Soil Biofilms in Pollutant Dynamics and Detoxification

Soil biofilms are structured, dynamic microbial consortia embedded within extracellular polymeric substances that regulate microscale physicochemical heterogeneity and drive biogeochemical transformations in soils. Despite increasing interest in biofilm-mediated remediation, current reviews have largely examined microbial ecology, engineered biofilm functions, and predictive modelling independently, limiting systems-level understanding of pollutant fate in complex soils. This review, therefore, proposes a revised conceptual framework integrating biofilm ecology, synthetic biology, and AI-driven predictive modelling to improve mechanistic and predictive understanding of emerging pollutant detoxification. Emerging pollutants, including pharmaceuticals, pesticides, per- and polyfluoroalkyl substances, micro- and nanoplastics, and heavy metals, exhibit persistence, bioaccumulation, and mixture-dependent effects that challenge conventional remediation strategies. Biofilm matrices function as reactive interfaces facilitating adsorption, sequestration, and enzymatic transformation, while steep redox and nutrient gradients support metabolically diverse processes such as cometabolism, syntrophic degradation, and biomineralisation. Increasing evidence indicates that quorum sensing, horizontal gene transfer, and low-abundance microbial taxa contribute significantly to adaptive responses and functional plasticity within biofilms. Advances in high-resolution imaging, spatial multi-omics, and microfluidic platforms have resolved previously inaccessible biofilm architectures and processes; however, integration with machine learning and process-based modelling remains limited, restricting field-scale prediction of pollutant behaviour and remediation outcomes. Synthetic biology enables targeted optimisation of biofilm functions, whereas AI-driven models enhance prediction of contaminant transport, transformation, and detoxification. Soil biofilms function both as sinks and catalytic hotspots, and resolving this duality through a predictive, systems-level framework represents a major advance beyond existing descriptive reviews.