Topological solitons propagating in mechanical metamaterials point toward potential applications in wave propagation, mechanical logic, and shape morphing. Inspired by the rich physics of topological solitons, researchers show here that this wave phenomenon can be also exploited to make a multistable metamaterial crawl. While previously proposed crawling machines usually require complex control of multiple actuators, this robot can be powered by a single actuator. All features needed for locomotion are embedded in the material and activated by the topological soliton, which here is the boundary between the two flexed forms of mechanical units linked in a chain.

Self-excitation of smart materials that respond to stimuli is important for building soft robots capable of self-sustainable untethered locomotion. However, in these machines the sophisticated interaction and energy flow between materials and stimuli are not well understood. The authors study a self-excited hydrogel cantilever oscillator irradiated by constant light, and find that the synergy between the photoinduced moment and oscillation ensures positive work input into the cantilever to overcome damping. Scaling analysis of the energy flow of the system reveals how the stable oscillation amplitude is governed by various geometric and material parameters.

Designing the propagation of single quanta at subwavelength scale is of major interest for integrated optical quantum information transfer. This study demonstrates the high potential of crystalline silicon wires with nanometric sections as low-loss nanochannels for quantum nanophotonics. Single photons from N-$V$ centers in nanodiamonds can be guided and spectrally filtered over several micrometers while conserving their quantum statistics. Moreover, these nanowaveguides can be designed so that only light from emitters of specific orientations is efficiently guided, enabling quantum state selectivity and modal control of single-photon transfer in subwavelength waveguides.

Key applications of quantum computing, including error correction, rely critically on the capability to measure (read out) the state of selected quantum bits, without disturbing other qubits or terminating the computation. Such $midcircuit$ $measurements$ are now being implemented in a few cutting-edge platforms, but their development is hindered by lack of a means to measure their performance and characterize the errors that they produce. The authors show how to extend a popular tomography method for precise characterization of midcircuit measurements, and using their QILGST protocol discover a pernicious kind of non-Markovian error, which once identified can be handled.

By its nature, a conventional computer based on deterministic bits is ill-matched to tasks such as sampling, inference, and optimization. A $probabilistic$ computer can be a natural tool for problems which are at base probabilistic. Unstable magnetic tunnel junctions (MTJs) provide compact, energy-efficient adjustable random-number generators for probabilistic computing, but an important practical concern is device variability. The authors show experimentally that even nonideal MTJs can function properly when utilizing $in$ $situ$ Boltzmann machine learning. The learning circuit here could be of interest for standalone devices capable of fast, efficient learning at the edge.

Meet Alice, Bob, and David. Establishing secure encryption keys between any two of these users is important for a long-range communication network, but with dubious Charlie also around, it becomes tricky. Twin-field quantum key distribution (TFQKD), while promising for this application, has been limited so far to two users, because of its requirement for phase stability of optical signals. In this work, the authors demonstrate a TFQKD network by connecting three users in a Sagnac fiber ring, a configuration with inherent phase stability. This experiment suggests that the Sagnac TFQKD system is an effective, practical approach to implementing a long-range secure communication network.

Collision phenomena are important in determining the structures and interactions of atoms and molecules. A broad concern is how to investigate collisions $in$ $situ$ under wide-ranging experimental conditions. To this end, the authors develop a spin-noise spectroscopic approach to measure key collision parameters in alkali atoms and inert atoms or molecules. The technique offers broad bandwidth (tens of GHz) and fine resolution (1 ppm), outperforming existing spin-noise methods. This approach is expected to aid research on a great range of collision phenomena under ambient conditions, such as nonperturbative study of cold atom-molecule collisions.

The nonlinear dynamics of current-driven skyrmions are essential for designing skyrmion-based spintronic devices. However, Thiele’s approach uses a rigid-body assumption that is not suitable here, as skyrmions may change shape at high speeds. This study extends Thiele’s equation by including the radial and tangential force components. With this approach, one may depict the inner structure of a current-driven skyrmion and propose a way to suppress deformation with an in-plane magnetic field. This extended analytic formulation may help to describe the nonlinear dynamics of other spin textures as well, and to promote applications based on such dynamics.

A Coulomb-blockade thermometer (CBT) provides calibration-free measurement of low temperatures, even in strong magnetic fields—but slowly. A standard conductance measurement with a CBT takes typically minutes, which is useless in applications where relatively fast changes in temperature are monitored. Besides, a slow measurement is prone to drifts, and low-frequency noise. The authors solve this issue by hooking a CBT sensor into a radio-frequency tank circuit, making the measurement about 1000 times faster. Their approach is expected to impact low-temperature thermometry, for $e.g.$ applications in quantum technology.

Topological solitons propagating in mechanical metamaterials point toward potential applications in wave propagation, mechanical logic, and shape morphing. Inspired by the rich physics of topological solitons, researchers show here that this wave phenomenon can be also exploited to make a multistable metamaterial crawl. While previously proposed crawling machines usually require complex control of multiple actuators, this robot can be powered by a single actuator. All features needed for locomotion are embedded in the material and activated by the topological soliton, which here is the boundary between the two flexed forms of mechanical units linked in a chain.

Self-excitation of smart materials that respond to stimuli is important for building soft robots capable of self-sustainable untethered locomotion. However, in these machines the sophisticated interaction and energy flow between materials and stimuli are not well understood. The authors study a self-excited hydrogel cantilever oscillator irradiated by constant light, and find that the synergy between the photoinduced moment and oscillation ensures positive work input into the cantilever to overcome damping. Scaling analysis of the energy flow of the system reveals how the stable oscillation amplitude is governed by various geometric and material parameters.

Designing the propagation of single quanta at subwavelength scale is of major interest for integrated optical quantum information transfer. This study demonstrates the high potential of crystalline silicon wires with nanometric sections as low-loss nanochannels for quantum nanophotonics. Single photons from N-$V$ centers in nanodiamonds can be guided and spectrally filtered over several micrometers while conserving their quantum statistics. Moreover, these nanowaveguides can be designed so that only light from emitters of specific orientations is efficiently guided, enabling quantum state selectivity and modal control of single-photon transfer in subwavelength waveguides.

Key applications of quantum computing, including error correction, rely critically on the capability to measure (read out) the state of selected quantum bits, without disturbing other qubits or terminating the computation. Such $midcircuit$ $measurements$ are now being implemented in a few cutting-edge platforms, but their development is hindered by lack of a means to measure their performance and characterize the errors that they produce. The authors show how to extend a popular tomography method for precise characterization of midcircuit measurements, and using their QILGST protocol discover a pernicious kind of non-Markovian error, which once identified can be handled.

By its nature, a conventional computer based on deterministic bits is ill-matched to tasks such as sampling, inference, and optimization. A $probabilistic$ computer can be a natural tool for problems which are at base probabilistic. Unstable magnetic tunnel junctions (MTJs) provide compact, energy-efficient adjustable random-number generators for probabilistic computing, but an important practical concern is device variability. The authors show experimentally that even nonideal MTJs can function properly when utilizing $in$ $situ$ Boltzmann machine learning. The learning circuit here could be of interest for standalone devices capable of fast, efficient learning at the edge.

Meet Alice, Bob, and David. Establishing secure encryption keys between any two of these users is important for a long-range communication network, but with dubious Charlie also around, it becomes tricky. Twin-field quantum key distribution (TFQKD), while promising for this application, has been limited so far to two users, because of its requirement for phase stability of optical signals. In this work, the authors demonstrate a TFQKD network by connecting three users in a Sagnac fiber ring, a configuration with inherent phase stability. This experiment suggests that the Sagnac TFQKD system is an effective, practical approach to implementing a long-range secure communication network.