Bizarre travelling flame discovery

(upbeat jazz music)

- Watch what happens when I pour lighter fluid

into this ring-shaped trough and then set light to it.

Look, the flame goes round and round.

Isn't that weird?

I really wanted to figure out what was going on,

and in doing so, I fell down a bit of a rabbit hole

about things called excitable mediums.

Excitable mediums are really interesting,

and by the end I had a handful of designs

that made the flames do really bizarre things.

This one in particular is a lot of fun,

but first, why was I dowsing my 3D prints

in lighter fluid in the first place?

Well, it's actually something one of my subscribers noticed,

and they sent me an email.

Rudy Stevens designed the part as a cap

for a small, closed ecosystem he was working on.

To make it a snug fit, he decided to heat it up

and mold it in place.

Using lighter fluid for the task,

he stumbled upon this weird phenomenon,

so, of course, I said, "That's amazing.

"Please send me the file so I can try it for myself,"

and, you know, it worked like a charm,

but it got me thinking, is this trough optimal?

Like, it can be quite hard to get the flame started,

and it doesn't always burn for very long once it is.

So, I asked Oscar Morris to design me

a parameterized CAD file so I could tweak the variables

and see what effect that had.

This is what Oscar does for a living, by the way.

The link to his website in the description.

See how I can change the size of the trough,

the width of the ring, the size of the opening.

I can also change the angle of the opening.

I wanted to have that ability,

because you'll notice in Rudy's design

that the opening is pointing inward slightly,

and I wondered if that was important.

The first thing I varied was just the width of the trough,

but I also varied the size of the ring slightly,

so I could nest them all together on the print bed

and print them at the same time.

I don't think that made much of a difference,

and it turns out there is a bit

of a sweet spot around five millimeters.

If the trough is too wide,

the whole thing just catches fire

instead of forming, like, a whizzing flame,

and if the trough is too small, well, it works,

but it doesn't last very long.

With the width of the trough fixed at five millimeters,

I then varied the size of the opening.

With a very small opening,

it's hard for the lighter fluid to catch on fire.

When it does, you get a small flame

that seems to travel very slowly for a short time

and then go out.

Conversely, if the trough is too open,

the whole thing tends to set on fire

instead of having a flame whizzing round.

It's a bit like when the trough was too wide,

you can get a whizzing flame eventually

when there's only a bit of lighter fluid,

but then it doesn't last very long.

I also thought it'd be interesting to try a small opening,

but where the width of the trough is increased

until that small opening is five millimeters,

but it didn't do anything interesting.

Like, it shows some of the same behavior,

but just in little bursts, and it's not sustainable.

Making the opening point inwards or outwards

didn't seem to make a difference, and in the end,

the optimum configuration seemed to be

a trough that was five millimeters wide

with a semi-circular cross-section.

I managed to get two flames going around at the same time

at one point, not sure how I managed that.

So, what's going on?

Well, when you burn lighter fluid,

you're not actually burning the liquid.

Instead, because the lighter fluid is really volatile,

it's constantly evaporating.

That's why it has such a strong smell.

So, above this little pool of lighter fluid,

there's a whole bunch of short-chain hydrocarbon molecules,

whizzing about and mixing with the oxygen in the air,

so when a flame is introduced,

that mixture of hydrocarbon gas and oxygen ignites.

The heat from the burning vapor

increases the rate of evaporation,

so that you have a steady source of new vapor

to keep the fire going.

It's similar to how a candle works, actually,

except that with a candle,

the flame first melts the solid wax,

which is then drawn up through the wick by capillary action

and then vaporized before being burnt in the air.

My hypothesis is that when you have

a very thin reservoir of lighter fluid like this,

then you only have a thin ring of vapor above the liquid.

So, when it's ignited, the flame isn't as hot,

which means the evaporation rate of the liquid fuel

doesn't increase that much,

and so the flame can't be maintained in that spot,

but to the left and right of that initial flame,

there's all this unburnt vapor,

so you get two flames running around the ring,

and they die when they meet at the opposite side.

So, that ring of vapor is a bit like a fuse,

arranged into a ring.

If you light it at one point, it travels around.

If you keep relighting it,

eventually, you get one flame going round on its own,

presumably because of some imbalance in the reservoir

or something like that.

I had some success forcing a single flame by putting my hand

over the ring just to one side of the ignition point,

then quickly removing my hand after it had blocked

one of the flames,

but why does it keep going round and round?

Well, brilliantly, in the time it takes for the flame to do

one full circuit of the ring,

enough lighter fluid has evaporated at the starting point

for the flame to be able to continue,

and so you get this flame that whizzes

round and round and round.

It's a bit like if you had a loop of fuse,

but somehow, by the time the flame got back round

to where it started,

the fuse had, like, regrown or something.

It's like knocking down a loop of dominoes,

except where you've got some mechanism running behind,

putting them back up again.

In other words, this mad contraption from JK Brickworks,

a link to that video in the card in the description.

The vapor explanation would also explain

why it doesn't work particularly well

when the trough is too wide.

A wider trough means more vapor, which means a hotter flame,

which means faster evaporation,

and so the flame can maintain itself in that position.

It also explains why a trough with high walls works better.

It puts a bit of distance between the flame and the liquid,

so that the heat from the flame

can't cause too much additional evaporation.

That's compared to what we saw when the trough was shallow,

and the flame was just everywhere all at once,

but presumably, if you close the trough too much,

then you have an issue with oxygen supply maybe,

I'm not sure.

So, this ring of lighter fluid, it turns out,

is an example of an excitable medium.

An excitable medium is something that has the potential

to switch to an excited state.

In this case, it's the hydrocarbon vapor

that has the potential to be on fire.

That's not the full definition, though.

There are three other criteria

that make it an interesting subject.

The first is that after it has been excited,

it can't be excited again right away.

The second characteristic is that

after a certain amount of time,

the medium is once again excitable.

That's called the refractory time,

and crucially, any part of the medium will become excited

if a neighboring part of the medium is excited.

It's such a simple definition, but it fully explains

the fire going round and round the ring,

and it's interesting that Rudy stumbled upon

just the right parameters for an excitable medium to emerge,

a thin enough channel,

so that the vapor isn't excitable immediately

after it catches fire,

and a ring with long enough circumference

that by the time one circuit is completed,

the refractory time of the medium has passed,

and the gas above the ring is once again excitable.

It's kind of reminiscent of cellular automata,

interesting movements that arise from simple rules.

You're probably familiar with the Game of Life,

for example, but this is a cellular automata model

of an excitable medium,

and it evolves in these interesting ways.

There's lots of examples in biology.

Signaling in organisms can often be modeled

as an excitable medium.

Heart fibrillation is an example

of a pathological excitable medium.

There's an infection called geographic tongue,

and it's believed to be a very slow example

of an excitable medium.

The infection spreads, the tongue recovers,

can't be infected again immediately,

but after some refractory time, the infection can come back,

and you get this complex pattern on the tongue.

Someone called Ollie sent me this video of the bubbles

on the surface of a hot chocolate collapsing.

The bubbles are slowly evaporating, so they're ready to pop,

which is to say they are excitable,

and to some extent each bubble is being supported

by its neighboring bubbles, so if one bubble collapses,

it's likely that the other bubbles around it will collapse,

and so it spreads as this pulse wave

across the surface of the hot chocolate.

Sometimes, you think, "That's interesting enough

"for someone to write a paper about,"

and then you find out they already have.

Mexican waves in sports stadiums are another example.

The people are the excitable medium in that case,

and just like with the lighter fluid ring,

it's possible to get a wave

that goes round and round and round, and forest fires.

It's possible for a forest to be

in the excited state of on fire,

and it fulfills all the other criteria, too,

it's just that the refractory time is incredibly long.

The forest has to regrow before another pulse wave

can pass through it.

On our podcast, Matt Parker asked the important question,

"How large would a ring-shaped forest need to be

"to sustain a continuous forest fire

"going round and the ring forest,

"just like the flame going round and round

"the ring of lighter fluid?"

The podcast is called, "A Podcast of Unnecessary Detail,"

by the way, and one of our listeners went ahead

and did the calculations.

This is very rough,

but the front of a forest fire progresses

at about 10 kilometers per hour,

which is surprisingly fast, actually,

and some forest fires happen as often as annually,

which means the refractory time can be as little

as one year or 8,760 hours.

So, if the fire front travels at 10 kilometers per hour,

and the refractory time of a forest is 8,760 hours,

then the circumference of a ring-shaped forest

would need to be 87,600 kilometers.

The circumference of the Earth is less than half that,

at 40,000 kilometers,

so if there was a forest all the way around

the circumference of the Earth,

it would not be long enough to sustain a never-ending fire

that goes round and round and round,

but I mean, you could zigzag the forest,

or maybe you could make the forest out of bamboo,

because that grows back quickly,

but anyway, there you go.

That is how excitable mediums can lead to spinning flames.

Finally, I got Oscar to design

a few more cool things for me.

I wanted to see what would happen if the ring split in two

and then came back together.

That's cool, isn't it?

It's interesting to see that you occasionally get,

just very briefly, a sustained flame in one position,

and it can cause other flames to change direction

or sometimes it can act almost like a spawning point,

so you get more and more flames

coming from that one location.

And here's a spiral.

(lighter clicks)

And here's a figure-eight-shaped trough.

I really didn't think this would work,

because obviously there's a raised section

where one track goes over the other,

so you have it slanting either side,

and I thought the lighter fluid would drain away from there.

Like, it definitely does,

but I guess enough of the lighter fluid

sticks around up there for the fire to go round and round

at least a few times before it goes out.

Then, there's this design, that has a flame

going round a central ring,

and every time it passes one of the arms,

a flame shoots out across the arm.

That's cool, isn't it?

And here's a modification of the spiral.

The issue with the spiral is it's not a closed loop,

so you set fire to one end, it travels round and round

until it reaches the other end, and then it stops.

Wouldn't it be good if you could keep it going somehow?

That's what the extra circle is for.

The flame that goes round around the circle

should keep firing flames down the spiral.

In reality, when you have these junctions,

you often end up with a sustained flame at that junction,

and that seems to be powering

the continuous supply of flames around the spiral,

but eventually, as the lighter fluid

dries out more and more, the design works as intended.

Now, there is still quite a bit of lighter fluid

in the middle, so there's a bit of strange behavior there.

It even sends one flame back in the opposite direction.

If you want to print out any of these for yourself,

there's a link in the description for that.

Lighter fluid is dangerous,

so that would be something for responsible adults only,

and if you have any ideas for tracks

that you would like to see, send me an email,

tell me on my Discord server.

For example, just before I was about to export this video,

Rudy sent me another design.

This is the flames solving a maze.

How cool is that?

It's actually quite nice T-shirt.

I don't really go for science T-shirts.

It's quite subtle, though.

This video is sponsored by Jane Street.

Jane Street is a quantitative trading firm

with offices in New York, London, Hong Kong,

Amsterdam, and Singapore.

They use techniques from machine learning,

distributed systems, programmable hardware,

and statistics to trade on markets around the world.

They're also a bunch of people who love

solving interesting problems, and they're passionate

about a lot of the same things that we are.

For an example of their nerdiness,

they designed this T-shirt, look.

It shows, like, how particles move in a cloud chamber.

You know, if you've got, like, maybe a magnetic field

and a particle whizzes in, it's going to spiral,

because of the magnetic field,

but some of them aren't, if they're not charged,

and it kind of slows them down, traps them,

and you use it in science, like in the large hadron collider

and things like that, but like embedded in there

is the Jane Street logo.

They currently have a few educational programs

that are accepting applications in both their New York

and London offices for pre-university

and university students.

These programs are part of Jane Street's commitment

to give as many people as possible a chance to learn

more about what they do.

The programs themselves are all a bit different,

but focus on topics that you might be interested in,

things like complex mathematical theory,

programming, statistics and more,

and it's definitely worth saying

you don't need a background in finance to apply.

Jane Street is looking for curious people

from any background with a passion for technology

and creative problem solving.

In other words, people like you.

Check out the link in the description now

to see what opportunities are currently open

and to see if any of them might be for you.

They currently have three programs open in NYC

and one in London,

and applications are closing in just a few weeks,

so check out Jane Street today.

Hope you enjoyed this video.

If you did, don't forget to hit subscribe

and the algorithm thinks you'll enjoy this video next.

(upbeat jazz music)