Oil-mist coalescing filters are commonly compared using filtration efficiency and pressure drop from short tests, although retained liquid continues to redistribute during operation. This critical review examines how liquid morphology, distribution and connectivity affect the interpretation of wet performance and service-life evidence. Imaging reveals substantial local saturation differences and extensive inter-fiber liquid structures. Within one hybrid–Janus comparison, whole-filter saturation ranged from 54.47% to 61.96%, whereas source-reported wet-to-clean pressure-drop ratios ranged from 3.32 to 22.36. These observations support considering liquid arrangement alongside bulk inventory, but do not isolate connectivity from pore geometry and wettability. Long-duration experiments also show that drainage can coexist with pressure-drop creep over 1100 h. We distinguish measured endpoints, observation durations and lifetime estimates extrapolated from short degradation records. A traceable service-life claim requires defined loading and operating conditions, concurrent resistance and outlet measurements, an explicit endpoint, and separation of measured duration from predicted life. A proposed reporting framework and a conditional pressure-scaling analysis identify the evidence needed to assess transfer from laboratory media to compressed-air cartridges.

Water-vapor removal from compressed air is governed by the product pressure dew point (PDP), dynamic bed utilization, hydraulic loss, regeneration demand, and cyclic reliability, yet activated-alumina studies are often interpreted from powder Brunauer–Emmett–Teller (BET) area or equilibrium uptake alone. This critical narrative review evaluates where evidence can and cannot be transferred across material, formed-particle, packed-bed, cycle, and field scales. It distinguishes equilibrium capacity, static test capacity, dynamic breakthrough capacity, and usable cycle capacity; conditionally compares desiccants and regeneration modes; and links PDP-defined breakthrough, mass-transfer-zone (MTZ)/length-of-unused-bed (LUB) measures, non-spherical-particle pressure drop, thermal waves, and regeneration endpoints. The available evidence supports several directional conclusions but not universal design values: alkali modification can increase uptake despite lower surface area; forming and binder chemistry alter accessible pores and strength; wall effects and particle geometry require measured hydrodynamic validation; and reported regeneration savings of approximately 27–40% are architecture-specific relative results rather than matched absolute energy benchmarks. A baseline-normalized diagnostic framework separates reversible regeneration faults and feed contamination from material aging and particle/bed degradation. The review identifies the central evidence gap as the absence of matched datasets reporting formed-body properties, adiabatic breakthrough, segmented pressure drop, regeneration energy, and long-cycle failure for the same material. It concludes with a standards-coverage map, a transparent screening sensitivity analysis, and a prioritized validation agenda.