The Sub-10 µm Detection Gap Driving the Entire Field

Microplastics — plastic particles smaller than 5 mm — have been detected in marine environments, freshwater bodies, soils, food chains, and human tissue, and the analytical tools available to characterise them have not kept pace with the scale of the problem. The core challenge, well documented in the academic literature retrieved for this landscape, is the detection of particles below 10 µm: a size class that defeats conventional optical microscopy and standard infrared methods. As one retrieved source states directly, "adequate analytical tools to sample, isolate, detect, quantify, and characterize small microplastics (<10 µm) are urgently needed."

This analytical gap is not merely academic. Patent filings across all nine jurisdictions in this dataset — KR, CN, US, EP, JP, IT, WO, IN, and FR — reflect direct engineering responses to it. Every major cluster in the landscape, from Raman spectroscopy to deep learning holography, traces its core technical claim to the problem of detecting, classifying, and quantifying particles at or below this threshold. According to WIPO, environmental technology patents have seen consistent growth over the past decade, and microplastics detection represents one of the most active emerging sub-fields within that trend.

The landscape spans filings from 2019 through early 2026 and encompasses five broad technical domains: spectroscopic identification (Raman, FTIR, NIR, UV, LIBS, photothermal IR); imaging-based analysis including hyperspectral, fluorescence, holographic, and CNN/deep learning approaches; electrochemical and microfluidic sensing; in-situ aquatic sampling and pretreatment; and physical/chemical removal systems including filtration, coagulation/flocculation, microbubble cleaning, and adhesive capture.

Spectroscopic Identification: The Most Patent-Dense Cluster

Spectroscopic identification — particularly Raman spectroscopy in combined or enhanced configurations — is the most patent-dense cluster in this dataset, with filings from China, South Korea, the United States, and Europe all targeting chemical identification of microplastic particles. Raman dominates because it can characterise polymer chemistry non-destructively, but its conventional form suffers from low throughput; the patent record reflects a sustained engineering effort to overcome this limitation.

Fudan University (CN, 2022) filed a Raman Spectral Imaging System that integrates stimulated and spontaneous Raman modes with laser scanning to overcome the throughput bottleneck of conventional spontaneous Raman. Ocean University of China (CN, 2020) combined fluorescence imaging with Raman scanning and a tunable laser to achieve high-throughput, wide-band spectral acquisition with automated target location for near-shore sediment analysis. Both filings target the same operational problem from different angles: speed without sacrificing chemical specificity.

Near-infrared (NIR) approaches appear in multiple filings targeting plastic type classification. Nextem Co. (KR, 2022) filed a system using similarity-scoring against a reference spectral database that self-updates with new confirmed measurements — a key operational advantage for field systems where polymer types encountered may vary. Wenzhou University (CN, 2020) addressed co-contaminant detection with a LIBS-based non-destructive method for single-particle microplastic composite heavy metal contamination, providing rapid full-element analysis without sample preparation.

The newest entry in the spectroscopic cluster is Photothermal Spectroscopy Corp. (US, 2025), which combines crossed-polarization or autofluorescence position detection with photothermal IR absorption to achieve sub-micron characterisation beyond the diffraction limit of conventional FTIR — directly addressing the sub-10 µm detection gap that defines the field's core challenge. Research into photothermal IR approaches is also tracked by organisations such as NIST, which maintains reference standards for polymer characterisation relevant to this emerging technique.

"Adequate analytical tools to sample, isolate, detect, quantify, and characterize small microplastics (<10 µm) are urgently needed" — retrieved academic literature, as cited in this patent landscape dataset.

AI and Imaging-Based Detection: The Fastest-Growing Frontier

Machine learning integration with optical imaging is the fastest-growing approach in this dataset, with the majority of filings in this cluster dated 2023–2026 — a clear signal that AI has moved from experimental to expected in new detection system design. Any new entrant building a microplastics detection platform without ML-based classification and correction is now positioned below the current technical frontier as evidenced by the 2023–2026 filing cluster.

Nanjing University (CN, 2024) filed a machine learning-based high-precision field monitoring system that builds a two-model correction architecture: a regression model for quantity correction and a classification model for type correction, both trained on physicochemical water parameters alongside in-situ Raman data, and calibrated periodically against laboratory FTIR standards. This architecture — field sensor plus ML correction plus periodic laboratory calibration — is becoming the reference design for deployable environmental monitoring systems.

South China Institute of Environmental Sciences (Ministry of Ecology, CN, 2023) applies a CNN model to water-surface microplastic imagery, using Wiener filter-based deblurring to correct for environmental lighting and wind effects prior to grayscale CNN inference — a practical engineering solution to the noise problem that has historically limited outdoor optical detection. Jinan University (CN, 2025) introduced a custom architecture with MFM (multi-scale feature fusion), FS-C3k2 (spatial-frequency feature extraction), and GSConvE (deep feature enhancement) modules for water body applications.

At the frontier, Shenzhen University of Technology (CN, 2026) filed a deep learning-based holographic microplastic screening system integrating high-speed holography with deep learning reconstruction for simultaneous high-throughput and high-precision screening — a label-free, high-throughput morphological classification approach using wavefront information rather than chemical spectroscopy. The University of Hong Kong (CN, 2024) filed a complementary polarization digital holography method and device for microplastic identification. These holographic approaches represent a genuine technical discontinuity from the spectroscopic mainstream.

Hyperspectral imaging for aquatic distribution mapping is covered by Nature and Technology Co. (KR, 2023) and Heon-ju Lee (KR, 2025), the latter using neural networks trained on standard spectra to profile particle position, material, and size in field samples. X Development LLC (US, 2024) — a Google subsidiary — filed a sensor fusion and model refinement architecture combining microplastics detection sensors with ancillary environmental sensors to continuously refine a trained detection model in deployment, marking a significant technology player's entry into this space. Research on AI applications in environmental sensing is also tracked by Nature, which has published extensively on machine learning approaches to environmental pollutant characterisation.

Removal Technology: The 5:1 Underdog and Its Opportunity

Removal technology is significantly underrepresented in the microplastics patent landscape relative to detection: among retrieved results, detection patents outnumber removal patents approximately 5:1. This gap is most acute for open-water and large-scale environmental remediation, where autonomous or passive removal systems remain technically immature and commercially underexplored — representing one of the most significant R&D and commercialization opportunities in the field.

Kemira Oyj (EP, 2023) uses fluorescence intensity and light scattering to evaluate optimal coagulant/flocculant chemistry for a given water matrix — linking detection capability directly to treatment optimization in a single workflow. IBM Corporation (JP, 2024) employs microbubble transducers in a cleaning chamber to dislodge and float microplastics for recovery by a filter chamber, with AI-modelled autonomous filtration workflows. These represent the more sophisticated end of the removal technology spectrum.

Procter & Gamble (CN, 2025) discloses pressure-sensitive adhesive articles repurposed from hygiene product manufacturing for removing microplastics and nanoplastics from wastewater effluent and laundry runoff. The patent cites data indicating that the textile washing pathway accounts for more than 20 km³ of microplastic-contaminated water annually — a figure that contextualises the scale of the problem the adhesive approach is designed to address. Universidad Autonoma de Madrid (BR, 2024) introduces magnetic iron material particles for selective microplastic separation followed by oxidative removal of organic non-plastic interferents, a compact chemically integrated approach suited for laboratory quantification workflows.

Autonomous aquatic collection platforms represent the most forward-looking segment of the removal cluster. NITTE University (IN, 2026) filed a GPS-guided autonomous floating treatment vessel integrating real-time water quality sensing with microplastic removal and mineral restoration functions. Morimoto Nobuyoshi (JP, 2024) filed a multi-vessel networked system to optimise microplastic harvest routes using inter-vessel data sharing — a coordinated fleet approach to ocean remediation. The OECD has identified marine plastic pollution as a priority environmental challenge, estimating that without intervention, plastic flows into aquatic environments will continue to grow through 2060.

Geographic Landscape: Korea Leads, China Focuses on Software, Nanoplastics Are Whitespace

South Korea is the dominant filing jurisdiction in this dataset by absolute count, with at least 25 distinct patent records. Korean assignees include Korea Institute of Industrial Technology, Korea Electronics Technology Institute, Korea Institute of Photonics Technology, Korea Institute of Construction Technology, University of Seoul, Pukyong National University, and multiple SMEs including Idham Environmental Technology Co., Cellabiotech, Nextem, and Nature and Technology Co. — reflecting a concentrated national research program distributed across public research institutes and universities.

China accounts for at least 14 retrieved records, with strong institutional clustering around environmental monitoring applications. Key CN assignees include Nanjing University, Fudan University, Ocean University of China, Guangdong University of Technology, South China Institute of Environmental Sciences (Ministry of Ecology), Southeast University, University of Hong Kong, and Shenzhen University of Technology. Critically, China's filings are uniformly directed at real-time water monitoring, AI-augmented detection, and contamination control modelling — a software-layer focus that creates a potential gap in Chinese domestic hardware IP and an opportunity for non-Chinese sensor companies to enter the Chinese market through licensing or joint venture arrangements.

Europe contributes detection systems and removal chemistry through LADAR Limited (EP/US), the most geographically prolific assignee in the dataset with filings across EP, US, and continuation records through 2025; Kemira Oyj (FI, via CA and EP filings) for coagulation chemistry; DWI Leibniz-Institut (WO, 2019) for biomolecular separation; and IFP Energies Nouvelles (FR, 2026) for pyrolysis quantification in soil matrices. The European Patent Office (EPO) has noted environmental technology as one of the fastest-growing patent categories in Europe, consistent with the activity seen in this dataset.

India shows early-stage but notable activity with NITTE University's autonomous floating treatment vessel (2026) and Tezpur University's smartphone-integrated fluorescence sensing device (2026) — both targeting the non-portability limitation of laboratory-grade detection systems and indicating emerging innovation capacity in the domain.

The most strategically significant gap in the geographic landscape is the near-absence of nanoplastics (<1 µm) filings. While the academic literature flags nanoplastics as a priority concern — research tracked by organisations including the WHO has raised questions about nanoplastic exposure routes in human health — the Korea Institute of Industrial Technology's 2024 nanoplastic detection system is one of very few filings specifically targeting this size class. This represents a high-value, low-competition IP whitespace for research organisations and companies willing to invest in the technical challenge of sub-micron characterisation.

Six Emerging Directions Shaping the 2026 Filing Wave

The 2024–2026 filing cohort in this dataset reveals six distinct emerging directions, each representing a departure from the established spectroscopic and sampling mainstream that dominated filings through 2022.

1. AI-Integrated Real-Time Field Monitoring

The convergence of in-situ Raman and NIR sensors with machine learning correction models signals a transition from laboratory-only spectroscopic identification to deployable environmental monitoring infrastructure. Nanjing University's two-model correction architecture (CN, 2024) — regression for quantity, classification for type — is the reference design for this approach. The patent landscape from PatSnap's innovation intelligence platform confirms that this architecture is being replicated and extended across multiple Chinese institutional filers.

2. Holographic and Polarization-Based Detection

The University of Hong Kong's polarization digital holography method (CN, 2024) and Shenzhen University of Technology's deep learning holographic screening system (CN, 2026) represent a push toward label-free, high-throughput morphological classification using wavefront information rather than chemical spectroscopy. This approach is technically distinct from all prior art in the spectroscopic cluster and may establish a new sub-cluster in the landscape.

3. Portable and Smartphone-Integrated Platforms

Tezpur University's portable smartphone-integrated sensing device (IN, 2026) combines nanostructure-assisted fluorescence with smartphone integration and on-site thermal enhancement — explicitly designed to overcome the non-portability of laboratory systems. This direction reflects a broader democratisation trend in environmental sensing, where cost and portability constraints are addressed through consumer hardware integration.

4. Novel Removal Chemistries and Materials

Procter & Gamble's adhesive-based removal (CN, 2025) and Universidad Autonoma de Madrid's magnetic separation (BR, 2024) both represent non-filtration capture mechanisms operable at scale in wastewater streams. IFP Energies Nouvelles' pyrolysis quantification (FR, 2026) addresses the measurement gap in soil matrices using controlled pyrolysis temperature sequencing to separate hydrocarbon contributions from microplastics and organic matter — a foundational method for agricultural and contaminated site assessment.

5. Multi-Vessel Autonomous Collection Networks

Morimoto Nobuyoshi's inter-vessel digital communication system for coordinated microplastic harvesting (JP, 2024) and NITTE University's GPS-guided autonomous treatment vessel (IN, 2026) indicate movement toward networked, autonomous ocean remediation infrastructure — a significant conceptual step beyond single-vessel collection systems.

6. Gene-Chip-Based Microplastic Source Tracing

Shenzhen AcuTe Environmental Technology Co. (CN, 2024) filed a gene detection method using DNA hybridisation arrays to identify microbial communities attached to microplastics, enabling source attribution by ecosystem. This forensic application has direct implications for regulatory enforcement: if the microbial community attached to a microplastic particle can identify its source ecosystem, regulators gain a new evidentiary tool for tracing plastic pollution to its origin. This direction has no close prior art in the dataset and represents a genuine technology discontinuity.

"Korea's public research institute ecosystem generates high filing volume but with fragmented assignee ownership — IP strategists targeting licensing opportunities in sensor hardware and sampling pretreatment should prioritise Korean research institutes."