Petrographic and petrophysical characterization of the main aerial and hydraulic mortars used in the construction and rehabilitation sectors

2.1. Materials

The studied materials in this work are an aerial lime, two types of natural hydraulic lime and an ordinary Portland cement. The air lime is a CL90-S PL (hydrated calcium lime in paste) from the company PROIESCON S.L. and the natural hydraulic limes a NHL2 and NHL5 (non-additivated, feebly, and eminently hydraulic binders) from the Saint Astier company. The limestone-additivated Portland cement (CEM II/ B-L 32.5 N) is an artificial hydraulic binder produced by Lafarge-Holcim. All types of limes are regulated by EN 459-1: 2015 standard (99. European Committee for Standardization (Comité Européen de Normalisation, CEN). EN 459-1. 2015. Building lime - Part 1: Definitions, specifications and conformity criteria.
) while cement is regulated by EN 197-1: 2011 (4141. European Committee for Standardization (Comité Européen de Normalisation, CEN). EN 197-1. 2011. Cement - Part 1: Composition, specifications and conformity criteria for common cements.
) and RC-16 (2016) (4242. Ministerio de Fomento. Secretario General Técnico. Comité Permanente del Cemento. RC-16. 2016. Instrucción para la recepción de cementos.
).

A total of 36 mortar samples were made with coarse siliceous sand aggregate (1-2 mm). Nine of each type of binder: 3 prismatic samples of 10x10x4 cm, 3 prismatic samples of 16x4x4 cm and 3 square plates of 5x5x1 cm. The binder: aggregate ratio was 2:3 and the curing period of the mortars prior to the start of the tests was 90 days. The curing conditions for all mortars were > 80% RH for the first 7 days and approximately 60% RH at a temperature of 20-25°C until 90 days. The 90-days curing period was selected on the basis of previous studies on both air and hydraulic limes (43-4643. Veiga MR, Aguiar J, Silva AS, Carvalho F. 2001. Methodologies for characterisation and repair of mortars of ancient buildings. In Proceedings of the 3rd International Seminar Historical Constructions: possibilities of numerical and experimental techniques, 353-362, Guimarães, Portugal.
44. Grilo J, Silva AS, Faria P, Gameiro A, Veiga R, Velosa A. 2014. Mechanical and mineralogical properties of natural hydraulic lime-metakaolin mortars in different curing conditions. Constr. Build. Mater. 51:287-294. https://doi.org/10.1016/j.conbuildmat.2013.10.045
45. Pavlík V, Užáková M. 2016. Effect of curing conditions on the properties of lime, lime-metakaolin and lime-zeolite mortars. Constr. Build. Mater. 102(1):14-25. https://doi.org/10.1016/j.conbuildmat.2015.10.128
46. Santhanam K, Ramadoss R. 2022. Sustainability development and performance evaluation of natural hydraulic lime mortar for restoration. Environ. Sci. Pollut. Res. Int. 29(6):79634-79648. https://doi.org/10.1007/s11356-022-21019-x
) and due to it was considered sufficient time for further handling and evaluation of the physical and mechanical behaviour of the limes, especially air lime.

The following abbreviations were used: CL90-S PL (air lime mortar), NHL2 (natural hydraulic lime NHL2 mortar), NHL5 (natural hydraulic lime NHL5 mortar) and OPC (ordinary Portland cement mortar).

2.2. Techniques and tests

The techniques and tests used in this study have been considered the most appropriate for the evaluation of the materials studied and are based on EN 17187: 2020 (4040. European Committee for Standardization (Comité Européen de Normalisation, CEN). EN 17187. 2020. Conservation of cultural heritage - Characterization of mortars used in cultural heritage.
), allowing for further comparisons with mortars used in cultural heritage.

3.1. Petrographic characterization

Macroscopically, all the mortars present a granular texture and a massive appearance, so their main differences are the colour and the cohesion between their aggregates and the binder (Figure 1). CL90-S PL is a white, not very dense, and nor competent mortar, which presents disaggregations or material losses on its surfaces as well as irregular pores and cracks. Whereas NHL2 and NHL5 are whitish-greyish mortars and dense and compact materials, also with rough surfaces. Their main differences are a higher cohesion between the aggregate and the binder and a lower porosity in NHL5 compared to NHL2. Finally, OPC has a greyish-greenish colour and a high density and hardness, as it is the only mortar with low porosity and smooth surfaces. Its porosity has a spheric morphology (vacuolar porosity), the same as in NHL2 and NHL5, but smaller in size.

Optical microscopy (POM) shows that all the mortars present a granular texture dominated by inert siliceous aggregate (Figure 2). The aggregate is well selected and consisted of mono- and polycrystalline quartz grains of subangular to subrounded morphology with a grain size of 0.7 to 1.4 mm (coarse to very coarse sand).

In CL90-S PL, the binder consists of a microcrystalline mass which presents shrinkage cracks and fissures because of the curing (hardening process). Porosity of irregular morphology is also identified due to possible dissolution processes (Figures 2A-B). This binder is not carbonated (CaCO₃), as it presents low birefringence colours (grey colours in cross polarized light), which belong to portlandite (C-H, Ca(OH)₂) (Figure 2B).

In the binders of NHL2 and NHL5, concentrated masses of carbonate in the form of nodules and the increment of the birefringence colours in cross polarized light (ochre and brown, Figures 2D-F) are observed. This indicates the advance of the carbonation process, larger in NHL5 (2-3 cm depth). A greater chemical interaction between the binder and the aggregate is also seen in NHL5. The quartz grains show diffuse and sinuous edges, due to strong chemical reactions produced by the high pH (pH > 10, portlandite) reached during the hydration phase. Under these chemical conditions, the edges of the quartz grains dissolve (pH > 9), while the carbonates of the binder precipitate (pH > 8) on those edges. This allows the binder to adhere well to the aggregate, preventing its disintegration. It has been noted that in NHL2 there are some shrinkage cracks, not present in NHL5, and vacuolar porosity. This type of porosity, which is dominant in NHL5 (Figures 2C-D), is due to fluid bubbles trapped during mixing. NHL5 also shows more calcium silicates crystals than NHL2. These calcium silicates clusters appear poorly defined and have low birefringence in cross-polarized light mode (Figures 2E-F).

Lastly, OPC presents a heterogeneous binder with the most advanced degree of carbonation among all the studied mortars (highest birefringence throughout the thin section). This mortar also shows vacuolar porosity, quartz grains with very diffuse and dissolved edges and many calcium silicates clusters (hatrurite/alite - C₃S (Ca₃SiO₅) and larnite/belite - C₂S (Ca₂SiO₄), Figures 2G-H). These calcium silicates clusters appear well developed (70-180 µm), with various tones (alite - white and belite - brown, in parallel-polarized light), and sometimes also surrounded by ferruginous films (brownmillerite or calcium ferritoalumanite - C₂AF (Ca₂(Al, Fe³⁺)₂O₅) ( Felite - C₄AF) (6868. Virella Torras A. 1964. Observación microscópica del clínker de cemento portland. Mater. Construcc. 14(113):23-30. https://doi.org/10.3989/mc.1964.v14.i113.1791
, 6969. Elsen J. 2006. Microscopy of historic mortars - a review. Cem. Concr. Res. 36(8):1416-1424. https://doi.org/10.1016/j.cemconres.2005.12.006
).

SEM results (Figure 3) confirm the observations made with POM. CL90-S PL presents a binder with a very homogeneous calcium composition (90-100% of CaO, by semi-quantitative EDS) and a lot of fissural porosity and microporosity (Figure 3A). The rest of the mortars, as they become more hydraulic, present in their binders more calcium silicates, carbonate nodules (95-100% of CaO) and vacuolar porosity (Figures 3B-D). In NHL2, monocalcium silicates (43% of SiO₂ and 57% of CaO ~ wollastonite - CS), bicalcium silicates (30-35% of SiO₂ and 60-65% of CaO ~ larnite/belite - C₂S) and some interstitial phases (20% of Fe₂O₃, 20% of Al₂O₃ and 50% of CaO ~ brownmillerite - C₂AF) are abundant (Figure 3B). In NHL5 and OPC, tricalcium silicates (20-25% of SiO₂ and 70-75% of CaO ~ hatrurite/alite - C₃S) also appear (Figures 3C-D), although in greater proportion in OPC, in which monocalcium silicates (wollastonite) is no longer detected.

Regarding the chemical composition of the binder, it shows an enrichment in SiO₂ and Al₂O₃ in the more hydraulic mortars, as can be seen in the three-phase triangle (Figure 4, < 3% of SiO₂ and 1.5% of Al₂O₃ for CL90-S PL, 11-12% of SiO₂ and (2% of Al₂O₃ for NHL2, 20-21% of SiO₂ and ~ 5% of Al₂O₃ for NHL5 and 24-25% of SiO₂ and 7-8% of Al₂O₃ for OPC).

Thus, a series of characteristics related to the different degree of hydraulicity of the binders can be observed (Figure 5).

The higher the hydraulicity degree, the faster the carbonation, which explains that OPC, with well-developed tricalcium silicates (alite), has the most advanced degree of carbonation. It also be noted that as the hydraulicity degree increases, does the aggressiveness of the dissolution-precipitation chemical reactions at the edges of the aggregate, which are due to the alkaline nature of the surrounding binder (77. Varas MJ, Álvarez de Buergo M, Fort R. 2005. Natural cement as the precursor of Portland cement: Methodology for its identification. Cem. Concr. Res. 35(11):2055-2065. https://doi.org/10.1016/j.cemconres.2004.10.045
).

Furthermore, water that is not able to react chemically with the components of the paste and that is used to increase the plasticity originates vacuolar pores, typical of hydraulic mortars. Natural hydraulic lime mortars (NHL2 and NHL5) differ in the degree of development of this type of porosity and in the aggressiveness of the alkaline reactions at the edges of the aggregate. Thus, the most hydraulic of them (NHL5) does not present shrinkage fissures, which impacts on a greater compactness and a less cohesion loss. Finally, CL90-S PL is the mortar that both micro and macroscopically shows the lowest degree of carbonation, which together with the high presence of shrinkage cracks results in high material losses or disaggregation.

In general, all lime mortars (especially CL90S-PL) do not show an advanced state of carbonation, which is explained by the curing time required by this type of materials (2424. Zhang D, Zhao J, Wang D, Xu C, Zhai M, Ma X. 2018. Comparative study on the properties of three hydraulic lime mortar systems: Natural hydraulic lime mortar, cement-aerial lime-based mortar and slag-aerial lime-based mortar. Constr. Build. Mater. 186:42-52. https://doi.org/10.1016/j.conbuildmat.2018.07.053
, 70-7270. Álvarez de Buergo Ballester M, González Limón T. 1994. Restauración de edificios monumentales. Estudio de materiales y técnicas instrumentales. Segunda edición. Monografía Centro de Estudios y Experimentación de Obras Públicas (CEDEX M-43).
71. Borrelli E, Laurenzi Tabasso M, Sánchez Martínez AE. 1996. Realización de muestras de mortero con cal aérea para el estudio de productos consolidantes: Propuesta metodológica. III Congreso Internacional de Rehabilitación del Patrimonio Arquitectónico y Edificación, 307-313, Granada, España.
72. Skoulikidis TH, Charalambous D, Tsakona K. 1996. Amelioration of the properties of hydrated lime for the consolidation of the surface or/and the mass of building materials of monuments or new buildings or statues and ornaments. In Proceedings of the 8th International Congress on Deterioration and Conservation of Stone, 1599-1605, Berlin, Germany.
). This high portlandite/calcite ratio was also confirmed by XRD, FTIR and TG-DTA analysis.

Thus, Figure 6 shows that all mortars present peaks of portlandite (4.90Å) and calcite (3.03Å) of variable intensity depending on their state of hydration and carbonation, as well as high intensity peaks of quartz (3.34Å) due to the aggregate fraction.

After 90 days of curing, CL90-S PL shows the highest amount of portlandite (C-H), followed by NHL2, NHL5 and OPC. In CL90-S PL, in addition to calcite, one of its metastable polymorphs is also detected: aragonite (3.39Å, 3.27Å, 2.71Å). The presence of this mineral phase has been found very frequently in fresh mortars in which carbonation occurs at relative humidities similar to the curing conditions of this study (~ 60% RH) (4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
, 7474. Ergenç D, Gómez-Villalba LS, Fort R. 2018. Crystal development during carbonation of lime-based mortars in different environmental conditions. Mater. Charact. 142:276-288. https://doi.org/10.1016/j.matchar.2018.05.043
, 7575. Steiner S, Lothenbach B, Proske T, Borgschulte A, Winnefeld F. 2020. Effect of relative humidity on the carbonation rate of portlandite, calcium silicate hydrates and ettringite. Cem. Concr. Res. 135:106116. https://doi.org/10.1016/j.cemconres.2020.106116
).

Regarding the characteristic phases of hydraulic binders (calcium silicates), they appear mainly in the range of 30-35º [2θ] (2.98-2.56Å). Although they are difficult to detect because of the overlap between their peaks and those from the aggregate (Figure 6) (77. Varas MJ, Álvarez de Buergo M, Fort R. 2005. Natural cement as the precursor of Portland cement: Methodology for its identification. Cem. Concr. Res. 35(11):2055-2065. https://doi.org/10.1016/j.cemconres.2004.10.045
, 4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
). The type of calcium silicates detected in all hydraulic binders is larnite/belite (C₂S), and only in OPC hatrurite/alite (C₃S) is also found. Larnite (2.79Å, 2.74Å and 2.33Å as distinctive peak) is the predominant hydraulic phase in natural hydraulic limes (7676. Allen G, Allen J, Elton N, Farey N, Holmes S, Livesey P, Radonjic M. 2003. Hydraulic lime mortar for stone. Brick and block masonry. Donhead Publishing Ltd. Dorset.
), as the most suitable temperatures for its formation are around 1000-1100 ºC (7777. Kozlovcev P, Přikryl R. 2015. Devonian micritic limestones used in the historic production of Prague hydraulic lime (‘pasta di Praga’): characterization of the raw material and experimental laboratory burning. Mater. Construcc. 65(319):e060. https://doi.org/10.3989/mc.2015.06314
), the considered ones as optimal to produce this binder (7878. Válek J, Van Halem E, Viani A, Pérez-Estébanez M, Sevcík R, Sasek P. 2014. Determination of optimal burning temperature ranges for production of natural hydraulic limes. Constr. Build. Mater. 66:771-780. https://doi.org/10.1016/j.conbuildmat.2014.06.015
, 7979. Kozlovcev P, Válek J. 2021. The micro-structural character of limestone and its influence on the formation of phases in calcined products: natural hydraulic limes and cements. Mater. Struct. 54:217. https://doi.org/10.1617/s11527-021-01814-7
).

On the other hand, hatrurite/alite is a characteristic mineral of higher hydraulicity binders, such as cement, that appears in the 32-33º [2θ] range with their most distinctive peaks at 1.76 Å (51.8º [2θ]) and 5.95Å (14.9º [2θ]) (4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
). Although this type of calcium silicate can be also generated in natural hydraulic limes if there are “hot spots” (~ 1250 ºC) in the kiln (8080. Gualtieri AF, Viani A, Montanari C. 2006. Quantitative phase analysis of hydraulic limes using the Rietveld method. Cem. Concr. Res. 36(2):401-406. https://doi.org/10.1016/j.cemconres.2005.02.001
) during the calcination of the raw materials.

In addition, OPC presents other phases such as (Figure 6): calcium ferritoaluminates (brownmillerite or felite -2.65Å and 7.19Å- Ca₂(Al,Fe³⁺)₂O₅) and calcium aluminates (celite -2.69Å- Al₂O₃·3CaO; C₃A) (4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
, 8181. Artioli G, Secco M, Addis A. 2019. The Vitruvian legacy: Mortars and binders before and after the Roman world. Artioli G, Oberti R (eds.). The contribution of mineralogy to cultural heritage. 151-202. https://doi.org/10.1180/EMU-notes.20.4
) as well as ettringite (9.16Å; Ca₆Al₂(SO₄)₃(OH)₁₂·26H₂O). It should be noted that the early/primary ettringite that precipitates in this system is not dangerous, although under site conditions (once the mortar has hardened and exposed to external agents: water, salts) it is susceptible to be transformed into delayed or secondary ettringite (8282. Collepardi M. 2003. A state-of-the-art review on delayed ettringite attack on concrete. Cem. Concr. Compos. 25(4-5):401-407. https://doi.org/10.1016/S0958-9465(02)00080-X
), which does give rise to expansion phenomena and the generation of gypsum (cracking, loss of mass and salt efflorescence pathologies) (8383. Palomo A, Blanco-Varela MT, Martínez-Ramírez S, Puertas F, Fortes C. 2002. Historic mortars: characterization and durability. new tendencies for research. Advanced Research Centre for Cultural Heritage: Interdisciplinary Projects. In Fifth Framework Programme Workshop, Brussels, Belgium.
).

Figure 7 shows the FTIR results. Thus, all mortars present a calcite stretching band domain (1450-1420 cm^-1) due to the carbonation process (CO₃ ^2-) and an asymmetric stretching band at 1070-1080 cm-¹ because of the quartz aggregate (Si-O-Si). Portlandite (3640 cm^-1) appears with higher intensity in CL90-S PL and decreases progressively in NHL2 and NHL5 but is not present in OPC. It should also be noted that in CL90S-PL there are aragonite stretching (1470 cm^-1) and bending (855 cm^-1) bands due to the early carbonation process under high humidity conditions (CO₃ ^2-).

In NHL2, NHL5 and OPC, the vibration bands of the quartz aggregate (1080-1190 cm^-1) mask the vibration bands of the calcium silicates. This causes them to appear to the right of the main quartz band (1070 cm^-1), at 990-1020 cm^-1, in the range reported in the literature (900-1120 cm^-1; (77. Varas MJ, Álvarez de Buergo M, Fort R. 2005. Natural cement as the precursor of Portland cement: Methodology for its identification. Cem. Concr. Res. 35(11):2055-2065. https://doi.org/10.1016/j.cemconres.2004.10.045
, 4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
, 8484. Farcas F, Touzé P. 2001. La spectrométrie infrarouge à transformée de Fourier (IRTF). Une méthode intéressante pour la caractérisation des ciments. Bull. Lab. Ponts Chaussées. 230(4350):77-88.
, 8585. Diekamp A, Stalder R, Konzett J, Mirwald PW. 2012. Lime mortar with natural hydraulic components: Characterisation of reaction rims with FTIR imaging in ATR-mode. Válek J, Hughes JJ, Groot CJWP (eds). In Historic Mortars: Characterisation, assessment and repair, RILEM Bookseries, vol 7. Springer, Dordrecht, 105-113. https://doi.org/10.1007/978-94-007-4635-0_8
)). The bands at 990-1020 cm^-1 and 515-520 cm^-1 correspond mainly to belite/larnite (C₂S), while the alite vibration bands (C₃S), at 880-950 cm^-1 according to previous authors (4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
, 8484. Farcas F, Touzé P. 2001. La spectrométrie infrarouge à transformée de Fourier (IRTF). Une méthode intéressante pour la caractérisation des ciments. Bull. Lab. Ponts Chaussées. 230(4350):77-88.
, 8585. Diekamp A, Stalder R, Konzett J, Mirwald PW. 2012. Lime mortar with natural hydraulic components: Characterisation of reaction rims with FTIR imaging in ATR-mode. Válek J, Hughes JJ, Groot CJWP (eds). In Historic Mortars: Characterisation, assessment and repair, RILEM Bookseries, vol 7. Springer, Dordrecht, 105-113. https://doi.org/10.1007/978-94-007-4635-0_8
), are not detected.

The TG-DTA curves show four main weight loss regions: at < 120 ºC, 120-200 ºC, 200-600 ºC and 600-900 ºC (2525. Ergenç D, Fort R, Varas-Muriel MJ, Álvarez de Buergo M. 2021. Mortars and plasters-How to characterize aerial mortars and plasters. Archaeol. Anthropol. Sci. 13:197. https://link.springer.com/article/10.1007/s12520-021-01398-x
, 3636. Moropoulou A, Bakolas A, Anagnostopoulou S. 2005. Composite materials in ancient structures. Cem. Concr. Compos. 27(2):295-300. https://doi.org/10.1016/j.cemconcomp.2004.02.018
, 3838. Ponce-Antón G, Cruz Zuluaga M, Ortega LA, Agirre Mauleon J. 2020. Petrographic and chemical-mineralogical characterization of mortars from the Cisterna at Amaiur Castle (Navarre, Spain). Minerals. 10(4):311. https://doi.org/10.3390/min10040311
, 3939. Fort R, Varas-Muriel MJ, Zoghlami K, Ergenç D, Zaddem A. 2024. Analytical characterisation of 1st- and 2nd-century Roman mortars at the Utica archaeological site (Tunisia): construction phases and provenance of the raw materials. J. Archaeol. Sci. Rep. 54:104404. https://doi.org/10.1016/j.jasrep.2024.104404
, 4848. Trezza MA, Scian AN. 2013. Aporte de las técnicas ATD/TG y espectroscopía FT-IR al estudio de la carbonatación de la matriz cementicia. Afinidad LXX, 562:112-117.
, 4949. Borsoi G, Santos Silva A, Menezes P, Candeias A, Mirão J. 2019. Analytical characterization of ancient mortars from the archaeological roman site of Pisões (Beja, Portugal). Constr. Build. Mater. 204:597-608. https://doi.org/10.1016/j.conbuildmat.2019.01.233
, 8686. Földvári M. 2011. Handbook of thermogravimetric system of minerals and its use in geological practice, Geological Institute of Hungary, Budapest.
). Table 1 shows the weight losses (%) in the different ranges.

The weight loss at < 120 ºC in all lime mortars (CL90-S PL, NHL2 and NHL5, < 3%) is due to the loss of water inside the mortars (T below 30 ºC). While in OPC the dehydration of characteristic cement minerals such as ettringite and tobermorite (C-S-H) also occurs (weight loss of 3.7% at 70-75 ºC, Table 1). Afterwards, in the range between 120 ºC and 200 ºC, in which the loss of crystallization water occurs, the weight loss of all the studied mortars is < 1%. These two ranges correspond to the dehydration phase of the samples (4949. Borsoi G, Santos Silva A, Menezes P, Candeias A, Mirão J. 2019. Analytical characterization of ancient mortars from the archaeological roman site of Pisões (Beja, Portugal). Constr. Build. Mater. 204:597-608. https://doi.org/10.1016/j.conbuildmat.2019.01.233
, 8686. Földvári M. 2011. Handbook of thermogravimetric system of minerals and its use in geological practice, Geological Institute of Hungary, Budapest.
).

In the next region, at 200-600 ºC, the weight loss is attributed to the loss of hydroxyl groups (dehydroxylation) (2525. Ergenç D, Fort R, Varas-Muriel MJ, Álvarez de Buergo M. 2021. Mortars and plasters-How to characterize aerial mortars and plasters. Archaeol. Anthropol. Sci. 13:197. https://link.springer.com/article/10.1007/s12520-021-01398-x
, 3636. Moropoulou A, Bakolas A, Anagnostopoulou S. 2005. Composite materials in ancient structures. Cem. Concr. Compos. 27(2):295-300. https://doi.org/10.1016/j.cemconcomp.2004.02.018
, 85-8785. Diekamp A, Stalder R, Konzett J, Mirwald PW. 2012. Lime mortar with natural hydraulic components: Characterisation of reaction rims with FTIR imaging in ATR-mode. Válek J, Hughes JJ, Groot CJWP (eds). In Historic Mortars: Characterisation, assessment and repair, RILEM Bookseries, vol 7. Springer, Dordrecht, 105-113. https://doi.org/10.1007/978-94-007-4635-0_8
86. Földvári M. 2011. Handbook of thermogravimetric system of minerals and its use in geological practice, Geological Institute of Hungary, Budapest.
87. Fort R, Varas-Muriel MJ, Ergenç D, Cassar J, Anastasi M, Vella NC. 2023. The technology of ancient lime mortars from the Zejtun Roman Villa (Malta). Archaeol. Anthropol. Sci. 15:15. https://doi.org/10.1007/s12520-022-01710-3
) and specifically to the thermal decomposition of portlandite at 400-415 ºC, which is the major component of the lime mortars (19-23% of loss), as well as OPC (17%).

Afterwards, the weight loss between 600 and 900 ºC corresponds to the loss of CO₂ because of the decomposition of calcium carbonate, which enables the determination of the CaCO₃ content within the sample (decarboxylation) (2525. Ergenç D, Fort R, Varas-Muriel MJ, Álvarez de Buergo M. 2021. Mortars and plasters-How to characterize aerial mortars and plasters. Archaeol. Anthropol. Sci. 13:197. https://link.springer.com/article/10.1007/s12520-021-01398-x
, 3636. Moropoulou A, Bakolas A, Anagnostopoulou S. 2005. Composite materials in ancient structures. Cem. Concr. Compos. 27(2):295-300. https://doi.org/10.1016/j.cemconcomp.2004.02.018
, 8787. Fort R, Varas-Muriel MJ, Ergenç D, Cassar J, Anastasi M, Vella NC. 2023. The technology of ancient lime mortars from the Zejtun Roman Villa (Malta). Archaeol. Anthropol. Sci. 15:15. https://doi.org/10.1007/s12520-022-01710-3
). In this way, all lime mortars (CL90-S PL, NHL2 and NHL5; 11-16%) present the highest loss percentages and OPC the lowest (~ 8%). Therefore, the total weight loss is higher in lime mortars (CL90-S PL, NHL2 and NHL5, 19-23%) than in OPC (17%) (Table 1).

Finally, the hydraulicity degree was calculated on the basis of the CO₂/H₂O ratio, so as this ratio decreases, the mortars are more hydraulic (2525. Ergenç D, Fort R, Varas-Muriel MJ, Álvarez de Buergo M. 2021. Mortars and plasters-How to characterize aerial mortars and plasters. Archaeol. Anthropol. Sci. 13:197. https://link.springer.com/article/10.1007/s12520-021-01398-x
, 3636. Moropoulou A, Bakolas A, Anagnostopoulou S. 2005. Composite materials in ancient structures. Cem. Concr. Compos. 27(2):295-300. https://doi.org/10.1016/j.cemconcomp.2004.02.018
, 3939. Fort R, Varas-Muriel MJ, Zoghlami K, Ergenç D, Zaddem A. 2024. Analytical characterisation of 1st- and 2nd-century Roman mortars at the Utica archaeological site (Tunisia): construction phases and provenance of the raw materials. J. Archaeol. Sci. Rep. 54:104404. https://doi.org/10.1016/j.jasrep.2024.104404
, 4949. Borsoi G, Santos Silva A, Menezes P, Candeias A, Mirão J. 2019. Analytical characterization of ancient mortars from the archaeological roman site of Pisões (Beja, Portugal). Constr. Build. Mater. 204:597-608. https://doi.org/10.1016/j.conbuildmat.2019.01.233
); Table 1 and Figure 8).

3.2. Petrophysical characterization

Surface properties: Colour is an important factor to consider, especially if mortars are going to be used on visible elements or facades. Based on the chromatic parameters, all the mortars are dull/matte (chroma C* < 6, white-grey and a* and b* values close to zero, grey) with high lightness values (L* = 70-92) and whiteness indexes (WI) greater than the yellowing ones (YI). CL90-S PL is the whitest mortar, followed by NHL2 and NHL5, while OPC is the most yellowish (Table 2). This is due to their mineralogy, in particular the presence of iron and aluminium phases such as brownmillerite in OPC and not in the lime mortars. Although there are other factors regarding the binders manufacturing process that impact on mortars colour, such as their raw materials composition and calcination temperatures.

The static contact angle test measurements provide information on the hydrophilic character of the mortars and therefore a first information on the behaviour of the mortar surface in relation to water.

Thus, CL90-S PL is classified as super-hydrophilic, that is, with a high affinity for water, as the droplet is quickly absorbed and not even the contact angle can be measured. While the rest of the mortars are hydrophilic, as they take longer to absorb the droplet and had θ < 90º. Therefore, the higher the hydraulicity of the binder and the droplet angle, the lower the hydrophilic character of the mortar. In consequence, OPC is the least hydrophilic mortar and presents the lowest standard deviation of all the mortars (Figure 9).

Dynamic properties: Table 3 shows the P-waves propagation values (Vp) of the mortars, which have a direct correlation with their hydraulic character.

On one hand, CL90-S PL and NHL2, as they present high porosities, shrinkage cracks and weak reactions between the paste and the aggregate, show an attenuation of the waves, presenting low Vp values (1258 m/s in CL90-S PL and 1352 m/s in NHL2). On the other hand, OPC presents a lower porosity and a higher density as well as cohesion, which impacts on Vp values > 3500m/s. While NHL5, with characteristics in between, has Vp values of 1617 m/s, more similar to those of CL90-S PL and NHL2. Thus, the differences in Vp values between lime and cement mortars reflects their different setting and curing speed.

Regarding the anisotropy, most of the mortars present a low complete anisotropy (dMm < 10) and similar total (dM) and relative anisotropy (dm) values. NHL2 and NHL5 are the most anisotropic mortars (dMm > 10%), and NHL2 shows the most different values between dM and dm (dM = 15.41% and dm = 2.93%), together with OPC (dM = 6.97% and dm = 0.69%). The latter denotes the existence of preferential orientations in these mortars and, therefore, planes of structural weakness possibly due to the manufacturing process. By contrast, CL90-S PL (dM = 2.39% and dm = 1.88%) is the least anisotropic mortar and shows a greater homogeneity.

Structural and hydric properties: Average values of the structural and hydric properties of the mortars studied are in Table 4. The real densities of all the mortars studied are high (~ 2460-2839 kg/m² by MIP), and higher than their bulk densities (~ 1492-2026 kg/m² by saturation), showing the presence of important percentages of porosity. Furthermore, the real and bulk density values obtained by MIP and by saturation are similar in most of the mortars, even though they were performed with different fluids and in different mortar samples.

In terms of total porosities by both PIM and saturation, all lime mortars (CL90-S PL, NHL2 and NHL5) have a porosity of around 40%, while in OPC it is half that (~ 20%). The highest percentages correspond to open porosity and Figure 10 shows the distribution curves of the different pore sizes of each mortar. In it, it has been noted that all the mortars are microporous (< 5 µm), with average pore sizes of 0.4 to 1.1 µm, and that the most microporous ones (94-97%) are the hydraulic mortars (NHL2, NHL5 and OPC). It should be noted that microporous mortars may tend to degrade if they have a low desorption capacity, as salts may precipitate inside them and the harmful effects are very difficult to detect, but very aggressive.

All the mortars studied have at least 3 pore-size populations in the microporosity range (0.01-0.1 µm, 0.1-1 µm and 1-3 µm, Figure 10) with average pore sizes between 0.36 µm and 1.13 µm depending on the mortar (Table 4).

The most remarkable intervals in NHL2, NHL5 and OPC are 0.1-1 µm and 1-3 µm, whereas CL90-S PL, in line with its appearance, presents a high percentage of macroporosity (> 5 µm, at least 1/3 of its total porosity). However, the value of the average pore size of CL90S-PL (0.36 µm) is lower than the actual average size, as the PIM technique does not consider pore sizes larger than 300 µm.

The mortars percentages of porosity are in relation to their compactness indexes (Table 4), around 0.6 in all lime mortars (CL90-S PL, NHL2 and NHL5) and 0.8 in OPC, the most dense and compact material.

Concerning the shapes of the pores, due to their large specific surfaces and small diameters (Table 4), all the mortars, especially the most hydraulic (NHL5 and OPC), have irregular pores and low (< 10) tortuosity values (capillary connections). This indicates straight connections that facilitate water ingress, so none of the mortars seem to offer water resistance, as previously verified by the static contact angle test. It also affects their water saturation capacity, absorption, desorption, capillary suction, and permeability.

As a result of the saturation test (Table 4), all lime mortars present similar values (28% of saturation in CL90-S PL and 24% in NHL2 and NHL5), whereas OPC has the lowest percentage of saturation (12%). This trend, in which lime mortars have a similar hydric behaviour while OPC does not, is also observed in the rest of the hydric tests.

First, during the water absorption test (Figure 11), all samples experience 90% water absorption in the first minutes. Their absorption coefficients are 24% (CL90-S PL), 22% (NHL2 and NHL5) and 11% (OPC). Thus, all lime mortars have a high absorption capacity, while it is the desorption phase that makes the difference between aerial and hydraulic mortars.

The CL90-S PL samples eliminate all their water content in approximately 24 hours, whereas the hydraulic mortars, and especially OPC, show a more distended desorption slope (much slower water loss) and leave a percentage of water retained inside. At the end of the test, NHL2 samples retain ~ 25% absorbed water, NHL5 ~ 35%, and OPC ~ 50% (Figure 11).

In the second place, during the capillary test, the lime mortars (CL90-S PL, NHL2 and NHL5) show a similar behaviour to the absorption test. The trend of their curves and their capillary absorption coefficients are different from those of OPC (Figure 12, Table 4). Thus, OPC absorbs up to 3 times less water than the lime mortars at the end of the test and shows a much slower and more distended absorption over time. OPC presents averages of 15.55 kg/m² of water per unit area, whereas the lime mortars have much steeper slopes and absorb the largest amounts of water (50.94 kg/m² - CL90-S PL, 56.18 kg/m² - NHL2, and 50.44 kg/m² - NHL5). The capillary absorption coefficients range from 0.15-0.16 kg/m²∙s^0.5 in lime mortars to (0,04 g/m²∙s^0.5 in OPC (Table 4), which is a significant difference regardless of the degree of hydraulicity of the limes.

Finally, the higher the hydraulicity, the lower the water vapour permeability of the mortars, which is directly related to the porosity (pore shape, average pore size, size distribution and capillary connections) (Table 4). CL90-S PL is the most permeable mortar (9.04∙10^-9 kg/m∙s∙Pa) with the highest values of water vapour flow rate, transmission rate and water vapour penetration, while NHL2 and NHL5 are half as permeable than CL90-S PL (4.37∙10^-9 and 4.30∙10^-9 kg/m∙s∙Pa). OPC is up to four times less permeable (2.17∙10^-9 kg/m∙s∙Pa) than CL90-S PL and half as permeable as NHL2 and NHL5.

Therefore, the only mortar sufficiently air permeable to perform the measurement with the air permeameter was CL90-S PL, which obtained a value of 3717 mD on average because of its high macroporosity (Table 4).

Mechanical properties: The mortars with the highest degree of hydraulicity show the highest hardness values after 90 days of curing (Table 5). In all cases, OPC doubles the hardness and strength of NHL2 and NHL5 and triples those of CL90-S PL, being the most suitable material for high mechanical stresses.

Thus, CL90-S PL shows the lowest values of Schmidt hardness (7 R) and microhardness (Leeb hardness, 213 HLD), while NHL2 and NHL5 present hardnesses of 18-20 R and microhardnesses of 260-267 HLD, although they are also classified as relatively soft materials. OPC, as expected, has the highest values with moderate micro-rebound (~ 428 HLD) and rebound (~ 30 R) resistance.

Finally, the calculated UCS values (Table 5) for the lime mortars (CL90-S PL ~ 13 MPa, NHL2 ~ 18 MPa and NHL5 ~ 19 MPa) were equivalent in strength to a soft stone (5-25 MPa) and for the OPC (~ 34 MPa) to a moderately hard stone (25-50 MPa).

In the case of CL90-S PL its UCS values are due to its high disintegration capacity, incomplete carbonation state and high porosity and cracking. NHL2 and NHL5 exhibit similar hardness and UCS values due to their porosity characteristics and slight-moderate chemical reactions. And the high values of OPC are mainly due to their higher curing speed (higher degree of carbonation) and lower porosity, in addition to the strong chemical reactions between the aggregate and the binder.