Randomly measured quantum particles and thermal noise
Victor Gurarie · 2025
Read Randomly measured quantum particles and thermal noise by Victor Gurarie (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-06382.txt.
Introduction
Victor Gurarie studies a quantum particle, in a 2025 preprint, whose position is measured at random times and places. Each measurement nudges the particle's motion. He contrasts that motion with a particle pushed by a random potential, like thermal noise. At first glance the two setups look alike: both jitter the particle in an unpredictable way. The question is whether any experiment can tell them apart. Gurarie computes averages of quantities built from the particle's density matrix.
For quantities that use the density matrix in a straight, linear way, no difference shows. The two motions are twins at that level. Nonlinear quantities, though, do separate them. There, measurement-driven motion leaves marks that noise-driven motion does not. The result matters for lab systems where detectors sample particles at random. What looks like plain noise may carry the fingerprint of measurement itself.
Catalog record
| CREATOR | Victor Gurarie |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-06382
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Randomly measured quantum particles and thermal noise Victor Gurarie Abstract We consider the motion of a quantum particle whose position is measured in random places at random moments in time. We contrast this motion with the motion of a quantum particle in a potential which varies randomly in space and in time, which could also be thought of as (possibly thermal) noise. We calculate expectations of observables both linear and nonlinear in the density matrix. We demonstrate explicitly that while linear observables cannot distinguish between random measurements and random noise, measurable distinctions can be seen in nonlinear observables. \graphicspath ./Figures/ \DeclareGraphicsRule *mps* It became widely appreciated in recent years that dynamic evolution of a quantum system, when it is subject to measurements, acquires new features Li et al. 2018 ; Skinner et al. 2019 . Unlike the Schrödinger evolution of a state of a quantum particle, described by a unitary evolution operator, measurements evolve the wave function according to Born’s rules, that is by application of nonunitary projection operators to the wave function. Evolution of the wave function by these abrupt projections are difficult to capture analytically. Instead, weak measurement operators can be introduced, describing the process where a system is coupled to an ancilla which is then projectively measured. The only condition that the weak measurement operators must satisfy is the Kraus operator condition \sum_{j}\hat{K}_{j}^{\dagger}\hat{K}_{j}=1, (1) where the index j labels the outcomes of the measurement. These operators can be chosen to be close to identity, thus providing us with wave function evolution continuous in time Murciano et al. 2023 ; Garratt et al. 2023 ; Poboiko et al. 2023 ; Guo et al. 2024 . For example, for a particle whose position is measured at random moments in time can be described by the Kraus operator \hat{K}\left[V\right]=\int dx\,e^{V(x)\delta t}\left|x\right>\left<x\right|, (2) where V(x) is a function corresponding to the outcomes of the measurement procedure (thus it replaces the index j above) and \delta t is the time interval between the measurements. In particular, choosing V(x)\delta t=1 for a<x<b and V(x)\rightarrow-\infty otherwise reduces \hat{K}[V] to a projection operator onto the interval [a,b] , so its application to the wave function describes the measurement outcome where the particle is confined to this interval. If instead V(x) remains arbitrary, the limit of small \delta t turns \hat{K}[V] into an operator very close to identity and allows for continuous evolution in time. Summation over j is now replaced by the integration over the functions V . Taking the limit of continuous time evolution results in the Schrödinger-like evolution equation i\frac{\partial\psi}{\partial t}=-\frac{1}{2m}\frac{\partial^{2}\psi}{\partial x^{2}}+iV(x,t)\psi, (3) with V(x,t) random in space and time, which can be taken to be a Gaussian random function with the average \left\langle\,V(x_{1},t_{1})V(x_{2},t_{2})\,\right\rangle=\lambda\,\delta(t_{1}-t_{2})\,W(x_{1}-x_{2}). (4) Importantly, when summation over the discrete index j is replaced by averaging over Gaussian random functions V , the operator \hat{K}[V] defined in ( 2 ) satisfies the Kraus condition ( 1 ), up to an overall constant which can be removed with proper normalization. We can take W in the form W(x)=\frac{1}{\ell}\,g\left(\frac{x}{\ell}\right), (5) where \int d\xi\,g(\xi)=1. (6) \ell can be thought of as a length of the interval where the particle is measured. \ell\rightarrow 0 limit corresponds to g reducing to a delta-function. A randomly measured quantum particle can be contrasted with a quantum particle moving in a potential, random in both space and time (a particle subject to noise, possibly thermal Kubo 1966 ; Landau and Lifshitz 1980 ; Kamenev 2011 ), described by the time-dependent Schrödinger equation i\frac{\partial\psi}{\partial t}=-\frac{1}{2m}\frac{\partial^{2}\psi}{\partial x^{2}}+V(x,t)\psi, (7) The only difference between the equations ( 3 ) and ( 7 ) is the imaginary unit in front of the random potential making the evolution described by the equation ( 3 ) nonunitary. Both setups involve a random potential, so to distinguish them we will call ( 3 ) random measurements and ( 7 ) random noise. In what follows we show that the expectation values of operators when averaged over random measurements such as encoded in the Schrödinger equation ( 3 ) coincide with the expectation values of operators averaged over noise, as encoded in ( 7 ). In particular, quantum mechanical average of the square of the position of the particle, further averaged over either random measurements or noise, as a function of time t is given by ( 17 ). To be able to distinguish between random measurements and random noise, we need look at higher powers of the quantum mechanical averages of operators, subsequently averaged over either measurements or noise. A quantum mechanical average of the position operator, which is then squared and averaged over measurements or noise, is given by either ( 32 ) or ( 34 ) respectively. Here is how we derive these results. If we are interested in measuring expectation value of some operator, such as the position operator of the particle, at a certain time t , we construct the density matrix \rho(x_{+},x_{-},t)=\psi(x_{+},t)\psi^{*}(x_{-},t) averaged over the random potential V . The averaging over the random potential can be done explicitly, which is especially easy to carry out in Feynman’s path integral approach to quantum mechanics. Indeed, the wave function is given by (in case of random measurements) \psi(y_{+},t)=\int_{x_{+}(t)=y_{+}}{\cal D}x_{+}(\tau)\,e^{i\int d\tau\left(\frac{m\dot{x}_{+}^{2}}{2}-iV(x_{+})\right)}, (8) while its complex conjugate is given by the complex conjugate counterpart of this functional integral \psi^{*}(y_{-},t)=\int_{x_{-}(t)=y_{+}}{\cal D}x_{-}(\tau)\,e^{-i\int d\tau\left(\frac{m\dot{x}_{-}^{2}}{2}+iV(x_{-})\right)}. (9) The x_{+}(\tau) and x_{-}(\tau) trajectories are often referred to as belonging to the forward and backward Keldysh contours Kamenev 2011 . To average the density matrix, we multiply these together and use \left\langle\,e^{\int d\tau V(x_{+})+\int d\tau V(x_{-})}\,\right\rangle=e^{\lambda\int d\tau\left(W(x_{+}-x_{-})+W(0)\right)}, (10) where brackets denote averaging over the random potential V by using ( 4 ). Note that W(0) is just an irrelevant constant which can be dropped. This immediately leads to the equation satisfied by the density matrix (it could be called a Schrödinger-like equation, or a Schrödinger equation for the density matrix in the Choi-Jamiołkowski representation Jiang et al. 2013 , or a Liouville equation) i\dot{\rho}=-\frac{1}{2m}\frac{\partial^{2}\rho}{\partial x_{+}^{2}}+\frac{1}{2m}\frac{\partial^{2}\rho}{\partial x_{-}^{2}}+i\lambda W(x_{+}-x_{-})\rho. (11) If alternatively we were interested in averaging over noise, as opposed to random measurements, we would write down \psi(y_{+},t)=\int_{x_{+}(t)=y_{+}}{\cal D}x_{+}(\tau)\,e^{i\int d\tau\left(\frac{m\dot{x}_{+}^{2}}{2}-V(x_{+})\right)}, (12) \psi^{*}(y_{-},t)=\int_{x_{-}(t)=y_{+}}{\cal D}x_{-}(\tau)\,e^{-i\int d\tau\left(\frac{m\dot{x}_{-}^{2}}{2}-V(x_{-})\right)}. (13) The averaging now leads to \left\langle\,e^{i\left(\int d\tau V(x_{-})-\int d\tau V(x_{+})\right)}\,\right\rangle=e^{\lambda\int d\tau\left(W(x_{+}-x_{-})-W(0)\right)}, (14) which in turn implies that the density matrix satisfies exactly the same equation ( 11 ), since W(0) , as was already pointed out, is an irrelevant constant. We now come to a striking conclusion that observables averaged over measurements or averaged over a random potential are indistinguishable from each other. Now the equation ( 11 ) can be solved quite generally. Suppose initially at t=0 the particle is confined to a particular region in space, \psi_{0}(x)=\exp\left(-x^{2}/(4\Delta^{2})\right)/\left(2\pi\Delta^{2}\right)^{1/4} . With this initial wave function, \left<\psi_{0}\right|\hat{x}^{2}\left|\psi_{0}\right>=\int dx\,\rho(x,x,0)\,x^{2}=\Delta^{2}. (15) In other words, we need to solve equation ( 11 ) with the initial conditions \rho(x_{+},x_{-},0)=\frac{1}{\sqrt{2\pi}\Delta}\exp\left(-\frac{x_{+}^{2}+x_{-}^{2}}{4\Delta^{2}}\right). (16) The solution can be found analytically for any function W . In particular it can be shown that at a time t the average square of the position of the particle is given Rosenbluth 1992 ; Gurarie 2024 by (see also Appendix A ) \int dx\,\rho(x,x,t)\,x^{2}=\Delta^{2}+\frac{\hbar^{2}}{4m^{2}\Delta^{2}}t^{2}+\frac{\lambda\left|W^{\prime\prime}(0)\right|}{3m^{2}}t^{3}. (17) Here to further elucidate the meaning of this answer the Planck constant \hbar was explicitly restored. The