Carl Sagan once famously wrote:  “We are made of star-stuff.” It’s certainly poetic, but it  doesn’t tell the whole story.

For one thing, it ignores the fact that some elements don’t have to be made in stars at all. Your body is roughly 10% hydrogen, and the bulk of that element was made moments after the Big Bang, millions and millions of years  before the first stars were born.

But “We are made of 90-ish percent  star stuff” doesn’t sound as nice. That’s not what this episode is about, though. Because the other issue I have with this saying is that it ignores just how complicated it is to make some of the other elements that are inside you.

And some elements that are technically  inside of you in trace amounts, like gold, are so hard to explain the origins of, that scientists don’t actually  know how it all got there. [♪ INTRO] Let’s pull out our handy dandy Periodic Table. Up here we’ve got your hydrogen and helium, which mostly came from the loose soup of subatomic particles that dominated the very early universe. The next lightest elements–lithium,  beryllium, and boron– can also be made without stars getting involved.

But for everything heavier than  that, we start to need those stellar engines, along with some  even more intense processes. Let’s start with the one that scientists  figured out first: nuclear fusion. In this super cool process, which  is actually super hot, atomic nuclei run into each other at such wildly  high speeds they smoosh together.

Fusion is why stars shine, and  they keep doing it for millions if not trillions of years, depending  on how massive they are. Now technically, there’s more than  one kind of stellar fusion process, producing different kinds of elements. And which one happens depends on what kind of star you have and  where it is in its life cycle.

But as impressive as it all sounds,  fusion has its element-creation limit. Eventually, you get less energy  out of making a heavier element than you have to put in, which grinds  that hot stellar furnace to a halt. In even the most massive stars, the  heaviest element it can reliably make is iron, which it does only at  the very last moment of its life.

You can get some nickel at the very,  very end of the fusion process, but it’s radioactive, and  doesn’t stick around very long. But if you grab your closest  copy of the periodic table, you’ll see there are way more elements  on it that are heavier than iron. Including some of humanity’s  favorites, like gold and platinum.

They’re not made by the Big  Bang, or by stellar fusion, but a secret third thing called neutron capture. There are also secret fourth, fifth, even sixth things to making elements, but we’ll have to leave them for another time. This process starts with a little seed nucleus made up of just a few protons and neutrons.

Then, if that nucleus finds  itself inside an environment where there’s a truly enormous amount of neutrons packed almost impossibly close together,  something wild can happen. Neutrons have neutral charges. So, unlike their proton  siblings, they aren’t repelled by each other’s electric charges,  or by those protons, either.

So if there’s a ton of them, they can  all start to glom onto that nucleus. But those neutrons aren’t stable.  Eventually, they decay into protons, plus some other subatomic particles  that don’t matter right now. And since it’s the number  of protons that determines what element you are, all those new protons turn the nucleus into a new, heavier, element.

Those neutrons don’t all decay at once, though. And all the while, even more neutrons outside the ball are avoiding the FOMO  and glomming on as well. That rate of glomming can be either slower or faster than the rate the neutrons are decaying.

Which is why this neutron capture  process comes in two main flavors: The first that astronomers  identified is called the s-process, which stands for slow process. Slow here meaning that new neutrons stick to their seed nucleus slower than  the nucleus forms protons. This version of neutron capture is known to happen in certain low-mass red giant  stars, called AGB stars… and, to a lesser extent, massive old stars.

But according to scientists,  if you look at the universe, and measure all the different  elements that could be made by the s-process, there’s too much matter  for the s-process to account for. So, there must be another  process making these elements.

Enter: the rapid process, or r-process. Now, the neutrons are being added to the seed nucleus way faster than  they can decay into protons. The r-process can only happen  in environments where the neutron density is way higher  than where you find the s-process. And those environments are much rarer in space, meaning that it was way harder to prove that the r-process was even real.

For a long time, scientists thought  the first evidence would come from supernovas: the violent  death of a massive star, or an equally violent outburst  from the core of a long dead star that accidentally siphoned too  much gas off a stellar neighbor. But upon closer examination, no dice. Supernovas don’t seem able  to facilitate the r-process.

We were back to square one. Until, that is, another idea that had been lurking in the background  stepped into the spotlight. In 1974, two astronomers  published a paper that analyzed the nuclear physics of black  hole-neutron star mergers.

Neutron stars are the remnants of stars that were massive enough to go supernova, but not so massive their cores completely collapsed into black holes. They’re chock full of neutrons packed as closely as the laws of physics allow. In theory, the short range gravitational  forces and neutron abundances combine to create the perfect  environment for the r-process.

And on further investigations, mergers between two neutron stars seemed to work, too. But basically no one believed the theory, because they were all so focused on supernovas. Eventually, thanks to modelling  software getting better and better, it became obvious that it is really hard to make the r-process work in supernovas.

There just aren’t enough neutrons present. And some even smaller particles, called neutrinos, tend to convert the neutrons that are there into protons before they can ball  up around their seed nuclei. But the merger scenario not only stayed possible, it started to seem pretty likely.

And eventually, scientists got the piece of  evidence they needed to confirm it. But before I can reveal what that evidence was, we’ve got to keep the lights on. So here’s an ad.

Thanks to JMP for supporting this SciShow video! JMP is a statistical analysis software designed to help you with all your analytics needs. And in the world of analytics,  there can be a lot of needs.

Like when you have so many variables that you need JMP’s predictor screening to  highlight the most meaningful ones and avoid wasting your  energy on the other ones. Or when you need JMP’s  Principal Components Analysis and Partial Least Squares  regression to figure out which of those variables have the  biggest impact on your dataset. From screening to transformations, JMP has you covered through your  whole data analysis process.

You can check them out with a  30-day free trial for anyone, anywhere at jmp.com/scishow. In 2017, the gravitational  wave detectors LIGO and VIRGO both captured an event known as GW170817. But researchers didn’t know  what caused that event.

At least not right away. It could have been just another  merger between two black holes. But upon further investigation,  it turned out to be the merger of two neutron stars.

Because when they smashed into each other, they didn’t just create a bunch of  gravitational ripples in space time. Unlike a pair of black holes, they also created an explosion called a kilonova. Kilonovas are less bright than supernovas, but they’re no less impactful.  Especially in this case.

Because if mergers like this really were somewhere the r-process could happen, the  radioactive decay that creates all those heavier elements would  also create a big ol’ kilonova. So, in order to confirm the  r-process is happening somewhere, you can watch for a kilonova. And after scientists detected GW170817, they found kilonova AT2017gfo coming from the exact same part of the sky.

Not with gravitational wave detectors, though. A kilonova emits a bunch of light,  so it was picked up by regular telescopes, like the VLT in  Chile and the Hubble up in space. It was a great example of  astronomers around the world, with different areas of  expertise, working together.

With this pair of astronomical signals in hand, astronomers were finally able to confirm the r-process happens in neutron star mergers. We finally had a source for  all those heavy elements. Including, if you saw any of the news  coverage, that shiny, glittery gold.

But there’s a problem. After years  of investigations, scientists still haven’t found the chemical signature  of gold in that kilonova light. They’ve found the tell-tale signs of other r-process elements, like tungsten, but not gold.

That doesn’t necessarily mean it’s not there. One study published in 2021  calculated there could be up to 3000 Earth’s worth of gold made  by just that one explosion, without our telescopes picking it up. P.

S. that’s not the amount  of gold on Earth times 3000. That’s the entire mass of the Earth times 3000. Roughly 18 octillion kilograms.

And that’s a lot by our puny human standards, but a cosmic speck by the universe’s. And binary neutron star  mergers seem to be pretty rare. And if you extrapolate, there’s  not enough gold being produced by these mergers to explain the total  amount of gold in the universe.

So scientists are back at the  drawing board, once again. Although this time they didn’t  have to erase the whole thing. They just added a second, smaller drawing board beside the original one.

For a while, they thought about attributing some r-process element production to supernovas. You know, again. But just like  before, that didn’t pan out.

Luckily astronomers have a few  alternates under consideration. Maybe mergers used to happen  more than they do now. Maybe the s-process is actually doing more work in pumping out heavier elements than we thought.

Or, maybe it’s magnetars, which are fancy, ultra-magnetic neutron stars. Back in 2004, astronomers detected a magnetar flare called SGR 1806–20. It was so bright, telescopes couldn’t even look at it directly to make some measurements.

Astronomers had to rely on  reflections off the Moon! And two decades later, a team decided to take another look at this monster. They had originally been digging  into a bunch of old telescope data, looking for evidence of the r-process.

Just a year before, a few members  on the team had discovered that magnetar flares, under the right  conditions, could trigger the r-process. And when they stumbled across this super bright magnetar flare, they  realized it met the criteria. Now, getting the r-process to happen in a magnetar is a bit more  complicated than in a merger.

The outer layers of a magnetar, which is where the r-process would have to happen,  is pretty rich in neutrons. Just not rich enough to do  all the element-synthesis we need it to do in the regular way. But according to this new  study, published in 2025, it can get around that deficit.

As I mentioned earlier, the  r-process requires two basic parts: a ton of spare neutrons, and some seed nuclei for those neutrons to glom onto. And in the case of magnetars,  the emphasis is on “spare.” The low-ish density and incredible  speed of a magnetar flare makes it impossible for too  many “seed nuclei” to form. But even though there’s not technically enough neutrons to kickstart the  r-process the normal way, there are still WAY more neutrons  than there are seed nuclei.

Because of that neutron supply and  demand, the r-process can still create big balls of neutrons that get  ejected out into the universe with the rest of the flare, and  eventually decay into heavy elements. The study estimates that over  the lifetime of our Milky Way, magnetar flares could be responsible for anywhere from 1–10% of all r-process elements. But according to models,  which admittedly were only based on this one event,  magnetar flares are only really able to produce elements on  the lighter end of the heavy element spectrum: like strontium and zirconium.

In other words, no gold. So, where all of the universe’s  gold comes from remains a mystery. Ditto with many of the other  elements far beyond iron of the Periodic Table, including ones that  are inside our bodies right now.

But it’s kinda understandable, given how extreme and infrequent the process  seems to be to make them. Our observational pool for  these sources is so small. While there are other kilonova  candidates besides the one associated with GW170817, none  has actually been confirmed.

And SGR 1806–20 was a once-in-a-lifetime magnetar flare in terms of scale. We still have a lot of work  to do to understand where exactly everything in the universe comes from. But for now, we can still  say we’re made of star stuff.

We’re just also made of a little more than that. [♪ OUTRO]