# Scientific papers Benchmarking Low-Cost Particulate Matter Sensors: Evaluating \ Performance Under Controlled Environmental Conditions Using \ Low-Cost Experimental Setups ---------------------------- **Title**: Benchmarking Low-Cost Particulate Matter Sensors: Evaluating Performance Under Controlled Environmental Conditions Using Low-Cost Experimental Setups\ **Authors**: Arianna Alvarez Cruz, Olivier Schalm, Luis Ernesto Morera Hernández, Alain Martínez Laguardia, Daniellys Alejo Sánchez, Mayra C. Morales Pérez, Rosa Amalia González Rivero, Yasser Morera Gómez\ **Affiliations**: Central University “Marta Abreu” of Las Villas (Cuba), Antwerp Maritime Academy (Belgium), Cienfuegos Center of Environmental Studies (Cuba)\ Project: Comparative assessment of low-cost PM sensors (NextPM vs OPC-N3) under laboratory and tropical field conditions\ **Publisher**: Atmosphere – MDPI, Basel, Switzerland (peer-reviewed journal)\ **Date**: February 2025\ **Document** Type: Peer-reviewed scientific article (open access, 22 pages)\ **DOI / URL**: This study benchmarked two low-cost particulate matter sensors—the Alphasense OPC-N3 and Tera Sensor's NextPM—through controlled lab experiments (clean air and water aerosols) and a 27-day field deployment in Cienfuegos, Cuba. The aim was to assess their stability, noise, humidity sensitivity, and reliability using raw data only, without post-processing or external calibration.\ The NextPM sensor demonstrated superior performance with **80% noise reduction** in clean-air tests (PM2.5: 0.3 ± 0.2 µg/m³ vs. OPC-N3: 0.4 ± 0.5 µg/m³) and **fewer outliers** (79 vs. 87 for PM10). In water aerosol tests, it recorded a **max PM10 of 11.3 µg/m³**, compared to **103.3 µg/m³ for the OPC-N3**, showing stronger rejection of liquid aerosols. In the field, NextPM’s PM2.5 average was slightly higher (6.4 µg/m³ vs. 4.5 µg/m³), but with **half the variability for PM10** (standard deviation: 5.6 vs. 12.1 µg/m³), and fewer extreme values, highlighting its **resistance to humidity-induced drift** thanks to its built-in heater {% embed url="" %} *** ## AQSPEC 2021 LAB EVALUATION **Title:** Laboratory Evaluation – Tera Sensor NextPM\ **Authors:** South Coast Air Quality Management District (AQMD)\ **Organization:** AQ-SPEC (Air Quality Sensor Performance Evaluation Center)\ **Date:** Report not explicitly dated, but based on tests conducted from September to November 2021 (and published after March 2022)\ **Publisher Location:** South Coast AQMD, Diamond Bar, CA, USA\ **URL (reference page):** **Document Type:** Technical laboratory evaluation report This study was conducted by the AQ-SPEC lab at South Coast AQMD to evaluate the performance of two NextPM sensors under controlled laboratory conditions for PM2.5 and PM10 measurements, following an initial field test phase. Results show outstanding correlation with reference-grade instruments (R² > 0.99), high and stable precision, 100% data recovery, and strong resistance to temperature and humidity changes, although the sensor consistently underestimates absolute concentrations compared to FEM standards. {% embed url="" %} *** ## AQSPEC 2021 FIELD EVALUATION **Title:** Field Evaluation – Tera Sensor NextPM\ **Authors:** South Coast Air Quality Management District (AQMD)\ **Organization:** AQ-SPEC (Air Quality Sensor Performance Evaluation Center)\ **Date:** Based on field deployment between September 29 and November 28, 2021\ **Publisher Location:** South Coast AQMD, Diamond Bar, CA, USA\ **URL (reference page):** **Document Type:** Field evaluation technical report (preliminary results) This field evaluation by the South Coast AQMD assessed the performance of three **NextPM** sensors over two months under real ambient conditions at the Rubidoux monitoring station, in comparison with reference-grade instruments such as the GRIMM and Teledyne T640. The **NextPM sensors demonstrated strong to very strong correlations** for PM1.0 and PM2.5 (R² > 0.95 on 24-hour averages), high data recovery (\~96%), and excellent inter-unit consistency, although they **systematically underestimated concentrations**, especially for PM10 where correlation dropped to moderate levels (R² \~ 0.65). {% embed url="" %} *** ## LCE - In Field Study of NextPM Sensor **Title:** In-Field Study of NextPM Sensor\ **Authors:** H. Wortham, L. Le Berre, R. Berdouni, C. Benadda, C. Chikhi, M. Mezzanotti\ **Reference:** RR-20210112-01\ **Version:** 1.0\ **Date:** February 22, 2021\ **Affiliations:** AtmoSud (Air Quality Monitoring Network, PACA region, France) and Laboratoire de Chimie de l’Environnement (LCE), Aix-Marseille University\ **Location:** Longchamp Super-site, Marseille, France\ **Document Type:** Scientific evaluation report (preliminary phase of a one-year field study) This field study, conducted at the highly instrumented Longchamp "super-site" in Marseille by AtmoSud and Aix-Marseille University, aimed to assess the performance of the **NextPM sensor** under real-world conditions, focusing on its ability to operate accurately across varied humidity, aerosol types, and meteorological scenarios. Results show **very high correlation** with a certified reference instrument for PM1 and PM2.5 (R² up to 0.93 daily, 0.86 hourly), a **100% data recovery rate**, excellent sensor reproducibility, and real-time capabilities nearly matching those of a regulatory analyzer—while **PM10 correlations were lower and more variable**, likely due to particle dynamics and airflow limitations. {% embed url="" %} *** REAL-TIME POLLUTANT IDENTIFICATION THROUGH OPTICAL \ PM MICRO-SENSOR --------------- **Title:** Real-Time Pollutant Identification through Optical PM Micro-Sensor\ **Authors:** Elie Azeraf, Audrey Wagner, Emilie Bialic, Samia Mellah, Ludovic Lelandais\ **Affiliation:** Capgemini Engineering\ **Status:** Preprint on arXiv (arXiv:2503.10724v1)\ **Date:** March 13, 2025\ **Corresponding Author:** \ **Document Type:** Scientific preprint (submitted to arXiv)\ **Link:** This study by Capgemini Engineering explores the use of machine learning (ML) techniques to identify specific pollution sources in real time—such as sand, ash, or candle smoke—using only data from the **NextPS** optical micro-sensor developed by Tera Sensor. By leveraging a novel classification framework based on particle size distribution ratios (rather than mass), the authors demonstrate that even a low-cost OEM sensor can enable pollutant discrimination with up to **82.44% accuracy**, particularly when coupled with **Hidden Markov Chain models**, confirming the **technical robustness of the Tera sensor platform** for embedded AI air quality applications. {% embed url="" %} *** ## Portable Sensors for Dynamic Exposure Assessments in Urban Environments: State of the Science **Title:** Portable Sensors for Dynamic Exposure Assessments in Urban Environments: State of the Science\ **Authors:** Jelle Hofman, Borislav Lazarov, Christophe Stroobants, Evelyne Elst, Inge Smets, Martine Van Poppel\ **Affiliations:** VITO – Environmental Intelligence Unit; Flanders Environmental Agency (VMM), Belgium\ **Journal:** *Sensors*, MDPI (ISSN 1424-8220)\ **Volume & Issue:** Vol. 22, No. 14, Article 5472\ **Date:** July 2022\ **DOI:** **Document Type:** Peer-reviewed scientific journal article As part of a comparative evaluation of 10 portable air quality sensors, the study tested the **PMSCAN device from Tera Sensor**, which integrates the **NextPM** optical sensor, for its ability to measure PM2.5 and PM10 in mobile, urban conditions. The PMSCAN exhibited a **good correlation with reference measurements for PM2.5** (R² up to 0.88), and was praised for its **high measurement resolution and robustness**, although like most sensors in the comparison, it required **post-processing** to correct for environmental influences such as humidity. *** {% embed url="" %} --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://tera-sensor.gitbook.io/tera-sensor/sensors/nextpm/scientific-papers.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.