Präsentation
Filtern
Erscheinungsjahr
- 2017 (4) (entfernen)
Dokumenttyp
- Vortrag (2)
- Posterpräsentation (2)
Sprache
- Englisch (4)
Referierte Publikation
- nein (4)
Schlagworte
- Chemometrics (4) (entfernen)
Eingeladener Vortrag
- nein (2)
Multivariate data analysis is a universal tool for the evaluation of process spectroscopic data. In process analytics, huge amounts of information are produced i) by the many variables contained in one spectrum often exceeding 1000 (wavenumbers, wavelength, shifts…), and ii) the high number of spectra that is generated within the measurement period. Multivariate data analysis, often also called Chemometrics, help to extract the relevant information which is needed to examine and even control processes. Therefore, calibration models must be precise and robust, and moreover, must cover a wide range of variation of factors posing an influence on the process.
In this pre-conference course an introduction to both, explorative data analysis by PCA (Principal Component Analysis), and regression analysis by the most frequently used method, i.e. PLSR (Partial Least Squares Regression) is given. In principal, all optical spectroscopic methods are suited for multivariate evaluation. It will be demonstrated that under certain preconditions, even process NMR spectra can be predicted by PLSR models.
At first, basic principles of multivariate data analysis will be provided. This includes a short introduction into the concept of model building and interpretation of results. Detailed aspects of data pretreatment and calibration & validation strategies for chemometric models will be provided with own data from NIR, Raman and NMR spectroscopy.
In a first example, the development of an online compatible method for the quantification of methanol in biodiesel by PLSR is presented. This also includes the classification of biodiesel feedstocks by PCA and statistical tools which allow for the estimation of a full uncertainty budget.
The Raman spectroscopic prediction of Hydroformylation reaction in a miniplant is used to discuss shortcomings and pitfalls which may occur with the transfer of off-line models to the real processes. Design of experiment and strategies for suitable lab-scale experiments are presented as a possible way to overcome problems.
In a third application the prediction of the reactants of an esterification reaction based on process NMR data is demonstrated.
For reaction monitoring and process control using NMR instruments, in particular, after acquisition of the FID the data needs to be corrected in real-time for common effects using fast interfaces and automated methods. When it comes to NMR data evaluation under industrial process conditions, the shape of signals can change drastically due to nonlinear effects. Additionally, the multiplet structure becomes more dominant because of the comparably low-field strengths which results in overlapping of multiple signals. However, the structural and quantitative information is still present but needs to be extracted by applying predictive models.
We present a range of approaches for the automated spectra analysis moving from statistical approach, (i.e., Partial Least Squares Regression) to physically motivated spectral models (i.e., Indirect Hard Modeling). By using the benefits of traditional qNMR experiments data analysis models can meet the demands of the PAT community (Process Analytical Technology) regarding low calibration effort/calibration free methods, fast adaptions for new reactants, or derivatives and robust automation schemes.
Annually vast amounts of plastics are produced world-wide. However, recycling and waste management is still insufficient resulting in large quantities of plastics being released into the environment. Degradation by sunlight, mechanical and biological factors lead to the breakdown of this waste into little fragments. By convention particles smaller than 5 mm are referred to as microplastics (MP). The occurrence of MP has been reported by researchers virtually all around the globe. Gaining knowledge on MP is currently a time-consuming process because analysis mainly relies on micro-infrared and micro-Raman methods. Prior to that the particles need to undergo purification and enrichment. Thus, only small numbers and volumes of samples can be investigated. Here we tested NIR spectroscopy combined with a multivariate data analysis as a means of speeding up the process of MP analysis.
Experiments were performed using the most abundant polymers polyethylene, polypropylene, polyethylene terephthalate and polystyrene. MP samples were obtained by adding the cryomilled and sieved (<125 µm) particles to approximately 1 g of standard soil at 0,5–10 mass%. Spectra were recorded with a fiber optic reflection probe connected to a FT-NIR spectrometer. 5–10 spectra recorded of each sample were used for the calibration of chemometric models (partial least squares regression, PLSR). “Unknown” test samples were then used to test the model’s capability to predict the type and amount of polymer.
In samples containing 1–5 % of the polymers the prediction yielded the highest degree of agreement with the gravimetric reference values. At low polymer loads some false positive results in the identification were observed. Large amounts of polymers limited the prediction capability by a nonlinear behaviour of the absorption. Further testing was done with real world samples such as compost and washing machine filters. Even though the calibration did not account for these highly complex sample compositions, satisfactory results could be achieved.
An alternative spectroscopic approach for the monitoring of microplastics in environmental samples
(2017)
The increasing pollution of terrestrial and aquatic ecosystems with plastic debris leads to the accumulation of microscopic plastic particles of still unknown fate. To monitor the degree of contamination and to understand the underlying processes of turnover, analytical methods are urgently needed, which help to identify and quantify microplastic (MP). Currently, costly collected and purified materials enriched on filters are investigated both by micro-infrared spectroscopy and micro-Raman. Although yielding precise results, these techniques are time consuming and restricted to sample aliquots in the order of micrograms precluding prompt and representative information on both, larger sample numbers and realistic material volumes. To overcome these problems, here we tested Raman and NIR process-spectroscopic methods in combination with multivariate data analysis.
For this purpose, artificial MP/soil mixture samples consisting of standard soils or sand with defined ratios of MP (0,5 – 10 mass% polymer) from polyethylene, polypropylene, and polystyrene were prepared. MP particles with diameters < 2 mm and < 125 µm were obtained from industrial polymer pellets after cryomilling. Spectra of these mixtures were collected by (i) a process FT-NIR spectrometer equipped with a reflection probe, (ii) by a cw-process Raman spectrometer and (iii) by a time-gated Raman spectrometer using fiber optic probes. For the calibration of chemometric models (partial least squares regression, PLSR) 5 – 10 spectra of defined MP/soil mixtures (consisting of 1 – 4 g material each) were collected. The obtained PLSR models served for the prediction of both, polymer type and content based on the spectra of “unknown” test samples.
Whereas MP could be detected by Raman spectroscopy in coastal sand at 0.5 mass%, in standard soils detection of MP was limited to 10 – 5 mass%. The sensitivity of Raman spectroscopy could be improved by mild treatment with hydrogen peroxide. FT-NIR was suitable for the investigation of MP in standard soils in the range of 5 – 1 mass%, however, here a non-linear effect was observed at higher polymer concentrations. When mixtures of several polymers at low concentration levels were milled together, FT-NIR spectroscopy yielded false positive polymers together with unprecise quantitative information. Recently, the investigation of “real-world” samples shall be tested and compared to the results obtained by micro-FTIR and micro-Raman.