MIT 8.04 Lecture 1: An overview of quantum mechanics.

L1.1 Quantum mechanics as a framework. Defining linearity (17:46)

L1.2 Linearity and nonlinear theories. Schrödinger’s equation (10:01)

L1.3 Necessity of complex numbers (07:38)

L1.4 Photons and the loss of determinism (17:20)

L1.5 The nature of superposition. Mach-Zehnder interferometer (14:30)

L1.1 Quantum mechanics as a framework. Defining linearity (17:46)

PROFESSOR: Very good. So it’s time to start. So today, I want to talk about general features of quantum
mechanics.

Quantum mechanics is something that takes some time to learn, and we’re going to be doing
some of that learning this semester. But I want to give you a perspective of where we’re going,
what are the basic features, how quantum mechanics looks, what’s surprising about it, and
introduce some ideas that will be relevant throughout this semester and some that will be
relevant for later courses as well. So it’s an overview of quantum mechanics.
So quantum mechanics, at this moment, is almost 100 years old. Officially-- and we will hear–
this year, in 2016, we’re celebrating the centenary of general relativity. And when will the
centenary of quantum mechanics be? I’m pretty sure it will be in 2025. Because in 1925,
Schrodinger and Heisenberg pretty much wrote down the equations of quantum mechanics.
But quantum mechanics really begins earlier. The routes that led to quantum mechanics
began in the late years of the 19th century with work of Planck, and then at the beginning of
the century, with work of Einstein and others,m as we will see today and in the next few
lectures. So the thoughts, the puzzles, the ideas that led to quantum mechanics begin before
1925, and in 1925, it suddenly happened.
So what is quantum mechanics? Quantum mechanics is really a framework to do physics, as
we will understand. So quantum physics has replaced classical physics as the correct
description of fundamental theory. So classical physics may be a good approximation, but we
know that at some point, it’s not quite right. It’s not only not perfectly accurate. It’s conceptually
very different from the way things really work.
So quantum physics has replaced classical physics. And quantum physics is the principles of
quantum mechanics applied to different physical phenomena. So you have, for example,
quantum electrodynamics, which is quantum mechanics applied to electromagnetism. You
have quantum chromodynamics, which is quantum mechanics applied to the strong
interaction.
You have quantum optics when you apply quantum mechanics to photons. You have quantum
gravity when you try to apply quantum mechanics to gravitation. Why the laughs?

And that’s what gives rise to string theory, which is presumably a quantum theory of gravity,
and in fact, the quantum theory of all interactions if it is correct. Because it not only describes
gravity, but it describes all other forces. So quantum mechanics is the framework, and we
apply it to many things.
So what are we going to cover today? What are we going to review? Essentially five topics–
one, the linearity of quantum mechanics, two, the necessity of complex numbers, three, the
laws of determinism, four, the unusual features of superposition, and finally, what is
entanglement. So that’s what we aim to discuss today.
So we’ll begin with number one, linearity. And that’s a very fundamental aspect of quantum
mechanics, something that we have to pay a lot of attention to. So whenever you have a
theory, you have some dynamical variables. These are the variables you want to find their
values because they are connected with observation.
If you have dynamical variables, you can compare the values of those variables, or some
values deduced from those variables, to the results of an experiment. So you have the
equations of motion, so linearity. We’re talking linearity. You have some equations of motion,
EOM. And you have dynamical variables.
If you have a theory, you have some equations, and you have to solve for those dynamical
variables. And the most famous example of a theory that is linear is Maxwell’s theory of
electromagnetism. Maxwell’s theory of electromagnetism is a linear theory.
What does that mean? Well, first, practically, what it means is that if you have a solution-- for
example, a plane wave propagating in this direction-- and you have another solution-- a plane
wave propagating towards you-- then you can form a third solution, which is two plane waves
propagating simultaneously. And you don’t have to change anything. You can just put them
together, and you get a new solution.
The two waves propagate without touching each other, without affecting each other. And
together, they form a new solution. This is extraordinarily useful in practice because the air
around us is filled with electromagnetic waves. All your cell phones send electromagnetic
waves up the sky to satellites and radio stations and transmitting stations, and the millions of
phone calls go simultaneously without affecting each other.
A transatlantic cable can conduct millions of phone calls at the same time, and as much data
and video and internet. It’s all superposition. All these millions of conversations go
simultaneously through the cable without interfering with each other.
Mathematically, we have the following situation. In Maxwell’s theory, you have an electric field,
a magnetic field, a charge density, and a current density. That’s charge per unit area per unit
of time. That’s the current density.
And this set of data correspond to a solution if they satisfy Maxwell’s equations, which is a set
of equations for the electromagnetic field, charged densities, and current density. So suppose
this is a solution, that you verify that it solves Maxwell’s equation.
Then linearity implies the following. You multiply this by alpha, alpha e, alpha b, alpha rho, and
alpha j. And think of this as the new electric field, the new magnetic field, the new charged
density, and the new current is also a solution. If this is a solution, linearity implies that you can
multiply those values by a number, a constant number, a alpha being a real number. And this
is still a solution.
It also implies more. Linearity means another thing as well. It means that if you have two
solutions, e1, b1, rho 1, j1, and e2, b2, rho 2, j2-- if these are two solutions, then linearity
implies that the sum e1 plus e2, b1 plus b2, rho 1 plus rho 2, and j1 plus j2 is also a solution.
So that’s the meaning, the technical meaning of linearity. We have two solutions. We can add
them. We have a single solution. You can scale it by a number.
Now, I have not shown you the equations and what makes them linear. But I can explain this a
little more as to what does it mean to have a linear equation. Precisely what do we mean by a
linear equation? So a linear equation.
And we write it schematically. We try to avoid details. We try to get across the concept. A linear
equation, we write this l u equal 0 where u is your unknown and l is what is called the linear
operator, something that acts on u. And that thing, the equation, is of the form l and u equal 0.
Now, you might say, OK, that already looks to me a little strange, because you have just one
unknown, and here we have several unknowns. So this is not very general. And you could
have several equations. Well, that won’t change much.
We can have several linear operators if you have several equations, like l1 or something, l2 on
something, all these ones equal to 0 as you have several equations. So you can have several
u’s or several unknowns, and you could say something like you have l on u, v, w equals 0
where you have several unknowns.
But it’s easier to just think of this first. And once you understand this, you can think about the
case where you have many equations. So what is a linear equation? It’s something in which
this l-- the unknown can be anything, but l must have important properties, as being a linear
operator will mean that l on a times u, where a is a number, should be equal to alu and l on u1
plus u2 on two unknowns is equal to lu 1 lu 2. This is what we mean by the operator being
linear.
So if an operator is linear, you also have l on alpha u1 plus beta u2. You apply first the second
property, l on the first plus l on the second. So this is l of alpha u1 plus l of beta u2. And then
using the first property, this is alpha l of u1 plus beta l of u2.
And then you realize that if u1 and u2 are solutions-- which means lu 1 equal lu 2 equals 0 if
they solve the equation-- then alpha u1 plus beta u2 is a solution. Because if lu1 is 0 and lu2 is
0, l of alpha u1 plus beta u2 is 0, and it is a solution. So this is how we write a linear equation.
Now, an example probably would help. If I have the differential equation du dt plus 1 over tau
u equals 0, I can write it as an equation of the form lu equals 0 by taking l on u to be defined to
be vu vt plus 1 over tau u.
Now, it’s pretty much-- I haven’t done much here. I’ve just said, look, let’s define l [? active ?]
[? on ?] u to be this. And then certainly, this equation is just lu equals 0. The question would be
maybe if somebody would tell you how do you write l alone-- well, l alone, probably we should
write it as d dt without anything here plus 1 over tau.
Now, that’s a way you would write it to try to understand yourself what’s going on. And you
say, well, then when l acts as the variable u, the first term takes the derivative, and the second
term, which is a number, just multiplies it. So you could write l as this thing.
And now it is straightforward to check that this is a linear operator. l is linear. And for that, you
have to check the two properties there. So for example, l on au would be ddt of au plus 1 over
tau au, which is a times du d tau plus 1 over tau u, which is alu.
And you can check. I asked you to check the other property l on u1 plus u2 is equal to lu 1
plus lu 2. Please do it.

L1.2 Linearity and nonlinear theories. Schrödinger’s equation (10:01)

MITOCW | watch?v=kiuwtaprFjk
PROFESSOR: So here is of something funny. You might say, OK, what is simpler? A theory that is linear or a
theory that is not linear? And the answer, of course, a linear theory is much simpler. General–
Maxwell’s equations are linear. Einstein’s theory of relativity is very nonlinear, very
complicated.
How about classical mechanics? Is classical mechanics linear or nonlinear? What do we think?
Can’t hear anyone. Linear, OK. You may think it’s linear because it’s supposed to be simple,
but it’s not. It’s actually is very nonlinear. Newton could solve the two body problem but he
couldn’t solve the three body problem. Already with three bodies, you cannot superpose
solutions that you get with two bodies. It’s extraordinarily complicated, classical mechanics.
Let me show you. If you have motion in one dimension, in 1D, you have the equation of
motion, motion in one dimension, and there are potential V of x, that this time independent-- a
particle moving in one dimension x with under the influence of a potential, V of x. The second–
the dynamical variable is x of t. The dynamical variable. And the equation of motion is-- so let
me explain this. This is force equal mass times acceleration. This is mass, this is acceleration,
the second derivative of the position, and V force is minus the derivative of the potential
evaluated at the position.
You know, derivatives of potentials-- if you think of a potential, the derivative of the potential is
here positive, and you know if you have a mass here, it tends to go to the left, so the force is
on the left, so it’s minus. So V prime is the derivative of V with respect to its argument. And the
problem is that while this, taking derivatives, is a linear operation. If you take two derivatives of
a sum of things, you take two derivatives of the first plus two derivatives of the second.
But yes, its-- this side is linear, but this side may not be linear. Because a potential can be
arbitrary. And that the reverse-- so suppose the potential is cubic in x. V of x goes like x cubed.
Then the derivative of V goes like x squared, and x squared is not a linear function. So this,
Newton’s equation, is not a linear equation. And therefore, it’s complicated to solve. Very
complicated to solve.
So finally, we can get to our case, quantum mechanics. So in quantum mechanics, what do we
have? Quantum mechanics is linear. First, you need an equation, and whose equation is it?
Schrodinger’s equation, 1925. He writes an equation for the dynamical variable, and the
dynamical variable is something called the wave function. This wave function can depend on t-

• depends on time-- and it may depend on other things as well. And he describes the
  dynamics of the quantum system as it evolved in time. There is the wave function, and you
  have an equation for this wave function.
  And what is the equation for this wave function? It’s a universal equation-- i hbar partial
  derivative with respect to time of psi is equal to H hat of psi, where H hat is called the
  Hamiltonian and it’s a linear operator. That’s why I had to explain a little bit what the linear
  operator is. This is the general structure of the Schrodinger equation-- time derivative and the
  linear operator. So if you wish to write the Schrodinger equation as an L psi equals 0, then L
  psi would be defined i hbar del/del t of psi minus H hat psi. Then this is the Schrodinger
  equation. This equation here is Schrodinger’s equation.
  And as you can see, it’s a linear equation. You can check it, check that L is a linear operator.
  Therefore, it is naturally linear, you can see, because you do it differently, because the
  derivative with respect to time is a linear operation. If you have the ddt of a number of times a
  function, the number goes out, you differentiate the function. ddt of the sum of two functions,
  you differentiate the first, you differentiate. So this is linear and H we said is linear, so L is
  going to be linear and the Schrodinger equation is going to be a linear equation, and
  therefore, you’re going to have the great advantage that any time you find solutions, you can
  scale them, you can add them, you can put them together, combine them in superpositions,
  and find new solutions.
  So in that sense, it’s remarkable that quantum mechanics is simpler than classical mechanics.
  And in fact, you will see throughout this semester how the mathematics and the things that we
  do are simpler in quantum mechanics, or more elegant, more beautiful, more coherent, it’s
  simpler and very nice.
  OK, i is the square root of minus 1, is the imaginary unit, and that’s what we’re going to talk
  next on the necessity of complex numbers. hbar, yes, it’s a number. It shows up in quantum
  mechanics early on. It it’s called Planck’s constant and it began when Planck tried to fit the
  black value spectrum and he found the need to put a constant in there, and then later, Einstein
  figured out that it was very relevant, so yes, it’s a number.
  For any physical system that you have, you will have a wave function and you will have a
  Hamiltonian, and the Hamiltonian is for you to invent or for you to discover. So if you have a
  particle moving on a line, the wave function will depend on time and on x. If you have a particle
  moving in three dimensions, it will depend on x vector. It may depend on other things as well
  or it maybe, like, one particle has several wave functions and that happens when you have a
  particle with spin.
  So in general, always time, sometimes position, there may be cases where it doesn’t depend
  on position. You think of an electron at some point in space and it’s fixed-- you lock it there
  and you want understand the physics of that electron locked into place, and then position is
  not relevant. So what it does with its spin is relevant and then you may need more than one
  wave function-- what is one describing the spin up and one describing the spin down?
  So it was funny that Schrodinger wrote this equation and when asked, so what is the wave
  function? He said, I don’t know. No physical interpretation for the wave function was obvious
  for the people that invented quantum mechanics. It took a few months until Max Born said it
  has to do with probabilities, and that’s what we’re going to get next.
  So our next point is the necessity of complex numbers in quantum mechanics.

L1.3 Necessity of complex numbers (07:38)

MITOCW | watch?v=f079K1f2WQk
PROFESSOR: So in quantum mechanics, you see this i appearing here, and it’s a complex number-- the
square root of minus 1. And that shows that somehow complex numbers are very important.
Well it’s difficult to overemphasize the importance of i-- is the square root of minus 1 was
invented by people in order to solve equations. Equations like x squared equals minus 1. And
it so happens that once you invent i you need to invent more numbers, and you can solve
every polynomial equation with just i. And square root of i-- well square root of i can be written
in terms of i and other numbers.
So if you have a complex number z-- we sometimes write it this way, and we say it belongs to
the complex numbers, and with a and b belonging to the real numbers. And we say that the
real part of z is a, the imaginary part of z is b. We also define the complex conjugate of z,
which is a minus i b and we picture the complex number z by putting a on the x-axis b on the
y-axis, and we think of the complex number z here-- kind of like putting the real numbers here
and the imaginary parts here. So you can think of this as ib or b, but this is the complex
number-- maybe ib would be a better way to write it here.
So with complex numbers, there is one more useful identity. You define the norm of the
complex number to be square root of a squared plus b squared and then this results in the
norm squared being a squared plus b squared. And it’s actually equal to z times z star. A very
fundamental equation-- z times z star-- if you multiply z times z star, you get a squared plus b
squared. So the norm squared-- the norm of this thing is a real number. And that’s pretty
important.
So there is one other identity that is very useful and I might well mention it here as we’re going
to be working with complex numbers. And for more practice on complex numbers, you’ll see
the homework. So suppose I have in the complex plane an angle theta, and I want to figure
out what is this complex number z here at unit radius. So I would know that it’s real part would
be cosine theta. And its imaginary part would be sine theta. It’s a circle of radius 1. So that
must be the complex number. z must be equal to cosine theta plus i sine theta.
Because the real part of it is cosine theta. It’s in that horizontal part’s projection. And the
imaginary part is the vertical projection. Well the thing that is very amazing is that this is equal
to e to the i theta. And that is very non-trivial. To prove it, you have to work a bit, but it’s a very
famous result and we’ll use it.
So that is complex numbers. So complex numbers you use them in electromagnetism. You
sometimes use them in classical mechanics, but you always use it in an auxiliary way. It was
not directly relevant because the electric field is real, the position is, real the velocity is real–
everything is real and the equations are real. On the other hand, in quantum mechanics, the
equation already has an i. So in quantum mechanics, psi is a complex number necessarily. It
has to be.
In fact, if it would be real, you would have a contradiction because if psi is real, turns out for all
physical systems we’re interested in, H on psi real gives you a real thing. And here, if psi is
real then the relative is real, and this is imaginary and you have a contradiction. So there are
no solutions that are real. So you need complex numbers. They’re not auxiliary. On the other
hand, you can never measure a complex number. You measure real numbers-- ammeter,
position, weight, anything that you really measure at the end of the day is a real number.
So if the wave function was a complex number, it was the issue of what is the physical
interpretation. And Max Born had the idea that you have to calculate the real number called
the norm of this square, and this is proportional to probabilities. So that was a great discovery
and had a lot to do with the development of quantum mechanics. Many people hated this. In
fact, Schrodinger himself hated it, and his invention of the Schrodinger cat was an attempt to
show how ridiculous was the idea of thinking of these things as probabilities. But he was
wrong, and Einstein was wrong in that way.
But when very good physicists are wrong, they are not wrong for silly reasons, they are wrong
for good reasons, and we can learn a lot from their thinking. And this EPR are things that we
will discuss at some moment in your quantum sequence at MIT. Einstein-Podolski-Rosen was
an attempt to show that quantum mechanics was wrong and led to amazing discoveries. It was
the EPR paper itself was wrong, but it brought up ideas that turned out to be very important.

L1.4 Photons and the loss of determinism (17:20)

MITOCW | watch?v=8OsUQ1yXCcI
PROFESSOR: Determinism. And it all begins with photons. Einstein reluctantly came up with the idea that
light was made of quanta-- quanta of light called photons. Now when you think of photons, we
think of a particle. So everybody knew that light was a wave. Maxwell’s equations had been so
successful. Nevertheless, photoelectric effect-- Planck’s work-- all were leading to the idea
that, in some ways, photons were also particles.
So when you think of a particle, however, there is an important difference between a particle in
the sense of Newton, which is an object with zero size that carries energy and has a precise
position and velocity at any time, and the quantum mechanical idea of particle, which is just
some indivisible amount of energy or momentum that propagates. So light was made of
photons-- packets of energy. And a photon is a particle-- a quantum mechanical particle. Not
in the sense that maybe it has position and velocity determined or it’s a point particle, but more
like a packet that is indivisible. You cannot decompose it in further packets.
So Einstein realized that for a photon, the energy was given by h nu. Where nu is the
frequency of the light that this photon is helping build up. So if you have a beam of light, you
should think it’s billions of photons. And according to the frequency of that light that is related
to the wavelength-- by the equation frequency times wavelength is velocity of light-- you
typically know, for light, the wavelength, and you know the frequency, and then you know the
energy of each of the photons.
The photons have very, very little energy. We have very, very little energy, but your eyes are
very good detectors of photons. If you’re in a totally dark room, your eye, probably, can take
as little as five photons if they hit your retina.
So it’s a pretty good detector of photons. Anyway, the thing that I want to explain here is what
happens if a beam of light hits a polarizer. So what is a polarizer? It’s a sheet of plastic or
some material. It has a preferential direction. Let me align that preferential direction with the xaxis, and that’s a polarizer. And if I send light that is linearly polarized along the x-axis, it all
goes through. If I send light linearly polarized along the y-axis, nothing goes through. It all gets
absorbed. That’s what a polarizer does for a living.
In fact, if you send light in this direction, the light that comes out is identical to the light that
came in. The frequency doesn’t change. The wavelength doesn’t change. It’s the same light,
the same energy. So far, so good. Now let’s imagine that we send in light linearly polarized at
some angle alpha. So we send an electric field E alpha, which is E0 cosine alpha x hat plus E0
sine alpha y hat.
Well, you’ve studied electromagnetism, and you know that this thing, basically, will come
around and say, OK, you can go through because you’re aligning the right direction, but you
are orthogonal to my preferential direction, or orthogonal I absorbed, so this disappears. So
after the polarizer, E is just E0 cosine alpha x hat. That’s all that is left after the polarizer. Well
here is something interesting-- you know that the energy on electromagnetic field is
proportional to the magnitude of the electric field square, that’s what it is. So the magnitude of
this electric field-- if you can notice, it’s the square root of the sum of the squares will give you
E0 as the magnitude of this full electric field.
But this electric field has magnitude E0 cosine alpha. So the fraction of power-- fraction of
energy through is cosine alpha squared. The energy is always proportional to the square. So
the square of this is E0 squared cosine squared alpha. And for this one, the magnitude of it is
E0, so you divide by E0 and cosine alpha is the right thing. This is the fraction of the energy. If
alpha is equal to 0, you get cosine of 01. You get all the energy 1.
If alpha is equal to pi over 2, the light is polarized along the y direction, nothing goes through–
indeed, cosine of pi over 2 is 0, and nothing goes through. So the fraction of energy that goes
through is cosine squared alpha. But now, think what this means for photons.
What it means for photons is something extraordinarily strange. And so strange that it’s almost
unbelievable that we get so easily in trouble. Here is this light beam over here, and it’s made
up of photons. All identical photons, maybe billions of photons, but all identical. And now, think
of sending this light beam over there-- a billion identical photons-- you send them one by one
into the state, and see what happens. You know what has to happen, because classical
behavior is about right. This fraction of the photons must go through, and 1 minus that must
not go through.
You see, it cannot be there comes a photon and half of it goes through, because there’s no
such thing as half of it. If there would be half of it, it would be half the energy and, therefore,
different color. And we know that after a polarizer, the color doesn’t change. So here is the
situation. You’re sending a billion photons and, say, one-third has to get through. But now, the
photos are identical.
How can that happen in classical physics? If you send identical photos, whatever happens to a
photon will happen to all, but the photon either gets absorbed or goes through. And if it gets
absorbed, then all should get absorbed. And if it goes through, all should go through because
they are all identical. And now you have found a situation which identical set of experiments
with identically prepared objects sometimes gives you different results. It’s a debacle. It’s a
total disaster.
What seems to have happened here-- you suddenly have identical photons, and sometimes
they go through, and sometimes they don’t go through. And therefore, you’ve lost
predictability. It’s so simple to show that if photons exist, you lose predictability. And that’s what
drove Einstein crazy. He knew when he entered these photons that he was getting in trouble.
He was going to get in trouble with classical physics.
So possible ways out-- people speculate about it-- people said, well, yes, the photos are
identical, but the polarizer has substructure. If it hits in this interatomic part, it goes through,
and in that interatomic part, it doesn’t go through. People did experiments many times. It’s not
true. The polarizer is like that. And then came a more outrageous proposition by Einstein and
others-- that there are hidden variables.
You think the photons are identical, but a photon has a hidden variable-- a property you don’t
know about. If you knew that property about the photon, you would be able to tell if it goes
through or it doesn’t go through. But you don’t know it, so that’s why you’re stuck with
probabilities.
It’s because the quantum theory is not complete. There are hidden variables. And once you
put the hidden variables, you’ll discover the photon has more something inside it, and they are
not the same, even though they look the same. And that’s a hidden variable theory. And it
sounds so philosophical that you would think, well, if you don’t know about them, but they are
there, these properties, how could you ever know they are there?
And the great progress of John Bell with the Bell inequalities is that he demonstrated that that
would not fix the problem. Quantum mechanics cannot be made deterministic with hidden
variables. It was an unbelievable result-- the result of John Bell. So that’s something we will
advance towards in this course but not quite get there. 805 discusses this subject in detail. So
at the end of the day, we’ve lost determinism. We can only predict probabilities.
So photons either gets through or not, and can only predict probabilities. Now we write, in
So photons either gets through or not, and can only predict probabilities. Now we write, in
classical physics, a beam like that. But how do we write the wave function of a photon? Well,
this is quite interesting. We think of states of a particle as wave functions. And I will call them,
sometimes, states; I will call them, sometimes, wave functions; and I sometimes will call them
vectors. Why vector? Because the main thing you do with vectors is adding them or multiplying
them by numbers to scale them.
And that’s exactly what you can do with a linear equation. So that’s why people think of states,
or wave functions, as vectors. And Dirac invented a notation in which to describe a photon
polarized in the x direction, you would simply write something like this. Photon colon x and this
object-- you think of it as some vector or wave function, and it represents a photon in the x
direction. And we’re not saying yet what kind of vector this is, but it’s some sort of vector. It’s
not just a symbol, it represents a vector. And that’s a possible state.
This is a photon polarized along x. And you can also have, if you wish, a photon polarized
along y. And linearity means that if those photos can exist, the superposition can exist. So
there can exist a state called cos alpha photon x plus sine alpha photon y, in which I’ve
superposed one state with another-- created a sum-- and this I call the photon state polarized
in the alpha direction.
So this is how, in quantum mechanics, you think of this-- photons-- we will elaborate that and
compare with this equation. It’s kind of interesting. What you lose here is this ease. There’s no
ease there because it’s one photon. When you have a big electric field, I don’t know how many
photons there are. I would have to calculate the energy of this beam and find the frequency
that I didn’t specify, and see how many photons. But each photon in this beam quantum
mechanically can be represented as this superposition. And we’ll talk more about this
superposition now because our next subject is superpositions and how unusual they are.
Well the hidden variable explanation failed because Bell was very clever, and he noted that
you could design an experiment in which the hidden variables would imply that some
measurements would satisfy an inequality. If the existed hidden variables and the world was
after all classical, the results of experiments would satisfy a Bell inequality. And then a few
years later, the technology was good enough that people could test the Bell inequality with an
experiment, and they figured out it didn’t hold. So the hidden variables lead to Bell inequalities
that are experimentally shown not to hold. And we will touch a little bit on it when we get to
untangle them.
After the polarizer, the photon is in the state photon x. It’s always polarized along the x
direction, so it’s kind of similar that this doesn’t go through. This goes through, but at the end
of the day, as we will explain very soon, the cosine alpha is not relevant here. When it goes
through, the whole photon goes through. So there’s no need for a cosine alpha. So that’s what
goes out of the polarizer.

L1.5 The nature of superposition. Mach-Zehnder interferometer (14:30)

MITOCW | watch?v=CR-eOhdxbes
PROFESSOR: Superposition is very unusual and very interesting. Now we’ve said about superposition that in
classical physics, when we talk about superposition we have electric fields, and you add the
electric fields, and the total electric field is the sum of electric fields, and it’s an electric field.
And there’s nothing strange about it. The nature of superposition in quantum mechanics is
very strange. So nature of superposition-- I will illustrate it in a couple of different ways.
One way is with a device that we will get accustomed to. It it’s called the Mach-Zehnder
interferometer, which is a device with a beam splitter in here. You send in a beam of light–
input- beam splitter and then the light-- indeed half of it gets reflected, half of it gets
transmitted. Then you put the mirror here-- mirror 1, you put the mirror 2 here, and this gets
recombined into another beam splitter. And then if there would be just a light going in, here
there would be two things going out. There’s another one coming from the bottom. There will
be two. There will be interference. So you put a detector D0 here and a detector E1 here to
detect the light.
So that’s the sketch of the Mach-Zehnder interferometer-- beam splitters and mirrors. Take a
beam, spit the light, go down, up, and then recombine it and go into detectors. This was
invented by these two people, independently, in the 1890s-- '91 to '92 apparently. And people
did this with light-- beams of light before they realized they’re photons.
And what happens with a beam of light-- it’s interesting-- comes a beam of light. The beam
splitter sends half of the light one way half of the light the other way. You already know with
quantum mechanics that’s going to be probabilistic some photons will go up maybe some
photons will go down or something more strange can happen.
If you have a superposition, some photons may go both up and down. So that’s what can
happen in quantum mechanics. If you send the beam, classical physics, it divides half and half
and then combines. And there’s an interference effect here. And we will design this
interferometer in such a way that sometimes we can produce an interference that everything
goes to D0 or everything goes to D1 or we can produce suitable interferences that we can get
fractions of the power going into D1 and D2-- D0 and D1.
So we can do it in different ways, but we should think of this as a single photon. Single photos
going one at a time. You see, whatever light you put in here, experimentally, the same
frequency goes out here. So what is interference? You might think, intuitively, that interference
is one photon interfering with another one, but it can’t be.
If two photos would interfere in a canceling, destructive interference, you will have a bunch of
energy. It goes into nothing. It’s impossible. If they would interfere constructively, you would
add the electric fields and the amplitude would be four times as big because it’s proportional to
the square. But two photos are not going to go to four photons. It cannot conserve energy.
So first of all, when you get light interference, each photon is interfering with itself. It sounds
crazy, but it’s the only possibility. They cannot interfere with each other. You can send the
photons one at a time and, therefore, each photon will have to be in both beams at the same
time. And then, each photon as it goes along, there will be an interference effect, and the
photon may end up here or end up there in a probabilistic way. So you have an example of
superposition.
Superposition. A single photon state a single photon is equal to superposition of a photon in
the upper beam and a photon in the lower beam. It’s like two different states-- a little different
from here, you had photons in two different polarizations states superposed. Here you have
photons in two different beams-- a single photon is in both beams at the same time. And
unless you have that, you cannot get a superposition and an interference that is consistent
with experiment. So what does that mean for superpositions?
Well, it means something that we can discuss, and I can say things that, at this moment, may
not make too much sense, but it would be a good idea that you think about them a little bit. We
associated states with vectors. States and vectors are the same thing. And it so happens that
when you have vectors, you can write them as the sum of other vectors.
So the sum of these two vectors may be this vector. But you can also write it as the sum of
these two vectors-- these two vectors add to the state. And you can write any vector as a sum
of different vectors, and that’s, actually, quite relevant. You will be doing that during the
semester-- writing a state a superposition of different things. And in that way you will
understand the physics of those states.
So for example, we can think of two states-- A and B. And you see, as I said, states wave
functions, vectors-- we’re all calling them the same thing. If you have a superposition of the
states A and B, what can happen? All right, we’ll do it the following way. Let’s assume if you
measure some property on A, you always get value A. So you measure something-- position,
momentum, angular momentum, spin, energy, something-- on A, it states that you always get
A. Suppose you measured the same property on B. You always get B as the value.
And now suppose you have a quantum mechanical state, and the state is alpha A plus beta B-

• it’s a superposition. This is your state. You superimpose A and B. And now you measure that
  property. That same property you could measure here, you measure it in your state. The
  question is, what will you get? You’ve now superimpose those states. On the first state, you
  always get A; on the second state, you always get B. What do you get on the superimposed
  states, where alpha and beta are numbers-- complex numbers in general?
  Well the most, perhaps, immediate guess is that you would get something in-between maybe
  alpha A plus beta B or an average or something. But no, that’s not what happens in quantum
  mechanics. In quantum mechanics, you always get A or you always get B. So you can do the
  experiment many times, and you will get A many times, and you may get B many times. But
  you never get something intermediate.
  So this is very different than in classical physics. If a wave has some amplitudes and you add
  another wave of different amplitudes, you measure the energy you get something
  intermediate. Here not! You make the superposition and as you measure you will either get
  the little a or the little b but with different probabilities.
  So roughly speaking, the probability to get little a is proportional to the number in front of here
  is alpha squared, and the probability to measure little b is proportional to beta squared. So in a
  quantum superposition, a single measurement doesn’t yield an average result or an
  intermediate result. It leads one or the other. And this should connect with this. Think of the
  photon we were talking about before.
  If you think of the photon that was at an angle alpha in this way, you could say that the
  polarizer is measuring the polarization of the object. And therefore, what is the possible result
  it may measure the polarizations say oh, if it’s in the x direction you get it right, and what is the
  probability that you get it to be in the x direction is proportional to cosine squared alpha-- the
  coefficient here squared. So the probability that you find the photon after measuring in the x
  direction is closer in squared alpha, and the probability that you’ll find that here is sine squared
  alpha.
  And after you measure, you get this state which is to say the following thing. The probability to
  get the value A is alpha squared, but if you get A, the state becomes A because this whole
  get the value A is alpha squared, but if you get A, the state becomes A because this whole
  state of the system becomes that. Because successive measurements will keep giving you the
  value A. If you get B, the state becomes B. So this is what is called the postulate of
  measurement and the nature of superposition. This is perhaps the most sophisticated idea
  we’ve discussed today, in which in a quantum superposition the results are not intermediate.
  So when you want to figure out what state you have, you have to prepare many copies of your
  state in this quantum system and do the experiment many times. Because sometimes you’ll
  get A, sometimes you’ll get B. After you’ve measured many times, you can assess the
  probabilities and reconstruct the state.