0:00:09.390,0:00:12.810
Michael Büker: Yes, alright, thank[br]you very much, okay. I’m glad

0:00:12.810,0:00:16.510
that you all found your way here[br]and it’s been mentioned already,

0:00:16.510,0:00:19.920
this is Comic Sans, which as you[br]know is the official type-font

0:00:19.920,0:00:23.930
of awesome particle physics stuff.[br]<i>laughter</i>

0:00:23.930,0:00:27.990
But in the interest of our mental[br]sanity, I will keep it to other fonts.

0:00:27.990,0:00:32.500
So from here on Comic Sans[br]is just a bad memory.

0:00:32.500,0:00:36.100
Okay, two things: First the[br]title, Breaking Baryons,

0:00:36.100,0:00:39.100
which of course is an allusion[br]to Breaking Bad, was inspired

0:00:39.100,0:00:44.520
by the wonderful talk from last year which[br]was called “How I Met Your Pointer”.

0:00:44.520,0:00:48.150
And which was also very successful[br]and you can check out that talk,

0:00:48.150,0:00:51.919
I got the link there. And this[br]talk goes especially well

0:00:51.919,0:00:56.090
with another talk that we’ll have[br]tomorrow by a real particle physicist,

0:00:56.090,0:00:58.720
at least a bit more than myself.

0:00:58.720,0:01:02.460
And it’s called “Desperately Seeking[br]SUSY” which is about particle theories

0:01:02.460,0:01:05.959
and the real cutting edge physical[br]questions. This is going to be

0:01:05.959,0:01:09.950
happening tomorrow. Allright, so[br]we’re going to start out with my talk

0:01:09.950,0:01:14.510
and I’m going to be talking about the[br]questions of “what are we doing?”,

0:01:14.510,0:01:18.330
“why?” and “what kind of stuff do we[br]use?”. And I’m gonna spend some time

0:01:18.330,0:01:22.190
on explaining this last part[br]especially. What is it that we do

0:01:22.190,0:01:27.780
and how does this work? So, what[br]we do is we give a very high energy

0:01:27.780,0:01:32.020
to small particles which[br]we call accelerating.

0:01:32.020,0:01:36.520
But from a certain level of energy[br]this doesn’t really make sense,

0:01:36.520,0:01:40.710
because we don’t actually make them go[br]faster. Once they reach the speed of light

0:01:40.710,0:01:43.930
they can’t go any faster. We just[br]turn up the energy and the speed

0:01:43.930,0:01:48.140
doesn’t really change. This is technically[br]useful but it also gives rise

0:01:48.140,0:01:53.520
to doubts about the term accelerating,[br]but anyway, we just call it ‘accelerate’.

0:01:53.520,0:01:56.710
There’s 2 basic types of devices that[br]you see there, you have storage rings,

0:01:56.710,0:02:00.450
which are the circular facilities that[br]most of you know. And then there is

0:02:00.450,0:02:03.270
linear accelerators which are in[br]comparison very boring, so I’m

0:02:03.270,0:02:07.720
not going to be talking about them[br]a lot. We make the particles collide

0:02:07.720,0:02:11.370
which is the reason for giving them high[br]energies, we want them to smash head-on.

0:02:11.370,0:02:14.950
And then this last part which is about[br]the most difficult thing is we just

0:02:14.950,0:02:19.680
see what happens. Which is not[br]at all as easy as it might sound.

0:02:19.680,0:02:23.520
So why are we doing this? You all[br]know this formula but I’m going to try

0:02:23.520,0:02:26.690
and put it in terms which are[br]a little bit closer to our hearts,

0:02:26.690,0:02:30.570
as we are here at Congress.[br]I might postulate that

0:02:30.570,0:02:35.760
parts, like electrical parts, building[br]parts, are actually the same as a device.

0:02:35.760,0:02:39.200
Now this is not quite wrong but it[br]doesn’t feel exactly right, either.

0:02:39.200,0:02:44.040
I mean, if you have some parts and then[br]build a device from it, it’s not the same.

0:02:44.040,0:02:49.220
It’s made from the same thing but you do[br]require a certain amount of conversion.

0:02:49.220,0:02:53.039
You have a building process, you have[br]specific rules how you can assemble

0:02:53.039,0:02:56.940
the parts to make a device and[br]if you do it wrong it will not work.

0:02:56.940,0:03:01.400
And this is actually pretty similar to the[br]notion of energy being equivalent to mass,

0:03:01.400,0:03:04.970
because energy can be converted into mass[br]but it’s not at all easy and it follows

0:03:04.970,0:03:08.910
a lot of very strict rules. But[br]we can use this principle

0:03:08.910,0:03:12.749
when we analyze how particle[br]reactions are used to take a look at

0:03:12.749,0:03:17.900
what mass and what energy forms[br]there are. Now suppose we are

0:03:17.900,0:03:21.840
thinking about a device[br]which is very, very rare,

0:03:21.840,0:03:26.640
such as a toaster that runs Net-BSD.[br]<i>laughter</i>

0:03:26.640,0:03:29.980
Now as you can see from the photo[br]and the fact that you see a photo,

0:03:29.980,0:03:33.520
I’m not making this shit up. There[br]is a toaster that runs Net-BSD but

0:03:33.520,0:03:37.100
that’s beside the point. Now if we[br]are particle physicists and we want

0:03:37.100,0:03:41.320
to research this question, we know[br]that parts are the same as a device,

0:03:41.320,0:03:45.630
so if we just get enough parts and[br]do the right kind of things to them,

0:03:45.630,0:03:49.340
there might just turn out, out of[br]nowhere a toaster that runs Net BSD.

0:03:49.340,0:03:54.580
So let’s give it a try. We produce[br]collisions with technical parts

0:03:54.580,0:03:58.790
and if we do enough of it, and if we[br]do it right, then there is going to be

0:03:58.790,0:04:02.520
this result. Now from these pictures[br]you can see, that doesn’t seem

0:04:02.520,0:04:07.250
to make a lot of sense. You will not[br]get a toaster from colliding vehicles.

0:04:07.250,0:04:10.590
<i>laughter</i>[br]But as particle physics go,

0:04:10.590,0:04:14.159
this is the best we can do. We[br]just smash stuff into each other

0:04:14.159,0:04:17.750
and we hope that some other stuff[br]comes out which is more interesting.

0:04:17.750,0:04:22.600
And that’s what we do. So to[br]put it in the technical terms,

0:04:22.600,0:04:26.569
we use storage rings which are this[br]one circular kind of accelerator

0:04:26.569,0:04:30.520
to produce collisions. Lots[br]of them with high energy.

0:04:30.520,0:04:34.879
And then we put some enormous[br]experimental devices there

0:04:34.879,0:04:38.480
and we use them to analyze what[br]happens. Now first let’s talk about

0:04:38.480,0:04:43.029
these storage rings. This schematic[br]view is what a storage ring is

0:04:43.029,0:04:47.009
mostly made of, and you can see right[br]away, that it’s not actually a circle.

0:04:47.009,0:04:50.279
And this is true for any storage ring.[br]If you look at them closely they are

0:04:50.279,0:04:54.229
not a perfect circle, you always[br]have acceleration parts which are

0:04:54.229,0:04:58.219
not actually curved. So we[br]have the 2 basic elements

0:04:58.219,0:05:02.270
of a curved part which is just “the[br]curve” and then you have a straight part

0:05:02.270,0:05:06.539
which is there for acceleration. Now you[br]have this separation, it would be nicer

0:05:06.539,0:05:11.009
to have a ring but it’s much more easy[br]this way. You have the acceleration

0:05:11.009,0:05:13.789
where it is straight and because it is[br]straight you don’t need to worry about

0:05:13.789,0:05:18.060
making the particles go on a curved[br]path. So you can just leave out

0:05:18.060,0:05:22.509
the magnetic fields. We[br]need magnetic fields

0:05:22.509,0:05:26.669
to keep them on a curve, but we need[br]electrical fields to accelerate them.

0:05:26.669,0:05:30.469
Now we could try and assemble these[br]into one kind of device. A device

0:05:30.469,0:05:34.409
that uses an electric field to accelerate[br]the particles and at the same time

0:05:34.409,0:05:38.040
uses a magnetic field to keep them on[br]a curved path. Now this is the first thing

0:05:38.040,0:05:41.280
that was tried. These kinds of[br]accelerators where called cyclotrons,

0:05:41.280,0:05:43.710
but they were very inefficient, you[br]couldn’t go to high energies, it was

0:05:43.710,0:05:47.839
very difficult. So the evolution went to

0:05:47.839,0:05:51.440
this way where you just[br]physically separate the 2 tasks.

0:05:51.440,0:05:54.990
You have a straight part for acceleration,[br]you have a curved part for the curve

0:05:54.990,0:05:57.890
and then that’s much more easy.[br]Okay, so let’s take a look at the

0:05:57.890,0:06:01.260
acceleration part of things. You[br]may know computer games

0:06:01.260,0:06:05.339
where you go racing about and then[br]you have some kind of arrows

0:06:05.339,0:06:08.559
on the ground and if you go over them in[br]the right direction they make you faster.

0:06:08.559,0:06:12.270
This is a kind of booster if you will.

0:06:12.270,0:06:16.229
If you happen to go around the wrong[br]way and you go onto these arrows,

0:06:16.229,0:06:19.599
they will slow you down, which makes sense[br]because you’re going the wrong way,

0:06:19.599,0:06:23.670
you shouldn’t be trying that. And this is

0:06:23.670,0:06:27.629
the same effect we can think of when we[br]think about what an electrical field does

0:06:27.629,0:06:31.589
to a charged particle. If a charged[br]particle moves through an electrical field

0:06:31.589,0:06:36.080
in the ‘right’ direction so to speak[br]it will speed the particle up,

0:06:36.080,0:06:38.889
taking energy from the field and to the[br]particle making it go faster. But if you

0:06:38.889,0:06:42.430
go the wrong way, then this particle[br]will slow down and it will

0:06:42.430,0:06:48.170
give off energy. If we where to try and…

0:06:48.170,0:06:51.610
let’s say we have a level editor,[br]right? And we can edit this level

0:06:51.610,0:06:55.610
where this little vehicle is going and[br]we want to make it go really fast.

0:06:55.610,0:06:58.060
So what do we do? We just take this[br]acceleration path, we just take

0:06:58.060,0:07:02.449
these arrows and we put them in a long[br]line. Let’s put 4, 5, 10 of them

0:07:02.449,0:07:06.159
in a row, so if we go over them[br]we’ll be really fast at the end.

0:07:06.159,0:07:09.280
Now suppose the level editor[br]does not allow this. It’s just

0:07:09.280,0:07:12.800
by the rules of the game it’s not possible[br]to put a bunch of arrows in a row.

0:07:12.800,0:07:17.659
Which sucks, because then we can’t[br]really make them go really fast.

0:07:17.659,0:07:22.349
But then we just ask an engineer[br]who’s got this shit together.

0:07:22.349,0:07:25.919
And what is he going to suggest?[br]You know what he’s going to suggest.

0:07:25.919,0:07:30.239
Can I hear it? Come on, “inverse the[br]polarity”, that’s what he always does!

0:07:30.239,0:07:39.050
<i>laughter and applause</i>

0:07:39.050,0:07:43.420
So we inverse the polarity. And we are[br]going to make our track look like this.

0:07:43.420,0:07:46.599
So we have an arrow which gives us a boost[br]in the right direction and then there’s

0:07:46.599,0:07:50.260
an arrow in the wrong direction.[br]If we go over the track in this way,

0:07:50.260,0:07:54.180
we’ll speed up and slow down and speed[br]up and slow down. And in the end

0:07:54.180,0:07:57.530
we won’t win anything. But here is where[br]Geordi comes into play, because

0:07:57.530,0:08:01.729
we’ll be switching polarities at just the[br]right moment and if we switch polarities

0:08:01.729,0:08:05.319
at the precise moment that we are[br]in between two of these fields,

0:08:05.319,0:08:09.069
then the next one will be an accelerating[br]field. And it goes on and on like this,

0:08:09.069,0:08:12.270
we always switch the direction[br]of the arrows at the right moment

0:08:12.270,0:08:16.429
when we are in between the two. And[br]from the point of view of the vehicle

0:08:16.429,0:08:19.849
it will look like there is an accelerating[br]field followed by an accelerating field,

0:08:19.849,0:08:23.720
followed by an accelerating field.[br]Which is the same as we tried to build

0:08:23.720,0:08:27.069
but which the game, or in the case[br]of real accelerators the universe

0:08:27.069,0:08:30.899
just wouldn’t allow. So we’re tricking[br]the universe by using Geordi’s tip

0:08:30.899,0:08:34.720
and inversing the polarity at just the[br]right moments. And this is what is done

0:08:34.720,0:08:41.580
in particle accelerators and this is[br]called Radio Frequency Acceleration.

0:08:41.580,0:08:44.450
Now this kind of device that you see[br]there is the device that is used

0:08:44.450,0:08:48.329
for this actual process in actual[br]accelerators. It’s about as big

0:08:48.329,0:08:51.670
as a human child, but it[br]weighs a bit more, it weighs

0:08:51.670,0:08:55.810
several hundred kilograms.[br]And in contrast to a child

0:08:55.810,0:08:59.779
it’s made of a metal called Niobium.[br]Now Niobium is a rare metal,

0:08:59.779,0:09:03.410
but it’s not super rare, and it fulfills

0:09:03.410,0:09:06.779
3 basic requirements that[br]we have for these devices.

0:09:06.779,0:09:10.480
It’s ductile, which means you can[br]easily shape it, because you see

0:09:10.480,0:09:14.170
that this shape is really weird, you got[br]these kind of cone things going on,

0:09:14.170,0:09:19.670
and they must be very precise. If these[br]cones on the inside of the cavity

0:09:19.670,0:09:25.130
are off by just micrometers the whole[br]thing won’t work. So you need a metal

0:09:25.130,0:09:28.389
which can be formed very well.

0:09:28.389,0:09:31.970
Then you must be able to make it[br]superconductive, to cool it down

0:09:31.970,0:09:35.770
to a temperature where it will[br]lose its electrical resistance.

0:09:35.770,0:09:39.029
The electrical resistance will go down[br]to almost zero, some nano-Ohms

0:09:39.029,0:09:43.169
is what’s left. So that’s the second[br]requirement for this metal,

0:09:43.169,0:09:46.220
and the third one is: it shouldn’t[br]be ‘super’ expensive. I guess

0:09:46.220,0:09:50.480
you could use platinum or something but[br]then you couldn’t pay for the accelerator

0:09:50.480,0:09:54.779
and as we are going to see, the[br]accelerator is expensive enough as it is.

0:09:54.779,0:09:58.060
So Niobium is what is used[br]for this kind of device and

0:09:58.060,0:10:04.690
as I said, we cool it down to about[br]4 Kelvins, which is -269°C

0:10:04.690,0:10:09.589
or 4°C above absolute Zero.[br]And at this temperature,

0:10:09.589,0:10:12.670
the electrical resistance of the metal[br]is almost zero which we need

0:10:12.670,0:10:16.569
for the high frequency[br]fields that we put in.

0:10:16.569,0:10:19.949
What we used to cool these things is[br]liquid helium, so when they’re in use

0:10:19.949,0:10:23.580
inside the accelerator they’re not[br]naked, exposed like you see here,

0:10:23.580,0:10:27.300
they are enclosed by huge tanks[br]which are super tight and must

0:10:27.300,0:10:31.890
hold on to large pressures and[br]be super temperature efficient,

0:10:31.890,0:10:36.190
very well insulating[br]because these must keep

0:10:36.190,0:10:39.839
the liquid helium inside. But on[br]the outside there is the tunnel

0:10:39.839,0:10:43.510
of the accelerator and that’s where people[br]walk around. Not while the accelerator is

0:10:43.510,0:10:46.750
running, but people walk around to do[br]maintenance and stuff. So you must have

0:10:46.750,0:10:51.060
a temperature differential between room[br]temperature next to the accelerator

0:10:51.060,0:10:55.779
and 4 Kelvin inside the tank[br]where this cavity is sitting.

0:10:55.779,0:10:59.300
So you have a temperature difference[br]of 300 degrees, which this tank

0:10:59.300,0:11:02.540
around the cavity must keep. So that’s[br]a very hard job, actually cooling

0:11:02.540,0:11:07.360
is one of the more difficult things

0:11:07.360,0:11:11.640
from an engineering point of view.[br]The thing which feeds the fields

0:11:11.640,0:11:16.089
– the actual changing electrical[br]fields are polarity switched –

0:11:16.089,0:11:19.660
into these cavities are called klystrons.[br]There’s a picture of a klystron,

0:11:19.660,0:11:23.890
it’s the longish device sitting on the[br]bottom. And they’re usually about

0:11:23.890,0:11:28.749
as big as a refrigerator or two.[br]And these klystrons produce

0:11:28.749,0:11:31.990
radio waves not very much unlike that

0:11:31.990,0:11:35.440
which you hear in your car when you just[br]turn on the radio. It’s not modulated

0:11:35.440,0:11:38.730
in the same way, so there’s no[br]sound information encoded,

0:11:38.730,0:11:42.290
but it’s extremely strong.[br]You can see on the bottom

0:11:42.290,0:11:46.290
that one of these klystrons as it is in[br]use at the LHC has a transmitting power

0:11:46.290,0:11:51.680
of 300 Kilowatts. Now if you think of the[br]transmitting power of the Fernsehturm

0:11:51.680,0:11:54.860
like the Hertz-Turm which is right next[br]- no, that way -

0:11:54.860,0:11:59.070
which is right next to the conference[br]center, or even the Fernsehturm in Berlin.

0:11:59.070,0:12:02.899
It has about half the transmitting[br]power of one of these klystrons.

0:12:02.899,0:12:06.389
Now for the LHC accelerator[br]16 of them are used.

0:12:06.389,0:12:09.490
So that’s a lot of transmitting power.[br]And because the power is so high

0:12:09.490,0:12:13.110
we don’t actually use cables.[br]Usually you transfer your…

0:12:13.110,0:12:15.629
when you have some oscillator and[br]you’re checking out some signals,

0:12:15.629,0:12:18.560
you just put cables between[br]your source and your device.

0:12:18.560,0:12:22.620
This is not what’s used here, because[br]cables get way too complicated

0:12:22.620,0:12:26.240
when you have these high energies.

0:12:26.240,0:12:28.750
So what is used, is waveguides and that[br]is what you can see on the top there

0:12:28.750,0:12:32.590
in this picture. It looks like an[br]air duct, it looks like there’s some

0:12:32.590,0:12:36.089
sort of air conditioning system and the[br]air moves through. That’s not what it is.

0:12:36.089,0:12:39.800
It is a waveguide which is designed[br]to have the radio waves inside

0:12:39.800,0:12:44.510
radiate in a certain direction.[br]Think of a series of mirrors,

0:12:44.510,0:12:50.620
long rectangular mirrors and you put[br]them all with the mirroring area inside.

0:12:50.620,0:12:54.119
So you have a tube which is mirroring[br]inside. And then at one side

0:12:54.119,0:12:57.090
you shine in a bright light. Now the[br]light can’t escape anywhere and it

0:12:57.090,0:13:00.089
always hits the mirrors so it[br]goes on in a straight path.

0:13:00.089,0:13:04.010
You’ve built yourself a waveguide[br]for light. Now this here,

0:13:04.010,0:13:07.220
this clunky looking metal[br]part is a waveguide

0:13:07.220,0:13:12.110
but for high frequency, high energy radio[br]waves which are fed into the cavities.

0:13:12.110,0:13:16.040
And that’s how acceleration happens.[br]Now let’s talk about the curves.

0:13:16.040,0:13:21.490
This is where it gets less[br]fidgety and more… boom!

0:13:21.490,0:13:24.639
So these devices you see here, there’s[br]2 devices sitting next to each other,

0:13:24.639,0:13:27.689
identical devices. These[br]are the cryo-dipoles.

0:13:27.689,0:13:30.290
Again, they have the word “cryo” in[br]them because they are also cooled

0:13:30.290,0:13:36.450
by liquid helium down to[br]a temperature of about -270°C.

0:13:36.450,0:13:40.129
They’re 40 meters long, they weigh[br]35 tons and each of these babies

0:13:40.129,0:13:44.750
costs about half a million Swiss Francs.

0:13:44.750,0:13:49.670
And as you can see one line above that,[br]there’s 1200 of these curve dipoles

0:13:49.670,0:13:54.719
in the LHC. So there you have[br]a cost of 1.5 to 2 billion dollars

0:13:54.719,0:13:57.850
in the curve magnets alone.[br]We’re not talking acceleration,

0:13:57.850,0:14:01.589
we’re not talking about power use, we are[br]not talking about the helium that you need

0:14:01.589,0:14:05.550
for cooling or the power that you need for[br]cooling. It’s just building these things,

0:14:05.550,0:14:08.769
just building the curve, 27 kilometers.

0:14:08.769,0:14:12.250
And that’s what you have there as a[br]cost. Now what they do is, they make

0:14:12.250,0:14:15.420
a huge magnetic field, because in[br]a magnetic field a charged particle

0:14:15.420,0:14:19.300
will go on a curve. That’s[br]what we want, right? But

0:14:19.300,0:14:23.970
to make these particles with a very high[br]energy and keep them on a tight curve…

0:14:23.970,0:14:27.009
now in particle physics’ terms[br]let’s say that 27 kilometers

0:14:27.009,0:14:30.920
to go around one way is a tight curve.

0:14:30.920,0:14:35.459
We need a current of 12,000 amps.[br]Which is a large current

0:14:35.459,0:14:38.579
that goes through these dipoles.[br]Which is the reason why we have them

0:14:38.579,0:14:44.850
superconductingly cooled, because[br]otherwise you put 12,000 amps

0:14:44.850,0:14:48.460
through a piece of metal and it just melts[br]away. You don’t get a magnetic field,

0:14:48.460,0:14:52.820
maybe for a microsecond or 2.[br]But you want to sustain a stable field

0:14:52.820,0:14:57.029
of 8.5 Tesla to make these[br]protons go around on a curve.

0:14:57.029,0:15:00.890
So, yeah, that’s a big thing.[br]There’s also niobium in there,

0:15:00.890,0:15:05.569
not the big clunky parts like the cavity[br]we saw, but thin niobium wires,

0:15:05.569,0:15:09.500
actually half niobium, half titanium[br]most of the time. But since

0:15:09.500,0:15:14.430
there are so many magnets and it’s[br]so long a curve, there is 600 tons

0:15:14.430,0:15:18.759
of atomic niobium in this[br]entire accelerator thing.

0:15:18.759,0:15:22.610
And this was a fourth of the[br]world production of niobium

0:15:22.610,0:15:26.950
which comes mostly from Brazil by the way.[br]This was a fourth of the world production

0:15:26.950,0:15:30.970
of niobium for 5 years.[br]So that’s where it all went.

0:15:30.970,0:15:35.769
It just went into the accelerator.[br]And now if we have this running,

0:15:35.769,0:15:39.259
we have it up, we have it cooled, we have[br]a large current going, we got our nice

0:15:39.259,0:15:43.109
big magnetic fields. And[br]there is energy stored.

0:15:43.109,0:15:46.910
I mean we put in a lot of power and the[br]magnetic fields are up and they’re stable

0:15:46.910,0:15:50.920
and that means that there’s magnetic[br]energy stored in this. And the amount

0:15:50.920,0:15:53.990
of energy that is stored in the curve[br]magnets alone of the LHC when it’s running

0:15:53.990,0:15:58.009
is 11 gigajoules. Sounds like a lot,

0:15:58.009,0:16:03.770
let’s compare it to something: If we[br]have an absurdly long freight train

0:16:03.770,0:16:07.699
with let’s say 15,000 tons. I hear that[br]normal freight trains in Germany

0:16:07.699,0:16:11.811
or England have about 5000 tons.[br]So let’s take a big freight train

0:16:11.811,0:16:18.369
and multiply it by 3. If this[br]freight train goes at 150 km/h,

0:16:18.369,0:16:21.899
then the kinetic energy, the[br]movement energy of this train

0:16:21.899,0:16:26.839
is equivalent to the magnetic[br]energy that is stored in the LHC.

0:16:26.839,0:16:29.969
And that is why we don’t want[br]any problem with the cooling.

0:16:29.969,0:16:33.740
<i>laughter</i>

0:16:33.740,0:16:39.740
Because if we get a problem with[br]the cooling, bad things happen.

0:16:39.740,0:16:44.469
This is a photograph of what at CERN[br]at the LHC they just call “the incident”.

0:16:44.469,0:16:47.089
<i>laughter</i>

0:16:47.089,0:16:50.059
Which was a tiny mishap that[br]happened just a few weeks

0:16:50.059,0:16:54.060
after the LHC was taken into[br]operation for the first time in 2008.

0:16:54.060,0:16:57.230
And it shut the machine[br]down for about 8 months.

0:16:57.230,0:17:00.390
So that was a bad thing. It’s[br]a funny story when they where

0:17:00.390,0:17:03.290
constructing these magnets; now[br]what you see here is the connection

0:17:03.290,0:17:07.850
between 2 of these magnets. I told you[br]that each of them weighs 35 tons.

0:17:07.850,0:17:12.740
So here you have a connection between[br]2 parts that are 35 tons in weight each.

0:17:12.740,0:17:18.100
And they’re shifted by almost half[br]a meter. So it takes a bit of boom.

0:17:18.100,0:17:21.630
So what happened was: the cooling broke[br]down and the helium escaped and

0:17:21.630,0:17:25.569
the sheer force of the helium expanding,[br]because if you have liquid helium

0:17:25.569,0:17:29.810
and it instantly evaporates into gaseous[br]helium then the volume multiplies

0:17:29.810,0:17:33.650
by a very large amount.[br]And what they had was…

0:17:33.650,0:17:36.720
what I hear is that the tunnel of the[br]LHC, which has a diameter of about

0:17:36.720,0:17:41.010
let’s say 6 or 7 meters was[br]filled with nothing but helium

0:17:41.010,0:17:44.510
which pushed away the air[br]for about 100 meters

0:17:44.510,0:17:48.140
around this incident. So the helium[br]evaporated, it pushed everything away,

0:17:48.140,0:17:52.960
it made everything really cold, some[br]cables broke and some metal broke.

0:17:52.960,0:17:57.010
And the funny thing now is, the[br]engineers that built the LHC,

0:17:57.010,0:18:00.210
before they did that, visited[br]Hamburg. Because here there is

0:18:00.210,0:18:03.511
a particle accelerator which is[br]not quite as large. The LHC

0:18:03.511,0:18:07.770
has 27 kilometers; here in Hamburg we[br]have a particle accelerator called HERA

0:18:07.770,0:18:12.490
which had 6.5 kilometers. So it’s[br]the same ballpark, it’s not as big.

0:18:12.490,0:18:15.750
And in HERA they had a safety system[br]against these kinds of cryo failures,

0:18:15.750,0:18:19.630
they’re called quenches.[br]They had a protection system,

0:18:19.630,0:18:23.480
which protects this exact part.[br]Now we’re talking about “Yeah,

0:18:23.480,0:18:26.710
how should we build this? Should[br]we have a quench-protection

0:18:26.710,0:18:31.030
at the connection between the dipoles?”[br]And the HERA people in Hamburg said:

0:18:31.030,0:18:34.690
“Well we have it, it’s a good thing,[br]you shouldn’t leave it out,

0:18:34.690,0:18:38.630
if you build the LHC.” Well,[br]they left it out. <i>laughter</i>

0:18:38.630,0:18:43.470
They ran out of time, they ran out of[br]money, the LHC project was under pressure.

0:18:43.470,0:18:45.950
Because they had promised to build[br]a big machine by that time and

0:18:45.950,0:18:49.450
they weren’t really finished, so they[br]cut some edges. Well this was

0:18:49.450,0:18:53.930
the edge they cut and it cost them[br]8 months of operation. Which says

0:18:53.930,0:18:59.360
that they really should have listened to[br]the people of Hamburg. Okay, so,

0:18:59.360,0:19:03.940
in summary of the operations of[br]a storage ring we can just say this:

0:19:03.940,0:19:07.040
They get perfectly timed kicks[br]with our polarity switching

0:19:07.040,0:19:11.700
at just the right moment by radio waves[br]generated in these large klystrons

0:19:11.700,0:19:16.110
from the funny looking metal[br]tubes that we called cavities.

0:19:16.110,0:19:18.461
And some big-ass superconducting[br]magnets keep them on a curve

0:19:18.461,0:19:22.780
when they are not being accelerated.[br]Now the trick is, one of these kicks

0:19:22.780,0:19:26.430
like moving through the cavity once, may[br]not give you all the energy you want,

0:19:26.430,0:19:30.160
in fact it doesn’t. But if you[br]make them go round in the ring,

0:19:30.160,0:19:34.150
they come by every couple of[br]nanoseconds. So you just have them

0:19:34.150,0:19:37.660
run through your acceleration all the[br]time. Which is the big difference

0:19:37.660,0:19:40.620
between the storage ring and a linear[br]accelerator. A linear accelerator

0:19:40.620,0:19:44.400
is basically a one shot operation but[br]here, you just give them an energy kick

0:19:44.400,0:19:48.880
every time they come around, which[br]is often, we’re going to see that.

0:19:48.880,0:19:52.880
So that’s the summary of what[br]the storage rings do. Now,

0:19:52.880,0:19:56.570
the machine layout, if you[br]look at a research center

0:19:56.570,0:20:01.160
which has a bunch of accelerators,[br]it almost always goes like this:

0:20:01.160,0:20:05.030
You have some old, small storage[br]rings and then they built

0:20:05.030,0:20:08.620
newer ones which were[br]bigger. So this is just

0:20:08.620,0:20:12.420
a historical development, first[br]you build small machines, then

0:20:12.420,0:20:14.920
techniques get better, engineering gets[br]better, you build bigger machines. But

0:20:14.920,0:20:18.640
you can actually use that, it’s very[br]useful because the older machines,

0:20:18.640,0:20:22.640
you can use as pre-accelerators.[br]For a variety of reasons it’s useful

0:20:22.640,0:20:26.370
to not put in your particles with[br]an energy of zero and then

0:20:26.370,0:20:30.180
have them accelerated up to the energy you[br]want. You want to pre-accelerate them,

0:20:30.180,0:20:33.460
make them a little faster at a time.[br]That’s what you do, you just

0:20:33.460,0:20:37.840
take the old accelerators. And if[br]we look at the accelerator layout

0:20:37.840,0:20:42.200
of some real world research centers,[br]you can actually see this. On the left

0:20:42.200,0:20:47.020
you have CERN in Geneva and on the[br]right you have DESY here in Hamburg.

0:20:47.020,0:20:51.140
And you can see that there are smaller[br]accelerators, which are the older ones,

0:20:51.140,0:20:54.140
and you have bigger accelerators[br]which are connected to them.

0:20:54.140,0:20:59.410
And that’s this layout of the machines.[br]Okay, now let’s talk about collisions.

0:20:59.410,0:21:03.411
This is a nice picture of a collision.[br]It’s not actually a proton collision

0:21:03.411,0:21:08.270
but a heavy-ion collision, which[br]they do part of the time in the LHC.

0:21:08.270,0:21:11.520
They are extremely hard to produce, we’re[br]going to see that, but still we make

0:21:11.520,0:21:15.690
an awful lot of them.[br]So let’s see, first of all

0:21:15.690,0:21:19.330
let’s talk about what the beam looks like,[br]because we’re going to be colliding beams.

0:21:19.330,0:21:23.220
So what are these beams? Is it[br]a continuous stream of particles?

0:21:23.220,0:21:27.780
Well it’s not. Because the acceleration[br]that we use, these radio frequency,

0:21:27.780,0:21:31.960
polarity shifting mechanisms, they[br]make the particles into bunches.

0:21:31.960,0:21:35.730
So you don’t have a continuous stream,[br]you have separate bunches.

0:21:35.730,0:21:38.610
But how large are these bunches?[br]Is there one particle per bunch?

0:21:38.610,0:21:41.150
You’ve got a particle, you wait[br]a while, there’s another particle?

0:21:41.150,0:21:44.650
Well, it’s not like that.[br]Because if it were like that,

0:21:44.650,0:21:49.300
if we had single particles coming after[br]one another, it would be impossible

0:21:49.300,0:21:52.750
to hit them. You have to aim[br]the beams very precisely.

0:21:52.750,0:21:56.620
I mean, think about it. One comes[br]around 27 kilometers around the ring.

0:21:56.620,0:21:59.950
The other comes around 27[br]kilometers going the other way.

0:21:59.950,0:22:03.480
And now you want them to hit. You have[br]to align your magnets very precisely.

0:22:03.480,0:22:07.060
You can think of it like this:[br]You have a guy in Munich

0:22:07.060,0:22:10.790
and you have a guy in Hamburg and[br]they each have a rifle. And the bullets

0:22:10.790,0:22:14.550
of the rifle are let’s say one centimeter[br]in size. So the guy in Hamburg

0:22:14.550,0:22:17.390
shoots in the air and the guy in Munich[br]shoots in the air, and they are supposed

0:22:17.390,0:22:22.490
to make the bullets hit in the[br]middle, over, let’s say Frankfurt.

0:22:22.490,0:22:25.720
Which they’re not going to manage.[br]And which is actually way too simple.

0:22:25.720,0:22:32.200
Because if the bullet is really[br]one centimeter in size,

0:22:32.200,0:22:37.360
then the equivalent distance that the two[br]shooters should be away from each other,

0:22:37.360,0:22:40.650
if we want to make it the same[br]difficulty as these protons,

0:22:40.650,0:22:45.050
would not be between Hamburg and Munich.[br]It would be from here to fucking Mars.

0:22:45.050,0:22:49.470
<i>laughter and applause</i>[br]I calculated that shit.

0:22:49.470,0:22:54.200
<i>applause</i>

0:22:54.200,0:22:57.650
We don’t even have rifles on Mars[br]anyway. <i>laughter</i>

0:22:57.650,0:23:01.690
So what we got is, we got large[br]bunches, very large bunches.

0:23:01.690,0:23:04.890
And in fact there’s 10^11[br]protons per bunch, which is

0:23:04.890,0:23:11.030
100 Billion. This is where I called Sagan[br]“ you going Millions of Millions“

0:23:11.030,0:23:15.120
Okay, so you got 100 Billion[br]protons in one bunch.

0:23:15.120,0:23:19.270
And the bunches go by one after the other.[br]Now, if you stand next to the LHC

0:23:19.270,0:23:23.160
and you were capable of observing these[br]bunches, you would see one fly by

0:23:23.160,0:23:28.170
every 25 nanoseconds. So you go “there’s[br]a bunch, now it’s 25 nanoseconds,

0:23:28.170,0:23:32.770
there is the next one”. And there’s about[br]7.5 meters between the bunches.

0:23:32.770,0:23:36.760
Now, 7.5 meters corresponds to[br]25 nanoseconds, you see that

0:23:36.760,0:23:42.940
the speed is very big and indeed[br]it’s almost the speed of light.

0:23:42.940,0:23:45.590
Which is just, we accelerate them[br]and at some point they just go

0:23:45.590,0:23:48.640
with the speed of light and we just push[br]up the energy, we don’t make them

0:23:48.640,0:23:53.750
go any faster actually. And if you[br]were to identify the bunches,

0:23:53.750,0:23:58.940
which actually you can, you would[br]see that there are 2800 bunches

0:23:58.940,0:24:02.890
going by; and then when[br]you have number 2809,

0:24:02.890,0:24:06.620
that’s actually the first one that you[br]counted which has come round again.

0:24:06.620,0:24:10.160
Per direction! So in total[br]we have over 5000 bunches

0:24:10.160,0:24:15.470
of 100 Billion protons each. So[br]that’s the beam we are dealing with.

0:24:15.470,0:24:19.610
Oh, and a funny thing: you get charged[br]particles moving, it’s actually a current,

0:24:19.610,0:24:22.680
right? In a wire you have[br]a current running through it,

0:24:22.680,0:24:27.150
there’s electrons moving or holes moving[br]and you get a current. If you were

0:24:27.150,0:24:31.800
to measure the current of the[br]LHC, it would be 0.6 milliamps,

0:24:31.800,0:24:34.330
which is a small current, but[br]we’re doing collisions anyway

0:24:34.330,0:24:38.270
and not power transmission,[br]so that’s fine. <i>laughter</i>

0:24:38.270,0:24:42.780
This is a diagram of what the actual[br]interaction point geometry looks like.

0:24:42.780,0:24:46.340
You get the beams from different[br]directions, think of it like the top one

0:24:46.340,0:24:50.010
coming from the right, the bottom[br]one coming from the left;

0:24:50.010,0:24:53.480
and they are kicked into intersecting[br]paths by magnets. You have

0:24:53.480,0:24:57.590
very complicated, very precise[br]magnetic fields aligning them,

0:24:57.590,0:25:01.850
so that they intersect. And it’s[br]actually a bit of a trying-out game.

0:25:01.850,0:25:05.970
I’ve heard this from[br]accelerator operators.

0:25:05.970,0:25:09.410
You shift the position of the beams[br]relative to each other by small amounts

0:25:09.410,0:25:12.880
and you just see where the collisions[br]happen. You go like: “Ah yeah, okay,

0:25:12.880,0:25:17.220
there’s lots of collisions, ah, now[br]they’re gone, I’m going back.”

0:25:17.220,0:25:20.440
And you do it like that. You can save the[br]settings and load them and calculate them

0:25:20.440,0:25:24.300
but it’s actually easier[br]to just try it out.

0:25:24.300,0:25:28.350
If we think of how much stuff we’ve[br]got going on: you got a packet,

0:25:28.350,0:25:31.240
a bunch of 100 Billion[br]protons coming one way,

0:25:31.240,0:25:35.100
you got another packet of 100 Billion[br]protons coming the other way.

0:25:35.100,0:25:39.640
Now the interaction point area is as small[br]as the cross section of a human hair.

0:25:39.640,0:25:43.270
You can see that, it’s one hundredth[br]of a square millimeter.

0:25:43.270,0:25:46.110
Now how many collisions do[br]you think we have? We’ve got…

0:25:46.110,0:25:48.120
Audience: Three![br]<i>Michael laughs</i>

0:25:48.120,0:25:51.850
Michael: …it’s actually not that bad.[br]We got about 20 in the LHC.

0:25:51.850,0:25:56.450
And the funny thing is, people[br]consider this a bit too much.

0:25:56.450,0:25:59.600
The effect is called pile-up. And the[br]bad thing about pile-up is you’ve got

0:25:59.600,0:26:03.590
beams intersecting, you’ve got bunches[br]‘crossing’ – that’s what we call it.

0:26:03.590,0:26:06.720
And there’s not just one collision which[br]you can analyze, there is a bunch of them,

0:26:06.720,0:26:10.110
around 20. And that makes that[br]more difficult for the experiments,

0:26:10.110,0:26:15.720
we’re going to see why. Well, and if we[br]have 20 collisions every bunch crossing

0:26:15.720,0:26:19.580
and the bunches come by every[br]25 nanoseconds, that gives us a total

0:26:19.580,0:26:24.690
of 600 Million collisions per[br]second. Per interaction point.

0:26:24.690,0:26:27.770
Which we don’t have just one of. We[br]have 4 experiments, each experiment

0:26:27.770,0:26:31.371
has its own interaction point. So[br]in total, we have about 2 Billion

0:26:31.371,0:26:36.660
proton-proton collisions happening[br]every second when the LHC is running.

0:26:36.660,0:26:39.580
Now let’s look at experiments.[br]<i>laughs</i>

0:26:39.580,0:26:44.070
Yeah, this is a photograph of one part of[br]the ATLAS experiment being transported.

0:26:44.070,0:26:47.690
And as for the scale of this thing, well,[br]in the physics community, we call this

0:26:47.690,0:26:53.700
a huge device.[br]<i>laughter</i>

0:26:53.700,0:26:57.150
I have a diagram of the experiment[br]where this is built in and

0:26:57.150,0:27:00.510
you’re going to recognize the part[br]which is the one I’ve circled there.

0:27:00.510,0:27:04.290
So the real thing is even bigger.[br]And down at the very bottom,

0:27:04.290,0:27:08.190
just to the center of the[br]experiment, there’s people.

0:27:08.190,0:27:12.860
Which if I check it like this,[br]they’re about 15 pixels high.

0:27:12.860,0:27:16.490
So that’s the scale of the experiment.

0:27:16.490,0:27:20.250
The experiment has the interaction point[br]at the center, so you got a beam line

0:27:20.250,0:27:23.570
coming in from the left, you got the other[br]beam line coming in from the right.

0:27:23.570,0:27:27.280
And in the very core of the experiment[br]is where the interactions,

0:27:27.280,0:27:31.140
where the collisions happen. And then[br]you got the experiment in layers,

0:27:31.140,0:27:35.240
like an onion, going around[br]them in a symmetrical way.

0:27:35.240,0:27:38.370
Inside you have a huge magnetic[br]field which is almost as big

0:27:38.370,0:27:42.470
as the curve magnets we were talking about[br]when I was describing the storage ring.

0:27:42.470,0:27:46.130
This is about 4 Teslas,[br]so it’s also a very big field.

0:27:46.130,0:27:50.160
But now we got a 4 Tesla field[br]not just over the beam pipe

0:27:50.160,0:27:54.340
which is about 5 centimeters in diameter,[br]but through the entire experiment;

0:27:54.340,0:27:58.080
and this thing is like 20-25 meters.[br]So you’ve got a 4 Tesla field

0:27:58.080,0:28:01.910
which should span more than 20 meters.

0:28:01.910,0:28:07.410
And, just for shits and giggles,[br]it’s got 3000 kilometers of cables.

0:28:07.410,0:28:11.060
Which is a lot; and if you just[br]pull some random plug

0:28:11.060,0:28:16.270
and don’t tell anyone which one it[br]was you’re making a lot of enemies.

0:28:16.270,0:28:19.980
So the innermost thing is what we[br]call the inner tracking. It is located

0:28:19.980,0:28:23.210
just centimeters off the beam line,[br]it’s supposed to be very very close to

0:28:23.210,0:28:26.290
where the actual interactions happen.

0:28:26.290,0:28:29.180
And this thing is made to leave the[br]particles undisturbed, they should just

0:28:29.180,0:28:32.590
fly trough this inner tracking detector.[br]And the detector will tell us

0:28:32.590,0:28:35.910
where they were, but not actually[br]stop them or deflect them.

0:28:35.910,0:28:40.050
This gives us precise location data,[br]as to how many particles there were,

0:28:40.050,0:28:44.030
what way they were flying,[br]and, from the curve,

0:28:44.030,0:28:47.570
what momentum they have. Outside[br]of that we’ve got calorimeters.

0:28:47.570,0:28:51.300
Now these are supposed to be stopping[br]the particles. A particle goes through

0:28:51.300,0:28:55.360
the inner tracking without being disturbed[br]but in the calorimeter it should stop.

0:28:55.360,0:28:58.970
And it should deposit all its energy there[br]and which is why we have to put around it

0:28:58.970,0:29:03.100
the inner tracking. You see, if we put the[br]calorimeter inside, it stops the particle,

0:29:03.100,0:29:07.770
outside of that nothing happens. So we[br]have the calorimeters outside of that.

0:29:07.770,0:29:12.070
And then we got these funny wing things[br]going on. That’s the muon detectors.

0:29:12.070,0:29:15.490
They are there for one[br]special sort of particle.

0:29:15.490,0:29:19.610
Out of the… 50, let’s say 60[br]– depends on the way you count –

0:29:19.610,0:29:22.860
elementary particles that we[br]have. These large parts are

0:29:22.860,0:29:26.250
just for the muons. Because the[br]muons have the property,

0:29:26.250,0:29:29.990
the tendency to go through all sorts of[br]matter undisturbed. So you just need to

0:29:29.990,0:29:33.270
throw a huge amount of matter[br]in the way of these muons, like:

0:29:33.270,0:29:36.750
“let’s have a brick wall and then[br]another one”. And then you

0:29:36.750,0:29:42.030
may be able to stop the muons,[br]or just measure them.

0:29:42.030,0:29:45.060
This is to give you an idea of the[br]complexity of the instrument

0:29:45.060,0:29:49.170
on the inside. This is the inner tracking[br]detector, it’s called a pixel detector;

0:29:49.170,0:29:52.730
and you see guys walking around in[br]protective suits. That is not for fun

0:29:52.730,0:29:56.920
or just for the photo, this is a very,[br]very precise instrument. But it’s sitting

0:29:56.920,0:30:00.100
inside this huge experiment which – again,

0:30:00.100,0:30:03.910
I calculated that shit – is about[br]as large as a space shuttle

0:30:03.910,0:30:07.420
and weighs as much as the[br]Eiffel Tower. And inside

0:30:07.420,0:30:12.030
they’ve got electronics, almost a ton[br]of electronics which is so precise

0:30:12.030,0:30:16.030
that it makes your smartphone[br]look like a rock. So there you go,

0:30:16.030,0:30:19.970
it’s a very, very complicated sort of[br]experiment. Let’s talk about triggering,

0:30:19.970,0:30:24.360
because as I said there’s 600 Million[br]events happening inside this.

0:30:24.360,0:30:27.600
That’s 40 Million bunch crossings.[br]Now: how are we going to analyze this?

0:30:27.600,0:30:31.720
Is there a guy writing everything[br]down? Obviously not.

0:30:31.720,0:30:35.540
So this experiment with all the tracking[br]and the calorimeters and the muons

0:30:35.540,0:30:39.800
and everything has about[br]100 Million electronic channels.

0:30:39.800,0:30:43.410
And one channel could be the measurement[br]of a voltage, or a temperature

0:30:43.410,0:30:47.330
or a magnetic field or whatever. So[br]we’ve got 100 Million different values,

0:30:47.330,0:30:52.540
so to speak. And that makes[br]about 1.5 Megabytes per crossing,

0:30:52.540,0:30:57.220
per every event readout. Which[br]gives us – multiplied by 40 Million –

0:30:57.220,0:31:01.260
gives us about 60 terabytes[br]of raw data per second.

0:31:01.260,0:31:05.610
That’s bad. I looked it up, I guess

0:31:05.610,0:31:10.340
the best RAM you can do is about[br]1 terabyte per second or something.

0:31:10.340,0:31:14.950
So we’re obviously not going to tackle[br]this by just putting in fast hardware,

0:31:14.950,0:31:18.690
because it’s not going[br]to be fast enough. Plus,

0:31:18.690,0:31:24.450
the reconstruction of an event is done[br]by about 5 Million lines of C++ code.

0:31:24.450,0:31:29.570
Programmed by some 2000-3000[br]developers around the world.

0:31:29.570,0:31:33.330
It simulates for one crossing[br]30 Million objects, which is

0:31:33.330,0:31:36.840
the protons and other stuff flying around.

0:31:36.840,0:31:44.410
And it is allocated to take 15 seconds[br]of one core’s computing time.

0:31:44.410,0:31:47.770
To calculate it all, you would[br]need about 600 million cores.

0:31:47.770,0:31:50.330
That’s not happening. I mean,[br]even if we took over the NSA

0:31:50.330,0:31:54.132
<i>laughter</i>[br]and used all of their data-centers

0:31:54.132,0:31:57.440
for LHC calculations, it still wouldn’t be[br]enough. So we have to do something

0:31:57.440,0:32:02.570
about this huge mass of data. And[br]what we do is, we put in triggers.

0:32:02.570,0:32:07.170
The trigger is supposed to reduce the[br]number of events that we look at.

0:32:07.170,0:32:10.830
The first level trigger looks at[br]every collision that happens.

0:32:10.830,0:32:13.840
And it’s got 25 nanoseconds[br]of time to decide:

0:32:13.840,0:32:17.410
Is this an interesting collision?[br]Is it not an interesting collision?

0:32:17.410,0:32:21.830
We tell it to eliminate[br]99.7% of all collisions.

0:32:21.830,0:32:26.480
So only every 400th collision[br]is allowed for this trigger to go:

0:32:26.480,0:32:30.280
“Oh, yeah, okay that looks interesting,[br]let’s give it to Level 2 trigger”.

0:32:30.280,0:32:34.150
So then we end up with about 100,000[br]events per second. Which get us

0:32:34.150,0:32:38.660
down to 150 Gigabytes per second. Now[br]we could handle this from the data flow,

0:32:38.660,0:32:43.450
but still we can’t simulate it. So[br]we’ve got another level trigger.

0:32:43.450,0:32:46.720
This is where the two[br]experiments at the LHC differ:

0:32:46.720,0:32:50.030
the CMS experiment has just a[br]Level 2 trigger; does it all there.

0:32:50.030,0:32:53.301
The ATLAS experiment goes the more[br]traditional way, it has a Level 2 trigger

0:32:53.301,0:32:57.500
and a Level 3 trigger. In the end these[br]combined have about 10 microseconds

0:32:57.500,0:33:01.450
of time, which is a bit more and it gives[br]them a chance to look at the events

0:33:01.450,0:33:05.920
more closely. Not just, let’s say:[br]“Was it a collision of 2 protons

0:33:05.920,0:33:09.300
or of 3 protons?”; “Were there[br]5 muons coming out of it

0:33:09.300,0:33:12.810
or 3 electrons and 2 muons?” This is[br]the sort of thing they’re looking at.

0:33:12.810,0:33:16.370
And certain combinations the triggers[br]will find interesting or not.

0:33:16.370,0:33:20.120
Let’s say 5 muons, I don’t give a shit[br]about that. “3 muons and 2 electrons?

0:33:20.120,0:33:23.480
Allright, I want to analyze it”. So[br]that’s what the trigger does.

0:33:23.480,0:33:27.640
Now this Level 2 and 3 trigger,[br]again, have to kick out about

0:33:27.640,0:33:31.070
99.9% of the events. They’re[br]supposed to leave us with

0:33:31.070,0:33:36.360
about 150 events per second. Which[br]gives a data volume of a measly

0:33:36.360,0:33:40.030
300 Megabytes per second and that’s[br]something we can handle. We push it

0:33:40.030,0:33:45.780
to computers all around the world.[br]And then we get the simulations going.

0:33:45.780,0:33:50.900
This is a display, this is[br]what you see in the media.

0:33:50.900,0:33:55.360
If you take one of these events – just[br]one of the interesting events which

0:33:55.360,0:34:00.740
actually reach the computers – because[br]those 40 million bunch crossings… well,

0:34:00.740,0:34:04.150
most of them don’t reach the computers,[br]they get kicked out by the triggers.

0:34:04.150,0:34:08.240
But out of the remaining 100 or 200[br]events per second, let’s say this is one.

0:34:08.240,0:34:12.849
It’s an actual event and it’s been[br]calculated into a nice picture here.

0:34:12.849,0:34:17.510
Now, normally they don’t do that, it’s[br]analyzed automatically by code

0:34:17.510,0:34:21.089
and it’s analyzed by the physics data.[br]And they only make these pretty pictures

0:34:21.089,0:34:25.339
if they want to show something to[br]the press. To the left you have

0:34:25.339,0:34:29.330
what’s called a Feynman Diagraph.[br]That’s just a fancy physical way

0:34:29.330,0:34:34.040
of saying what’s happening there. And[br]it involves the letter H on the left side,

0:34:34.040,0:34:37.180
which means there’s a Higgs involved.[br]Which is why this event was particularly

0:34:37.180,0:34:42.280
interesting to the people[br]analyzing the data at the LHC.

0:34:42.280,0:34:47.230
And you see a bunch of tracks, you see[br]the yellow tracks all curled up inside,

0:34:47.230,0:34:51.290
that’s a bunch of protons hitting[br]each other. The interesting thing is

0:34:51.290,0:34:55.710
what happens for example above[br]there with the blue brick kind of things.

0:34:55.710,0:35:00.050
There’s a red line going through[br]these bricks. This indicates a muon.

0:35:00.050,0:35:05.480
A muon which was created in[br]this event there in the center.

0:35:05.480,0:35:08.980
And it went out and the[br]bricks symbolize the way

0:35:08.980,0:35:13.140
the reaction was seen by the experiment.

0:35:13.140,0:35:16.880
There was actually just a bunch of bricks[br]lighting up. You got, I don’t know,

0:35:16.880,0:35:21.320
500 bricks around it and brick 237[br]says: “Whoop, there was a signal”.

0:35:21.320,0:35:24.300
And they go: “Allright, may have been[br]a muon moving through the detector”.

0:35:24.300,0:35:28.700
When you put it all together you[br]get an event display like this. Okay,

0:35:28.700,0:35:32.590
so we got to have computers analyzing[br]this. And with all the 4 experiments

0:35:32.590,0:35:36.570
running at the LHC, which is not just[br]CMS and ATLAS I mentioned but also

0:35:36.570,0:35:41.630
LHCb and ALICE, they produce about[br]25 Petabytes of data per year.

0:35:41.630,0:35:46.230
And this cannot be stored at CERN alone.[br]It is transferred to data centers

0:35:46.230,0:35:50.780
around the world by what is called[br]the LHC Optical Private Network.

0:35:50.780,0:35:55.530
They’ve got a network of fibers going from[br]CERN to other data-centers in the world.

0:35:55.530,0:36:00.430
And it consists of 11 dedicated[br]10-Gigabit-per-second lines

0:36:00.430,0:36:04.410
going from CERN outwards. If we[br]combine this, it gives us a little over

0:36:04.410,0:36:08.330
100 Gigabits of data[br]throughput, which is about

0:36:08.330,0:36:11.880
the bandwidth that this congress has.

0:36:11.880,0:36:14.560
Which is nice, but here it’s dedicated[br]to science data and not just porn

0:36:14.560,0:36:20.250
and cat pictures.[br]<i>laughter and applause</i>

0:36:20.250,0:36:23.930
<i>applause</i>

0:36:23.930,0:36:27.580
From there it’s distributed outwards[br]from these 11 locations to about

0:36:27.580,0:36:31.490
170 data centers in all the[br]world. And the nice thing is,

0:36:31.490,0:36:35.090
this data, these 25 Petabytes[br]per year, is available

0:36:35.090,0:36:38.310
to all the scientists working[br]with it. There’s about… well,

0:36:38.310,0:36:41.440
everybody can look at it, but there’s[br]about 3000 people in the world

0:36:41.440,0:36:45.270
knowing what it means. So all these[br]people have free access to the data,

0:36:45.270,0:36:48.900
you and I would have free access to the[br]data, just thinking it’s cool to have

0:36:48.900,0:36:53.260
a bit of LHC data on your harddrive maybe.[br]<i>laughter</i>

0:36:53.260,0:36:57.850
All in all, we have 250,000[br]cores dedicated to this task,

0:36:57.850,0:37:01.990
which is formidable. And about[br]100 Petabytes of storage

0:37:01.990,0:37:05.730
which is actually funny, because[br]25 Petabytes of data are accumulated

0:37:05.730,0:37:10.090
per year and the LHC has been[br]running for about 4 years.

0:37:10.090,0:37:13.600
So you can see that they buy the[br]storage as the machine runs. Because

0:37:13.600,0:37:17.540
100 Petabytes, okay, that’s what we have[br]so far. If we want to keep it running,

0:37:17.540,0:37:21.730
we need to buy more disks. Right! Now,

0:37:21.730,0:37:25.380
what does the philosoraptor[br]say about the triggers?

0:37:25.380,0:37:29.110
If the triggers are supposed to eliminate[br]those events which are irrelevant,

0:37:29.110,0:37:33.420
which is not interesting, well,[br]who tells them what’s irrelevant?

0:37:33.420,0:37:37.230
Or to put it in the terms[br]of Conspiracy-Keanu:

0:37:37.230,0:37:43.120
“What if the triggers throw away the[br]wrong 99.something % of events?”

0:37:43.120,0:37:48.230
I mean, if I say: “If there’s an event[br]with 5 muons going to the left,

0:37:48.230,0:37:52.500
kick it out!”. What if that’s actually[br]something that’s very, very interesting?

0:37:52.500,0:37:56.010
How should we tell? We need to[br]think about this very precisely.

0:37:56.010,0:37:59.320
And I’m going to tell you about[br]an example in history where

0:37:59.320,0:38:02.800
this went terribly wrong, at least for[br]a few years. We’re talking about

0:38:02.800,0:38:06.820
the discovery of the positron.[br]A positron is a piece of anti-matter;

0:38:06.820,0:38:10.770
it is the anti-electron. It was[br]theorized in 1928, when

0:38:10.770,0:38:15.440
theoretical physicist Dirac put up a bunch[br]of equations. And he said: “Right,

0:38:15.440,0:38:20.030
there should be something which is like[br]an electron, but has a positive charge.

0:38:20.030,0:38:22.470
Some kind of anti-matter.” Well,[br]that’s not what he said, but that’s

0:38:22.470,0:38:26.740
what he thought. But it was[br]only identified in 1931.

0:38:26.740,0:38:30.310
They had particle experiments back then,[br]they were seeing tracks of particles

0:38:30.310,0:38:34.090
all the time. But they couldn’t[br]identify the positron for 3 years,

0:38:34.090,0:38:37.210
even though it was there on paper.[br]So what happened? Well,

0:38:37.210,0:38:41.230
you see the picture on the left. This[br]is the actual, let’s say baby picture

0:38:41.230,0:38:44.460
of the positron. I’m going to[br]build up a scheme on the right

0:38:44.460,0:38:48.440
to show you a bit more, to[br]give you a better overview of

0:38:48.440,0:38:52.150
what we are actually talking about.[br]In the middle you’ve got a metal plate.

0:38:52.150,0:38:55.200
And then there’s a track which is bending[br]to the left, which is indicated here

0:38:55.200,0:39:01.890
by the blue line. Now if we analyze[br]this from a physical point of view,

0:39:01.890,0:39:05.270
it tells us that the particle[br]comes from below,

0:39:05.270,0:39:08.310
hits something in the metal plate[br]and then continues on to the top.

0:39:08.310,0:39:12.900
So the direction of movement[br]is from the bottom to the top.

0:39:12.900,0:39:17.310
The amount by which its curvature[br]reduces when it hits the metal plate

0:39:17.310,0:39:21.780
tells us it has about the mass of[br]an electron. Okay, so far so good.

0:39:21.780,0:39:26.020
But then it has a positive charge.[br]Because we know the…

0:39:26.020,0:39:29.580
we know the orientation of the magnetic[br]field. And that tells us: “Well,

0:39:29.580,0:39:33.280
if it bends to the left, it[br]must be a positive particle.”

0:39:33.280,0:39:37.020
So we have a particle with the mass of[br]an electron, but with a positive charge.

0:39:37.020,0:39:43.190
And people were like “Wat?”.[br]<i>laughter</i>

0:39:43.190,0:39:46.160
So then someone ingenious came[br]up and thought of a solution:

0:39:46.160,0:39:48.480
‘They developed the picture[br]the wrong way around!?’

0:39:48.480,0:39:52.300
<i>laughter and applause</i>

0:39:52.300,0:39:59.470
<i>applause</i>

0:39:59.470,0:40:02.780
It’s what they thought. Well it’s wrong,[br]of course, there’s such a thing as

0:40:02.780,0:40:08.500
a positron. And it’s like an electron,[br]but it’s positively charged. But…

0:40:08.500,0:40:13.520
to put it in a kind of summary maybe:[br]you can only discover that

0:40:13.520,0:40:17.180
which you can accept as a result.[br]This sounds like I’m Mahatma Gandhi

0:40:17.180,0:40:23.200
or something but it’s just what we call[br]science. <i>laughter</i>

0:40:23.200,0:40:27.740
Okay, so to recap: What have we[br]seen, what have we talked about?

0:40:27.740,0:40:32.210
We saw from the basic principle,[br]that if we have energy in a place,

0:40:32.210,0:40:36.190
then that can give rise to other forms of[br]matter, which I called ‘parts = a device’.

0:40:36.190,0:40:39.360
You got your little parts, you do[br]some stuff, out comes a device.

0:40:39.360,0:40:43.100
We have storage rings which give[br]a lot of energy to the particles

0:40:43.100,0:40:46.700
and in which they move around in huge[br]bunches. Billions of billions of protons

0:40:46.700,0:40:51.020
in a bunch and then colliding. Which[br]gives in the huge experiments

0:40:51.020,0:40:55.390
that we set up an enormous amount of data[br]ranging in the Terabytes per second

0:40:55.390,0:40:59.740
which we have to program triggers[br]to eliminate a lot of the events

0:40:59.740,0:41:03.750
and give us a small amount of data which[br]we can actually work with. And then

0:41:03.750,0:41:07.190
we have to pay attention to the[br]interpretation of data, so that

0:41:07.190,0:41:11.500
we don’t get a fuck-up like with the[br]positron. Which is a very hard job.

0:41:11.500,0:41:16.780
And I hope that I could give you[br]a little overview of how it’s fun.

0:41:16.780,0:41:20.250
And it’s not just about building[br]a big machine and saying:

0:41:20.250,0:41:24.180
“I’ve got the largest accelerator of[br]them all”. It’s a collaborative effort,

0:41:24.180,0:41:28.600
it’s literally thousands of people working[br]together and it’s not just about

0:41:28.600,0:41:32.390
two guys getting a Nobel Prize. You[br]see this picture on the top left, that’s

0:41:32.390,0:41:36.900
about 1000 people at CERN watching[br]the ceremony of the Nobel Prize

0:41:36.900,0:41:40.600
being awarded. Because everybody felt[br]there’s two people getting a medal

0:41:40.600,0:41:45.230
in Sweden, but it’s actually an[br]accomplishment… it’s actually an award for

0:41:45.230,0:41:49.190
everybody involved in this enormous thing.[br]And that’s what’s a lot of fun about it

0:41:49.190,0:41:53.991
and I hope I could share some of this[br]fascination with you. Thank you a lot.

0:41:53.991,0:42:19.000
<i>huge applause</i>

0:42:19.000,0:42:22.410
Before we get to Q&A, I’m going to be[br]answering questions that you may have.

0:42:22.410,0:42:25.560
My name is Michael, I’m @emtiu on[br]Twitter, I’ve got a DECT phone,

0:42:25.560,0:42:29.550
I talk about science, that’s[br]what I do. I hope I do it well.

0:42:29.550,0:42:32.210
And you can see the slides and[br]leave feedback for me please

0:42:32.210,0:42:36.770
in the event tracking system. And[br]tomorrow, if you have the time

0:42:36.770,0:42:39.720
you should go watch the “Desperately[br]seeking SUSY” talk which is going to be

0:42:39.720,0:42:43.480
talking about the theoretical side of[br]particle physics. Okay, that’s it from me,

0:42:43.480,0:42:46.540
now on to you.[br]Herald: Okay, if you have questions,[br]

0:42:46.540,0:42:50.240
please line up, there’s a mic there and[br]a mic there. And if you’re on the stream,

0:42:50.240,0:42:53.770
you can also use IRC and[br]Twitter to ask questions. So

0:42:53.770,0:42:55.820
I’m going to start here,[br]please go ahead.

0:42:55.820,0:43:00.490
Question: Thanks a lot, it was a very[br]fascinating talk, and nice to listen to.

0:43:00.490,0:43:04.030
My question is: Did HERA[br]ever suffer a quench event

0:43:04.030,0:43:08.030
in which the quench protection[br]system saved the infrastructure?

0:43:08.030,0:43:11.250
Michael: No, actually it didn’t. There[br]were tests where they provoked

0:43:11.250,0:43:15.040
a sort of quench event in order to[br]see if the protection worked. But

0:43:15.040,0:43:18.100
even if this test would have failed it[br]would not have been as catastrophic.

0:43:18.100,0:43:22.020
But there were failures in the[br]operation of the HERA accelerator

0:43:22.020,0:43:25.790
and there was one cryo failure. Which[br]is actually a funny story. Which is

0:43:25.790,0:43:30.140
where one part of the[br]helium tubing failed

0:43:30.140,0:43:33.680
and some helium escaped[br]from the tubing part

0:43:33.680,0:43:36.790
and went into the tunnel. Now what[br]happened was that the air moisture,

0:43:36.790,0:43:41.180
just the water in the[br]air froze at this point.

0:43:41.180,0:43:45.450
And the Technical Director of the HERA[br]machine told us this: at one point

0:43:45.450,0:43:49.020
he sat there with a screwdriver and[br]a colleague, picking off… the ice

0:43:49.020,0:43:53.120
off the machine for half the night before[br]they could replace this broken part.

0:43:53.120,0:43:56.480
So, yeah, cryo failures[br]are always a big pain.

0:43:56.480,0:44:01.790
Herald: Do we have questions[br]from the internet? …Okay.

0:44:01.790,0:44:04.490
Signal Angel: We have[br]one question that is:

0:44:04.490,0:44:09.500
“How are the particles[br]inserted into the accelerator?”

0:44:09.500,0:44:13.420
Michael: They mostly start[br]in linear accelerators.

0:44:13.420,0:44:19.310
Wait, we’ve got it here. So you[br]got the series of storage rings

0:44:19.310,0:44:23.780
there at the top in the middle and[br]you have one small line there.

0:44:23.780,0:44:26.900
That’s a linear accelerator. To get[br]protons is actually very easy.

0:44:26.900,0:44:30.400
You buy a bottle of hydrogen which[br]is just a simple gas you can buy.

0:44:30.400,0:44:34.380
And then you strip off the electrons.[br]You do this by ways of exposing them

0:44:34.380,0:44:38.280
to an electric field. And what you’re left[br]with is the core of the hydrogen atom.

0:44:38.280,0:44:42.670
And that’s a proton. Then you[br]accelerate the proton just a little bit

0:44:42.670,0:44:47.650
into the linear accelerator and from there[br]on it goes into the ring. So that means

0:44:47.650,0:44:52.780
basically at the start of these colliding[br]experiments is just a bottle of helium

0:44:52.780,0:44:56.590
that somebody puts in there. And[br]at the LHC it’s about, you know,

0:44:56.590,0:45:00.430
a gas bottle. It’s about this big and it[br]weighs a lot. At the LHC they use up

0:45:00.430,0:45:03.531
about 2 or 3 bottles a year for[br]all the operations, because

0:45:03.531,0:45:07.760
a bottle of hydrogen[br]has a lot of protons in it.

0:45:07.760,0:45:11.020
Herald: You please, over there.

0:45:11.020,0:45:15.120
Question: Actually I have[br]2 questions: One part is,

0:45:15.120,0:45:18.790
you said there are 2 beams[br]moving in opposite directions.

0:45:18.790,0:45:22.680
And you explained the way where you[br]switched polarity. How can this work

0:45:22.680,0:45:26.010
with 2 beams opposing each other?

0:45:26.010,0:45:31.160
Michael: That’s a good question. Now, if[br]I show you the picture of the cryo dipole,

0:45:31.160,0:45:36.980
you will see that these 2 beams[br]are not actually in the same tube.

0:45:36.980,0:45:40.650
There we go. You see a cryo dipole and

0:45:40.650,0:45:44.210
on the inside of this blue tube, you[br]see that there’s actually 2 lines.

0:45:44.210,0:45:47.760
You can’t see it very well but[br]there’s 2 lines. So they are

0:45:47.760,0:45:51.980
inside the same blue tube, but then[br]inside that is another small tube,

0:45:51.980,0:45:56.040
which has a diameter of just about[br]a Red Bull bottle. Say 5 or 6 centimeters

0:45:56.040,0:45:58.860
in diameter. And this is where the beam[br]happens. And they are just sitting

0:45:58.860,0:46:02.480
next to each other. So the beams[br]are always kept separate

0:46:02.480,0:46:06.310
except from the interaction points[br]where they should intersect.

0:46:06.310,0:46:10.090
And the acceleration happens[br]obviously also in separate cavities.

0:46:10.090,0:46:11.740
Herald: You had a second question?

0:46:11.740,0:46:15.890
Question: The second question is: The[br]experiments, where are they placed,

0:46:15.890,0:46:18.750
on the curve or on the acceleration part?

0:46:18.750,0:46:22.610
Michael: The interaction points are[br]placed between the acceleration

0:46:22.610,0:46:25.930
on the straight path. Because, again,[br]it’s much easier if you had the protons

0:46:25.930,0:46:30.130
going straight for 200m; then you[br]can more easily aim the beam.

0:46:30.130,0:46:34.240
If they come around the curve then they[br]have – you know they have a curve motion,

0:46:34.240,0:46:38.000
you need to cancel that. That[br]would be much more difficult.

0:46:38.000,0:46:39.410
Herald: And the left, please.

0:46:39.410,0:46:42.630
Question: Okay, so you got yourself[br]a nice storage ring and then

0:46:42.630,0:46:44.970
you connect it to the power plug[br]and then your whole country

0:46:44.970,0:46:48.120
goes dark. Where does the power come from?

0:46:48.120,0:46:52.510
Michael: Well, in terms of power[br]consumption of, let’s say

0:46:52.510,0:46:56.950
households, cities, or aluminum plants:

0:46:56.950,0:47:00.620
accelerators actually don’t[br]use that much power. I mean

0:47:00.620,0:47:03.370
most of us don’t run an aluminum[br]plant. So we’re not used to this

0:47:03.370,0:47:07.370
sort of power consumption. But’s it’s not[br]actually all that big. I can tell you about

0:47:07.370,0:47:11.290
the HERA accelerator that we had here[br]in Hamburg, which I told you is about

0:47:11.290,0:47:15.880
6.5 kilometers, not the 27, so you[br]can sort of extrapolate from that.

0:47:15.880,0:47:20.230
It used with the cryo and the[br]power current for the fields

0:47:20.230,0:47:25.030
and everything – it used about[br]30 MW. And 30 Megawatts is a lot,

0:47:25.030,0:47:29.270
but it’s not actually very much in[br]comparison to let’s say aluminum plants,

0:47:29.270,0:47:34.140
our large factories. But in fact,[br]the electricity cost is a big factor.

0:47:34.140,0:47:38.530
Now you see the LHC is located at the[br]border between Switzerland and France.

0:47:38.530,0:47:41.770
It gets most of its power from France.[br]

0:47:41.770,0:47:45.020
And you always have an annual shutdown of[br]the machine. You always have it off about

0:47:45.020,0:47:47.890
1 or 2 months of the year. Where you do[br]maintenance, where you replace stuff,

0:47:47.890,0:47:51.690
you check stuff. And they always[br]take care to have this shutdown

0:47:51.690,0:47:55.500
for maintenance in winter. Because[br]they get their power from France.

0:47:55.500,0:47:59.660
And in France many people use[br][electrical] power for heating.

0:47:59.660,0:48:03.670
There’s not Gas heating or Long[br]Distance heat conducting pipes

0:48:03.670,0:48:07.480
like we have in Germany e.g. The people[br]just use [electrical] power for heat.

0:48:07.480,0:48:11.500
And that means in winter the electricity[br]price goes up. By a large amount. So

0:48:11.500,0:48:15.410
they make sure that the machine is off in[br]winter when the electricity prices are up.

0:48:15.410,0:48:18.050
And it’s running in the summer where[br]it’s not quite as bad. So it’s a factor

0:48:18.050,0:48:21.890
if you run an accelerator. And you[br]should tell your local power company

0:48:21.890,0:48:25.130
if you’re about to switch it on![br]<i>laughter</i>

0:48:25.130,0:48:28.820
But actually, it won’t make the grid off,[br]even a small country like Switzerland

0:48:28.820,0:48:30.890
break down or anything.

0:48:30.890,0:48:35.150
Herald: Do we have more questions from[br]the internet? Internet internet, no,

0:48:35.150,0:48:39.970
no internet. Okay. Then just[br]go ahead, Firefox Girl.

0:48:39.970,0:48:43.000
Question (male voice): So you see a lot[br]of events. And I guess there’s many

0:48:43.000,0:48:48.210
wrong ones, too. How do you select if[br]an event you see is really significant?

0:48:48.210,0:48:51.470
Michael: Well, you have different kinds[br]of analysis. Like I told you there is

0:48:51.470,0:48:57.750
100 Mio. channels you can pick from.

0:48:57.750,0:49:01.960
With the simplest trigger that[br]you have, the Level 1 trigger,

0:49:01.960,0:49:06.560
it can’t look at the data in much[br]detail. Because it only has 25 ns.

0:49:06.560,0:49:09.910
But as you go higher up the chain,[br]as the events get more rare,

0:49:09.910,0:49:13.320
you can look at them more closely. And[br]what we end up in the end, these 100,

0:49:13.320,0:49:17.890
maybe 200 events per second, you can[br]analyze them very closely. And they get…

0:49:17.890,0:49:20.990
they get a full-out computation. You[br]can even make these pretty pictures

0:49:20.990,0:49:26.560
of some of them. And then it’s basically,[br]well, theoretical physicists’ work,

0:49:26.560,0:49:29.161
to look at them and say: “Well, this[br]might have been that process…”, but

0:49:29.161,0:49:33.060
still a lot of them get kicked out. When[br]the discovery of the Higgs particle

0:49:33.060,0:49:37.540
was announced, it was ca. 1 1/2 years ago…

0:49:37.540,0:49:42.470
Well, the machine had been running[br]for 2 1/2 years. And, like I told you,

0:49:42.470,0:49:46.390
there’s about 2 Billion proton collisions[br]per second. Now the number of events

0:49:46.390,0:49:51.150
that were relevant to the discovery[br]of the Higgs – the Higgs events –

0:49:51.150,0:49:54.890
it was not even 100.[br]Out of 2 Billion per second.

0:49:54.890,0:50:00.490
For 2 1/2 years. So you have to sort out[br]a lot. Because it’s very very, very rare.

0:50:00.490,0:50:03.400
And that’s just the work of[br]everybody analyzing, which is why

0:50:03.400,0:50:06.849
it’s a difficult task,[br]done by a lot of people.

0:50:06.849,0:50:08.380
Herald: The right, please.

0:50:08.380,0:50:13.060
Question: What I’m interested in: You[br]say ‘one year of detector running’.

0:50:13.060,0:50:16.460
How much time in this year does[br]this detector actually run…

0:50:16.460,0:50:18.140
…is it actually running?

0:50:18.140,0:50:21.560
Michael: Well, yeah, like I said, we[br]have the accelerator off for about

0:50:21.560,0:50:25.670
1 or 2 months. Then if something[br]goes wrong it will be off again.

0:50:25.670,0:50:29.450
But you want to keep it running[br]for as long as possible, which…

0:50:29.450,0:50:33.760
in the real world… let’s say it’s[br]9 months a year. That’s about it.

0:50:33.760,0:50:35.260
Question: Straight through?

0:50:35.260,0:50:38.570
Michael: Straight through – ah, well,[br]not in a row. But it’s always on

0:50:38.570,0:50:41.350
at least for a week. And then you[br]get maybe a small interruption

0:50:41.350,0:50:46.459
for a day or two, but you can also have[br]a month of straight operation sometimes.

0:50:46.459,0:50:47.810
Herald: Internet, please!

0:50:47.810,0:50:51.580
Signal Angel: Yeah, another question:[br]what would happen if they actually find

0:50:51.580,0:50:54.820
what you are looking for?[br]<i>Michael laughs</i>

0:50:54.820,0:50:58.690
Do we throw the LHC in the[br]dumpster or what do we do?

0:50:58.690,0:51:01.930
Michael: That’s a good question![br]It would be one hell-of-a waste

0:51:01.930,0:51:06.310
of a nice-looking tunnel! <i>laughs</i>[br]You might consider using it for

0:51:06.310,0:51:10.160
– I don’t know – maybe swimming[br]events, or bicycle racing.

0:51:10.160,0:51:13.050
Well, but actually that’s a very good[br]question because the tunnel

0:51:13.050,0:51:17.700
which the LHC sits in, this 27 km[br]tunnel, it was not actually dug,

0:51:17.700,0:51:21.220
it was not actually made just for the LHC.[br]There was another particle accelerator

0:51:21.220,0:51:25.620
inside before that. It had less energy,[br]because it didn’t accelerate protons

0:51:25.620,0:51:30.030
but just electrons and positrons.[br]That’s why the energy was a lot lower.

0:51:30.030,0:51:34.060
But they said: “Well, okay, we’re going[br]to build a very large accelerator,

0:51:34.060,0:51:38.200
does anyone have a[br]30 km tunnel, maybe?”

0:51:38.200,0:51:41.460
and then someone came up with:[br]“Yeah, well, we got this 27 km tunnel

0:51:41.460,0:51:45.450
where this LEP accelerator is sitting in.[br]And when it’s done with its operations

0:51:45.450,0:51:47.470
in…” – I don’t know, by that time,[br]let’s say in – “…10 years, we’re going

0:51:47.470,0:51:51.900
to shut it off. Why don’t we put the next[br]large accelerator in there?” So you try[br]

0:51:51.900,0:51:55.860
to reuse infrastructure, but of course[br]you can’t always do that. The next big,

0:51:55.860,0:52:00.470
the next huge accelerator, if we get the[br]money together as a science community,

0:52:00.470,0:52:03.540
because the politicians are[br]being a bitch about it…

0:52:03.540,0:52:06.920
if we get the money it’s going to be[br]the International Linear Collider.

0:52:06.920,0:52:10.900
And that’s supposed to have[br]100 km of particle tubes

0:52:10.900,0:52:16.240
and, well, you need to build[br]a new tunnel for that, obviously.

0:52:16.240,0:52:20.050
Question: First off, couldn’t[br]you use it in something

0:52:20.050,0:52:23.829
like material sciences, like[br]example with DESY?

0:52:23.829,0:52:27.240
Well okay, if you are done with[br]leptons you can still use it

0:52:27.240,0:52:30.590
for Synchrotron Laser[br]or something like this.

0:52:30.590,0:52:33.500
Michael: That was thought of. The HERA[br]accelerator at DESY was shut off

0:52:33.500,0:52:37.170
and people were thinking about if they[br]could put a Synchrotron machine inside it.

0:52:37.170,0:52:41.670
But the problem there is the HERA[br]accelerator is 25 m below the ground.

0:52:41.670,0:52:44.960
This is not enough space.[br]With particles accelerating

0:52:44.960,0:52:48.730
you just need a small tube. But for[br]Synchrotron experiments you need

0:52:48.730,0:52:51.810
a lot of space. So you would have[br]to enlarge the tunnel by a lot,

0:52:51.810,0:52:56.210
and this was not worth it, in the case of[br]the HERA accelerator. But interestingly,

0:52:56.210,0:53:00.000
one of the pre-accelerators of HERA,[br]one that was older is now used

0:53:00.000,0:53:04.100
for Synchrotron science, which is[br]PETRA. Which used to be just an

0:53:04.100,0:53:08.200
old pre-accelerator, and now it’s one of[br]the world’s leading Synchrotron machines.

0:53:08.200,0:53:11.960
So, yeah, you try to reuse things[br]because they were expensive.

0:53:11.960,0:53:15.630
Question: And may I just[br]ask another question?

0:53:15.630,0:53:21.830
You said you get… you use just the matter

0:53:21.830,0:53:25.420
from a bottle of hydrogen[br]or a bottle of helium.

0:53:25.420,0:53:29.980
Well, most helium or hydrogen is protons

0:53:29.980,0:53:33.850
or, in the case of helium, helium-4. But

0:53:33.850,0:53:37.350
you have a little bit helium-3 or deuterium.

0:53:37.350,0:53:41.150
And well, you are looking for[br]interesting things you don’t expect.

0:53:41.150,0:53:44.880
So how do you differentiate if it’s really

0:53:44.880,0:53:50.320
something interesting or: “Oh, one of[br]these damn deuterium nuclides, again!”

0:53:50.320,0:53:54.100
Michael: You don’t get wrong isotopes[br]because you just use a mass spectrometer

0:53:54.100,0:53:58.290
to sort them out. You have a magnetic[br]field. You know how large it is. And

0:53:58.290,0:54:03.380
the protons will go and land – let’s say[br]– 2 micrometers next to the deuterons,

0:54:03.380,0:54:07.230
and they just sort them out.

0:54:07.230,0:54:11.240
Question: I have 2 questions. One is:

0:54:11.240,0:54:15.100
I guess you mentioned that[br]basically once the experiment

0:54:15.100,0:54:19.550
runs at speed of light you[br]just put more energy into it.

0:54:19.550,0:54:22.380
But what is actually the meaning[br]of the energy that you put into it?

0:54:22.380,0:54:25.230
What does it change in the experiment?[br]Like the Higgs was found

0:54:25.230,0:54:28.260
at a particular electron volt…

0:54:28.260,0:54:33.410
Michael: Yeah, it was[br]found at 128 GeV. Well,

0:54:33.410,0:54:37.610
it’s more of a philosophical question.[br]There is a way of interpreting

0:54:37.610,0:54:41.480
the equations of special relativity where[br]you say that, when you don’t increase

0:54:41.480,0:54:45.930
the velocity you increase the mass.[br]But that’s just a way of looking at it.

0:54:45.930,0:54:50.260
It’s more precise and it’s more[br]simple to say: you raise the energy.

0:54:50.260,0:54:53.130
And at some low energies that[br]means that you raise the velocity.

0:54:53.130,0:54:55.890
And at some high energies it means[br]the velocity doesn’t change anymore.

0:54:55.890,0:55:00.110
But overall you add more energy.[br]It’s one of the weird effects

0:55:00.110,0:55:07.769
of special relativity and there[br]is no very nice explanation.

0:55:07.769,0:55:10.950
Question: Let’s assume there is[br]an asteroid pointing to earth.

0:55:10.950,0:55:14.410
<i>Michael laughs</i>[br]Could you in theory point this thing

0:55:14.410,0:55:17.980
on the asteroid and destroy it,[br]or would it be too weak?

0:55:17.980,0:55:19.830
<i>laughter</i>

0:55:19.830,0:55:24.290
<i>applause</i>

0:55:24.290,0:55:26.750
Michael: I’m going to help you out.[br]Because it wouldn’t actually work

0:55:26.750,0:55:30.430
because between the accelerator and the[br]asteroid there’s the earth atmosphere.

0:55:30.430,0:55:33.750
And that would stop all the particles.[br]But even if there were no atmosphere:

0:55:33.750,0:55:37.690
no, it would be much too weak. Well,

0:55:37.690,0:55:40.620
you’d have to keep it up for a long time[br]at least. There was this one accident

0:55:40.620,0:55:46.210
at the HERA accelerator where the[br]beam actually went off its ideal path

0:55:46.210,0:55:50.300
and it went some 2 or 3 cm[br]next to where it should be.

0:55:50.300,0:55:54.550
And it hit a block of lead – just,[br]you know, the heavy metal lead –

0:55:54.550,0:55:59.070
and the beam shot into this[br]lead thing and the entire beam,

0:55:59.070,0:56:02.960
which was a couple of Billions of[br]protons, was deposited into this lead

0:56:02.960,0:56:06.670
and some kilograms of lead[br]evaporated within microseconds

0:56:06.670,0:56:10.630
and there was a hole like pushed by[br]a pencil through these lead blocks.

0:56:10.630,0:56:15.160
So, yeah, it does break stuff apart. But[br]even if you managed to hit the asteroid

0:56:15.160,0:56:19.339
you would make a very small hole.[br]But you wouldn’t destroy it.

0:56:19.339,0:56:26.800
It would be a nice-looking asteroid then.[br]<i>laughter</i>

0:56:26.800,0:56:30.820
Question: Before you turned on the LHC[br]the popular media was very worried

0:56:30.820,0:56:34.220
that you guys were going[br]to create any black holes.

0:56:34.220,0:56:39.080
Did you actually see any black holes[br]passing by? <i>Michael laughs</i>

0:56:39.080,0:56:43.080
Michael: Well, there may have been[br]some, but they were small, and

0:56:43.080,0:56:48.810
they were insignificant. The interesting[br]thing is… sorry, I’m going to recap, yeah.

0:56:48.810,0:56:52.010
The interesting thing is that whatever[br]we can do with the LHC – where

0:56:52.010,0:56:56.869
we make particles have large energies[br]and then collide – is already happening!

0:56:56.869,0:57:00.830
Because out in space there is black[br]holes with enormous magnetic fields

0:57:00.830,0:57:04.450
and electrical fields. And these[br]black holes are able to accelerate

0:57:04.450,0:57:08.320
electrons to energies much, much[br]higher than anything we can produce

0:57:08.320,0:57:12.340
in any accelerator. The LHC[br]looks like a children’s toy

0:57:12.340,0:57:16.370
in comparison to the energies that[br]a black hole acceleration can reach. And

0:57:16.370,0:57:21.170
the particles which are accelerated in[br]these black holes hit earth all the time.

0:57:21.170,0:57:24.630
Not a lot, let’s say one of these[br]super-energetic particles they come around

0:57:24.630,0:57:28.840
about once a year for every[br]square kilometer of earth.

0:57:28.840,0:57:31.470
But still, they’ve been hitting[br]us for Millions of years.

0:57:31.470,0:57:34.900
And if a high-energy particle[br]collision of this sort were able

0:57:34.900,0:57:39.140
to produce a black hole that swallows[br]up the earth it would be gone by now.

0:57:39.140,0:57:45.499
So: won’t happen.[br]<i>applause</i>

0:57:45.499,0:57:48.190
Question: Maybe more interesting[br]for this crowd: you talked about

0:57:48.190,0:57:52.580
the selection process of the events.

0:57:52.580,0:57:56.750
So I guess these parameters[br]are also tweaked to kind of

0:57:56.750,0:58:00.430
narrow down like what[br]a proper selection procedure.

0:58:00.430,0:58:04.040
Is there any kind of machine[br]learning done on this to optimize?

0:58:04.040,0:58:07.230
Michael: Not that I know of. But there is[br]a process which is called ‘Minimum Bias

0:58:07.230,0:58:11.690
Data Collection’. Where you[br]actually bypass all the triggers

0:58:11.690,0:58:15.290
and you select a very small portion[br]of events without any bias.

0:58:15.290,0:58:19.990
You just tell the trigger: “Take[br]every 100 Billionth event”

0:58:19.990,0:58:22.940
and you just pass it through no matter[br]what you think. Even if you think

0:58:22.940,0:58:28.150
it’s not interesting, pass it through.[br]This goes into a pool of Minimum Bias Data

0:58:28.150,0:58:32.830
and these are analyzed especially in order[br]to see the actual trigger criteria

0:58:32.830,0:58:37.230
are working well. So yeah,[br]there is some tweaking. And

0:58:37.230,0:58:41.230
even for old machines[br]we have data collected

0:58:41.230,0:58:44.910
and sometimes we didn’t know what we[br]were looking for. And some 20 years later

0:58:44.910,0:58:48.800
some guy comes up and says: “Well,[br]we had this one accelerator way back.

0:58:48.800,0:58:52.249
There may have been this and that[br]reaction. Which we just theorize about.

0:58:52.249,0:58:56.200
So let’s look at the old data and see[br]if we see anything of that in there

0:58:56.200,0:58:59.420
now, because it’s limited because[br]it goes through all the filters”.

0:58:59.420,0:59:03.600
You can’t do this all the time with[br]great success. But sometimes,

0:59:03.600,0:59:06.810
in very old data you find new[br]discoveries. Because back then

0:59:06.810,0:59:11.980
people weren’t thinking about looking[br]for what we are looking now.

0:59:11.980,0:59:16.470
Question: I always asked myself about[br]repeatability of those experiments.

0:59:16.470,0:59:20.480
Seeing as the LHC is the biggest one[br]around there, so there’s no one out there

0:59:20.480,0:59:23.320
who can actually repeat the[br]experiment. So how do we know

0:59:23.320,0:59:26.440
that they actually exist, those particles?

0:59:26.440,0:59:30.150
Michael: That’s a very good question.[br]I told you that there is 2 main

0:59:30.150,0:59:33.940
large experiments. Which is the CMS[br]experiment and the ATLAS experiment.

0:59:33.940,0:59:39.020
Now these both sit at the same ring.[br]They have some 10 km between them

0:59:39.020,0:59:41.740
because they’re on opposite ends[br]of the ring. But still, obviously,

0:59:41.740,0:59:46.690
they’re on the same machine. But these 2[br]groups, the ATLAS and the CMS experiment,

0:59:46.690,0:59:51.910
operate completely separately. It’s not[br]the same people, not the same hardware,

0:59:51.910,0:59:55.250
not the same triggers,[br]not even the same designs.

0:59:55.250,0:59:58.760
They build everything up from scratch,[br]separate from each other. And

0:59:58.760,1:00:02.700
it’s actually funny because when you[br]look at a conference and here is CMS

1:00:02.700,1:00:05.570
presenting their results and here is[br]ATLAS presenting their results,

1:00:05.570,1:00:08.300
they pretend like the other[br]experiment is not even there.

1:00:08.300,1:00:11.730
And that’s the point of it: they’re[br]not angry at each other. It must be

1:00:11.730,1:00:16.070
2 separate experiments because obviously[br]you can’t build a second accelerator.

1:00:16.070,1:00:18.720
So you try to have redundancy in order

1:00:18.720,1:00:22.900
for one experiment to confirm[br]what the other finds.

1:00:22.900,1:00:27.900
Herald: Okay. It’s midnight[br]and we’re out of time.

1:00:27.900,1:00:31.400
So please thank our awesome speaker![br]<i>applause</i>

1:00:31.400,1:00:39.163
<i>Subtitles created by c3subtitles.de[br]in the year 2016. Join and help us!</i>