Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How can I define the fractional derivative of the Delta function?

I mean $D^{\alpha}= \frac{d^{\alpha}}{dx^{\alpha}} $ where $\alpha$ can be any real number, then if we define $D^{\alpha} \delta (x) $ how can we define it in the sense of distribution?

Applying formal integration by parts $ \alpha $ times I guess that

$$ \int_{-\infty}^{+\infty}D^{\alpha}\delta (x) g(x)dx= (-1)^{[ \alpha]}\int_{-\infty}^{+\infty}D^{\alpha}g(x)\delta(x)dx= (-1)^{[ \alpha]}D^{\alpha}g(0) $$

for any test function $ g(x) $.

share|cite|improve this question
You might want to explain applying formal integration by parts $\alpha$ times when $\alpha$ is not a nonnegative integer. – Did Sep 24 '12 at 13:27
Hint: Look at the Fourier transform. – Nate Eldredge Sep 24 '12 at 14:18
@NateEldredge can you elaborate on that? – Navin Sep 26 '12 at 22:54
A general principle is that taking derivatives in regular space is equivalent to multiplying by powers of $i \xi$ in frequency space. Additionally, the Fourier transform of the delta distribution is constant. So, one would reasonably define $D^\alpha \delta := \frac{1}{\sqrt{2 \pi}}\check{(i \xi)^\alpha}$. This inverse Fourier transform certainly exists as a tempered distribution, but it's not clear (to me) if there's any simpler form for it. – Nick Alger Oct 27 '12 at 12:58
Up to a constant, the inverse Fourier transform mentioned by @Nick Alger is a linear combination of $|x|^{-1-\alpha}$ and $\hbox{sign}(x)\cdot |x|^{-1-\alpha}$, computed by analytic continuation. – paul garrett Nov 26 '12 at 14:08
up vote 2 down vote accepted

I don't know whether the current integral representation of the fractional derivative works on delta function or not. If it works:

$D^\alpha\delta(x)=\begin{cases}\dfrac{1}{\Gamma(\lceil\alpha\rceil-\alpha)}\dfrac{d^{\lceil\alpha\rceil}}{dx^{\lceil\alpha\rceil}}\int_0^x(x-t)^{\lceil\alpha\rceil-\alpha-1}\delta(t)~dt&\text{when}~\alpha>0~\text{and}~\alpha~\text{is not an integer}\\\dfrac{1}{\Gamma(-\alpha)}\int_0^x(x-t)^{-\alpha-1}\delta(t)~dt&\text{when}~\alpha<0\end{cases}=\begin{cases}\dfrac{1}{\Gamma(\lceil\alpha\rceil-\alpha)}\dfrac{d^{\lceil\alpha\rceil}}{dx^{\lceil\alpha\rceil}}(x^{\lceil\alpha\rceil-\alpha-1}H(x))&\text{when}~\alpha>0~\text{and}~\alpha~\text{is not an integer}\\\dfrac{x^{-\alpha-1}H(x)}{\Gamma(-\alpha)}&\text{when}~\alpha<0\end{cases}$

share|cite|improve this answer

$$\delta(x)=\frac{ e^{-(x/\varepsilon)^2}}{\varepsilon\sqrt{\pi }}$$

where $\varepsilon$ is infinitesmal.

Now take fractional derivative of it. By substitution it comes to the fractional derivative of normal distribution.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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