Introduction

The evaluation of changes in hydric properties of stone materials caused by the application of restoration treatments is a key task when analysing the effectiveness of conservation interventions, as water is probably the most dangerous agent of degradation regarding archaeological remains. Usually, techniques that require sampling are employed for this kind of tests, such as capillarity water absorption (UNE-EN 15801:2010 standard, AENOR 2010a) or water saturation tests (UNE-EN 1936:2007standard, AENOR 2007), which must be performed in laboratory. Other NDT, such as Karsten tube absorption test (UNE-EN 16302:2016 standard, AENOR 2016a) or water-substrate contact angle (UNE-EN 15802:2010 standard, AENOR 2010b), as well as the sponge test method (Ribeiro et al. 2022; Bison et al. 2023), may be useful in this sense, although providing only superficial and not bulk information.

Restoration treatments, such as consolidating and protective ones, can cause adverse effects resulting in damage of the substrate caused by the creation of a non-breathable surface barrier (Pérez Ema and Álvarez de Buergo 2013). In this case, the consolidant product does not penetrate adequately into the porous system creating a plastic barrier or crust that prevents the internal moisture to get out. Same effect can arise from the application of protective products, especially when they are not permeable to water vapour (Charola 2003; Laborde Márquez 2013; UNE-EN 16581:2016 standard, AENOR 2016b). Thus, the crystallization of salts would occur in the layer immediately below the surface and could cause detachment.

IRT is a non-destructive and contactless technique, traditionally used for civil engineering and architecture, as well as, particularly, for cultural heritage. In this sense, applications are wide: from diagnosis studies to conservation interventions and monitoring tasks (inspection of moisture distribution; detection of cracks, delamination, detachment or voids; to verify the presence of different building materials, etc. (Avdelidis and Moropoulou 2004; Moropoulou et al. 2001, 2013; Sansonetti et al. 2012; Paoletti et al. 2013; Menendez 2016; Forestieri et al. 2017; Forestieri and Álvarez de Buergo 2018, Kirimtat and Krejcar 2018, Vázquez et al. 2015).

An IR camera detects the heat emitted by the observed objects in form of IR radiation and converts it into an electronic image considering the objects as “black bodies”, i.e., hypothetical perfect absorbers and radiators of energy, with no reflecting power. Although some corrections can be made to calculate absolute surface temperature calculations for very homogeneous materials, most application of IRT are based on relative changes of apparent surface temperature.

There are two methodological approaches regarding IRT investigation: single-shot passive thermography and multi-image thermography. In passive thermography, an image of the object is taken, registering differences of apparent surface temperature at a given moment. Multi-image thermography aims to analyse temperature changes as a function of space and, particularly, time. One of the methods used is active thermography, in which an external heat source is used to “excite the sample” and obtain a series of consecutive images after the application of the stimulus (Gómez Heras et al. 2014) to enhance the differences in thermal behaviour during cooling or heating. Sequential thermography focuses on studying how a certain process is reflected in spatial and temporal changes of apparent surface temperature. Specifically, time-sequential thermography is often used to determine heat fluxes from either buildings or the ground, by computing series of images taken at a certain time (Hoyano et al. 1999; Meier et al. 2010).

Water has a very high specific latent heat of evaporation. Therefore, the evaporation process lowers intensely the surface of the evaporating material, even more if this is porous. Hence, IRT is a very suitable technique for studying evaporation in porous materials (Chauvet et al. 2010; Avdelidis et al. 2003; Ludwig et al. 2004, 2018; Sansonetti et al. 2012).

In this work, a simplified approach regarding the application of time-sequential IRT for the study of how restoration interventions modify the evaporative behaviour is presented.

There is an evident lack of analytical methods for the on-site study of hydric properties (Menendez 2016; Delgado Rodrigues and Ferreira Pinto 2022), existing in the bibliography numerous attempts to find innovative methods and techniques (Van Hees et al. 1995; Vallet and Vallet-Coulomb 1999; Pardini, and Tiano 2004; Svahn 2006; Drdácky et al. 2011; Vandevoorde et al. 2013, Vogel et al. 2020; D’Alvia et al. 2022). In this sense, the particular convenience of performing capillarity tests is notorious (Peruzzi et al. 2003), so it is necessary to find alternative methodologies in order to analyse this phenomenon in the archaeological context with greater precision.

The importance of assessing changes in the hydric properties of the material after restoration interventions is crucial, since water interfere in most deterioration processes. The capillary absorption and evaporation tests (capacity of the material to transfer water from the interior to the exterior) are determinant, but hardly feasible on-site without sampling. This work aims, in addition, to carry out a methodological approach to the in situ evaluation of hydric properties through IRT. Treatments, however, must be always tested in the laboratory before being applied to a real structure. Therefore, this NDT could be used to select the best treatments in terms of water performance and then apply them in situ.

This study is part of a PhD Project titled Evaluation of effects arising from restoration interventions carried out on stone material: Roman Theatre and Mitreo’s house of Merida (Perez Ema 2017) carried out between the Technical University of Madrid (UPM) and the Institute of Geosciences-IGEO (Spanish Research Council and Complutense University of Madrid, CSIC, UCM). In this study, different portable and non-destructive techniques have been used in order to assess changes in the behaviour of different conservation treatments applied on stone archaeological materials of Merida archaeological ensemble (Figs. 1 and 2) (Perez Ema et al. 2013).

Fig. 1
figure 1

Location of Augusta Emerita on Roman Hispania map. Map by Maria José de Almeida (Cortesão Silva and Márquez Pérez 2020)

Full size image
Fig. 2
figure 2

(Source: Consorciomerida.org)

Aerial view of the Roman theatre area

Full size image

Goals

The main general goals pursued with this study are the following:

  • • To analyse the changes in water absorption and evaporation capacity of the treated elements in comparison with non-treated ones;

  • • Check the effectiveness of restoration treatments, especially the protective ones;

  • • Visual monitoring of the development of tests in real time in order to get quantitative and qualitative information.

Material and methods

Stone

The front stage of the Roman Theatre of Merida is mainly built with granite and white marble (Fig. 3), both high quality stone. The first is a porphyritic granite (Mota López 2015: 106), constituted by quartz, potassium feldspar and plagioclase minerals, as well as biotite. Its mechanical strength and low alterability are largely due to its texture, compactness of its grains and reduced porosity (3%) (fissural-type porosity).

The white marble, fine-grained, is mainly composed of calcite minerals, with traces of dolomite and muscovite. Its high quality lies in its low porosity, high strength and compactness. Both lithotypes show a remarkable purity due to their mineralogical uniformity, with high concentrations of CaO in the case of white marble, around 50%, while granite is characterized by a concentration of SiO2 close to 67% (Perez Ema 2018).

Granite is employed mainly as structural element, located on the podium and plinth of the upper colonnade. Precisely, due to such mechanical resistance, granite is located on the podium, which determines two main types of potential risk: on the one hand, mechanical deterioration, since it holds the weight of the structural loads of the upper colonnades; on the other, physical and chemical deterioration, being in contact with the ground floor, hence decay arising from capillary absorption of moisture (Merida is located in the Guadiana basin, the groundwater level is at a shallow depth, so the humidity of the ground is a potential agent of deterioration).

White marble was selected for this study as it is the principal decorative element of the front stage, located in capitals and cornices, as well as external facing with marble panels. Apart from this one, there are two more varieties present in the front stage: grey marble and an ophicalcite (misclassified as pink marble).

Fig. 3
figure 3

Front stage of the Roman Theatre of Merida and details of the decorative elements of white marble (top right) and granite from podium (bottom right)

Full size image

Both materials undergo deterioration processes related to the action of water, despite its low open porosity (1,4%) (Perez Ema 2018), mainly granular disintegration, related to the presence of salt efflorescence and biodeterioration (Figs. 4, 5, 6 and 7).

Fig. 4
figure 4

Deterioration by salt cristalization on the surface of granite’s ashlar

Full size image
Fig. 5
figure 5

Granular disagregation of white marble cornice

Full size image
Fig. 6
figure 6

Biodegradation processes on a granite chapitel

Full size image
Fig. 7
figure 7

Pigeon droppings on a white marble cornice

Full size image

Samples and elements of the original material were collected in an adjacent area some tens of meters far from the theatre, where excavated, and not relocated, blocks of stone material were accumulated for years as stockpiling materials (this area is currently cleared of stony remains).

Ten cubic specimens of 50 ± 5 mm side from these collected stone materials were tested, five from each lithotype (marble and granite). The specimens tested in the laboratory showed a good state of preservation, given that it was taken from the interior of the stone blocks, removing the external decayed surface to obtain regular specimens. Ideally, material with a higher level of degradation would have been preferred, despite this it was valid for the comparative study pursued for this work.

Restoration treatments

  1. A)

    Selection criteria

Different consolidating and protective (water repellents) treatments have been selected. One of the criteria for its selection has been its widespread use in the archaeological and architectural field. Among them, some of the most commonly used products for stone material restoration are organosilicic (tetraethyl orthosilicate, TEOS) and siloxane-based products. Both treatments have been traditionally applied also in archaeological sites of Merida, according to some archive reports consulted at the Roman Archaeology Museum (Merida). Estel©1000 was applied directly as commercially available, in white spirit D40 solution, while Tegosivin©HL-100 was diluted in ethanol at 50% concentration.

In addition to these products, an experimental treatment based on nanotechnology was included. Merida Consortium required innovative alternatives to traditional treatments since, in some cases, these last ones manifested adverse effects.

To this end, the collaboration with the Nanomaterials Research group TEP-243, from the University of Cadiz (Department of Chemistry-Physics, Faculty of Sciences) has been carried out. This group develops new restoration products with double consolidating and hydrophobic capacity, through the synthesis of molecular sieve type of nanomaterials (Mosquera 2013).

  1. B)

    Products

Three different products have been applied on the selected stone materials (Table 1) regarding above mentioned criteria:

Table 1 Restoration products tested
Full size table

Restoration products were applied in laboratory environmental conditions (± 20 °C and RH range between 50 and 60%) by impregnation with brush until saturation (two hands within 24 h of each other) on the whole surface of each sample (the 6 faces of each specimen), alone or mixed with another product (this is the case of samples treated with both ethyl silicate first and siloxane later). It was decided to apply it in this way in order to be able to test the effect of the treatment on the material, as accurately simulating the conditions in situ is more complex and difficult to assess.

The nomenclature given to each of the four treatments is the following: non-treated (ST), treated with Estel©1000 (S), treated with Tegosivin©HL (H), treated with Estel©1000 + Tegosivin© (SH) and treated with UCAT-10P© (N).

Analytical techniques

The use of IRT for the evaluation of restoration treatments is based on the fact that the thermal properties of building materials change with the addition of conservation products (consolidant and protective products), also in the presence, and movement, of water, resulting in surface temperature variations, depending on the type and penetration depth of the treatment (Avdelidis et al. 2003; Forestieri and Álvarez de Buergo 2018).

The equipment, available in the Petrophysics Laboratory of the Institute of Geosciences-IGEO (CSIC-UCM), is the ThermaCAM B4 camera from FLIR Systems, with software Flir Quickreport for analysing and processing of images.

Methodology

In this study, a simplified method of time-sequential thermography has been used as it allows us to monitor the evolution of the hydric test and the differences in materials with different treatments (Ibarra-Castanedo et al. 2009; Sansonetti et al. 2012; Pucci et al. 2015). This simplified approach analyses the average apparent surface temperature for each treated block instead of computing images pixel by pixel, as it is normally the case in time-sequential thermography. This approach is considered more adequate for its use in conservation practice, as it is more accessible, does not rely on any other mean than the proprietary software of the thermal camera and allows comparing different sets of samples without the need of having the thermal camera static.

  • Tests in laboratory:

The difference in the thermal behaviour of the stone samples with different restoration treatments, and samples without any treatment, was analysed. Two types of tests have been performed, related to the hydric properties of the material: capillary water absorption test and water absorption-evaporation test. In both cases, tests were carried out in two stages, 12 months after the application of treatments (POST-T(12)); and after 24 months (POST-T(24)). The specific methodology for each test is described as follows:

  1. I.

    Capillary water absorption test: marble and granite specimens (with and without treatments) were placed on a constant moisture source (a wet wipe soaked in water) oriented according to the anisotropic index of each sample in order to minimize differences due to the stone anisotropy. This index was previously analysed by determining ultrasound pulse velocity (UPV) along the three spatial axes in order to settle X, Y and Z axes, being the X axis the one that shows a higher propagation velocity (Fort et al. 2008; Perez Ema 2017). Hence, all samples have been placed with the X-axis of the specimen perpendicular to the bearing surface (Figs. 8 and 9).

In order to monitor the evolution of the test, a sequence of infrared images has been recorded during 12 consecutive hours: during the first hour, every 5 min; during the second hour, every 30 min; and the rest of the time, every hour. A total of 25 images per material (marble and granite) have been obtained, resulting in a total amount of 100 images between POST-T(12) and POST-T(24). IRT allows us to visualize the progression of capillary ascent of water in each sample in the short and mid-term, with different treatments, and compare it with the non-treated samples.

Fig.8
figure 8

Set-up for the capillary water absorption test

Full size image
Fig.9
figure 9

Set-up for the absorption-evaporation test

Full size image

If traditional capillarity tests allow to obtain quantitative data (through the value of capillary water absorption coefficient-AC), IRT provides an added value to the qualitative analysis allowing to visualize, at the same time, how the capillary test progresses in samples with different treatments and how the water rise takes place.

As the stone suctions water from the base, the humidity value rises, causing a decrease in the average temperature of the material. The more water absorbed, the greater the decrease in temperature, until saturation, when adaptation to room temperature begins. Recording thermographic images at uniform intervals enables to visualize the process through time.

  1. II.

    Absorption-Evaporation test: After the capillary water absorption test, the samples were immersed in water for 24 h. Afterwards, they were removed from water and placed on a grid tray, thus avoiding the accumulation of water in the base. A sequence of 18 images per material was recorded during 5 h (until all the samples reached room temperature), obtaining a total of 72 images between POST-T(12) and POST-T(24).

Results

  • I. Capillary water absorption test:

Graphics in Figs. 11 and 13 are indicative of the temperature variation of samples through time. The results for each material are analysed below.

  • I. a. Marble:

In POST-T(12) test, both the untreated specimen (ST) and the specimen treated with Estel©1000 (S) show a greater water suction than the rest of the specimens, reaching its maximum height (capillary water front ascension level) after 25–30 min (Fig. 10a), which is the maximum height of the specimen (50 mm). Moreover, among samples treated with water-repellent products (H, SH and N), the one with the lowest water suction is the one with double treatment (SH), showing no change until minute 30, while H and N perform a similar evolution with a slightly larger suction, but scarce anyway.

In POST-T(24) stone specimens some interesting variations occur (Fig. 10b) comparing to POST-T(12) samples. The water sucked up by sample S is greater in POST-T(24). After 12 h of test, sample treated with Estel©1000 has suctioned more water than in POST-T(12). In addition, S sample shows greater water retention capacity than in POST-T(I).

Fig.10
figure 10

Thermographic images showing the progress of the POST-T (12) and POST-T (24) capillarity test by IRT on white marble specimens (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P.©; ST: non-treated)

Full size image

Graphs in Fig. 11 a-b show temperature variation over time for each marble sample in POST-T(12) test. According to these data, water saturation is reached after 4 h of the beginning of the test, when the greatest decrease in temperature, respecting to the initial value, is registered. Samples with hydrophobic treatments (H, SH and N) show a minimal decrease in temperature, as they barely suction water.

In POST-T(24) test, the variation of temperature over time is minimal in specimens treated with water-repellents, while the specimen treated with consolidant (S) registers a remarkable decrease in temperature, greater than the one registered in the untreated specimen (ST). This decrease continues, in addition, after 12 h of testing (Fig. 11b).

Fig. 11
figure 11

Variation of apparent surface temperature of treated and untreated samples during capillarity test carried out on white marble, 12 months -POST-T(12) (a)- and 24 months -POST-T(24) (b)- after application of treatments (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P.©; ST: non-treated)

Full size image

  • I. b. Granite:

Results obtained in the granite specimens show greater differences between treated samples. Besides, differences between POST-T(12) and POST-T(24) specimens (Fig. 12 a and b) are only remarkable in the case of Estel©1000.

Fig.12
figure 12

Thermographic images showing the progress of the POST-T(12) and POST-T(24) capillarity test by IRT on granite specimens (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P.©; ST: non-treated)

Full size image

The non-treated specimen (ST) is the one that performs the fastest capillarity suction, reaching the level of water saturation after 25 min. Instead, samples treated with consolidant Estel©1000 do not reach saturation until 45 min.

Specimens H and N, with hydrophobic treatments, show little changes until 1.5 h later, while SH specimen (with double treatment) shows the beginning of cooling process in the area in contact with humidity between the 4th and 6th hour.

Both, thermographic images and the graphs in Fig. 13 a-b, allow to determine the water saturation point of ST specimen after 2 h within POST-T(12) test. While, in POST-T(24) test, the specimen treated with ethyl silicate is the first one reaching saturation level 2 h after the beginning of the test; this means that water suction velocity increased in POST-T(24), resulting higher even than the one registered in the non-treated specimen.

The protection barrier created by the hydrophobic treatment is more effective in the case of the double SH treatment than applied alone (H), both in POST-T(12) and in POST-T(24) tests.

Fig. 13
figure 13

Variation of the apparent surface temperature over time of treated and untreated specimens during capillarity test carried out on granite, 12 months -POST-T(12) (a)- and 24 months –POST-T(24) (b)- after the application of treatments (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P©; ST: non-treated)

Full size image

  • II. Evaporation tests

Evaporation test analyses the drying process of the samples, in other words, how each specimen loses the water absorbed after complete saturation. As the evaporation process progress, temperature first decreases rapidly, increasing afterwards in order to reach room temperature. The more water absorbed, the greater the decrease of temperature and the longer it will take to evaporate for the tested specimens.

Evaporation rate depends on the difference between the humidity/saturation of the material and the relative humidity (RH) of the environment: the greater this difference, the higher the rate of evaporation (although other factors may influence, such as the type and size of pores).

The sequence of thermographic images enables to check the amount of water absorbed during imbibition, as well as to analyse the water retention capacity of the samples during evaporation.

Evaporation process takes place faster than the capillary absorption, so the tests last less, 5 h in total.

  • II. a. Marble:

S and ST marble specimens absorbed a large amount of water during water imbibition process, which causes a noticeable decrease of temperature, greater than the one registered on specimens H, SH and N, both in POST-T(12) and in POST-T(24) (Fig. 14 a and b).

Fig. 14
figure 14

Thermographic images showing the progress of the POST-T(12) and POST-T(24) Evaporation test by IRT on marble specimens (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P.©; ST: non-treated)

Full size image

However, while in POST-T(12) test all the specimens reached room temperature after 1 h, in POST-T(24) specimens treated with Estel©1000 show a slighter capacity of water evaporation, even smaller than the non-treated specimen (Fig. 15 a-b). In addition, specimens with hydrophobic treatments (H, SH and N) hardly absorbed water in the imbibition process regarding POST-T(24) test, meaning that its protective effectiveness increased with respect to POST-T(12).

Fig.15
figure 15

Variation of apparent surface temperature of treated and untreated specimens during the evaporation test carried out on white marble 12 months -POST-T(12) (a)- and 24 months –POST-T(24) (b)- after the application of treatments

Full size image

  • II. b. Granite

The differences in the evolution of the Evaporation test are considerable in the case of granite specimens, clearly distinguishing the effects in 3 groups, from greater to lesser variation: S and ST on the one hand; H and N; and, finally, SH, which shows minimal variation.

This disparity is especially visible in the sequence of thermographic images that ranges from minute 25 to minute 40 (Figs. 16 and 17).

Fig. 16
figure 16

Thermographic images showing the progress of the POST-T(12) and POST-T(24) evaporation test by IRT on granite specimens (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P.©; ST: non-treated)

Full size image

Although all the specimens absorb certain amount of water, the evaporation rate is higher in POST-T(12) respecting POST-T(24). SH specimen shows a very fast evaporation rate, evaporating most of the surface water from minute 25 in POST-T(24) test. However, in POST-T(12), that point is not reached until minute 50. S and ST specimens remain completely saturated after the first hour in POST-T(24) test.

The most interesting variation between POST-T(12) and POST-T(24) is shown by the specimen treated with Estel©1000. In POST-T(12) test it begins to dry from minute 40 (Fig. 17a); by contrast, ST specimen remains water saturated after minute 50. In POST-T (24) test, specimen S shows greater capacity of retention than the specimen without treatment (ST), being particularly evident in the measurement recorded after 3 h (Fig. 17b). After 5 h, all the specimens lost most of the superficial moisture and adapted to room temperature.

Fig. 17
figure 17

Variation of apparent surface temperature over time of treated and untreated specimens during the Evaporation test carried out on granite 12 months -POST-T(12) (a)- and 24 months –POST-T(24) (b)- after the application of treatments (S: Estel©1000; H: Tegosivin©HL-100: SH: Estel©1000 + Tegosivin©HL-100; N: UCAT-10P.©; ST: non-treated)

Full size image

Discussion

Previous analysis studying the state of conservation carried out on the Roman Theater of Mérida (Perez Ema 2015 and 2017) verify the significant implication that water plays on the degradation processes that affect the monument, from capillary rise as well as from infiltration and rain. For this reason, the evaluation of changes in hydric properties of the materials caused by the application of restoration treatments is a key task when assessing the effectiveness and compatibility of used products.

The results obtained by IRT are quite homogeneous and clear. Hydrophobic treatments are, in general terms, effective in the creation of protective barriers against external water ingress, highlighting the effectiveness of the double treatment (SH) over the rest. The technique used provides a fairly complete view of the effects that treatments have on the absorption and permeability capacity, in marble and granite.

Both materials, marble and granite, show a low open porosity which largely determines the penetration capacity of the treatments. Thus, while in marble the treatments barely penetrate into the most superficial layers of the specimens, the absorption capacity of the granite and the high connectivity of its porous system facilitate a greater penetration of the product. The water-repellent products create a merely superficial protective barrier on marbles, which barely penetrates the inner layers due to their low porosity. In granite, however, the protective layer penetrates a bit deeper than in marble.

The consolidative treatment based on ethyl silicate, Estel©1000 (S), hampers, in both stone materials, the Evaporation process, retarding the evaporation of water, which is retained for longer inside the porous system. The product created an outer barrier, on the surface, which prevents transpiration. This barrier allows the entrance of liquid water, but delays the exit of water vapor from the inside. This effect, in addition, increases with time, being more evident in the medium term (POST-T(24)) than in the short term (POST-T(12)).

In the case of siloxane-based treatments, in those with Tegosivin© (H and SH), the efficacy against the entrance of water is evident. The thermographic images show that the application of a previous consolidation product in SH specimen, increases its protective efficacy. However, studies carried out by Alonso and Esbert (1994) showed that the overlap of the double consolidating + hydrophobic treatment is more harmful than only the hydrophobic one, when applied on porous materials in the presence of salts. It is therefore essential to determine the salt content of the material beforehand.

The experimental nano-product UCAT-10P© performs a similar behaviour to Tegosivin©, standing out more for its protective function and less for the consolidating one. Although the results are less uniform, it shows an evident water-repellent effectiveness. The effect of this product presents a progressive effectiveness, being higher in the medium term than in the short one.

Tests carried out demonstrate a limited durability of the water-repellent products, decreasing with time in most cases. This disadvantage has been mentioned by other authors (Ferreira Pinto and Delgado Rodrigues 2008; Doehne and Price 2010), emphasizing the need for maintenance or retreatment of the materials to be protected from the outside water ingress. Despite, it is possible to differentiate a greater durability of the protective effect in the case of UCAT-10P© (N), being lower in the case of Tegosivin©, in the SH specimen particularly.

This work demonstrated that the application of Estel©1000 delays the evaporation process in the case of granite and marble (either by the creation of an external barrier, or by the transformation of the porous system, or by the combination of both). When analysing possible effects on materials placed on site, these results would allow us to hypothesize that the application of this product could lead to damage by disintegration caused by crystallization of salt efflorescence in the layers of the stones closest to the treated surface of the stones. This damage would be more severe in the boundary between the treated external surface and the untreated layers underneath, where salts would concentrate and finally crystallize as the water in which they are dissolved finds the way out during the evaporation process. The same effect would take place on surfaces treated with water-repellent products, by reducing the exit of water vapor. This effect is not visible in laboratory tests using IRT because the water-repellent barrier, applied to the entire surface of the specimen, diminishes the entry of water during the imbibition process, and, therefore, the evaporation is very fast. In situ, on the other hand, capillary absorption moisture from the ground is assured, since not all the surfaces of the stone elements would be treated, just the exposed surface.

Related to the lack of testing methods for on-site determination of hydric properties, a useful methodological approach is proposed, by the alternative use of a simplified approach to time-sequential IR thermography as a portable and non-destructive method for the evaluation of both capillary suction and evaporation capacity of stone materials. The first step in order to raise a new methodological approach and to demonstrate its applicability to on-site studies is to determine whether there is a correlation between data obtained with techniques, capillarity and IRT tests. For this purpose, we compared the value of the water gain (Δg/m2) in the capillarity test (Fig. 18), with the variation in temperature (ΔoC) registered by IRT and its evolution over time. It is also clarifying to calculate the Pearson correlation coefficient (r), which is represented in Fig. 19 for each of the treatments applied on granitic material in POST-T(12) test.

Fig. 18
figure 18

Capillary absorption curve of granite, without any treatment, showing the increase in weight (% absorbed water) over time

Full size image

Pearson’s correlation coefficient (r), allows determining the relationship between different variables (Pita Fernández and Pértegas Díaz 2001). To exemplify a single case, granite has been used because it is the material that shows the most uniform and representative results. When there is a negative linear relationship between two variables, the Pearson’s coefficient ranges between 0 and -1, and allows quantifying the intensity of the relationship between these variables (Vila et al. 2016). The values shown next to the graphs correspond to each treatment.

Fig. 19
figure 19

Correlation between the results obtained by traditional capillarity test and capillary test monitored by IRT. Graphics represent the calculation of the difference between the water weight gain (Δg/m2) (Y axe) and the temperature variation (ΔoC) (X axe) per unit of time (h) (the increase in time is represented by a progressive dimming of the colors assigned to each treatment). Pearson’s correlation coefficient (r) is indicated in the lower left angle of each graph for each of the analysed variables

Full size image

In all cases, it is proved that there is a correlation between results, being a correlation of negative linear type in most cases (H, SH and N), while in the case of S there is a non-linear correlation, existing correlation in any case. This means that the weight increases of the specimens during the capillarity test, linked to the increase of weight per unit area, implies the reduction in apparent surface temperature. So, both variables have a significant direct relationship, with a Pearson coefficient close to 0.

Conclusions

Conclusions regarding the effects arising from the application of treatments on marble and granite indicate that all of them are effective in fulfilling their intended functions. However, not all of them perform it in the same extent and, in some cases, despite this effectiveness, they show more disadvantages and potential damage than benefits. The protective efficacy of the different treatments applied is ordered in ascending way as follows: S → H → N → SH.

In direct relation with this last statement, it is worth noting that one of the most obvious disadvantages in the use of the Estel©1000 is the creation of a superficial barrier that, on the one hand, increases the speed of capillary suction, on the other hand prevents the escape of water vapor from inside to outside. This can promote internal deterioration processes related to the presence of salts.

There is a direct correlation between data obtained through capillarity and IRT, which allows opening a new experimental route for the evaluation of new methods of analysis and on-site evaluation in order to measure the hydric properties of materials in the archaeological context through non-destructive techniques.

The IRT technique offers an outstanding alternative for the determination of on-site hydric properties of materials in archaeological sites and for the on-site evaluation of consolidation and protection treatments as it is portable, non-destructive and allows the visualization in real time. Although this paper presents the main theoretical and practical basis to assess the effect of conservation treatments on hydric behaviour, further studies are necessary to determine how to implement the methodology of these tests when carried out onsite.