Development of durable commercial/heavy fuel cell (FC) vehicles is required to decrease carbon dioxide emission from commercial/heavy vehicles. The mileage of heavy vehicles is far higher than that of passenger vehicles. Because FC stacks in vehicles deteriorate during long-term operation, their lifetime should be maximized by increasing durability. To improve FC stack durability, we must avoid catalyst degradation. When the hydrogen supply on the anode catalyst surface is inadequate, carbon of the cathode catalyst is oxidized with water. This occurs where there is insufficient hydrogen, such as at anode outlet during cell starting or water flooding. It is therefore important to prevent deterioration of FC operation caused by inadequate hydrogen. In stacks where the anode gas is circulated, nitrogen permeates and remains at the anode until it is exhausted, which lowers hydrogen concentration. Moreover, hydrogen is consumed in the piling direction, so its concentration is lower downstream. This leads to degradation at the anode outlet, so a high hydrogen flow rate is required during operation.To optimize hydrogen supply through feedback control, it is beneficial to measure the local hydrogen partial pressure. Hydrogen distributions exist in the reactive plane, gas manifolds, and piping system. To avoid FC deterioration caused by insufficient hydrogen, hydrogen concentration needs to be measured in the in-plane direction. However, it is difficult to measure hydrogen distribution in the in-plane direction. Some commercial sensors may affect FC materials. General hydrogen sensors are unsuitable for hydrogen concentrations over 4% inside FC systems.Raman spectroscopy is a useful hydrogen detection method. Generally, leaked hydrogen gas is intermittently detected with Raman spectroscopy. The intensity of Raman scattering is about 10−4 times that of Rayleigh scattering, and low-density gas scattering is weak compared with that of a liquid or solid. However, the Raman shift of hydrogen vibration (4160 cm−1) is far larger than that of other materials and does not require high wavelength resolution. Moreover, the hydrogen partial pressure in FC systems is high and easily detected.We developed a new method to measure the hydrogen Raman scattering from the reacting surface of an operating FC. During operation, hydrogen is consumed in the anode channel and its partial pressure drops near the outlet. To confirm this phenomenon with Raman spectroscopy, we needed to detect scattering light from the anode channel during cell operation. This measurement also required optics able to resolve very weak hydrogen Raman scattering. For these reasons, we fabricated a visualization cell that transmitted pump and scattered light without gas leakage. Pump light was introduced into a flow field in a channel with a depth of 1 mm, so the channel worked as an optical path . The optical path was from a mirror on the cathode flow field toward the anode to avoid interference with sealing materials. Sapphire windows were included in the anode flow field to allow visualization of scattered light.Using this cell and an Intensified Charge-Coupled Device (ICCD ) camera, we vertically observed Raman scattering from hydrogen. We included a bandpass filter with a center wavelength of 420-nm and halfwidth of 10 nm inside the camera lens to capture images of 417-nm scattered light from hydrogen excited by 355-nm pump light through the filter. A 355-nm notch filter was also added to remove the optical noise from Rayleigh scattering, allowing visualization of only hydrogen.Before the measurement, scattering intensity was calibrated using 100% hydrogen and nitrogen to clarify the relation between scattering intensity and partial pressure. The flow field was filled with 100% hydrogen (100 kPa) and an intensity distribution was obtained. Pure nitrogen was also measured to evaluate the dark noise and the correlation between intensity and hydrogen partial pressure. Local partial pressure was determined from the intensity at a desired point.We then measured the scattered light during cell operation at 100 kPaG. Hydrogen with a stoichiometric ratio of 1.1 diluted 45% by nitrogen was supplied to the cell to mimic stack operation. The cell operated at 0.5 A/cm2. The scattering distribution was measured at 10 Hz and ten measurements were integrated to determine hydrogen partial pressure distribution from the scattered light intensity . Hydrogen partial pressure decreased from 80 to 40 kPa along the in-plane flow, consistent with our rough predictions of 90 and 38 kPa, respectively. This data was thought to be affected by the sensitization noise from the photomultiplier tube elements in the ICCD camera. Suppressing optical and electronic noise or conducting an imaging process with dispersion treatment will allow detection of various gases with this Raman method.