We analyze three distinct approaches to time of arrival (TOA) distributions for discrete quantum systems using a spin-12 particle in a constant magnetic field as a paradigmatic example. We argue that these distributions should not be regarded as competing predictions for the same notion of arrival time, but rather relate to fundamentally different notions whose relevance depends on the physical context. These results are used to analyze information propagation arrival time distributions in XX spin chain systems, and discuss potential applications in quantum information science.
This paper investigates the question of arrival time at a given state in a discrete quantum system. As a paradigmatic example, we ask how long it takes for a spin-12 particle at rest in a constant magnetic field, initially prepared say in the spin-up state, to be detected in the spin-down state. Clearly, the question “When does a spin flip?” must be understood operationally not in terms of the evolution of an underlying actual state, but in terms of measurement outcomes. In the absence of measurement, the spin undergoes unitary dynamics, evolving into a coherent superposition of up and down states, with the Schrödinger equation deterministically fixing the periodic times at which the transition probability reaches unity. Such times, however, do not correspond to observable events. A spin-flip time can therefore only be defined through measurement statistics, namely as the time at which the target state is first detected within a specified measurement protocol. The present work does not address ontological questions concerning the fundamental nature of time in quantum mechanics, but instead focuses on the operational definition and experimental reconstruction of time arrival distributions within the standard formalism.
If the Born rule provides the probability density for measuring a quantum state at a fixed time, the standard formalism of quantum mechanics, unfortunately, offers no direct prescription for the probability density of a time measurement associated with a fixed state. This asymmetry underlies the so-called time-of-arrival (TOA) problem, which stems from the treatment of time as a parameter rather than an observable and from the absence of a self-adjoint operator canonically conjugate to the Hamiltonian [1], while extensive efforts have been devoted to define arrival-time distributions for continuous systems, with no consensus emerging so far [2,3], comparatively little systematic attention has been devoted to time measurements in discrete quantum systems, especially from a unified operational perspective that explicitly contrasts distinct measurement protocols. Most approaches developed for continuous position measurements do not readily extend to this setting, to the notable exceptions of [4,5], where an application to spin systems is explicitly mentioned, and [6,7] building on earlier work for continuous systems [8,9,10].
Our view is that differences in predicted TOA distributions for both continuous and discrete systems do not merely result from competing approaches to a single notion of TOA, but often reflect divergences or ambiguities in the underlying definition of a quantum arrival time itself. In classical dynamics, a straightforward procedure exists for obtaining an unambiguous notion of arrival time: one may simply continuously monitor the system and record the first hitting time of a target state without altering the motion. In quantum mechanics, however, measurement unavoidably disturbs the system. Repeated projective measurements of a target state collapse the wavefunction and can freeze the dynamics through the quantum Zeno effect [11,12]. Conversely, weak or continuous measurements reduce backaction but lead to non-unitary open-system evolution describable by stochastic master equations or POVMs (positive-operator-valued measures) [13,14]. Consequently, any meaningful definition of a time-of-arrival (TOA) distribution must explicitly specify the measurement protocol under which it is defined.
In this paper, we restrict our attention to the unperturbed unitary evolution of the system and propose a systematic comparison of three operationally distinct notions of arrival time within a single discrete-spin framework. Accordingly, we focus on approaches in which arrival statistics are extracted from single-shot measurements performed at prescribed times, rather than from repeated strong projections or continuous weak monitoring that modify the dynamics. In this context, we extend the existing literature on time-of-arrival (TOA) distributions in discrete quantum systems by examining three distinct notions of TOA for a spin system, analyzing the corresponding probability distributions, and discussing their potential applications. We first show that cumulative time distributions for discrete systems can be inferred from transition probabilities given by the Born rule, which are experimentally accessible through projective measurements performed at discrete times on independently prepared systems. Differentiation of these cumulative distributions yields a first notion of TOA, referred to as the quantum flow TOA distribution (or QF distribution for short), closely related to earlier flux and stochastic approach to TOA in [8,9,10], and to the time-of-flow concept introduced operationally in [6,7]. We compare this notion with the quantum clock (QC) approach. The term “quantum clock” is used in different ways in the literature, ranging from physical timekeeping devices employed in quantum metrology to relational constructions in which time is inferred from correlations between subsystems. In the present work, we adopt the operational perspective developed in [4,5], where a clock degree of freedom is introduced and time is defined conditionally from joint measurements on the system and the clock. Within this framework, one asks: given that the system is detected in a specified arrival state, what is the probability distribution of the corresponding clock reading? The resulting time distribution is therefore conditional upon successful detection of the target state, and differs conceptually from flow-based or first-detection constructions. We emphasize that our use of the quantum clock approach serves as a benchmark notion of arrival time within a unified operational comparison, rather than as the central focus of the present analysis. We finally introduce another notion, termed the quantum waiting time distribution (or QW distribution for short), defined as the waiting time distribution until a first detection of the target state occurs under single measurements on an ensemble of identical systems. Since each measurement is performed on an independent copy of the system, the resulting QW TOA distribution characterizes the stopping time of an experimental protocol rather than a fundamental property of the system’s dynamics itself such as the first-arrival time of a single system undergoing successive strong measurements. As such, the QW TOA distribution represents a protocol-dependent observable capturing the latency of a measurement procedure. We also provide an explicit experimental protocol for the empirical reconstruction of each distribution via discrete-time projective measurements, in the spirit of stroboscopic measurement schemes discussed in [6,7,8,9,10,15]. We use these results to analyze information propagation time distributions in XX spin chain systems via the quantum wait TOA distribution and analyze its relation with the Lieb–Robinson bound [16]. Finally, we conclude by emphasizing that the operational relevance of each notion of arrival time is context dependent, and that each approach naturally gives rise to distinct applications in quantum sensing, computing, and communication.
