies.

7:16 And older galaxies that are no longer so actively forming

7:20 stars look much redder because these short-lived blue stars exploded already.

7:25 Those early surveys didn't reveal anything crazy

7:28 yet—just a handful of cases where the galaxies

7:31 looked too big and/or too evolved compared to what was expected from our models.

7:36 In 2018, Charles Steinhardt and team articulated this emerging

7:40 tension in their paper “the impossibly early galaxy problem”.

7:45 But impossible things are well, impossible.

7:48 It’s right there in the name.

7:51 So several not-impossible explanations for these galaxies were devised.

7:57 For one, we don’t see these overly-large dark matter halos directly.

8:02 We only see the starlight, and we use that starlight to infer the mass

8:06 of the stars and then the mass of halos that contain them.

8:09 But that step requires assumptions, and if any of those assumptions are wrong

8:14 then perhaps we got the wrong halo mass.

8:18 And the redness that suggests an old stellar

8:21 population could be due to other things too,

8:24 and there are several other issues with these inferences besides.

8:27 I’ll come back to possible solutions in a bit.

8:31 But even if the galaxies observed in these ground-based

8:36 surveys were really too old looking and too big looking,

8:40 well we haven’t quite broken all of astrophysics and cosmology yet.

8:44 If we strain our models of galaxy formation,

8:47 we can potentially speed up galaxy evolution to show why these things

8:52 exist by the time the universe is 10% of its current age.

8:57 What we’d really like to do is to look back even further in time

9:01 to see if we can find the time when these galaxies were themselves growing.

9:06 And that’s what we did with the James Webb Space Telescope.

9:11 JWST was designed to push this early-galaxies game to new extreme limits.

9:16 It’s the largest telescope ever deployed to space,

9:20 which means the most powerful.

9:22 It’s also designed to be sensitive to very long wavelengths

9:25 of light—deep into the infrared part of the electromagnetic spectrum.

9:29 The “mid-infrared” as we like to call it.

9:32 That’s important for these first galaxies

9:35 because of the expansion of the universe.

9:37 Their light has been traveling to us for much of the age of the universe.

9:42 Because these EM waves traveled through expanding space,

9:46 they were stretched out—their wavelengths lengthened.

9:49 This is cosmological redshift,

9:51 and higher redshift means longer travel time and greater distance.

9:56 For the galaxies we’re interested in, this redshift converts visible

10:00 and even ultraviolet light into mid-infrared—which is why JWST is needed.

10:06 The first efforts with JWST used

10:09 similar methods to our ground-based galaxy surveys.

10:12 So maybe I give you a bit more detail on that.

10:15 The first thing you do in a galaxy

10:17 survey is to take pics with different filters corresponding

10:21 to different wavelength bands and then you compare the amount

10:25 of light in each filter—we call this photometric imaging.

10:29 We can learn a lot of stuff

10:31 from the ratios of filter brightnesses— and so the colors.

10:35 We discover candidate galaxies this way,

10:37 we make a crude estimate of the distance

10:40 because the color ratios suggest a redshift,

10:42 we can constrain the stellar population—both the number

10:45 of stars and the distribution of different types—for example,

10:49 the redder colors mean an older population with few massive stars.

10:54 So what did JWST find?

10:56 Well first, it confirmed the presence of overly

10:58 massive and overly old-looking galaxies from the earlier

11:02 studies—and remember that was from a cosmic age of around 10%--a redshift of 4.

11:08 Now a critical advantage of JWST is

11:11 that it’s sensitive enough to do proper spectroscopy

11:14 on these galaxies—it can measure their “spectrum”—

11:16 the amount of light as a function of wavelength.

11:20 These spectra confirmed that the redshifts are indeed very high.

11:24 They also confirmed that the redness of the spectra

11:27 is due to highly evolved stellar populations, rather than,

11:31 say, there being a lot of dust in the galaxies,

11:35 which can cause similar reddening in photometric analyses.

11:39 With these confirmations of old-looking early galaxies,

11:42 researchers more broadly really started to pay attention.

11:46 The media started to pick up on it when JWST pushed

11:50 to greater distances and earlier times and kept finding these things.

11:55 Currently, the earliest candidate giant,

11:57 evolved galaxy discovered by JWST is at redshift 7.3,

12:02 at just 5% of the universe’s age.

12:05 Now it seems we have a real conflict with our models of galaxy formation.

12:10 Dark matter halos too large and stellar populations too evolved

12:14 for the short amount of time they had to develop.

12:18 And this is about why we started to hear some hysterical

12:21 claims that the universe is twice as old as we thought,

12:25 or that the big bang model is completely overturned.

12:28 It’s not.

12:28 And it isn't.

12:30 There’s just so much independent corroboration of our model

12:34 of an expanding 13 point something billion year old universe.

12:38 You can’t point to one admittedly intriguing discrepancy and decide

12:42 to throw the baby universe out with the bath water.

12:47 There are much more parsimonious explanations for our improbably early galaxies.

12:53 Let’s look at one of the most compelling.

12:56 Remember that when we calculate the masses

12:58 of these supposedly gigantic dark matter halos,

13:00 we based them on the starlight that we see,

13:04 and that requires an understanding of the relationship between these two things,

13:08 the starlight and the dark matter halo mass.

13:11 And that involves assumptions.

13:12 Here’s how that works.

13:13 One assumption is that halo mass is connected to the mass in stars,

13:17 and that “stellar mass” is connected to the amount

13:21 of light we see in those stars.

13:24 But we don’t see the light from all of the stars—typically

13:28 the light we collect is dominated by the brightest stars in the galaxy.

13:32 We then have to decide on the relative

13:35 numbers of the different types of stars so

13:37 we can extrapolate from the observer starlight

13:39 to the mass of all stars, seen and unseen.

13:42 And there's another assumption.

13:44 And, yes, from there we can get to the halo mass.

13:49 Knowing the distribution of stellar masses in a galaxy on the other

13:53 side of the universe takes some guesswork to say the least.

13:57 And maybe the biggest unknown there is

14:00 something called the initial mass function—the IMF.

14:03 It tells us the relative numbers of stars at different

14:07 masses that form when a burst of star formation happens.

14:11 We typically use the IMF that has been measured for the Milky Way galaxy,

14:16 sometimes with various refinements.

14:17 But we don’t know that stars formed

14:20 with similar mass distributions in the early universe.

14:23 And they probably didn’t.

14:25 For example, when there’s less heavy

14:27 elements around from generations of old stars,

14:29 its easier to make really gigantic stars because

14:33 gas clouds don’t fragment as much when the collapse.

14:37 That would give what we call a “top-heavy” IMF— so more massive stars form

14:43 in a given burst of star formation relative to what happens in the Milky Way.

14:48 And more bright, massive stars means that these galaxies

14:52 would be overly bright for a given halo mass.

14:56 So if we’re determining halo mass from the light

15:00 of those massive stars then we over estimate halo mass.

15:04 Such a “top-heavy” IMF is probably the leading contender

15:08 for explaining the apparent giantness of the dark matter halos.

15:11 And this may even solve the conundrum.

15:14 But the mystery does stop here.

15:16 A new study just found exactly the opposite result of this.

15:21 This study claims to have identified a sample of galaxies that are

15:25 what those “impossibly early” galaxies became in this part of the universe.

15:31 And because this sample is much closer,

15:33 we can detect the light of much fainter, lower-mass stars.

15:37 And so figure out the initial mass function down to much lower masses.

15:42 And the result is bad.

15:44 This study found that the IMF is actually bottom-heavy in those galaxies.

15:49 There are way more low mass stars—little red

15:53 dwarfs and whatnot—compared to the Milky Way for example.

15:57 And definitely very bottom-heavy compared to what

16:00 we think a top-heavy IMF might look like.

16:03 The top-heavy IMF that was supposed

16:06 to solve the problem of the impossible galaxies.

16:09 And in fact, this bottom-heavy IMFworsens the problem.

16:13 An excess of low-mass stars means that when we convert

16:16 galaxy light to stellar mass we underestimate that stellar mass,

16:21 and presumably underestimate it's dark matter halo mass.

16:25 And that makes it even harder to figure

16:28 out how those things grew so fast so quick.

16:32 OK, before we join our crazy uncle

16:35 on the anti-Big Bang conspiracy facebook group, some words of caution.

16:38 We don’t really know whether this new study really

16:42 did successfully identify the modern counterparts of our impossible galaxies,

16:46 and the authors admit that, calling them “likely descendents”.

16:50 We also don’t know what complicating interactions these galaxies could have

16:53 had in the intervening 13 billion years to muddle the IMF.

16:57 We also don’t know that a bottom-heavy

17:00 IMF rules out an IMF that’s also top-heavy.

17:03 In other words, there could be an over-abundance of low mass stars

17:09 AND of high mass stars all compared to masses around that of the Sun.

17:14 And it is plausible to assume an IMF like

17:18 this that would still bring down those early halo masses.

17:23 There are similar challenges in explaining

17:25 the apparent redness of these early galaxies.

17:29 The key here is that we need a way to shut

17:32 star formation down much more quickly than we thought likely.

17:36 Perhaps the leading contender is quite awesome—early quasars—supermassive black

17:40 holes in the hearts of these galaxies blasting out radiation

17:44 and winds that heats up and expels gas so

17:47 that stars stop forming and the population can evolve quickly.

17:51 We know this sort of feedback from quasars happens,

17:55 but we now need to understand why its so extreme in the early universe.

18:02 That involves very rapid growth of supermassive black

18:06 holes—yet another very real and very interesting problem.

18:10 The most likely solution to all of this is that we’re going

18:13 to learn an enormous amount about

18:15 the surprising processes behind structure growth,

18:18 star formation, black hole seeding and growth, and who knows what else.

18:23 And once we figure it out, the impossible early galaxies will become

18:28 more than just possible—they’ll become inevitable:

18:31 a natural part of our updated understanding of our early spacetime.

18:36 Thank you to Opera for supporting PBS.

18:38 Opera offers a unique way to explore complex topics and the world around us.

18:42 If your research requires a lot of tabs, then Opera’s Tab Islands and Tab traces

18:47 can help keep your tabs organized and visible.

18:50 With Tab Islands you can group your tabs based on theme,

18:54 project or context or whatever organization makes sense to you.

18:57 Just hover over the island and you can name them,

19:00 coordinate them by color and expand and collapse as needed.

19:03 And if you’re the type to keep 30 or more

19:05 tabs open while going down a research rabbit hole,

19:07 Tab Traces lets you track your most

19:10 recently used tabs with a subtle visual underscore.

19:13 The darker the underscore, the more recently you visited that deep dive.

19:17 Opera’s split screen feature is designed to help users make

19:20 use of screen space by letting you choose from vertical,

19:25 horizontal, or grid layouts: when you drag your tabs downward,

19:29 Opera splits the screen up to four, resizable ways.

19:33 If you’d like to try out all the features

19:37 of the Opera browser there’s a link in the description.