Résumé : Observations over recent decades suggest that sea ice plays a significant role in global biogeochemical cycles, providing an active biogeochemical interface at the ocean-atmosphere boundary. Sea ice is a semisolid matrix permeated by a network of channels and pores. The brine-filled spaces are colonized by sympagic (ice-associated) communities that are both taxonomically diverse and metabolically active, with multiple trophic levels, efficiently consuming, reprocessing, and redistributing chemicals within the ice and exchanging with both the overlying atmosphere and the underlying ocean. Analyzing biogeochemical properties in sea ice is fundamentally complicated by its inherent heterogeneity and multiphase nature. This is especially illustrated by the lack of robust estimates for the most basic biogeochemical fluxes, such as primary production and carbon export. Measurements of sea-ice primary production are scarce and challenging. Accumulation of organic matter being trapped within sea ice during the growth season is likely to provide a conservative estimate of the net community production. More than 20 years ago, Legendre et al. [1] used the few available observations to infer Antarctic sea-ice primary productivity. We will revisit this estimation by using a much larger database (n = 421 ice-cores). Based on this compilation, a preliminary estimate for Antarctic sea-ice primary productivity is 35 Tg C yr-1, representing roughly 20% of the primary productivity in the seasonal ice zone. Sympagic communities are exposed to major biogeochemical and physical changes during their lifetime into the ice. A key adaptive response is the formation of biofilms, which play multiple roles in the entrapment, retention and survival of microorganisms. We will review the growing body of evidence that suggests that the biofilm is also playing a major role in sea-ice biogeochemical dynamics (e.g., macro- and micro-nutrient storage; microenvironments with distinct biogeochemical properties; …). Current methods for collecting sea-ice samples mostly involve melting an ice core, disrupting both geochemical continuum and equilibrium of the sample and erasing any spatial information. Although analyzing bulk liquids may provide information on centimeter-scale, this is insufficient to detect or describe microbial processes, operating on submillimetre scale in biofilms, yet there is currently no alternative option for these measurements. To overcome this issue, we will also introduce a passive sampling technique of Diffusive Gradients in Thin-films for 2-dimensional images of labile metals in soil/sediment that we are developing for sea-ice application. This new approach may provide new conceptual and quantitative understanding of biogeochemical processes regulating the distribution of key elements and solutes in the structurally multiphase complex sea-ice matrix.