Degradation assessment of archaeological oak (Quercus spp.) buried under oxygen-limited condition

2.1 Sample preparation

The archaeological oak samples were 1 m long beams presumed to be from a fence from an 14th-century archaeological site called Olita Mate located on Armenia Street in Iasi city, Romania. The archaeological oak wood beams were used for paving and arranged vertically embedded in the clay soil of the central street and covered with macadam paving (cubic stone with dimensions of approx. 150 × 150 × 150 mm) of granite and diorite. The stone pavement was eventually destroyed, and the archaeological oak beam samples were buried to a depth of approximately 2 m. 250 archaeological beams were extracted on site and 6 of them were brought to the ARHEOINVEST laboratory (Iasi, Romania) for dendrochronological analysis and degradation assessment. Despite the sample age, the wood structure was intact but darkly discoloured by the presence of iron that had reacted with the tannins to form iron tannate (Figure 1), as described previously (Ghavidel et al. 2020d). The beams were in the form of trunks with a diameter of less than 300 mm, covered with tree bark. The outer 15 mm of each sample was heavily encrusted with soil and was cut away to facilitate sample preparation. Three discs were extracted from each beam, and then samples with size of 10 × 10 × 30 (±0.5) (R × T × L) mm³ were cut from each disc. The control samples were obtained from a 45 mm by 90 mm by 4 m long defect-free board obtained from the Iasi region in Romania. Only defect-free samples (without knots, cracks, or reaction wood, and with an annual growth ring slope of less than 5°) were used for the testing. The moisture contents at the time of collection of the archaeological and recently-cut oak were 9.3 and 7.6%, respectively. The respective oven dry density of the recently-cut and archaeological oak samples were 730 kg/m³ and 660 kg/m³. Samples from heartwood were used for both recently-cut and archaeological wood.

Figure 1:

Recently-cut (A) and archaeological oak (B).

2.2 X-ray photoelectron spectroscopy (XPS)

Chemical compositions of wood samples were examined at room temperature using a concentric hemispherical electron energy analyser (CHA) (Specs model EA10 plus, Bestec Co, Berlin, Germany) with an Al-Kα radiation source, a photonic energy of 1486.6 eV, and a base pressure of 2 μPa, as described previously (Ghavidel et al. 2020a, 2020b). Three samples were selected randomly from different parts of archaeological and recently-cut samples, two measurements were performed per sample, and the average value was reported (n = 6).

2.3 X-ray diffraction analysis (XRD)

The crystalline structure of the samples was measured on wood ground to pass a 10 µm mesh screen (P7 premium, Fritsch) on a Philips X’Pert MPD PW 3040 X-ray diffractometer (Netherlands) equipped with a PW 3050/10 goniometer, divergence slit 0.5°, anti-scatter slit 0.5°, receiving slit 0.6 mm, secondary graphite monochromator, mask 15 mm, 40 kV activity and 30 mA with Cu Kα radiation. Analysis was carried out in a continuous mode using the scanning range of 5–50°2θ with a phase width of 0.02°2θ as previously described (Ghavidel et al. 2020a, 2020b). The sample rotating speed was 1 rpm, and each move took 4 s to count. The crystallinity index (CrI) and crystallite size were determined according to Equations (1) and (2) (Ghavidel et al. 2020b):

where I_am is the minimum intensity peaks at 2 = 16.14°, assigned to the amorphous phase, and I₂₀₀ is the intensity of the crystalline peak at 2θ = 22.5°, allocated to both crystalline and amorphous parts.

where K is a constant with the value of 0.94, λ is the X-ray wavelength, β is the full width at half maximum of the diffraction band and θ is the Bragg angle corresponding to the (200) plane.

2.4 Water sorption analysis

Water vapour sorption behavior of ground wood samples (passed through a 20 mesh sieve) was investigated using a Q5000 SA automated vapor sorption (AVS) analysis apparatus (Q5000 SA, TA Instruments) at a constant temperature of 25 °C (Ghavidel et al. 2020a; Hosseinpourpia et al. 2018, 2019, 2020). Sorption isotherm was measured from 0 to 95% relative humidity (RH) and then reversed to 0 percent RH with dm/dt criteria of less than 0.005% per 10 min. The weight at equilibrium was measured using a micro-balance at each RH, and the equilibrium moisture content (EMC) was calculated accordingly.

2.5 Microscopic analyses

The degree of damage was assessed using both light and scanning electron microscopy. Three samples from different parts of archaeological oak were selected and thin sectioned for light microscopy from the transverse and longitudinal faces of the water-soaked archaeological wood. The transverse sections were stained with 1% (w/v) safranin O in ethanol while the longitudinal sections were stained with 0.1% (w/v) aniline blue (Ghavidel et al. 2021b, 2020c). Samples were imaged with a Zeiss Axioscope equipped with AxioCam and Axio vision software (Carl Zeiss MicroImaging GmbH, Germany). The archaeological oak samples were still relatively sound and could be hand-sectioned for staining and analysis, while a large number of sections were prepared to carefully study the degradation patterns.

Recently-cut and archaeological samples were further examined using a Hitachi S-3400 N (Tokyo, Japan) scanning electron microscope operated at an 8 kV accelerating voltage. Samples (5 mm cubes) were surfaced with a razor blade to create a smooth surface and mounted on aluminium stubs. The surfaces were examined directly without sputter coating (Ghavidel et al. 2020d, 2022). A TSCAN (Brno, Cheque Republic) FESEM was used where the higher resolution was required following a similar sample preparation method.

3.1 Chemical structure and crystallinity of archaeological oak

X-ray photoelectron spectroscopy (XPS) analysis of the recently-cut and archaeological oak samples produced C1s energy level spectra with four sub-peaks, e.g., C1, C2, C3, and C4, and O1s energy level spectra with two sub-peaks of O1 and O2 (Figure 2).

Figure 2:

High–resolution spectra of the C1s and O1s energy levels of recently-cut (A, C) and archaeological samples (B, D) of oak.

The C1 signal (Figure 2A, B) which mainly represents the presence of extractives and lignin increased substantially in the archaeological oak with a simultaneous decrease in the concentration of oxygenated C atoms. Increasing proportions of lignin reflect the tendency for the hollocellulose components to be degraded in archaeological samples, leaving a lignin-rich residual (Capano et al. 2015; Christiernin et al. 2009). The C2 sub-peak which represent cellulose and hemicellulose was slightly lower in the archaeological oak, while the C3 and C4 sub-peaks were not markedly changed. These changes reflect the loss of low molecular weight components including extractives, hemicelluloses, and even a small portion of amorphous cellulose, which typically occurs during the first 150 years of ageing (Ghavidel et al. 2020a, 2020b; Popescu et al. 2009). The higher resistance of oak to microbial attack might help explain the lower losses of these wood components. Popescu et al. (2009) reported a significant decrease in the C1 sub-peak and an increasing C2 sub-peak in archaeological lime wood as the result of an increase in C-O groups caused by oxidation and hydrolysis reactions that occurred during the ageing process. The results in the present work, however, suggest that neither oxidation nor hydrolysis occurred in the archaeological oak (Table 1). For the oxidation of lignin, oxygen is required, and the low access to oxygen in the archaeological site largely limited that process. Additionally, oxidation and hydrolysis are mutually reliant and catalyse one another. The β-glycosidic link of carbohydrates is more vulnerable to hydrolysis processes than oxidation (Łojewska et al. 2005). The oxygen/carbon ratio and C_ox/C_unox were much lower in the archaeological sample (Table 1) which could be attributed to bacterial degradation of the carbohydrates. Thus, hydrolysis may be unlikely to occur in the archaeological oak sample. The C1 sub-peak corresponding to lignin and extractives in wood was also higher in archaeological oak, supporting the premise that carbohydrates in this fraction have been preferentially degraded in archaeological oak. The decreased C_ox/C_unox also suggests an increase in C–C bonds, which tend to be more closely associated with lignin.

The O1 and O2 sub-peaks (Figure 2C, D) are related to the oxygen atoms that are linked with double and single bonds to carbon atoms, respectively (Popescu et al. 2009). The O2 sub-peak of the archaeological oak decreased, potentially as a result of hemicellulose degradation (Hedges 1989; Nzokou and Pascal Kamdem 2005; Popescu et al. 2009). Hemicellulose is considered to be the most susceptible of the three polymers to microbial degradation and can be degraded by bacteria in the absence of oxygen, while lignin degradation is oxidative and requires the presence of some oxygen (Björdal et al. 2000; Goodell et al. 2020). Oxygen is generally limited in buried substrates. These results are consistent with anaerobic bacterial degradation.

The crystalline structure and degree of crystallinity of recently-cut and archaeological oak wood samples are shown in Figure 3 and Table 2. I₂₀₀ and I_am represent the maximum and minimum intensities of the crystalline cellulose, respectively. The crystalline index is shown by CrI and the crystalline size of the cellulose is represented by L. There were small differences in the crystallinity of cellulose between recently-cut and archaeological oak that may reflect the leaching of extractives from the archaeological materials. A small reduction in the crystalline size of the archaeological wood could be related to moisture absorption, chemical reactivity, or swelling (Lionetto et al. 2012). Although, it is known that the degradation of hemicellulose and amorphous cellulose increases the crystallinity index, the total crystallinity percentage may decrease when available degraded carbohydrates are consumed by degrading agents (Howell et al. 2009).

Figure 3:

X-ray diffractograms from archaeological and recently-cut oak.

Degradation patterns differ by deterioration agent with white rot fungi tending to erode cell walls, brown rot fungi producing a general attack that can be difficult to detect microscopically, and soft rot fungi producing either erosion of the cell wall or diamond-shaped cavities within the S₂ cell wall layer. Bacterial decay can be characterised as either erosion of the wood cell wall from the lumen outward or tunnelling within the cell wall. Bacterial decay in anoxic environments tends to start on the surface and slowly extend inwards (Björdal et al. 2000; Pedersen et al. 2021; Singh et al. 2022), degrading the wood polymers and influencing the crystalline structure of wood.

3.2 Moisture sorption of archaeological oak

Moisture sorption reflects the availability of hydroxyls on the hemicellulose and amorphous cellulose fractions of the wood and would be expected to decline as bacteria degrade this fraction (Han et al. 2020). The isotherm curves of recently cut and archaeological oak samples illustrated the typical type II sigmoid character of cellulosic materials (Figure 4). The ageing process altered the sorption behaviour of archaeological wood samples through the whole adsorption and desorption runs especially at higher relative humidities (RHs), e.g., over 80%. The archaeological oak wood reached slightly higher equilibrium moisture content (EMC) values during the sorption and desorption processes across the entire hygroscopic range. This might be attributed to the slightly lower crystallinity in the archaeological oak samples in comparison with the recently-cut one, as indicated by the XRD analysis. Figure 4B demonstrates the sorption hysteresis of oak samples. As with the sorption isotherms, the hysteresis values were similar for both sample types. These results differ from the higher EMC values found during adsorption/desorption processes with elm and poplar samples exposed for prolonged periods under low oxygen conditions (Ghavidel et al. 2020a). The differences in hysteresis may reflect variations in volumetric swelling as a result of kinetic retardation during shrinkage/swelling (Hosseinpourpia et al. 2016). The slightly higher values in sorption hysteresis of the archaeological oak might be related to reduced flexibility of wood cells and/or fewer hydroxyl groups, as a result of the degradation and ageing processes. Degradation of amorphous cellulose and hemicellulose over time might also result in a more stable and lower degradation rate of crystalline cellulose levels (Howell et al. 2009). Moreover, the degradation of hemicellulose in archaeological wood reduces the length of the crystal regions, which may result in the appearance of new absorption sites (Guo et al. 2018), and, consequently, increment of equilibrium moisture content.

Figure 4:

Adsorption and desorption isotherms (A) and sorption hysteresis (B) of recently-cut and archaeological oak at changing RH levels.

3.3 Morphology of the degraded wood

The optical microscopy images were shown in Figure 5A–C. Light microscopic examination of transverse -sections of degraded cells indicated that decay started in the lumen and progressed outwards until it reached the middle lamella where the damage stopped. Lignin levels are typically highest in the primary cell wall and middle lamella, but this region is not directly affected by the bacterial attack. Lignin is present at proportionally lower levels in the secondary cell walls, but its presence appears to reduce cellulose degradation. In addition, lignin creates a physical barrier that affects and reduces the biological degradation of cellulose (Singh et al. 2022). This pattern again supports the premise that bacterial decay largely affected the carbohydrates such as cellulose and hemicellulose, leaving a lignin-rich fraction. This attack pattern is consistent with an attack by erosion bacteria (EB) as previously described (Cha et al. 2021; Nilsson and Björdal 2008; Singh et al. 2022). A similar pattern was also observed in the longitudinal sections as the characteristic fine ducts could be found in the longitudinal section (Figure 5C, arrows). No fungal hyphae were detected in the transverse or tangential sections.

Figure 5:

Light microscopic images of transverse sections of archaeological oak (A, B) showing cell wall erosion (arrows) and a tangential section using polarised light showing cell wall damage (C).

The scanning electron (SEM) micrographs of archaeological oak showed no evidence of extensive damage to the rays, vessels, or fibers (Figure 6). Even the tyloses were intact in the vessels (yellow arrow in Figure 6C). Since the archaeological oak was underground for over 600 years, there was evidence of soil infusion into the wood cells, especially in the parenchyma, which is consistent with prolonged timber burial (Huisman et al. 2008; Smith and Lillie 2007). No evidence of fungal activity was found in SEM images. Partial cell wall degradation was infrequent, suggesting that bacterial degradation was still at a relatively early stage (green arrow in Figure 6B). The absence of fungal hyphae does not completely eliminate them as possible degradation agents but the absence of any hyphae suggests that they were not primary agents under these conditions.

Figure 6:

Scanning electron (SEM) micrographs of transverse (A, B), and tangential sections (C, D) of archaeological oak showing bacteria, cell wall erosion, and collapsed ray cells (Arrows). Field emission scanning electron microscograph (FESEM) (E, F) of archaeological oak wood showing erosion bacteria (orange arrows) on the wood cell wall.

The rod-shaped erosion bacteria were observed using field emission SEM (FESEM) (Figure 6E, F). These bacteria often had pointed ends and were 102 µm long, which was consistent with previous reports (Daniel 2014). Nilsson and Björdal (2008) found that erosion bacteria required special conditions for isolation and were only active under limited-oxygen conditions. The degradation patterns of EB are comparable to the soft-rot, where the bacteria align with the underlying cellulose microfibrils in the cellulose-rich S₂ cell wall layer. The location of EB in this cell wall zone may help explain how they primarily degrade cellulose (Daniel 2014; Nilsson and Björdal 2008; Singh et al. 2022).