Non-Destructive In Situ Investigation of the Study of a Medieval Copper Alloy Door in Canosa di Puglia (Southern Italy)

Abstract

This paper reports the analyses carried out on the medieval copper alloy door (1111–1118 AD) of the mausoleum of Boemondo d’Altavilla in Canosa di Puglia (Southern Italy). The studied door is the smallest medieval bronze door extant in Italy and, unlike the other Byzantine doors, was most probably made in Canosa di Puglia and not in Constantinople. Analyses were performed to assess the chemical composition of the alloy patinas using a portable energy dispersive X-ray fluorescence (ED-XRF) instrument designed at the University of Salento. The experimental results suggested that the two door leaves have the same chemical composition, even if they appear different in both style and size. Furthermore, the alloy used for the door is different from the other previously-analyzed Byzantine bronze doors. The obtained results can be used in the future to compare the chemical composition of other Byzantine doors in order to better understand the manufacture of these precious artifacts.

1. Introduction

Nowadays, only about thirty medieval doors made of copper alloy remain to be admired in Europe. Unfortunately, many of the doors have been destroyed due to corrosive processes caused by atmospheric agents, and even more so to the reuse of the precious metal of which they are made for other purposes, e.g., the construction of weapons [1,2,3].

Most of the bronze doors dating from the Ancient period belong to the last centuries of the Western Roman Empire. These doors are characterized by the simplicity of the decorative plant or animal motifs that adorn them. Clear examples of this type of door are the bronze doors of Hadrian’s Pantheon, the door of Divine Romulus’ temple, and the valves with acanthus spirals of the Curia Julia.

In Italy, after the year 1000 between the second half of the 11th and the first half of the 12th century there was an increase in the use of doors made of copper alloy. These masterpieces were usually manufactured in Constantinople and then exported to Italy, or made in Italian workshops imitating the original Byzantine doors.

The oldest bronze door in Italy is the one in Amalfi Cathedral (1060 AD). In addition to this door, seven bronze doors were installed in Italy over a period of about fifty years.

Most of these doors are found in central-southern Italy: in Monte Cassino (1066 AD) [4,5], Rome (1070 AD) [6], Monte Sant’Angelo (1076 AD) [7], Atrani (1087 AD), Salerno (1085 AD), Benevento (1150–1151 AD), Canosa di Puglia (1111–1118 AD), Trani (1111–1118 AD), Troia (1119–1127 AD) and Monreale (1140 AD). Only two bronze doors, the one in St Clement’s (ca. 1080 AD) and the one in the atrium of St Mark’s Church (ca. 1112 AD), are in northern Italy (Venice).

Despite being known as “bronze doors”, Byzantine doors are made of orichalcum, a particular type of brass consisting of a copper alloy with a low zinc content [8,9]. These doors were probably called “bronze” because, over time, brass becomes green, much as bronze does.

From a structural point of view, Byzantine doors are made of small modular units of orichalcum called panels, fixed with nails to a wooden core. This method of construction made possible the fabrication of doors that were light and manageable and therefore easy to transport.

The most widely used technique for casting bronze (or brass) in the Middle Ages was the “lost wax” process. First, a chalk model of the object to be created was made and then the whole piece was covered with wax. Later, this kind of mould, made of chalk and wax, was covered with refractory soil and baked in an oven. The temperature caused the melting of the wax (hence the term “lost wax”) and the hardening of the refractory soil coating. The small cavity left by the melted wax was filled by the molten metal, which assumed the hollow shape of the desired artefact. This technique was very suitable for single part doors, although expensive and laborious for multi-part doors such as the bronze ones placed in Italy.

According to the studies of Angelucci, it is possible to assume that Byzantine door leaves were made using the so-called “sand” or “moulded” casting method [10,11]. This technique differs from the previously-used one as it did not involve the use of wax. To prepare the moulds of the piece to be made, a frame called a casting flask was used. At its center was placed the model, generally made of wood, having the shape of the piece to be cast. Sand was pressed around the model to create a cavity in which, once the model was removed, the casting could take place. This technique allowed the production of numerous replicas of the same piece without having to prepare a wax mould each time, an advantage that met the needs of Byzantine door manufacture.

Byzantine doors were decorated with precious goldsmith’s work, made using the techniques of inlay and niello [12,13,14,15]. From the iconographic point of view, these doors contain representations of Christ, the Virgin Mary, the Archangels, or the titular saints of the church. The role attributed to them is that of intercessors to God for the entrance of the faithful into the Kingdom of Heaven [1,16].

It has been important to highlight that there has been very little scientific research concerning the chemical composition of these precious artifacts [7,11,17,18,19,20,21,22,23].

The aim of this work is the non-destructive and in situ analysis of the medieval door of the mausoleum of Boemondo d’Altavilla in Canosa di Puglia (Southern Italy) using portable energy dispersive X-ray fluorescence (ED-XRF) to evaluate the chemical composition of the alloy patinas.

The obtained experimental results may help to increase the knowledge about the chemical composition of the medieval Byzantine “bronze doors” and could be used, in the future, to determine their complex production methods, which are still shrouded in mystery.

2. Materials and Methods

2.1. Description of the Analyzed Door

The investigated copper alloy door is placed at the entrance of the mausoleum of Boemondo I of Altavilla, located in the city of Canosa di Puglia (Southern Italy).

The sepulchral monument contains the mortal remains of a Norman leader of the First Crusade, Marco Boemondo I of Altavilla, Prince of Antioch, who died in 1111 AD. It is an exceptional and unique example in the West of Islamic or Syrian-inspired architecture, reproducing the Holy Sepulchre in Jerusalem. The mausoleum was commissioned by Marcus Bohemond’s mother, Alberada, and is located next to the Cathedral of St Sabino.

From an architectural point of view, this monumental tomb is characterized by an octagonal drum decorated with small columns and marbles [24]. The entrance door has a total area of 2.31 m² and appears asymmetrical, as it consists of two leaves of different sizes. Figure 1 shows a picture of the door with measurement points.

The left leaf is 200.5 cm high, 58.5 cm wide and has a thickness ranging from 1.5 cm to 7.5 cm, with a total weight of 317 Kg. The right leaf, on the other hand, is 202.7 cm high, 56 cm wide and has a thickness ranging from 1.5 cm to 5.9 cm, with a total weight of 168.5 Kg.

Although it recalls early medieval Byzantine manufacture, the door was most probably not made in Constantinople but rather in Canosa di Puglia [21]. Restored in 1914 and again in 2001, it is the smallest of the medieval bronze doors in Italy and the only one made of copper alloy and decorated on both sides.

The two leaves, probably made of reused materials, are different not only in size but in style and manufacturing process as well [11,25].

The left leaf is a single piece made with the sand-casting technique, and shows a framing with slightly protruding geometric designs.

The right leaf, on the other hand, is composed by four tiles with the following dimensions: 49.7 cm × 56.5 cm, 51.3 cm × 56 cm, 52.2 cm × 56 cm, 49.5 cm × 55.5 cm. It is reasonable to assume that the four panels were made using the same technique employed for the left leaf, this time using a double case as the tiles have decorations obtained during casting on both the front and the back.

The iconography of the bronze door blends Byzantine stylistic features with oriental elements. The motifs on the two leaves are so different that, in order to harmonise the composition, Ruggero da Melfi, the craftsman who made the door, created a sort of vegetal frame that was identical for both leaves.

The left leaf shows three rosettes decorated with geometric shapes. In the upper one there was a relief, no longer extant, representing a Madonna and Child. Only an inscription has survived (Maria mater Domini Ihesus filius Marie), along with the non-passing holes that fixed the relief.

The central rose window has a lion protome in the centre, while the lower one is decorated with a complex six-petalled flower [26].

The left leaf has two fractures, the first in the lower area, which has been roughly patched in bronze (Figure 2a), and the second in the middle area, corresponding to the lion protome (Figure 2b).

Figure 3a shows a picture of the upper rose window of the left leaf, Figure 3b the right leaf, Figure 3c the lower rose window of the left leaf, and Figure 3d the right leaf.

Figure 4a shows pictures of the second panel of the right door leaf and Figure 4b shows a picture of the third panel of the same leaf.

The reverse, with its rough surface indicative of sand casting, shows three longitudinal modern crossbars, which assemble the three fragments that make up the left wing.

There are sixteen verses engraved with a burin. Above the upper rose window are these six verses:

• Unde boat mundus, quanti fuerit Boamundus:
• Graecia testatur, Syria dunumerat.
• Hanc expugnavit, illam protexit ab hoste;
• hinc rident Graeci, Syria, damna tua.
• Quod Graecus ridet, quod Syrus luget, uterque
• iuste, vera tibi sit, Boamunde, salus.

Immediately below the upper rose window, engraved by burin with a smaller font size than above, are two pairs of couplets:

• Vicit opes regum Boamundus opesque potentum
• et meruit dici nomine iure suo
• intonuit terris. Cui cum succumberet orbis,
• non hominem possum dicere, nolo deum.

Then, leaving a few centimetres of space from the previous inscription, follow four more verses, this time hexameters, with characters of a similar size to the previous four:

• Qui vivens studuit, ut pro Christo moreretur,
• promeruit, quod ei morienti vita daretur.
• Hoc ergo Christi clementia conferat isti,
• militet ut celis suus hic adleta fidelis.

Finally, under the central rose window, two more hexameters are engraved with a larger graphic form:

• Intrans cerne fores; videas, quid scribitur; ores,
• ut celo detur Boamundus ibique locetur.

All of the verses are engraved in such a way as to occupy an entire line of writing, however, as we have said, the graphic modules appear to be of different sizes, probably to facilitate reading for those entering the mausoleum. A possible explanation of the sixteen inscriptions could be the following:

• For this reason, the world resounds of who Boemondo was:
• Greece attests it, Syria enumerates it.
• He conquered this one, he protected that one from the enemy;
• So the Greeks laugh at your damage, O Syria.
• What the Greek laughs at, what the Syrian mourns,
• the one and the other rightly are for you true salvation, O Boemondo.
• Boemondo overcame the power of kings and the work of the mighty men
• and rightly deserved that from his name it was said that he thundered in the lands.
• And, since the world succumbed to him,
• I cannot call him man, but I will not call him God.
• Whoever living endeavoured to die for Christ,
• deserved that the one who was dying should be given life.
• Therefore, may the clemency of Christ grant him this,
• that his faithful athlete may be a soldier in heaven.
• You who enter observe the door;
• see what is written;
• pray that Boemondo may be given to heaven and placed there.

The right leaf, as previously mentioned, consists of four tiles. Those at the top and bottom are decorated with rosettes with more elaborate motifs than those on the left leaf. The central panels, on the other hand, show silver ageminature. In particular, the top-central tile depicts Boemondo I and his half-brother Ruggero Borsa, who long disputed the inheritance of their father Roberto il Guiscardo, kneeling in front of a crucifix, now lost.

Although for several years they were rivals in life, they are depicted in the tile as reconciled in death. A message of peace, which continues in the centre-bottom tile where the heirs of Boemondo I and Ruggero Borsa (Boemondo II and Guglielmo II) are shown holding hands with their uncle Tancredi of Altavilla, probably as a sign of reconciliation between the families.

As regards the dating of the door, it must be considered that Boemondo I died, as did Ruggero Borsa, in early 1111 AD. This suggests that the door was probably commissioned not long after the death of the brothers.

The central tiles of the right leaf show hammered burin engravings similar to those on the left one;, these were probably to hold a niello decoration [11]. In the lower panel of the right leaf, between the upper frame and the rose window, there is an inscription:

• Sancti Sabini Canusii Rogerius Melfie campanarum fecit has ianuas et candelabrum

This reveals the name of the craftsman who made the door, Ruggero da Melfi. According to some experts, this inscription may also suggest that the door was not made for Boemondo’s Mausoleum, but for the adjacent cathedral of St. Sabino [21].

2.2. ED-XRF Analysis

The analysis was performed by using an energy dispersive X-ray fluorescence (ED-XRF) portable instrument designed at the University of Salento [27,28].

It was composed of an X-ray tube produced by MOXTEK^® with a Pd-anode operating at 1–40 kV of voltage and 0–100 µA of current, a Si-PIN detector produced by AMPTEK^® model XR_100CR, thermoelectrically cooled, with a beryllium window of 25 µm. It had a resolution of 150 eV at 5.9 keV and a pocket multi-channel analyzer produced by AMPTEK^® model MCA8000A, interfaced with a laptop. The diameter of the X-ray beam was an ellipse with axes equal to 2 mm and 3 mm.

Cu, Pb, Zn and Fe were quantitatively determined at high energy: 20 kV of voltage and 3 µA of current, with an acquisition time of 60 s. Sn was determined at low energy: 6 kV of voltage and 40 µA of current, with an acquisition time of 60 s.

The calibration was carried out by analyzing five standards consisting of CuO, CuCO₃, Cu/Sn (80/20% w/w), Pb₃O₄, Zn, Fe and SnO₂ in different ratios. The concentration of each standard was chosen according to the composition of the studied alloy.

All chemical compounds were of analytical grade and purchased from Sigma-Aldrich^® except the Cu/Sn (80/20% w/w) standard, which was purchased from Goodfellow^®.

Each standard was prepared by mixing the compounds in different weight percentages. In particular, the chemical compounds were weighted using an analytical balance KERN^® model ABT 100–5M, subsequently mixed and homogenized in an agate mortar for ten minutes, and finally compressed at 200 bar for ten minutes. The homogeneity of elements in the standard met the requirements for the ED-XRF quantitative analysis. Moreover, the samples analysed were assumed to have infinite thickness and the quantitative results expressed in terms of weight percentage (% wt).

Table 1. Chemical composition of the standards used for calibration in ED-XRF analysis.

Standard	Cu	Sn	Pb	Zn	Fe
	(% wt)
1	73.5	2.2	12.5	4.8	0.5
2	65.0	4.9	8.4	7.5	1.1
3	58.1	8.4	5.9	12.4	2.2
4	51.5	10.7	3.6	18.4	2.9
5	48.9	11.8	2.5	21.2	3.3
Standard deviation	0.5	0.3	0.5	0.3	0.2

shows the quantitative chemical composition of the five analyzed standards for ED-XRF analysis.

2.3. Statistical Analysis

Experimental results obtained by ED-XRF analysis were elaborated with multivariate statistical analysis [29,30,31] in order to identify possible correlations and/or differences existing among the samples analysed. In particular, the Statistica version 10 software package (StatSoft^® Inc., Tulsa, OK, USA) was used, applying the method of principal component analysis (PCA). This statistical method calculates the orthogonal linear combinations of the autoscaled variables by employing a correlation matrix based on the maximum variance criterion. Such linear combinations are called principal component scores, and the coefficients of the linear combinations are called loadings. The numerical loading value for each variable of a given principal component shows how much the variable has in common with that component. PCA can provide information about the similarities and groupings of the samples considered, and if a trend exists, it can evaluate the possibility of classifying the samples.

3. Results and Discussion

Fortunately, the precious artifact does not show serious deterioration phenomena thanks to the previous restoration works carried out. In addition, we performed the analyzes on homogeneous and regular areas, as the level of surface irregularity on an old manufact can be quite high and this is a critical point that must be carefully considered.

Table 2. Brief description of the analysed samples and their composition obtained by using ED-XRF.

Sample	Description	Cu	Sn	Pb	Zn	Fe	Trace
		(% wt)
R01	Alloy	73.3	7.5	8.7	<0.3	2.3
R02	panel	71.3	10.5	10.8	<0.3	<0.5
R03	alloy	72.2	9.1	9.0	<0.3	3.1
R04	alloy	73.3	7.3	8.0	<0.3	2.3
R05	alloy	71.9	8.8	6.7	<0.3	3.0	Ca
R06	panel, green area	71.0	7.7	8.8	<0.3	2.3	Cl
R07	panel	68.2	10.5	10.5	<0.3	<0.5
R08	panel, dark area	49.7	27.4	10.2	<0.3	<0.5
R09	lion protome	72.7	4.5	10.4	2.3	<0.5
R10	alloy	71.1	8.0	9.8	<0.3	1.3
R11	alloy	66.2	6.8	14.2	<0.3	1.1	Ca, Cl
R12	panel	69.2	10.5	10.2	<0.3	<0.5
R13	panel, green area	69.0	7.0	13.9	<0.3	1.1	Cl
R14	panel	70.1	10.4	10.0	<0.3	<0.5
R15	top door hinge	69.4	4.3	2.4	10.4	0.7
R16	lower door hinge	75.8	4.8	2.5	11.4	0.7
L01	alloy	67.2	6.6	18.1	<0.3	0.9
L02	upper rose window	68.4	7.3	9.6	<0.3	<0.5
L03	upper rose window	67.3	6.2	10.2	<0.3	<0.5
L04	handle (lion)	70.5	5.7	8.8	2.7	<0.5
L05	lion protome	63.8	8.8	20.9	<0.3	<0.5
L06	alloy	73.8	5.2	10.2	<0.3	2.5	Ca, Cl
L07	edge	61.4	11.1	16.2	<0.3	<0.5
L08	alloy	64.2	12.2	12.2	<0.3	<0.5
L09	alloy, red area	73.0	3.3	7.0	<0.3	1.0
L10	alloy, green area	66.5	12.2	8.0	<0.3	0.5	Ca, Cl
L11	alloy, green area	64.7	11.2	11.0	<0.3	1.1	Ca, Cl
L12	nail	2.0	5.2	3.5	<0.3	82.0
L13	closed hole	1.5	< 1.0	82.5	<0.3	<0.5
L14	nail	4.0	4.6	3.0	1.5	80.0
L15	alloy, near the fracture	46.8	16.8	17.4	7.8	1.8
L16	alloy, near the fracture	49.9	15.1	15.8	7.4	3.7
L17	lower door hinge	56.8	<1.0	14.3	1.6	<0.5
L18	top door hinge	76.5	3.7	2.7	13.9	0.7
L19	center of the fracture	5.3	48.5	8.9	<0.3	<0.5	Ca
	Standard deviation	0.5	0.3	0.5	0.3	0.2

summarizes the experimental results obtained by analysing sixteen samples on the right leaf and nineteen samples on the left leaf using the ED-XRF apparatus. The samples on the right leaf are indicated by the letter R, while the samples on the left leaf are indicated by the letter L.

The two leaves are both made of a bronze alloy with a comparable chemical composition. Indeed, the average concentrations of the right leaf are Cu (70.6% wt), Sn (8.7% wt), Pb (10.1% wt), Zn (<0.3% wt) and Fe (1.4% wt), while the average concentrations of the left leaf are Cu (68.6% wt), Sn (8.3% wt), Pb (10.4% wt), Zn (0.6% wt) and Fe (0.6% wt). These results allow us to affirm that the alloy used for the two door leaves is the same, although their style and dimensions appear different.

The analysed nails (L12 and L14) are mainly composed of Fe (80.0–82.0% wt), in addition to Cu (2.0–4.0% wt), Sn (4.6–5.2% wt), Pb (3.0–3.5% wt) and Zn (<0.3–1.5% wt). This proves that they were not original, but were added during subsequent restoration work.

Three door hinges (R15, R16 and L18) have similar composition, while the bottom door hinge L17 has a lower concentration of Cu (56.8% wt), Sn (<1.0% wt) and Zn (1.6% wt) and a higher concentration of Pb (14.3% wt). This last hinge was probably replaced during subsequent restoration work, as it is located near the lower fracture of the left door leaf.

The chemical composition of the three lion protomes (R09, L04 and L05) is different. In particular, sample L05 provided lower concentrations of Cu (63.8% wt) and Zn (<0.3% wt) and higher concentrations of Sn (8.8% wt) and Pb (20.9% wt).

The samples of alloy near the lower fracture of the left door leaf (L15 and L16) show different compositions than the door alloy. This confirms that the door has been restored and that the section added on the left bottom is not original.

Similarly, sample L19, located in the center of the central fracture of the left leaf, has a composition very different from the door alloy, with a high concentration of Sn (48.5% wt).

The four green patinas (R06, R13, L10 and L11) found visually did not show an appreciable variation in their composition in relation to the door alloy, although all showed traces of chlorine, certainly attributable to processes of chlorine deposition due to degradation phenomena and/or marine aerosols.

On the other hand, the dark patina (R08) provided a different composition compared to the door alloy, with a very low concentration of Cu (49.7% wt) and a higher concentration of Sn (27.4% wt) as well as Pb (10.2% wt), while Zn and Fe were both lower than the respective limits of detection.

The red patina (L09) has comparable concentration to the door alloy for Cu (73.0% wt), Pb (7.0% wt) and Zn (<0.3% wt), excepting only Sn content which is very low (3.3% wt).

The closed hole (L13) shows very high Pb concentration (82.5% wt), and was therefore probably added as a filler during previous restoration work.

Principal cluster analysis was applied to the obtained data to better highlight the similarity or compositional differences among the measurement points. Principal component analysis was applied on a data set of thirty-five cases (sixteen samples on the right leaf and nineteen samples on the left leaf) and three variables (concentration of Cu, Sn and Pb). Zn and Fe concentrations were not considered in the statistical treatment as they were low values or were often below the detection limit.

Then, three principal components were extracted covering 100% of the cumulative variance (48.28%, 37.67% and 14.05%, respectively). The loading of the variables on the first two principal components (Figure 5a) shows that Pb and Sn are the dominant variables for the positive values of the first principal component (Factor 1) and Cu is the dominant variable for the negative values of the first principal component. Moreover, Pb is the dominant variable for the positive values of the second principal component (Factor 2), whereas Sn is the dominant variable for the negative values.

The scatter plot of the scores for the first two principal components, PC1 and PC2 (Figure 5b), shows that most of the samples can be grouped into a cluster (indicated with a dashed line), excepting fourteen samples which are not of the alloy and which can be further separated. In fact, sample L13 (closed hole) has a positive value on both PC1 and PC2, and L19 (center of the fracture) has a positive value on PC1 and negative on PC2.

Furthermore, the nails (L12 and L14), the areas close to the fracture (L15 and L16), the lion protome (L05) and an area of the door edge (L07) all have a positive value on PC1, while the door hinges (L17, L18, R15 and R16) and the red area (L09) have negative values on the same principal component. A different chemical composition is present in the sample R08 (dark area), which has a positive value on PC1 and negative value on PC2.

Therefore, the statistical treatment of the analytical data made it possible to more easily determine the various measurement points. The synergy between the ED-XRF technique and statistical analysis has once again proven to be a valid methodological approach in the field of cultural heritage for highlighting differences and/or similarities among samples.

Furthermore, the mean concentrations of the studied alloy were compared with the values from other Byzantine doors obtained in previous research in order to understand similarities or differences among these alloys.

Table 3. Comparison between the alloy composition obtained in this work and those of other Byzantine doors.

Byzantine Door	Cu	Sn	Pb	Zn	Reference
	(% wt)
Amalfi (1060 AD)	62	Traces	19	17	[16]
Monte Cassino (1066 AD)	80	7	3	10	[11]
Rome, S. Paul’s outside the Walls (1070 AD)	73.5	0.12	8.48	17.9	[11]
Monte Sant’Angelo (1076 AD)	70.2	0.8	11.2	14.2	[7]
Venice, S. Clemente (1080 AD)	72.4	2.2	8.6	16.8	[11]
Salerno (1085 AD)	77	3.5	11	4.5	[16]
Venice, central door (1112 AD)	78	5	8	9	[11]
Canosa di Puglia (1111-1118 AD)	69.6	8.5	10.2	0.3	This work

shows a comparison between the composition of the door alloy analysed in this work and those of other seven Byzantine doors which can be admired in Italy: Amalfi Cathedral, Monte Cassino, S. Paul’s outside the Walls at Rome, Monte Sant’Angelo, S. Clemente at Venice, Salerno, and the central door at Venice.

It is evident that the chemical composition of the alloy of the studied door is very different from that of the other seven doors. In particular, the main difference is found in the percentage of Zn used, which is much lower than that of the other Byzantine doors. We hypothesise that this result is due to the fact that, unlike the other doors, the Canosa door was made in Apulia and not in Constantinople.

4. Conclusions

This study allowed us to determine the chemical composition of the alloy used to craft the medieval door of Canosa di Puglia (1111–1118 AD) through a series of non-destructive and in situ analyses performed using a portable energy dispersive X-ray fluorescence (ED-XRF) instrument.

The results of these investigations showed that the two door leaves, despite their different style and size, are made of the same ternary Cu-based alloy (Cu-Sn-Pb).

The statistical elaboration of the obtained experimental results also showed that some parts of the door, such as the nails, closed hole, fractures, lion protomes, door hinges, red areas and dark areas have a different chemical composition, probably because of subsequent restoration work.

In addition, it was verified that the alloy used to fabricate this door has a different composition compared to those of seven other Byzantine doors in Italy (Amalfi Cathedral, Monte Cassino, S. Paul’s outside the Walls at Rome, Monte Sant’Angelo, S. Clemente at Venice, Salerno and the central door at Venice).

In the future, the results obtained in the present work could be an important resource for the study of the production process of bronze doors. Nevertheless, it would be desirable to have data on the chemical composition of all the Byzantine doors made in Constantinople and placed in Europe (better known as “bronze doors”) in order to reconstruct their history and shed light on the manufacturing techniques used to make these invaluable masterpieces.