In the 2nd and 3rd post, we discuss two forms of translations - rotations and time-evolutions. But wait, we haven’t even talked about linear translation!
Before even talking about the translation operator, we need to think about what space, or what basis states we should use. Obviously, translations are closely related to positions, and it would be convenient to have the position space. In this post, we will first deal with 1D space for simplicity. We define a position operator x^:
x^∣x⟩=x∣x⟩
where
∣x⟩ is an eigenstate of x^in the position space, and x is the 1D position. Unlike angular momentum space we’ve dealt with, the position space have infinite many eigenstates, and we need to be more careful with the mathematical notations.
For a general state ψ, we can represent it as the superposition of position eigenstates:
∣ψ⟩=∫−∞∞dx∣x⟩⟨x∣ψ⟩
and its conjugate,
⟨ψ∣=∫−∞∞dx⟨ψ∣x⟩⟨x∣
If we apply this relationship to an eigenstate
∣x′⟩, we will obtain an important relationship:
∣x′⟩=∫dx∣x⟩⟨x∣x′⟩⟹⟨x∣x′⟩=δ(x−x′)
From here, we can also further explore the first postulate of QM we talked about in the first post, that a quantum mechanical state is represented by a probability wave function.
Let ψ be a normalized state, i.e.
⟨ψ∣ψ⟩=1, we have
This is where we get the physical explanation that
⟨x∣ψ⟩2 is the probability that the particle is at position x. Further, notice that
⟨x∣ψ⟩, the probability of amplitude, is a function of x. Here we have the wave function:
ψ(x)=⟨x∣ψ⟩
The Translation Operator
Now we can talk about the translation operator T^, which should act as follows:
T^(a)∣x⟩=∣x+a⟩
Literally, T^(a) should “move” the state by a.
We can act T^ on
∣ψ⟩ to obtain
∣ψ′⟩=T^∣ψ⟩=∫dx′∣x′+a⟩⟨x′∣ψ⟩
with this definition and see how ψ(x) changes to ψ′(x):
This may seem wrong, but think about what happens when you shift a function f(x) to the left by a: you obtain f(x−a).
Now let’s investigate the representation of the translation operator. Think about the rotation operator Ri^ and the time-evolution operator U^ that we’ve discussed in the 2nd and 3rd posts, it shouldn’t be a surprise that we need a generator of translation, which happens to be the momentum operatorpx^.
T^(dx)=1−ℏipx^dxT^(a)=e−ipx^a/ℏ
The Commutator of the Position and Momentum Operator
The commutator of x^ and px^ very useful and fundamental relationship to obtain:
By applying this infinitesimal limit definition above, we have:
(x^T^(δx)−T^(δx)x^)∣ψ⟩=(ℏ−iδx)[x^,px^]∣ψ⟩
By applying the translation operator and position operator directly, we obtain
where in the last step we keep only the leading-order term of the wave function. Equate the two results, we find that:
[x^,px^]=iℏ⟹ΔxΔpx≥2ℏ
The Ehrenfest Theorem and The Correspondence Principle
Though your intuition for QM now should be enough to convince you that p^x is the momentum operator, let’s validate this by calculating some momentum related quantities. Specifically, let’s start with the energy.
In classical mechanics, we have total energy is equal to mechanical energy and potential energy, or E=2mpx2+V(x), which is translated to the Hamiltonian with the corresponding eigenvalues:
H^=2mpx^2+V(x^)
Recall the result about the time dependence of expectation from post 3,
dtd⟨A⟩=ℏi⟨ψ(t)∣[H^,A^]∣ψ(t)⟩+⟨ψ(t)∣∂t∂A^∣ψ(t)⟩
Using the fact that operators x^ and px^ are time-independent, we can obtain two relationships,
The next to last step of the second derivation requires writing out V(x^) as Taylor series and find out [x^n,px^] by the mathematical induction.
These two relationships together,
dtd⟨x⟩=m⟨px⟩dtd⟨px⟩=⟨−dxdV⟩
are called the Ehrenfest theorem. It may seem to suggest the expectation values ⟨x⟩,⟨px⟩ satisfy the Newton’s second law, but that will require ⟨V(x)⟩=V(⟨x⟩), which is not always true. Nevertheless, for the systems widely studied by the classical mechanics, this holds true. The Ehrenfest theorem manifests the correspondence principle, the principle that quantum mechanics reproduce the classical mechanics results in large limits.
The Momentum Operator in the Position Space
Last but not the least, let’s consider the momentum operator’s representation in the position space.
First, we directly apply the translation operator in the position space:
Then, observe the definition of the translation operator:
T(δx)=1−ℏipx^δx
We conclude that, in the position space:
px^=iℏ∂x∂
The Momentum Space
In the momentum space, we have basis states as the eigenstates of the momentum operator:
px^∣p⟩=p∣p⟩
Similarly to the position space, we have the following representations and relationships:
∣ψ⟩=∫dp∣p⟩⟨p∣ψ⟩⟨p∣p′⟩=δ(p−p′)∫dp∣⟨p∣ψ⟩∣2=1
What’s interesting and useful is to find out
⟨x∣p⟩ so that we can make transitions between the position space and the momentum space.
Luckily, we know the representation of the momentum operator in both spaces:
⟨x∣px^∣p⟩=p⟨x∣p⟩=iℏ∂x∂⟨x∣p⟩
Solving the differential equation, we get:
⟨x∣p⟩=Neipx/ℏ
where N is a constant. From this expression, we can already derive the de Broglie wavelength. Notice that the expression is a oscillatory function of x with the complex phase p/ℏ. Hence the wavelength λ is p/ℏ2π=p2πℏ, or simply
λ=ph
To find the value of the constant, we represent a momentum eigenstate in the position space:
where we use the Dirac delta’s representation δ(x)=2π1∫dkeikx.
Now we’ve found that N=2πℏ1 (I know 2πℏ=h, but we the quantum physicists just don’t like h…), we can write:
⟨x∣p⟩=2πℏ1eipx/ℏ
The Solutions to the Schrödinger Equation
Now we have a representation of the momentum operator in the position space, we can first write out the Hamitonian in the position space:
H^=−2mℏ2∂x2∂2+V(x^)
So the time-dependent Schrödinger Equation in the position space is:
[−2mℏ2∂x2∂2+V(x^)]ψ(x,t)=iℏ∂t∂ψ(x,t)
Suppose ψ(x) is an energy eigenfunction of the Hamiltonian and ψ(x,t)=ψ(x)e−iEt/ℏ, we have the time-independent Schrödinger Equation:
[−2mℏ2∂x2∂2+V(x^)]ψ(x)=Eψ(x)
With this equation and a specific potential function, we can solve the wave equation.
There are many typical 1D problems to solve such as the finite and infinite potential well, scattering, tunneling, the delta function potential, the quantum oscillator. Check out resources online like this post from UVA for more references.
References
[1] Zajc, William. PHYS GU4021-4022 Quantum Mechanics I-II, Department of Physics, Columbia University. 2019-2020.
[2] Townsend, John. A Modern Approach to Quantum Mechanics. 2nd Edition.