How to derive Heisenberg Uncertainty Principle relation?

  • 2
    $\begingroup$ Have you tried googling the second part of the question? $\endgroup$ May 30, 2015 at 8:34
  • 1
    $\begingroup$ Also, this link will help: physics.stackexchange.com/q/24116 $\endgroup$ May 30, 2015 at 8:35
  • $\begingroup$ Hi Pushkar Soni. Welcome to Phys.SE. You should only ask one question per post, so I removed your second question (because there already appears an answer to the first question). $\endgroup$
    – Qmechanic
    May 30, 2015 at 11:04
  • 1
    $\begingroup$ Even Wikipedia gives a proof of the general Schrödinger-Robertson uncertainty principle. Do you want something else? $\endgroup$
    – ACuriousMind
    May 30, 2015 at 13:24

2 Answers 2


Let's see...that's a good revision exercise:

1) To begin with, one should define the quantities involved, namely the variance or its square root the standard deviation

(As I see it, the framework is that the expectation value is the evaluation of a state $|\psi\rangle\in\mathcal{H}$ on an observable $A\in\mathcal{B}(\mathcal{H})$ self-adjoint, i.e. $ \langle A\rangle := \langle \psi|A|\psi\rangle$ (or more generally $\omega(A)$). It can really be interpreted as an expectation value in the sense of probability theory, i.e. as the integral of a random variable with some probability density over the spectrum of A which is interpreted as the possible outcomes.)

The variance of an observable $A\in\mathcal{B}(\mathcal{H})$ is defined by $$ \Delta A^2 := \left\langle (A - \langle A\rangle)^2\right\rangle = \langle A^2\rangle - \langle A\rangle^2$$ where the dependence in the state $|\psi\rangle$ is not explicitly expressed.

Let $A,B\in\mathcal{B}(\mathcal{H})$ be two observables, the Heisenberg uncertainty relation can be stated $$ \Delta A \cdot \Delta B \geq \frac{1}{2} \left|\langle [A,B]\rangle\right| $$

(Copy from some lecture notes... no references sorry) One possibility to proove it is to consider the following quantity $\left|\langle A'B'\rangle\right|^2$ (and find an inequality that holds for any two observables). One has $$A'B'=\frac{1}{2} \{A',B'\} + \frac{1}{2} [A',B']$$

  • The anticommutator is self-adjoint, so by a general property of states the expectation value $\langle \{A',B'\}\rangle \in\mathbb{R}$ (is real) and the commutator is "anti"-self-adjoint, i.e. $[A',B']^{\dagger}= -[A',B']$ so that the expectation value $\langle [A',B']\rangle \in i\mathbb{R}$ (is purely imaginary). Thus the square module of the expectation value reads $$\left|\langle A'B'\rangle\right|^2 =\frac{1}{4} \langle \{A',B'\}\rangle^2 + \frac{1}{4} \left|\langle[A',B']\rangle\right|^2 $$
  • The second essential property that one needs is the Cauchy-Schwarz inequality for states (on a C*-algebra) $$ \left|\langle A'B'\rangle\right|^2 \leq \langle A'^2\rangle \langle B'^2\rangle$$

Combining both (leaving out the anticommutator $\frac{1}{4} \langle \{A',B'\}\rangle^2$) yields $$\frac{1}{4} \left|\langle[A',B']\rangle\right|^2 \leq \langle A'^2\rangle \langle B'^2\rangle$$

  • the last ingredient is to take $A':= A-\langle A\rangle \Bbb{1},\ B':= B-\langle B\rangle 1$. One checks that $$ \Delta A^2 = \langle A'^2\rangle ,\quad \Delta B^2 = \langle B'^2\rangle \quad \text{and}\quad \langle[A,B]\rangle = \langle[A',B']\rangle$$

then take the square root.

2) I leave it to more knowledgeable people...


Robertson Uncertainty Relation: $$(\Delta\hat{A})^2(\Delta\hat{B})^2\geq\left(\frac{1}{2i}\langle[\hat{A},\hat{B}]\rangle\right)^2$$ We have $[\hat{x},\hat{p}]=\hat{x}\hat{p}-\hat{p}\hat{x}$ and $\hat{p}=-i\hbar\frac{\partial}{\partial x}$

so $\hat{x}\hat{p}g(x)=x(-i\hbar)\frac{\partial g}{\partial x}$

and $\hat{p}\hat{x}=-i\hbar\frac{\partial}{\partial x}(\hat{x}g(x))=-i\hbar g(x)-i\hbar x\frac{\partial g}{\partial x}$

This gives us $[\hat{x},\hat{p}]g(x)=\hat{x}\hat{p}g(x)-\hat{p}\hat{x}g(x)=i\hbar g(x)$

so $[\hat{x},\hat{p}]=i\hbar$

$$(\Delta\hat{x})^2(\Delta\hat{p})^2\geq\left(\frac{1}{2i}\langle\Psi|\left[\hat{x},\hat{p}\right]\Psi\rangle\right)^2$$ $$=\left(\frac{1}{2i}ih\langle\Psi|\Psi\rangle\right)^2$$ $$=\frac{\hbar^2}{4}$$ Therefore $$(\Delta\hat{x})(\Delta\hat{p})\geq\frac{\hbar}{2}$$

  • $\begingroup$ I believe that you meant '\hbar' instead of 'h' in the third line up...? $\endgroup$ Jul 2, 2018 at 9:55

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.