I am puzzling about an exercise in the book "Electromagnetic Theory" by Ferraro (p.543).

An electron (mass $m$, charge $-e$) in a monopole magnetic field $\vec{B}\left(\vec{r}\right)=g\frac{\vec{r}}{r^3}$ has due to the Lorentz force $\vec{F}_L=-e\vec{v}\times\vec{B}$ the following equation of motion:


Since $\vec{J}=\vec{r}\times\vec{p}+eg\frac{\vec{r}}{r}$ is constant, the electron moves on the surface of a cone, which tip is in the origin at $\vec{r}=0$. Mathematically, $\vec{r}\cdot\vec{J}=egr=rJ\cos\vartheta$, thus the angle $\vartheta$ between $\vec{r}$ and $\vec{J}$ is constant.

Now I use spherical coordinates:

$$\vec{r}=r\mathbf{e}_{r} $$ $$\vec{v}=\dot{\vec{r}}=\dot{r}\mathbf{e}_{r}+r\dot{\vartheta}\mathbf{e}_{\vartheta}+r\dot{\varphi}\sin\vartheta\mathbf{e}_{\varphi}$$ $$\vec{a}=\ddot{\vec{r}}=\left(\ddot{r}-r\dot{\vartheta}^{2}-r\dot{\varphi}^{2}\sin^{2}\vartheta\right)\mathbf{e}_{r}+\left(2\dot{r}\dot{\vartheta}+r\ddot{\vartheta}-r\dot{\varphi}^{2}\sin\vartheta\cos\vartheta\right)\mathbf{e}_{\vartheta}+\left(2\dot{r}\dot{\varphi}\sin\vartheta+2r\dot{\vartheta}\dot{\varphi}\cos\vartheta+r\ddot{\varphi}\sin\vartheta\right)\mathbf{e}_{\varphi} $$

and obtain the following equations of motion (using $\dot{\vartheta}=0$):

$$\ddot{r}=r\dot{\varphi}^2\sin^2\vartheta \qquad(1)$$ $$r\dot{\varphi}^2\cos\vartheta=\frac{eg}{mr}\dot{\varphi} \qquad(2)$$ $$2\dot{r}\dot{\varphi}+r\ddot{\varphi}=0 \qquad(3)$$ with the boundary conditions: $$r\left(0\right)=r_0\qquad\varphi\left(0\right)=0.$$

In order to solve (1) for $r\left(t\right)$, one has to use the fact that the kinetic energy $\frac{1}{2}m\vec{v}^2$ is conserved. This holds due to the vector potential $\vec{A}=\frac{1}{2}\vec{B}\times\vec{r}=0$. Therefore, $$\vec{v}^2=\dot{r}^2+r^2\dot{\varphi}^2\sin^2\vartheta\equiv u^2=\text{const}$$

Using this in (1) one gets the simplified differential equation: $$r\ddot{r}=u^2-\dot{r}^2\qquad(4)$$

Does anyone know how to derive the following solution of (4)? $$r\left(t\right)=r_0\sqrt{1+\left(\frac{ut}{r_0}\right)^2}\qquad(5)$$

Solving (2) for $\dot{\varphi}$: $$\dot{\varphi}=\frac{eg}{mr^2\cos\vartheta}$$

and inserting (5) I obtain: $$\varphi\left(t\right)=\frac{eg}{mr_0u\cos\vartheta}\arctan{\frac{ut}{r_0}}.$$

Now the solutions says: $$r\left(\varphi\right)=\frac{r_0}{\cos\left(\varphi\sin\vartheta\right)}.$$

But I get something similar, but not equal: $$r\left(\varphi\right)=\frac{r_0}{\cos\left(\frac{mr_0u\cos\vartheta}{eg}\varphi\right)}.$$

Has anyone an idea where I could have done something wrong?


Actually, I don't have to use the conservation of the kinetic energy. With a slightly different procedure, one gets another simplified differential equation.

Inserting (2) in (1):

$$\ddot{r}=\frac{e^2 g^2 \tan^2\vartheta}{m^2} \frac{1}{r^3}$$

Now look at:

$$\frac{\text{d}}{\text{dt}}\dot{r}^2=2\dot{r}\ddot{r}=\frac{e^2 g^2 \tan^2\vartheta}{m^2} \frac{2\dot{r}}{r^3}=\frac{\text{d}}{\text{dt}}\left(\frac{e^2 g^2 \tan^2\vartheta}{m^2}\frac{-1}{r^2}\right)$$

Thus, a simpler differential equation is obtained ($C=\text{const}$):

$$\dot{r}^2=\frac{e^2 g^2 \tan^2\vartheta}{m^2}\frac{-1}{r^2}+C$$

and can be solved by separation of variables:

$$\text{dt}=\frac{r\text{ dr}}{\sqrt{Cr^2-\frac{e^2 g^2 \tan^2\vartheta}{m^2}}}$$

$$t=\frac{1}{C}\sqrt{Cr^2-\frac{e^2 g^2 \tan^2\vartheta}{m^2}}+t_0$$

With $t_0=0$ and $r\left(0\right)=r_0$:

$$r\left(t\right)=\sqrt{\frac{e^2 g^2 \tan^2\vartheta}{m^2r_0^2}t^2+r_0^2}$$

Now solving for $\varphi\left(t\right)$ from (2):


$$\text{d}\varphi=\frac{m\cos\vartheta r_0^2}{eg\sin^2\vartheta}\frac{\text{dt}}{t^2+\left(\frac{m r_0^2}{e g \tan\vartheta}\right)^2}$$

Using the arctangent integral:


and $\varphi\left(0\right)=0$, one obtains:

$$\varphi\left(t\right)=\frac{1}{\sin\vartheta}\arctan\frac{e g \tan\vartheta\ t}{m r_0^2} $$


$$r\left(\varphi\right)=r_0 \sqrt{\tan^2\left(\sin\vartheta \varphi\right)+1}=\frac{r_0}{\cos\left(\sin\vartheta\varphi\right)}$$

  • $\begingroup$ When you do your integration of dr dt, I believe that you are missing a constant of integration. $\endgroup$
    – Loonuh
    Oct 27 '15 at 5:41
  • $\begingroup$ This solution is only valid if you assume that at t=0, the particle is at the distance of closest approach, at which point $\dot{r}=0$; otherwise, you will have a constant of integration which depends on the angular momentum, the energy, $r_0$, and $eg$. $\endgroup$
    – Loonuh
    Oct 27 '15 at 5:59
  • $\begingroup$ Do you mean the integration of the 4th equation? Of course, on each side I have a constant after integration, but these two constants I simply summarize in $t_0$. $\endgroup$
    – Andy
    Nov 1 '15 at 14:08
  • $\begingroup$ You can only shift time to $t_0 = 0$ with $r(0) = r_0$ and remain consistent with your answer if $r_0$ is the distance of closest of approach. Otherwise a term with order $t^1$ will appear under your square root in your r(t). $\endgroup$
    – Loonuh
    Nov 5 '15 at 7:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.