**4. Discussion**

This article describes a generally applicable approach to evaluation whether an imaging spectroscopic instrument is suitable as a tool for monitoring of archival documents. Depending on the object type to be investigated, instruments are typically selected based on their spectral range and resolution, their field-of-view and spatial resolution. However, in order to be useful for monitoring, an IS instrument has to fulfil two additional requirements:


We have shown that Requirement 1 can be tested by subjecting Blue Wool Standards to repeated measurements, corresponding to the intended duration of the monitoring schedule. The induced spectral change needs to be determined with an independent, sufficiently sensitive method, such as a spectrophotometer.

From a conservation point of view, the decision whether the predicted spectral changes induced by IS measurements themselves are acceptable or not must become part of the overall risk assessment of an institution's exhibition and monitoring schedule. A reasonable guideline would be that the spectral changes induced by measurements must be significantly less than those expected from exposure during exhibitions.

In the case study discussed here, the "SEPIA" hyperspectral imaging system was tested for an intended monitoring period of 20 years with an average of two measurements per year. The spectral changes induced by the tested IS instrument in Blue Wool Standard Grade 1 remained below the threshold for visually perceptible colour changes of ∆E2000 = 0.7, which fulfils Requirement 1.

In order to test Requirement 2, sample materials representative of the monitored objects should be used in addition to standard reference materials. This is because the wavelength range and degree of spectral change depend on the material composition. Furthermore, the strength and spectral characteristics of light-induced change typically depend significantly on the irradiation spectrum. It is therefore advisable that the sample materials are exposed to accelerated degradation under conditions similar to those they are exposed to during exhibitions, in order to induce the same type of spectral change.

In our case study, four archival materials were subjected to accelerated ageing by 30 different combinations of light intensity and duration, corresponding to light doses of up to 500 klx·h. The deployed lamp type was the same as that used in the exhibition room *De Verdieping van Nederland* at the Nationaal Archief, to ensure that the same type of spectral change is induced as expected for exhibitions. The samples were irradiated with only moderate light intensities of up to 5000 lx and forced-air cooling was applied to minimize their temperature increase. The induced spectral changes should not be distorted significantly by any non-linearity of the material response with respect to the intensity (reciprocity-failure [36]) or by very high local sample temperature and low humidity. The spectral changes in the experiment are therefore expected to be comparable to those incurred by the materials over many years due to the normal exhibition schedule.

The changes were evaluated by calculating colour differences ∆E2000 and standardized Euclidian distances ∆Euclid between the sample spectra measured before and after accelerated light ageing. The minimum light dose required to induce a detectable spectral change varied considerably from material to material. For the most light-sensitive Blue Wool Standard Grade 1, the detection limit for light-induced colour change corresponded to an exhibition period of ~10 days.

The instrument was thus shown to have an acceptably small impact on the materials, fulfilling Requirement 1, and acceptable high detection sensitivity for changes induced by exhibition lighting, i.e., also fulfilling Requirement 2.
