The identification of the stains on the Leclanché glass jar was successfully achieved by Guilhermina Cadeco et al., where the presence of manganese dioxide on the stained glass was verified. XRF analyses demonstrated that manganese was the most abundant element (~25%) and XPS confirmed the presence of MnO2 [26].
The chemical characterization of black-blue stains on granite at Santa Marinha, São Tiago and N. Sra. of Guadalupe churches were attained through XRF, FTIR and XPS analysis, demonstrated by the results reported hereinafter.
Characterization of black-blue stains in Vila Real churches (north of Portugal)
Unusual Mn-rich black-blue stains have been causing aesthetical changes on granite walls of Vila Real churches. These black-blue stains are mainly visible on granite support at all churches and in some areas of their frescoes. After an exhaustive searching of churches with Mn-rich stains, were found a typology of black-blue stains at Sta. Marinha (Vila Marim), St. Tiago (Folhadela) and Nossa Senhora de Guadalupe (Mouçós) churches, all located in Vila Real district (Fig. SI1).
Table 1. Iron and Manganese percentage of black-blue stains on granite and CIELab color coordinates. Un = unstained; St = stained; L* = lightness; a* = red/green value; b* = yellow/blue value
|
|
Elemental Analysis
|
CIELab coordinates
|
Churches
|
Granite
|
Fe (%)
|
Fest/Feun
|
Mn (%)
|
Mnst/Mnun
|
L*
|
a*
|
b*
|
Sta. Marinha
|
Un
|
1.01±0.19
|
0.96
|
0.08±0.02
|
18.73
|
|
|
|
St
|
0.97±0.11
|
1.50±0.77
|
23.86±1.03
|
3.36±0.34
|
3.90±1.20
|
St. Tiago
|
Un
|
0.72±0.21
|
0.85
|
0.08±0.04
|
17.25
|
|
|
|
St
|
0.61±0.30
|
1.38±0.53
|
21.16±1.39
|
3.48±0.47
|
3.07±0.37
|
N.S. Guadalupe
|
Un
|
0.79±0.31
|
1.31
|
0.05±0.03
|
23.35
|
|
|
|
St
|
1.04±0.20
|
1.17±0.11
|
24.04±2.40
|
3.59±0.74
|
4.15±1.98
|
Table 1 shows the similarity between Mn-rich stains found in three churches of Vila Real. These black-blue stains are also characterized by the presence of iron because its chemical similarity with manganese. In terms of elemental percentage, was uncovered that stained granite presents a manganese percentage 17.25 to 23.35 times higher than the unstained granite and the levels of iron are almost identical, resulting in Mn/Fe ratios of 0.08±0.02 and 1.64±0.57 for unstained and stained granite, respectively. Concerning to colorimetric characterization, the lightness of these stains is 23.02±1.61 being concordant with the visual evaluation that revealed an apparent brightness (common in biofilms). The parameters a* and b* revealed values of 3.48±0.12 and 3.70±0.57 being also plausible with stains ranging around black colorations.
Were performed analytical techniques at Sta. Marinha, St. Tiago and N.S. Guadalupe churches to unveil the chemical composition of these type of stains on granite. Black-blue stains were analysed by FTIR (Fig. SI2) and is observed the presence of symmetric and asymmetric stretching vibrations of νCH3 and νCH2 between 2920 and 2850 cm-1, and the stretching vibration of νOH at 3433±5 cm-1. The bands at 2360, 2337 and 1385 cm-1 are due to the absorption of atmospheric CO2. The range of bands between 1100 and 670 cm-1 may be ascribed to silicate vibrations (e.g. quartz, typically present on granite). The assignments between 528 and 420 cm-1 suggest the presence of Mn oxides, equivalent to Mn oxide coatings found on Lascaux cave (520, 470 and 420 cm-1) [5] and Paranhos spring water tunnel (529, 468 and 425 cm-1) [4].
These stains were also analysed with XPS on samples of the three churches (Fig. SI3 and Table SI1). The binding energies at 532.40 (O1s) (Fig. SI3e), 102.92 (Si2p3) (Fig. SI3g) and 74.38 eV (Al2p3) (Fig. SI3h) indicate the presence of silicates and aluminosilicates, concordant assignments with the granite substrate. The presence of organic elements is also shown at 284.55 (C1s) (Fig. SI3d) and 406.90 eV (N1s) (Fig. SI3f), corresponding to adventitious carbon and nitrates, respectively. The presence of halite (NaCl) at 1071.35 eV (Na1s) (Fig. SI3j) might be an evidence of the migration of some ions from groundwater, passing through the porous granite, and precipitating at the surface. The existence of manganese oxides in black-blue stains could be proved at 641.24 and 643.88 eV (Fig. SI3b), both assigned to Mn(II/III) (e.g. Mn3O4) and Mn (IV) (e.g. MnO2), respectively. The same occurs for iron oxides at 709.52 and 712.45 eV (Fig. SI3c) attributed to Fe (II/III) (e.g. Fe3O4) and Fe (III) (e.g. Fe2O3), respectively. The spectral interpretation for samples collected at S. Tiago and N.S. Guadalupe is similar and is summarized in Table SI1. It is also observed the presence of calcite, confirmed by the spin-dublet of Ca 2p region at 347.13 (Ca 2p1/2) and 350.65 eV (Ca 2p3/2), giving a chemical shift of Δ Ca 2p ~ 3.5 eV and an intensity peak ratio approximately 2:1, respectively (Fig. SI3i) [27]. The presence of calcite might indicate that some granite ashlars were whitewashed with lime, probably covered with frescoes, a common practice at that time.
Cleaning assays
Determination of soluble Mn by AAS
This study intended to ascertain the suitable reactant able to reduce insoluble Mn oxides (Mn4+) to soluble manganese oxidation states (Mn2+/3+). Thiourea Dioxide (TD), Hydroxylamine Hydrochloride (HH) and Hydroxymethanesulfinic Acid (HSA) were assayed. HSA has been proposed as a reductive bleaching for textile, paper and wood pulp industries; HH is commonly used to extract selectively manganese and iron from soils and ores [28]; TD has been widely used as a green and affordable reducing agent in industries like paper, photography, leather processing and textile printing because of its outstanding safety and nontoxic properties [29], [30].
The results obtained by AAS showed that TD and HH are the most effective Mn4+ reducing agents (Fig. SI4). At the first 15 minutes, ~80% of Mn4+ was reduced by TD and ~90% by HH. Mn4+ was almost completely reduced at 75 and 60 minutes by TD and HH, respectively, being the best candidates to be used in the cleaning of Mn-rich stains, inversely to HSA that showed a weak reductive capacity, reaching ~10% after 75 minutes.
Application of Chitosan: TD gel on different Mn-stained substrates
The previous study showed that TD and HH are effective in the reduction of Mn4+ to soluble Mn2+/3+. Nevertheless, HH is known by its acute toxicity and carcinogenicity, inversely to TD that is considered a green reducing agent.
The optimal gel preparation is described in the Section 2.3 and the main premise for its formulation is the use of green products, namely, the chitosan, acetic acid, phosphoric acid and thiourea dioxide. Chitosan is a cationic natural polymer and its remarkable and unique properties such as biodegradability, biocompatibility and antimicrobial activity has been proved [31]. Nonetheless, chitosan needs an acid solution to be solubilized and acetic acid is commonly used for this effect. Aqueous solutions of acetic acid are ranked high as relatively green solvents [32]. Phosphoric acid was added to chitosan gel solution because it acts as co-solvent, aiding the dissolution of insoluble heavy metals (e.g. Fe and Mn) [33]. Phosphoric acid is considered a green product because it respects the energy efficiency principle, concerning to the 12 principles of Greener Chemistry. The proposed green gel was applied on different Mn-stained substrates to test its cleaning efficiency.
Application on stained granite samples
Cleaning efficiency of green gel was assayed, applying it on three different granite samples (Fig. 3 and Fig. SI5). The black stains labelled inside yellow marks showed Mn and Fe percentages ranging from 1.87 to 2.97% and 0.86 to 2.23%, respectively. These percentages originated a Mn/Fe ratio of 1.22±0.56, decreasing 35% in comparation with 1.64±0.57 from Table 1. This Mn/Fe ratio seems to influence the changing of CIELab coordinates because it also led to different L*, a* and b* parameters (Table 2). Moreover, granite samples collected at historical quarry appears to present higher Mn percentages (2.53±0.58%) when compared with stained granite at the indicated churches (1.35±0.17%), being perfectly suitable to the proposed cleaning study.
Table 2. Iron and Manganese percentage of stained granite samples and CIELab color coordinates. L* = lightness; a* = red/green value; b* = yellow/blue value
Elemental analysis
|
CIELab coordinates
|
Granite
Sample
|
Fe (%)
|
Mn (%)
|
Mn/Fe
|
L*
|
a*
|
b*
|
I
|
1.68±0.09
|
2.97±0.24
|
1.77
|
27.29±0.40
|
3.42±0.18
|
9.74±0.32
|
II
|
2.23±0.02
|
2.76±0.09
|
1.24
|
38.69±0.03
|
3.50±0.01
|
6.74±0.01
|
III
|
0.86±0.04
|
1.87±0.31
|
0.65
|
38.27±0.30
|
3.28±0.19
|
9.55±0.40
|
As can be seen in the granite samples I, II and III (Fig. 3 and Fig. SI5), the cleaning gel effectively removed the black stains within the yellow circles. The images captured by Dino-lite before and after the cleaning procedure, substantiate that these Mn and Fe oxides are overlapped at the granite surface, similar to a patina, and the cleaning cycles repetition, where the gel is applied and after removed, is directly associated to the thickness of these superimposed Mn and Fe oxides. Due to the abrasive effect of cotton swabs, is also observed some remaining cotton fibres on the cleaned area (Fig. 3B2).
In situ application on stained granite churches
Mn-stains characterized in Section 4.1 were also submitted to cleaning assays. The areas where the gel was applied are labelled in Fig. SI1 with dashed yellow circles. As expected, the green gel showed its cleaning efficiency by removing effectively black-blue stains on granite, within the yellow marks, at the three churches (Fig. 4 and Fig. SI6). Following the same procedure adapted from granite samples, the cleaning cycles were repeated until full stain removal, with reference to the thickness of the patinas that seem to differ. The patina porosity and gel viscosity are factors that influence the penetration depth of the gel. The dissolution/reduction processes promoted by phosphoric acid and thiourea dioxide are enabled on thin layers of manganese oxide. Thus, it is estimated that the number of cleaning cycles is proportional to the porosity and thickness of the patina.
Application on stained Leclanché cells
Nowadays, Leclanché cells are being used as museum artifacts due to its relevance in the electrochemistry history. Were provided two Mn-stained glass jar cells by ISEP´s Museum (Fig. 5) to test the green gel efficiency, intending to recover the materials originality.
Photos were taken before and after gel application on a stained face and two points were selected to get RGB data (1-8 points) using a Photoshop software. The same procedure was done with a white paper sheet and its RGB parameters (I-IV points) were setting as a cleaned material.
RGB parameters taken at 1-8 and I-IV points are represented in the Table 3. The colorimetric methodology to reveal the cleaning gel efficiency is based on a formula used to calculate the magnitude of a vector (|υ|) with two points (paper sheet coordinates as reference), exemplified by the following formula (2):
(2) |υ|1 = ((RI-R1)2+(GI-G1)2+(BI-B1)2)1/2
Table 3. RGB data before (1, 2, 5 and 6) and after (3, 4, 7 and 8) the cleaning assay. R = red, G = green, B = blue, |υ|ac = magnitude of a vector after cleaning and |υ|bc = magnitude of a vector before cleaning.
|
|
|
|
|
|
Points
|
R
|
G
|
B
|
|υ|
|
|υ|ac/|υ|bc
|
1
|
150
|
89
|
26
|
292
|
0.40
|
3
|
209
|
185
|
159
|
117
|
I
|
249
|
249
|
249
|
|
|
2
|
209
|
173
|
124
|
146
|
0.42
|
4
|
229
|
216
|
194
|
62
|
II
|
246
|
246
|
245
|
|
|
5
|
195
|
162
|
110
|
171
|
0.26
|
7
|
233
|
227
|
212
|
44
|
III
|
248
|
248
|
248
|
|
|
6
|
133
|
99
|
70
|
256
|
0.28
|
8
|
211
|
209
|
195
|
72
|
IV
|
246
|
247
|
245
|
|
|
The points 1, 3; 2, 4; 5, 7 and 6, 8 are compared with paper sheet RGB coordinates in which the points are I, II, III and IV, respectively. A high |υ| indicates a more stained glass, in reverse, a low |υ| suggests a more cleaned glass. The same logic is followed by the ratio |υ|ac/|υ|bc, where a low ratio value is indicative of a high cleaning efficiency. Considering the thickness of the glass and the inclination of the light, the results show, in general, an effective cleaning activity of the Chitosan:TD gel on glass substrates stained with manganese oxide.