Going 3D: micro-XRF Scanning of Highly-Topographic Cultural Heritage Objects

Unlike paintings, most Cultural Heritage objects do not always possess a plane surface, and even paintings themselves are often not flat. Highly topographic objects range from metal or wooden sculptures to ceramics, glass, stone objects, and various forms of contemporary art. So far, topography is usually considered the main limiting factor of imaging analysis with a high spatial resolution. Recent approaches to overcoming this challenge include constructing equipment that can follow the surface topology [1, 2] and, for open beam systems measuring directly in the air, the post-processing of data to correct working distance variations using the Ar-signal [3]. Following the surface topography is impractical for intricate or highly irregular objects, where the spectrometer cannot always be positioned in an ideal perpendicular geometry to the surface. Even correction based on the Ar signal has its limitations.

We are adopting a different approach. In this case, a double 60 mm² SDD M6 JETSTREAM system is used. This instrument, equipped with a polycapillary lens, maintains good spatial resolution due to its low divergence design. The polycapillary lens irradiates the sample with a perpendicular beam (Fig. 1a), typically with a smallest spot size of approximately 30-50 µm (Fig. 1b). Due to the beam's low divergence, the minimum spot size is achieved at ~3 mm closer to the sample surface than the first calibrated position at a 100 µm spot size.

Generally, the M6 JETSTREAM has five calibrated spot size positions based on increased working distances and the corresponding spot divergence of the polycapillary lens. It is important to note that a longer distance results in a loss of intensity (Fig. 1b).

Fig. 1: Measurement set-up of the M6 JETSTREAM.

The M6 JETSTREAM's patented Aperture Management System (AMS) further reduces the divergence of the X-ray beam. The AMS is a specially designed aperture that can be positioned automatically between the X-ray source and the polycapillary lens entrance. The selected aperture narrows the diameter of the primary radiation entering the lens, excluding the outer capillaries from the transmission process (Fig. 2a). This approach significantly reduces beam divergence. As shown in Figure 2, the standard divergence of the M6 JETSTREAM lens is ~30 µm/mm; when using the smallest aperture, this is reduced to 10 µm/mm.

The AMS comes in two variants: the AMS 1000 and the AMS 500, which differ in the diameter of the opening shaping the primary beam before it enters the lens. Effectively, at a distance of 10 cm, the spot size would be around 4 mm wide, but the AMS 1000 and AMS 500 can reduce this to 2.5 mm and 1.2 mm, respectively (Fig. 2b). However, this reduction in divergence comes with a trade-off: the AMS reduces intensity by approximately 1/3 for the AMS 1000 and 1/7 for the AMS 500. Therefore, the AMS is only available for double 60 mm² SDD systems to help compensate for this loss in intensity.

Fig. 2a: Schematic of the patented Aperture Management System (AMS).

Fig. 2b: The effect of the AMS on beam divergence.

The AMS in Practice: micro-XRF Analysis of 3D Cultural Heritage Objects

How does this now apply in practice? Measuring an object like the polychrome painted wooden sculpture of Jesus depicted below, the working distance of the M6 JETSTREAM is limited by the highest position of the sculpture, in this case the sculpture’s knee . This means, that the hands of the small sculpture are 5 cm out of focal plane.

Micro-XRF elemental distribution images of a M6 JETSTREAM scan without the AMS reveals the presence of lead white (Pb) and vermilion (Hg) in areas below the focal plane, but longer measurement times and the AMS 500 enables the visualization of cracks in the paint on the hand as well as the fine brushwork of the vermilion-red paint above the lead white flesh tone.

Polychrome Sculpture of Jesus

Fig. 3a: Measuring three dimensional objects like this painted wooden sculpture of Christ on a Cross from a private collection, the beam divergence across the different planes of the sculpture significantly widens in areas below the focal point, here being the sculpture’s knee.

Fig. 3b: Micro-XRF mapping of the sculpture without (left) and with (right) the AMS 500. Lower intenisties are accounted for by increased pixel time. The elemental distribution of Pb, Fe, Cu and Hg maintains a high resolution when measuring with the AMS.

Fig. 3c: Even at 5 cm out of the focal plane, the elemental distribution of Pb and Hg is highly resolved.

The same principles apply to all types of materials which either have a strongly diverging surface topography or cannot be aligned parallel to the measuring head. Even very fine structures and trace amounts can be visualized to a better degree.

The following cases reveal the power of the AMS feature:

Iron Gall Ink Manuscript

Fig. 4a: The AMS feature applied to a fragment of a manuscript written with iron gall ink. The sample is intentionally misaligned: While the left-hand side is in focus, the uttermost right is about 2 cm away from the focal plane.

Fig. 4b: Due to the intended misalignment, the Ca signals on the right-hand side can hardly be assigned to the ink writing.

Fig. 4c: With the AMS and longer measurement times, the single letters are clearly visible in the Ca distribution.

Fig 4d: The same applies when looking at the Fe distribution. The right side shows both a lower intensity and resolution…

Fig. 4e: …while the writing is far higher resolved when scanning the fragment with the AMS.

Glass Beads

Fig. 5a: Two Millefiori glass beads with a diameter of 10 mm with a typical pattern decoration. The different glass colors are inter alia created by the addition of different chemical elements to the glass.

Fig. 5b: Measurement of the two glass beads without the AMS.

Fig. 5c: Using the AMS, surface structures not only within the depths but also in the focal plane can be displayed more clearly.

Fig. 5d: Zoom into the upper part of the right glass bead wihout…

Fig. 5e: …and with AMS, showing how the decorative pattern continues along fractures.

Fig. 5f: This becomes even more cleary visible when comparing the two scans of the left glass bead, measured without AMS…

Fig. 5g: …and with AMS. The multi-colored decoration is applied around a homogeneous glass core. The AMS scan further reveals small damages and air bubbles across the surface.

A Tinned Copper Cauldron

Fig. 6a: A tinned copper cauldron, reworked into a water vessel (left). The strong surface topography becomes especially visible when looking at the measurement setup of the metallic cauldron below the M6 JETSTREAM (right). Parts of the surface are up to 10 cm below the focal plane.

Fig. 6b: The Cu distribution clearly highlights the benefits of measuring this 3D object with the AMS feature.

Fig. 6c: The remnants of the tinning visible in the Sn-Kα map can be more clearly allocated to specific areas in the scan conducted with the AMS.

Using the AMS to Study Inaccessible Areas

As discussed above, using the AMS reduces the intensity up to 1/7 for the AMS 500. Nevertheless, as in the case of the tea cup, even without enhancing the measurement time per pixel, a gain in resolution is notable. This means, that even if the overall time available for conducting an analytical campaign cannot be prolonged, the AMS offers an opportunity to enhance results without having to consider the lower intensity.

The first measurement of the tea cup shown below was conducted with excitation conditions of 50 kV and 600 µA. The pixel size is fixed to 250 µm, while the dwell time per pixel is at 20 ms. In the following example, the cup was remeasured without changing any of the measurement parameters but using the AMS 500. While only 1/7 of the intensity is expected, the rose decoration still becomes visible to a high resolved extent. With a 7x times longer measurement time per pixel (140 ms), application traces of the various colourants are clearly notable.

Decorated Teacup

Fig. 7a: The AMS allows to measure and visualize areas of interest, that re inaccessible for focusing. In this porcelain cup, the very fine flower decoration is located at a depth of 4 cm. Studying the colourants used for creating the decoration requires a sophisticated analytical approach.

Fig. 7b: Scanning the cup without the AMS at 50 kV, 600 µA with a pixel size of 250 µm and a dwell time per pixel of 20 ms, the gold decoration on the outer edges of the cup and elemental signals, here displayed on the example of Pb, can be detected in areas beyond the focal plane, though the resolution is very poor.

Fig. 7c: When only using the AMS feature while maintaining the same measurement conditions as before, the decoration of the cup beyond the focal plane can be displayed to a higher resolved degree.

Fig. 7d: Increasing the measurement time by a factor of 7 to cope for the intensity loss that comes with the AMS, allows to get a clear image of the application of the tea cup’s décor.

Conclusion

As shown in the examples above, the AMS is a powerful feature that allows users to retain a high spatial resolution of the micro-XRF elemental distribution images. Though its use results in a decrease of the relative intensity, this can be easily circumnavigated by increasing measurement time or incoming flux. Even if no additional time is available, a higher resolution in areas far out of focus can be gained.

For more information watch our on-demand webinar:

Webinar: New Horizons of micro-XRF in Art and Conservation

Discover M6 JETSTREAM Contact an expert

References

Thomas Calligaro, L. Arean, C. Pacheco, Q. Lemasson, L. Pichon, B. Moignard, C. Boust, L. Bertrand, S. Schoeder, M. Thoury, L. Rosta, L. Szentmiklósi, J. Füzi, Z. László, V. Heirich: A new 3D positioner for the analytical mapping of non-flat objects under accelerator beams. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms, Volume 467, 2020, https://doi.org/10.1016/j.nimb.2020.01.028.
Alfeld, M. W. E. M., Tempel, P., & van der Wijk, V. (2023). Cable Robots as Conventional Linear Stage. Alternatives for the Investigation of Complex-Shaped Objects via Macroscopic X-ray Fluorescence Imaging. Quantum Beam Science, 7(4), Article 37. https://doi.org/10.3390/qubs704003
Alfeld M, Gonzalez V, van Loon A. Data intrinsic correction for working distance variations in MA-XRF of historical paintings based on the Ar signal. X-Ray Spectrom. 2021;50:351–357. https://doi.org/10.1002/xrs.3198