Czytaj książkę: «Monument Future», strona 20
Discussion
Due to the lime inclusions, the RM mortar can be classified as a dry slaked mortar, that is known due to its sandwich-like preparation (Wedekind 2014). The RM mortar was made in a traditional way, but its durability is limited and not comparable with a hot mixed mortar by using amorphous aggregates, like the two described historical mortars. However, 142the RM mortar from the 1941 restoration work was still preserved in huge amounts. The RM mortar does not show any harmful side effects unlike the modern, cement-containing mortars used for the main restoration works in Zvartnots.
Figure 5: Damage mapping of a wall at the Zvartnots Cathedral ruin.
The two different historical mortars can be classified as hot-prepared mortars. In this case, the burned lime was probably mixed and slaked together with the amorphous siliceous aggregate. This could be the reason why no visible lime accumulation is visible. It is known, that the solubility of amorphous silicate increases sharply at a high pH in combination with heat (Iler 1979). All conditions are present during the mortar preparation, heat, a high pH of 14 and a lot of amorphous silicate material used as aggregate.
The high temperature reached by the exothermic reaction of the burned lime slaked with water and the high pH (14) seems to soften and chemically desolve the fine amorphous silicious particles within the aggregate and to form calicite-silica compounds. These cement-like, but salt-free and highly porous mortars seem to be the main reason, why very large parts of the Armenian cultural heritage is still preserved. These mortars also show a perfect compatibility with the highly porous building stones (Tab. 1). The tuff material has a high water uptake capacity and shows a comparably high resistance agains salt weathering (Pötzl et. al 2018).
Conclusions and further investigations
The high values of sulfate measured at the Zvartnots Cathedral could probably be due to the use 143of cement mortar, implemeted during restoration works at the ruins in the 1950s–1980s (Fig. 2f). Rain water could also be one possible source, since immense infiltration took place at many walls of the ruins. However, aggressive magnesium sulfpher salts could be formed if a dolomitic lime was used for the historical mortar. This should be clarified by further investigations.
Figure 6: Damage mapping of the right half of the main façade at the Avan Cathedral.
A proper desalination method has to be tested and applied. The first experimental conservation works during the summer school 2019 using poultices and the sprinkling method (Wedekind, Ruedrich 2006) have shown promising results.
The observations and results show that the restoration mortar, which is comparable to the historical mortar could be a proper material for both historical objects.
To discover the secret of its eternity and to prepare and develop a suitable and sustainable mortar for restoration, should be the main goal of future research.
Acknowledgements
We would like to thank all the participants of the 1th Summer School “Stones in Armenian Architecture” for their great work. We also thank the administration of heritage preservation of the Zvartnots side for their friendly cooperation, especially Grigor Nalbandyan as well as the Armenian Apostolic Church. We are grateful to Alfons van den Kerkof for allowing us access to the cathodoluminescence microscope at the Geoscience Centre of the University of Göttingen and for his friendly help and cooperation.
This work was generously supported by the Volkswagen Foundation (AZ93919).
References
Brandi, C. (2006) Theorie der Restaurierung (trad. Schäedler-Saub, U., Jakobs, D.), Munich.
Iler, R. K. (1979) The Chemistry of Silica, Wiley (Interscience, New York.
Maroutyan, T. (1976) Avani tachary ev hamanman husharcanner [Cathedral of Avan and same monuments], 1976, Yerevan
Pötzl, Chr., Siegesmund, S., Dohrmann, R., Koning, J. M., Wedekind, W. (2018) Deterioration of volcanic tuff rocks from Armenia: constraints on salt crystallization and hydric expansion, Environmental Earth Sciences 77:660, https://doi.org/10.1007/s12665-018-7777-8.
Wedekind, W., Harutyunyan, E., Novakovic, N., Siegesmund, S. (2020) Experimental Conservation and first Investigations on the Weathering of Geghard Monastery (Armenia). In Siegesmund, S. and Middendorf, B. (Eds.), Monument future: Decay and conservation of stone, Göttingen, Kassel, Mitteldeutscher Verlag, Halle.
Wedekind, W. (2014) Schwierige Ruinen – Zur Erhaltung der Ruinen und Felsmonumente an der Unstrut. In.: Siegesmund, S., Hoppert, M., Epperlein, K. (Eds.) Natur – Stein – Kultur – Wein – 144zwischen Saale und Untrut. Mitteldeutscher Verlag, p. 293–320.
Wedekind, W., Ruedrich, J. (2006) Salt-Weathering, Conservation Techniques and Strategies to protect the rock cut Facades in Petra/Jordan. In: R. Fort, M. Álvarez de Buergo, M. Gomez-Heras & C. Vazquez-Calvo (Eds.). Heritage, Weathering and Conservation. Taylor & Francis, London, p. 261–268. terscience, New York.
Maroutyan, T. (1976) Avani tachary ev hamanman husharcanner [Cathedral of Avan and same monuments], 1976, Yerevan
145
NOTES OF A BOWING BEHAVIOR ON LIMESTONE
Luís Sousa1,2, Johanna Menningen3, Siegfried Siegesmund3
IN: SIEGESMUND, S. & MIDDENDORF, B. (EDS.): MONUMENT FUTURE: DECAY AND CONSERVATION OF STONE.
– PROCEEDINGS OF THE 14TH INTERNATIONAL CONGRESS ON THE DETERIORATION AND CONSERVATION OF STONE –
VOLUME I AND VOLUME II. MITTELDEUTSCHER VERLAG 2020.
1 University of Trás-os-Montes e Alto Douro, Quinta de Prados, 5001-801 Vila Real, Portugal
2 Centro de Geociências da Universidade de Coimbra, Coimbra, Portugal
3 Geoscience Centre of the Georg August University Göttingen, Germany
Abstract
Eleven Portuguese limestones were evaluated regarding their bowing behavior. Only one variety showed meaningful bowing values under thermohydric conditions. The slabs of this variety, named Valverde, show a constant rate of deformation and reach bowing values of about 8 mm/m. While under dry heating cooling cycles no deformation was observable, the addition of water lead to increasing deformation values. On the contrary, alternating dry and wet cycles enable the limestone samples to recover during the dry cycles. Only continuously wet conditions imprint permanent bowing.
Petrographic observations and dilatometry tests allow to identify several stylolites filled with iron oxides as a possible reason for this behavior. These irregular fractures open under thermohydric conditions. Beyond a certain opening limit, the asperities inhibit the minerals to retake the initial position and deformation will be permanent. Therefore, this stone should not be used under wet conditions combined with higher temperatures. Those conditions would cause a bowing behavior as reported in several cases.
Keywords: limestone; bowing; petrographic properties
Introduction
Bowing behavior has been described for different rock types. Nevertheless, marble is the stone mostly prone to such a permanent deformation. The specific textural characteristics of marble together with anisotropic calcite crystal properties causes thermal strain within grain to grain contacts and originates bowing (Siegesmund et al. 2000; Siegesmund et al. 2008). Although this phenomena was described for many rock types, as for example 146even weathered granites can display bowing behavior (Siegesmund et al. 2018), there are only scarce cases reported for limestones (Siegesmund 2008).
Figure 1: Macroscopic appearance of Valverde limestone (honed surface; size of the photo: 10 cm × 10 cm).
Table 1: General characteristics and classification (according Dunham (1962) and Folk (1962)) of the selected limestones. The sample VAV is the Valverde variety.
| Sample | General characteristics | Classification |
| ALP | Grey limestone composed of a micritic groundmass (95 %) and 5 % of components | Pelagic Mudstone (after Dunham, 1962) and Micrit (after Folk, 1962) |
| ATAZ | Light cream limestone with 50 % groundmass and 50 % components | Peloidal wackestone (after Dunham, 1962) and Pelmicrit (after Folk, 1962) |
| ATCR | Cream coloured limestone with 60 % groundmass and 40 % components | Peloidal wackestone and packstone (after Dunham, 1962) and Pelmicrit (after Folk, 1962) |
| CODFV | Light grey limestone composed of 60 % groundmass and 40 % components | Ooid-peloid grainstone (after Dunham, 1962) and Oopelsparit (after Folk, 1962 |
| LIOZ | Fully recrystallized limestone composed of calcite microcrystals (sparit and microsparit) | Dolosparit (after Folk, 1962) |
| MCCT | Light cream limestones composed of 50 % groundmass and 50 % components | Bioclastic grainstone (after Dunham, 1962) and Biopelsparudit (after Folk, 1962) |
| SBM | Light cream limestone composed of 40 % groundmass and 60 % components | Peloidal grainstone (after Dunham, 1962) and Pelsparit (after Folk, 1962) |
| SBR | Light cream limestone composed of 40 % groundmass and 60 % components | Ooid grainstone (after Dunham, 1962) and Oosparudit (after Folk, 1962) |
| VAV | Light to medium grey limestone composed of 70 % groundmass and 30 % components. | Bioclastic packstone/grainstone (after Dunham, 1962) and Biosparit/Biodismicrit (after Folk, 1962). |
| VPAZ | Medium grey limestone composed of a micritic groundmass (> 90 %) and less than 10 % of components | Pelagic Mudstone (after Dunham, 1962) and Micrit (after Folk, 1962) |
| VPCR | Light grey limestone composed of 60 % groundmass and 40 % components | Bioclastic floatstone (after Dunham, 1962) and Biomicrudit (after Folk, 1962). |
For a Portuguese limestone (usually named Valverde) several stoneworkers described bowing phenomena. In order to evaluate these observations systematically, eleven limestone varities were selected and analysed: Alpinina (ALP), Ataíja Azul (ATAZ), Ataíja Creme (ATCR), Semi-rijo Codaçal (cut pararellel to the sedimentary lamination, CODFV), Lioz (LIOZ), Moca 4M (cut perpendicular to the sedimentary lamination, MCCT), Semi-rijo Branco Mais(i) (SBM), Semi-rijo Branco Real (SBR), Vidraço Azul Valverde (VAV); Vidraço Portela Azul (VPAZ); Vidraço Portela Creme (VPCR).
The limestone samples used in this study where collected in the Maciço Calcário Estremho (MCE), a Jurassic limestone massif located in the Portuguese Lusitanian Basin, with the exception of one sample (LIOZ) collected in a Cretaceous Unit near Lisbon (Silva, 2017). MCE is one of the world’s leading regions producers of limestones for ornamental purposes. The quarries develop along the hillside, with several benches, can reach depths around 50 m and are grouped in several exploitation sites. Several lithostratigraphic units are exploited and originate different ornamental varieties, from cream laminated calcarenites to grey calciclastic micritic limestones (Carvalho and Lisboa, 2018). The samples selected from MCE show this high variability.
Figure 2: Microstylolites (right in the photo) and non-disturbed matrix (left in the photo) in the sample VAV (crossed polars).
147Besides a petrographic characterization, the hydric expansion caused by moisture exposure was evaluated as well as the bowing potential and thermal expansion measurements were performed.
Material
The detailed petrographic characterization of the Valverde sample highlights some interesting features; the limestone shows a partially recrystallized groundmass, consisting of around 50 % micritic and 20 % of dismicritic components and peloidal microsparit.
Microscopically it is possible to recognize that there are zones (patches) composed of micritic matrix and an other one composed of sparitic cement, giving the rock an inhomogeneous effect similar to bioturbation. All components are smaller than 2 mm composed of bioclasts, pelloids, pseudo-pelloids, benthic foraminifers, echinoderms and several spicules of crinoids. The fabric of the sample varies locally from mud- to grain- supported.
An interesting aspect of this sample is the presence of disturbed areas where parallel microstylolites appear, affecting the groundmass. The microstylolites consist of an opaque material (black) probably accumulations (laminaes) of organic material and clays.
According to Flügel (2004) this limestone belongs to the RMT 14, Bioclastic packstone and wackestone with skeletal grains, various amounts of intraclasts and some ooids (shoal-near). These sediments are common for the open marine, inner ramp zone.
Methods
The hydric expansion was determined for all limestones on cylindrical samples (Ø 15 mm × 100 mm). The length change of those cylinders was measured under water-saturated conditions. The samples were completely immersed in demineralized water for 24 hours, meanwhile the expansion of the samples was measured with a displacement transducer. The resolution of the transducer is 0.1 µm and the accuracy is about 0.5 µm.
Figure 3: Evolution of the moisture expansion during a 24 hours test.
Figure 4: Bowing behavior of the selected limestones under wet and dry cycles.
Figure 5: Bowing effect after heating cycles.
To determine the bowing potential and its directional dependence, slabs of 40 × 10 × 3 cm were exposed to heating cycles (20–80–20 °C per day) simulating in an accelerated procedure the conditions at the building. Additionally, the specimens were 148exposed to thermohydric conditions, as the slabs were moistened on one side and cyclically heated on the reverse slab surface. The slabs were heated for 6 hours before they cool down to ambient temperature and the water was refilled 2 hours before the next heating cycle started. The bowing was measured by means of a measuring bridge with an accuracy of l µm/35 cm. A total of 91 cycles were performed (for details see Koch and Siegesmund 2002).
Figure 6: Progressive increase of residual strain for 20–90–20 °C cycles of the VAV samples. Investigated as a function of the number of heating cycles under dry (3 cycles, red background) and wet (4 cycles, blue background) conditions.
Thermal expansion measurements took place in the temperature range of 20–90 °C using a pushrod dilatometer (for details see Koch and Siegesmund 2004). The specimens had a length of 50 mm and a diameter of 15 mm. The residual strain (RS) was determined within seven heating–cooling cycles: 3 dry cycles 20–90–20 °C and 4 wet cycles 20–90–20 °C, in the Z and XY directions. RS is determined by the ratio between the sample length change after the heating–cooling cycles and the initial length (Shushakova et al. 2013).
Results and Discussion
The values of hydric dilatation are lower than 0.09 mm/m, with excepetion of the Valverde (VAV) samples which reach 0.22–0.26 mm/m.
Usually limestones show low hydric dilatation values, because these rocks have low clay content (Siegesmund and Dürrast, 2014). The dilatation of the VAV sample could be explained by some possible clay content within the dark material observed in microstylolites. Further investigations must be performed to clarify this hypothesis.
The bowing tests show the absence of permanent changes in almost all of the selected limestone. In fact, only the Valverde (VAV) sample is affected by this experiment. In the first 24 wet cycles the bending increases continuosly and reaches the value of 7 mm/m. Followed by dry conditions, the sample shows some recovery down to 6 mm/m within the first cycles. This value remains stable during the next 41 dry cycles. When the wet cycles start again a new increase of the bowing is evident with a maximum expansion around 9 mm/m.
149High values of bowing are frequently observed in marbles (e. g. Siegesmund et al. 2008) but unknown in limestones. Furthermore, as the mineralogy of all selected limestones is similar the bowing of the VAV sample turns strange. The only different factor are the microstylolites, which therefore seem to be the reason for the bowing potential. Since no bending occurs when dry cycles are performed, the water is an important factor. The interaction of heat and water seems to open and offset the microstylitic cracks, avoiding the recontraction and resulting in an irreversible deformation.
The thermal expansion measurements for the VAV sample show some residual strain about 0.25 mm/m after 4 wet cycles, while the performed dry cycles do not exhibit permanent deformation. Additionally, the sample has been measured in two directions, Z and XY, perpendicular and parallel to the microstylolites considering any effect caused by directional dependence. As depicted in Fig. 4, the expansion of both directions is comparable. Consequently no directional dependence of the expansion can be observed.
Residual strain is only observed under wet conditions as well as the bowing behavior is only evident under wet conditions. Despite the values of residual strain are lower than the range of values found in marbles exhibiting bowing (Mennigen et al. 2018), they confirm the bowing behavior of this sample. The areas with parallel microstylolites affecting the groundmass are aleatory distributed in the limestone and each one possibly contributes the total bending of the sample.
These preliminary results show the influence of the water in the bowing behavior under thermal cycles. However the nature of the microstylolites needs to be completely analysed, especially if clays are present. Therefore, more studies are necessary to understand completely the anomalous bowing behavior of the Valverde limestone.
Conclusions
The boeing potential of eleven varieties of Portuguese limestones were studied. Only one variety (Valverde) show evident bowing, reaching values around 9 mm/m after 91 cycles. This variety is also the only one showing hydric dilatation with values around 0.26 mm/m after 24 hours. Residual strain happens only in wet conditions, with values around 0.25 mm/m.
Petrographic observations of this variety highlight the presence of disturbed areas with parallel microstylolites. The microstylolites are of an opaque material, probably accumulations of organic material and clays. Both water and heat may cause the opening and offsetting of the microstylolites causing irreversible deformation.
Further studies are necessary to fully explain this bowing behavior. Meanwhile, this stone should not be used as a cladding material in a humid and warm environment, where bowing behavior can occur more easily.
