When the universe was about nine billion years old, on the eve of the formation of our own solar system, after spending ages in calm complacency slowly spinning the cosmic web from loose threads of galaxies and dark matter, a crisis erupted as a hidden force began to wrest control of cosmological fate from the hands of gravity.

In the 1990s on Earth, the geopolitical cold war had finally ended with the collapse of the Soviet Union, but growing tensions within the astronomical community were now running hot. The stakes were never so high. The ultimate future of the universe hung in the balance. Well, to be fair, the universe is going to do whatever the universe was going to do anyway, but we'd like to understand it before it actually, you know, happens.

Cosmology, the science of the universe, is a game of taking almost stupidly simple questions and asking them to the whole entire cosmos as a single physical object. But, as we've seen, simple doesn't mean easy, and straightforward questions like “Hey, how much does that galaxy over there weigh?” can turn out to have some frustratingly surprising answers. So astronomers, in fine academic form, instead of taking the difficulties of measuring the mass of objects like galaxies and clusters as a proper warning from nature to back off, ramped up the observations and attempted the same feat on cosmological scales. Because science.

And immediately they ran into problems.

To explain the problems, I need to talk about geometry. It turns out Kepler was sort of right hundreds of years ago—the universe is ruled by geometry that can predict the future of the universe, just not divine geometry that can predict who my true love is. So, half right. Not so bad.

The language of the cosmos at the very largest scales is general relativity, and as I'm totally sure you remember from a few chapters ago, general relativity connects stuff (matter, energy, etc.) to geometry (the bending, warping, and flexing of space-time). This language can be applied to small systems like planets orbiting stars and black holes, and it can be applied to large systems, like the whole flipping universe. Seriously, like Einstein himself did shortly after inventing the concept. The equations of general relativity aren't picky (insert “general” pun here); to solve cosmological problems, the recipe is simple.

Tally up all the matter and energy contents of the universe. Be sure to include any nonvisible matter, radiation, and neutrinos. Everything. Stuff that into one side of Einstein's equations. Solve said equations (this is the messy part, but lucky for you, we get to skip over it in this narrative). General relativity then tells you how the whole entire universe bends, warps, and flexes.

Now what? Well, again with your handy general relativity tool kit, you take that knowledge of bending, warping, and flexing, and you can predict how it will eventually evolve. Voilà: the future fate of the universe in your hot little hands. We know that the universe is expanding, thanks to Hubble, but just how fast is it expanding? Is it speeding up or slowing down? Will it stop and turn around? Will it coast to a stop? If we fast-forward a hundred billion years from now, what will the universe look like?

The answers to these questions depend on how much of what kind of stuff the universe contains. Heaps of matter will drive a different expansion rate than heaps of radiation, and both of those will certainly result in a different fate than just a thin sprinkling of matter. So to get some clarity on the issue, it's absolutely essential to get a complete and total census of the material population of the cosmos.1

But wait, there's more. You can flip the equations of general relativity around. If you know—or can measure—the geometry of the universe, you turn the relativistic crank and learn the contents. All it takes is being able to measure the geometry of the universe, which, as you might imagine, could be a little challenging. Go ahead and make a bet now on which is harder: measuring the contents or measuring the geometry.

When I say geometry, I don't just mean circles and triangles; the geometry you learned in high school and promptly forgot until you had to paint your living room is only a subset of the full picture. That's the realm of what's called Euclidean geometry, which as you might have guessed was developed by Euclid himself, more than two thousand years ago. And that picture is all about flatness: circles, triangles, angles, lines, and everything all living in a plain, flat universe.

But general relativity is all about curvature—so, step number one, how do you define “curvature”? You have an intuition that the Earth's surface is curved, but your own backyard is relatively flat. How can we define curvature in such a way that we can apply it to any system that we care about, and specifically to the system that we care about the most: the universe?

Here's one way: parallel lines. If you draw two parallel lines on a piece of paper, and extend those lines to infinity, the lines will stay parallel. That's kind of the definition of “parallel,” so I hope that's not a big issue for you. But try drawing some parallel lines on a globe. Start at the equator with two tiny lines, perfectly parallel. Advance each line northward in a perfectly straightforward way, never turning left or right. Soon enough, despite your best efforts, the lines will intersect—at the north pole.

You guessed it: the globe, and the Earth, is curved.

Another method is triangles. If you draw a triangle on a flat piece of paper, the interior angles will add up to 180 degrees. That's also the definition of a triangle, so we should be good. Now draw a triangle on a globe. Add up the interior angles. You'll get a number larger than 180, I guarantee it. Don't trust me? Do it yourself, right now. I dare you.

If you happen to have a horse saddle lying around (I won't judge), you can play the same games, but you'll find the opposite result: initially, parallel lines will spread farther apart, and triangle angles will fall short of 180 degrees.

The beauty of these definitions of curvature is that they can be applied to any number of dimensions—most important for us, three: the spatial dimensions of our universe.

That's great! Easy-peasy, we just need to bust out the markers and see how triangles and parallel lines on billion-light-year scales behave. Thankfully, nature provided the measuring tools already: the cosmic microwave background. We know, based on our knowledge of the nuclear realm, about the sizes and scales of the minute bumps and wiggles in that afterglow light pattern—we know how big they were when they were formed. And we know how big they are now, eons later, because they're right there, projected onto our sky.

Beams of light make for excellent cosmic Sharpies. Two beams of light, initially parallel and given enough distance, will eventually trace out the geometry of the universe on the grandest of scales. Sure, little things like galaxies (and yes, in a cosmological sense entire galaxies are considered “little”) will distort their paths, but we want to paint a much bigger picture. In the intervening billions of years between generation and acquisition of those light beams, they will either remain parallel, spread apart, or converge. By comparing what we know about the typical sizes of patches in the cosmic microwave background to what we actually measure, we can literally measure the geometry of the entire universe.

It's flat.

Like a pancake. Like Kansas. Like a board. The universe is flat. Parallel lines stay parallel over the course of billions of years and billions of light-years. Our universe appears totally geometrically flat to a ridiculously small margin of error—we're talking one part in a million here, folks.2

When this answer—that we live in a flat universe—first started to become apparent in the 1990s with the first high-resolution maps of the cosmic microwave background, it was totally fine for the theorists. Why, they had been crowing about a flat universe for decades now, and they were glad the observers were finally catching up. Their reasoning came from inflation, that poorly understood but seemingly necessary process in the very early universe, when the entire cosmos blew up many orders of magnitude in the blink of an eye.

The process of inflation sent the true scale of the universe far, far, beyond our relatively pathetically small observable bubble. The universe could have any curvature it liked, like the surface of the Earth or a saddle, but it wouldn't matter. Our observable patch is so small that we're essentially guaranteed to measure a flat universe. Just like the Earth is round, but your backyard is so small it appears flat (insert usual caveats about small-scale deviations like groundhog holes not mattering for our measurements of the bigger picture).

Inflation went one better. To measure a certain degree of flatness today, with a giant universe, meant that the universe had to be even more flatter when it was smaller: the dilution of matter should have driven us far, far away from perfect flatness long ago. Inflation solves that problem by making the true universe so gigantic that we can't help but measure flatness. Nice.

If the inflation story is correct—and even though we don't know the characters, we understand the general plotline—then we must live in a (locally) flat universe. We don't have any other choice. So it's no surprise that the cosmic microwave background reveals precisely that answer.

But nobody else, most especially the astronomers, bought into that argument. Hence the mild disagreements of the 1990s.

The problem was that the measurement of flatness was simultaneously also a measurement of the total contents of the universe, due to all that general relativity stuff you just read about a minute ago. And the astronomers were already busy running around weighing all the galaxies and clusters they could train their telescopes on, and they were falling far short of the predicted number.

Like, less than a third. The total mass of everything in the universe, including even dark matter, was less than 30 percent of the matter necessary to match up with the observations of a flat cosmos.3

Hence, argument. The theorists accused the observers of being lazy and not working hard enough in their measurements, because surely they were missing something. The observers countered that theorists should take a break from the chalkboards for once and actually look at the universe around them that was abundantly not agreeing with their predictions—perhaps they forgot to carry the two in one of their calculations? And the cosmic background is pretty far away, even for astronomy. You sure you got it right?

You know, the usual stuff.

In comes a third approach, also tied to general relativity: the expansion history. Matter and energy tell space-time how to bend, and the bending of space-time tells matter how to move. So you can measure geometry, contents, or behavior to get at the underlying physics. And the behavior in this case is, of course, the expansion of the universe.

What Hubble managed to measure back in the early twentieth century was the local expansion rate, based on a relatively small sample of nearby galaxies. And in cosmology, “local” is synonymous with “today.” The light from those galaxies didn't take that long to get here—it might as well have been from just last week, cosmologically speaking. So Hubble's measurement, and any other measurement based on close galaxies, is the current expansion rate of the universe right now at this very moment.

But the universe could have had different expansion rates in the past. Indeed, the entire concept of inflation depends on that ludicrous-speed expansion in its early moments. And with radiation and then matter taking center stage at different times with different densities, that can affect how quickly (or, for that matter, slowly) the universe expands.

To get these measurements we have to go deep into the universe, pushing further with surveys and observations than we ever have before. Redshift by redshift, galaxy by galaxy, we need to reconstruct the expansion history of the universe. Of course there will be limitations—the dark ages won't have much to offer us, at least not yet—but galaxies arrived on the scene pretty early in cosmic history. Surely there's some method to capture their distance and velocity, and to combine that with the data from as many other galaxies as possible.

That's the key; that will resolve the tension, one way or another, between the camps of scientists who have spent decades refining their techniques and sharpening their arguments but haven't come to a solution. It's the Kepler-Brahe and Shapley-Curtis debates all over again, centuries later.

To reach tall heights you need a ladder, and to pierce deeper into the celestial realm than our ancestors would have ever dreamed possible, you need…

…a cosmic ladder. No, that's not a joke. I mean, I totally set it up as a joke, but “cosmic distance ladder” is a real-life phrase used by real-life scientists to refer to a real-life concept. The challenge with measuring truly astronomical distances is that objects can be very far away even for an astronomer.

We've already met parallax, which after centuries of frustration finally led to an accurate distance to another star. Less than a century later, Edwin Hubble used Cepheid variable stars to confirm the remoteness of Andromeda. If you recall (which, by the way, is an academic code phrase meaning “you should have recalled”), Hubble couldn't use parallax for his groundbreaking work because parallax just wasn't good enough: the distances were too great, and the back-and-forth wiggles were too small to measure.

But even Cepheids, as useful as they are, can only be used to capture distances to nearby galaxies. While that technique swaps out a very difficult measurement (a distance) for a relatively easier one (brightness variations), you still need enough raw light to get the job done. If the Cepheid is too far away, you simply don't have enough photons to work with, and you're back to being hopeless.

So we need something else, some other way to hook into the distance of an object without actually having to measure the distance. And preferably we need some of these to overlap with at least a few known Cepheids. We could use Cepheids to reach farther than parallax because we had a few of those variable stars near enough that we could practically break out the astronomical measuring tape and pin them down. Once the method was safely confirmed, we could extend the Cepheids into the unknown depths without losing too much sleep over the issue.

So parallax and Cepheids are the first two rungs of a ladder, where we (hopefully) use a series of different objects and techniques to (hopefully) take us deep into the cosmic depths, each more distant method overlapping with a closer, vetted technique. A ladder that lets us measure cosmic distances. A cosmic distance ladder. See, I told you it was a real thing in real life.

Say you have two light bulbs of identical make, manufacture, wattage, color, and so on. Keep one bulb next to you, and throw the other one as hard as you can—or better yet, to continue with the experiment, walk it across the room and gently place it down. Turn on both lights. Break out your brightness-o-meter (I know you have one). Measure the brightness of the bulb next to you. Measure the brightness of the bulb far away from you.

Hopefully the distant bulb will be dimmer, and if you think a little bit about the relationship between dimness and distance, you could reasonably argue that if we had a bunch of such light bulbs scattered around the universe, we could use their relative dimnesses to calculate distances.

Unfortunately, flying around placing light bulbs at strategic cosmic locations would render the whole distance-measurement game moot, so we need nature to manufacture some for us. Many folks back in the day, like Newton and Galileo, had hoped/assumed that stars were of equal true brightness, allowing them to be used like our light bulb analogy, but sadly, that didn't work out well for anybody.

If stars don't cut it, what can? Whatever this standard candle could be, it has to be (a) common, so we can get a lot of measurements, (b) very bright, so we can see it reliably from far away, and (c) actually standard, so we can compare the measured to the true brightness and get work done.

In almost all circumstances, nature is sneakily cruel to us, offering only tantalizing hints of the mysteries of the cosmos, usually not enough to slake our thirst. But when it comes to cosmological questions, we have caught a couple of lucky breaks (it's still up for debate if nature willingly intended this or if they were just lucky accidents that slipped by her usually vigilant gaze). One case was the cosmic microwave background, which confirmed the big bang picture and unlocked the modern age of cosmology, and the other has to do with supernovae.

Stars are born, stars live, and stars die. Most stars, like our sun, simply exhaust themselves and extinguish their nuclear flames with only a modest flare or two to signify the ends of their lives. Some stars, the big ones, mark the end of their short, furious lives with explosions that literally light up entire galaxies—the supernovae.

Astronomers had been sporadically spotting new stars for millennia, and it was Tycho Brahe who named them novae when he extensively wrote about one that appeared in his sky one evening. By the late 1800s, it was realized that some novae were much brighter than others. Superlatively so, and the word supernova came into fashion.

As is the usual custom in science, the more a phenomenon like migration patterns or beetle larvae is studied, the more divisions, classifications, and subgroups get attached. It's no different with (super)novae, which, thanks in part to the work of our dear friend Fritz Zwicky, are helpfully organized into several different classifications with unhelpful naming schemes.4

Honestly, we don't need to care about the naming schemes and their origins and definitions. I know three of you are now going to throw away this book in disgust because I'm not going into detail on the subject, but for the rest of us, we're just going to pay attention to one particular kind of supernova: the type Ia. We're going to care about it the most because that's the one that cosmologists care about the most, and this is a book on cosmology.

In the 1990s astronomers noted, probably because they were looking for something exactly like this, that type Ia supernovae have a few desirable properties. While these kinds of supernovae are relatively rare, popping up a handful of times per century per galaxy, there are so many galaxies in the universe that they're basically happening all the time. They are exceedingly bright too: for a few weeks, a single supernova detonation will release more energy than an entire galaxy. That's a few hundred billion stars, for those keeping score. This means that we can see type Ia supernovae all over the place, even from incredibly distant galaxies.

So they're common and they're bright. Two out of three criteria met—but what about the third? You know, the most important one: do they have the same brightness? No, they don't. Well, almost, kind of.

When a supernova goes off, it quickly reaches a peak brightness, then slowly fades down over the course of a few days or weeks. Different type Ia's have different peak brightnesses, and they also have different cooldown times, which at first blush isn't surprising: supernovae are going to do whatever they dang well feel like. At first this might look hopeless as a standard candle, but the ever-ingenious astronomers noticed a pattern: the brighter a supernova reached at its peak, the longer it took to cool off. And not in a general, vague, hand-wavey sort of way; there was a very clear mathematical relationship between those two quantities.

That was the key that opened the door to the distant universe. Type Ia supernovae aren't standard candles, but they are standardizable candles. With a little bit of finagling,5 you can capture a random supernova in some remote stretch of the far-flung cosmos and calculate its true, inherent brightness, the same brightness you would measure if you were right there in front of it, having your face melted by the blast. You can then compare the true brightness to the brightness you measure safely on Earth, and calculate a distance.

Even at a distance of fifty million light-years, this type Ia supernova still manages to dazzle, briefly expending more energy than its entire host galaxy. (Image courtesy of NASA / ESA.)

But wait, there's more! The magnificent explosions are cacophonies of light, featuring the usual assortment of absorption and emission spectral lines that astronomers of the nineteenth century learned to love. A reliable spectral line means you can get a redshift, which means a velocity (and if we can't use the spectrum from the supernova itself, we can always use the one from its host galaxy). This is the exact same technique that was unlocked more than a hundred years prior, was utilized to great effect to discover the motion of stars, and decades later revealed the expanding universe to Hubble's amazed eyes. And now it was brought to its ultimate conclusion, its final form: studying distant dying stars to prize out one last datum of cosmological significance from the fading embers of those cataclysmic explosions. Their stellar sacrifice is our gain, thanks to a practice developed in the steam age.

Two key pieces of info for the price of one (difficult) measurement: an unlikely gift from nature. If a type Ia supernova goes off in a distant galaxy, you can capture not only its distance but its speed. So not only do you get to extend the cosmic distance ladder to new celestial heights, you can also get a handle on the expansion rate between us and a distant point. And by figuring out the expansion rate at different points in deep time, you can tease out the matter and energy contents of the cosmos and figure out why all the other astronomers are hating on each other.

Of course, there's way more to the distance ladder—it's more than three rungs. There are more steps in between Cepheids and supernovae with fantastic names like RR Lyrae variables and the Tully-Fisher relation. Those aren't as useful for cosmological purposes directly, but they do serve as important cross-checks: each rung on the ladder isn't entirely separate, but overlaps with others, so we have multiple independent ways of measuring and calibrating distances. For the craziness I'm about to share with you, it's important to mention that part, because what cosmologists have recently revealed about our universe is easy to dismiss (that's how weird it is) if you don't know how robust the measurement really is.

I'm also purposely keeping you in the dark (for now) about the actual physics of type Ia supernovae, because as I said with Cepheids, it doesn't matter. What matters for the ultimate goal—measuring the expansion history of the universe—is that we can identify a standard candle. Who cares how the candle is made? All we need to know is that we can estimate its true brightness to a certain level of accuracy. Let's not make our lives messier than need be. Words of wisdom from the cosmological community indeed.

So here's the game plan: measure a bunch of type Ia supernovae (only type Ia behave in this friendly way, so unfortunately we have to stick to them). Measure a bunch more. Toss a few more in for good statistics. Calculate their distance and speed. Estimate the expansion history of the universe over the past few billion years. Use general relativity to calculate what the matter and energy contents have been over those past few billion years. Resolve dispute among astronomers and cosmologists. Sit back and relax.

Two independent teams in the 1990s went about exactly that, and their goal was to measure the deceleration, however slight, of the universe's expansion. That's because any amount of matter, however slight, would eventually pull in the reins on our current expansionary phase. It may not be enough to fully stop it—that requires a lot more matter than we could ever find—but the presence of matter in the universe slows down expansion over the course of billions of years. So by carefully pinning down that deceleration, they could tell everybody else what's out there in the cosmos.

The two teams, working independently, managed to measure a deceleration—but with a minus sign in front. The distant supernovae that they collected were too dim. They were farther away than they ought to be. It was 1998, and after checking their work over and over again, and a few furtive phone calls between the two groups (“Are—are you seeing what we're seeing?”), the world came to know that the expansion of our universe is accelerating.6

The collaborations also quietly scribbled out the word “deceleration” from their project subtitles.

We live in an expanding universe; it's getting bigger and bigger every day. We've had a century to become grudgingly comfortable with the concept. But now, at this very moment, the expansion is accelerating. It's getting bigger and bigger faster and faster every day. To really drive the point home, the universe is expanding faster than if it were totally empty of all matter. Like finding your favorite comfort food at the buffet, there's no going back.

In a replay of the Shapley-Curtis debates from generations past, the supernova results were the ultimate compromise. They satisfied everybody by satisfying nobody. Everyone was right because they were all wrong. The answers clicked into place in a way that nobody enjoyed. You go into the doctor with a broken finger and a nosebleed, and you come out with your amputated finger sewn onto your nostril. You're technically healed, but not in the most pleasant or straightforward way.

The universe is indeed flat—parallel lines stay parallel, and triangles add up to an agreeable 180 degrees. The inflation theorists and cosmic microwave background observers were vindicated. And the universe is only about 30 percent matter. The astronomers had done a fine job accounting for all the mass after all. What made up the difference, what filled up the universe from edge to edge, was a previously unknown substance (or, at least, an effect). This substance is causing the expansion of the universe to accelerate unbridled.

We had a good thing going with dark matter, sounding all cool and mysterious and a little sassy, so why not extend the concept? We have absolutely no clue what's behind this accelerated expansion, so let's call it—drumroll, please—dark energy.

Welcome to the modern universe. It doesn't make any sense.

What is dark energy? There isn't much to say on the subject, because if you didn't catch it the first time, we have no idea what it is. “Dark energy” is a sweet name for an observed phenomenon—the accelerated expansion of the universe—but that doesn't really illuminate (snicker) the cause of that expansion. It's worse than our current problems with dark matter; at least there we have some ways of navigating the theories and putting experiments online. With dark energy, even twenty-plus years after the initial detection, we're still in the groping-around phase.

But let's paint the picture, hazy as it is, as we've got it right now. We can start with the reason we call it “energy,” and it's not because “matter” was already taken and there are only two kinds of things in the universe (and really, if you want to get pedantic—and who doesn't?—matter and energy are two sides of the same coin, but that's another book).

It's your birthday and I get you a present, because I'm pretending to be your friend for the purposes of this explanation. You undo the intricate bow and carefully unwrap the package, opening the box to find…nothing. Completely empty. A pure vacuum, in fact. No matter, no radiation. Not a single particle, not a single photon. Absolutely pure nothing.

You politely mask your disappointment and thank me, internally rolling your eyes that I'm using a special event like your birthday for another stupid science thought experiment. You sigh, preparing yourself for the inevitable monologue. Here it comes.

Your empty box is anything but—it's actually full of dark energy. Dark energy fills the cosmos to the brim. It's a property of the vacuum of space itself. Have a small empty box? You have a little bit of dark energy. Big empty box? A lot more dark energy. Since we live in an expanding universe, we're getting more and more vacuum, more empty space, and hence more dark energy every day.

Now it just so happens that when you take a material with this property—that you get more of it when you expand the volume of its container—the math of general relativity produces a surprising result, and you can probably guess what it is: you get accelerated expansion. The very fact that dark energy is a property of the vacuum means that it drives the continued creation of itself. The universe has dark energy. It expands. It has more dark energy. This pushes the expansion a little bit faster. Even bigger universe. Even more dark energy. Even more accelerated expansion. Rinse and repeat if desired.

And before you interrupt me, I'll interrupt myself: you may be familiar with the concept of conservation of energy (usually incanted in some fashion like “Energy cannot be created or destroyed, it can only change forms”), and you're probably asking yourself why this first pass at explaining dark energy doesn't violate rule 1 of How Things Work. The answer is very simple and very annoying: energy isn't always conserved. It's conserved in a few special systems, especially systems found in homework problems of physics textbooks, but nature is much more subtle than that. In a changing universe (like, say, one that's expanding), energy can be added at will. It's just not a big deal, and we're all going to have to get comfortable with that if we're going to make any progress, OK?

Anyway, this accelerated expansion business didn't get started in earnest until about five billion years ago. It's not like dark energy just poofed into existence by some poorly understood process. Dark energy has always been here. It's a property of the vacuum of space-time. It's there, right in front of your face and inside your very bones. It's been with the universe since the earliest moments of the big bang itself. But it's been hidden, in the background, unimportant.

It's a game of densities: the same game that's been played for billions of years in our expanding universe. Matter finally won out over radiation, eventually causing the release of the cosmic microwave background, because the expansion diluted the radiation and dropped its density well below thresholds where anybody would notice or care. And while matter had its eventual—and inevitable—triumph, its reign was not forever. There's only a fixed amount of particles—dark or otherwise—available in the universe. Day by day, cubic meter by cubic meter, the density of matter has been dropping.

But dark energy's distinguishing feature is constant density. The bigger your universe, the more total dark energy there is. While matter is dominating, its gravitational attraction overwhelms any inclination dark energy might have to accelerate the cosmos. Indeed, the expansion of the universe slowed down over the course of the first few billion years. But like too little butter spread over too much bread, matter lost its grip on the fate of the cosmos. It's simply not a player anymore. Starting five billion years ago, the density of matter slipped below that of dark energy and continued falling. Dark energy dominates the modern-day universe. It's the single largest component. It's in the driver's seat now, and it doesn't even know where the brake pedal is.

I think it's important to remind you that you can sleep soundly tonight—dark energy only makes itself noticeable on large scales. Clusters are still gravitationally bound together. Galaxies safely keep their stars within their embrace. The planets still twirl in their waltzes around their suns. Dark energy is persistent but weak—any place where another force is stronger, the accelerated expansion can't take hold. Just as an expanding universe doesn't mean the Earth is expanding, an accelerated expansion only takes place in cosmological settings.

Unfortunately for the cosmic web, it can't sleep soundly tonight. Beyond the scale of already-formed clusters, there's not enough attractional oomph to overwhelm the tearing tendency of dark energy. Like a swimmer caught in the riptide, getting farther away from the shores no matter how hard she fights.

It's happening in the voids first. They're already empty of most matter, and it's there that dark energy is most dominant, expanding and inflating them, pushing their boundaries ever larger. The tenuous cosmic web exists now only as gossamer strands and almost-transparent walls between the indomitable nothingness of the voids. Eventually those grand structures will dissolve. As the cosmic web becomes unspun, only the isolated clusters, mere pockets of gravitational attraction strong enough to stand fast, will remain, becoming remote islands separated by vast black wastelands ruled by one thing and one thing only: dark energy.

Pretty dark stuff (pun most definitely intended). I'll get to the future fate of the cosmos in more gruesome detail later, but for now I want to clear up a few loose ends. First off, this business of energy. Hmmm, what else fills up the vacuum, permeating all of space-time within the entire universe? Good question—I'm glad you asked. Quantum fields do! The buzzing and vibrating substrate has all the needed properties to explain dark energy. And it is an energy too—a vacuum energy. At the ground state, the bare minimum allowable configuration, the quantum fields that constitute our reality have some base energy, and they are perhaps ultimately responsible for the accelerated expansion of the universe itself.

We've broken open the quantum field In Case of Emergencies box before, to explain the driving force behind inflation and to seed the growth of the largest structures in the universe, so at first blush, this seems like a natural extension of those strategies. Have a mystery in the universe? Just blame quantum fields—look, look, they're practically oozing guilt on their faces. It's easy. Everyone will believe you.

One small, tiny, niggling caveat: it doesn't work. I have a confession to make. When physicists go to calculate the actual amount of vacuum energy—like, actually produce a number instead of just talking about it vaguely—they get infinity. That's right: infinity. It turns out that all the wiggling and jiggling those quantum fields do at a subatomic level keeps adding and adding to itself without end. The fundamental (ha!) problem is that the wiggles can be as small as they want. It's like trying to listen to an orchestra with an infinite number of instruments that can play at every frequency imaginable. All the sounds add up to an infinite amount of energy and your ears bleed and your head explodes.

That's…kind of a problem, which we saw earlier in the saga of physicists first trying to extend the work of Dirac and marry special relativity to quantum mechanics. Fortunately for most calculations of importance, the absolute value of this vacuum energy doesn't matter. I know it's weird and nonintuitive to think about, and I'm truly sorry about that, but that's physics for you. You can do your particle science on the first floor or the tenth floor or the infiniteenth floor; it doesn't matter.

Except when it matters, as in the case with dark energy. We know very well (by now, at least) how much vacuum energy, if it is the culprit here, is driving the accelerated expansion. It's a very tiny and definitely not infinite number: somewhere in the ballpark of 10⁻³⁰ grams in every cubic centimeter of the cosmos—ten measly hydrogen atoms per cubic meter, give or take, spread across the entire universe. That's all it takes to give us the spectacular and stupefying accelerating-cosmos results.

When we attempt a naïve prediction of this number based on quantum field theory by, say, only adding up frequencies to some reasonable threshold like, I dunno, the Planck scale or whatever, out pops a number clocking in at 10⁹⁰ grams per cubic centimeter. That's wrong. Very wrong. Deeply, uncomfortably, you're-obviously-missing-something-fundamental-about-the-universe wrong.

So maybe quantum fields aren't the best path forward to explaining dark energy. But even if we can't get the number right (or even close to kind of right), dark energy still acts as if it were a vacuum energy. Even if the true source is something else, accelerated expansion has the behavior—just unfortunately not the magnitude—of space-time-filling quantum fields permeating all of the cosmos.

Basically, we're stuck.

There's another way to frame this problem in general relativity, which might or might not be helpful, depending on your level of pessimism. This framing can be sourced directly to old Einstein himself. Remember how, when he first manufactured relativity in the general sense, he didn't know about the expanding universe? And how the equations gave him the flexibility to toss in an extra constant to maintain a static cosmos? And then as Hubble astonished the world, he scrubbed out the so-called cosmological constant when nobody was looking?

Well, here we are again, finding ourselves in a situation where it looks like we need to add that constant back in after all. Not to maintain a static universe against any movement, but to power the accelerated expansion. The reason this might help is that it's an alternative way to formulate the conundrum, which might lead somewhere promising, since a cosmological constant sits on the “curved space-time” side of the relativity fence. The downside is that a cosmological constant is perfectly identical to a vacuum energy, which sits on the “matter and energy” side of the equations of general relativity. So you haven't actually moved anywhere.

Still, could dark energy really be a fault of gravity rather than some vague, poorly understood part of our universe? It's the same question we had with dark matter, and it has the same problems. As nauseating as vacuum energy is, there appears to be (at the time of this writing) no modification and/or extension of general relativity that competently explains dark energy.7

Part of the problem is that we're in the dark (ha!) not just theoretically but observationally. It's been a couple of decades since dark energy first burst onto the scene in scores of supernova detonations seen across the universe. Astronomers are not idle folks, and in those years they've collected heaps of evidence and come to the conclusion that yup, it's still there.

Take, for instance, those baryon acoustic oscillations, the fancy-pants term for the rockin’ sound waves in the early universe. When recombination smoothed the troubled waters of the young cosmos, those waves got frozen in, leading to a slight—but detectable—impression in the arrangement of galaxies on very (very) large scales. That impression acts as a standard ruler, as if you tossed a bunch of yardsticks (metersticks in the international print of this book) around the cosmos. Since we know the size of the oscillations, we can compare that to the size we measure in the sky, and reconstruct the expansion history at those points.

It's a completely independent way of targeting dark energy, and it reveals…dark energy. Indeed, if dark energy hadn't happened to our universe, galaxies would have continued their cosmological building program, erecting ever-larger structures, and that process would have washed out any primordial imprints altogether. The very fact that we can still detect that faint impression leads us to conclude that dark energy is happening.

Despite decades of our best observational efforts, we've only been able to, at best, confirm the existence of dark energy, without actually learning a lot more about it. Has it changed with time, or is it truly a cosmological constant? Is it connected at all to dark matter? Are there multiple sources of dark energy? If it's a vacuum energy, why can't we get the numbers to come out right? Is dark energy related to inflation, which was driven by another (mysterious) quantum field, and if so, why are these effects separated by billions of years and orders of magnitude in energy?

All questions, no answers.

Our first dart thrown at the dark energy board is miles off, but we don't have any other decent (even half-decent; heck, we'd even accept 1 percent decent at this stage) ideas.

The modern picture of our universe, as painted by hundreds of observations and experiments independently searching for answers and cross-checking each other, from the cosmic microwave background to distant supernovae to the weights of clusters of galaxies, is cold and bleak: 13.8 billion years old, composed of less than 5 percent normal (light-loving) matter. One-quarter dark matter and three-quarters dark energy. It's geometrically flat, but the expansion is accelerating, for reasons we don't fathom.

We call it “concordance cosmology,” as it's the result of many different lines of research that all point to the same bleak conclusions.8 It's a completely different universe from the one explored hundreds of years ago. That universe was complicated and messy, but small and hot. Cozy and alive. The universe revealed in the modern age is old and slow, well past its prime and dominated by mysteries piled on mysteries. It turns out that the efforts of generations of scientists over the course of centuries have barely even scraped the cosmological surface.