Rapid detection of microplastics in food and environment using hyperspectral imaging technology.pptx

Rapid detection of microplastics in food and environment using hyperspectral imaging technology.pptx

1. Rapid detection of microplastics in food and environment using hyperspectral imaging technology ———— Jiajia Shan ———— Dalian University of Technology School of Ocean Science & Technology

2. Introduction Microplastics pollution Microplastics in different food Methods of microplastic detection Hyperspectral imaging spectroscopy (HSI) technology Main work on microplastics detection based on HSI Detection of Microplastics in soil Detection of Microplastics in seawater Detection of Microplastics in sea fish CONTENTS 2

3. Introduction——Microplastics pollution 3

4. Introduction ——Microplastics pollution 4 Blood Heather, et al. Environment International , 2022,163:107199 Lung Lauren, et al. Science of the Total Environment, 2022,831:154907 Placenta Braun, et al. Pharmaceutics ， 2021,13 （ 7 ） :921 Breast milk

5. Introduction ——Microplastics pollution 5 Sridhar, A et al., Chemosphere, 2022 (286), 131653

6. Mamun, AA et al., Science of The Total Environment, 2023(858), 159834  

7. Introduction ——Microplastics in different food 7 What kinds of food are contaminated by microplastic??

8. Introduction——Microplastics in aquatic organism 8 Jin , MK et al. Journal of food science,2021,7. Aquatic organism

9. Introduction——Microplastics in different food 9 Jin , MK et al. Journal of food science,2021,7.

10. Introduction——Microplastics in different food 10 Particle size distribution MPs particles ranging from 20 μm to 5 mm. MPs smaller than 1000 μm accounted for more than two-thirds of the total number of MPs in fish samples. The proportion of MPs with a size of less than 250 μm was 17% to 79% of all MPs. Compositions ： polyethylene terephthalate (PET), polypropylene (PP) and polyethylene (PE), polyvinyl chloride (PVC), polymethyl methacrylate (PMMA).

11. Introduction——Microplastics in different food 11 Fish Bivalves Source Commonly, MPs in aquatic organisms originate from living habitats, fishing line/net, domestic washing, feeding habits and packaging. Among them, the main source of MPs in aquatic organisms is water bodies (living habitats), where MPs mainly originate from domestic sewage and fisheries .

12. Introduction——Microplastics in different food 12 Figure (a) Abundance of MPs found in the table salt; X axis presents salt brand numbers of Sea salt and Well salt. (b) Abundance of MP detected in sea salt from different regions. KR: Korea; CN: China; CT: Chinese Taipei; TH: Thailand; IN: India; VN: Vietnam; FR: France; IT: Italy; BG: Bulgaria; US: the United States; SN: Senegal. S n : Sample number. Salt

13. Introduction——Microplastics in different food 13 The main source of MPs is the original water . For rock salt and well salt, they are usually collected underground where it is difficult for synthetic polymers to penetrate, so the transportation process to the surface is the main way to introduce MPs. MPs are also introduced in the product processing such as MPs pollution from the air during the evaporation process and packaging . Sources

14. Introduction——Microplastics in different food 14 Tap water: Researchers have investigated tap water from 14 countries, and anthropogenic debris was found from 81% of the 159 tested samples. The range of MPs particles was from 0 to 61 particles/L, with an average mean of 5.45 particles/L. Kosuth , M.; Mason, S. A.; Wattenberg, E. V., Anthropogenic contamination of tap water, beer, and sea salt. Plos One 2018, 13 (4). Tap water

15. Introduction——Microplastics sources in different food 15 Tap water :raw water and the process of purification and transportation. The MPs in raw water mainly come from the degradation of various plastic products and the discharge of domestic water. the storage tanks in drinking water treatment plant (DWTPs) are coated with epoxy resin to prevent corrosion. The conveying pipes are mostly made of PVC or PE, and the pipe fittings are made of PA . Plastic wear during storage and transportation is also one of the sources of MPs. Kosuth , et al. Plos One 2018, 13 (4). Source

16. Introduction——Microplastics distribution in different food 16 Schymanski , et al. Water Research 2018, 129 , 154-162. Jin , MK et al. Journal of food science,2021,7. Bottled water 11 Globally-sourced brands of bottled water purchased from 19 locations in nine countries have been tested for the detection of MPs.

17. Introduction——Microplastics in different food 17 Beverage MPs distribution in different beverages Particle size distribution of MPs Jin , MK et al. Journal of food science,2021,7.

18. Introduction——Microplastics in different food 18 Fresh water. The packaging materials for beverages are mainly plastic, which can also cause microplastic pollution during the packaging process or improper cleaning In addition to the above factors, the materials and equipment used in the production process (mainly filters) can also cause pollution. Sources

19. Introduction——Microplastics in different food 19 Hernandez, et al. immersed empty plastic tea bags in reverse osmosis water for 5 minutes at 95 °C. Analysis showed that approximately 11.6 billion MPs and 3.1 billion NPs were detected into a single cup of the beverage. Du, et al. collected four types of take-out containers (polypropylene, PP; polystyrene, PS; polyethylene, PE; polyethylene terephthalate, PET) from five cities in China. The results showed that MPs were found in all take-out containers with abundance from 3 to 29 items/container . The highest value occurred in PS containers with rough surface. Hernandez , et al ., Environ Sci Technol 2019, 53 (21), 12300-12310. Du, et al. , Journal of hazardous materials 2020, 399 , 122969-122969. Package food

20. Introduction——Microplastics sources in different food 20 Food trays that are often made from extruded polystyrene (XPS). Kedzierski , et al. purchased four different brands meat products (chicken) packed in extruded polystyrene trays and analyzed. 4.0 to 18.7 items/kg were found from packaged meat. The second source of MPs may be produced when opening the package . Sobhani , et al. investigated the MPs generated during the opening of plastic packaging. A shopping bag, packaging film, plastic bottle, glove and packaging foam were opened by scissoring with scissors and tearing with hands. In these processes, 0.46-250 MPs/cm could be generated, depending on various conditions such as stiffness, thickness, density of plastic materials, and size of MPs. Kedzierski , et al. Food Packaging and Shelf Life 2020, 24. Sobhani, et al . Scientific reports 2020, 10 (1), 4841-4841. Karami, et al . Science of the Total Environment 2018, 612 , 1380-1386. Source

21. Introduction——Microplastics in other food 21 Karami , et al. investigated the potential existence of MPs in 20 brands of canned sardines and sprats . The experimental results demonstrated that there were plastic particles found from 4/16 brands with 1-3 particles. Liebezeit and Liebezeit also analyzed five commercial sugars , and found that transparent and colored fibers (average 217±123/kg of sugar) and fragments (32±7/kg of sugar) in refined sugar, with 560 fibers and 540 fragments in unrefined cane sugar per kilogram of honey.

22. Introduction——Microplastics in other food 22 Conti, et al. showed that besides the very high vascularization of the fruit pulp , the greater size and complexity of the root system and age of the trees were also the reason resulting in more MPs in fruits than vegetables . Conti, et al. Environmental Research 2020, 109677.

23. Introduction——Microplastics in other food 23 Liebezeit et al. (2013) detected 19 honey samples from 5 countries, and found that fibers and fragments are already present in the bees’ feed and are transferred from the blossoms to the hive by the insects. Liebezeit , et al. Food Addit Contam Part A Chem Anal Control Expo Risk Assess, 30(12), 2136-2140.

24. Introduction——Methods of microplastic detection 24 Yu, CX et al., Analytical and Bioanalytical Chemistry, 2022 (414)volume 414: 4591–4612

25. Introduction——Methods of microplastic detection 25 Flotation is the commonly used method to separate MPs from the liquid phase using the density difference between target components and impurities. The main reagents used are saturated NaCl 、 NaI and ZnCl 2 . Saturated NaCl solution (density ≈ 1.2 g·cm −3 ) is cheap and non-toxic, mainly used to extract low-density microplastics, such as PP 、 PS 、 PE, which is suitable for the extraction of MPs in water and table salt. Sample Preparation

26. Introduction——Methods of microplastic detection 26 Sample Preparation Evaluation of MPs in tissues of aquatic organisms, digestion is the necessary step to remove large amounts of organic impurities from the solid samples. Digestion procedures include acid digestion, alkaline digestion and enzyme digestion. For biotic samples, 30% hydrogen peroxide (H 2 O 2 ) and 65% nitric acid (HNO 3 ) have been frequently used to digest the organic interference. Besides, nitric acid (HNO 3 ), perchloric acid (HClO 4 ), hydro-chloric acid (HCl) or a mixture of the above are used to digest the biological tissues.

27. Introduction——Methods of microplastic detection 27 Sample Preparation Filtration is an effective approach commonly used to separate MPs from liquid phase (water, beverage, milk and beer), or supernatant solutions obtained from density separation using a filter membrane with a smaller pore size for separation. Filter medium including glass fiber, nitrocellulose filter, polycarbonate membranes are used to allow only liquid to pass through. The pore size of filters varies from 0.45 to 20 μm (Crawford & Quinn, 2017). For beverage and package food, filtration is the most commonly used pretreatment method, often carries out together with washing.

28. Introduction——Methods of microplastic detection 28 Sample Preparation Dissolution and coagulation/flocculation are also pretreatment method. Oßmann et al. (2018) added an ethylene diamine tetraacetic acid tetrasodium salt (EDTA) solution when processing the sample, and dissolved the particles composed of calcium carbonate or magnesium carbonate by complexing calcium ions and magnesium ions with EDTA

29. Introduction——Methods of microplastic detection 29 Detection and Characterization Fourier Transform Infrared spectroscopy (FT-IR) and Raman spectroscopy are commonly used to identify chemical compositions of the suspicious MPs. Infrared spectrum (IR) of the measured MPs presents characteristic peaks which correspond to specific chemical bonds. Other potential methods for MPs identification, such as pyrolysis gas chromatography/mass spectrometry (PY-GC/MS), scanning electron microscope-energy dispersive X-ray (SEM-EDAX) and environmental scanning electron microscopy-energy dispersive spectroscopy (ESEM-EDS), have also been investigated and shown to be highly sensitive and precise and able to detect nanoscale MPs.

30. Introduction——Microplastic detection 30 Yang, L et al., Science of the Total Environment, 780:146546

31. Introduction——Hyperspectral imaging technology 31 Hyperspectral technology contains both of spectral and spatial information, which means every pixel in hyperspectral imaging is associated with a range of wavelength. Siche et al. Food Eng Rev (2016) 8: 306-322

32. Introduction——Hyperspectral imaging technology 32 Chemical compounds Shan et al. Food analytical methods 2018, 11(6): 1701-1710. Application of HSI Technology

33. Introduction——Hyperspectral imaging technology 33 Application of HSI Technology Wei et al. Journal of Food Engineering 2019, 248: 89-96. Feng et al. Science of The Total Environment 2019, 659: 1021-1031.

34. 34 Part I Detection of Microplastics in soil

35. Detection of Microplastics in soil 35 Separation based on density Sampling Sieving Microplastics collected Density separation (saturated sodium chloride, NaCl) was used to separate microplastics from soil based on densities difference. The mixture was stirred for 0.5 h and stranded 12 h until the upper suspension clear. Then the extracts floating on the surface of the saturated NaCl solution were picked out visually using a stainless steel needle.

36. Detection of Microplastics in soil 36 (a) (b) After separation, microscopy and Laser Confocal Micro-Raman Spectrometer were used to observe the morphology and identify the chemicals of microplastics, individually. Traditional spectroscopy analysis How can we detect / identify all the MPs or MPs suspicious at the same time, saving laborious?

37. Detection of Microplastics in soil 37 (a) (b) Experimental devices Hyperspectral imaging system Wavelength ： 400nm-1000nm Sample Sample

38. Detection of Microplastics in soil 38 Original hyperspectral imaging Region of interest Spectral information Fresh leaves could be easily identified from other components through spectral information, due to chlorophyII . Compared to fresh leaves, the difference among other components was not obvious. Difficult to identify different substances by naked eyes, according to spectral information.

39. Detection of Microplastics in soil 39 Supervised classification methods, such as mahalanobis distance (MD), maximum likelihood (ML) and support vector machine (SVM) were used to identify microplastics among the other environmental surroundings from the hyperspectral images.

40. Detection of Microplastics in soil 40 Table. The classification results for soils samples with microplastics (1-5 mm) by using ML, MD and SVM algorithms Classification results of PE of particle size 1-5 mm through hyperspectral images Figure. The visual classification results of ML, MD and SVM algorithms

41. Detection of Microplastics in soil 41 Figure. Hyperspectral images of eight types of MPs with particle sizes from 0.1- 3 mm (A1-A11); NIR spectroscopy of PS with particle sizes from 0.1-3 mm (B) Spectral features of MPs became very weak as the particle size decreases, increasing the difficulty in detection. Taking PS as an example, the evident spectral features were noted from 900-955 nm, 1100-1160 nm, 1200-1260 nm, and 1650-1700 nm. Spectral features of PS became weaker and weaker when particle sizes decreased from 3 mm to 0.1 mm. No obvious peak was observed when particle sizes decreased to 0.2 mm. The LOD might be limited by the spatial resolution of the system.

42. Detection of Microplastics in soil 42 It was difficult to classify microplastics from other components including rocks, wilted leaves, fresh leaves, and branches separately when particle size of microplastics was < 1 mm . Because the background noise with the similar size as microplastics (<1 mm) would have a serious effect on the classification. Morphology processing methods – erosion and dilation, were used to improve the original images , to strengthen the information of white and black microplastics, and to eliminate the background noise in hyperspectral images.

43. Detection of Microplastics in soil 43 Figure. ROIs of white (A) and black (B) microplastics and SVM classification results for white microplastic on original hyperspectral (A1) and black microplastic on original hyperspectral (B1); classification results for white microplastic on dilation processed hyperspectral image (A2) and black microplastic on erosion processed hyperspectral image (B2) Table. The SVM classifications of microplastics (0.5-1 mm) obtained from original and preprocessed hyperspectral images

44. Detection of Microplastics in soil 44 For particle size from 1-5 mm, the six kinds of microplastics revealed the satisfactory classification results, with the average P from 79% to 100% and R from 90% to 100%. For particle size from 0.5-1 mm, the classification results for the six microplastics were obtained, with the average P from 86% to 99% and R from 79% to 85%. Six kinds of household polymers with different colors and chemicals were used to validate the method.

45. Detection of Microplastics in soil 45 1-5mm 0.5-1mm Figure. ROIs of environmental surroundings and microplastics particle 1-5 mm (A1) and 0.5-1 mm (B1); and SVM classification results for 1-5 mm microplastics (A2) and 0.5-1 mm microplastics (B2) on the hyperspectral image

46. Jiajia shan , Junbo Zhao, Lifen Liu, Yituo Zhang, Xue Wang, Fengchang Wu. A Novel Way to Rapidly Monitor Microplastic in soil by Hyperspectral Imaging Technology and Chemometrics [J]. Environmental Pollution , 2018, 238: 121-129.

47. 47 Part Ⅱ Detection of Microplastics in seawater

48. Detections of Microplastics in seawater 48 Wavelength range ： 900nm-1700nm Polyethylene (PE), polystyrene (PS) and polypropylene (PP) are the most common polymers encountered in marine environments. The contaminated sea water with microplastics was simulated by adding the above three types. As we all known, water has a strong absorption in the NIR region. In order to eliminate the water effect on the collected spectra, the above contaminated sea water samples with microplastics were simply filtered through a qualitative filter paper ( ф 9cm, 30 ～ 50 μ m pore size) MPs contaminations in marine Aim to monitor MPs contamination by hyperspectral imaging system?? Simulated experiment

49. Detections of Microplastics in seawater 49 Microplastics presented the similar spectral peak as background (majority sea water) with the broad peaks around 1050 nm and 1250 nm, indicating microplastics features were overlapped by background. There was no obvious peak found from the background spectrum in blue because background (mainly water) domination was strongly eliminated.

50. Detections of Microplastics in seawater 50 Sea water samples Filtration samples The results suggest that the performance of the described method is significantly improved by filtering away water before hyperspectral images acquirement. With the aid of spectral and spatial information, microplastics can be separated and identified from the images, instead of real experiment extraction procedures including ingestion and filtration.

51. Detections of Microplastics in seawater 51 The SVM detection models for the eight types of microplastics were established and P and R values for each type of microplastics were calculated respectively. Increasing the polymers types

52. Detections of Microplastics in seawater 52 A high recovery rate of the tested types of microplastics could be identified when the particle sizes were larger than 1 mm. P and R values dropped sharply when the particle sizes were smaller than 0.2 mm, indicating the described method was highly robust for more types of microplastics detection.

53. Detections of Microplastics in seawater 53 Five household microplastics were used to validate the classification model. The results showed that household microplastics could be detected accurately, where average P values could reach 97% . Phytoplankton, Marine organics, Household microplastic

54. Simple and Rapid Detection of Microplastics in Seawater using Hyperspectral Imaging Technology [J]. Jiajia shan ; Junbo Zhao; Yituo Zhang; Lifen Liu; Fengchang Wu; Xue Wang. Analytica Chimica Acta , 2019, 1050: 161-168.

55. 55 Part Ⅲ Detection of Microplastics in seafish

56. Detection of Microplastics in seafish 56 8 of Sea Bass ( Lateolabrax japonicas ), 9 of Redeye Mullets ( Liza haematocheila ) and 3 of Yellow Goosefishes ( Lophius litulon ) were caught from the Yellow Sea near Dalian, China in November 2018. The intestinal tracts were dissected and the intestinal tracts content was weighted and spread on the Teflon substrate for the HSI measurement. Besides, 20 field fish samples were spiked with filed polymers to validate the recovery of the proposed method. One or two field polymers (identified by Raman spectroscopy) were randomly spiked into the ITC to test the recovery of field polymers in the field fish.

57. Detection of Microplastics in seafish 57 It was difficult to observe MPs (especially particles < 0.4 mm) from the ITC, and it was impossible to identify MPs type by naked eyes. HSI obtained from the original and dried MPs-contaminated.

58. Detection of Microplastics in seafish 58 Problem ： interferences from non-plastic substances such as fish bones, fish scales, shell and so on Fish scale suspicious Fish bone suspicious Shell suspicious Shell suspicious parasite bone

59. Detection of Microplastics in seafish 59 To improve the performance of the SVM model for field samples identification, spectral variables of the suspicious particles such as scales, bones and shells were added to extend the model library as much as possible for the model development. The efficiency of the updated SVM model was significantly improved for characterizing and quantifying MPs in the ITC of field fish, with few non-polymers falsely identified as polymers.

60. Detection of Microplastics in seafish 60 In total, 5 plastic items were found in 4 of the 20 (20%) investigated fishes. Considering the three species, 2 out of 3 (66.7%) analyzed Yellow Goosefishes ( Lophius litulon ) presented plastic debris in the ITC, with a total of 2 particles (one PP fragment and one PE fiber). Regarding Sea Bass ( Lateolabrax japonicas ), one PE fiber was found from the 8 individual (12.5%). And 2 plastic particles (PP and PE fragment) were found from the 1 out of 9 Redeye Mullets ( Liza haematocheila ) individual (11.1%)

61. Yituo Zhang, Xue Wang, Jiajia Shana , Junbo Zhao, Wei Zhang, Lifen Liu, Fengchang Wu. Hyperspectral Imaging-Based Method for Rapid Detection of Microplastics in the Intestinal Tracts of Fish [J]. Environmental Science & Technology, 2019, 53: 5151-5158.

62. Conclusions 62 Advantages: HSI technology may facilitate the MPs detection by eliminating the complicated preprocessing steps, which are simple sample preparation, high efficiency, time and labor saving. Limitations: High spatial resolution hyperspectral system, such as Micro-HSI system, should be considered for small MPs identification. A huge spectral database containing as much polymers (types, aged) existing in the environment as possible should be built.

63. Thank you for your attention and suggestion