A) Chemical composition

One of the vitrifiers in the samples was SiO₂, 13 of these samples fell within the range of 55 to 65 wt.%, figures similar to those quantified by other authors in 12^th to the 18^th centuries glass ^{(Carmona, García-Heras, Gil & Villegas, 2005, pp. 251-258}; ^{Müller, Torge & Adam, 1994, pp. 45-48}; ^{Schalm, 2004, pp. 1647-1656}; ^{Schalm Janssen, Wouters & Caluwé, 2007, pp. 663-668)}, one of these is below the minimum value; five, slightly above the maximum; and one, significantly exceeds it (the latter is considered to be of more recent production). This indicates that the samples studied were made with original recipes between the 9^th and 18^th centuries that did not change over time. In relation to the presence of P₂O₅, considered vitrifying since the 20^th century ^{(Kreidl & Weyl, 1941, pp. 372-378)}, was found in 13 of the samples in concentrations between 0.04 and 0.72 wt.%, figures that do not agree with those of the 9^th and 18^th centuries glass, as they exceed the figures of 2.1 wt.%. Sample 7 contained 2.8 and 18, 6.0 wt.%. This discrepancy in the content of P₂O₅ could reflect the type of raw material used or the type of their production site; therefore, it is ruled out that they were made using modern production processes.

Regarding the flux material, 19 samples correspond to the so dium glass variant and one-sample 22-, to the potassium variant, while the high concentration of K₂O of the latter suggests that it corresponds to an object manufactured in the 18^th century. Most likely, dry oak, poplar, pine, birch, and ashes were used as raw materials ^{(Peralta, 2018, pp. 3-29)}, a primitive recipe of European origin. It is considered that most of the sodium-type pieces used ashes from the Salsola tragus, known as barrilla in Spanish (a plant of the Chenopodiaceae family that grows in salty soils) or tequesquite (from Nahuatl tequixquitl) ^{(Peralta, 2018, pp. 3-29)}, material obtained from the vicinity of Lake Texcoco ^{(Tylor, 1861, pp. 129-161)} (there was also a glass factory near the lake since mid-18^th century), or sodium carbonate, obtained by the Solvay process, used since 1861.

The use of wood ash and other components to produce glass was made in some European regions, while the use of barrilla plants, common in Eastern cultures since Mesopotamian times, was spreading ^{(Tait, 2004, pp. 78-79)}. With the latter, the melting temperature was reduced; however, the generated vitreous mass was more viscous, and the pieces produced were more fragile, but with potassium oxide, the glass acquired greater brightness ^{(Fernández, 2002, pp. 315-330)}. The use of the barrilla was gradual since this knowledge was reserved, and it was not until the early 19^th century that its use was replaced by the Leblanc system, and this, in turn, by the Solvay. Meanwhile, some glassmakers used the K₂O present in the ashes of shrubs.

The considered stabilizing components contributed to the reinforcement of the molecular structure generated by the alkaline ions (Na and K). In relation to CaO, it appears to have a higher concentration in samples 21 and 22, which is consistent with what is recognized in glass from the 9^th and 18^th centuries originating in Germany, France, and Spain ^{(Carmona, García-Heras, Gil & Villegas, 2005, pp. 251-258}; ^{Müller, Torge & Adam, 1994, pp. 45-48}; ^{Schalm, Janssens, Wouters & Caluwé, 2007, pp. 663-668)}. On the other hand, in samples 6, 10, and 15 the concentration was below 1 wt.%, 14 of these presented it within the range of 1.48 to 7.13 wt.%, reduced values, but not identical to those identified in samples of the mentioned centuries ^{(Cagno, Janssens & Mendera, 2008, pp. 1389-1395}; ^{Lima, Medici, Pires de Matos & Verita, 2012, pp. 1238-1248}; ^{Marrocchino et al., 2020, pp. 819-827}; ^{Verita, Reiner & Zecchin, 2002, pp. 261-271}; ^{Wolf et. al, 2015, pp. 660-667)}; however, studies of Italian samples from the 13^th to 16^th and 16^th to 17^th centuries reported ranges lower than those from the 9^th to 18^th centuries, also of the same provenance ^{(Cagno, Mendera, Jeffries & Janssens, 2010, pp. 3030-3036}; ^{Raedt et al., 2002, pp. 1912-1917)}; only sample 5 lacks that oxide, so, it can be deduced that a more recent recipe was used for its elaboration. From the above, it can be inferred that the source of calcium extraction was different according to the artisan’s experience, incorporating a greater or lesser amount.

In relation to magnesium oxide, the values recognized in the fragments were greater than the unity, regardless of whether they were sodium or potassium glass; the same is true in Italian and European glass from the 9^th to the 18^th centuries ^{(Cagno, Janssens & Mendera, 2008, pp. 1389-1395}; ^{Cagno, Mendera, Jeffries & Janssens, 2010, pp. 3030-3036}; ^{Carmona, García-Heras, Gil & Villegas, 2005, pp. 251-258}; ^{Lima, Medici, Pires de Matos & Verita, 2012, pp. 1238-1248}; ^{Marrocchino et al., 2020, pp. 819-827}; ^{Müller, Torge & Adam, 1994, pp. 45-48}; ^{Schalm, Wouters & Caluwé, 2004, pp. 1647-1656}, ^{Schalm Janssens, Wouters & Caluwé, 2007, pp. 663-668)}. Exclusively, two previous studies of Italian glass from the 6^th to the 12^th centuries and from the late 13^th century report that they do not exceed the unity. On the other hand, the values of Al₂O₃ are remarkably high in soda glass, except in samples 9 and 19, which do not exceed unity in wt.%; 6, 13, and 22 presented it above 10 wt.%; and the remaining in a range of 1.5 to 6 wt.%. Most studies about this oxide from the 8^th to 18^th centuries show that it has a weight of between 2.0 and 7.0 grams. This is similar to what other studies have found, which usually have values of between 1.50 and 1.50 wt.% ^{(Barrera & Velde, 1989, pp. 48-54}; ^{Cagno, Janssens & Mendera, 2008, pp. 1389-1395}; ^{Carmona, García-Heras, Gil & Villegas, 2005, pp. 251-258}; ^{Marrocchino et al., 2020, pp. 819-827}; ^{Müller, Torge & Adam, 1994, pp. 45-48}; ^{Raedt et al., 2002, pp. 1912-1917}; ^{Schalm, Janssens, Wouters & Caluwé, 2007, pp. 663-668)}. The above could be because the raw material was of different provenance or, as with the Italian glass, was intentionally changed to produce glass of different aesthetic characteristics.

With respect to Fe₂O₃, it was absent in samples 5 and 6, 14 samples were between the range of 0.08 and 1.00 wt.%, 3 between 1.25 and 2.19 wt.%, and one sample with an elevated value of 23.52 wt.%. This last value is partly similar to what has been detected in Italian glass ^{(Cagno, Janssens & Mendera, 2008, pp. 1389-1395}; ^{Lima, Medici, Pires de Matos & Verita, 2012, pp. 1238-1248}; ^{Marroscchino, 2020, pp. 819-837}; ^{Raedt et al., 2002, pp. 1012-1017}; ^{Verita, Reiner & Zecchin, 2002, pp. 261-271)}, dated between the 7^th and 17^th centuries, and, to a lesser extent, in Belgian, Swiss, and Spanish glass from the 8^th to the 18^th centuries ^{(Carmona, García-Heras, Gil & Villegas, 2005, pp. 251-258}; ^{Schalm, Janssens, Wouters & Caluwé, 2007, pp. 663-668}; ^{Wolf et al., 2015, pp. 660-667)}.

The above indicates that, most likely, the difference lies in the concentration of other oxides that were not identified in the present study. The manganese detected in 13 samples was in the range of 0.01-0.20 wt.%, 4 between 0.21-1.00 wt.%, and it was absent in sample 5. The low range mentioned is similar to that detected in the Italian glass, and higher in the rest of the European countries discussed. Although it is considered that the presence of iron alone gives the glass its green coloration, this is relative since it will be found in different concentrations when reacting with oxygen. Generating other oxides, such as ferrous and ferric, which, depending on the proportion in which they are found, will give the glass a blue-green to green or yellow-green to yellow coloration. Manganese can provide violet coloration depending on other oxides present and not only act as a decolorant ^{(García, Gil, Carmona & Villegas, 2003a, pp. 21-34}; ^{Mirti, Davit & Gulmini, 2020, pp. 221-239)}. Other metals present in the sand can generate diverse tonalities. Therefore, it is necessary to expand the detection of more oxides to correlate them with the color of the glass or, otherwise, with the action of other metals with decolorizing functions.

B) Microscopic physical deterioration

Glass is apparently considered to be resistant to chemical agents; however, when subjected to environmental changes, such as humidity, temperature, pH, and microorganisms, its integrity is affected. Prolonged exposure to humidity causes the formation of a silica gel layer or film on its surface. The environment and the temper ature generate hydronium ions (H₃O⁺), and along with them, an acid medium that interacts with the alkaline ions (Na and K) of the glass matrix. The removal of the latter (de-alkalinization) determines the concentration of OH^- ions, which increases the pH and breaks the bond between silica and oxygen that forms the molecular lattice, or the cause of glass degradation. However, continued deterioration through an influx of heavy metals from the subsurface has an effect on the physical appearance ^{(García Rincón, Jimeno & Villegas, 2003b, pp. 173-81}; ^{García, Gil, Carmona & Villegas, 2003a, pp. 21-34}; ^{Silvestri, Molin & Salviulo, 2005, pp. 1338-1349)}. Previous observations have revealed notable morphological modifications attributed to microorganisms that exhibit affinity and adhesion to the glassy surface. This results in the formation of a biofilm comprising single cells, hyphae, and filaments, a polymeric extracellular material that, in conjunction with the slime, retains moisture and favors the release of metals. Consequently, the chemical degradation induced by acid and chelate production alters the pH, thereby favoring the persistence of microorganisms and the establishment of others ^{(Rölleke et al., 1999, pp. 107-114}; ^{Stockmann et al., 2012, 1-18)}.

With the naked eye, there were detected yellowish and bluish tonalities on the surface of the samples analyzed in this study, as well as a detachment of thin fractions in the form of flakes, light-brown adherence, and scratches. When small fractions were observed under the scanning electron microscope, manifestations in the relief of the surface were detected. Samples 5 and 6 presen ted irregular surfaces with an angulation pattern (Figure 3) and curvilinear shapes in the manner of puzzle pieces, an aspect that was not observed in the rest of the fragments. It is probable that the prominent variations in the chemical constituents detected in these samples have an impact on the molecular organization and, consequently, on the microscopic physical structure, as evidenced by the high percentage of SiO₂ in sample 5 and the hyalith glass fragment in sample 6.

(Micrographs: Sc. D. Arturo García Bórquez, 2020; courtesy: Escuela Superior de Física, IPN, Mexico)

Figure 3 Photomicrograph of sample 5 at 100X, corresponding to the base portion of a dark gray container with a decanter. Its high SiO₂ content gives it the molecular arrangement shown in this image, which does not resemble the rest of the samples analyzed. 

A portion of sample 7 classified as opal glass presented, due to its macroscopic characteristics, a rough surface very similar to a sample previously studied ^{(García, Rincón, Jimeno & Villegas, 2003b, pp. 173-181)}. Its authors attribute this characteristic to the high content of SiO₂ and CaO, which coincides to some extent with the values of SiO₂ in the samples analyzed in the present investigation, but not with those of CaO, which are much lower.

The rest of the samples, regardless of whether they are soda or calcium glass, show modifications in the relief of the surfaces caused by small punctiform excavations, some with regular edges as in sample 13 (Figure 4A) and others with irregular edges interposed between tortuous grooves, as in samples 10 and 19 (Figure 4B). Previous works have reported that soda glass is more res istant to the environment than potassium glass ^{(Bettembourg, 1976, pp. 36-42)}; however, this was not the case in the present study, which could be due to the characteristics of the medium and the time of exposure to it. Sample 11 showed areas with wide excavations of uniformly curved edges, which, presumably, originated from air bubbles that were the result of inadequately mixing the raw material, or of insufficient temperature for the melting of the glassy material.

(Micrographs: Sc. D. Arturo García Bórquez, 2020; courtesy: Escuela Superior de Física, IPN, Mexico)

Figure 4 Photomicrograph A, of sample 13 at 2000X, of a small portion of the neck of a bottle with low NaO₂ content exhibits surface deterioration that differs from that of image B, which is of sample 19, at 125X, of the neck and shoulder fraction of a container with much higher content of the referenced oxide. The difference in deterioration is probably dissimilar due to the time of exposure to moisture and subsoil components. 

In samples 4 and 20 (Figures 5A and 5B respectively), excavations occupied by polyhedral structures can be seen, which may be inorganic crystals of calcium, oxalates, oxides, sulphates or other components formed by the interaction of the environment. Likewise, on the smooth surfaces of samples 6 and 10, adhesion in the form of a bar is observed, and on sample 14, a spherical structure. In samples 10, 16, 17, and 18 (Figure 6) the adhesions are of irregular shape, so, it is assumed, they are likely of biological origin. Previous research on this matter does not report adhesions of this type; however, the photographic records do not present similarities to what was observed in the present investigation, so it is possible that the variations, result from the presence of different bacteria or other microorganisms that should be cultivated in similar conditions for their identification. The fragment of sample 9 presents lumpy and filamentous adherence on its surface, which could be the manifestation of a fungal colony (Figure 7), similar to that reported in previous research ^{(Andrejevna & Aldona, 1999, pp. 181-191)}.

(Micrographs: Sc. D. Arturo García Bórquez, 2020; courtesy: Escuela Superior de Física, IPN, Mexico, 2024)

Figure 5 Photomicrograph A is of sample 4 at 8000X, a portion of a liquorice vat of hemispherical shape with ample deterioration of its surface; it has a reduced content of SiO₂ and a higher content of Fe₂O₃ than sample 20, whose image B is at 500X and belongs to the liquor fraction. The incrustations on its surfaces are polyhedral. 

(Micrographs: Sc. D. Arturo García Bórquez, 2020; courtesy: Escuela Superior de Física, IPN, Mexico, 2024)

Figure 6 Micrographs A, B, C, and D, respectively, of samples 10 (bottom of liquor box), 16 (portion of knitting needle), 17 (fragment of container), and 18 (fraction of lens). They exhibit surface adhesions of particular morphology that could be of biological type; the first three images are 8000X and the last one is 16000X. 

(Micrographs: Sc. D. Arturo García Bórquez, 2020; courtesy: Escuela Superior de Física del IPN, Mexico, 2024)

Figure 7 Photomicrograph of sample 9 at 2000X, apparently part of the base of a cup and with high and K₂O content, showing hyphal growth. 

A particular characteristic was observed in samples 9, 19, 22, and 23: this was the straight laminar organization, while in samples 8, 18, and 20 they were concentric with a wide radius (Figures 8A and 8B). In the work of ^{Cox and Ford (1993, pp. 5637-5647)} this characteristic was reported in potassic glass, but was concentric with a short radius, so it is proposed that it is a consequence of moisture erosion in punctual zones by sodium, potassium, and magnesium leaching. However, the variants observed in this research could correspond to the area of the fragment of the piece of glass; that is, if the laminar organization is straight, it could be a piece of the wall of a bottle, and if the distribution is concentric, it could be a piece belonging to the area of the shoulder or lip of a bottle.

(Micrographs: Sc. D. Arturo García Bórquez, 2020; courtesy: Escuela Superior de Física del IPN, Mexico, 2024)

Figure 8 Micrographs of samples 8:A, which corresponds to the bottom of a circular decanter, and 18:B, which is the base of a cup, show that the first is sodic and the second is potassic with high SiO₂ content. In both images at 1000X, the molecular arrangement is evident. In image A, the straight arrangement is observed in the center and without cleavage, and in image B, the convex linear arrangement is the probable result of the shape of the zones of each one of the pieces. 

C) Description of fragments and color

The determination of the tonality of the samples was considered as a macroscopic particularity of reference, and not with the purpose of determining the oxide or oxides that generate it.