Grantee Research Project Results
Final Report: A Fieldable, Portable, Reagent-Free Microplastic Sensor Enabling Rapid Readout and Modular Operation
EPA Contract Number: 68HERC22C0008Title: A Fieldable, Portable, Reagent-Free Microplastic Sensor Enabling Rapid Readout and Modular Operation
Investigators: Hemami, Sheila
Small Business: Triple Ring Technologies
EPA Contact: Richards, April
Phase: I
Project Period: December 1, 2021 through May 31, 2022
Project Amount: $99,949
RFA: Small Business Innovation Research (SBIR) Phase I (2022) RFA Text | Recipients Lists
Research Category: Small Business Innovation Research (SBIR) , SBIR - Water
Description:
A comprehensive understanding of microplastic pollution is significantly hampered by the unavailability of low-cost, robust, accurate, and rapid analysis techniques. Appropriate measurement could provide valuable information as to the source of the pollution, enabling mitigating actions. Currently these measurements are impossible outside of the laboratory, and are made tediously and at a high cost. Because of the barriers to accurate measurement, we are largely ignorant of the density and distribution of microplastic particles in the water column and uptake by wildlife and humans, and therefore our ability to manage the risks or take action to mitigate this growing pollutant is severely hindered. There is an urgent need for such measurement information, toward informing the public, policy makers, and toward developing and managing effective mitigation strategies for plastic pollution in the world's water bodies.
The state-of-the-art for microplastic detection consists of tedious lab techniques that require expensive precision instrumentation and chemicals to dissolve biological materials. After water samples have been collected and concentrated via filtering in the field, plastics are separated from biological material via chemical digestion. The remaining microplastics are then analyzed with optical spectroscopic techniques such as Raman or Fourier Transform Infrared Spectroscopy (FTIR). This process spans multiple days, and requires environmentally harmful chemicals. Our efforts address the unmet need for rapid portable microplastic characterization without requiring laboratory instrumentation or harmful chemicals. This would allow real-time feedback in the field for identifying microplastics to permit users to dynamically adjust their sampling methods and locations.
The basic sensing technology is impedance spectroscopy as described in Colson & Michel, ACS Sensors, "Flow-Through Quantification of Microplastics Using Impedance Spectroscopy," January 2021. The original published approach uses impedance spectroscopy to characterize the electrical properties of individual particles directly in a continuous flow of water, allowing rapid categorization of particle material (plastic, biological, bubbles). The bench demonstration utilized controlled synthetic samples (single plastic, single shape) with successful sample characterization for particles > 300 µm in diameter. Figure 1 illustrates the experimental system.
Figure 1: Benchtop impedance measurement system from Colson & Michel, 2021.
The purpose of this SBIR was to de-risk integrating the technology into a fieldable unit by addressing practical considerations. This was addressed through three objectives. The first involved analytical and experimental confirmation of system behavior and performance as a function of the probing frequencies, and the subsequent impact on performance and robustness of the particle categorization algorithm. Data was collected and analyzed with a wide range of frequencies, and multiple particle categorization algorithms were implemented to characterize system accuracy and to ensure that desired performance criteria of various end users could be met. The second objective encompassed initial development of a fluidic pumping mechanism and bubble mitigation techniques. This included prototyping and testing liquid sample delivery methods as well as developing bubble detection algorithms, to avoid bubbles being erroneously counted as plastic particles. The final objective was to assess the system performance on heterogeneous samples (differing in plastic type and also in shape), to identify and develop mitigation strategies for challenges posed by the heterogeneity of real-life microplastics. Different types and form factors of microplastics were acquired and analyzed.
Summary/Accomplishments (Outputs/Outcomes):
Efforts undertaken in this SBIR effort have successfully de-risked the practical considerations outlined above. The system can achieve accurate counts of heterogenous microplastics greater than 300 µm in diameter, while successfully detecting and correcting for bubbles. Counts are accurate to within 10%. Heterogeneity refers to shape (beads, crushed beads, irregular shavings, pieces of fishing line, weathered beads), and plastic type (4 kinds of plastic). Bubble mitigation is entirely software based and is based upon the unique impedance characteristics of bubbles.
This performance improvement results from technical progress on several dimensions. The system was extensively characterized across a range of frequencies and an increased number and variety of particles, resulting in a ground truth dataset that was over 20 times larger than the original experimental data set. The introduction of "blank" measurements also facilitated noise characterization and baseline performance. Improved analog signal processing, encompassing both peak detection and bubble detection, resulted in a more robust data set for the classifier training and ultimately contributed to the better performance. Importantly, preprocessing in software to identify and remove bubbles proved to be significantly more successful than attempting to classify bubbles as a particle type. While we do not believe that software-only bubble mitigation will be sufficient for the ultimate system design, we have demonstrated that physical bubble mitigation need not be perfect, but rather only needs to reduce the bubbles to a level such that they can be removed via software processing.
Example ground truth impedance data sets are shown in Figure 2 for two selected frequencies, where real and imaginary impedance is shown among different particle classes.
Figure 2: Real and imaginary impedance at 2 frequencies for plastic, organic, and inorganic particles, illustrating the data in 4 dimensions. A total of 6 frequencies are used in the classification algorithm (12 dimensions), resulting in repeatable counts of microplastic particles in aqueous solution.
We were fortunate to obtain field-collected microplastic samples from several sources and are currently in the process of analyzing these materials. Figure 3 shows an example of in-field coastal particle samples collected for microplastic characterization.
Figure 3: Northern Massachusetts coastal surface water samples collected for microplastic identification.
Finally, we have created several prototypes for liquid sample introduction and pumping at a controlled rate through the impedance sensor. Efforts are ongoing towards integrating the automated fluidics into a self-contained unit. Requirements of the pumping mechanism include the ability to flow particles of varying buoyancy, achieving a controllable and stable flow rate, and providing bidirectional flow to allow multiple measurements of a single sample.
Conclusions:
The work completed during this phase has shown relatively significant performance for a first prototype of realistic microplastic counting for particles greater than 300 µm in diameter. Steps towards portable unit integration have begun, including identification of candidate pumping mechanisms, preliminary electronics integration, and control software development. Progress is ongoing for these efforts towards an alpha prototype for on-site measurements with real-time feedback with the goal of commercial development in a variety of environmental markets which currently there exists no comparable device. Subsequent phases of funded work will include improvements in the fluidics and pumping system, custom amplifier electronics for embedded control, and particle sorting capabilities, culminating with integration into a portable and fieldable prototype.
Our commercialization objectives for the Phase 1 SBIR included continuous engagement with customers, updating our network of data and hardware customers, and socializing the technology and market opportunity with potential investors and advisors. These objectives were in service of defining a minimum viable product for early adopters, cultivating alpha- and beta-test partners and early adopters, and in preparing for assessing a commercialization pathway (licensing to a larger company, or spinning out a company for ultimate acquisition).
The applications for a fieldable microplastic unit span multiple markets, including government use through the municipal water systems, wastewater treatment systems, and regions involved in stormwater monitoring. Additionally, private and public laboratories use fairly high sophisticated methods for microplastic counting and would benefit from benchtop versions of the unit. The research market serves as an additional customer as the number of publications within microplastics has increased by a factor of 30 in the last 10 years. Finally, personal environmental monitoring, including at-home methods for monitoring drinking water, has increased due to the heightened awareness of microplastic pollution. Based upon our findings in this SBIR effort, we believe that a single core system that can serve these markets can be brought to market within the next 2 years. The next step will be prototype construction and demonstration, aiming for strategic deployment with alpha-test partners.
The perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.