Is an operator (integral) Hermitian?

In summary: Gotcha. Very clear. Thank you for taking the time to explain it to me.In summary, the conversation discusses the concept of Hermitian operators and the attempt to prove that the operator ##\int_{-\infty}^{\infty} |x><x| dx## is Hermitian using the property ##\hat{Q} = \hat{Q}^{\dagger}##. The individual struggles to understand the expression and its relation to the definition of a Hermitian operator. The expert explains the role of the wave function and its representation in the basis of position eigenstates. Finally, it is shown that the operator is indeed Hermitian.
  • #1
happyparticle
394
20
Homework Statement
Is the operator ##\int_{-\infty}^{\infty} |x><x| dx## Hermitian
Relevant Equations
##\hat{Q} = \hat{Q}^{\dagger}##
##<f|\hat{Q}|g> = <\hat{Q}f|g> ##
Knowing that to be Hermitian an operator ##\hat{Q} = \hat{Q}^{\dagger}##.
Thus, I'm trying to prove that ##<f|\hat{Q}|g> = <\hat{Q}f|g> ##.
However, I don't really know what to do with this expression.
##<f|\hat{Q}g> = \int_{-\infty}^{\infty} [f(x)^* \int_{-\infty}^{\infty} |x> <x| dx f(x)] dx##

Am I on the right track?

Thank you
 
Last edited:
Physics news on Phys.org
  • #2
In general $$\bra f Q \ket g = \bra g Q^{\dagger} \ket f^*$$
 
  • #3
I'm facing the same issue. I'm not sure how to deal with the integral as operator.
I'm using Griffith's equation 3.20 ##
<f|\hat{Q}g> = <\hat{Q}f|g>
##
 
  • #4
happyparticle said:
However, I don't really know what to do with this expression.
##<f|\hat{Q}g> = \int_{-\infty}^{\infty} [f(x)^* \int_{-\infty}^{\infty} |x> <x| dx f(x)] dx##
I can only make sense of the operator ##\hat{Q}## if it operates on kets. So I would keep using the Dirac notation (no ##f(x)## or ##g(x)##, but ##\bra{f}## and ##\ket{g}##), while introducing the definition of the operator ##\hat{Q}##.
 
  • Like
Likes PeroK and malawi_glenn
  • #5
Write an arbitrary state ##\left|\psi\right.\rangle## in the basis of position eigenstates ##\left|x\right.\rangle## and show that the operator in the exercise acts on it like the identity operator, ##\hat{Q}\left|\psi\right.\rangle = \left|\psi\right.\rangle##. The identity operator is clearly hermitian, as any state is an eigenstate with eigenvalue 1. The wave function ##\psi(x)## is the multiplier of ##\left|x\right.\rangle## in the integral representation of the arbitrary state in basis of position eigenstates.
 
  • Like
Likes strangerep and malawi_glenn
  • #6
Alright, I still have few questions.

I'm doing what helbert2 said.
##\int_{-\infty}^{\infty} dx |x><x| \psi##

Is ##\int_{-\infty}^{\infty} dx (<x|)^t |x> \psi## = ##\int_{-\infty}^{\infty} dx <x|x> \psi = \psi##
correct? If so I'm not sure to understand why ##(<x|)^t = |x>##. I just saw that, but without further explanation.
 
  • #7
happyparticle said:
Alright, I still have few questions.

I'm doing what helbert2 said.
##\int_{-\infty}^{\infty} dx |x><x| \psi##

Is ##\int_{-\infty}^{\infty} dx (<x|)^t |x> \psi## = ##\int_{-\infty}^{\infty} dx <x|x> \psi = \psi##
correct? If so I'm not sure to understand why ##(<x|)^t = |x>##. I just saw that, but without further explanation.
By definition ##\bra x## is the bra associated with the ket ##\ket x##.
 
  • #8
Do you mean ##|x \rangle \langle x| = \langle x |x \rangle## ? I'm not so sure to see the link with my questions.
 
  • #9
happyparticle said:
Do you mean ##|x \rangle \langle x| = \langle x |x \rangle## ? I'm not so sure to see the link with my questions.
I definitely don't mean that. The LHS is an operator and the RHS is a number.

You need to revise what all these symbols mean. You start with a ket space, then introduce the idea of a dual space or bra space in Dirac notation.
 
  • #10
I'm not sure to understand what ##|x \rangle \langle x|## is exactly.
for a vector ##|x \rangle##

Is it correct to say that
##|x \rangle \langle x|## is a n x 2 matrix and ##\langle x |x \rangle## is a dot product?
 
  • #11
happyparticle said:
Alright, I still have few questions.

I'm doing what helbert2 said.
##\int_{-\infty}^{\infty} dx |x><x| \psi##

Is ##\int_{-\infty}^{\infty} dx (<x|)^t |x> \psi## = ##\int_{-\infty}^{\infty} dx <x|x> \psi = \psi##
correct? If so I'm not sure to understand why ##(<x|)^t = |x>##. I just saw that, but without further explanation.
The wave function ##\psi (x)## of state ##\left|\psi\right.\rangle## is defined as the inner product with a position eigenstate ##\left|x\right.\rangle##: ##\psi (x) = \langle \left.x\right|\psi\rangle##. This is visible in the calculation

##\displaystyle\langle \left.x\right|\psi\rangle = \langle \left.x\right|\int_{-\infty}^{\infty}\psi (x')\left|x'\right.\rangle dx' = \int_{-\infty}^{\infty}\psi (x')\langle \left.x\right|x'\rangle dx' = \int_{-\infty}^{\infty}\psi (x')\delta (x-x') dx' = \psi (x)##.

A position eigenstate ##\langle \left.x\right|## or ##\left|x\right.\rangle## can be moved in or out of an integral that doesn't integrate over variable ##x##. If you apply this and the property ##\langle x\left|x'\right.\rangle = \delta (x-x')## on the quantity ##\hat{O}\left|\psi\right.\rangle##, you see that it's the same as ##\left|\psi\right.\rangle##.
 
  • Like
Likes strangerep
  • #12
happyparticle said:
Homework Statement:: Is the operator ##\int_{-\infty}^{\infty} |x><x| dx## Hermitian
Relevant Equations:: ##\hat{Q} = \hat{Q}^{\dagger}##
##<f|\hat{Q}|g> = <\hat{Q}f|g> ##

Knowing that to be Hermitian an operator ##\hat{Q} = \hat{Q}^{\dagger}##.
Thus, I'm trying to prove that ##<f|\hat{Q}|g> = <\hat{Q}f|g> ##.
However, I don't really know what to do with this expression.
##<f|\hat{Q}g> = \int_{-\infty}^{\infty} [f(x)^* \int_{-\infty}^{\infty} |x> <x| dx f(x)] dx##

Am I on the right track?

Thank you
I don't know, how you get two integrals here since you have
$$\langle f|\hat{Q} g \rangle=\int_{\mathbb{R}} \mathrm{d} x \langle f |x \rangle \langle x |g \rangle = \int_{\mathbb{R}} f^*(x) g(x)=\langle f|g \rangle.$$
Since this holds for any ##f## and ##g##, this implies that ##\hat{Q}=...##?
 
  • Like
Likes Maarten Havinga
  • #13
happyparticle said:
I'm not sure to understand what ##|x \rangle \langle x|## is exactly.
for a vector ##|x \rangle##

Is it correct to say that
##|x \rangle \langle x|## is a n x 2 matrix and ##\langle x |x \rangle## is a dot product?
You really do need to go back to basics and digest this material.

The outer product ##\ket x \bra y## is an operator that maps a ket ##\ket z## to the ket ##(\ket x)\braket {y|z}##.

Note that it's usual to drop the round brackets here, which I've only included for emphasis.
 
Last edited:
  • Like
Likes vanhees71
  • #14
vanhees71 said:
I don't know, how you get two integrals here since you have
$$\langle f|\hat{Q} g \rangle=\int_{\mathbb{R}} \mathrm{d} x \langle f |x \rangle \langle x |g \rangle = \int_{\mathbb{R}} f^*(x) g(x)=\langle f|g \rangle.$$
Since this holds for any ##f## and ##g##, this implies that ##\hat{Q}=...##?
I thought because the operator was an integral and the inner product of 2 functions is an integral as well.

I didn't know ##
\psi (x) = \langle \left.x\right|\psi\rangle
##
If I understand griffith (p.90 third edition) this is like a dot product between 2 vectors?
 
  • #15
happyparticle said:
I thought because the operator was an integral and the inner product of 2 functions is an integral as well.

I didn't know ##
\psi (x) = \langle \left.x\right|\psi\rangle
##
If I understand griffith (p.90 third edition) this is like a dot product between 2 vectors?
You want to recognize that ##f(x)## and ##\lvert f \rangle##, while related, are not the same thing and therefore are not interchangeable. ##\lvert f \rangle## denotes a vector while ##f(x)## represents the vector in the position basis. In your initial attempt, you mixed these two types of objects up, which led to the confusing mess you started with.

Say you have two vectors ##\mathbf{a}## and ##\mathbf{b}## in an ##n##-dimensional vector space ##V##, and you want to calculate ##\mathbf{a}\cdot\mathbf{b}##. One way to do that it to first choose an orthonormal basis ##\{\mathbf{e}_1, \mathbf{e}_2, \dots, \mathbf{e}_n\}## and then represent ##\mathbf{a}## and ##\mathbf{b}## by the ##n##-tuples ##(a_1, a_2, \dots, a_n)## and ##(b_1, b_2, \dots, b_n)##. In other words, you have
\begin{align*}
\mathbf{a} = \sum_{i=1}^n a_i \mathbf{e}_i \\
\mathbf{b} = \sum_{i=1}^n b_i \mathbf{e}_i
\end{align*} Then the dot product would be given by
$$\mathbf{a}\cdot\mathbf{b} = \sum_{i=1}^n a_i b_i.$$

Compare that last expression to the one you learned for ##\langle f \vert g \rangle##.
$$\langle f \vert g \rangle = \int f^*(x)g(x)\,dx.$$ ##\langle f \rvert## and ##\lvert g \rangle## corresponds to ##\mathbf{a}## and ##\mathbf{b}##. The integration corresponds to summing over ##i##. And ##f(x)## and ##g(x)## correspond to the ##a_i##'s and ##b_i##'s.

Just as ##(a_1, a_2, \dots, a_n)## is merely a representation of ##\mathbf{a}## with respect to the orthonormal basis, ##f(x)## is merely a representation of ##\lvert f \rangle## in the position basis. In other words,
$$\lvert f \rangle = \int \psi(x) \lvert x \rangle\,dx.$$ Compare that expression to the sum above where ##\mathbf{a}## was expanded in terms of the chosen basis. The integral corresponds to the summation; ##\psi(x)## to ##a_i##, and ##\lvert x \rangle## to ##\mathbf{e}_i##.

Finally, note that
$$\mathbf{e}_i \cdot \mathbf{a} = \mathbf{e}_i \cdot \sum_{j=1}^n a_j \mathbf{e}_j = \sum_{j=1}^n a_j (\mathbf{e}_i \cdot \mathbf{e}_j) = \sum_{j=1}^n a_j \delta_{ij} = a_i.$$ The analogous calculation In Dirac notation is
$$\langle x \vert f \rangle = \langle x \rvert \int f(x')\lvert x' \rangle\,dx' =
\int f(x')\langle x \vert x' \rangle\,dx' =
\int f(x')\delta(x-x')\,dx' = f(x)$$

I glossed over some technicalities since you're generally dealing with complex numbers in quantum mechanics, but the gist should be clear. Dirac notation might look a little strange at first, but once you get used to it, you'll find it's a much more convenient way of expressing the ideas you learned in linear algebra.
 
Last edited:
  • Like
Likes vanhees71, happyparticle, PeroK and 1 other person
  • #16
happyparticle said:
I thought because the operator was an integral and the inner product of 2 functions is an integral as well.

I didn't know ##
\psi (x) = \langle \left.x\right|\psi\rangle
##
If I understand griffith (p.90 third edition) this is like a dot product between 2 vectors?
You should read about the foundations of the theory in terms of the Dirac formulation again. Griffiths's quantum textbook seems to lead to a lot of confusion, as this thread demonstrates once more. I recommend J. J. Sakurai, Modern Quantum Mechanics, Revised Edition
 
  • Like
Likes happyparticle
  • #17
Alright, thank you for the recommendation.
Thank you all for the help. Sorry if I'm slow to understand, thank you for your patience.
 
  • Like
Likes vanhees71
  • #18
Don't worry. To understand quantum theory it takes some time!
 

1. What does it mean for an operator to be Hermitian?

An operator is Hermitian if it is equal to its own conjugate transpose. In other words, if A is a Hermitian operator, then A = A, where A is the conjugate transpose of A.

2. Why is it important for an operator to be Hermitian?

Hermitian operators have many important properties that make them useful in quantum mechanics and other areas of physics. For example, they have real eigenvalues and their eigenvectors form an orthonormal basis. This allows us to easily solve certain mathematical problems and make predictions about the behavior of physical systems.

3. How can I determine if an operator is Hermitian?

To determine if an operator is Hermitian, you can check if it satisfies the definition of Hermitian, which is A = A. This can be done by calculating the conjugate transpose of the operator and comparing it to the original operator. Alternatively, you can also check if all of its eigenvalues are real.

4. Can a non-square operator be Hermitian?

No, a non-square operator cannot be Hermitian. The definition of a Hermitian operator requires it to be equal to its own conjugate transpose, which is only possible if the operator is square (has the same number of rows and columns).

5. Are all physical observables represented by Hermitian operators?

Yes, in quantum mechanics, all physical observables are represented by Hermitian operators. This is because the eigenvalues of a Hermitian operator correspond to the possible outcomes of a measurement, and the eigenvectors represent the states in which the observable has a definite value. Therefore, Hermitian operators play a crucial role in the study of physical systems.

Similar threads

  • Advanced Physics Homework Help
Replies
4
Views
1K
Replies
2
Views
1K
  • Advanced Physics Homework Help
Replies
4
Views
3K
  • Advanced Physics Homework Help
Replies
1
Views
170
  • Advanced Physics Homework Help
Replies
10
Views
396
  • Advanced Physics Homework Help
Replies
11
Views
1K
Replies
2
Views
3K
Replies
1
Views
724
  • Advanced Physics Homework Help
Replies
9
Views
1K
  • Advanced Physics Homework Help
Replies
1
Views
1K
Back
Top