During the century from the publication of the work by Einstein (1905 Ann. Phys. 17 549) Brownian motion has become an important paradigm in many fields of modern science. An essential impulse for the development of Brownian motion theory was given by the work of Langevin (1908 C. R. Acad. Sci., Paris 146 530), in which he proposed an ‘infinitely more simple’ description of Brownian motion than that by Einstein. The original Langevin approach has however strong limitations, which were rigorously stated after the creation of the hydrodynamic theory of Brownian motion (1945). Hydrodynamic Brownian motion is a special case of ‘anomalous Brownian motion’, now intensively studied both theoretically and in experiments. We show how some general properties of anomalous Brownian motion can be easily derived using an effective method that allows one to convert the stochastic generalized Langevin equation into a deterministic Volterra-type integro-differential equation for the mean square displacement of the particle. Within the Gibbs statistics, the method is applicable to linear equations of motion with any kind of memory during the evolution of the system. We apply it to memoryless Brownian motion in a harmonic potential well and to Brownian motion in fluids, taking into account the effects of hydrodynamic memory. Exploring the mathematical analogy between Brownian motion and electric circuits, which are at nanoscales also described by the generalized Langevin equation, we calculate the fluctuations of charge and current in RLC circuits that are in contact with the thermal bath. Due to the simplicity of our approach it could be incorporated into graduate courses of statistical physics. Once the method is established, it allows bringing to the attention of students and effectively solving a number of attractive problems related to Brownian motion.

The dynamics of flexible polymers in dilute solutions is studied taking into account the hydrodynamic memory, as a consequence of fluid inertia. As distinct from the Rouse-Zimm (RZ) theory, the Boussinesq friction force acts on the monomers (beads) instead of the Stokes force, and the motion of the solvent is governed by the nonstationary Navier-Stokes equations. The obtained generalized RZ equation is solved approximately using the preaveraging of the Oseen tensor. It is shown that the time correlation functions describing the polymer motion essentially differ from those in the RZ model. The mean-square displacement (MSD) of the polymer coil is at short times approximately t(2) (instead of approximately t). At long times the MSD contains additional (to the Einstein term) contributions, the leading of which is approximately t. The relaxation of the internal normal modes of the polymer differs from the traditional exponential decay. It is displayed in the long-time tails of their correlation functions, the longest lived being approximately t(-3/2) in the Rouse limit and t(-5/2) in the Zimm case, when the hydrodynamic interaction is strong. It is discussed that the found peculiarities, in particular, an effectively slower diffusion of the polymer coil, should be observable in dynamic scattering experiments.

The attenuation function S(t) for an ensemble of spins in a magnetic-field gradient is calculated by accumulation of the phase shifts in the rotating frame resulting from the displacements of spin-bearing particles. The found S(t), expressed through the particle mean square displacement, is applicable for any kind of stationary stochastic motion of spins, including their non-markovian dynamics with memory. The known expressions valid for normal and anomalous diffusion are obtained as special cases in the long time approximation. The method is also applicable to the NMR pulse sequences based on the refocusing principle. This is demonstrated by describing the Hahn spin echo experiment. The attenuation of the NMR signal is also evaluated providing that the random motion of particle is modeled by the generalized Langevin equation with the memory kernel exponentially decaying in time.

Stochastic motion of charged particles in the magnetic field was first studied almost half a century ago in the classical works by Taylor and Kurşunoğlu in connection with the diffusion of electrons and ions in plasma. In their works the long-time limits of the mean square displacement (MSD) of the particles have been found. Later Furuse on the basis of standard Langevin theory generalized their results for arbitrary times. The currently observed revival of these problems is mainly related to memory effects in the diffusion of particles, which appear when colored random forces act on the particles from their surroundings. In the present work an exact analytical solution of the generalized Langevin equation has been found for the motion of the particle in an external magnetic field when the random force is exponentially correlated in the time. The obtained MSD of the particle motion across the field contains a term proportional to the time, a constant term, and contributions exponentially decaying in the time. The results are more general than the previous results from the literature and are obtained in a considerably simpler way applicable to many other problems of the Brownian motion with memory. PACS numbers: 05.40.Jc, 05.40.Ca, 05.10.Gg, 66.10.cg I. INTRODUCTIONThe stochastic motion of charged particles in external magnetic fields is an old problem, the interest to which appeared in connection with the diffusive processes in plasma. In the work [1] Taylor considered the diffusion of ions in plasma across the magnetic field with the stochasticity arising from the fluctuations of the electric field. The equations of motion for the ions were the Langevin equations with the dynamic friction on an ion proportional to its velocity. The solution for the mean square displacement (MSD) was found in the limit of long times. Kurşunoğlu [2,3] derived the formulas for the diffusion of charged particles across and along the magnetic field as for the motion of Brownian particles with fluctuating accelerations and found the MSD from the probability density distribution for the particle velocity. In the paper [3] also a possible anisotropy in the dynamical friction has been considered by introducing the friction anisotropy matrix. In this case the MSD, obtained also for long times, and the diffusion coefficients in two perpendicular directions