WEBVTT

00:00:05.755 --> 00:00:08.954
Good morning everyone!
I'll speak about time measurement.

00:00:08.955 --> 00:00:14.508
To talk about measuring time,
I will ask an obvious question, which is:

00:00:14.509 --> 00:00:16.059
what time is it?

00:00:16.059 --> 00:00:19.041
It may seem like a trivial question,

00:00:19.042 --> 00:00:23.262
but if we performed an experiment today,
and had everyone looking at their watch,

00:00:23.263 --> 00:00:26.989
everyone would have a different
time from his or her neighbor.

00:00:27.509 --> 00:00:30.962
Particularly, you would have
a different time from the one

00:00:30.963 --> 00:00:33.970
on this rather strange clock.

00:00:33.971 --> 00:00:37.982
This is a clock that gives you
the atomic time.

00:00:37.983 --> 00:00:40.912
I wasn't able to bring an atomic clock
along with me today.

00:00:40.913 --> 00:00:43.639
The atomic time is built
at the Paris Observatory.

00:00:43.640 --> 00:00:48.341
It is broadcast by radio waves,
and here, we receive that information.

00:00:48.390 --> 00:00:52.897
If you compare the time
- that's the date, November 27 -

00:00:52.956 --> 00:00:56.647
to the time here,
you see there's a difference.

00:00:56.744 --> 00:01:01.527
My presentation is not only going
to explain to you how to reset the clock.

00:01:02.490 --> 00:01:07.257
I am going to explain to you
how to measure time with high precision.

00:01:07.258 --> 00:01:10.372
You will see that the magnitudes
in the precision are astonishing.

00:01:10.960 --> 00:01:14.954
To provide you with some original
and fascinating applications,

00:01:15.753 --> 00:01:21.134
I'll start with a very simple thing.
To measure time, we use a ruler.

00:01:21.135 --> 00:01:24.632
I use the analogy between a time ruler
and a spatial ruler.

00:01:24.633 --> 00:01:28.754
To measure a distance, you take
a ruler that's been calibrated.

00:01:28.755 --> 00:01:32.692
You'll count the number of graduations,
for example, in centimeters,

00:01:32.693 --> 00:01:34.387
so if you count five graduations,

00:01:34.817 --> 00:01:37.241
assuming that one graduation
equals one centimeter,

00:01:37.242 --> 00:01:40.002
you will deduce
a length of five centimeters.

00:01:40.003 --> 00:01:44.250
With time, it's going to be the same.
We are going to use a temporal ruler.

00:01:44.501 --> 00:01:47.569
A temporal ruler can take
the form of an oscillator.

00:01:47.781 --> 00:01:50.759
An oscillator is a physical device

00:01:50.760 --> 00:01:53.417
that gives you
a periodic signal with timing

00:01:53.418 --> 00:01:56.507
- whose parameter is reproduced
in a periodic way with time.

00:01:56.508 --> 00:02:00.503
I brought one with me,
Professor Calculus's pendulum.

00:02:00.504 --> 00:02:01.897
This is an oscillator.

00:02:01.907 --> 00:02:05.857
As you see, we can count time
by counting the number of round trips.

00:02:05.867 --> 00:02:12.848
If we say a round trip takes one second,
we can count one, two, the time passing.

00:02:13.292 --> 00:02:15.027
Using this time ruler,

00:02:15.161 --> 00:02:17.766
which elementary calibration
- called period -

00:02:17.800 --> 00:02:19.628
we are able to measure time.

00:02:19.678 --> 00:02:22.772
We can imagine
that if we want a better precision,

00:02:22.972 --> 00:02:25.517
we will need more graduations.

00:02:25.577 --> 00:02:29.441
It is equivalent
to what we have with a ruler.

00:02:29.491 --> 00:02:32.549
If your ruler, instead of being
calibrated in centimetres,

00:02:32.592 --> 00:02:34.357
is calibrated in millimeters,

00:02:34.358 --> 00:02:37.268
and that you measure 51
- tiny millimeter calibrations -

00:02:37.418 --> 00:02:40.891
you can assume that your length
is 5.1 centimeters.

00:02:41.111 --> 00:02:43.346
It is exactly the same for measuring time.

00:02:43.526 --> 00:02:47.540
If you take a faster oscillator
than the one I showed you,

00:02:47.670 --> 00:02:50.897
therefore, having a shorter period,
meaning a frequency,

00:02:50.898 --> 00:02:53.152
a greater number of pulses per second

00:02:53.442 --> 00:02:56.660
you will achieve a much better
resolution in time measure.

00:02:56.900 --> 00:03:00.504
This is what has led research

00:03:00.514 --> 00:03:04.025
since the invention
of time measure with oscillators.

00:03:04.045 --> 00:03:08.304
Typically, given the magnitudes,
you take a mechanical clock,

00:03:08.314 --> 00:03:10.360
it doesn't need to be a Swiss clock,

00:03:10.361 --> 00:03:14.354
just a mechanical one,
similar to my little oscillator pendulum,

00:03:14.364 --> 00:03:16.454
it beats at a pulse per second,

00:03:16.474 --> 00:03:20.199
so you don't have a huge precision,
when you want to measure a second.

00:03:20.200 --> 00:03:23.670
If you take an oscillator
that you carry with you,

00:03:23.690 --> 00:03:25.985
like your watch or your mobile phone.

00:03:25.986 --> 00:03:29.209
It has a quartz oscillator
using piezoelectricity,

00:03:29.210 --> 00:03:33.730
it has the shape of a vibrating diapason
at a millimeter scale.

00:03:33.731 --> 00:03:37.723
It is going to beat
at 32,768 pulses per second.

00:03:37.733 --> 00:03:41.950
You are going to cut one second,
into 32,768 small elementary periods.

00:03:42.130 --> 00:03:43.890
Why such a weird number?

00:03:43.980 --> 00:03:46.777
Because it is easy to divide
by two, 15 times,

00:03:46.907 --> 00:03:51.899
to get to one pulse per second,
and get the tick-tock of your watch.

00:03:52.059 --> 00:03:57.986
And if we go to the ultimate,
to the fastest oscillators known today,

00:03:58.030 --> 00:04:02.064
called lasers - oscillators,
in the optics field -

00:04:02.076 --> 00:04:06.189
you see that a laser is an oscillator
that gives you an electromagnetic wave

00:04:06.359 --> 00:04:08.300
beating extremely fast until it cuts

00:04:08.513 --> 00:04:12.775
your second to 500,000
billions little pulses.

00:04:13.159 --> 00:04:16.555
You can see the elementary calibration
is extremely small

00:04:16.815 --> 00:04:20.459
We will count 500,000 billion and say
that one second has passed,

00:04:20.459 --> 00:04:22.882
we will count again 500,000 billion etc.

00:04:22.982 --> 00:04:24.892
You see, in measuring time,

00:04:24.933 --> 00:04:30.420
having a frequency as high as possible
is what gives you greater precision.

00:04:30.421 --> 00:04:34.502
We can say we have
almost solved the problem.

00:04:34.912 --> 00:04:39.582
Not at all! In fact, what confidence
can we have in this measure?

00:04:39.842 --> 00:04:42.029
I go back to the example with two rulers.

00:04:42.219 --> 00:04:46.464
You buy two rulers, in two different
places, different countries, and measure.

00:04:46.484 --> 00:04:48.333
Make the experiment and you will see.

00:04:48.334 --> 00:04:51.674
It might not be as blatant,
but you will see it works very well.

00:04:52.304 --> 00:04:55.548
For the same length,
you won't have the same calibrations.

00:04:55.608 --> 00:04:57.099
So which should we trust?

00:04:57.159 --> 00:05:00.398
Which rule should we trust,
which one has the right measure?

00:05:00.758 --> 00:05:03.285
With the oscillators,
whilst measuring time,

00:05:03.475 --> 00:05:04.927
we face the same problem.

00:05:05.377 --> 00:05:08.099
Your oscillator,
two different oscillators,

00:05:08.279 --> 00:05:09.785
won't give you exactly

00:05:09.895 --> 00:05:12.576
the same number of period
for measuring a given duration;

00:05:12.577 --> 00:05:14.669
each little calibration is different.

00:05:14.789 --> 00:05:16.890
Or, if you take an oscillator,

00:05:17.137 --> 00:05:19.923
according to the place,
or the moment you use it,

00:05:20.053 --> 00:05:22.025
it won't give you the same measurement.

00:05:22.295 --> 00:05:24.272
For example, the pendulum I showed you,

00:05:24.612 --> 00:05:28.409
whether you use it
at the equator or at the poles,

00:05:28.429 --> 00:05:32.991
because the oscillation's period depends
on the gravitational force,

00:05:33.411 --> 00:05:36.515
after one year, you will have
around two days' difference

00:05:36.705 --> 00:05:38.531
between the measures... that's huge.

00:05:38.581 --> 00:05:42.487
Maybe not in everyday life,
though two days is significant.

00:05:42.537 --> 00:05:44.858
But for the applications
I am going to show you,

00:05:45.038 --> 00:05:46.934
it is something very annoying.

00:05:47.034 --> 00:05:48.808
How do we solve this problem?

00:05:48.888 --> 00:05:50.993
This is where we build atomic clocks.

00:05:51.233 --> 00:05:53.795
The atom is the solution to this problem,

00:05:53.975 --> 00:05:56.334
since the atom is going
to be our reference.

00:05:56.464 --> 00:06:00.032
What happens in an atomic clock?
It is relatively simple.

00:06:00.112 --> 00:06:04.712
You still have an oscillator,
but we will compare its frequency

00:06:04.852 --> 00:06:09.658
to one that is infinitely stable,
universal, and extremely well-known.

00:06:09.678 --> 00:06:13.810
It is the frequency of resonance
to hop from one atomic level to the next.

00:06:13.811 --> 00:06:16.431
Why is this atomic frequency
very well known?

00:06:16.432 --> 00:06:21.260
Well, because quantum mechanics
tells us that the states of energy

00:06:21.261 --> 00:06:26.037
that is, energy levels
between which atoms transit,

00:06:26.038 --> 00:06:27.682
these states of energy have

00:06:27.683 --> 00:06:30.720
extremely stable
and well determined values.

00:06:30.721 --> 00:06:34.082
Thus, the frequency of resonance
to go from a level to another,

00:06:34.083 --> 00:06:37.123
will too be extremely well fixed.

00:06:37.336 --> 00:06:42.027
Here you have a photo of the atomic clock

00:06:42.397 --> 00:06:44.397
which is at the Paris Observatory.

00:06:44.912 --> 00:06:48.275
Today, using atoms
which are a bit specific,

00:06:48.635 --> 00:06:50.578
since they are cold atoms.

00:06:50.579 --> 00:06:53.982
We cool them by laser,
to extremely low temperatures,

00:06:53.983 --> 00:06:56.092
and we trap them with the laser light,

00:06:56.093 --> 00:06:59.510
using optical oscillators
beating extremely fast.

00:06:59.512 --> 00:07:02.344
We manage to have a precision
in measuring time

00:07:02.354 --> 00:07:04.499
which is very impressive,

00:07:04.500 --> 00:07:08.681
since a clock amongst the best
in the world today,

00:07:09.031 --> 00:07:12.117
will only go off one second
after 3 billion years.

00:07:12.297 --> 00:07:16.116
In other words, we are capable of giving

00:07:16.117 --> 00:07:19.770
the value of a small graduation,
or of the frequency of the clock,

00:07:19.771 --> 00:07:21.836
with 17 digits after the decimal point.

00:07:21.837 --> 00:07:23.959
As you see, it is an application,

00:07:23.960 --> 00:07:28.821
that is very highly impressive,
a very high level of stability

00:07:28.822 --> 00:07:31.212
and which further more
has many many applications.

00:07:31.234 --> 00:07:33.839
The first application
is the speaking clock.

00:07:33.840 --> 00:07:37.047
It is an application
which generally speaks to the public.

00:07:37.048 --> 00:07:39.201
Where does the speaking clock come from?

00:07:39.202 --> 00:07:42.005
It was created at
the Paris Observatory in 1933,

00:07:42.006 --> 00:07:45.329
At that time, it was the role
of astronomers to give the time.

00:07:45.330 --> 00:07:46.978
It was not atomic physics yet.

00:07:46.979 --> 00:07:49.955
The line of the Paris Observatory
always took care of it,

00:07:49.956 --> 00:07:52.320
Because everybody
called Ernest Esclangon,

00:07:52.321 --> 00:07:55.632
who was the Director of the Observatory,
to get the time.

00:07:55.633 --> 00:08:01.640
Ernest Esclangon had the idea
of developing this speaking clock.

00:08:01.642 --> 00:08:04.177
There has been several generations
of speaking clock.

00:08:04.178 --> 00:08:07.431
Today, the speaking clock presented
here gives you the time

00:08:07.432 --> 00:08:10.615
with 50 milliseconds of uncertainty.

00:08:10.616 --> 00:08:14.769
As metrology is an experimental science,
we will call the speaking clock.

00:08:14.770 --> 00:08:18.191
I have the authorization to keep it
connected, it is my privilege!

00:08:18.192 --> 00:08:22.575
It is always a risk, in experiments:
it might not work.

00:08:22.578 --> 00:08:26.512
Clock: It is 5 pm, 7 minutes 10 seconds.

00:08:27.642 --> 00:08:29.590
ND: We are going to wait just a little,

00:08:29.591 --> 00:08:32.669
But you can see the red lights there.

00:08:32.671 --> 00:08:37.258
Speaking clock: It is 17 hours,
7 minutes, and 20 seconds.

00:08:37.259 --> 00:08:39.453
ND: There, it works! Thank you!

00:08:40.904 --> 00:08:42.335
(Applause)

00:08:44.475 --> 00:08:48.340
This application might be harmless,
but it is important especially

00:08:48.341 --> 00:08:51.309
at the time of changes
between summer time and winter time.

00:08:51.310 --> 00:08:56.943
If we want to give the time
in a more precise way,

00:08:56.944 --> 00:09:03.611
we can also use Internet,
telecommunication satellites, or GPS, etc.

00:09:03.612 --> 00:09:05.479
This is the first application.

00:09:05.480 --> 00:09:08.267
The second application,
which is very fashionable,

00:09:08.268 --> 00:09:09.834
and is also very important

00:09:09.835 --> 00:09:14.059
is the use the atomic clocks
to test Einstein's law of relativity

00:09:14.060 --> 00:09:17.669
that tells you, for the last 100 years,
that time is not absolute.

00:09:17.670 --> 00:09:20.313
That is, if you take identical clocks,

00:09:20.314 --> 00:09:22.927
and you put them
in different frames of reference,

00:09:22.928 --> 00:09:25.577
which move relative to one another,

00:09:25.578 --> 00:09:28.130
or which have different
environmental parameters,

00:09:28.131 --> 00:09:30.558
you will find and measure differences

00:09:30.559 --> 00:09:33.324
between the times
and frequencies of the clocks.

00:09:33.325 --> 00:09:37.889
This non-absolute character of time
is already tested on the ground.

00:09:37.890 --> 00:09:40.453
We are going to test it
extremely precisely in space,

00:09:40.454 --> 00:09:44.797
by installing in a few years,
an ultra-precise clock

00:09:44.798 --> 00:09:47.329
aboard the International Space Station.

00:09:47.330 --> 00:09:51.596
And by comparing the time
and frequency of this clock in space

00:09:51.597 --> 00:09:55.184
with the time and frequency
of clocks situated all around the Earth,

00:09:55.185 --> 00:09:57.922
it will become possible
to validate Einstein's theory.

00:09:57.923 --> 00:09:59.851
Knowing that all modern theories

00:09:59.852 --> 00:10:02.471
predict a violation of Einstein's theory.

00:10:02.472 --> 00:10:04.930
So there is a real
scientific benefit in doing that.

00:10:04.931 --> 00:10:09.668
We are going to test various aspects
of general relativity.

00:10:09.669 --> 00:10:14.525
For instance, we will test
a rather interesting property, which says:

00:10:14.526 --> 00:10:16.933
fundamental constants are constant.

00:10:16.934 --> 00:10:18.249
This is not trivial!

00:10:18.250 --> 00:10:21.839
In physics, a whole set of constants
is supposed to be constant.

00:10:21.840 --> 00:10:26.996
In fact, all modern theories predict
these constants vary in time and in space.

00:10:27.142 --> 00:10:29.355
We will be able to test this precisely.

00:10:29.356 --> 00:10:32.241
We will also test an original effect
of general relativity,

00:10:32.242 --> 00:10:35.709
time passes at a different rhythm
according to the altitude.

00:10:36.992 --> 00:10:39.713
For instance, you, who are
sitting in the first row,

00:10:39.714 --> 00:10:43.770
you do not age at the same speed
as those sitting in the last row,

00:10:43.771 --> 00:10:46.085
since you are sitting
at different altitudes.

00:10:46.195 --> 00:10:49.679
But to reassure you,
on the length of my presentation

00:10:49.689 --> 00:10:52.886
the difference of aging
is around one picosecond.

00:10:52.887 --> 00:10:56.960
10 to the power of -12 seconds,
a billionth of a billionth of a second.

00:10:56.961 --> 00:11:00.070
So, there is no need to run up and down,
remain in your seats!

00:11:00.071 --> 00:11:03.874
We will also test
the speed of light is constant.

00:11:03.875 --> 00:11:07.250
This is an extremely strong postulate
of special relativity:

00:11:07.251 --> 00:11:10.513
the speed of light is independent
from the frame of reference

00:11:10.514 --> 00:11:12.414
against which the measure is taken.

00:11:12.415 --> 00:11:15.400
This is an extremely important property,

00:11:15.401 --> 00:11:20.555
which is used to measure distances
from time measurements.

00:11:20.556 --> 00:11:25.906
If you want to measure a distance
you use a signal which will propagate,

00:11:25.913 --> 00:11:30.900
and by knowing the time of propagation,
knowing the speed of propagation,

00:11:30.901 --> 00:11:35.230
which is the case with the speed of light,
you can infer the distance.

00:11:35.231 --> 00:11:36.772
One could say there is no need

00:11:36.773 --> 00:11:40.298
for an ultra-stable clock to do that.
But yes, there is a need for it.

00:11:40.299 --> 00:11:43.011
Light goes fast, at 300,000 km per second.

00:11:43.012 --> 00:11:46.983
If you make a nanosecond error,
a billionth of a second,

00:11:46.984 --> 00:11:49.375
you are wrong by 30 centimeters.

00:11:49.376 --> 00:11:55.252
Typically, this kind of application,
measuring distances from measures of time,

00:11:55.253 --> 00:11:57.752
is used to measure Earth-Moon distance.

00:11:57.753 --> 00:11:59.651
By sending impulses to the Moon,

00:11:59.652 --> 00:12:04.447
which are reflected by retro-reflectors
installed by the Apollo missions

00:12:04.448 --> 00:12:07.553
to measure the Earth-Moon distance
better than to the centimeter.

00:12:08.330 --> 00:12:11.914
When we have a distance to measure,
we know where to position.

00:12:11.916 --> 00:12:13.238
How do we do this?

00:12:13.239 --> 00:12:14.958
With the GPS, for instance,

00:12:14.959 --> 00:12:19.560
if you have a cluster of satellites
with synchronized atomic clocks,

00:12:19.561 --> 00:12:22.044
by measuring the travel time of each wave,

00:12:22.045 --> 00:12:25.471
from each satellite to your receptor,

00:12:25.472 --> 00:12:28.981
you measure your distance
from each satellite,

00:12:28.982 --> 00:12:31.511
and by triangulation,
you measure your position.

00:12:31.512 --> 00:12:34.385
You need four satellites,
because in the time-space

00:12:34.386 --> 00:12:37.316
there are four coordinates: x, y, z and t,

00:12:37.317 --> 00:12:40.426
since time is also needed
to position one self in time-space.

00:12:40.427 --> 00:12:42.663
You can see that the GPS's applications

00:12:42.664 --> 00:12:45.106
are not only to position
oneself in one's car

00:12:45.107 --> 00:12:47.754
an area where we need
a resolution of a few meters.

00:12:47.755 --> 00:12:50.265
There are also applications in geophysics.

00:12:50.266 --> 00:12:53.846
We will be able to analyze
the movement of tectonic plates

00:12:53.847 --> 00:12:57.206
with resolutions down
to a few centimeters per year,

00:12:57.207 --> 00:12:59.007
which is an excellent resolution.

00:12:59.008 --> 00:13:00.087
It is interesting,

00:13:00.088 --> 00:13:03.755
since from time measurements,
we know the functioning of the Earth,

00:13:03.756 --> 00:13:06.490
we infer fluctuations
of the rotation of the Earth.

00:13:06.491 --> 00:13:09.741
It is interesting because historically,
it was the exact opposite.

00:13:09.747 --> 00:13:12.168
It was the Earth's rotation
which gave the hour.

00:13:12.169 --> 00:13:14.182
At present it is the opposite.

00:13:14.183 --> 00:13:18.673
The measurement of time gives us
the fluctuations of the Earth's rotation.

00:13:18.674 --> 00:13:24.018
Another thing you might have heard,
on the radio or on television,

00:13:24.019 --> 00:13:26.057
are the famous intercalary seconds.

00:13:26.058 --> 00:13:29.170
As the Earth does not go
perfectly round, as we all know,

00:13:29.171 --> 00:13:31.700
and the atomic time is infinitely stable,

00:13:31.701 --> 00:13:35.496
meaning that both time scales linked
to Earth's rotation and to atomic clocks

00:13:35.497 --> 00:13:37.759
are going to diverge from one another.

00:13:37.760 --> 00:13:39.458
To prevent them diverging too much,

00:13:39.496 --> 00:13:41.724
we voluntarily add,
at the international level,

00:13:41.725 --> 00:13:44.284
a extra second, called
the intercalary second.

00:13:44.285 --> 00:13:47.743
This means that generally every two years,

00:13:47.744 --> 00:13:52.478
either on June 30, or December 31,
- December 31 is less bothersome -

00:13:52.479 --> 00:13:55.189
one minute is actually made of 61 seconds.

00:13:55.198 --> 00:13:58.737
This leap must be made
everywhere, all over the Earth.

00:13:58.738 --> 00:14:01.821
So as a conclusion,

00:14:01.846 --> 00:14:06.850
I would like to show you

00:14:06.874 --> 00:14:10.577
time measurement has left
the field of astronomy

00:14:10.601 --> 00:14:14.879
to land in the domains of atomic physics,
and of quantum mechanics.

00:14:14.880 --> 00:14:18.033
Since the invention of clocks,
around the mid-20th century,

00:14:18.034 --> 00:14:20.493
we have gained
a factor of 10 every 10 years.

00:14:20.494 --> 00:14:24.565
It is really impressive progress.

00:14:24.566 --> 00:14:29.544
Each time we improved the precision,
we said to ourselves:

00:14:29.545 --> 00:14:32.634
there is no need for all these
figures after the decimal point.

00:14:32.635 --> 00:14:33.663
It is not true.

00:14:33.664 --> 00:14:38.222
Each time, an application appeared,
10 years later, 20 years later,

00:14:38.223 --> 00:14:40.511
which used that precision.

00:14:40.512 --> 00:14:42.942
I think that I'll conclude and say

00:14:42.943 --> 00:14:45.650
that those who measure time
are ahead of their time

00:14:45.650 --> 00:14:47.419
is totally appropriate in this case.

00:14:47.419 --> 00:14:48.759
Thank you very much.

00:14:48.759 --> 00:14:49.949
(Applause)