Non-destructive preservation state estimation of waterlogged archaeological wooden artifacts.

Non-destructive preservation state estimation is an essential prerequisite for the preservation and conservation of waterlogged archaeological wooden artifacts. Herein, Near Infrared (NIR) spectroscopy coupled with orthogonal partial least squares discriminant analysis (OPLS-DA) were applied to assess sixty-four waterlogged archaeological woods collected from seven excavation sites in the period range of 2900 BCE-1912 CE, aiming at developing a non-destructive, accurate and rapid preservation state estimation methodology. The role of non-decayed recent wood of relevant species on preservation state estimation was studied in prior, showing the use of non-decayed recent wood could not improve the predictive ability. Besides, the high variability in terms of chemical structure between archaeological softwoods and archaeological hardwoods did affect the preservation state estimation. Thus, a simple OPLS-DA model of non-destructively distinguishing archaeological hardwoods from softwoods, R2Xcum of 0.659, R2Ycum of 0.836 and Q2cum of 0.763, was established to avoid and overcome destructive approach for wood identification. Then, the well-defined three grouped separations of slightly-decayed, moderately-decayed and severely-decayed waterlogged archaeological woods were revealed in OPLS-DA models, providing R2Xcum of 0.793, R2Ycum of 0.738, Q2cum of 0.680, and R2Xcum of 0.780, R2Ycum of 0.901, Q2cum of 0.870, for waterlogged archaeological hardwoods and waterlogged archaeological softwoods respectively. Potential predictive wood spectral bands were screened and tentatively identified as hydroxyls of crystalline cellulose, acetyl groups of hemicelluloses, C-H bands of lignin, which guaranteed the elimination of non-structural compounds, such as water and inorganic components interference. Furthermore, the developed NIR methodology was validated by an extensively used destructive method consisting of anatomical characteristics, maximum water content and basic density analyses. The results indicated that NIR coupled to chemometrics could non-destructively and accurately predict the preservation states of waterlogged archaeological wooden artifacts and avoid the interference of water and inorganic deposits.

Evaluation of the Deterioration State of Archaeological Wooden Artifacts: A Nondestructive Protocol based on Direct Analysis in Real Time - Mass Spectrometry (DART-MS) Coupled to Chemometrics.

Evaluating the deterioration state of archeological wood is obligatory before the preservation of archeological wooden artifacts. Herein, a nondestructive, accurate, and rapid methodology is first developed via direct analysis in real time-mass spectrometry (DART-MS) with chemometrics to classify archeological wood and recent wood into 3 groups according to their deterioration states. As water in wooden artifacts probably affected the ion fragmentation process during DART-MS, ions responsible for evaluating the deterioration state were separately screened toward waterlogged archeological wood and dried archeological wood by partial least-squares discriminant analysis (PLS-DA). The well-defined separation of severely decayed archeological wood, moderately decayed archeological wood and recent wood was revealed in PLS-DA models. Twenty and 27 wood fragment ions were further screened as key variables to evaluate the deterioration state of waterlogged archeological wood and dried archeological wood, respectively. They were tentatively identified as ions of lignin monomeric compositions, lignin dimers, lignin trimers, and oligosaccharides. Results strongly suggested that differences in the structure and relative abundances of wood cell wall components accounts for the evaluation of deterioration state by DART-MS coupled to chemometrics. PLS-DA models provided R2Y = 0.836, Q2 = 0.817, and R2Y = 0.754, Q2 = 0.682 were then established separately using mass spectral fingerprints of respective potential predictive wood fragment ions. Furthermore, archeological woods, consisting of Castanopsis, Quercus, Idesia, Populus, and Cunninghamia species and with an average MWC range of 103-465%, were used as an external validation set and evaluated with the methodology developed herein and the MWC criteria. Results showed that DART-MS coupled to chemometrics could accurately predict the inhomogeneous deterioration states of archeological wooden artifacts and avoid the interference of inorganic deposits, in comparison with the MWC criteria.