The appearance of particle tracks in detectors (Q2035921)

Ever since its observation the seemingly innocuous appearance of the atomic particle tracks in detectors has puzzled the physicists as one of the most striking manifestations of the weird properties of the quantum world. To make a long story short, in the quoted words of M.\ Born at the 1927 Solvay conference: ``If one associates a spherical wave with each emission process, how can one understand that the track of each \(\alpha\)-particle appears as a (very nearly) straight line? In other words: how can the corpuscular character of the phenomenon be reconciled here with the representation by waves?'' In particular, the \(S\)-wave of the outgoing \(\alpha\)-particle is spherically symmetric, but the particle tracks are not. How is, then, that this initial symmetry is broken? In a sense the typical answer -- given in a celebrated W.\ Heisenberg 1927 thought experiment -- is well known: it is an effect of the collapse of the wave packet due to the (repeated) measurements of the particle position. But this is famously an answer that begs a lot of explanation and that, ever since its formulation, has elicited heated discussions not to be summarized here. Admittedly however the present paper deals neither with a physical model for the atom ionization and the subsequent drop formation in a cloud chamber (along the lines, for example, of the quoted N.F.\ Mott 1929 paper), nor with an explanation of the strange nature of the quantum measurements, so that in particular no new insight is to be found about the wave packet collapse, or the \textit{Heisenberg cut}, notions that are simply accepted and used along the paper. For instance, in the words of the authors (page 438): ``To describe the effect of an instantaneous measurement of the approximate position of the particle on its state we follow the conventional wisdom of quantum mechanics: In the course of such a measurement whose result is given by some vector \(\mathbf q\in\mathbb{R}^d\), the state \(\rho\) of the particle changes according to'' the usual quantum rule of the wave packet collapse summarized in the subsequent equation (29). The focus of the discussion instead is to ``present a mathematically rigorous analysis of the appearance of particle tracks'' within the framework of the said ``conventional wisdom of quantum mechanics,'' namely to show that the track appearance is well accounted for with a scrupulous application of the quantum formalism. This is in any case a commendable task, and not a very easy one to carry out as the complexity of the subsequent discussion shows. More precisely the authors want to ``present a theoretical analysis of a gedanken experiment of the sort Heisenberg had in mind in 1927,'' with repeated position measurements every \(\tau\) seconds. In their discussion however they do not deal with \textit{idealized} quantum measurements, but they take instead the considerable trouble of discussing the case of \textit{approximate} measurements. The general states of the charged particle of mass \(M\) are here density operators \(\rho\) in a Hilbert space \(\mathcal H\) where \(\mathbf X, \mathbf P\) are the position and momentum operators, while the state vectors \(\Omega\) of the electromagnetic (EM) field plus photomultipliers live in another Hilbert space \(\mathfrak H\). The values \(\mathbf q\in\mathbb{R}^d\) (representing the approximate position measurements) of suitable operators \(\mathbf Q\) in \(\mathfrak H\) are then ``supposed to be tightly correlated with the positions, \(\mathbf x\in\mathbb{R}^d\), of the charged particle.'' Before each measurement the EM field and photomultipliers always are in the state \(\Omega_{in}\); then during the light-scattering the state changes according to a propagator \(U_t(\mathbf x)\) (\(\mathbf x\) being the position of the charged particle during the scattering process, and \(t\ll\tau\) the scattering time span) and quickly relaxes back to \(\Omega_{in}\) long before the subsequent measurement is performed. It is therefore possible to define the transition amplitude \(V_{\mathbf q}(\mathbf x)\) of equation (7) representing the probability density amplitude of finding \(\mathbf q\) when the particle is in \(\mathbf x\). The operators \(V_{\mathbf q}(\mathbf X)\) will then constitute a positive-operator-valued measure (POVM) that turns out to be instrumental to implement the wave packet reduction of every approximate position measurement. On the other hand the evolution between two measurements of the position-momentum pair \(\mathbf X,\mathbf P\) of the freely moving particle is described by a propagator \(U_S\) associated to a symplectic matrix \(S\) in the phase space \(\Gamma\) With the Gaussian \textit{ansatz} of equation (9) for the transition amplitude, a sequence of approximate measurements \(\mathbf q_0,\dots,\mathbf q_n\) falling in the subsets \(\Delta_0,\dots,\Delta_n\) entails a change in the initial density matrix \(\rho\) produced by a total operator \(W_n(\mathbf q_0,\dots,\mathbf q_n)= U_{S^{n+1}}V_{\mathbf q_n} (\mathbf X_{n\tau}) \ldots V_{\mathbf q_0} (\mathbf X)\) given as a repeated combination of \(V_{\mathbf q}\) and \(S\). Finally this gives rise to a probability measure \(\mathbb P_\rho\) on the process of the sequences \(\mathbf q_n\) that can be used to calculate the probability that the position of the particle at the times \(n\tau\) is within \(\Delta_n\). The aim of the paper now is to show that (page 435) ``with high probability, the cells \(\Delta_0,\dots,\Delta_n\) which indicate the positions of the particle at times \(0, \tau,\dots , n\tau\) , are centered in points ``close'' to \(\mathbf x(0), \mathbf x(\tau), \dots , \mathbf x(n\tau)\), respectively, where \(\mathbf x(t) =\mathbf x+t\mathbf v, t \in [0, n\tau]\), is the trajectory of a freely moving classical particle.'' Here \(\mathbf v\) is a value of \(\mathbf V=\mathbf P/M\) Without retracing in this short review all the details of their exhaustive discussion we will only recall next that the authors on the one hand study (by means of a suitable family of coherent states \(|W,\zeta\rangle \) centered around phase space points \(\zeta\in\Gamma\)) ``the stochastic dynamics of a (quasi-) freely moving quantum particle subjected to repeated measurements of its approximate position;'' and on the other they ``introduce a stochastic process [equation (39)] with values in the classical phase space of the particle that indexes the trajectory of coherent states occupied by the particle under the forward dynamics.'' Their first main result is then summarized in the Theorem 2.2 that ``relates the sequence of measurement data of approximate particle positions to the sequence of phase space points determined by the stochastic process in Eq. (39)'' by establishing an equality in law between the classical positions \(\xi_n\) of the particle -- plus an independent Gaussian noise \(\eta_n\) -- and the measurement results \(\mathbf Q_n\). In the Theorem 2.4 they next ``determine the best guess of the initial condition of a phase space trajectory of the stochastic process introduced in (39) from its tail,'' and finally the Theorem 2.8 ``relates the positive operator-valued measure (POVM) induced by sequences of approximate particle position measurements to a POVM taking values in the space of coherent states.'' The bulk of the paper is thereafter devoted to a long and rather convoluted sequence of technical arguments peppered with a great deal of lemmas and propositions needed to prove the said results, and ends finally with a few examples of free particles, harmonic oscillators and particles in a constant magnetic field to show the efficacity of the method.