Subtitles (313)
0:00This is a tiny piece of metal
just three millimeters across.
0:03And here's what happens if
you just keep zooming in:
0:14Each of these blobs is an actual atom.
0:18I saw this the other day at
the University of Sydney,
0:20and it kind of blew my mind
because up until just 30 years ago,
0:24directly seeing atoms like this
was thought to be impossible.
0:29The rooms that you're going to see here are
perhaps the most shielded rooms on campus,
0:34or even in the whole of Sydney, I would say.
0:37And perhaps also the most expensive.
0:41So why is it so hard to see atoms?
0:45Well, you can't actually see
atoms with visible light.
0:48That's because while light has wavelengths
between 380 and 750 nanometers,
0:52an atom is still over 3000 times
smaller, just 0.1 nanometers.
0:57And if the wavelength of light is much
bigger than the thing you're trying to see,
1:00the light will just diffract or bend
around it so you won't be able to see it.
1:06So if you want to see atoms, you need
something with a much, much smaller wavelength.
1:11The best candidate isn't even light.
1:17In 1924, a French physicist named Louis de Broglie
worked out that everything was sort of wavelike.
1:24Not just light, but matter too.
1:26Atoms, molecules, even you
yourself have a wavelength.
1:30And the formula for this
wavelength is Planck's constant,
1:33divided by the object's momentum,
that is, mass times velocity.
1:37So here what you actually see,
1:39that's the column of the microscope where
we accelerate the 300 kV electrons down.
1:44300 kilovolts, these electrons?
1:46So they are relativistic particles.
1:49How fast are they moving?
99% of the speed of light?
1:51Around 80.
80% the speed of light?
1:54Yeah.
So what would be their wavelength?
1:56The wavelength is the Planck
constant over the momentum, right?
1:59So if we calculate that, we come to
around between 2 to 3 picometers.
2:05That's over 100,000 times
smaller than visible light.
2:08So theoretically, you get
100,000 times more resolution.
2:13Shortly after de Broglie's discovery, a
group of scientists in Germany started
2:17working on a microscope that would
use these high-speed electrons.
2:21The only problem is you can't
bend electrons using glass lenses.
2:25So how do you focus them?
2:28Hans Busch, a German physicist, suggested that
an electromagnetic lens might do the trick.
2:33He published his results in 1926,
but never actually built one.
2:38Fortunately, a copy of his paper fell into the
hands of an eager young PhD student, Ernst Ruska.
2:46Ruska built his first prototype by coiling up
2:48some wire and surrounding it with iron —
taking care to leave a gap in the middle.
2:53Then, when he passed a current through the coil,
2:55it induced a donut-shaped magnetic field
through the metal and across this gap.
3:02To test it, Ruska first boiled
electrons off a tungsten filament,
3:06the same kind of filament you'd
find in an incandescent light bulb.
3:09He accelerated these free electrons through a
3:11positively charged anode down
to his electromagnetic lens.
3:15As an electron approaches the lens, the
magnetic field exerts a force on it.
3:20So if an electron is traveling in the y direction
and the magnetic fields are in the x direction,
3:24this force, called the Lorentz
force, pushes it in the z direction.
3:28But as the electron moves this way, it encounters
other magnetic field lines along the donut shape,
3:34which constantly point its motion in a circle.
3:37But then this circular motion means
the Lorentz force starts pushing the
3:40electron inwards as well, spiraling
it into the center of the lens.
3:45Now, if you trace the path of
the whole beam of electrons,
3:48you'll see they all get steered
into the center, focusing the beam.
3:54By 1931, Ruska and his colleague,
Max Knoll, used this kind of design
3:59to build the first working electron microscope.
4:02It was pretty basic, made of brass
roughly bolted together, but it worked.
4:10The image itself was created once the focused
4:12electron beam hit a sample
sitting at the focal point.
4:15The sample needed to be incredibly
thin, only around 100 nanometers thick.
4:20More electrons would make it through
the thinner parts of the sample than
4:23the thicker parts, creating an
electron imprint of the sample.
4:27Then a second electromagnetic lens magnified this
4:29imprint down onto a fluorescent
detector, producing the final image.
4:34This was known as a transmission
electron microscope or TEM.
4:39Now, early versions of the
microscope barely magnified at all.
4:43In fact, it wasn't even better
than an optical microscope.
4:46But Ruska was determined.
4:50Over the next few years, he experimented with
4:52adding more lenses onto the microscope
to create bigger and bigger images.
4:56By the mid-1930s, Ruska had gotten the
TEM way past 10,000 times magnification.
5:02It could produce close-ups of insects, bacteria,
5:05and even viruses at a level far
surpassing the optical microscope.
5:11But right as Ruska's TEM was taking
off, a German physicist named Otto Scherzer
5:16published a paper claiming that the
microscope was about to hit a brick wall.
5:23There was a flaw in the electromagnetic lens,
he wrote, that was completely unavoidable.
5:30For an electron to make it
to the focus of the lens,
5:33it needs to be deflected by a specific amount.
5:36If you simplify its trajectory, you can define
that ideal deflection with this angle theta.
5:42This angle depends on the horizontal distance
5:44of the electron from the optical
axis, and how far down the axis the focus is,
5:49also known as the focal
length.
5:51The shorter the focal length, the stronger the magnification.
5:54If you graph this angle as a function
of distance to the optical axis,
5:58you'll see that it can be approximated as linear.
6:01The problem is that the magnetic
field doesn't scale linearly.
6:05It's much stronger near the edges of the magnet.
6:08So if you plot the curve for the
actual deflection of the electrons,
6:12you'll see that the magnetic field
overdeflects the electrons further out.
6:16Their angles are bigger than they should be,
6:17so they end up focusing
before the rays in the middle.
6:21And as a result, the focus is spread across the
6:23optical axis instead of being
contained in a single point.
6:28The blur starts out around the edges of the image,
but it gets worse the higher the magnification.
6:33This is called spherical aberration, and it
distorts every radially symmetric magnetic lens.
6:40In fact, it doesn't just affect magnetic lenses.
6:44Every spherical lens, from a camera to a telescope
to a magnifying glass also suffers from it.
6:50But there is a surprisingly simple way to
minimize spherical aberration. Just add
6:55a second lens, one that diverges
light instead of converging it.
7:00Now, a diverging lens also
suffers from spherical aberration.
7:04But if it has the same amount of aberration
as your converging lens just in reverse,
7:09you can stack the two to essentially
cancel out their effects.
7:12And that removes the aberrations almost entirely.
7:16Almost all modern lens systems in cameras
7:18and microscopes use some sort
of correcting divergent lens.
7:22So you might imagine that the TEM
simply needs its own version of a
7:26diverging spherical lens to magnify further.
7:29But with magnets, this is physically impossible.
7:33Every magnet has two poles, a North and a
South. It's impossible to just have one.
7:39Even if you split a magnet down the middle,
7:41it creates two smaller magnets,
both with a North and South.
7:45And all magnetic field lines have to start at one
pole and end at the other, forming a closed loop.
7:50It's a direct result of the second
Maxwell equation because the field
7:54that you create has field lines that
start and end at the same magnet.
7:59So the electrons will always
cross through two lines.
8:03The first time it passes through, by the Lorentz
force, it's brought into the spiraling motion.
8:08And then the second time
from that spiraling motion,
8:11which has then a slightly different
direction, it's pushed towards the axis.
8:16That's why all electromagnetic lenses by default
will converge that beam, and never diverge it.
8:24Even if you shot electrons in from the other
side of the lens, they would still get focused.
8:30This is what Otto Scherzer's paper proved
in 1936, stopping progress on the TEM.
8:35It is impossible to produce a radially
symmetric magnetic lens that diverges.
8:41And this was, of course, a big roadblock
for the development of electron microscopy,
8:45because people saw, okay, we can
accelerate electrons as much as we want,
8:50the presence of spherical aberration
will always be in the way.
8:55Because of this roadblock, advancements in the
microscope's resolution slowed significantly.
9:00By 1955, another microscope
beat the TEM to the punch and
9:04took the first generally accepted image of atoms.
9:09This was called the field ion microscope,
9:11and it worked by shooting helium or neon
atoms at an atomically sharp needle tip.
9:16The tip was positively charged. So
when the gas atoms hit the needle,
9:20they got ionized and were ejected
off perpendicular to the surface.
9:24And that could form an impression
of the atomic structure of the tip.
9:29But this method was limited.
9:31You could only get a sense of the atomic
structure of the very tip of the needle.
9:36And the images weren't all that impressive.
9:39Luckily, Ruska's electron microscope
9:41wouldn't stay stuck in the realm
of insects and bacteria forever.
9:46Now, you might not be an insect getting
bombarded with relativistic electrons,
9:50but it can sometimes feel
like it when you're getting
9:53bombarded with spam calls and targeted ads.
9:56It's a real problem when we're
researching for our videos.
9:59I was reading up on lenses and optics,
10:01and a few days later I started getting
targeted ads for glasses and eye surgery.
10:06So someone out there is probably
selling my browsing data.
10:10But this is where today's
sponsor, Incogni, comes in.
10:13With your permission, Incogni will fight the data
10:16brokers on your behalf by finding out
who has your data, which laws apply,
10:20and then they request, with the appropriate
legal language, your information be deleted.
10:25You just sign up and they'll give you a list of
all the companies that have your information,
10:29how severe each claim is, and
the status of all your requests.
10:33Now, I signed up last April, and since
then they filed 317 requests for me,
10:37281 of which have been completed,
saving me over 210 hours of work.
10:42And now I'm no longer getting
ads for glasses I don't need.
10:45So to try it Incogni and fight against the
data brokers, visit Incogni.com/Veritasium.
10:50You can also click the link in the
description or scan this QR code.
10:54Make sure to use the code Veritasium to get 60%
10:57off your annual subscription to
take control of your data today.
11:00That's Incogni.com/Veritasium.
11:03And now back to the electron microscope.
11:06Despite Scherzer's aberration limit
work on the transmission electron microscope continued.
11:11During the next four decades.
11:13People tried boosting the resolution with
clever workarounds, and perhaps none more
11:17so than British-American physicist Albert Crewe.
11:21His idea was to replace the tungsten filament,
11:23which fired off electrons at
random, with a more directed source.
11:27So instead of boiling electrons off the surface,
11:30he tried pulling them off with
a stronger electric field.
11:34And, by sharpening the tungsten into a fine tip,
11:37he was able to create a narrow beam which was
over a thousand times brighter than before.
11:43He paired his new narrow beam
with an unlikely technology.
11:48The cathode ray tube TV.
11:51These TVs worked by scanning an
electron beam across a screen. The
11:55screen was coated in a phosphor that
produced light when hit by electrons.
11:58And by varying the intensity of the electron beam,
12:01you could vary the brightness of the
screen, giving you a black-and-white image.
12:07Crewe was inspired to design a
similar electron beam for the TEM
12:10that would scan across the nanoscopic sample.
12:13So instead of creating an imprint
of the whole sample at once,
12:17Crewe's electron beam made smaller imprints,
mapping the sample out bit by bit.
12:22This wasn't the first time someone had
tried to make a scanning version of the TEM.
12:27German researcher Manfred von Ardenne
built an early prototype in the 1930s,
12:33but it was destroyed during WWII.
12:36When Crewe revived Ardenne's design,
he made several drastic improvements,
12:40and by 1970 he had this.
12:43The first image of single atoms
taken with the electron microscope.
12:50Researchers quickly jumped to employ his
tech, producing countless images of atoms.
12:56After nearly a century of improvements
from Ruska, Crewe, and many others,
13:00the magnification upgrades on
the TEM had reached their peak.
13:05But Scherzer's problem persisted.
13:08Spherical aberration set a hard
limit on how small you could see.
13:12Even Crewe himself gave up on trying to
get around it after over ten years of work.
13:17"Unfortunately, we could never make it
work. After many heartbreaking attempts,
13:22we were forced to admit defeat."
13:25Around this time, other microscopes
emerged that could also image atoms.
13:29These probe microscopes work by gliding an
incredibly small stylus across the sample.
13:34The stylus detects variations in quantum effects
13:37or nanoscale forces to then map the
surface structure of the sample.
13:41These were easier to build, and
because they didn't use any lenses,
13:45they weren't limited by spherical aberration.
13:48Their images were even 3D.
13:51But the looming issue was that
these probes weren't really
13:54seeing atoms. It was more like feeling atoms.
13:58Throughout the 80s and 90s, this was all we had.
14:01But what if there was another way?
14:04Scherzer's theorem proof that a diverging,
radially symmetric lens isn't possible.
14:09But if you're willing to give up on that
symmetry, well, the theorem no longer applies.
14:14The problem is that radial symmetry is
arguably the most important property of
14:18any lens, because if you break the
symmetry, you also break the image.
14:26But three maverick scientists
thought there might be a way.
14:29Knut Urban, Max Haider, and Harold Rose were known
14:32in the electron microscope
community as troublemakers,
14:36and for years barely anyone had been
interested in their research or,
14:40more importantly, in funding it.
14:43And for a good reason too.
Their idea was kind of crazy.
14:46I mean, they purposely wanted to break the
image using a lens that wasn't symmetric.
14:51Their hope was that there would be a small part
14:54of this distorted image that
would be slightly diverging.
14:57And maybe, just maybe, this
small part could correct the
15:01spherical aberration of the original lens.
15:08To distort the image, they used a massive nest
of electromagnets with six,
15:12eight, or even ten separate coils and
magnets with bumpy magnetic fields.
15:18These were known as the hexapole,
octopole, and decapole magnets.
15:22So as the electron beam passed through a hexapole,
15:25it would twist and squeeze the flat
2D image into a triangular saddle.
15:31And the circumference of the original beam
would be pushed into the three corners,
15:35with the rest of the interior stretched out.
15:38But now the middle of the image
would have a slight concave bow,
15:42giving the effect of a small divergence.
15:45Then Rose, Haider, and Urban forced the
beam through a second hexapole , one that worked
15:49the opposite way, so it would unbend the
distorted image back into a circular shape.
15:55But now they, calculated this
new image might have the remnants
15:59of that tiny divergence still in its center, with
spherical aberration pointing in the opposite way.
16:05So if they got their maths and engineering
exactly right, they could feed an image with
16:09spherical aberration through these two lenses
to almost completely counteract the effect.
16:15And I imagine a lot of people in the field thought
it was a crazy idea when it was proposed, right?
16:19Not only the concept, but, that this is,
like, technically feasible, I believe.
16:25It was thought that this is not possible.
16:29By May 1997, the group had just two months of
16:32development time left before their
last sponsor withdrew their backing.
16:36And to make matters worse, their latest lens
iteration was still just on the drawing board.
16:41But somehow, by the 23rd of July,
16:43just a week before their funding ran
out, the new lens was ready to test.
16:47They gingerly placed it into the microscope,
16:50but like every time before it,
the lens was unstable and failed.
16:54So they decided to switch off the equipment
for 24 hours to allow the magnets to settle.
17:00And then at 2 a.m. on the
24th, they turned it on again,
17:04and almost magically, the
picture started to stabilize.
17:10Suddenly, there was no aberration.
Only beautiful, clear images of atoms.
17:19After more than 60 years of failed attempts,
17:22Urban, Rose, and Haider pulled
off the seemingly impossible.
17:25With this method, they cut the resolution
of the TEM down to only 0.13 nanometers.
17:31An average TEM image went from
looking like this to this.
17:38A few months after the group's breakthrough,
17:40Knut Urban attended a microscopy
conference to share these results.
17:43But because of the group's reputation,
17:45he was relegated to a small back
room that barely anyone noticed.
17:50Soon, however, word spread that against all odds,
17:53his pictures seemed real. Then
a crowd of hundreds formed.
17:59People were lining up outside, hoping to get
a glimpse of their stunningly sharp images.
18:09So we're going to get a sample holder.
18:13Yeah, so we got a sample holder out.
18:15We put that under the optical microscope.
18:18So the sample itself is a small lamella that
you can't see without the optical microscope.
18:26Yeah.
Have a look through that.
18:30Beautiful. On top of the B there's
a prong. Yeah. And on the very top
18:34of that on the left-hand side,
looks like a little bit of dust.
18:38That's our actual sample.
18:47Okay, now I simply go up with the magnification
and I do a very few, like, more basic alignments.
18:55In this electron microscope because it's
called transmission electron microscope,
18:59the electrons always transmit the sample.
19:01Here we look through our entire sample at the same time.
19:05 And that's why it's so important
that we align the sample.
19:08If you imagine atoms in a high-symmetry
direction are lined up like pearls on a string.
19:14When we look down it, well, we can see an image.
19:16But if we are in, like, some random direction
that everything would be just blurred.
19:20So that's why we have to do
some tilting in the edge.
19:23And this is where the actual sample
starts — the strontium titanate.
19:27And this is a thin region where we
hope to get atomic resolution.
19:32So this is 5000 times? Yes
19:36And we see strontium, titanium. We see
oxygen. We see carbon. That's contamination.
19:42So most likely what we're looking
at is carbon contamination.
19:46When you do this focus, what
are you really looking for?
19:49I look for this edge to become sharp.
19:59What? Just like that. That's wild.
20:07Shortly after, the group successfully corrected
20:11
Orndrej Krivanek independently achieved
20:13the same for Crewe’s version of
the microscope, the scanning TEM.
20:18And in 2020, all four were awarded
the prestigious Kavli Prize in
20:22Nanoscience for accomplishing what
so many others thought impossible.
20:27Through their persistence and ingenuity,
seeing atoms like this is, well, normal.
20:36How big of a difference does
aberration correction make?
20:39If you want to see atoms and if you want to,
for example, measure in atomic distances,
20:45and if you want to learn what type of atoms
you have, you need aberration correction.
20:50Any research that's material science.
Materials engineering, chemical engineering...
20:57You need to see what's happening at the atomic level
20:59because you want to relate
the properties to the structure.
21:03If you can't see the structure at the atomic
level, you only have half of the information.
21:07So that was a game changer.
21:08That's why nowadays every university in
principle needs a microscope like that.
