When velocity autocorrelations mirror force autocorrelations: Exact noise-cancellation in interacting Brownian systems
Anton Lüders, Suvendu Mandal, Thomas Franosch · 2026
Read When velocity autocorrelations mirror force autocorrelations: Exact noise-cancellation in interacting Brownian systems by Anton Lüders, Suvendu Mandal, Thomas Franosch (2026) 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-2603-08918.txt.
Introduction
Anton Lüders and colleagues prove when a noise-cleaning trick for Brownian walks is exact. The method splits paths into free motion plus interaction shifts, dropping cross terms without prior proof. In thermal equilibrium the cross terms vanish, since velocity links track negative force links exactly, while out of equilibrium they survive as a mark of nonequilibrium physics.
Catalog record
| CREATOR | Anton Lüders, Suvendu Mandal, Thomas Franosch |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 95729 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.08918
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
When velocity autocorrelations mirror force autocorrelations: Exact noise-cancellation in interacting Brownian systems Anton Lüders, Suvendu Mandal, Thomas Franosch Abstract Resolving the mean-squared displacement (MSD) and velocity autocorrelation function (VACF) of interacting Brownian particles remains a central challenge in simulations of soft-matter systems, especially at low densities where particle-particle interactions are sparse and signals are dominated by thermal noise. A recently proposed noise-cancellation (NC) algorithm [Mandal et al. Phys. Rev. Lett. 123, 168001 (2019)] addresses this by decomposing particle trajectories into two components: free Brownian motion and interaction-induced displacements. The NC approximation enhances signal clarity by neglecting cross-correlations between the total particle displacements and the extracted interaction-induced contributions of the trajectories — an assumption that has so far lacked rigorous theoretical justification. In this work, we establish an exact theoretical relation between the VACF, the force autocorrelation function (FACF) characterizing the interaction-induced contributions, and these cross-correlations, which is valid for Brownian systems. We show that in thermal equilibrium, the cross-correlations vanish for Brownian systems because the VACF is strictly proportional to the negative FACF, which establishes the NC algorithm as an exact method. In contrast, for Brownian nonequilibrium systems, the cross-correlations remain finite, providing a direct fingerprint of nonequilibrium physics in such systems and a criterion to distinguish equilibrium from nonequilibrium states. Here, suitable corrections must be applied for the NC method to remain accurate. Our results expand the scope of the NC algorithm to a broad range of soft-matter systems in and out of equilibrium, where it has the potential to strongly enhance the resolution of VACF data obtained through simulations in future studies. Phys. Rev. E 113, 035305 (2026). I Introduction Understanding how microscopic fluctuations and their correlations decay over time is essential to characterize the transport properties of liquids, complex fluids, and colloidal suspensions. A central quantity in this context is the velocity autocorrelation function (VACF), which measures how the velocity of a tagged particle remains correlated with itself over time. The VACF not only reveals the underlying dynamical processes at play but also connects directly to macroscopic transport coefficients, such as the diffusivity, via the Green-Kubo relations Mouas et al. 2012 ; Heptner and Dzubiella 2015 ; Hansen and McDonald 2013 ; Kubo et al. 2012 . In simple molecular fluids, the VACF encodes rich dynamical information across time scales. Classical kinetic theory predicts an exponential decay of the VACF in dilute gases due to uncorrelated binary collisions Williams 1972 ; Cichocki and Felderhof 1994 . However, groundbreaking simulations by Alder and Wainwright revealed a drastic departure from this behavior: in Newtonian many-body systems, the VACF exhibits a long-time algebraic decay as t^{-3/2} in three dimensions Alder and Wainwright 1970 . This long-time tail, now firmly established both theoretically and experimentally Ernst et al. 1970 ; Dorfman and Cohen 1970 ; Lukić et al. 2005 ; Jeney et al. 2008 ; Franosch et al. 2011 ; Huang et al. 2011 , arises from the hydrodynamic backflow resulting from momentum conservation — a hallmark of collective fluid behavior. In colloidal systems, by contrast, this picture changes fundamentally. Due to strong viscous damping from the surrounding solvent, momentum is rapidly dissipated and no longer conserved at the particle scale. Instead, long-time dynamics are governed by the particle-number conservation Mandal et al. 2019 . This leads to a steeper but negative VACF decay, typically scaling as t^{-5/2} in three dimensions Hanna et al. 1981 ; Ackerson and Fleishman 1982 ; Felderhof and Jones 1983 [or t^{-(d+2)/2} in the d -dimensional case]. Accurately resolving this long-time behavior is crucial for determining transport coefficients such as the diffusivity, yet extracting the VACF from stochastic simulations remains a formidable challenge. The difficulty arises because the VACF signal decays rapidly and becomes increasingly buried in noise at long times, especially in overdamped dilute systems. As a result, achieving statistically converged estimates typically requires a prohibitively large number of independent simulations Frenkel 1987 ; Mandal et al. 2019 ; Ramírez et al. 2010 . This is a long-standing problem in computational statistical mechanics, where various strategies, such as fluctuation-subtraction techniques Oettinger 1994 ; Melchior and Öttinger 1996 ; Wagner and Öttinger 1997 , have been developed to reduce statistical variance. In the specific context of time-dependent correlation functions, several optimized sampling approaches have been proposed to enhance statistical efficiency Siems and Nielaba 2018 ; Ramírez et al. 2010 ; Dubbeldam et al. 2009 ; Frenkel 1987 . A promising recent advance is the noise-cancellation (NC) algorithm introduced by two of the current authors and collaborators Mandal et al. 2019 for the calculation of the mean-squared displacement (MSD) and the VACF in Brownian dynamics (BD) simulations. Building on earlier ideas by Frenkel Frenkel 1987 , the NC method decomposes particle displacements into two components: the free Brownian motion and a contribution due to interparticle forces. By neglecting the cross-correlations between the total particle displacements and the extracted interparticle-force contributions, the method reconstructs the VACF from the second time derivative of the mean-squared displacement (MSD), drastically improving the signal-to-noise ratio at long times. This gain in signal is achieved because the remaining non-trivial terms of the MSD differ from zero only for direct particle-particle interactions, eliminating the need to average out the Brownian fluctuations Rusch et al. 2024 . Accordingly, no explicit dependence on the Brownian noise persists. This strategy proved to increase the resolution of the long-time tails of the VACF by several orders of magnitude Mandal et al. 2019 , and it was shown that numerical MSD results obtained by the NC procedure achieve a higher precision than standard algorithms when studying dilute or weakly interacting systems Rusch et al. 2024 . However, the theoretical basis for the NC approximation remains incomplete. While numerical evidence suggests that the ignored cross-correlation terms are indeed negligible in equilibrium and dilute conditions, the underlying reasons for this remain unclear. Moreover, in nonequilibrium or strongly interacting Brownian systems, these cross-correlations could be significant, potentially undermining the accuracy of the NC method. In this work, we develop a rigorous theoretical foundation for the NC algorithm that is applicable to systems that can be described by overdamped Langevin equations (i.e., Brownian systems). We derive an exact relation connecting the VACF, the force autocorrelation function (FACF), and the cross-correlation terms omitted in the NC method. Our key result is that, in thermal equilibrium, the NC algorithm is an exact expression for Brownian systems, meaning that the cross-correlation term always vanishes under such conditions. This result not only explains the remarkable accuracy of the NC method in BD simulations but also provides the first rigorous justification for its application to equilibrium soft-matter setups. In striking contrast, once Brownian systems are driven out of equilibrium, the cross-correlations remain finite and cannot be ignored. These non-vanishing contributions encode a distinct fingerprint of nonequilibrium physics, thereby offering a powerful tool for distinguishing equilibrium from nonequilibrium states. Lastly, we find that, even in nonequilibrium, the NC algorithm can sometimes be adjusted to give reasonable or even exact results, which expands the computation scheme of Ref. Mandal et al. 2019 to Brownian nonequilibrium systems for the first time. This work is structured as follows: First, we summarize the fundamentals of the NC algorithm as discussed in Refs. Mandal