0
$\begingroup$

I'm trying to compute to quantities with Hamiltonian and Dirac delta function but I don't how to do it properly. I'm stuck calculating the following quantity

$$ \frac{d}{dE} \left[ \theta(E-H(x,p;V)) \right]. \tag{1}$$ I know that the derivative of the Heaviside step function $\theta(x)$ is the dirac delta function $\delta(x)$ $$ \frac{d\theta(x)}{dx}= \delta(x)$$

So my guess would be the following:

$$ \frac{d\theta(E-H)}{dE}= \delta(E-H).$$

However if I use the chain rule I get: $$ \frac{d\theta(E-H)}{d(E-H)}\frac{d(E-H)}{dE}= \delta(E-H) \left(1-\frac{dH}{dE}\right).$$

The text I'm studying about areas in phase planes gives the following but leaves out a lot of intermediate steps:

$$ \frac{d}{dE} \left[ \theta(E-H(x,p;V)) \right]=\delta (p^2/2m+\phi-E).$$

Where $H=p^2/2m+\phi(x;V).$

The next step makes me equally confused, it gives the following equation:

$$ \int dp \delta (p^2/2m+\phi-E)=2m/p(x) \ \ \ \ \ \ \ (2)$$

Where the text tells it uses the formula $$\delta(f(p)) = \sum_i \delta(p-p_i)/|f'(p_i)|.$$

Questions How to understand the following equations:

$$ \frac{d}{dE} \left[ \theta(E-H(x,p;V)) \right] =\delta (p^2/2m+\phi-E)) \ \ \ \ \ \ \ \ \ \ (1)$$

$$ \int dp \delta (p^2/2m+\phi-E)=2m/p(x) \ \ \ \ \ \ \ (2)$$

$\endgroup$

1 Answer 1

1
$\begingroup$

The first is simple I think: $$\frac{d}{dE}\theta[E-H(x,p;V)]=\delta[E-H(x,p;V)]=\delta[H(x,p;V)-E]= \delta\Big[\frac{p^2}{2m}+\phi(x;V)-E\Big]$$ where in the third step I used the symmetric property of the $\delta$ function, i.e. $\delta(-x)=\delta(x)$.

As far as the second question is concerned, one can proceed in the following way $$\int dp\delta(p^2/2m+\phi-E)=\int dp \Big[\frac{1}{|p_+(x)/m|}\delta(p_+^2/2m+\phi-E) +\frac{1}{|p_-(x)/m|}\delta(p_-^2/2m+\phi-E)\Big]$$ where $p_{\pm}=\pm\sqrt{2m(E-\phi)}$ are the two solutions to the equation $f(p)=p^2/2m+\phi-E=0$. The $x$ dependence of $p$ indicates that $p$ is a function of $x$ through $\phi$. Performing the integrals and identifying that the magnitude of $p_+(x)$ equals the magnitude of $p_-(x)$, yields $$\int dp\delta(p^2/2m+\phi-E)=\Big[\frac{1}{|p_+(x)/m|} +\frac{1}{|p_-(x)/m|}\Big]=\frac{2m}{p(x)}$$ where $p(x)=|p_+(x)|=|p_-(x)|=\sqrt{2m(E-\phi)}$.

I hope this helps! If anything is unclear, please comment!

$\endgroup$
2
  • $\begingroup$ Thank you very much for your quick and lucid answer. I have one question, what is gained by swapping $\delta(E-H)=\delta(H-E)$? Or is this a common thing to do? $\endgroup$ Commented Jul 31, 2022 at 9:12
  • 1
    $\begingroup$ Hi! I spotted a mistake in my answer and I have edited the answer! I actually substituted $p_+(x)$ with $|p_+(x)|$ and $p_-(x)$ with $|p_-(x)|$! As far as your question is concerned, nothing is gained by writting $\delta(E-H)$ instead of $\delta(H-E)$, as they are the same! I do not know why one would prefer one instead of the other, but the reasons must be aesthetic or random probably, because it does not matter which one is preferred... I hope this helps $\endgroup$
    – schris38
    Commented Jul 31, 2022 at 9:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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