Development of an online hydrogen fuel quality analyzer with gas chromatography-pulsed discharge helium ionization detector for applying hydrogen infrastructures.

Hydrogen fuel, which is essential for the hydrogen economy, including hydrogen cell vehicles, must be of high quality for optimal hydrogen cell use. Currently, hydrogen fuel quality control is mainly done by offline analysis with periodic sampling. However, with the anticipated surge in hydrogen charging stations, there's a pressing need for cost-effective, high-throughput online analysis systems. Additionally, the miniaturization of these analytical instruments for field application is also a challenge. In this study, we present a compact, real-time hydrogen fuel analyzer based on gas chromatography with a pulsed discharge helium ionization detector. Its dual-column system efficiently analyzes major impurities in hydrogen fuel in less than 30 min. Indicator species (CO, CO2, CH4, O2, N2, and additional hydrogen sulfide [H2S]) are determined by examining hydrogen production and supply processes. The analyzer's measurement capability is consistent with µmol/mol-level analysis, providing valuable real-time information for hydrogen infrastructure managers. Additionally, it can analyze H2S, a crucial marker of sulfur compounds acting as catalytic poisons in fuel cells. This real-time analyzer offers efficient, informed decision-making support for hydrogen infrastructure managers, enhancing the overall reliability of hydrogen fuel in fuel-cell electric vehicles.

Spontaneous emergence of rudimentary music detectors in deep neural networks.

Music exists in almost every society, has universal acoustic features, and is processed by distinct neural circuits in humans even with no experience of musical training. However, it remains unclear how these innate characteristics emerge and what functions they serve. Here, using an artificial deep neural network that models the auditory information processing of the brain, we show that units tuned to music can spontaneously emerge by learning natural sound detection, even without learning music. The music-selective units encoded the temporal structure of music in multiple timescales, following the population-level response characteristics observed in the brain. We found that the process of generalization is critical for the emergence of music-selectivity and that music-selectivity can work as a functional basis for the generalization of natural sound, thereby elucidating its origin. These findings suggest that evolutionary adaptation to process natural sounds can provide an initial blueprint for our sense of music.

Learning Entropy Production via Neural Networks.

This Letter presents a neural estimator for entropy production (NEEP), that estimates entropy production (EP) from trajectories of relevant variables without detailed information on the system dynamics. For steady state, we rigorously prove that the estimator, which can be built up from different choices of deep neural networks, provides stochastic EP by optimizing the objective function proposed here. We verify the NEEP with the stochastic processes of the bead spring and discrete flashing ratchet models and also demonstrate that our method is applicable to high-dimensional data and can provide coarse-grained EP for Markov systems with unobservable states.

Peptoid-Conjugated Magnetic Field-Sensitive Exciplex System at High and Low Solvent Polarities.

The magnetic field effect (MFE) in exciplex emission (ExE) has been studied for decades, but it has been observed to occur only in solvents with a limited range of polarity. This limitation is mainly due to the reversible interconversion collapse between two quenching products of the photoinduced electron transfer, the exciplex and magnetic field-sensitive radical ion pair (RIP) beyond that polarity range. In a nonpolar solvent, the formation of RIPs is suppressed, whereas in a polar solvent, the probability of their re-encounter forming the exciplexes decreases. In this study, we developed new exciplex-forming (phenyl-phenanthrene)-(phenyl-N,N-dimethylaniline)-peptoid conjugates (PhD-PCs) to overcome this limitation. The well-defined peptoid structure allows precise control of the distance and the relative orientation between two conjugated moieties. Steady-state and time-resolved spectroscopic data indicate that the PhD-PCs can maintain the reversibility, which allows MFEs in ExE regardless of the solvent polarity. Subtle differences between the ExEs of the PhD-PCs were observed and explained by their exciplex geometries obtained through time-dependent density functional theory (TD-DFT) calculations.

Quantitative imaging of magnetic field distribution using a pyrene-based magnetosensing exciplex fluorophore.

Quantitative imaging of magnetic field distribution was carried out using a pyrene-based magnetosensing exciplex fluorophore, pyrene-(CH2)12-O-(CH2)2-N,N-dimethylaniline (Py-12-O-2-DMA), on a conventional fluorescence microscope with an off-the-shelf LED lamp. No continuous sample supply was required for the process. The solvent system (anisole : DMF, 50 : 50 (v/v)) was carefully selected for monitoring the extent of modulation caused by the external magnetic field. The emission from Py-12-O-2-DMA increased by ca. 1.5 times under an external magnetic field of 50 mT. The pyrene-based reporter was ca. 24.7 times brighter than a previously reported phenanthrene-based complex when excited by using the widely available 355 nm excitation. Moreover, the maximum wavelength up to which Py-12-O-2-DMA could be excited (up to 380 nm) was longer than the wavelength up to which Phen-12-O-2-DMA could be excited. The combined advantages allowed the capture of magnetic field images with a high S/N ratio under milder conditions such as low illumination power, reduced sample concentration, and simpler optical setup. The system was also found to be feasible for 3D magnetic field distribution imaging by two-photon fluorescence microscopy.