

Summary
The protection of the food supply and the minimization of harmful mycotoxin contamination are closely linked. Recent evidence suggests that the socioeconomic impact of these substances is increasing globally, with significant disparities across regions and populations. Access to reliable, target-oriented detection methods is therefore essential. Portable detection systems enable timely decision-making at critical points in the food supply chain and improve monitoring and risk management. As a result, the demand for portable mycotoxin detection systems is growing. Previous research has shown that mid-infrared (MIR) spectroscopy is suitable for portable mycotoxin detection, but limitations remain due to a lack of measurement certainty and detection capacity. The research in this thesis describes how this limitation can be addressed by introducing innovative methods that integrate sample preparation and enhance MIR spectroscopy, enabling the direct detection of mycotoxins in food at regulatory levels. These innovations focus on the development of a new, 3D-printed device for sample preprocessing and the development of 'selective paper-enhanced infrared spectroscopy' (S-PEIRS). In Chapter 1, the motivation and scientific background of the work in this thesis were introduced, covering the topics of mycotoxin analysis, integrated sample preparation using 3D printing, paper-based microfluidics, and IR spectroscopy.
In Chapter 2, I described the development of an affordable 3D-printed device that fits into an existing method for measuring mycotoxins outside the laboratory. Contaminated wheat kernels were ground with the device, equipped for this purpose with a universal cordless drill, and the particles were internally sieved. Grinding and obtaining sieved wheat particles took only five minutes. In addition, ten first-time users were able to operate the device with minimal instructions. Due to the modular design, extraction of deoxynivalenol (DON) could be carried out directly in the device. After applying the extract to a lateral flow test, DON could be detected, and the full workflow could be completed in 15 minutes. The efficiency of the 3D-printed device was compared with a laboratory mill, and the new workflow showed a strong correlation with laboratory analysis. These developments clearly show the potential of 3D printing for integrated sample preparation outside the laboratory.
In Chapter 3, a new methodological framework was introduced to overcome the limitations of MIR spectroscopy for detection at regulatory levels. In the European Union, the allowed limit for DON in raw wheat is set at a maximum of 1000 µg/kg. At these low concentrations, DON molecules are not detectable with standard MIR spectroscopy instruments, and certainly not in complex matrices such as wheat. The new method, S-PEIRS, involves the following steps: (i) wheat extract purification using simple immunoaffinity columns (IACs), (ii) enrichment of the DON molecules from the purified extract in a paper tip, (iii) paper-based MIR analysis of the enriched solutes in the paper tip and (iv) in-depth data analysis to detect the concentration-dependent DON molecular feature at a specific wavenumber and down to 1000 µg/kg in various types of wheat. S-PEIRS illustrates the potential of paper-based IR spectroscopy for portable food safety analysis, as innovative detection method with molecular certainty.
In Chapter 4, the S-PEIRS framework established in Chapter 3 was advanced to achieve the quantitative detection of another, more strictly regulated mycotoxin; ochratoxin A (OTA). OTA occurs in many food products such as cereals, wines, coffee and spices and the ML set by the EU for wines is only 2 µg/kg, which is 500 times lower than the ML of DON in wheat. OTA molecules exhibit different charge states at varying pH. This characteristic was used to transfer OTA from a large wine volume into a small extract, significantly increasing its concentration. Then, S-PEIRS was adjusted to detect the concentration-dependent OTA feature at a specific wavenumber. Due to further purification of the extract and enrichment of the OTA molecules in the paper tip following the S-PEIRS framework, the quantitative detection of OTA down to 2 µg/kg in wine was achieved. This is the first time that such low concentrations could be measured quantitatively using MIR spectroscopy and broadens the further application of MIR for strictly regulated compounds, also beyond the field of food safety.
In Chapter 5, a general discussion on the key thesis topics of integrated 3D-printed sample preparation and S-PEIRS was provided, including major challenges and achievements, and an outlook for future research with prelimenary results.

















