Research Group Inorganic Trace Analysis



» Winter Plasma 2022

Poster Presentation Award

Author: Jakob Willner Email: Position: PhD Student Division: Corrosion Analysis & Prevention KAI Kompetenzzentrum für Automobil- und Industrieelektronik GmbH Webpage: Supervisors: Univ.Prof. Dr. Andreas Limbeck2 Univ.Prof. Dr. Günter Fafilek2 Dr. Silvia Larisegger1 1 KAI 2 TU Wien

© Jakob Willner

Willner et al.: Humidity and temperature dependence of the Sulfur uptake behaviour of different polymers examined via LA-ICP-MS depth profile measurements

» AOFKA 2021


2nd Best Poster Award

Results and Discussion Introduction Experimental Maximilian Weiss +43 1 58801 15183 • Sputtering is a simple way to apply copper to a sample • CuF emmsion higly linear corrrelated to fluorine concentration in standards • Sensitifity high enough for single shot images • Fluorine mappings in good aggrement with reference values • The method provides high discrimation power of PTFE from the background Conclusion Contact PTF E PTF E resin Cu sputter target vacuum chamer e e 120 nm Cu layer Laser CuF CuF CuF molecular emmission embedding sputter coating LIBS measrument Parameter Value Laser energy [mJ] 1.6 Spot size [μm] 100 Frequency [Hz] 10 Grating grooves/mm 300 Stage velocity [mm/s] 2 Spot spacing [μm] 200 Gate delay (ICCD) [μs] 7 Gate width (ICCD) [μs] 10 Figure 2: Single shot spectrum of the copper coated PTFE. The highlighted area is the integrated and background corrected region of the CuF C 1Π -->X 1Σ+ band. Figure 1: Schematics of the measurement process, from embedding of the PTFE tubes, over sputter depostion of copper to LIBS measurement. Figure 3: a Calibration graph of the pressed powder standards b: Microscopic image of a laser crater in Cu coated PTFE; c: Surface profile of the crater from b; d: Broadband LIBS spectrum, annotated regions were used for the PCA. Figure 1: Left: Image of the sample: five PTFE tubes embedded in acrylic resign (Struers Versocit-2), right: Pressed powder standards. Table 1: LIBS measuring parameters a b c d LIBS analysis of Fluorine in solid samples via measurement of molecular emission bands Maximilian Weiss1, Zuzana Gajarska1, Hans Lohninger1, Georg Ramer1, Bernhard Lendl1 and Andreas Limbeck1 1 Institute of Chemical Technologies and Analytics, TU Wien, 1060 Vienna, Austria Figure 4 : Reconstructed chemical maps from emission intensities for the CuF molecule, scale bar = 1 mm. Figure 5: Scores of the first principal component from the broadband six channel detector recorded simultaneously with the molecular CuF band. Figure 7: Sample images based on the Youdens J threshold. The threshold images of both signals resemble the fluorine distribution present in the optical images. The growing importance of fluoropolymers in high-tech applications and green technologies results in a rising need for their characterization. In contrast to conventional methods, laser-induced breakdown spectroscopy (LIBS) provides the advantage of a spatially resolved analysis. Nevertheless, the high excitation energy of fluorine results in low sensitivity of the atomic F(I) lines, which limits the feasibility of its LIBS-based analysis. We present a novel approach, in which a thin film of copper is deposited on top of the sample via sputter coating. In the late-stage LIBS plasma copper atoms recombine with fluorine to the CuF molecule, which strongly emits in the visible range. We show that this method allows a quantitative, as well as spatial resolved assessment of the fluorine content in the sample. To access the applicability of the method, two kinds of samples were produced: For quantitative measurements pressed powder standards were produced from mixtures of PTFE and cellulose powder. To access the imaging capabilities PTFE tubes with an outer diameter of 4 mm and an inner diameter 2 mm were embedded in a Versocit-2 acrylic resin (Struers, Germany). Prior to analysis, the sample surface was polished using a series of silicon carbide (SiC) grinding papers. Copper thin films with a thickness of 120 nm were deposited via a magnetron sputter coater (Baltech MED-020, Liechtenstein) using a copper target. LIBS experiments were performed with a Applied Spectra J200 Tandem LIBS spectrometer equipped with a 266 nm ns Nd:YAG laser and a 6 channel CCD spectrograph covering the spectral range from 188 to 1048 nm. Further an Acton SP2750 spectrometer with a PIMAX2 (ICCD) detector (both Princeton Instruments, USA). All measurements were carried out under an argon gas flow of 1 L/min, the measurement parameters are in Table 1. Surface profiles of craters were recorded with a Dektak XT stylus profilometer (Bruker, USA). Images were reconstructed from raw data using Epina ImageLab 3.45. The data were processed using baseline-corrected integrals. The quantitative analysis of the pressed powder standards was performed in OriginPro 2020, graphs were prepared using the python (v 3.7.6) programming language and the matplotlib (3.2.3), numpy (1.18.1) and scipy (1.4.1) packages. Copper was chosen for this work, as it is a rather noble metal frequently applied for magnetron sputtering. Moreover, the CuF molecule emissions in the visible range are free of interference from atomic copper lines, as can be seen in Fig 1, were the integrated and background corrected area, used for further analysis is highlighted. To assess the sensitivity and linearity of the method, the pressed powder standards made of cellulose and PTFE were evaluated. Here the signal of 60 single shots was accumulated. In Figure 3a, the calibration curve is shown, exhibiting a high linearity (r2=0.99) and giving a LOD of 160 μg/g fluorine, indication a high enough sensitivity for single shot imaging. In Figure 3b the crater of a single laser with 100 μm spot size shot on Cu coated PTFE is shown. Due to the bad adhesion of the film a larger area of Cu is removed, limiting the imaging resolution to 200 μm. With profilometry (Fig. 3c) it was confirmed that the actual crater has a diameter of 100 μm, corresponding to the laser spot size. Using the molecular emission signal a image of the fluorine content could be reconstructed (Fig.4), which resembles the optical image. To access if the defects in the image (marked with arrows) stem from the sample or the measurement, a PCA was performed the co-recorded broadband spectrum (see Brunnbauer et. al. 2020), showing the same artifact. In order to compare separation power of the fluorine from the background, a mask representing the PTFE rings and a mask representing the background region (acrylic resin) was created for both samples using the microscopic image and ImageLab software. The emission signals from Fig. 4 were plotted in a histogram (Fig.5) representing the distribution of the molecular signal intensities in the region of the PTFE tubes (orange) and the background (blue). The dotted line in the histograms represents the PTFE-background threshold determined with Youdens J. To combine the spatially resolved with the statistical information, an image was created (Fig6.). Using this threshold, pixels above the threshold are set to black and below to white. Figure 6: Histograms representing the distribution of the integrated CuF molecular band intensities in the region of PTFE ring (orange) and acrylic resin (blue).

© Maximilian Weiss

Weiss et al.: LIBS analysis of Fluorine in solid samples via measurement of
molecular emission bands

» Winter Plasma 2020

Poster Presentation Award

MATRIX INDEPENDENT QUANTIFICATION OF TRACE ELEMENTS IN POLYMERS USING LASER INDUCED BREAKDOWN SPECTROSCOPY (LIBS) Lukas Brunnbauer1, Silvia Larisegger2, Michael Nelhiebel2 and Andreas Limbeck1 1 TU Wien, Institute of Chemical Technologies and Analytics, Getreidemarkt 9/164-IAC, 1060 Vienna, Austria 2 KAI Kompetenzzentrum Automobil- und Industrieelektronik GmbH, Technologiepark Villach, Europastraße 8, 9524 Villach, Austria Polymers are widely used in different industrial applications such as food packaging, construction materials and as insulating layers in electronic devices. In all these fields, the trace metal composition of polymers is gaining more and more interest as trace metals heavily influence the required specifications of the polymers. Conventional determination of the trace metal content of polymers includes complete digestion of the sample. As polymers are usually rather hard to dissolve, the procedure often requires a mixture of various strong acids and oxidants. Subsequently, the trace metal content of the solution is determined using e.g. liquid ICP-MS or ICP-OES measurement. As this approach reveals only the bulk trace metal content and is very laborious and error prone, direct solid-sampling methods are of great interest to determine the trace metal content in polymers. In this work, we demonstrate the capabilities of laser induced breakdown spectroscopy (LIBS) combined with multivariate data evaluation strategies for matrix independent quantification of trace metals in polymers. • Different polymers show different slopes in their univariate calibrations • Multivariate calibration approach eliminates matrix effects when doing quantitative polymer analysis • Application of the model to quantify polymers that are not part of the model The author gratefully acknowledges the financial support funded by the Austrian Research Promotion Agency (FFG, Project No. 863947). Lukas Brunnbauer Introduction Experimental Results Conclusion and Outlook Contact Acknowledgements 0 50 100 150 200 0.00 0.05 0.10 0.15 0.20 0.25 0.30 PAN (R2 = 0.90) PMMA (R2 = 0.97) PSO (R2 = 0.96) PVA (R2 = 0.97) PVC (R2 = 0.96) PVP (R2 = 0.96) Intensity Li @ 670 nm (a.u.) Concentration Li (ppm) 1. Polymers (PAN, PMMA, PSO, PVA, PVC, PVP) obtained from Acros Organics (Geel, Belgium) dissolved in NMP are spiked with NMP solutions containing Li, Na, K, Ca 2. Spiked polymer solutions are applied to a high purity silicon wafer (10 x 10 mm) 3. Samples are cured at 120 °C for 3 h to remove NMP from the samples LIBS Analysis: LIBS J200 Applied Spectra (Fremont, CA) Laser Energy (mJ) 2.88 Spot size (μm) 100 Repetition rate (Hz) 10 Gate delay (μs) 0.5 Accumulated spectra 35 Measurement spots/sample 6 Atmosphere Ar 200 400 600 800 1000 0E+00 1E+05 2E+05 3E+05 4E+05 5E+05 Intensity (a.u.) Wavelength (nm) PAN PMMA PSO PVA PVC PVP C(I) 193.09 nm C(I) 247.88 nm CN violet band C2 swan band Ca(I) 393.00 nm Na (I) 589.00 nm H(I) 656.28 nm Li(I) 670.78 nm O(I) 777.42 nm K(I) 766.49 nm Figure 2: Representative LIBS spectra for the 6 investigated polymers with marked emission signals used for data evaluation Table 1: LIBS measurement parameters • 6 line scans (35 shots) with a total length of 6 mm were analyzed for each polymer standard • Recorded spectra of each line scan are accumulated • Average spectrum of 6 line scans is evaluated • Emission signals marked in Figure 2 are used for data evaluation Figure 1: Schematic sample preparation procedure Polymer Standards Preparation: Univariate Calibration Li: Figure 3: Univariate calibrations for Li of the 6 investigated polymers • Different slopes of the calibrations are caused by matrix effects (Figure 3) Multivariate Calibration: • Multivariate calibration model (PLS) was build using analyte emission signals and polymer specific emission signals • Cross validation was calculated (Figure 4/5) with a test size of 6 (10% of the data) and 5 repetitions to determine the optimal number of factors for PLS and the RMSEP • Obtained model allows quantification of polymer types that are part of the model Figure 5: Actual concentration vs estimated concentration of the PLS model for Li Figure 4: Actual concentration vs estimated concentration of the PLS model for K 0 50 100 150 200 250 -50 0 50 100 150 200 250 PAN PMMA PSO PVA PVC PVP Estimated concentration K (ppm) Actual concentration K (ppm) R2 = 0.9285 RMSEP = 19.9 ppm 0 50 100 150 200 0 100 200 300 PAN PMMA PSO PVA PVC PVP Estimated concentration Li (ppm) Actual concentration Li (ppm) R2 = 0.9223 RMSEP = 19.7 ppm Li K

© Lukas Brunnbauer

Brunnbauer et al.: Matrix independent quantification of trace elements in polymers using laser-induced breakdown spectroscopy

Oral presentations

» CSI XLII 2022