- #1
Pouramat
- 28
- 1
Summary:: hodge Duality...
Does anybody has idea for solving these 2 problems?
Does anybody has idea for solving these 2 problems?
Last edited by a moderator:
TSny said:See the errata for Carroll's textbook corresponding to page 92 as given below. It fixes the wording of problem 9.
View attachment 274846
got this in my solution.
And did you solve it in 4-dim Minkowski space? I want to know the coorect solution and answer. tnxTSny said:See the errata for Carroll's textbook corresponding to page 92 as given below. It fixes the wording of problem 9.
View attachment 274846
In the homework forum, we require that you show your work. We don't provide answers, but we can try to provide some guidance.Pouramat said:And did you solve it in 4-dim Minkowski space? I want to know the coorect solution and answer. tnx
This is not the correct answer. Can you explain how you got this result? This will help us see how you are thinking about the problem.Pouramat said:9.a. I did this part and get the answer as q sin theta dr d theta d phi
##q## is just a constant. So, ##\dot q = 0##.Pouramat said:$$
d\star F = (q \sin \theta) dr \wedge d \theta \wedge d \phi + (\dot q \cos \theta +q (d/dt (cos \theta) )) dt \wedge d \theta \wedge d \phi
$$
but I 'm not sure about my solution.
##F## is related to ##\ast (\ast F)##. See equation (2.83) in Carroll: ##\ast \ast A \equiv \ast (\ast A) = (-1)^{s+p(n-p)}A ##In part (b) I have no idea how to read F form *F. can you help me?
Dear TSnyTSny said:##q## is just a constant. So, ##\dot q = 0##.
The time derivative ##d/dt (\cos \theta)## should be a partial derivative ##\large \frac{\partial \cos \theta}{\partial t}##. What does this evaluate to? How did ##\sin \theta## become ##\cos \theta## in the last part of your answer?
Shouldn't there be a partial derivative with respect to ##r## in the first part of your answer?
See Example 1 here.
##F## is related to ##\ast (\ast F)##. See equation (2.83) in Carroll: ##\ast \ast A \equiv \ast (\ast A) = (-1)^{s+p(n-p)}A ##
So, try to work out ##\ast\ast F## by applying the Hodge star operator to ##\ast F##.
Ok. But, what is the value of ##\frac{\partial \sin \theta }{\partial t}##?Pouramat said:Dear TSny
so the correct answer is :
$$
d \ast F = q \frac{\partial \sin \theta }{\partial t} dt \wedge d \theta \wedge d \phi
$$
The exponents on the ##(-1)## look good. And the occurrence of ##dt \wedge dr## also looks good. But the ##\sin \theta## is not correct. Things get a little complicated because we're working in spherical coordinates. From Carroll's equation (2.82) $$\left(\ast \ast F\right)_{\alpha \beta} = \frac{1}{2!}{\epsilon^{\mu \nu}}_{\alpha \beta} (\ast F)_{\mu \nu}$$and for part (b),
$$
\ast \ast F = (-1)^{1+2(2)}(q sin \theta) dt \wedge dr
$$
Thank you. But you might want to hold that thought.You are a good teacher...
Daer TSny;TSny said:Ok. But, what is the value of ##\frac{\partial \sin \theta }{\partial t}##?
The exponents on the ##(-1)## look good. And the occurrence of ##dt \wedge dr## also looks good. But the ##\sin \theta## is not correct. Things get a little complicated because we're working in spherical coordinates. From Carroll's equation (2.82) $$\left(\ast \ast F\right)_{\alpha \beta} = \frac{1}{2!}{\epsilon^{\mu \nu}}_{\alpha \beta} (\ast F)_{\mu \nu}$$
Since you know that the only nonzero components of ##\ast F## are ##(\ast F)_{\theta \phi}## and ##(\ast F)_{\phi \theta}##, convince yourself that the above equation reduces to $$\left(\ast \ast F\right)_{\alpha \beta} = {\epsilon^{\theta \phi}}_{\alpha \beta} (\ast F)_{\theta \phi}$$ So, now we have to work out ##{\epsilon^{\theta \phi}}_{\alpha \beta} ##. Write it as
$${\epsilon^{\theta \phi}}_{\alpha \beta} = g^{\theta \rho} g^{\phi \sigma} \epsilon_{\rho \sigma \alpha\beta} $$ Work this out using what you know about the metric components in spherical components and using Carroll's equation (2.69), $$\large \epsilon_{\rho \sigma \alpha\beta} = \sqrt{|g|} \, \tilde \epsilon_{\rho \sigma \alpha\beta}$$ Here, ##\large \tilde \epsilon_{\rho \sigma \alpha\beta}## is the Levi-Civita symbol defined in Carroll's equation (2.65). See what you get.Thank you. But you might want to hold that thought.
Yes. So, ##d(\ast F) = 0## at all spacetime points except possibly at the origin of the spherical coordinate system. ##\sin \theta## is not well-defined at the origin, so I don't think we can necessarily say that ##d(\ast F) = 0## at the origin. This might be relevant to the result that you will get for the electric field in part (d).Pouramat said:(a) $$
d(\ast F) = \partial_t (q \sin \theta) dt \wedge d \theta \wedge d \phi + \partial_r (q \sin \theta) dr \wedge d \theta \wedge d \phi = 0$$
Ok(b)
I took ##\ast F = A## so:
$$
(\ast A)_{\rho \sigma} = \frac{1}{p!} {\epsilon^{\mu \nu}}_{\rho \sigma} A_{\mu \nu}
$$
This is not correct. Note that you have the index ##\mu## appearing three times on the right hand side. Likewise for ##\nu##. Also, ##\alpha## and ##\beta## do not appear on the left. So, they each need to appear twice as dummy summation indices on the right side. Consider the example ##T^{\mu \nu} = g^{\mu \alpha} g^{\nu \beta} T_{\alpha \beta}##. See if you can correct your expression for ##(\ast A)_{\rho \sigma}##now I should lower the indices of ##\epsilon##, that is:
$$
(\ast A)_{\rho \sigma} = \frac{1}{2!} g^{\mu \alpha} g^{\nu \beta}\epsilon _{\mu \nu \rho \sigma} A_{\mu \nu}
$$
Yes, except for the sign of the first "1" in ##diag (1,1,r^2 , r^2 \sin^2 \theta)##.because we are working in spherical coordinate:
$$
g_{\mu \nu} = diag (1,1,r^2 , r^2 \sin^2 \theta) ; \sqrt{|g|} = r^2 \sin \theta
$$
The matrix with elements ##g^{\mu \nu}## is the inverse of the matrix with elements ##g_{\mu \nu}##. Since the matrix with elements ##g_{\mu \nu}## is diagonal, the matrix with elements ##g^{\mu \nu}## will also be diagonal and the diagonal elements are ##g^{\mu \mu} = 1/g_{\mu \mu}##.I really do'nt know how to evaluate ##g^{\mu \alpha} g^{\nu \beta}##
Ooooh yes I Know:TSny said:Yes. So, ##d(\ast F) = 0## at all spacetime points except possibly at the origin of the spherical coordinate system. ##\sin \theta## is not well-defined at the origin, so I don't think we can necessarily say that ##d(\ast F) = 0## at the origin. This might be relevant to the result that you will get for the electric field in part (d).
Ok
This is not correct. Note that you have the index ##\mu## appearing three times on the right hand side. Likewise for ##\nu##. Also, ##\alpha## and ##\beta## do not appear on the left. So, they each need to appear twice as dummy summation indices on the right side. Consider the example ##T^{\mu \nu} = g^{\mu \alpha} g^{\nu \beta} T_{\alpha \beta}##. See if you can correct your expression for ##(\ast A)_{\rho \sigma}##
Yes, except for the sign of the first "1" in ##diag (1,1,r^2 , r^2 \sin^2 \theta)##.The matrix with elements ##g^{\mu \nu}## is the inverse of the matrix with elements ##g_{\mu \nu}##. Since the matrix with elements ##g_{\mu \nu}## is diagonal, the matrix with elements ##g^{\mu \nu}## will also be diagonal and the diagonal elements are ##g^{\mu \mu} = 1/g_{\mu \mu}##.
I believe your final result is correct even though I don't quite follow your individual steps. The subscripts ##\rho \sigma## should be ##t r## on the left side of your last two equations. Your final result will simplify greatly.Pouramat said:Ooooh yes I Know:
$$
(\ast A)_{\rho \sigma} = \frac{1}{p!} g^{\mu \alpha} g^{\nu \beta} \epsilon_{\alpha \beta \rho \sigma} A_{\mu \nu}
$$
$$
(\ast A)_{\rho \sigma} =\frac{1}{2!} \sqrt{|g|} g^{\mu \alpha} g^{\nu \beta} \tilde\epsilon_{\alpha \beta \rho \sigma} A_{\mu \nu}
$$
so here I should put the exact elemet in ##g^{\mu \alpha}## and ##g^{\nu \beta}## and sum over?
because ##g^{\mu \nu}## is diagonal so:
$$
(\ast A)_{\rho \sigma} =\frac{1}{2!} \sqrt{|g|} g^{\mu \mu} g^{\nu \nu} \tilde\epsilon_{\mu \nu \rho \sigma} A_{\mu \nu}
$$
ok, now by replacing the indices we have:
$$
(\ast A)_{\rho \sigma} =\frac{1}{2!} (r^2 \sin \theta) g^{\theta \theta} g^{\phi \phi} \tilde\epsilon_{\theta \phi t r} A_{\theta \phi}
$$
$$
(\ast A)_{\rho \sigma} =\frac{2}{2!} (r^2 \sin \theta) (\frac{1}{r^2}) (\frac{1}{r^2 \sin^2}) (q \sin \theta)
$$
the factor 2 comes from the change of indices ##\theta## with ##\phi## and r with t.
Finally we get the expression for ##F## as following:TSny said:I believe your final result is correct even though I don't quite follow your individual steps. The subscripts ##\rho \sigma## should be ##t r## on the left side of your last two equations. Your final result will simplify greatly.
Here's another way to get this. Start with
$$
(\ast A)_{\rho \sigma} =\frac{1}{2!} \sqrt{|g|} g^{\mu \alpha} g^{\nu \beta} \tilde\epsilon_{\alpha \beta \rho \sigma} A_{\mu \nu}
$$ We know that all ##A_{\mu \nu} = 0## except ##A_{\theta \phi}## and ##A_{\phi \theta}##. So, we have $$(\ast A)_{\rho \sigma} =\frac{1}{2!}\left[ \sqrt{|g|} g^{\theta \alpha} g^{\phi \beta} \tilde\epsilon_{\alpha \beta \rho \sigma} A_{\theta \phi} + \sqrt{|g|} g^{\phi \alpha} g^{\theta \beta} \tilde\epsilon_{\alpha \beta \rho \sigma} A_{\phi \theta} \right ]$$ Using the antisymmetry properties of ##A_{\mu \nu}## and ##\tilde\epsilon_{\alpha \beta \rho \sigma}##, the two terms in the bracket are equal. So, $$(\ast A)_{\rho \sigma} = \sqrt{|g|} g^{\theta \alpha} g^{\phi \beta} \tilde\epsilon_{\alpha \beta \rho \sigma} A_{\theta \phi} $$ Since the metric tensor is diagonal, this reduces to $$(\ast A)_{\rho \sigma} = \sqrt{|g|} g^{\theta \theta} g^{\phi \phi} \tilde\epsilon_{\theta \phi \rho \sigma} A_{\theta \phi} $$
Therefore, the only nonzero elements of ##(\ast A)_{\rho \sigma}## are ##(\ast A)_{t r} = -(\ast A)_{r t}##, where $$(\ast A)_{t r} = \sqrt{|g|} g^{\theta \theta} g^{\phi \phi} \tilde\epsilon_{\theta \phi t r} A_{\theta \phi} = \sqrt{|g|} g^{\theta \theta} g^{\phi \phi} A_{\theta \phi} $$ which agrees with your result.
So, after simplification, what do you get for the components of the two-form ##F##?
Ok. Looks good except that the factor of ##q## is missing.Pouramat said:Finally we get the expression for ##F## as following:
$$
(F)_{tr} = \frac{1}{r^2}
$$
We should not forget the minus sign we'd evaluated before:
$$
(\ast \ast F)_{\alpha \beta} = (-1)^{-1+2(2)} {\epsilon^{\mu \nu}}_{\alpha \beta}(\ast F)_{\mu \nu}
$$
$$
F = -\frac{1}{r^2} dt \wedge dr = \frac{1}{r^2}dr \wedge dt
$$
In evaluating the partial derivative ##\partial_t(\sin \theta) ##, the variables ##r##, ##\theta##, and ##\phi## are held constant. So, ##\partial_t(\sin \theta) = 0##.and a point about part (a):
$$
\partial_t(\sin \theta) = \frac{\partial \theta}{\partial t} \frac{\partial}{\partial \theta}(\sin \theta ) = \dot \theta (\sin \theta)
$$
$$
d(\ast F) = \partial_t (q \sin \theta) dt \wedge d \theta \wedge d \phi + \partial_r (q \sin \theta) dr \wedge d \theta \wedge d \phi = 0
$$
Can I ask your opinion about integration in part (d), you can see my attempt at solution in post #13
The volume ##V## is a ball in 3-dimensional space. So, the integral on the left is over a 3-dimensional region. You then wrote that as an integral over a 2-dimensional region parameterized by ##\theta## and ##\phi##. If the ball is centered on the origin, then I think that what you wrote is correct. But can you state the theorem that you used to go from a 3-dimensional integral to a 2-dimensional integral? And can you describe the 2-dimensional region that you are integrating over? The integration should be a definite integral with specific limits for the variables ##\theta## and ##\phi##.(d) Integration in Euclidean 3 space:
$$
\int_V d \ast F = q\int \sin \theta d\theta d\phi = -q \phi \cos \theta
$$
The signs can be confusing here. From Carroll's equation (1.94), ##F^{0i} = E^i##. Lowering the index ##i## on both sides gives ##F_{0i} = E_i## (without a negative sign on the right side). The usual way to express the components of the electric field vector ##\mathbf E## would be to use upper indices ##E^{i}##. See Carroll's equation (1.93), for example. Thus, ##\mathbf E = E^r \hat r + E^{\theta} \hat \theta + E^{\phi} \hat \phi##. Here, ##E^r = F^{01} = F^{tr} = -F_{tr}##.Pouramat said:Using Field strength tensor we can read electrical and magnetic indices as following:
$$
F_{\mu \nu} =
\begin{pmatrix}
0 & -E_1 & -E_2 & -E_3 \\
E_1 & 0 & B_3 & -B_2 \\
E_2 & -B_3 & 0 & B_1\\
E_3 & B_2 & -B_1 & 0
\end{pmatrix}
$$
and now ##F## is :
$$ F_{\mu \nu}= q
\begin{pmatrix}
0 & \frac{1}{2r^2} & 0 & 0 \\
\frac{-1}{2r^2} & 0 & 0 & 0 \\
0 & 0 & 0 & 0\\
0 & 0 & 0 &0
\end{pmatrix}
$$
It seems that we just have ##E_1 = \frac{-q}{2r^2}## (which is correct in its dimension).
Yes(d)About integration, I think it is sounds some how like Stokes' theorem:
$$
\int_V d\omega = \int_{\partial V} \omega
$$
which ##\phi## goes from 0 to ##2 \pi## and ##\theta ## goes from 0 to ##\pi##
I think you are right. Thanks for being such a star.TSny said:I'm still not following how you are handling the signs. In many applications outside of GR, we use lower indices for the components of vectors. Thus, the cartesian components of the velocity of a particle might be written ##V_x##, ##V_y##, and ##V_z##. However, in tensor analysis, we distinguish between vectors and one-forms (or dual vectors) as in sections 1.4 and 1.5 of Carroll. Components of vectors are written with upper indices, as in Carroll's equation (1.37) while the components of one-forms are written with subscripts, as in Carroll's (1.47). In Carroll, the components of the electric field vector in spherical coordinates would be denoted with upper indices as ##E^r##, ##E^\theta##, and ##E^\phi##.
So, I think part (c) of problem 9 is asking for these upper-index quantities. And these are given by Carroll's equation (1.94): ##E^i = F^{0i}##. So, for example, part of the answer to part (c) would be to give the radial component of the electric field: ##E^r = F^{01} = F^{tr}##.
Recall that you found ## F = - \large \frac{q}{r^2} \normalsize dt \wedge dr .\,\,\,\,## So you know that ##F_{tr} = - \large \frac{q}{r^2}##.
Therefore, to find the radial component of the electric field vector, you just need to raise the indices ##t## and ##r## in order to get ##F^{tr}##.
The confusing thing is that ##\vec{E}## and ##\vec{B}## are only vector components under rotations but not under the full Lorentz group. That's why one usually uses the formalism used above, i.e., the ##(1/2,1/2)##-representation of the Lorentz group, i.e., the antisymmetric Faraday tensor with the corresponding components.TSny said:The signs can be confusing here. From Carroll's equation (1.94), ##F^{0i} = E^i##. Lowering the index ##i## on both sides gives ##F_{0i} = E_i## (without a negative sign on the right side). The usual way to express the components of the electric field vector ##\mathbf E## would be to use upper indices ##E^{i}##. See Carroll's equation (1.93), for example. Thus, ##\mathbf E = E^r \hat r + E^{\theta} \hat \theta + E^{\phi} \hat \phi##. Here, ##E^r = F^{01} = F^{tr} = -F_{tr}##.
How did you get a factor of ##2## in the denominator? You should end up with a very familiar electric field.
Yes
A manifold is a mathematical object that is locally similar to Euclidean space. It can be thought of as a curved space that can be described using coordinates, similar to latitude and longitude on a globe.
These questions help to develop a deeper understanding of the properties and structure of manifolds, which are essential concepts in mathematics and physics. They also serve as important exercises for students to practice applying the concepts learned in the chapter.
Manifolds are used in physics to describe the geometry of spacetime, particularly in the theory of general relativity. They allow for the mathematical description of curved spaces, which is necessary for understanding the effects of gravity.
A smooth manifold is one that is infinitely differentiable, meaning that it has no sharp corners or edges. A differentiable manifold may have some points where it is not smooth, such as a kink or a cusp.
Manifolds have both theoretical and practical applications. In addition to their use in physics, they are also used in fields such as computer graphics, robotics, and computer vision. They are also used in data analysis and machine learning to represent complex data sets.