Technological evolution of constructed wetlands: a big data patent analysis

Constructed wetlands (CWs) have been demonstrated as an efficient nature-based technology for treating various types of wastewaters. Although extensive reviews in the CW field focus predominantly on academic literature, systematic analyses grounded in patent documentation are notably limited. In fact, patent literature analysis proves indispensable for elucidating the developmental trajectories, research priorities, and innovation pathways of CW technologies. This study retrieved a dataset of 8579 patents spanning three decades from the Derwent Innovation Index database, employing tailored search strings and IPC classifications. An integrated methodology combining bibliometric analysis and text mining was adopted to examine both structured and unstructured patent datasets. Through the application of latent Dirichlet allocation (LDA), an unsupervised machine learning method, six key technical domains were identified and analyzed: substrate optimization, ecological effects, nitrogen and phosphorus removal mechanisms, configuration improvements, process coupling, and industry expansion. Technological advancements have followed a trajectory from standalone treatment technologies to high-efficiency optimization processes and, eventually, to integrated coupled treatment systems. Principal component analysis (PCA) identified three promising directions for future technological innovation: reducing greenhouse gas emissions while enhancing carbon sequestration, synergistic control and resource recovery of emerging contaminants, and the development and regulation of intelligent wetland systems. This systematic patent-based analysis offers valuable decision-making support and strategic guidance for driving innovation and advancing technological development in the field of CWs.