# Find the divergence of the function

## Homework Statement

Let V = (sin(theta)cos(phi))/r Determine:
(a) ∇V
(b) ∇ x ∇V
(c) ∇∇V

## The Attempt at a Solution

#### Attachments

• EPSON015.jpg
9.1 KB · Views: 457

rude man
Homework Helper
Gold Member
Your formulas are all wrong. You need to get the right formulas as you finally did with your gradient exercises.

rude man
Homework Helper
Gold Member

No time to check it thoroughly now, but 2 cooents:
1. you're using i j k again when you shouldn't.
2. curl(grad f) = 0 always. The curl of the gradient of a function is always zero. (Not all functions have gradients).
3 your writing is hard for me to read but it looks like you wrote grad when you meant div.

Is my formula even close to correct? If I take off I,j, and k and does it make it a curl when you take the cross product of a gradient?

Sorry that was one long run on sentence. I meant to say will it be ok if I take off I,j, and k?

rude man
Homework Helper
Gold Member
Sorry that was one long run on sentence. I meant to say will it be ok if I take off I,j, and k?

No. The gradient of a function is a vector. Therefore you need the three unit vectors. Do't use i j k unless you're in an xyz cartesian coordinate system. Use the ones I gave you previously.

The curl can only be taken of a vector. There is no such thing as the curl of a scalar.

It's a mathematical identity that curl (grad f) = 0 for any f. It's very good to remember that identity.

Ok is my formula ok? I'm gonna go look again but I believe Wikipedia had that listed I can add unit vectors that's no problem.

rude man
Homework Helper
Gold Member
Ok is my formula ok? I'm gonna go look again but I believe Wikipedia had that listed I can add unit vectors that's no problem.

If wikipedia listed it I'm sure it was right. But they must have had the unit vectors too?
(Now I know how you did so well with the gradients, right?) Ugggg that is all Wikipedia lists I'm gonna go to my instructors web page and check there

WannabeNewton
##\hat{i},\hat{j},\hat{k}## are just labels for unit vectors. They traditionally represent the unit vectors in Cartesian coordinates but you can call them whatever you want, they're just labels. Generally one would write ##\hat{e}_1,\hat{e}_2,\hat{e}_3## to represent unit vectors in arbitrary coordinates. If you are using spherical coordinates, then you need to use the correct formulas for ##\nabla f, \nabla \cdot V, \nabla \times V, \nabla^2 f## in spherical coordinates. You can find them easily online; for example ##\nabla = \hat{e}_r\partial_r + \frac{1}{r}\hat{e}_{\theta}\partial_{\theta} + \frac{1}{r\sin\theta} \hat{e}_{\varphi}\partial_{\varphi}##.

rude man
Homework Helper
Gold Member
Ugggg that is all Wikipedia lists I'm gonna go to my instructors web page and check there

Never mind, I screwed you up & forgot you're now doing divergences, not gradients.

Divergence of a vector is a scalar so you don't attach the unit vectors.

Sorry!

rude man
Homework Helper
Gold Member
actually I screwed up the problem never said anything about divergence! That being said I corrected what I did and have uploaded it hopefully its ok!!

Oh, OK. I think I mistook this for another of your threads whichI think did have div's in them.

I am too lazy to check all your math steps but for curl[grad(f)] the answer must be zero so I'm afraid something went awry there.

Why don't you run grad(f) thru wolfram alpha to make sure you got that part right.

ok how do I use wolfram for gradients?

rude man
Homework Helper
Gold Member
ok how do I use wolfram for gradients?

Stand by, I'll give it a shot, haven't done that myself yet ...

EDIT: OK, punch in " gradient of sin(theta)*cos(phi)/r " in their top window.

They don't include the unit vectors. They give you the r, theta and phi components in that order, separated by commas, instead. Dumb, but better than screwing up solving it ourselves!

Last edited:
rude man
Homework Helper
Gold Member
ok uploaded something else I can could use some assistance with

Back in 1 hr.

Evaluate the divergence of the following:
(A) A=xyUx+y^2Uy-xzUz
(B) B=pz^2Up+psin^2(phi)Uphi+2pzsin^2(phi)Uz
(C) C=rUr+rcos^2(theta)Uphi

(A) y+2y-x
(B) z^2+sin(2phi)+2psin^2(phi)
(C) 1

rude man
Homework Helper
Gold Member
(A) y+2y-x
(B) z^2+sin(2phi)+2psin^2(phi)
(C) 1

Unfortunately, B and C are incorrect. They are tricky so work carefully. And check your results with wolfram.

For example, B=pz2Up + psin2(phi)Uphi + 2pzsin2(phi)Uz

so what is ∂Bρ/∂ρ + Bρ/ρ? Not z2.

Last edited:
div B=pz^2+psin^2(phi)+2pzsin^2(phi) I entered that in wolfram it didnt work what syntax should I use.

rude man
Homework Helper
Gold Member
div B=pz^2+psin^2(phi)+2pzsin^2(phi) I entered that in wolfram it didnt work what syntax should I use.

look at what I came up with for the gradient and improvise! Play around with it for a while. I pointed out that it gave the gradient in comma-separated chunks, omitting the unit vectors. Wolfram will have examples of how to enter a vector to get the divergence.

I would suggest using r instead of rho.

Bp/p

Whoops I meant what is Bp/p was that in the formula

div (rho(z)^2,(rho)sin^2(phi),2(rho)zsin^2(phi)) tried this on wolfram worked but gave the answer in Cartesian

rude man
Homework Helper
Gold Member
div (rho(z)^2,(rho)sin^2(phi),2(rho)zsin^2(phi)) tried this on wolfram worked but gave the answer in Cartesian

Did you look carefully? When I ran grad V3 it gave it in cartesian but also in spherical.

rude man
Homework Helper
Gold Member
Whoops I meant what is Bp/p was that in the formula

sure was! ∂Bρ/∂ρ + Bρ/ρ is the Bρ part of div B.

Hmm Bp/p isn't a derivative I am failing to see what I can do with this. I am going to go on a long shot here and guess that the unit vectors cancel as they multiply together, and p cancels the other p and leaves me with z^2 again? Then z^2+z^2 heh that can't be right, what am doing wrong anymore hints?