In-depth phenolic characterization of iron gall inks by deconstructing representative Iberian recipes

Iron-gall ink is one of the most important inks in the history of western civilization. The deep black colour results from Fe3+ complexes with phenolic compounds available in gall extracts. Unfortunately, it induces the degradation of both ink and support over time. Furthermore, our knowledge of these complex molecular structures is limited. This work aims to overcome this gap, revealing essential information about the complex structures of these pigments and dyes that will create a breakthrough in the next generation of conservation treatments. It presents the first in-depth phenolic identification and quantification of extracts and inks, prepared with and without gum arabic (an essential additive in medieval recipes). Five representative Iberian recipes were selected and prepared. Their phenolic profile was analysed by HPLC–DAD and HPLC–ESI–MS, which revealed that the phenolic compounds present in higher concentration, in the gall extracts, are pentagalloylglucose and hexagalloylglucose (0.15 ± 0.01–32 ± 3 mg/mL), except for one recipe, in which gallic acid is the main phenolic. The influence of the ingredients is also discussed by deconstructing the recipes: extracts of additives as pomegranate peel and solvents used in the extraction of the galls (vinegar and white wine) were characterized.

: SIM-MS (selected ion monitory -mass spectrometry) chromatogram of Hexagalloylglucose isomers present in the Braga extract ( Figure 2 and Table 2

Principal component analysis report
Principal component analysis (PCA) models were developed from Gallic acid, PGG+HGG and sum of phenolic compounds concentrations (Table 3). Models for extracts, inks with and without arabic gum were performed in Matlab Version 8.6 (R2015b) (The Mathworks, Natick, MA) and the PLS Toolbox Version 8. 2.1 (Eigenvector Research,Manson,WA).
Autoscaling was applied to all three variables included in the PCA model (scaling was applied for extracts, inks with and without arabic gum independently). More than 95% of variance was captured in all three models with two principal components. This can be explained by the correlation found between the sum of phenolic compounds and the PGG+HGG concentrations. The concentration of gallic acid was found to be mainly uncorrelated with the remaining especially in the case of the extracts. The variability found in terms of sum of phenolic compounds and the PGG+HGG concentrations is mainly expressed in all models in the first principal component while the variability of gallic acid is expressed in the second principal component. For instance, Montpellier and Guadalupe recipes show a higher variability in terms of gallic acid concentrations. On the other hand, recipes from Cordoba, despite presenting substantially lower amounts of phenolic compounds and PGG+HGG are also more consistent and present less intrinsic variability.
Madrid samples present also very consistent results for all samples (also showing lower amounts of gallic acid), independently of the analyzed dataset. Braga samples present a nonnegligeable variability in terms of phenolic compounds and PGG+HGG. Considering the different recipes and the different datasets it is possible in many cases to cluster the results according to the recipe, except probably samples from Braga that show some intrinsic variability, do not forming a noticeable cluster in the first two components.          Table S4: Concentration of all individual phenols (expressed in mg/mL of equivalents of gallic acid) found in all extracts and inks following the recipe Guadalupe.   1.43g of galls were crushed and added to 50mL of millipore water and left to extract at room temperature for 3 days. After this time, the solution was heated and reduced to 1⁄4 and then filtered. 0.95g of ground arabic gum were added, under stirring until dissolution, and, after cooling, 0.95g of FeSO 4 were added. It was left to settle for one day and then filtered.
1.918g of well crushed galls and 0.959g of pomegranate peel were added to 50mL of millipore water and left to extract for 8 days at room temperature. After this period, the solution was slowly heated until boiling point and after cooling it was filtrated. 1.91g of FeSO4 were placed in a glass cup and covered in water and dissolved, and this solution was later added to the extraction. The solution was left to settle for one day and then filtrated. After, gum arabic was added and left to settle for two days.
5.6g of crushed galls, 3.75g of FeSO 4 and 1.87g of gum arabic were placed in three separate glass cups. 37.5mL of millipore water and white wine (12.5mL of wine) were distributed by the cups in the following manner: enough solution to cover the FeSO 4 and the gum arabic and the rest for the galls. The solutions were left to settle for 6 days, stirring every day with a dried branch from a fig tree. After this time, the galls solution was heated with low temperature for 10 minutes (60ºC) and filtered. After a while, and without letting it be too cold, the FeSO 4 and gum arabic solutions were added, always under stirring. After being at room temperature, 1.87g of CuSO 4 were added and the solution was filtrated. •