2.1.2. Acquisition of Spectra

As the Bruker Bravo does not include a camera, the measurement preparation demanded precise positioning. To locate and define the measuring spot on the paint-outs, a Melinex® template in the shape of the optical head was used. Measurements were taken at the closest possible distance without touching the surface, which was approximately half a millimetre. Since this study was undertaken to investigate the possibility of analysing SOPs in real artworks where focusing on well-defined paint spots is mandatory, it was decided not to use the de-focusing tip. Handheld Raman spectra were recorded with initial settings of 100 ms and 20 acquisitions within the spectral range of 300–2000 cm−1. These settings proved to be efficient to verify the Raman signal and fluorescence. Depending on the signal to noise ratio of the spectrum, these settings were tuned to: acquisition times 100–300 ms; 50–200 acquisitions.

A common limitation of portable Raman devices is the reduced possibilities to control the measurement settings [19]. For the Bravo spectrometer, the lack of direct control of the laser energy was considered potentially problematic. As the maximum energy reaching the surface is quite high in comparison to the Perkin Elmer microRaman instrument (Table 1), the measurement spot was checked under the microscope before and after analysis in order to be able to track possible laser induced damage.

Measurements were recorded without removing the varnish layer from the paintouts at the upper middle edge of the samples, in order to measure a spot with high pigment density.

#### 2.1.3. Interpretation of Raman spectra

Handheld Raman spectra were interpreted using reference spectral databases. Raman reference spectra of SOPs have been published listing the peaks with their relative intensity or in digital databases [10,14,29]. Furthermore, the continuously updated in-house built Raman spectral library was used.

It should, however, be taken into account that reference spectra are mostly acquired on pure materials, like pigment powders. When compared with paint systems, differences in heat dispersion or fluorescence behaviour may lead to slight differences in the spectra [43,44]. The excitation wavelength and power of the laser can influence the reference spectra too. As KIK-IRPA documents the exact measurement settings for each reference spectrum (https://soprano.kikirpa.be, accessed on 21 August 2020), their database was used as the main reference for the interpretation of the spectra, in addition to the inhouse database.

In some cases, it was not possible to interpret the Raman spectrum and assign it to a defined pigment. In these specific cases, additional methods, such as ultra-high performance liquid chromatography mass spectrometry (UHPLC-MS), were used to try to identify the unknown samples. These results were verified by analysing corresponding references from the in-house reference collection.
