Theoretische Methoden zur Quantifizierung und Klassifizierung von Mikroplastik
Loading...
Date
relationships.isAuthorOf
Publisher
item.page.supervisor-of-thesis
Abstract
Although research on microplastics in rivers began more than ten years ago in 2014, standardized sampling strategies are still missing, leading to poor comparability and reproducibility of study results. Two important aspects related to this are addressed in this thesis: First, when reviewing several study results a recurring error became apparent that causes the reported measurement uncertainties to underestimate the true uncertainties. This problem arises because important sources of measurement uncertainty have not been taken into account. Second, it has been shown that the distribution of microplastics in rivers varies strongly both spatially and temporally; nevertheless, the design of most measurement studies does not adequately reflect this complex microplastic distribution in rivers. To address the first problem, this work defines and discusses two different sources of measurement uncertainty (intrinsic and extrinsic) that affect microplastic measurements. It is demonstrated that both have a strong influence on the measured microplastic distribution in water and sediment. Consequently, both types must be considered in order to determine an accurate estimate of the true measurement uncertainty. The second problem, concerning the complex microplastic distribution in rivers, is then qualitatively analysed using two complementary models. It is shown that microplastic distribution can fluctuate markedly in all three spatial directions. A comparison of microplastic particle numbers in the water column and sediment revealed that particle numbers in sediment depend heavily on the sedimentation‑ and erosion‑rate coefficients. Therefore, particle numbers at two sediment sampling sites can already differ substantially - even in the absence of external microplastic sources or sinks - because the local environmental conditions that govern the sedimentation and erosion of microplastics are different. Consequently, differences in sediment particle numbers provide only limited or no information about the presence of external sources of microplastics. Moreover, it was shown that particle numbers in water and sediment are coupled when the river is in a transient state. Based on these results, the following recommendations for measuring microplastics in rivers are derived: (1) Multiple samples should be taken along all three spatial axes and analysed independently in the laboratory in order to estimate measurement uncertainties accurately. (2) Samples should be collected from both water and sediment. (3) Differences in sediment particle numbers should not be used as evidence for the existence of microplastic sources or sinks.
The second part of this thesis deals with the classification of the polymer type of microplastics, because this information can help identify important microplastic sources. Typically, spectroscopic methods such as Raman or FTIR spectroscopy are used to determine polymer type, although other spectroscopic techniques are also possible. A potential alternative is Laser‑Induced Breakdown Spectroscopy (LIBS), especially because of its speed, potential portability for field measurements, and the possibility to analyse samples without extensive pretreatment. Classifying polymer type based on LIBS spectra, however, poses a particular challenge: on the one hand, spectra of different polymer types look overall very similar, and on the other hand the data are extremely high‑dimensional. Other research groups that faced the same challenges have successfully used the amplitudes of spectral lines of a few manually selected chemical species as features for classification. This suggests that the amplitudes of signal peaks in a spectrum constitute important classification features. Nevertheless, manual selection of a few peaks is inefficient because a large portion of the information contained in the rest of the spectrum—information that may also be important for classification—is discarded. Therefore, in the second part of this thesis a method was developed that can help identify signal peaks automatically in spectra. Unlike other approaches that often rely on amplitude‑based procedures to detect peaks, this method adopts an order or rank‑based approach derived from permutation‑entropy analysis of time series. The approach can be applied directly to raw spectra, enables amplitude‑independent comparison of multiple spectra, and allows the development of a statistical test to separate signal from noise. Initial tests of the method on real LIBS microplastic spectra yielded very promising results: almost all signal peaks, regardless of their amplitude or spacing, were identified.
Review
Metadata
Contributors
Supervisor:
Dates
Issued: 2025-11-28
Faculty
FB13:Physik
Language
de
Keywords
MikroplastikMicroplasticPeak DetectionPeak DetektionMicroplastics in riversMikroplastik in Flüssenplastic classificationKlassifikation der PlastikartLIBSsediment samplingplastic identification
DFG-subjects
3.22-01 - Statistische Physik, Nichtlineare Dynamik, Komplexe Systeme, Weiche und fluide Materie, Biologische Physik
show more
Fuhrmann, Joshua: Theoretische Methoden zur Quantifizierung und Klassifizierung von Mikroplastik. : 2025-11-28.
License
Except where otherwised noted, this item's license is described as Attribution-NonCommercial-ShareAlike 4.0 International