A significant number of food recall incidents occur globally each year due to potential foreign objects found in food products. How can we quickly and accurately identify the source of contamination and monitor nutritional components? The Bruker M4 TORNADO micro X-ray fluorescence spectrometer (micro-XRF) provides an efficient solution.

Foreign object contamination in food—such as metal fragments, plastic, glass, or insects—can originate from issues in production and processing, pest infestation during storage, or contamination during transportation. This type of contamination not only poses a threat to consumer health but also leads to economic losses and brand reputation risks for companies.

China's "Food Safety Law" and related standards impose strict limits on food contaminants. Existing screening methods often struggle to simultaneously achieve material identification and size measurement, a gap precisely filled by micro-XRF technology.

Bruker M4 TORNADO

The Bruker M4 TORNADO is a micro-beam, large-area elemental imaging spectrometer with the following characteristics:

It uses polycapillary focusing optics to concentrate the excitation beam onto a very small area (<20 µm), achieving excellent spatial resolution for elemental imaging analysis.Different types of samples can be analyzed with simple preparation or even directly without preparation.

This Micro-XRF spectrometer provides a non-destructive, high-throughput quantitative characterization solution for large-area material composition distribution, analyzing samples from multiple angles: single point, multiple points, line scan, and area scan.

Detailed Application Cases

The analyzed objects included: a raw potato for detecting contaminants, freeze-dried sections of several potato varieties for studying nutritional component distribution, and a potato chip for understanding the distribution of seasoning and salt content. All measurements were performed in area scan mode; specific measurement conditions are shown in Table 1.

Table 1: Sample and Measurement Conditions

Analysis of a Contaminated Raw Potato

Figure 1: Contamination analysis on the surface of a raw potato.

A raw potato containing metallic contaminants was analyzed using the M4 TORNADO. After measurement, various software tools were used for analysis. An object region was drawn around the metallic contaminant on the potato surface (Fig. 1a) to obtain a spectrum representative of that specific area. Using a spectral matching tool, spectra similar to the contaminant's spectrum were searched for in a known standard library. This library was compiled from alloy standards measured in point mode under identical conditions. As shown in Fig. 1b, the stainless-steel alloy SS 408 showed a higher match to the spectrum of the surface contaminant in Object 1.

Fig. 1a: Elemental distribution map of a potato with different metal contaminants.

Fig. 1b: Overlay comparison of the spectrum from Object 1 matched against a database of possible alloys.

Nutritional Component Analysis of Potato Slices

For more in-depth research (e.g., in R&D), appropriate sample preparation can significantly enhance the understanding of a product and its properties. Therefore, thin potato slices, cut from potatoes and subsequently freeze-dried, were analyzed. The benefit of the thin-section method is the utilization of high spatial resolution. In bulkier samples, the incident focused beam begins to diverge after passing the focal plane (Fig. 2). Depending on the distance of a particle from the focal plane, objects become increasingly blurry. If a particle lies below the focused surface, the beam broadens again, effectively increasing the object's size.

Figure 2: Schematic of the detection volume and resulting particle size.

The green rectangle represents a thin section, showing resolved particles.

The pink rectangle represents a bulk sample, showing defocused particles.

For prepared thin sections, the source volume is naturally reduced to the cross-section of the beam and the sample. This enables higher spatial resolution, allowing small inhomogeneities down to 20 µm to be resolved. As the system creates a hyperspectral data cube containing signals for all visible elements, the distribution map for each element can be extracted.

Figure 3a shows the multi-element distribution maps of a freeze-dried thin section of a fresh potato. Individual element distributions can be easily extracted. Superimposing these element maps helps understand correlations between elements in the sample and draw conclusions about the sample's chemistry, i.e., which type of salt was used. To visualize the small concentration changes of Cl in the image, a false-color display (Fig. 3b) was used to show the full dynamic range of the chlorine signal intensity. It is evident that Cl is concentrated in the center of the potato slice.

Fig. 3a: Multi-element distribution maps of a freeze-dried slice of fresh potato.

Fig. 3b: Heat map (arbitrary units) of normalized Cl intensity in the potato slice sample, visualizing Cl distribution.

If sample preparation is consistent and measurement conditions are similar, semi-quantitative comparisons of composition are possible. Therefore, integrated spectra from similarly sized objects were selected, and the count rates for each element were compared (Table 2). The count rates reflect compositional variations between different varieties. This can allow inferences about the nutritional value of different varieties. Rh is an artifact from excitation (backscatter), reflecting the density of the detected volume.

Table 2: Elemental count rates (cps) for entire slices of different potato varieties.

Depending on the element analyzed, detection limits can be as low as a few ppm (for elements with atomic number 22 < Z < 42). Using the XMethod software add-on package and appropriate standards, empirical-based quantitative analysis can be easily established.

Analysis of Industrially Produced Potato Chips

As demonstrated with the contaminated potato sample (Fig. 1), the M4 TORNADO is a valuable tool for assessing product quality control. Its advantages include a small spot size, enabling the inspection of inhomogeneities down to the µm-scale, and the acquisition of detailed elemental distribution maps.

An industrially produced potato chip (Fig. 4a) was analyzed to show its elemental distribution. No sample preparation was required. The chip had significant curvature. The highest point (an edge) of the sample was brought into focus for measurement. Although parts of the sample were out of focus, they still provided elemental distribution maps. The height variation also created a quasi-3D effect for the sample, with parts farther from the focal plane providing less signal (inverse square law of distance between excitation, sample, and detector).

Figure 4: Potato chip analysis

a) Mosaic image of the potato chip with the measured area.

b) Na and Cl elemental distribution maps.

c) Na elemental distribution map with defined objects.

d) Spectra corresponding to the objects in c.

The M4 TORNADO offers multiple opportunities for analysis within the food industry. These include elemental distribution as well as the identification and quantification of small inclusions. Micro-XRF technology is non-destructive and requires minimal to no sample preparation.

Analyzing contaminants, like metallic inclusions in a raw potato, allows for positive alloy identification. Differences in nutritional component distribution between several potato varieties can be visualized and their absolute contents compared. Variations in the elemental composition of a potato chip can be easily detected, revealing unevenness in seasoning application.

Authors

Rebecca Novetsky, Micro-XRF Application Scientist, Bruker AXS Inc., Madison, USA

Dr. Max Bigler, Micro-XRF Application Scientist, Bruker Nano GmbH, Berlin, Germany