Is the Chain Rule Applied to Spherical Polar Coordinates Different?

physconomics
Messages
22
Reaction score
0
Homework Statement
##x=r sin\theta cos\phi, y=rsin\theta sin\phi, z=rcos\theta##
Find ##\delta f/\delta x##and##\delta r/\delta x## treating r as a function of the cartesian coordinates
Given f is a function of r only, show
##\delta f^2/\delta x = (x/r)(df/dr)##
and
##\delta ^2 *f/\delta x^2 = (1/r)(df/dr) + (x^2/r)(d/dr)(1/r)(df/dr)##
Relevant Equations
?
Ive found ##\delta x/\delta r## as ##sin\theta cos\phi##
##\delta r/\delta x## as ##csc\theta sec\phi##
But unsure how to do the second part? Chain rule seems to give r/x not x/r?
 
Last edited:
on Phys.org
physconomics said:
Ive found ##\partial f\over \partial x## as ##\sin\theta \cos\phi##
You can't mean that. Independent of ##f## ?
 
physconomics said:
Ive found ##\delta f/\delta x## as ##sin\theta cos\phi##
##\delta r/\delta x## as ##csc\theta sec\phi##
But unsure how to do the second part? Chain rule seems to give r/x not x/r?

Please show your actual work, not just your end result.

Also, note that ##\partial## is written \partial in LaTeX.
 
BvU said:
You can't mean that. Independent of ##f## ?
Sorry, I meant ##\frac{\partial x}{\partial r}##
 
physconomics said:
Sorry, I meant ##\frac{\partial x}{\partial r}##
It is not true for partial derivatives that ##\partial x/\partial r = (\partial r/\partial x)^{-1}##
 
Orodruin said:
Please show your actual work, not just your end result.
And then we will try to work our way through this very relevant exercise :smile: .
 
Orodruin said:
It is not true for partial derivatives that ##\partial x/\partial r = (\partial r/\partial x)^{-1}##
Okay, so how else would I get ##\partial x/\partial r##?
 
Invert the transformation
$$x=r \sin\theta \cos\phi, \quad y=r\sin\theta \sin\phi, \quad z=r\cos\theta$$ to $$r = ...(x, y,z),\quad \phi = ...,\quad \theta = ...$$

(advice: use \sin and \cos, \log etc in ##\LaTeX##)
 
BvU said:
Invert the transformation

x=rsinθcosϕ,y=rsinθsinϕ,z=rcosθx=rsin⁡θcos⁡ϕ,y=rsin⁡θsin⁡ϕ,z=rcos⁡θ​

to

r=...(x,y,z),ϕ=...,θ=...r=...(x,y,z),ϕ=...,θ=...​
(advice: use \sin and \cos, \log etc in LATEXLATEX)
Okay so I get
r=√x2+y2+z2r=x2+y2+z2
##\theta = \cos^{-1} (\frac {z} {\sqrt{x^2 + y^2 + z^2})##
ϕ=tan−1(yx)ϕ=tan−1⁡(yx)

Am I right in thinking this makes

∂r∂x=sinθcosϕ∂r∂x=sin⁡θcos⁡ϕ?
 
  • #10
But there are tricks available -- ah you are on a good track. Differentiating ##r^2 = x^2 + y^2 + z^2## is easier to deal with.

Use \sqrt { ...} and always preview to check parentheses are right.
 
  • #11
BvU said:
But there are tricks available -- ah you are on a good track. Differentiating ##r^2 = x^2 + y^2 + z^2## is easier to deal with.

Use \sqrt { ...} and always preview to check parentheses are right.
Oh god of course, thank you.
So ##\frac{\partial r}{\partial x} = \frac{x}{r}##
And then from that I can show
##\frac{\partial f}{\partial x} = \frac{\partial f}{\partial r} \frac{\partial r}{\partial x} = \frac{df}{dr}\frac{x}{r}##
Im guessing I just use the chain rule for the next part?

Sorry for the state of my LaTeX maths
 
  • #12
Good ! Keep going ...

physconomics said:
Sorry for the state of my LaTeX maths
No need. You are obviously eager to learn and we are equally eager to assist.
 
  • #13
BvU said:
Good ! Keep going ...

No need. You are obviously eager to learn and we are equally eager to assist.

Right so I'm a bit stuck, I'm getting ##\frac{\partial^2 f}{\partial x^2} = \frac{\partial}{\partial x}(\frac{df}{dr}\frac{x}{r}) = \frac{\partial}{\partial x}(\frac{df}{dr}) \frac{x}{r} + \frac{df}{dr}\frac{1}{r} ##
where
##\frac{\partial}{\partial x}(\frac{df}{dr}) = \frac{\partial}{\partial x}(\frac{r}{x} \frac{\partial f}{\partial x}) = (\frac{-r}{x^2} \frac{\partial f}{\partial x}) + (\frac{r}{x}\frac{\partial^2 f}{\partial x^2})##
Is this correct? It doesn't lead me to the right answer
 
  • #14
BvU said:
Invert the transformation
$$x=r \sin\theta \cos\phi, \quad y=r\sin\theta \sin\phi, \quad z=r\cos\theta$$ to $$r = ...(x, y,z),\quad \phi = ...,\quad \theta = ...$$

(advice: use \sin and \cos, \log etc in ##\LaTeX##)

It should be noted that you can do this without actually inverting the coordinate relations. The general theory is based on the chain rule. For a function of one parameter only, say ##y(x)##, you would have
$$
1 = \frac{dx}{dx} = \frac{dy}{dx} \frac{dx}{dy}.
$$
This is the reason that ##dx/dy = (dy/dx)^{-1}## in this case (just solve for ##dx/dy##). However, for functions of several variables, say ##y^i(x^1, \ldots, x^n)##, the chain rule would instead give you
$$
\delta^i_j = \frac{\partial y^i}{\partial y^j} = \sum_{k} \frac{\partial y^i}{\partial x^k} \frac{\partial x^k}{\partial y^j}.
$$
Thus, from here you cannot just divide by a factor from the RHS to obtain ##\partial x^k/\partial y^j## since what you have is a sum. However, if you define the matrices ##A## and ##B## defined as
$$
A = \begin{pmatrix}\frac{\partial x^1}{\partial y^1} & \frac{\partial x^1}{\partial y^2} & \ldots \\
\frac{\partial x^2}{\partial y^1} & \frac{\partial x^2}{\partial y^2} & \ldots \\
\vdots & \vdots & \ddots \end{pmatrix}, \qquad
B = \begin{pmatrix}\frac{\partial y^1}{\partial x^1} & \frac{\partial y^1}{\partial x^2} & \ldots \\
\frac{\partial y^2}{\partial x^1} & \frac{\partial y^2}{\partial x^2} & \ldots \\
\vdots & \vdots & \ddots \end{pmatrix},
$$
then the relation can be rewritten
$$
1_n = BA,
$$
where ##1_n## is the ##n \times n## unit matrix. In other words, ##A = B^{-1}##. Thus, if you compute all the derivatives ##\partial y^i/\partial x^j##, you can find ##\partial x^i/\partial y^j## by matrix inversion.

However, if you are unfamiliar with matrices, it is also fine to solve the linear system of equations. For example, in polar coordinates on ##\mathbb R^2##: ##x = r\cos\phi##, ##y = r\sin\phi##, you would find
$$
1 = \frac{\partial x}{\partial x} = \frac{\partial x}{\partial r} \frac{\partial r}{\partial x} + \frac{\partial x}{\partial \phi} \frac{\partial \phi}{\partial x} = \cos\phi \frac{\partial r}{\partial x} - r \sin\phi \frac{\partial \phi}{\partial x}
$$
and
$$
0 = \frac{\partial y}{\partial x} = \frac{\partial y}{\partial r} \frac{\partial r}{\partial x} + \frac{\partial y}{\partial \phi} \frac{\partial \phi}{\partial x} = \sin\phi \frac{\partial r}{\partial x} + r \cos\phi \frac{\partial \phi}{\partial x}
$$
from which you can solve directly for ##\partial r/\partial x## and ##\partial \phi/\partial x##.
 
  • Like
Likes   Reactions: BvU
  • #15
physconomics said:
Right so I'm a bit stuck, I'm getting ##\frac{\partial^2 f}{\partial x^2} = \frac{\partial}{\partial x}(\frac{df}{dr}\frac{x}{r}) = \frac{\partial}{\partial x}(\frac{df}{dr}) \frac{x}{r} + \frac{df}{dr}\frac{1}{r} ##
where
##\frac{\partial}{\partial x}(\frac{df}{dr}) = \frac{\partial}{\partial x}(\frac{r}{x} \frac{\partial f}{\partial x}) = (\frac{-r}{x^2} \frac{\partial f}{\partial x}) + (\frac{r}{x}\frac{\partial^2 f}{\partial x^2})##
Is this correct? It doesn't lead me to the right answer
Note that
$$
\frac{\partial}{\partial x} \frac{x}{r} \neq \frac 1r
$$
as ##r## depends on ##x##.
 
  • #16
Orodruin said:
Note that
$$
\frac{\partial}{\partial x} \frac{x}{r} \neq \frac 1r
$$
as ##r## depends on ##x##.
Is ##\frac{\partial}{\partial x} \frac{x}{r} = \frac{r - \frac{x^2}{r}}{r^2}##
 
  • #17
physconomics said:
Is ##\frac{\partial}{\partial x} \frac{x}{r} = \frac{r - \frac{x^2}{r}}{r^2}##

Yes, although writing:

##\frac{\partial}{\partial x} \frac{x}{r} = \frac{r^2 - x^2}{r^3}##

seems a bit more logical.
 

Similar threads

Replies
8
Views
2K
  • · Replies 7 ·
Replies
7
Views
1K
Replies
1
Views
3K
  • · Replies 10 ·
Replies
10
Views
2K
  • · Replies 1 ·
Replies
1
Views
4K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
6
Views
3K
Replies
6
Views
2K
  • · Replies 9 ·
Replies
9
Views
5K