Is there a simpler way to write the cross product of curl?

Click For Summary

Discussion Overview

The discussion revolves around finding a simpler expression for the cross product of a vector field \(\vec B\) with the curl of another vector field \(\vec A\). Participants explore various mathematical manipulations and identities related to this expression, focusing on theoretical aspects and vector calculus identities.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Exploratory

Main Points Raised

  • One participant seeks a neat way to express \(\vec B \times (\vec \nabla \times \vec A)\) and presents calculations for the x-component, suggesting a potential expression involving the dot product of \(\vec B\) and the gradient of \(\vec A\).
  • Another participant references an identity from Griffiths' "Introduction to EM" that relates the gradient of the dot product of two vectors to their curls and gradients, hinting at possible simplifications.
  • A third participant manipulates the expression using the Levi-Civita symbol and arrives at a form that separates terms involving \(\vec B\) and \(\nabla A_m\), indicating a potential simplification.
  • A fourth participant advises the use of Cartesian tensors when proving vector identities, suggesting a methodological approach to the problem.

Areas of Agreement / Disagreement

Participants present various approaches and manipulations without reaching a consensus on a single simplified expression. Multiple competing views and methods remain in the discussion.

Contextual Notes

Some mathematical steps and assumptions are not fully resolved, and the discussion relies on specific vector calculus identities that may have limitations based on context or definitions.

Galileo
Science Advisor
Homework Helper
Messages
1,980
Reaction score
7
Is there any neat way/rule to write:

[tex]\vec B \times (\vec \nabla \times \vec A)[/tex]
?

I've tried it myself and found (e.g) for the x-component:

[tex]\left(B_x\frac{\partial A_x}{\partial x}+B_y\frac{\partial A_y}{\partial x}+B_z\frac{\partial A_z}{\partial x}\right)-\left(B_x\frac{\partial A_x}{\partial x}+B_y\frac{\partial A_x}{\partial y}+B_x\frac{\partial A_x}{\partial z}\right)[/tex]

I can write the last terms with the minus sign as: [itex]\vec B \cdot \nabla A_x[/itex], but I can't find a way to do something nice to the first term, except maybe:

[tex]\left(\vec B \cdot \frac{\partial}{\partial x}\vec A\right)[/tex]
I've never seen such an expression before though.
The other 2 components are similar:
[tex]\left[\vec B \times (\vec \nabla \times \vec A)\right]_y=\left(\vec B \cdot \frac{\partial}{\partial y}\vec A\right)-\left(\vec B \cdot \nabla A_y\right)[/tex]
[tex]\left[\vec B \times (\vec \nabla \times \vec A)\right]_z=\left(\vec B \cdot \frac{\partial}{\partial z}\vec A\right)-\left(\vec B \cdot \nabla A_z\right)[/tex]

I figured I may see something if I combined them all into the general expression:

[tex]\left(B_x\frac{\partial A_x}{\partial x}+B_y\frac{\partial A_y}{\partial x}+B_z\frac{\partial A_z}{\partial z}\right)\hat x +\left(B_x\frac{\partial A_x}{\partial y}+B_y\frac{\partial A_y}{\partial y}+B_z\frac{\partial A_z}{\partial y}\right)\hat y+\left(B_x\frac{\partial A_x}{\partial z}+B_y\frac{\partial A_y}{\partial z}+B_z\frac{\partial A_z}{\partial z}\right)\hat z-(\vec B \cdot \vec \nabla)\vec A[/tex]
There's definitely a pattern in the first 3 terms, but the best I could come up with is writing these terms as:
[tex]B_x\nabla A_x+B_y\nabla A_y+B_z\nabla A_z[/tex]
That has condensed it a lot. Looks like a dot product with B, but...
 
Physics news on Phys.org
identity: (from Griffiths, Introduction to EM)

grad(A dot B)=A cross (curl B) + B cross (curl A) + (A dot grad)B + (B dot grad)A
 
Playing around with it more:

[tex]\vec B \times(\nabla \times \vec A)<br /> =\epsilon_{ijk}B_j ( \epsilon_{klm}\nabla_l A_m)<br /> =B_m \nabla_i A_m - B_l \nabla_l A_i[/tex]
where I used
[tex]\epsilon_{ijk}\epsilon_{klm}=\delta_{il}\delta_{jm}-\delta_{im}\delta_{jl}[/tex]

So, you've essentially got it.
 
HINT:ALWAYS use cartesian tensors when proving vector identities...With objects from R^{n},of course.

Daniel.
 

Similar threads

  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 3 ·
Replies
3
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 10 ·
Replies
10
Views
1K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 3 ·
Replies
3
Views
4K