Assessing the Degradation Status of the Imperial Doors of the Ascension Church, Grindu Commune, Romania

Abstract

To asses the degradation status of the Imperial Doors of the early 19th century Ascension Church iconostasis, a complex study consisting of micro-optical and scanning electron microscopy followed by energy-dispersive X-ray fluorescence and Fourier transform infrared spectroscopy was performed. Accordingly, the entire left door and some small fragments of gilded wood were investigated. The final results evidenced a certain degree of degradation of the lime wood and gilded surfaces, mainly due to the bacterial and fungi attacks given the increased humidity and the presence of more than a century of candle soot. Also, some unsuccessful restorations performed using brass paint instead of gold foils were evidenced. Overall, this study permitted elaborating more appropriate procedures for the iconostasis’ full restoration to its initial form, given that the Ascension Church is classified as a historical objective of national and universal value.

1. Introduction

Modern restorers/conservatives need to know more about the remaining original portions of any work of art and understand how the piece of art has been treated over the years. Also, a major problem facing conservatives is caused by improper conservation treatments from the past [1].

The methods of art restoration used in earlier periods were closely linked to the techniques known at that time. Advances in science and technology and the development of conservation as a profession during the 20th century have led to safer and more effective approaches to studying, preserving, and restoring objects. Modern conservation practice adheres to the principle of reversibility, which dictates that treatments should not cause permanent alteration to the object. Art conservation has become an important tool of research; it is a standard practice among professional conservators to document treatments with photographs and written reports [1,2].

Thus, there is a real opportunity for science to further integrate into the preservation as well as restoration of modern art, especially in diagnosing art, because the appearance of a work of art reveals only its current state.

Over the past few decades, the diagnosis of artworks has become less invasive thanks to the development of advanced non-invasive techniques, such as X-ray fluorescence (XRF) [3,4,5], Fourier transform infrared (FTIR) [5,6,7], and attenuated total reflectance Fourier transform infrared (ATR-FTIR) [8,9] spectroscopy, or scanning electron microscopy (SEM) [10,11], to enumerate only those methods that were used in the present study.

These techniques can provide insight into the history of a work of art and evaluate previous interventions, which, in turn, can help with conservation and recovery efforts [12].

In many cases, the diagnosis will reveal the condition of a painting and may indicate its vulnerabilities. Sometimes, the diagnosis will reveal a ‘pentimento’, or an early change of mind by the artist [13]. In other cases, the diagnosis reveals all the interventions of other artists or restorers who hide or destroy part of the original work. Depending on the philosophy of restoration, the work of art is preserved in its current state, restored to its original state, or left as a pastiche of old and new [14].

At the same time, cultural heritage is increasingly threatened by damage and even destruction from natural causes, as well as social and economic instability. In this regard, painted wood carvings or panels are very vulnerable to changes in the microclimate [15].

Before any restoration or consolidation, it is important to carefully examine the object to determine its current state of preservation with maximum precision so as to choose the most suitable treatments [16,17].

Given the highlighted aspects, the assessment of the state of degradation of a work of art should be substantiated as a concept of mandatory diagnostic methodology. This must be materialized by a protocol based on the fusion of historical and stylistic data on the one hand and scientific data, which should include local microclimate conditions, the nature of aggressive agents, and their effects, on the other [18,19].

The protocol must start from the knowledge of the materials used in the creative process, as well as in previous interventions, together with data based on historical and stylistic documentation about aggressors who have acted over time on the work of art. To determine the degree of degradation, it is necessary to establish a correct relationship between the microclimate and aesthetic changes. At the same time, the proposal for appropriate chemical reactions becomes a well-founded matter of the protocol. In addition, it is necessary to know in depth the details of alteration processes and accelerated aging on replica samples. Finally, one can choose the most appropriate complementary techniques to achieve the best results in the validation of restoration data, in a minimum time, and at a low cost, by providing accurate data and not approximate proposals.

We applied this concept of diagnosis in a case study represented by the 19th century “Imperial Doors”, a golden carved and painted door representing the central entrance of the iconostasis of the Ascension Church, Grindu commune, Romania. Therefore, we investigated the two main components of these doors, that is, the wooden support and the gilded layers, choosing the most representative fragments.

The Church of Ascension Iconostasis

The Church of Ascension (Grindu commune, Ialomita county) is a class A historic objective of national and universal value, classified by the LMI-IL-II-m-A-14130 code according to the Romanian Ministry of Culture classification [20].

The church was rebuilt between 1838 and 1842 by the Great Clucer (Lord Stewart) Ispas Fagarasanu on the ruin of an older sanctuary founded by Dragan Fagarasanu in 1816, almost destroyed by the earthquake of 1837. All data are attested by the inscription of the current church written in Cyrillic letters and engraved on a votive stone in the porch of the church (“pisania” in Romanian), placed above the entrance door. The iconostasis of this church has great artistic and historical value. The blessing of the church was made in 1841 by Metropolitan Neophyte, the ruler of Wallachia being Alexandru Dimitrie Ghica (1834–1842), both mentioned, along with several generations of founders.

In Eastern Christianity, the iconostasis (Greek: $\epsilon \iota \kappa o\nu \sigma \tau \mathsf{ά}\sigma \iota o\nu$) is represented by the wall that spares the sanctuary church ship. Developed in the early period of the Byzantine empire, the iconostasis was originally a modest screen of wood, but, by the 12th century, in most churches, wood replaced stone and the iconostasis rose in height [21,22,23], becoming a characteristic feature of the churches of the Byzantine and Coptic rite. A traditional iconostasis has three doors: an “Imperial” or “Holy Door” in the center, and north and south ones on both sides [24,25].

The iconostasis of the Ascension Church was realized in a combination of Baroque, Rococo, and Classical motifs at the beginning of the 19th century. It represents a monumental artwork (height 6.70 m, width 7.70 m) made of gilded lime wood at one of the monasteries of Mount Athos. The entire iconostasis is supported by eight pillars with ornaments in the form of vines and grapes. Following the pure Byzantine style, it is adorned with six rows of ornaments and seven rows of icons, of which four imperial icons on both sides of the central “Imperial Door” are coated in silver. The pillars have capitals on each row of ornaments.

Above the imperial icons, six pillars hold six chalices with pansies flowers. The second line of the capitals sustains eight winged doves ready to fly. Each icon is framed by ornamental leaves, buds, and bouquets of roses. Eight icons, including the “Imperial” one and depicting scenes from Lord Jesus’s life, were painted by an unknown Athonite craftsman, as shown in the inscriptions written in Greek on the back of the icons.

The painting dating from the middle of the 19th century is of Byzantine style with Western influences. Tempera grassa combined with gold leaves is used for the Saints’ aureole while clothes are gilded by colloidal gold [26]. A photographic image of the iconostasis is reproduced in Figure 1, while Figure 2 illustrates the “Imperial Doors” and more details regarding its actual conservation status.

2. Materials and Methods

2.1. Imperial Doors

Each half of the “Imperial Doors” has an eagle head with the body of a coiled snake holding in its beak a vine with grapes from which other eagles pinch (Figure 2).

Concerning the gilding, the entire iconostasis was initially gilded following a Classical technique consisting of the application of multiple layers of gesso containing a certain amount of lead white, followed by a layer of Armenian bole [26,27,28], spread with gilder’s liquor binder consisting of a solution of water, alcohol, and rabbit skin glue and, finally, while the surface is still wet, gold leaf. Over time, the iconostasis was damaged and, in 1924, it was partially restored. Unfortunately, the original gilding was replaced by brass painting so only a few portions of the initial gilding remained. After that, the iconostasis has passed through several stages of non-professional conservation and restoration processes to improve the appearance. As a result, the imperial doors suffered profound alterations that also compromised the beauty of the Ascension Church.

2.2. Sampling

For a better assessment of the conservation status of the “Imperial Door”, investigations were mainly performed in situ on whole, rather than detached, sections (Figure 2), while more fragments of a wooden support and gilded wood (Figure 3, Figure 4 and Figure 5) were detached and individually investigated.

In the case of gilded wood, the main information concerning the actual conservation status of gilded surfaces (original and restored) as well as of the traces of painting layers deposited on these were obtained by XRF and ATR-FTIR (see the next section).

2.3. Analytical Techniques

The in situ examination was performed by visual inspection as well as by a CANON EOS-R6 MarkII photo camera while the most representative detached samples were investigated by a handheld Dino-Lite Digital Microscope AM 4013TL-M40 (http://www.dinolite.us/, accessed on 21 August 2024), with a variable magnification of 1 to 37X.

Elemental analysis was performed on detached samples of painted and gilded wood by XRF spectroscopy by a hand-held Innov-X Alpha Series (https://www.pine-environmental.com/products/innovx_6000_alpha_series, accessed on 21 August 2024) spectrometer. The instrument, provided with a W anode at 45 kV anodic potential and 6 $\mathsf{\mu }$A current, has an energy resolution of 230 eV for the 5.95 keV Mn ${\mathrm{K}}_{\alpha }$ line.

Wood fiber was investigated by a Quanta 200 FEI environmental scanning electron microscope (SEM) (https://www.chalmers.se/en/infrastructure/cmal/instruments/electron-microscopy/sem-fei-quanta/, accessed on 21 August 2024) at an accelerating voltage of 12.5 kV and an ambient pressure of 200 Pa.

To estimate the degradation status of the gilded surfaces as well as of the wooden material of the Imperial Door, a Bruker Helios Fourier transform infrared (FTIR) microscope (https://www.bruker.com/en/products-and-solutions/infrared-and-raman/ft-ir-routine-spectrometer.html, accessed 21 August 2024) equipped with a cryogenic HgCdTe detector was used. In the case of wooden detached fragments, selected specimens were finely ground, the resulting powder was embedded in KBr pellets, and the corresponding FTIR spectra were acquired in the range of 4000–600 cm⁻¹ at a 4 cm⁻¹ resolution with automatic background subtraction. Brucker OPUS 6.5 software (https://www.bruker.com/en/products-and-solutions/infrared-and-raman/opus-spectroscopy-software.html, accessed 21 August 2024) that performs atmospheric compensation, vector normalization, and baseline correction as well as additional rubber band correction was used too.

3. Results and Discussion

3.1. Church Microclimate

To assess the action of aggressor agents, the indoor microclimate of the Ascension Church, e.g., temperature, humidity, ambient lighting, and ${CO}_2$ concentration, was monitored monthly during one year.

As can be seen from Figure 6, the church inner space is characterized by a higher than optimum humidity during spring, autumn, and winter, a phenomenon that becomes very intense at lower temperatures. Along the summer months, inside the church, the temperature never exceeds 22 °C while the relative humidity remains relatively high due to inadequate ventilation. The illumination remains all the time at about 10 to 100 times lower than 50 lx, the value considered optimal inside a museum [29], but this is a characteristic of the majority of Byzantine stile churches.

At its maximum value, the ${CO}_2$ concentration, whose source could be the presence of a great number of people attending the Holy Liturgy, showed to fluctuate between 0.5 and 5 ppm during the summer while, in the winter time, due to properly adjusted gas stoves, its concentration never exceeded 15 ppm. As these values fall within the technical regulations, we consider that ${CO}_2$ could generate only minor alterations [30].

According to oral testimonies, the main pollutants indoor of the church over the years were carbonaceous particulate matter resulting from burning candles and incense, especially during the Holy Liturgy. As a result, various combustion compounds such as olefins, aromatic hydrocarbons, carbonyl and carboxylic moieties, and carbonaceous particles were released into the atmosphere during religious services [31].

It is worth mentioning that microclimate conditions, coupled with a high bio-receptivity of the lime wood, could favor the fungal attack too [31]. Also, in our opinion, the mentioned aggressor agents can stimulate over time a weathering phenomenon responsible for the physical and chemical transformations of the historic materials, resulting in visible aesthetic changes (Figure 2 insets) [32,33,34].

At a first visual inspection, the “Imperial Doors” present some of the most representative effects of decay and degradation due to the combined action of aging agents and weathering (Figure 2, Figure 3, Figure 4 and Figure 5), which will be further analyzed and discussed separately for the wooden support and gilded layers.

3.2. Wooden Material

Wooden support is recognized as a very sensitive composite toward relative humidity and biologic attack due to its organic origin. Its anisotropy, along the three principal anatomical axes—longitudinal, parallel to fibers, radial, and tangential—determines an uneven distribution of moisture as well as the biologic attack and oxidative aging, mainly developed in the tangential direction and less pronounced in the radial one [35]. With regard to the influence of temperature and humidity fluctuation due to the church microclimate, it can be observed on the wooden details illustrated in Figure 4 that show a variety of deep cracking and cracks, respectively, along with a considerable number of instances of swelling as well as contractions.

The multi-annual cyclic fluctuation in the temperature and humidity determines cyclic swelling and expansions as well as drying and contractions, which, in our opinion determines the apparition superficial to deep cracks, which finally manifests in the loss of matter (Figure 2 and Figure 3) as well as the appearance of embrittlement (Figure 4).

In turn, the embrittlement (Figure 4), which represents a specific degradation of natural polymeric materials, is usually caused by oxidative aging and/or by biologic attack. Oxidative aging is associated with an increased consumption of oxygen [36], better described by our proposed reactions (1)–(4), whose degradation products are carbonyl (2) and carboxylic (3) compounds, rich in oxygen:

$$\begin{array}{c}\hfill {\mathrm{O}}_2+\mathrm{h}\nu \to 2\mathrm{O}·\end{array}$$

(1)

$$\begin{array}{c}\hfill -{\mathrm{CH}}_2-{\mathrm{CH}}_2-+2\mathrm{O}·\to {-\mathrm{CH}}_2-\mathrm{HCO}+\mathrm{HO}·\end{array}$$

(2)

$$\begin{array}{c}\hfill -{\mathrm{CH}}_2-{\mathrm{CH}}_2-+2\mathrm{O}·\to {-\mathrm{CH}}_2-\mathrm{COOH}+\mathrm{H}·\end{array}$$

(3)

$$\begin{array}{c}\hfill \mathrm{HO}·+\mathrm{H}·\to {\mathrm{H}}_2\mathrm{O}\end{array}$$

(4)

At the same time, during the biologic attack, mold and bacteria feed cellulose, hemicelluloses, and less lignin (more resistant due to its aromatic moieties), releasing metabolites that stain the surface in shades of green, grey, or black (Figure 5a,c). Through a loss of backbone polymers, the wood is softened to embrittlement due to cellulose and hemicellulose degradation [37,38] described by our proposed reactions:

$$\begin{array}{c}\hfill -{\left({\mathrm{C}}_6{\mathrm{H}}_{10}{\mathrm{O}}_5\right)}_n-+bacteria\to -{\left({\mathrm{C}}_6{\mathrm{H}}_{10}{\mathrm{O}}_5\right)}_p-p<n\end{array}$$

(5)

$$\begin{array}{c}\hfill -{\left({\mathrm{C}}_5{\mathrm{H}}_8{\mathrm{O}}_4\right)}_m-+bacteria\to -{\left({\mathrm{C}}_5{\mathrm{H}}_8{\mathrm{O}}_4\right)}_q-q<m\end{array}$$

(6)

where Equations (5) and (6) refer to cellulose and hemicellulose degradation, respectively. As a result of bacterial action, in both cases, the polysaccharide chains become shorter.

Given the church microclimate, it is most plausible that the presence of a dark-greenish (Figure 3a) or black–brown (Figure 5c) stain is due to a biologic attack.

Concerning the origin of wooden material used for iconostasis, we should remark on the existence of a good agreement between the FTIR spectra of the iconostasis wooden fragment and the corresponding FTIR spectra of actual and dry lime wood (Figure 7). This finding confirms that the “Imperial Door” is made of lime wood, in accordance with historical documentation.

It is worth mentioning that the FTIR technique allowed, besides a confident identification of the nature of wooden material, a more precise investigation of the results of both biologic attack and any oxidative aging responsible for wooden support decay. For this reason, in analyzing the wood FTIR spectra, we have chosen the most relevant infrared domains as lying between 3500 and 2800 cm⁻¹ and 1800 and 600 cm⁻¹, (Figure 7). These domains of IR contain the most relevant “fingerprints” bands for attesting the nature of lime wood support [39] as previously reported in [40,41,42] on one hand and for identifying the major decay products on the other. After carefully analyzing the wood spectra illustrated in Figure 7, it can be remarked that almost all “fingerprints” bands in the IR bands mentioned before are present in all three types of lime wood, e.g., fresh, dry, and sampled from iconostasis, but with different amplitudes depending on the degree of alteration; see

Table 1. The assignments of FTIR bands used to differentiate different types of lime wood as well as to evidence the degree of degradation of the iconostasis wooden material. Wave numbers are expressed in cm⁻¹.

Assignments	Reference Lime Wood	Dried Wood	Iconostasis Wood	Remarks
$\nu$O-H	3337 i	3345 m	3338 w	Cellulose chain damage
$\gamma$C-H aliphatic	2900–2800 sh	2900–2800 w large	2900–2800 vw	Idem
${\gamma }_{asim}$ C-O-C cellulose and hemicellulose	1033 i–1160 sh	1033 m–1160 sh	1033 w–1160 sh	Decreasing the amount of cellulose
$\nu$C=O unconjugated hemicellulose	1735	1737	1737 wl	Loss of hemicellulose
$\omega$-CH2 –cellulose	1319 m	1319 w	1319 vw	Drastic decrease in cellulose and hemicellulose
$\nu$C=O embedded in aromatic skeleton	1593 w	1593 w	1593 vw	Damage to lignin aromatic skeleton
$\gamma$C=O acetyl and carboxyl groups of xylan	1503 m	1503 m	1503 vw	Damage to lignin aromatic matrix
$\nu$C=O lignin and acetyl in carboxylic vibration of xylan	1234 m	1234 w	1234 vw	Damage to lignin skeleton
${\delta }_{asym}$ C-H aromatic skeleton of lignin, splitting bands	1456 m; 1423 m	1456 w; 1423 w	1456 vw; 1423 vw	Decrease in lignin amount
$\delta$H-O-H of absorbed water	1624 m	1624 w	absent	Decrease in hydration degree

i—intense, m—medium, w—weak, sh—shoulder, vi—very intense, vw—very weak, wl—weak and large bands.

and the comments below.

According to [43], between 3600 and 2800 cm⁻¹, there are valence vibrations assigned to the cellulose skeleton corresponding to (i). $\nu$O-H, which is slightly moved from 3337 cm⁻¹ intense in fresh wood to 3345 cm⁻¹ medium in dried wood and to 3338 cm⁻¹ weak in iconostasis wood (Figure 7); (ii). aliphatic chains $\nu$C-H at 2900 cm⁻¹ medium with a shoulder at 2800 cm⁻¹ in fresh wood and 2850 cm⁻¹ weak and large in dried wood, while the 2800 cm⁻¹ line becomes very weak and flattened in the iconostasis (Figure 7). These changes could be attributed to water loss in the case of dried wood or through the action of the microclimate, where the poly-condensation reactions take place between the OH groups grafted on the cellulosic chains. Consequently, the orderly arrangement of the polymer chains is destroyed, so the smaller the number of OH groups, the lower the amplitude of corresponding infrared bands, which significantly diminishes in the case of iconostasis wood (Figure 7). All of these changes that denote significant damage to the cellulose chains of the decayed iconostasis wood are also evidenced in Table 1.

Moreover, between 1800 and 600 cm⁻¹, we have observed a similar behavior of more specific lines attributed to (i). $\nu$C=O in unconjugated hemicelluloses at 1735 cm⁻¹; (ii). the $\delta$H-O-H line of absorbed water molecules on the cellulose chains at 1624 cm⁻¹ is completely absent in the case of iconostasis wood, suggesting a drastic dehydration compared with the dried wood; (iii). $\nu$C=O embedded in the aromatic skeleton of the lignin line at 1593 cm⁻¹ is significantly attenuated in the case of iconostasis wood, most probably related to a decrease in the lignin amount; (iv). the valence vibration assigned to $\omega$-CH– cellulose at 1319 cm⁻¹ shows the same tendency of a major decrease in the cellulose and hemicelluloses amount; (v). $\nu$C=O lignin belongs to acetyl in the carboxylic moiety of xylan at 1234 cm⁻¹, testifying also the degradation of the lignin skeleton (Table 1).

The significant loss of cellulose and hemicellulose and the deterioration of the lignin matrix, as well as the presence of black stains, could be related to bacteria and mold attacks, which feed themselves by the consumption of cellulose and lignin, respectively [37]. For this reason, it is most plausible that the released metabolites are responsible for the gray and black stains, which can be remarked on in the wood samples illustrated in Figure 3a and Figure 5c.

In this regard, the FTIR spectra of the iconostasis wood (Figure 7) did not show the presence of infrared bands corresponding to $\nu$C=O carbonyl compounds, as well as those of $\nu$C-OH and $\nu$COOH, all of them properly corresponding to carboxylic acids related to the oxidative aging decaying product reactions according to Equations (2) and (3).

Complementary to FTIR, SEM furnished high-resolution images of wood fibers and their alteration, such as the loss of cell wall components or their decay (Figure 8), as well as traces of foreign material, most possibly fragments of gesso due to former restorations (Figure 8a,c). The loss of woody material is visible through the deterioration of the regular structure and the frayed appearance of the cell walls. Due to the microclimate conditions, the biologic attack was, according to literature data [42], most likely produced by soft rot fungi such as Chaetomium sp., Xylaria sp., Alternaria sp., or Humicola sp., recognized as cellulosolithic and ligninolytic mold. Their exoenzyme activities can cause more serious damage by decomposition of the polymeric fibers, or manifesting, as mentioned before, as embrittlement, or by the presence of colored fruiting bodies and the production of colored metabolites as illustrated in Figure 4a and Figure 5c.

These considerations, in the absence of corresponding images that could help to identify fungus genera, should be considered, until future SEM investigations, as hypothetical.

Another issue of the wooden support consists of the presence of greenish stains on the fragment illustrated in Figure 4a. Initially, the stain was considered as produced by fungus metabolites; therefore, we investigated by ATR-FTIR spectroscopy the greenish area on the wooden fragment illustrated in the Figure 4a sample, and an infrared spectrum (Figure 9) typical for a mixture of organic and inorganic compounds was obtained. As can be seen, a large intense band lies between 3500 and 3000 cm⁻¹, corresponding to stretching vibration $\nu$O-H in inorganic hydroxides, and there is a very narrow intense peak at 1677 cm⁻¹ corresponding to $\nu$-COO- in copper acetate moiety [43], while a split medium shoulder at 2982 cm⁻¹ is characteristic of the stretching vibration $\nu$C-H in aliphatic hydrocarbon chains. The other bands lie between 1500 and 600 cm⁻¹ and correspond to different interactions of organic moieties with inorganic components.

In this case, the most plausible explanation of the presence of the $\nu$-COO- band can be related to the use of copper acetate Cu(CH₃COO)₂·Cu(OH)₂·5H₂O as an artisanal antifungal agent during the previous interventions. The copper acetate was most probably suspended in an oily aliphatic binder that embedded the wooden fibers. This hypothesis is in good agreement with the presence in the FTIR spectrum of a split medium shoulder at 2982 cm⁻¹ (Figure 9) characteristic of the stretching vibration $\nu$C-H in aliphatic hydrocarbon chains [42,43].

Further, the presence of copper acetate can partially explain the corresponding XRF spectrum (Figure 10a), which evidenced intense Cu ${\mathrm{K}}_{\alpha }$ and ${\mathrm{K}}_{\beta }$ together with less intense ${\mathrm{K}}_{\alpha }$ and ${\mathrm{K}}_{\beta }$ Zn lines, the last one due most probably to the use of brass powder dye [44,45] as a cheaper substituent of gold foils during early 20th century iconostasis restorations, when brass or bronze powder dyes were used [46].

Another aspect related to the wooden surfaces, gilded or not, consists of the presence of an appreciable amount of particulate matter of which its composition had to be taken into account, using the appropriate solvents in removing them. Known as superficial dirt or grime, refs. [47,48], it represents an accumulation of particulate matter from the atmosphere as a complex mixture of water-soluble inorganic salts and insoluble mineral dust and carbonaceous material consisting of soluble and insoluble organics as a greasy and/or oily component, as well as elemental carbon embedded on the surface [49,50].

Given the church interior microclimate as well as the religious ritual services, which imply burning candles and incense, sources of volatile organic polar compounds and/or carbonaceous particles, the ATR-FTIR technique showed to be one of the most appropriate. In this regard, the FTIR spectrum of the thick blackened layer on the surface of the wooden fragment illustrated in Figure 10a,b is reproduced in Figure 11 [50].

In this case, the sharp medium band at 3622 cm⁻¹ corresponds to a stretching frequency of $\nu$O-H, the large band at 3300 cm⁻¹ corresponds to $\nu$N-H in aliphatic amine, while the stretching valence vibration $\nu$C-H of aliphatic hydrocarbon chains appears as medium split bands at 2958 cm⁻¹, 2917 cm⁻¹, and 2849 cm⁻¹. In turn, the aldehyde and ketone spectrum showed a shoulder of the carbonyl moiety $\nu$C=O at the characteristic wave number 1795 cm⁻¹, while the carbonyl frequency $\nu$C=O in carboxylic functions of organic acids R-COOH appears as a very intense band at 1675 cm⁻¹ [43].

The secondary amines, presenting a very intense band of stretching valence $\nu$N-H at 1342 cm⁻¹, were evidenced too, while the unsaturated hydrocarbon chains were evidenced by a medium band at 916 cm⁻¹ proportional to $\delta$C=C as well as at 813 cm⁻¹ [43].

All of these bands corresponding to the moieties belonging to aliphatic chains of amines, carboxylic acids, and ketones/aldehydes sustain the hypothesis that the main source of evidenced particulate mater consists of a mixture of organic compounds resulting from burning candles and incense, typical for the church interior atmosphere.

3.3. Gilded Layer

According to the historical tradition [51], a gilded wood surface means a structure of several layers successively applied on a wooden surface, such as the following: (i). the preparative layer known as white gesso covers the first wooden support; (ii). Armenian bole silicate or red gesso is the second layer; (iii). on the surface is applied gilded leaf or colloidal gold suspended in rabbit skin glue.

The white gesso is based on a mixture of chalk, lead white, and/or white zinc after the last decades of the 19th century and gypsum according to the historical documentation, which, in chemical terms, means CaC$O_3$, 2Pb(C$O_3$)₂·Pb(OH)₂, ZnO, and CaSO₄. These components are dispersed in a binder, usually rabbit skin glue, a natural protein compound that contains different amino-acids, such as glycine, proline, hydroxyproline, or histidine, corresponding to the general formula -NH-CHR-CO-.

Concerning the white gesso, its ATR-FTIR spectrum reproduced in Figure 12 shows bands belonging to more inorganic and organic moieties:

(i). An inorganic hydroxide OH⁻ band that appears as a narrow left shoulder of the broad band between 3000 and 3400 cm⁻¹;

(ii). The shoulder at 1791 cm⁻¹ corresponds to $\nu$C=O in phospholipids;

(iii). The group of bands correspond to the amide skeleton amide I 80% $\nu$C in the (C=ONH) moiety as a more intense shoulder at 1662 cm⁻¹ and amide II 60% $\delta$NH₂ and a 40% CN intense band at 1531 cm⁻¹;

(iv). A weak and broad shoulder at about 1400 cm⁻¹ that could be attributed to the ${\nu }_{asim}$${CO}_3^{2-}$ anion;

(v). A medium band at 1160 cm⁻¹ assigned to $\nu$C-O in R-COO-R′ in nucleic acids [43].

(vi). The medium bands at 883 cm⁻¹ and 717 cm⁻¹ denote the presence of $\delta$O-C-O in the carbonate anion;

(vii). The weak bands at 530 cm⁻¹ and 440 cm⁻¹ correspond to the bond valence vibration of Zn-O in zinc oxide and characteristic bands belong to the organic components;

It should be remarked that the 883 cm⁻¹ and 717 cm⁻¹ bands are shifted by 10 cm⁻¹ or less with respect to pure calcite [52,53,54] but correspond to dolomite as reported in [55]. In our opinion, this peculiarity could be explained by the existence of significant deposits of dolomite and dolomitic marble in Northeastern Greece, from where the Athonite craftsman could obtain it and use it in the preparation of white gesso [56,57].

As dolomite limestone is a mineral, its presence in work of arts could be used in provenance studies, especially in the case of Athonite manufactures.

It is worth mentioning [58] that the characteristic intense band of the tetrahedral sulphate anion vibration mode, usually slightly split in anhydrite, is absent in the ATR-FTIR spectrum of white gesso, suggesting that, in our case, the white gesso did not contain gypsum. This hypothesis is also sustained by the absence of Sr lines in both the white gesso and red XRF spectrum (Figure 10), Sr being a general Ca companion in gypsum [59].

The elemental composition of white gesso as evidenced by the XRF spectrum (Figure 10) shows the presence of Ca, Cu, Zn, Pb, and traces of Au. At the same time, the corresponding ATR-FTIR spectrum suggests a mixture of chalk CaCO₃ with small amounts of white lead 2PbCO₃·Pb(OH)₂ and ZnO, as the $\nu$O-H in hydroxides and the $\nu$Zn-O are present (Figure 12).

At the same time, the Cu and Zn lines in the white gesso XRF spectrum could be attributed to brass powder dye used during previous low-quality restoration. On the other hand, the presence in the XRF spectrum of the ${\mathrm{L}}_{\alpha }$ line of Au may be related to the initial gilded surface, covered, as mentioned, with a layer of low-quality brass dye.

The ATR-FTIR spectrum of the red gesso appears as quite different (Figure 13) and is characterized by the same narrow left shoulder of the broad band between 3000 and 3400 cm⁻¹ due to the presence of water (also not structural) as in the infrared spectrum of white gesso, while, regarding organic moieties, 1638 cm⁻¹ and 1457 cm⁻¹ correspond to amide I $\nu$CO in (CO-NH moiety) and to amide II ($\delta$NH and $\nu$CN), respectively, proving an interaction with Fe²⁺ cations.

The tetrahedral silicate moiety (SiO₄) Si-O is characterized by four characteristic frequencies at 1150 cm⁻¹, 903 cm⁻¹, 874 cm⁻¹, and 810 cm⁻¹, slightly shifted as a consequence of being neighbors within the crystaline lattice [60]. At the same time, the three prominent peaks at 607 cm⁻¹, 600 cm⁻¹, and 533 cm⁻¹ could be assigned to the Feⁿ⁺-O valence vibration mode in the organic moiety (Figure 13) [61].

The presence of red gesso, which contains a significant amount of Armenian bole, a silicate with the general formula of Ca_m(Fe²⁺/Fe²⁺,Ti)_n(SiO₄)(OH)_r, is confirmed by the Ca, Fe, and Ti lines in its XRF spectrum (Figure 10a), as well as by the characteristic bands of the Si-O and Fe²⁺-O in its ATR-FTIR spectrum (Figure 13). The simultaneous presence of these lines is due to the tetrahedral SiO₄ ordered in a crystalline lattice that allows for Fe-O interaction and the inclusion of a hydroxyl anion and Fe²⁺ and Fe³⁺ cations for charge balance.

Moreover, the band between 1400 and 1500 cm⁻¹ appears to be high due only to amide II (see Figure 12 for comparison), so it is most probable that the presence of calcite as a secondary mineral in the Armenian bolus, a kaolinite clay mineral rich in iron oxides, could explain the proportional inverse intensities of amide I and amide II in this sample.

Also, the stretching vibration of $\nu$C-O in the R-COO-R of nucleic acids at 1160 cm⁻¹, usually evidenced by a narrow band in the white gesso spectrum (Figure 12), becomes large, slightly shifted, and split in a weak shoulder at 1015 cm⁻¹ in the case of red gesso (Figure 13). In our opinion, most plausibly, these peculiarities are due to different interactions between the organic binder and metallic cations, Fe, and, to a lower extent, the Zn, which is worth a detailed investigation, as well as the influence of Armenian bolus secondary minerals on the red gesso FTIR spectrum.

The small hydrolytic processes, favored by the wet church microclimate, permit the amino acid molecules to react with the existing free cations Ca²⁺, Pb²⁺, Fe²⁺, and Zn²⁺, giving simple salts [62] or coordinative compounds [63] according to the proposed reactions:

(7)

The reaction is generally possible due to the decrease in the protean skeleton helicity simultaneous with an increase in anti-parallel $\beta$ sheets [64]. These kinds of interactions explain the alteration in the matrix, cracks, and loss of components of the gilded layers evidenced in Figure 2, Figure 4, and Figure 5.

On the other hand, it can be noted that the ATR-FTIR bands of white gesso (Figure 12) and red gesso (Figure 13), which correspond to amide I and amide II, although different in shape and resolution, prove the presence of protein material and, indirectly, the use of an organic binder, most likely extracted from rabbit skin [63].

4. Conclusions

Assessing the state of degradation of “Imperial Doors” means a cooperation between restorers, art historians, chemists, physicists, and biologists. Through the investigation and monitoring of the Church’s microclimate, the aesthetic changes could be attributed to aggressive agents and chemical transformation, including the contribution of particulate matter.

The use of several non-invasive analytical techniques, such as optical microscopy (OM), attenuated total reflectance–Fourier transform infrared spectroscopy (ATR-FTIR), X-ray fluorescence (XRF), invasive environment electron microscopy scanning (SEM), or Fourier transform infrared spectroscopy (FTIR), makes it possible not only to assess the degradation of the Imperial wooden Doors of the Church of the Ascension in the 19th century but also to highlight the chemical alteration induced by the microclimate of the church.

It was possible to confirm that the iconostasis was made of linden wood according to local Athonite tradition, while the original surface was partially covered with brass powder paint during previous unsuccessful restorations.

The influence of the microclimate of the church was highlighted not only by the visual examination of wood damages, such as the abrasion of painted layers, loss of gold coating, or cracks of the gesso, but also at the molecular level, by highlighting the degradation of both lignin and cellulose, mainly through a biological attack.

At the same time, as the ATR-FTIR investigation has shown, an important role in the damage to the gilded surface was played by the internal atmosphere of the church, an atmosphere rich in organic residues due to the burning of oil lamps, candles, and other aromatic substances during religious services.

The issue of the consolidation and preservation of works of art is of paramount importance with regard to the protection and preservation of cultural heritage, and the data provided by this study are available on demand for any restorers, which could be useful for the good management of future interventions and restorations of works of religious art.

The identification of dolomite limestone as one of the main components of the white gesso could be regarded as a characteristic of Athonite religious art workshops.