Abstract. Over recent decades the highest rates of water column warming and sea ice loss across the Arctic Ocean have been observed in the Barents Sea. These physical changes have resulted in rapid ecosystem adjustments, manifesting as a northward migration of temperate phytoplankton species at the expense of silica-based diatoms. These changes will potentially alter the composition of phytodetritus deposited at the seafloor, which acts as a biogeochemical reactor and is pivotal in the recycling of key nutrients, such as silicon (Si). To appreciate the sensitivity of the Barents Sea benthic system to the observed changes in surface primary production, there is a need to better understand this benthic–pelagic coupling. Stable Si isotopic compositions of sediment pore waters and the solid phase from three stations in the Barents Sea reveal a coupling of the iron (Fe) and Si cycles, the contemporaneous dissolution of lithogenic silicate minerals (LSi) alongside biogenic silica (BSi), and the potential for the reprecipitation of dissolved silicic acid (DSi) as authigenic clay minerals (AuSi). However, as reaction rates cannot be quantified from observational data alone, a mechanistic understanding of which factors control these processes is missing. Here, we employ reaction–transport modelling together with observational data to disentangle the reaction pathways controlling the cycling of Si within the seafloor. Processes such as the dissolution of BSi are active on multiple timescales, ranging from weeks to hundreds of years, which we are able to examine through steady state and transient model runs. Steady state simulations show that 60 % to 98 % of the sediment pore water DSi pool may be sourced from the dissolution of LSi, while the isotopic composition is also strongly influenced by the desorption of Si from metal oxides, most likely Fe (oxyhydr)oxides (FeSi), as they reductively dissolve. Further, our model simulations indicate that between 2.9 % and 37 % of the DSi released into sediment pore waters is subsequently removed by a process that has a fractionation factor of approximately −2 ‰, most likely representing reprecipitation as AuSi. These observations are significant as the dissolution of LSi represents a source of new Si to the ocean DSi pool and precipitation of AuSi an additional sink, which could address imbalances in the current regional ocean Si budget. Lastly, transient modelling suggests that at least one-third of the total annual benthic DSi flux could be sourced from the dissolution of more reactive, diatom-derived BSi deposited after the surface water bloom at the marginal ice zone. This benthic–pelagic coupling will be subject to change with the continued northward migration of Atlantic phytoplankton species, the northward retreat of the marginal ice zone and the observed decline in the DSi inventory of the subpolar North Atlantic Ocean over the last 3 decades.