I've been given the lagrangian
$L(\vec{x},\dot{\vec{x}}) = -mc^2 \sqrt{1-\frac{\dot{\vec{x}}}{c^2}} + \frac{e}{c} \vec{A} \cdot \dot{\vec{x}} - e \phi$
and I have to determine the hamiltonian and the equations of motion as well as proving that the time derivative of the relativistic kinetic energy is equal to
$\frac{dE_{\text{kin}}}{dt} = e \vec{E} \cdot \dot{\vec{x}}$.
I've done the first part and got for the Hamiltonian
$H = c \sqrt{m^2 c^2 + (\vec{p} - \frac{e}{c} \vec{A}})^2 + e \phi$
which then gives me the equations of motion
$\dot{\vec{p}} =-\frac{\partial H}{\partial t} = \frac{e}{c} (\vec{\nabla} \cdot \vec{A}) \dot{\vec{x}} - e \vec{\nabla} \phi$
$\dot{\vec{x}} = \frac{\vec{p} - (e/c) \vec{A}}{\sqrt{m^2 + (1/c^2)(\vec{p} - (e/c)\vec{A})}}$.
Then the kinetic energy is given as
$E_{\text{kin}} := \frac{mc^2}{\sqrt{1-(\dot{\vec{x}}^2/c^2)}}$
but I have been stuck trying to find its time derivative since hours. I figured that the total relativistic energy was equal to the hamiltonian and this led me to
$E_{\text{rel}} = H = c\sqrt{m^2c^2 + (\vec{p}-(e/c)\vec{A}}+e\phi = c\sqrt{m^2c^2 + \vec{p}_{\text{mec}}^2} + e\phi = mc^2 \gamma + e\phi = E_{\text{kin}}+e\phi$
$\implies E_{\text{kin}} = H - e\phi$
$\implies \frac{d E_{\text{kin}}}{dt} = \frac{dH}{dt} - e \frac{d \phi}{dt}$
At this point I've used the lagrangian to calculate the time derivative of the hamiltonian since
$\frac{dH}{dt} = -\frac{dL}{dt}$
which is
$\frac{dH}{dt} = -m \gamma \dot{\vec{x}} \cdot \ddot{\vec{x}} - (e/c) \dot{\vec{x}} (\dot{\vec{x}} \cdot \vec{\nabla}) \vec{A} - (e/c) \dot{\vec{A}} + e \frac{d \phi}{dt}$
so
$\dot{E}_{\text{kin}} = \vec{p} \cdot \ddot{\vec{x}} - \frac{e}{c} \big( \dot{\vec{x}} (\dot{\vec{x}} \cdot \vec{\nabla}) \vec{A} + \dot{\vec{A}} \big)$.
I think I've been wandering in circles since I still need $\phi$ for the electric field and the first term is again a problem. Does anybody have a hint before I lose five more hours?
Thanks a lot in advance, your help is appreciated!
Julien.