# How to prove scaled delta function relation mathematically?

I am working through Shankar's Introduction to Quantum Mechanics. I have come across exercise 1.10.1, which asks the reader to show that:

$$\delta(ax)=\frac{\delta(x)}{|a|}.$$

I can understand it intuitively, in the sense that $$x\to ax$$ will 'compress' the delta function in the x-axis direction, and it has to be $$|a|$$ because the delta function is an even function.

However, how can this relationship be shown mathematically? I have been playing with it for a few days and I seem to be struggling.

Edit:

To expand on the method I have been trying, I am starting with the following completeness relation that Shankar gives:

$$\int_{-\infty}^{\infty}\vert x'\rangle\langle x'\vert dx'=I$$

I then create a new variable, $$x''$$, such that:

$$x''=ax'$$ $$dx''=adx'$$

then I substitute that in:

$$\int_{-\infty}^{\infty}\vert x''\rangle\langle x''\vert \frac{dx''}{a}=I$$

then I perform his operation (that he uses to show $$\int\delta(x)=1$$), where he takes the inner product on one side with $$\vert f \rangle$$ and on the other with $$\langle x \vert$$:

$$\int_{-\infty}^{\infty}\langle x \vert x''\rangle\langle x''\vert f \rangle\frac{dx''}{a}=\langle x \vert I \vert f \rangle = f(x)$$

$$f(x)\int_{-\infty}^{\infty}\delta(x,x'') dx'' = a f(x)$$

However, this seems to lead to the opposite conclusion, that $$\delta(ax)=a\delta(x)$$.

Where have I gone wrong?

## migrated from physics.stackexchange.comSep 20 at 18:24

This question came from our site for active researchers, academics and students of physics.

• Put it in an integral, make a variable substitution, consider cases for when $a>0$ and $a<0$ – Aaron Stevens Sep 19 at 18:46
• Would Mathematics be a better home for this question? – Qmechanic Sep 19 at 18:48
• @AaronStevens I have been trying to do that, but I don't quite see how. – Time4Tea Sep 19 at 19:01
• I thought $$\int_{-\infty}^{\infty}\langle x \vert x''\rangle\langle x''\vert f \rangle\frac{dx''}{a}=f(x)\int_{-\infty}^{\infty}\delta(x,x'') \frac{dx''}{a}$$ – Aaron Stevens Sep 19 at 19:21
• @AaronStevens that seems equivalent to what I got. However, doesn't that mean that $\delta(ax)$ is a factor of $a$ bigger than $\delta(x)$, because you have to divide it by $a$ to get the same result? Perhaps I'm interpreting it wrong ... – Time4Tea Sep 19 at 19:24

You're working much too hard. The delta function is defined by

$$\int_a^b \delta(x) f(x) dx = \begin{cases}f(0) & 0 \in (a,b) \\ 0 & \text{otherwise}\end{cases}$$

Start with $$\int_a^b \delta(ax)f(x) dx$$, perform a variable substitution with $$u= ax$$, and then apply the definition of the delta function. As Aaron said, you should separately consider the cases when $$a<0$$ and $$a>0$$, and you should arrive at the result right away.

• Thanks a lot, this is really helpful. This problem has been bugging me for days ... – Time4Tea Sep 19 at 21:27

J. Murray's answer was very helpful. So, just to make sure I've got this right, this is the definition of the delta function:

$$\int_a^b \delta(x) f(x) dx = \begin{cases}f(0) & 0 \in (a,b) \\ 0 & \text{otherwise}\end{cases}$$

and we want:

$$\int_b^c \delta(ax) f(x) dx$$

So, substitute $$u=ax$$, $$du=adx$$:

$$\int_{ab}^{ac} \delta(u) f(u/a) \frac{du}{a}=?$$

$$\frac{1}{a}\int_{ab}^{ac} \delta(u) f(u/a) du=\begin{cases}\frac{f(0)}{a} & 0 \in (ab,ac) \\ 0 & \text{otherwise}\end{cases}$$

by the definition of the delta function. Then, we have to switch the variables back:

$$\frac{1}{a}\int_{b}^{c} \delta(ax) f(x) a dx=\begin{cases}\frac{f(0)}{a} & 0 \in (b,c) \\ 0 & \text{otherwise}\end{cases}$$

$$\int_{b}^{c} \delta(ax) f(x) dx=\begin{cases}\frac{f(0)}{a} & 0 \in (b,c) \\ 0 & \text{otherwise}\end{cases}$$

which is what we want; however, that assumes that $$a>0$$. If $$a<0$$:

$$\frac{1}{a}\int_{ab}^{ac} \delta(u) f(u/a) du=\begin{cases}-\frac{f(0)}{a} & 0 \in (ac,ab) \\ 0 & \text{otherwise}\end{cases}$$

because $$ab$$ is now larger than $$ac$$, so the integral is 'reversed'. So, that is what gives the general case of:

$$\int_{b}^{c} \delta(ax) f(x) dx=\begin{cases}\frac{f(0)}{|a|} & 0 \in (b,c) \\ 0 & \text{otherwise}\end{cases}$$

Is that correct?

• Looks right to me! – J. Murray Sep 19 at 21:32