Uncertainty principle

Part of a series of articles about
Quantum mechanics
${\displaystyle i\hbar {\frac {d}{dt}}|\Psi \rangle ={\hat {H}}|\Psi \rangle }$ Schrödinger equation
Introduction Glossary History
Background Classical mechanics Old quantum theory Bra–ket notation Hamiltonian Interference
Fundamentals Complementarity Decoherence Entanglement Energy level Measurement Nonlocality Quantum number State Superposition Symmetry Tunnelling Uncertainty Wave function Collapse
Experiments Bell's inequality CHSH inequality Davisson–Germer Double-slit Elitzur–Vaidman Franck–Hertz Leggett inequality Leggett–Garg inequality Mach–Zehnder Popper Quantum eraser Delayed-choice Schrödinger's cat Stern–Gerlach Wheeler's delayed-choice
Formulations Overview Heisenberg Interaction Matrix Phase-space Schrödinger Sum-over-histories (path integral)
Equations Dirac Klein–Gordon Pauli Rydberg Schrödinger
Interpretations Bayesian Consistent histories Copenhagen de Broglie–Bohm Ensemble Hidden-variable Local Superdeterminism Many-worlds Objective-collapse Quantum logic Relational Transactional Von Neumann–Wigner
Advanced topics Relativistic quantum mechanics Quantum field theory Quantum information science Quantum computing Quantum chaos EPR paradox Density matrix Scattering theory Quantum statistical mechanics Quantum machine learning
Scientists Aharonov Bell Bethe Blackett Bloch Bohm Bohr Born Bose de Broglie Compton Dirac Davisson Debye Ehrenfest Einstein Everett Fock Fermi Feynman Glauber Gutzwiller Heisenberg Hilbert Jordan Kramers Lamb Landau Laue Moseley Millikan Onnes Pauli Planck Rabi Raman Rydberg Schrödinger Simmons Sommerfeld von Neumann Weyl Wien Wigner Zeeman Zeilinger
v t e

The uncertainty principle, also known as Heisenberg's indeterminacy principle, is a fundamental concept in quantum mechanics. It states that there is a limit to the precision with which certain pairs of physical properties, such as position and momentum, can be simultaneously known. In other words, the more accurately one property is measured, the less accurately the other property can be known.

More formally, the uncertainty principle is any of a variety of mathematical inequalities asserting a fundamental limit to the product of the accuracy of certain related pairs of measurements on a quantum system, such as position, x, and momentum, p.^[1] Such paired-variables are known as complementary variables or canonically conjugate variables.

First introduced in 1927 by German physicist Werner Heisenberg,^[2][3][4][5] the formal inequality relating the standard deviation of position σ_x and the standard deviation of momentum σ_p was derived by Earle Hesse Kennard^[6] later that year and by Hermann Weyl^[7] in 1928:

where ${\displaystyle \hbar ={\frac {h}{2\pi }}}$ is the reduced Planck constant.

The quintessentially quantum mechanical uncertainty principle comes in many forms other than position–momentum. The energy–time relationship is widely used to relate quantum state lifetime to measured energy widths but its formal derivation is fraught with confusing issues about the nature of time. The basic principle has been extended in numerous directions; it must be considered in many kinds of fundamental physical measurements.

It is vital to illustrate how the principle applies to relatively intelligible physical situations since it is indiscernible on the macroscopic^[8] scales that humans experience. Two alternative frameworks for quantum physics offer different explanations for the uncertainty principle. The wave mechanics picture of the uncertainty principle is more visually intuitive, but the more abstract matrix mechanics picture formulates it in a way that generalizes more easily.

Mathematically, in wave mechanics, the uncertainty relation between position and momentum arises because the expressions of the wavefunction in the two corresponding orthonormal bases in Hilbert space are Fourier transforms of one another (i.e., position and momentum are conjugate variables). A nonzero function and its Fourier transform cannot both be sharply localized at the same time.^[9] A similar tradeoff between the variances of Fourier conjugates arises in all systems underlain by Fourier analysis, for example in sound waves: A pure tone is a sharp spike at a single frequency, while its Fourier transform gives the shape of the sound wave in the time domain, which is a completely delocalized sine wave. In quantum mechanics, the two key points are that the position of the particle takes the form of a matter wave, and momentum is its Fourier conjugate, assured by the de Broglie relation p = ħk, where k is the wavenumber.

In matrix mechanics, the mathematical formulation of quantum mechanics, any pair of non-commuting self-adjoint operators representing observables are subject to similar uncertainty limits. An eigenstate of an observable represents the state of the wavefunction for a certain measurement value (the eigenvalue). For example, if a measurement of an observable A is performed, then the system is in a particular eigenstate Ψ of that observable. However, the particular eigenstate of the observable A need not be an eigenstate of another observable B: If so, then it does not have a unique associated measurement for it, as the system is not in an eigenstate of that observable.^[10]

The uncertainty principle can be visualized using the position- and momentum-space wavefunctions for one spinless particle with mass in one dimension.

The more localized the position-space wavefunction, the more likely the particle is to be found with the position coordinates in that region, and correspondingly the momentum-space wavefunction is less localized so the possible momentum components the particle could have are more widespread. Conversely, the more localized the momentum-space wavefunction, the more likely the particle is to be found with those values of momentum components in that region, and correspondingly the less localized the position-space wavefunction, so the position coordinates the particle could occupy are more widespread. These wavefunctions are Fourier transforms of each other: mathematically, the uncertainty principle expresses the relationship between conjugate variables in the transform.

Plane wave

Wave packet

Propagation of de Broglie waves in 1d—real part of the complex amplitude is blue, imaginary part is green. The probability (shown as the colour opacity) of finding the particle at a given point x is spread out like a waveform, there is no definite position of the particle. As the amplitude increases above zero the curvature reverses sign, so the amplitude begins to decrease again, and vice versa—the result is an alternating amplitude: a wave.

According to the de Broglie hypothesis, every object in the universe is associated with a wave. Thus every object, from an elementary particle to atoms, molecules and on up to planets and beyond are subject to the uncertainty principle.

The time-independent wave function of a single-moded plane wave of wavenumber k₀ or momentum p₀ is^[11] $${\displaystyle \psi (x)\propto e^{ik_{0}x}=e^{ip_{0}x/\hbar }~.}$$

The Born rule states that this should be interpreted as a probability density amplitude function in the sense that the probability of finding the particle between a and b is $${\displaystyle \operatorname {P} [a\leq X\leq b]=\int _{a}^{b}|\psi (x)|^{2}\,\mathrm {d} x~.}$$

In the case of the single-mode plane wave, ${\displaystyle |\psi (x)|^{2}}$ is 1 if ${\displaystyle X=x}$ and 0 otherwise. In other words, the particle position is extremely uncertain in the sense that it could be essentially anywhere along the wave packet.

On the other hand, consider a wave function that is a sum of many waves, which we may write as $${\displaystyle \psi (x)\propto \sum _{n}A_{n}e^{ip_{n}x/\hbar }~,}$$ where A_n represents the relative contribution of the mode p_n to the overall total. The figures to the right show how with the addition of many plane waves, the wave packet can become more localized. We may take this a step further to the continuum limit, where the wave function is an integral over all possible modes $${\displaystyle \psi (x)={\frac {1}{\sqrt {2\pi \hbar }}}\int _{-\infty }^{\infty }\varphi (p)\cdot e^{ipx/\hbar }\,dp~,}$$ with ${\displaystyle \varphi (p)}$ representing the amplitude of these modes and is called the wave function in momentum space. In mathematical terms, we say that ${\displaystyle \varphi (p)}$ is the Fourier transform of ${\displaystyle \psi (x)}$ and that x and p are conjugate variables. Adding together all of these plane waves comes at a cost, namely the momentum has become less precise, having become a mixture of waves of many different momenta.^[12]

One way to quantify the precision of the position and momentum is the standard deviation σ. Since ${\displaystyle |\psi (x)|^{2}}$ is a probability density function for position, we calculate its standard deviation.

The precision of the position is improved, i.e. reduced σ_x, by using many plane waves, thereby weakening the precision of the momentum, i.e. increased σ_p. Another way of stating this is that σ_x and σ_p have an inverse relationship or are at least bounded from below. This is the uncertainty principle, the exact limit of which is the Kennard bound.

We are interested in the variances of position and momentum, defined as $${\displaystyle \sigma _{x}^{2}=\int _{-\infty }^{\infty }x^{2}\cdot |\psi (x)|^{2}\,dx-\left(\int _{-\infty }^{\infty }x\cdot |\psi (x)|^{2}\,dx\right)^{2}}$$ $${\displaystyle \sigma _{p}^{2}=\int _{-\infty }^{\infty }p^{2}\cdot |\varphi (p)|^{2}\,dp-\left(\int _{-\infty }^{\infty }p\cdot |\varphi (p)|^{2}\,dp\right)^{2}~.}$$

Without loss of generality, we will assume that the means vanish, which just amounts to a shift of the origin of our coordinates. (A more general proof that does not make this assumption is given below.) This gives us the simpler form $${\displaystyle \sigma _{x}^{2}=\int _{-\infty }^{\infty }x^{2}\cdot |\psi (x)|^{2}\,dx}$$ $${\displaystyle \sigma _{p}^{2}=\int _{-\infty }^{\infty }p^{2}\cdot |\varphi (p)|^{2}\,dp~.}$$

The function ${\displaystyle f(x)=x\cdot \psi (x)}$ can be interpreted as a vector in a function space. We can define an inner product for a pair of functions u(x) and v(x) in this vector space: $${\displaystyle \langle u\mid v\rangle =\int _{-\infty }^{\infty }u^{*}(x)\cdot v(x)\,dx,}$$ where the asterisk denotes the complex conjugate.

With this inner product defined, we note that the variance for position can be written as $${\displaystyle \sigma _{x}^{2}=\int _{-\infty }^{\infty }|f(x)|^{2}\,dx=\langle f\mid f\rangle ~.}$$

We can repeat this for momentum by interpreting the function ${\displaystyle {\tilde {g}}(p)=p\cdot \varphi (p)}$ as a vector, but we can also take advantage of the fact that ${\displaystyle \psi (x)}$ and ${\displaystyle \varphi (p)}$ are Fourier transforms of each other. We evaluate the inverse Fourier transform through integration by parts: $${\displaystyle {\begin{aligned}g(x)&={\frac {1}{\sqrt {2\pi \hbar }}}\cdot \int _{-\infty }^{\infty }{\tilde {g}}(p)\cdot e^{ipx/\hbar }\,dp\\&={\frac {1}{\sqrt {2\pi \hbar }}}\int _{-\infty }^{\infty }p\cdot \varphi (p)\cdot e^{ipx/\hbar }\,dp\\&={\frac {1}{2\pi \hbar }}\int _{-\infty }^{\infty }\left[p\cdot \int _{-\infty }^{\infty }\psi (\chi )e^{-ip\chi /\hbar }\,d\chi \right]\cdot e^{ipx/\hbar }\,dp\\&={\frac {i}{2\pi }}\int _{-\infty }^{\infty }\left[{\cancel {\left.\psi (\chi )e^{-ip\chi /\hbar }\right|_{-\infty }^{\infty }}}-\int _{-\infty }^{\infty }{\frac {d\psi (\chi )}{d\chi }}e^{-ip\chi /\hbar }\,d\chi \right]\cdot e^{ipx/\hbar }\,dp\\&=-i\int _{-\infty }^{\infty }{\frac {d\psi (\chi )}{d\chi }}\left[{\frac {1}{2\pi }}\int _{-\infty }^{\infty }\,e^{ip(x-\chi )/\hbar }\,dp\right]\,d\chi \\&=-i\int _{-\infty }^{\infty }{\frac {d\psi (\chi )}{d\chi }}\left[\delta \left({\frac {x-\chi }{\hbar }}\right)\right]\,d\chi \\&=-i\hbar \int _{-\infty }^{\infty }{\frac {d\psi (\chi )}{d\chi }}\left[\delta \left(x-\chi \right)\right]\,d\chi \\&=-i\hbar {\frac {d\psi (x)}{dx}}\\&=\left(-i\hbar {\frac {d}{dx}}\right)\cdot \psi (x),\end{aligned}}}$$ where ${\displaystyle v={\frac {\hbar }{-ip}}e^{-ip\chi /\hbar }}$ in the integration by parts, the cancelled term vanishes because the wave function vanishes at infinity, and the final two integrations re-assert the Fourier transforms. Often the term ${\textstyle -i\hbar {\frac {d}{dx}}}$ is called the momentum operator in position space. Applying Plancherel's theorem and then Parseval's theorem, we see that the variance for momentum can be written as $${\displaystyle \sigma _{p}^{2}=\int _{-\infty }^{\infty }|{\tilde {g}}(p)|^{2}\,dp=\int _{-\infty }^{\infty }|g(x)|^{2}\,dx=\langle g\mid g\rangle .}$$

The Cauchy–Schwarz inequality asserts that $${\displaystyle \sigma _{x}^{2}\sigma _{p}^{2}=\langle f\mid f\rangle \cdot \langle g\mid g\rangle \geq |\langle f\mid g\rangle |^{2}~.}$$

The modulus squared of any complex number z can be expressed as $${\displaystyle |z|^{2}={\Big (}{\text{Re}}(z){\Big )}^{2}+{\Big (}{\text{Im}}(z){\Big )}^{2}\geq {\Big (}{\text{Im}}(z){\Big )}^{2}=\left({\frac {z-z^{\ast }}{2i}}\right)^{2}.}$$ we let ${\displaystyle z=\langle f|g\rangle }$ and ${\displaystyle z^{*}=\langle g\mid f\rangle }$ and substitute these into the equation above to get $${\displaystyle |\langle f\mid g\rangle |^{2}\geq \left({\frac {\langle f\mid g\rangle -\langle g\mid f\rangle }{2i}}\right)^{2}~.}$$

All that remains is to evaluate these inner products.

$${\displaystyle {\begin{aligned}\langle f\mid g\rangle -\langle g\mid f\rangle &=\int _{-\infty }^{\infty }\psi ^{*}(x)\,x\cdot \left(-i\hbar {\frac {d}{dx}}\right)\,\psi (x)\,dx-\int _{-\infty }^{\infty }\psi ^{*}(x)\,\left(-i\hbar {\frac {d}{dx}}\right)\cdot x\,\psi (x)\,dx\\&=i\hbar \cdot \int _{-\infty }^{\infty }\psi ^{*}(x)\left[\left(-x\cdot {\frac {d\psi (x)}{dx}}\right)+{\frac {d(x\psi (x))}{dx}}\right]\,dx\\&=i\hbar \cdot \int _{-\infty }^{\infty }\psi ^{*}(x)\left[\left(-x\cdot {\frac {d\psi (x)}{dx}}\right)+\psi (x)+\left(x\cdot {\frac {d\psi (x)}{dx}}\right)\right]\,dx\\&=i\hbar \cdot \int _{-\infty }^{\infty }\psi ^{*}(x)\psi (x)\,dx\\&=i\hbar \cdot \int _{-\infty }^{\infty }|\psi (x)|^{2}\,dx\\&=i\hbar \end{aligned}}}$$

Plugging this into the above inequalities, we get $${\displaystyle \sigma _{x}^{2}\sigma _{p}^{2}\geq |\langle f\mid g\rangle |^{2}\geq \left({\frac {\langle f\mid g\rangle -\langle g\mid f\rangle }{2i}}\right)^{2}=\left({\frac {i\hbar }{2i}}\right)^{2}={\frac {\hbar ^{2}}{4}}}$$ or taking the square root $${\displaystyle \sigma _{x}\sigma _{p}\geq {\frac {\hbar }{2}}~.}$$

with equality if and only if p and x are linearly dependent. Note that the only physics involved in this proof was that ${\displaystyle \psi (x)}$ and ${\displaystyle \varphi (p)}$ are wave functions for position and momentum, which are Fourier transforms of each other. A similar result would hold for any pair of conjugate variables.

In matrix mechanics, observables such as position and momentum are represented by self-adjoint operators.^[12] When considering pairs of observables, an important quantity is the commutator. For a pair of operators  and ${\displaystyle {\hat {B}}}$, one defines their commutator as $${\displaystyle [{\hat {A}},{\hat {B}}]={\hat {A}}{\hat {B}}-{\hat {B}}{\hat {A}}.}$$ In the case of position and momentum, the commutator is the canonical commutation relation $${\displaystyle [{\hat {x}},{\hat {p}}]=i\hbar .}$$

The physical meaning of the non-commutativity can be understood by considering the effect of the commutator on position and momentum eigenstates. Let ${\displaystyle |\psi \rangle }$ be a right eigenstate of position with a constant eigenvalue x₀. By definition, this means that ${\displaystyle {\hat {x}}|\psi \rangle =x_{0}|\psi \rangle .}$ Applying the commutator to ${\displaystyle |\psi \rangle }$ yields $${\displaystyle [{\hat {x}},{\hat {p}}]|\psi \rangle =({\hat {x}}{\hat {p}}-{\hat {p}}{\hat {x}})|\psi \rangle =({\hat {x}}-x_{0}{\hat {I}}){\hat {p}}\,|\psi \rangle =i\hbar |\psi \rangle ,}$$ where Î is the identity operator.

Suppose, for the sake of proof by contradiction, that ${\displaystyle |\psi \rangle }$ is also a right eigenstate of momentum, with constant eigenvalue p₀. If this were true, then one could write $${\displaystyle ({\hat {x}}-x_{0}{\hat {I}}){\hat {p}}\,|\psi \rangle =({\hat {x}}-x_{0}{\hat {I}})p_{0}\,|\psi \rangle =(x_{0}{\hat {I}}-x_{0}{\hat {I}})p_{0}\,|\psi \rangle =0.}$$ On the other hand, the above canonical commutation relation requires that $${\displaystyle [{\hat {x}},{\hat {p}}]|\psi \rangle =i\hbar |\psi \rangle \neq 0.}$$ This implies that no quantum state can simultaneously be both a position and a momentum eigenstate.

When a state is measured, it is projected onto an eigenstate in the basis of the relevant observable. For example, if a particle's position is measured, then the state amounts to a position eigenstate. This means that the state is not a momentum eigenstate, however, but rather it can be represented as a sum of multiple momentum basis eigenstates. In other words, the momentum must be less precise. This precision may be quantified by the standard deviations, $${\displaystyle \sigma _{x}={\sqrt {\langle {\hat {x}}^{2}\rangle -\langle {\hat {x}}\rangle ^{2}}}}$$ $${\displaystyle \sigma _{p}={\sqrt {\langle {\hat {p}}^{2}\rangle -\langle {\hat {p}}\rangle ^{2}}}.}$$

As in the wave mechanics interpretation above, one sees a tradeoff between the respective precisions of the two, quantified by the uncertainty principle.

Consider a one-dimensional quantum harmonic oscillator. It is possible to express the position and momentum operators in terms of the creation and annihilation operators: $${\displaystyle {\hat {x}}={\sqrt {\frac {\hbar }{2m\omega }}}(a+a^{\dagger })}$$ $${\displaystyle {\hat {p}}=i{\sqrt {\frac {m\omega \hbar }{2}}}(a^{\dagger }-a).}$$

Using the standard rules for creation and annihilation operators on the energy eigenstates, $${\displaystyle a^{\dagger }|n\rangle ={\sqrt {n+1}}|n+1\rangle }$$ $${\displaystyle a|n\rangle ={\sqrt {n}}|n-1\rangle ,}$$ the variances may be computed directly, $${\displaystyle \sigma _{x}^{2}={\frac {\hbar }{m\omega }}\left(n+{\frac {1}{2}}\right)}$$ $${\displaystyle \sigma _{p}^{2}=\hbar m\omega \left(n+{\frac {1}{2}}\right)\,.}$$ The product of these standard deviations is then $${\displaystyle \sigma _{x}\sigma _{p}=\hbar \left(n+{\frac {1}{2}}\right)\geq {\frac {\hbar }{2}}.~}$$

In particular, the above Kennard bound^[6] is saturated for the ground state n=0, for which the probability density is just the normal distribution.

Position (blue) and momentum (red) probability densities for an initial Gaussian distribution. From top to bottom, the animations show the cases Ω = ω, Ω = 2ω, and Ω = ω/2. Note the tradeoff between the widths of the distributions.

In a quantum harmonic oscillator of characteristic angular frequency ω, place a state that is offset from the bottom of the potential by some displacement x₀ as $${\displaystyle \psi (x)=\left({\frac {m\Omega }{\pi \hbar }}\right)^{1/4}\exp {\left(-{\frac {m\Omega (x-x_{0})^{2}}{2\hbar }}\right)},}$$ where Ω describes the width of the initial state but need not be the same as ω. Through integration over the propagator, we can solve for the full time-dependent solution. After many cancelations, the probability densities reduce to $${\displaystyle |\Psi (x,t)|^{2}\sim {\mathcal {N}}\left(x_{0}\cos {(\omega t)},{\frac {\hbar }{2m\Omega }}\left(\cos ^{2}(\omega t)+{\frac {\Omega ^{2}}{\omega ^{2}}}\sin ^{2}{(\omega t)}\right)\right)}$$ $${\displaystyle |\Phi (p,t)|^{2}\sim {\mathcal {N}}\left(-mx_{0}\omega \sin(\omega t),{\frac {\hbar m\Omega }{2}}\left(\cos ^{2}{(\omega t)}+{\frac {\omega ^{2}}{\Omega ^{2}}}\sin ^{2}{(\omega t)}\right)\right),}$$ where we have used the notation ${\displaystyle {\mathcal {N}}(\mu ,\sigma ^{2})}$ to denote a normal distribution of mean μ and variance σ². Copying the variances above and applying trigonometric identities, we can write the product of the standard deviations as $${\displaystyle {\begin{aligned}\sigma _{x}\sigma _{p}&={\frac {\hbar }{2}}{\sqrt {\left(\cos ^{2}{(\omega t)}+{\frac {\Omega ^{2}}{\omega ^{2}}}\sin ^{2}{(\omega t)}\right)\left(\cos ^{2}{(\omega t)}+{\frac {\omega ^{2}}{\Omega ^{2}}}\sin ^{2}{(\omega t)}\right)}}\\&={\frac {\hbar }{4}}{\sqrt {3+{\frac {1}{2}}\left({\frac {\Omega ^{2}}{\omega ^{2}}}+{\frac {\omega ^{2}}{\Omega ^{2}}}\right)-\left({\frac {1}{2}}\left({\frac {\Omega ^{2}}{\omega ^{2}}}+{\frac {\omega ^{2}}{\Omega ^{2}}}\right)-1\right)\cos {(4\omega t)}}}\end{aligned}}}$$

From the relations $${\displaystyle {\frac {\Omega ^{2}}{\omega ^{2}}}+{\frac {\omega ^{2}}{\Omega ^{2}}}\geq 2,\quad |\cos(4\omega t)|\leq 1,}$$ we can conclude the following (the right most equality holds only when Ω = ω): $${\displaystyle \sigma _{x}\sigma _{p}\geq {\frac {\hbar }{4}}{\sqrt {3+{\frac {1}{2}}\left({\frac {\Omega ^{2}}{\omega ^{2}}}+{\frac {\omega ^{2}}{\Omega ^{2}}}\right)-\left({\frac {1}{2}}\left({\frac {\Omega ^{2}}{\omega ^{2}}}+{\frac {\omega ^{2}}{\Omega ^{2}}}\right)-1\right)}}={\frac {\hbar }{2}}.}$$

A coherent state is a right eigenstate of the annihilation operator, $${\displaystyle {\hat {a}}|\alpha \rangle =\alpha |\alpha \rangle ,}$$ which may be represented in terms of Fock states as $${\displaystyle |\alpha \rangle =e^{-{|\alpha |^{2} \over 2}}\sum _{n=0}^{\infty }{\alpha ^{n} \over {\sqrt {n!}}}|n\rangle }$$

In the picture where the coherent state is a massive particle in a quantum harmonic oscillator, the position and momentum operators may be expressed in terms of the annihilation operators in the same formulas above and used to calculate the variances, $${\displaystyle \sigma _{x}^{2}={\frac {\hbar }{2m\omega }},}$$ $${\displaystyle \sigma _{p}^{2}={\frac {\hbar m\omega }{2}}.}$$ Therefore, every coherent state saturates the Kennard bound $${\displaystyle \sigma _{x}\sigma _{p}={\sqrt {\frac {\hbar }{2m\omega }}}\,{\sqrt {\frac {\hbar m\omega }{2}}}={\frac {\hbar }{2}}.}$$ with position and momentum each contributing an amount ${\textstyle {\sqrt {\hbar /2}}}$ in a "balanced" way. Moreover, every squeezed coherent state also saturates the Kennard bound although the individual contributions of position and momentum need not be balanced in general.

Consider a particle in a one-dimensional box of length ${\displaystyle L}$. The eigenfunctions in position and momentum space are $${\displaystyle \psi _{n}(x,t)={\begin{cases}A\sin(k_{n}x)\mathrm {e} ^{-\mathrm {i} \omega _{n}t},&0<x<L,\\0,&{\text{otherwise,}}\end{cases}}}$$ and $${\displaystyle \varphi _{n}(p,t)={\sqrt {\frac {\pi L}{\hbar }}}\,\,{\frac {n\left(1-(-1)^{n}e^{-ikL}\right)e^{-i\omega _{n}t}}{\pi ^{2}n^{2}-k^{2}L^{2}}},}$$ where ${\textstyle \omega _{n}={\frac {\pi ^{2}\hbar n^{2}}{8L^{2}m}}}$ and we have used the de Broglie relation ${\displaystyle p=\hbar k}$. The variances of ${\displaystyle x}$ and ${\displaystyle p}$ can be calculated explicitly: $${\displaystyle \sigma _{x}^{2}={\frac {L^{2}}{12}}\left(1-{\frac {6}{n^{2}\pi ^{2}}}\right)}$$ $${\displaystyle \sigma _{p}^{2}=\left({\frac {\hbar n\pi }{L}}\right)^{2}.}$$

The product of the standard deviations is therefore $${\displaystyle \sigma _{x}\sigma _{p}={\frac {\hbar }{2}}{\sqrt {{\frac {n^{2}\pi ^{2}}{3}}-2}}.}$$ For all ${\displaystyle n=1,\,2,\,3,\,\ldots }$, the quantity ${\textstyle {\sqrt {{\frac {n^{2}\pi ^{2}}{3}}-2}}}$ is greater than 1, so the uncertainty principle is never violated. For numerical concreteness, the smallest value occurs when ${\displaystyle n=1}$, in which case $${\displaystyle \sigma _{x}\sigma _{p}={\frac {\hbar }{2}}{\sqrt {{\frac {\pi ^{2}}{3}}-2}}\approx 0.568\hbar >{\frac {\hbar }{2}}.}$$

Assume a particle initially has a momentum space wave function described by a normal distribution around some constant momentum p₀ according to $${\displaystyle \varphi (p)=\left({\frac {x_{0}}{\hbar {\sqrt {\pi }}}}\right)^{1/2}\exp \left({\frac {-x_{0}^{2}(p-p_{0})^{2}}{2\hbar ^{2}}}\right),}$$ where we have introduced a reference scale ${\textstyle x_{0}={\sqrt {\hbar /m\omega _{0}}}}$, with ${\displaystyle \omega _{0}>0}$ describing the width of the distribution—cf. nondimensionalization. If the state is allowed to evolve in free space, then the time-dependent momentum and position space wave functions are $${\displaystyle \Phi (p,t)=\left({\frac {x_{0}}{\hbar {\sqrt {\pi }}}}\right)^{1/2}\exp \left({\frac {-x_{0}^{2}(p-p_{0})^{2}}{2\hbar ^{2}}}-{\frac {ip^{2}t}{2m\hbar }}\right),}$$ $${\displaystyle \Psi (x,t)=\left({\frac {1}{x_{0}{\sqrt {\pi }}}}\right)^{1/2}{\frac {e^{-x_{0}^{2}p_{0}^{2}/2\hbar ^{2}}}{\sqrt {1+i\omega _{0}t}}}\,\exp \left(-{\frac {(x-ix_{0}^{2}p_{0}/\hbar )^{2}}{2x_{0}^{2}(1+i\omega _{0}t)}}\right).}$$

Since ${\displaystyle \langle p(t)\rangle =p_{0}}$ and ${\displaystyle \sigma _{p}(t)=\hbar /({\sqrt {2}}x_{0})}$, this can be interpreted as a particle moving along with constant momentum at arbitrarily high precision. On the other hand, the standard deviation of the position is $${\displaystyle \sigma _{x}={\frac {x_{0}}{\sqrt {2}}}{\sqrt {1+\omega _{0}^{2}t^{2}}}}$$ such that the uncertainty product can only increase with time as $${\displaystyle \sigma _{x}(t)\sigma _{p}(t)={\frac {\hbar }{2}}{\sqrt {1+\omega _{0}^{2}t^{2}}}}$$

Starting with Kennard's derivation of position-momentum uncertainty, Howard Percy Robertson developed^[13][1] a formulation for arbitrary Hermitian operator operators ${\displaystyle {\hat {\mathcal {O}}}}$ expressed in terms of their standard deviation $${\displaystyle \sigma _{\mathcal {O}}={\sqrt {\langle {\hat {\mathcal {O}}}^{2}\rangle -\langle {\hat {\mathcal {O}}}\rangle ^{2}}},}$$ where the brackets ${\displaystyle \langle {\hat {\mathcal {O}}}\rangle }$ indicate an expectation value of the observable represented by operator ${\displaystyle {\hat {\mathcal {O}}}}$. For a pair of operators ${\displaystyle {\hat {A}}}$ and ${\displaystyle {\hat {B}}}$, define their commutator as $${\displaystyle [{\hat {A}},{\hat {B}}]={\hat {A}}{\hat {B}}-{\hat {B}}{\hat {A}},}$$

and the Robertson uncertainty relation is given by^[14] $${\displaystyle \sigma _{A}\sigma _{B}\geq \left|{\frac {1}{2i}}\langle [{\hat {A}},{\hat {B}}]\rangle \right|={\frac {1}{2}}\left|\langle [{\hat {A}},{\hat {B}}]\rangle \right|.}$$

Erwin Schrödinger^[15] showed how to allow for correlation between the operators, giving a stronger inequality, known as the Robertson-Schrödinger uncertainty relation,^[16][1]

where the anticommutator, ${\displaystyle \{{\hat {A}},{\hat {B}}\}={\hat {A}}{\hat {B}}+{\hat {B}}{\hat {A}}}$ is used.

Proof of the Schrödinger uncertainty relation

The derivation shown here incorporates and builds off of those shown in Robertson,^[13] Schrödinger^[16] and standard textbooks such as Griffiths.^[17]: 138 For any Hermitian operator ${\displaystyle {\hat {A}}}$, based upon the definition of variance, we have $${\displaystyle \sigma _{A}^{2}=\langle ({\hat {A}}-\langle {\hat {A}}\rangle )\Psi |({\hat {A}}-\langle {\hat {A}}\rangle )\Psi \rangle .}$$ we let ${\displaystyle |f\rangle =|({\hat {A}}-\langle {\hat {A}}\rangle )\Psi \rangle }$ and thus $${\displaystyle \sigma _{A}^{2}=\langle f\mid f\rangle \,.}$$

Similarly, for any other Hermitian operator ${\displaystyle {\hat {B}}}$ in the same state $${\displaystyle \sigma _{B}^{2}=\langle ({\hat {B}}-\langle {\hat {B}}\rangle )\Psi |({\hat {B}}-\langle {\hat {B}}\rangle )\Psi \rangle =\langle g\mid g\rangle }$$ for ${\displaystyle |g\rangle =|({\hat {B}}-\langle {\hat {B}}\rangle )\Psi \rangle .}$

The product of the two deviations can thus be expressed as

${\displaystyle \sigma _{A}^{2}\sigma _{B}^{2}=\langle f\mid f\rangle \langle g\mid g\rangle .}$

(1)

In order to relate the two vectors ${\displaystyle |f\rangle }$ and ${\displaystyle |g\rangle }$, we use the Cauchy–Schwarz inequality^[18] which is defined as $${\displaystyle \langle f\mid f\rangle \langle g\mid g\rangle \geq |\langle f\mid g\rangle |^{2},}$$ and thus Equation (1) can be written as

${\displaystyle \sigma _{A}^{2}\sigma _{B}^{2}\geq |\langle f\mid g\rangle |^{2}.}$

(2)

Since ${\displaystyle \langle f\mid g\rangle }$ is in general a complex number, we use the fact that the modulus squared of any complex number ${\displaystyle z}$ is defined as ${\displaystyle |z|^{2}=zz^{*}}$, where ${\displaystyle z^{*}}$ is the complex conjugate of ${\displaystyle z}$. The modulus squared can also be expressed as

${\displaystyle |z|^{2}={\Big (}\operatorname {Re} (z){\Big )}^{2}+{\Big (}\operatorname {Im} (z){\Big )}^{2}={\Big (}{\frac {z+z^{\ast }}{2}}{\Big )}^{2}+{\Big (}{\frac {z-z^{\ast }}{2i}}{\Big )}^{2}.}$

(3)

we let ${\displaystyle z=\langle f\mid g\rangle }$ and ${\displaystyle z^{*}=\langle g\mid f\rangle }$ and substitute these into the equation above to get

${\displaystyle |\langle f\mid g\rangle |^{2}={\bigg (}{\frac {\langle f\mid g\rangle +\langle g\mid f\rangle }{2}}{\bigg )}^{2}+{\bigg (}{\frac {\langle f\mid g\rangle -\langle g\mid f\rangle }{2i}}{\bigg )}^{2}}$

(4)

The inner product ${\displaystyle \langle f\mid g\rangle }$ is written out explicitly as $${\displaystyle \langle f\mid g\rangle =\langle ({\hat {A}}-\langle {\hat {A}}\rangle )\Psi |({\hat {B}}-\langle {\hat {B}}\rangle )\Psi \rangle ,}$$ and using the fact that ${\displaystyle {\hat {A}}}$ and ${\displaystyle {\hat {B}}}$ are Hermitian operators, we find $${\displaystyle {\begin{aligned}\langle f\mid g\rangle &=\langle \Psi |({\hat {A}}-\langle {\hat {A}}\rangle )({\hat {B}}-\langle {\hat {B}}\rangle )\Psi \rangle \\[4pt]&=\langle \Psi \mid ({\hat {A}}{\hat {B}}-{\hat {A}}\langle {\hat {B}}\rangle -{\hat {B}}\langle {\hat {A}}\rangle +\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle )\Psi \rangle \\[4pt]&=\langle \Psi \mid {\hat {A}}{\hat {B}}\Psi \rangle -\langle \Psi \mid {\hat {A}}\langle {\hat {B}}\rangle \Psi \rangle -\langle \Psi \mid {\hat {B}}\langle {\hat {A}}\rangle \Psi \rangle +\langle \Psi \mid \langle {\hat {A}}\rangle \langle {\hat {B}}\rangle \Psi \rangle \\[4pt]&=\langle {\hat {A}}{\hat {B}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle +\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle \\[4pt]&=\langle {\hat {A}}{\hat {B}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle .\end{aligned}}}$$

Similarly it can be shown that ${\displaystyle \langle g\mid f\rangle =\langle {\hat {B}}{\hat {A}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle .}$

Thus, we have $${\displaystyle \langle f\mid g\rangle -\langle g\mid f\rangle =\langle {\hat {A}}{\hat {B}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle -\langle {\hat {B}}{\hat {A}}\rangle +\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle =\langle [{\hat {A}},{\hat {B}}]\rangle }$$ and $${\displaystyle \langle f\mid g\rangle +\langle g\mid f\rangle =\langle {\hat {A}}{\hat {B}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle +\langle {\hat {B}}{\hat {A}}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle =\langle \{{\hat {A}},{\hat {B}}\}\rangle -2\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle .}$$

We now substitute the above two equations above back into Eq. (4) and get $${\displaystyle |\langle f\mid g\rangle |^{2}={\Big (}{\frac {1}{2}}\langle \{{\hat {A}},{\hat {B}}\}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle {\Big )}^{2}+{\Big (}{\frac {1}{2i}}\langle [{\hat {A}},{\hat {B}}]\rangle {\Big )}^{2}\,.}$$

Substituting the above into Equation (2) we get the Schrödinger uncertainty relation $${\displaystyle \sigma _{A}\sigma _{B}\geq {\sqrt {{\Big (}{\frac {1}{2}}\langle \{{\hat {A}},{\hat {B}}\}\rangle -\langle {\hat {A}}\rangle \langle {\hat {B}}\rangle {\Big )}^{2}+{\Big (}{\frac {1}{2i}}\langle [{\hat {A}},{\hat {B}}]\rangle {\Big )}^{2}}}.}$$

This proof has an issue^[19] related to the domains of the operators involved. For the proof to make sense, the vector ${\displaystyle {\hat {B}}|\Psi \rangle }$ has to be in the domain of the unbounded operator ${\displaystyle {\hat {A}}}$, which is not always the case. In fact, the Robertson uncertainty relation is false if ${\displaystyle {\hat {A}}}$ is an angle variable and ${\displaystyle {\hat {B}}}$ is the derivative with respect to this variable. In this example, the commutator is a nonzero constant—just as in the Heisenberg uncertainty relation—and yet there are states where the product of the uncertainties is zero.^[20] (See the counterexample section below.) This issue can be overcome by using a variational method for the proof,^[21][22] or by working with an exponentiated version of the canonical commutation relations.^[20]

Note that in the general form of the Robertson–Schrödinger uncertainty relation, there is no need to assume that the operators ${\displaystyle {\hat {A}}}$ and ${\displaystyle {\hat {B}}}$ are self-adjoint operators. It suffices to assume that they are merely symmetric operators. (The distinction between these two notions is generally glossed over in the physics literature, where the term Hermitian is used for either or both classes of operators. See Chapter 9 of Hall's book^[23] for a detailed discussion of this important but technical distinction.)

In the phase space formulation of quantum mechanics, the Robertson–Schrödinger relation follows from a positivity condition on a real star-square function. Given a Wigner function ${\displaystyle W(x,p)}$ with star product ★ and a function f, the following is generally true:^[24] $${\displaystyle \langle f^{*}\star f\rangle =\int (f^{*}\star f)\,W(x,p)\,dx\,dp\geq 0~.}$$

Choosing ${\displaystyle f=a+bx+cp}$, we arrive at $${\displaystyle \langle f^{*}\star f\rangle ={\begin{bmatrix}a^{*}&b^{*}&c^{*}\end{bmatrix}}{\begin{bmatrix}1&\langle x\rangle &\langle p\rangle \\\langle x\rangle &\langle x\star x\rangle &\langle x\star p\rangle \\\langle p\rangle &\langle p\star x\rangle &\langle p\star p\rangle \end{bmatrix}}{\begin{bmatrix}a\\b\\c\end{bmatrix}}\geq 0~.}$$

Since this positivity condition is true for all a, b, and c, it follows that all the eigenvalues of the matrix are non-negative.

The non-negative eigenvalues then imply a corresponding non-negativity condition on the determinant, $${\displaystyle \det {\begin{bmatrix}1&\langle x\rangle &\langle p\rangle \\\langle x\rangle &\langle x\star x\rangle &\langle x\star p\rangle \\\langle p\rangle &\langle p\star x\rangle &\langle p\star p\rangle \end{bmatrix}}=\det {\begin{bmatrix}1&\langle x\rangle &\langle p\rangle \\\langle x\rangle &\langle x^{2}\rangle &\left\langle xp+{\frac {i\hbar }{2}}\right\rangle \\\langle p\rangle &\left\langle xp-{\frac {i\hbar }{2}}\right\rangle &\langle p^{2}\rangle \end{bmatrix}}\geq 0~,}$$ or, explicitly, after algebraic manipulation, $${\displaystyle \sigma _{x}^{2}\sigma _{p}^{2}=\left(\langle x^{2}\rangle -\langle x\rangle ^{2}\right)\left(\langle p^{2}\rangle -\langle p\rangle ^{2}\right)\geq \left(\langle xp\rangle -\langle x\rangle \langle p\rangle \right)^{2}+{\frac {\hbar ^{2}}{4}}~.}$$

Since the Robertson and Schrödinger relations are for general operators, the relations can be applied to any two observables to obtain specific uncertainty relations. A few of the most common relations found in the literature are given below.

• Position–linear momentum uncertainty relation: for the position and linear momentum operators, the canonical commutation relation ${\displaystyle [{\hat {x}},{\hat {p}}]=i\hbar }$ implies the Kennard inequality from above: $${\displaystyle \sigma _{x}\sigma _{p}\geq {\frac {\hbar }{2}}.}$$
• Angular momentum uncertainty relation: For two orthogonal components of the total angular momentum operator of an object: $${\displaystyle \sigma _{J_{i}}\sigma _{J_{j}}\geq {\frac {\hbar }{2}}{\big |}\langle J_{k}\rangle {\big |},}$$ where i, j, k are distinct, and J_i denotes angular momentum along the x_i axis. This relation implies that unless all three components vanish together, only a single component of a system's angular momentum can be defined with arbitrary precision, normally the component parallel to an external (magnetic or electric) field. Moreover, for ${\displaystyle [J_{x},J_{y}]=i\hbar \varepsilon _{xyz}J_{z}}$, a choice ${\displaystyle {\hat {A}}=J_{x}}$, ${\displaystyle {\hat {B}}=J_{y}}$, in angular momentum multiplets, ψ = |j, m⟩, bounds the Casimir invariant (angular momentum squared, ${\displaystyle \langle J_{x}^{2}+J_{y}^{2}+J_{z}^{2}\rangle }$) from below and thus yields useful constraints such as j(j + 1) ≥ m(m + 1), and hence j ≥ m, among others.

• For the number of electrons in a superconductor and the phase of its Ginzburg–Landau order parameter^[25][26] $${\displaystyle \Delta N\,\Delta \varphi \geq 1.}$$

The derivation of the Robertson inequality for operators ${\displaystyle {\hat {A}}}$ and ${\displaystyle {\hat {B}}}$ requires ${\displaystyle {\hat {A}}{\hat {B}}\psi }$ and ${\displaystyle {\hat {B}}{\hat {A}}\psi }$ to be defined. There are quantum systems where these conditions are not valid.^[27] One example is a quantum particle on a ring, where the wave function depends on an angular variable ${\displaystyle \theta }$ in the interval ${\displaystyle [0,2\pi ]}$. Define "position" and "momentum" operators ${\displaystyle {\hat {A}}}$ and ${\displaystyle {\hat {B}}}$ by $${\displaystyle {\hat {A}}\psi (\theta )=\theta \psi (\theta ),\quad \theta \in [0,2\pi ],}$$ and $${\displaystyle {\hat {B}}\psi =-i\hbar {\frac {d\psi }{d\theta }},}$$ with periodic boundary conditions on ${\displaystyle {\hat {B}}}$. The definition of ${\displaystyle {\hat {A}}}$ depends the ${\displaystyle \theta }$ range from 0 to ${\displaystyle 2\pi }$. These operators satisfy the usual commutation relations for position and momentum operators, ${\displaystyle [{\hat {A}},{\hat {B}}]=i\hbar }$. More precisely, ${\displaystyle {\hat {A}}{\hat {B}}\psi -{\hat {B}}{\hat {A}}\psi =i\hbar \psi }$ whenever both ${\displaystyle {\hat {A}}{\hat {B}}\psi }$ and ${\displaystyle {\hat {B}}{\hat {A}}\psi }$ are defined, and the space of such ${\displaystyle \psi }$ is a dense subspace of the quantum Hilbert space.^[28]

Now let ${\displaystyle \psi }$ be any of the eigenstates of ${\displaystyle {\hat {B}}}$, which are given by ${\displaystyle \psi (\theta )=e^{2\pi in\theta }}$. These states are normalizable, unlike the eigenstates of the momentum operator on the line. Also the operator ${\displaystyle {\hat {A}}}$ is bounded, since ${\displaystyle \theta }$ ranges over a bounded interval. Thus, in the state ${\displaystyle \psi }$, the uncertainty of ${\displaystyle B}$ is zero and the uncertainty of ${\displaystyle A}$ is finite, so that $${\displaystyle \sigma _{A}\sigma _{B}=0.}$$ The Robertson uncertainty principle does not apply in this case: ${\displaystyle \psi }$ is not in the domain of the operator ${\displaystyle {\hat {B}}{\hat {A}}}$, since multiplication by ${\displaystyle \theta }$ disrupts the periodic boundary conditions imposed on ${\displaystyle {\hat {B}}}$.^[20]