- Email: [email protected]
Technology of multilayer mortars applied in ancient floor mosaic substrates
Journal of Archaeological Science: Reports 20 (2018) 683–691
Contents lists available at ScienceDirect
Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep
Technology of multilayer mortars applied in ancient floor mosaic substrates ⁎
T
Vasiliki Pachta , Maria Stefanidou Lab. of Building Materials, Dept. of Civil Engineering, Aristotle University of Thessaloniki, PO Box: 482, 54124 Thessaloniki, Greece
A R T I C LE I N FO
A B S T R A C T
Keywords: Multilayer mortars Floor mosaics Substrate Historic mortars
Ancient floor mosaics started to be used from the 2nd millennium BCE, while their structure was established during the Classic Era. From the 4th century BCE they were wide spread and systemized around the Mediterranean Basin, as a mean to decorate public buildings, private houses or open spaces (i.e. atriums). Their substrate, as well as the successiveness of the layers played an important role on their structure, enhancing their durability to environmental factors and external loading. In the present study four multilayer mortar samples from Hellenistic and Roman floor mosaic substrates of monuments from Northern Greece, were analyzed. The methodology concerned in situ, macroscopic and microstructure observation, as well as determination of the physical, mechanical and chemical properties of each layer but at the same time of the structure as a whole. The evaluation of the test results led to the identification of the characteristics of each mortar layer, such as the binding system type, the Binder/Aggregate ratio and the application technique followed. According to the study, all multilayer samples presented similar characteristics, such as: stratigraphy of four very well compacted layers, whose thickness was decreased towards the surface layer, binding system based on lime and natural pozzolan, increase of the Binder/Aggregate ratio upwards and decrease of aggregates' maximum size and porosity. A significant aspect was the good adhesion of the mortar layers, which was attributed to the layers' compaction, as well as to different technological factors.
1. Introduction Floor mosaics were used from the 2nd millennium BCE as a mean to enhance the aesthetic and architectural value of constructions (Bruneau, 1987; Dumbabin, 1999; Rossi, 1970). An early example was found in Akrotiri, Thera (1700 BCE), decorated with colored, natural pebbles, inserted in a lime-based mortar layer (Palivou, 1999). From the 4th century BCE and on the building materials and techniques applied in floor mosaics were evolved and their construction started to follow specific criteria. Tesserae comprised of natural pebbles or shaped small pieces of stone of various colors (glass or brick tesserae was also used) depicting mythological scenes, animal, vegetal or geometric patterns (Lysandrou et al., 2017; Pachta, 2011; Robinson, 1930). At the same time, famous mosaic artists were acknowledged, such as Gnossis (4th–3rd cent. BCE), Sossos (3rd–2nd cent. BCE), Asklipiadis (2nd–1st cent. BCE), Dioskouridis (1st cent. BCE) and Aspasios (2nd–1st cent. BCE) (Pachta, 2011). The art of floor mosaics was wide spread during Roman times and started to decline from the 6th century CE (Early Christian times), when the technique of wall mosaics started to be evolved (Pachta, 2011). Diachronically the substrate of floor mosaics was a significant part of their structure, since it guaranteed their durability. Vitruvius (1960)
⁎
Corresponding author. E-mail addresses: [email protected] (V. Pachta), [email protected] (M. Stefanidou).
https://doi.org/10.1016/j.jasrep.2018.06.018 Received 13 April 2018; Received in revised form 31 May 2018; Accepted 14 June 2018 2352-409X/ © 2018 Published by Elsevier Ltd.
and Pliny (1971) were the first ancient authors (1st cent. CE), who described floor mosaics' substrate and proposed technical guidelines for their construction (materials, application). Vitruvius (1960) suggested three mortar layers (supra nucleus, nucleus and rudus) layered on a subground of pebbles (statumen), above a very well compacted subground. Supra nucleus, in which the tesserae were inserted, consisted of marble dust and fine sand, while nucleus contained crushed brick and rudus coarse aggregates. The thickness of the layers was decreased towards the surface, as well as the aggregates' maximum size. In all cases, lime was mixed with natural pozzolan (rudus, nucleus) and brick dust (nucleus, supra nucleus). Pliny on the other hand, proposed two layers applied in well-compacted earth (Pliny, 1971). The first consisted of pebbles or crushed brick and the second of lime, sand, charcoal and ash (thickness 15 cm). Later (15th century), Alberti (1965) suggested four layers: a layer of river pebbles (dimension less than a palm), a mortar layer with crushed brick (5–8 cm thick), a layer of tiles mixed with linseed oil and finally a well compacted mortar layer made of lime, pozzolan, sand and brick dust. The multilayer substrate, as suggested by the ancient authors, has been identified in various floor mosaics located around the Mediterranean basin, dated from the Classic era (beginning of the 4th
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 1. Studied floor mosaics substrates. a, b: Aiges Palace, c, d: Archaeological site of Pella, e, f: Archaeological site of Dion, g, h: Galerius complex.
creating random or ordered lines). Generally the increase of the effective contact area of the layers improves their adhesion. - the pore system of the subjacent layer (rudus), in order for the covering layer (nucleus) to be able to penetrate, as well as the rheological characteristics of the covering mortar - the layering (compaction) technique applied.
cent. BCE) until the Early Christian period (6th cent. CE). Some examples were found at Olynthos, Greece (Classic, beginning of the 4th cent. BCE, four layers) (Robinson, 1930), Pella, Greece (Hellenistic, 3rd cent. BCE, four layers) (Pachta, 2011), Herculaneum and Stabiae, Italy (Roman, 1st cent. CE, three layers) (Leone et al., 2016; Izzo et al., 2016), Histonium, Spain (Roman, 2nd cent. CE, three layers) (Puertas et al., 1994), Pafos, Cyprus (Roman, 2nd cent. CE, six layers) (Getty Conservation Institute, 1991), Jarash, Jordan (Early Christian, 6th cent CE, three to four layers) (Hamarneh and Abu-Jaber, 2017). The importance of the mortar layers' good compaction, in order to enhance the durability of these masterpieces, has been highlighted by the ancient authors (ICCROM, 1980; Vitruvius, 1960). On the other hand, there are many contemporary research studies regarding the adhesion of multilayer systems that focus on various parameters (i.e. mechanical strength, physical properties, chemical reaction), resuming that it is both a result of inter-layer adhesion and mechanical cohesion of the mortar layers (Botas et al., 2017; Goodwin and West, 1980; Groot, 1993; Stefanidou et al., 2017; Stolz et al., 2016; Thamboo and Dhanasekar, 2015). In most cases, it is pointed out that the enhancement of the cohesion of two or more mortar layers depends on a synergy of factors, such as (Groot, 1993; Stefanidou et al., 2017; Stolz et al., 2016):
In the present study, the analysis results of four multilayer mortar samples from Hellenistic and Roman floor mosaics' substrates of Northern Greece are presented and comparatively evaluated. The aim was to identify the materials used and the constructional techniques applied in these special technological artifacts and determine whether their construction followed the diachronic principles given by the ancient authors' instructions. For the better understanding of the results, the analyzed layers were named after Vitruvius, as statumen (substructure layer of pebbles), rudus (lower coarse mortar layer), nucleus (intermediate fine mortar layer) and supra nucleus (upper fine mortar layer that undertook tesserae). The current state of the substratum's was identified, as correlated to the layers' structural and physico-mechanical properties. The results showed the significant role that the substrate played for the maintenance of floor mosaics that whereas subjected to extreme environmental conditions (i.e. weathering, extreme humidity, loading) are still preserved.
- the consistency of the mortar layers - the high ratio of the effective contact area to the total surface (through scoring of the fresh mortar of the subjacent layer and 684
Roman
Aiges Palace, Room M2 (350–340 BCE)
Hellenistic
685
Galerius Complex, Basilica (3rd cent. CE)
Archaeological site of Dion, Polygon (3rd cent. CE)
Archaeological site of Pella, Helens abduction Villa (325–300 BCE)
Monument
Historic period
Table 1 Stratigraphy and characteristics of the substrates.
2.5YR 7/3 light red 7.5YR 7/4 pink 2.5R 8/1 white 2.5Y 7/1 light gray 10R 8/1 white 2.5R 7/2 light gray 2.5Y 6/2 light brownish gray 2.5Y 8/1 white 10R 7/2 pale red 10R 8/2 pinkish white 5YR 8/1 white 10YR 7/1 light gray 10R 6/4 pale red 10R 7/2 pale red 10YR 7/3 very pale brown 7.5YR 7/1 light gray
Rudus (8 cm) Statumen (8 cm) Supra nucleus (0.8 cm) Nucleus (2 cm) Rudus (3–4 cm) Statumen (6 cm) Supra nucleus (0.3 cm) Nucleus (2.5 cm) Rudus (4 cm) Statumen (7 cm) Supra nucleus (1 cm) Nucleus (5 cm) Rudus (7 cm) Statumen (8 cm)
Color (Munsel chart)
Supra nucleus (0.15 cm) Nucleus (4 cm)
Substrate layer/ thickness
Well compacted with low proportion of fine, natural, siliceous aggregates and crushed brick (0–1 mm) Very well compacted mortar with high proportion of natural, siliceous aggregates (0–8 mm). Very good adhesion of paste-aggregates Very well compacted mortar with high proportion of coarse, natural, siliceous aggregates (0–16 mm). Very good adhesion of paste-aggregates Coarse pebbles of calcitic origin, dimensioned around 8 × 10 cm Well-compacted mortar with low proportion of angular, fine aggregates of siliceous origin (0–4 mm). Good adhesion of paste-aggregates Very well compacted mortar with high proportion of angular aggregates of siliceous origin (0–6 mm). Very good adhesion of paste-aggregates Very well compacted mortar with high proportion of angular aggregates of siliceous origin (0–8 mm). Very good adhesion of paste-aggregates Coarse pebbles of calcitic origin, dimensioned around 6 × 8 cm Very well compacted mortar with low proportion of natural, siliceous aggregates and crushed brick (0–4 mm). Very good adhesion of paste-aggregates Very well compacted mortar with high proportion of natural, calcitic aggregates and crushed brick (0–6 mm). Very good adhesion of paste-aggregates Well compacted mortar with high proportion of natural aggregates of calcitic origin (0–8 mm). Unshaped stones of marble, dimensioned around 6 × 8 cm Very well compacted mortar with low proportion of fine, natural, siliceous aggregates and crushed brick (0–4 mm). Well compacted mortar with high proportion of natural, siliceous aggregates and crushed brick (0–8 mm). Well compacted mortar with high proportion of coarse, natural aggregates of siliceous origin (0–16 mm). Unshaped pieces of marble dimensioned around 4 × 8 cm.
Microscopic observation
V. Pachta, M. Stefanidou
Journal of Archaeological Science: Reports 20 (2018) 683–691
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 2. Preservation state of floor mosaics. a, b. Galerius Palace, Basilica, c, d. Archaeological site of Dion, Polygon.
Fig. 3. Sampling, macroscopic and microscopic photos of the Aiges Palace floor mosaic substrate (d., g. graphical scale in cm).
686
Journal of Archaeological Science: Reports 20 (2018) 683–691
25.96 35.17
-
1/2 0–16 2.2
-*: Not measured due to the low quantity of the sample.
33.2 Rudus (10 cm)
1.412
0–8 2.8 1.444 31.6
0–16 0–2 – -* 1.650 -* 30.4 -* Galerius Complex, Basilica
Rudus Supra nucleus (1–1.5 cm) Nucleus (7 cm)
0–8 1.684 26.2 Nucleus
Roman
Archaeological site of Dion, Polygon
Archaeological site of Pella, Helens abduction Villa
3.41
0–8 0–16 0–1 0–4 0–8 0–2 4.04 2.96 -* -* 1.45 -* 11.7 16.1 -* 18.6 25.8 20.2 Nucleus Rudus Supra nucleus Nucleus Rudus Supra nucleus
1.979 1.838 -* 1.725 1.689 1.701
0–2 -* 1.873 10.8 Supra nucleus Aiges Palace, Room M2 Hellenistic
Aiges Palace, Vergina (Room M2) Archaeological site of Pella (Dionysus villa) Archaeological site of Dion (Polygon Building) Galerius Complex, Thessaloniki (Basilica)
Regarding the monuments, the ancient city of Aiges was firstly inhabited during the 3rd millennium BCE and was the first capital of the Macedonian Kingdom. During Hellenistic times (4th cent. BCE) remained as the strategic landmark of the Macedonian Empire (Andronikos, 1984; Pantermalis, 1987). The Palace of the Emperor Filippos B′ (350–340 BCE), was located at Aiges, covering a total area of 9.300 m2. The building was square and organized around the central atrium with several rooms decorated with floor mosaics (Andronikos, 1984; Pantermalis, 1987). The analyzed substratum was sampled from the mosaic of Room M2, which was covered with angular limestone pieces dimensioned around 9 cm (Fig. 1a–b). The ancient city of Pella (located 50 km far from Aiges) was the capital of the Macedonian Empire from the end of the 5th century BCE (Siganidou and Lilimpaki-Akamati, 1996). It was constructed according to the Hippodamian system, consisting of the large agora, the palaces and various private houses (villas) (Siganidou and Lilimpaki-Akamati, 1996). The villa named after ‘Helen's abduction’ had a total area of 2.350 m2 and was organized around the central atrium. In the northern part of the villa there were three symposium halls (‘andronas’) decorated with elaborate floor mosaics. The ‘Helen's abduction’ mosaic, from where the mortars were sampled, covered the larger hall (325–300 BCE) and depicted the mythological scene of Helen being abducted by Theseus (Fig. 1c–d) (Siganidou and Lilimpaki-Akamati, 1996). It was attributed to the artist Gnossis, through an inscription (‘ΓΝΩΣΙΣ ΕΠΟΗΣΕΝ’) (Siganidou and Lilimpaki-Akamati, 1996). It was decorated with small sized tesserae of various colors (black, gray, white, red). The ancient city of Dion, located at the foothills of Mountain Olympus and sacred to Zeus, was known from the end of the 5th century BCE and was rebuilt during the 2nd century BCE (Pantermalis, 1997). It comprised of the large agora, the ancient theatre, city walls, multiple sanctuaries, baths, private houses and public buildings (Pantermalis, 1997). The Polygon building (3rd cent. CE) covered an area of 1300 m2 and was organized around a central atrium. The floor was decorated with a central scene of multicolored square tesserae dimensioned around 0.8 cm and perimetrically with zones of geometric and animal patterns with angular stone pieces (around 5 cm), from where the mortars were sampled (Fig. 1e–f). The Galerius Complex was built in the center of Thessaloniki during the 3rd century CE and concerned many monuments such as the palace of Galerius and the Basilica (Athanasiou et al., 2015). Basilica was the reception building of the complex, decorated with floor mosaics (Athanasiou et al., 2015). The mosaic, from where the mortar substratum was sampled, depicted geometric patterns constructed with multicolored (black, gray, white, red) tesserae dimensioned around 0.8 cm (Fig. 1g–h).
L:P
37.42 29.63 1/1.5
L:P:B
25.52 31.81 37.2 33.54 1/2.5 1/1
L:P L:P:B
43.55 24.52 L:P:B 1/2
27.54 31.69 27.24 34.03 27.33 27.07 33.44 38.20 36.60 33.30 35.80 21.55 L:P L:P L:P L:P L:P L:P:B 1/2 1/3 1/1 1/1.5 1/2.5 1/1
31.69 38.23
The analyzed mortars were sampled from four floor mosaic substrates of Hellenistic and Roman monuments of Northern Greece. The monuments concerned:
Siliceous, crushed brick Siliceous Siliceous Siliceous Siliceous Siliceous Calcitic, crushed brick Calcitic, crushed brick Calcitic Natural, crushed brick Natural, crushed brick Natural
1/1
L:P:B
SiO2 + Al2O3 + Fe2O3 Origin Gradation (mm)
Aggregates Comp. strength (MPa) Ap. spec. gravity Porosity (%) Layer Monument Hist. period
Table 2 Physical, mechanical and chemical and characteristics of the multilayer mortar samples.
2. Archaeological sites
CaO
B/A ratio (by weight)
Binding system L: lime P: pozzolan B: brick dust
Total oxides (% w/w)
V. Pachta, M. Stefanidou
3. Materials and methods Four multilayer substrates (one from each floor mosaic) were analyzed, leading to totally twelve mortar samples, including supra nucleus, nucleus and rudus (Table 1). All materials were carefully selected and sampled, in order to be authentic and not of subsequent repair works (relevant information was given by the sites' responsible). On site inspection assisted to understand the stratigraphy of the mosaic 687
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 4. Brick dust and crushed brick presence in nucleus of the Galerius Palace floor mosaic substrate.
Fig. 5. Aggregates and inclusions (piece of wood) found in rudus of Aiges Palace floor mosaic substrate.
(Pachta and Papayianni, 2016; Papayianni et al., 2012; Stefanidou et al., 2013). These remarks were based upon the comparative evaluation of the physico-chemical analysis results (microstructure analysis, aggregates' gradation, chemical analysis). All results were comparatively evaluated in order to identify the structural characteristics of the substrates, as well as the aspects that were taken into account during their construction (Pachta and Papayianni, 2016; Papayianni et al., 2012).
substrates, the general characteristics of each layer (thickness, color), as well as their state of conservation. A series of laboratory tests were conducted in order to determine their physical, mechanical and chemical properties (Papayianni et al., 2012; Stefanidou et al., 2013). The tests were performed in each mortar layer, which was separated carefully (in order not to create additional cracks and voids) with lancet. Microstructure observation was performed with stereoscope (Leica Wild M10) assisted by image analysis (ProgRes). Aggregates gradation was tested by hand grinding and sieving, according to EN1015-1:1998. Porosity and apparent specific gravity were performed according to RILEM CPC 11.3, while compressive strength was tested in cube-shaped samples of each layer, dimensioned around 4 × 4 × 4 cm (Papayianni et al., 2012). Finally, wet chemical analysis assisted by Atomic Absorption was performed in a fine fraction of the sample (< 75 μm), coming from the finest sieve of the aggregates' gradation testing. Additional comparative observations among samples were made, regarding their structure and constituents, such as binding system type, Binder/Aggregate (B/A) ratio and texture
4. Results and discussion According to the in situ survey, all mosaics were situated in a highly leveled ground, while special attention seemed to have been taken to the surface inclinations in order to avoid water accumulation. Taking into account the direct contact to humidity and the exposition to environmental factors, their current state was generally good. Detachment of the substrates' layers was mainly recorded on the edges of the mosaics, while loosening of tesserae was also partially detected (Fig. 2a–d). 688
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
4–8 cm thick, nucleus 2–5 cm and supra nucleus 0.5–1 cm), with the total thickness of the substrates to vary from 12 cm to 22 cm. In Table 2 the physical, mechanical and chemical characteristics of the mortar samples are presented, as well as their correlating properties, such as B/A ratio and binding system type. Fig. 3 illustrates the macroscopic and microscopic characteristics of the mortar layers of the substrate from the Aiges Palace mosaic. According to the evaluation of the results, all mortar layers were based on lime and pozzolan in a proportion ranging from 1:0.5 to 1:1. The ratio of pozzolan was usually increased in the intermediate mortar layer (nucleus). The systematic addition of pozzolan in these mortars probably emerged from the need to be resistant to intense humidity, as well as durable to loading (Stefanidou et al., 2013; Theodoridou et al., 2013). In some cases and only in the external layers (supra nucleus, nucleus), brick dust was added probably in an effort to enhance the hydraulic properties of the mortars and increase their resistance to humidity (Fig. 4) (Stefanidou et al., 2013). However, brick dust also played an aesthetic role in mosaics, since the reddish hue of the tesserae joints, made the colors and patterns of the mosaic decoration to be more vivid and realistic (Fig. 2a–d). Relevant research has been assessed in wall mosaics' case studies (Baraldi et al., 2016). Aggregates were natural, coming usually from nearby river deposits and were of siliceous and calcitic origin. Their shape was mostly rounded, while angular aggregates were also detected, enhancing the paste-aggregates' bonding (Fig. 3e, i) (Stefanidou and Papayianni, 2006). Crushed brick was added mainly in the upper layers (supra nucleus, nucleus), also in order to enhance mortars' hydraulic character (Fig. 4) (Stefanidou et al., 2013). The aggregates' maximum size (Table 2) was generally decreased towards the surface and their gradation ranged from 0 to 8 mm to 0–16 mm for rudus, 0–4 mm to 0–8 mm for nucleus and 0–1 mm to 0–2 mm for supra nucleus (Fig. 5). Additionally, the B/A ratio was increased towards the upper layers, ranging from 1/2 to 1/3 for rudus, 1/1.5 to 1/2 for nucleus and 1/1 for
Fig. 6. Aggregates' gradation of rudus and nucleus from the Aiges Palace floor mosaic substrate.
All substrates consisted of four layers (Table 1, Fig. 1b, d, f, h), reflecting the typical floor mosaic substrate (Baraldi et al., 2016; Getty Conservation Institute, 2003; ICCROM, 1980; Vitruvius, 1960). In Table 1 the stratigraphy of the substrates and the structural characteristics of each layer are presented. In the Hellenistic substrates (Aiges Palace, Pella), statumen was made of coarse, natural pebbles, while in the Roman ones (Dion, Galerius Complex) angular pieces of marble were used. The diameter of the stone pieces used in the statumen of both period monuments was in the range of 6–10 mm. Regarding the mortar layers (Fig. 3a, b, d, e, g, h), their thickness was decreased towards the surface (rudus was
Fig. 7. Correlation of the porosity and apparent specific gravity of the mortar layers, according to their stratigraphy. 689
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 8. Microscopic photo of the contact area of the mortar layers (nucleus-rudus) of Aiges Palace (left) and archaeological site of Pella (right) mosaic substrates.
Other parameters that could influence the adherence between layers, as defined by literature (Stefanidou et al., 2017; Stolz et al., 2016) could be the increase of the effective contact area through the deliberate roughness of the subjacent mortar layer, its higher porosity and the rheological characteristics of the covering layer. According to the study, the porosity of the mortar layers was generally decreased to the surface, while in many cases extreme roughness of the contact surface between layers was observed. Resuming the evaluation of the test results it can be concluded that for > 7 centuries (4th cent. BCE to 3rd cent. CE) ancient floor mosaics presented similar characteristics regarding their substrate. Four, very well compacted layers were usually observed, according to the ancient authors descriptions (statumen, rudus, nucleus, supra nucleus), while several parameters were taken into account for the construction of the multilayer mortar systems (i.e. binding system type, B/A ratio, aggregates' type and gradation, application techniques).
supra nucleus. Taking into account the decrease of the upper layers' thickness, the binder-rich mortars in combination with the decrease of their aggregates' size was necessary for their efficient application, their good compaction and probably the increase of their cohesion to the subjacent layers. In some cases, inclusions were added in the matrix, such as pieces of wood, straw and reed, carefully selected and of specific geometry, probably as a mean to increase the stability of the layers and decrease their cracking tendency (Fig. 6) (Stefanidou et al., 2012). Generally, inclusions were inspected in the lower mortar layer (rudus), which was thicker and probably more prone to cracking. Regarding physical properties (Table 2), porosity was generally decreased towards the upper layer, probably as a result of the aggregates' size decrease and also the increase of the B/A ratio (Stefanidou and Papayianni, 2006) (Fig. 7). The values of porosity were generally low, compared to the ones found in historic mortars of other applications (i.e. structural mortars) (Pachta and Papayianni, 2016; Papayianni et al., 2012; Stefanidou et al., 2013), depicting probably the high compaction of the layers during their application. Apparent specific gravity varied according to porosity, B/A ratio and type - gradation of aggregates, but was generally high. Compressive strength on the other hand ranged from 1.5 MPa to 4 MPa (Table 2). These values were higher than the strength recorded in historic lime-pozzolan mortars of these periods and of other applications (i.e. structural mortars) which were usually around 0.8–2 MPa (Pachta and Papayianni, 2016; Papayianni et al., 2012; Stefanidou et al., 2013). Keeping in mind the continuous contact with the ground and the loads that floor mosaics suffered, the high values of compressive strength could be closely related with the special attention given to the selection of the raw materials and the application techniques followed. A significant parameter that should be taken into account when studying multilayer mortar systems as defined by literature (Botas et al., 2017; Stefanidou et al., 2017) is the contact area between the layers. In the studied mosaic substrates, the interface between the layers was in most cases extremely stable, which led to a very difficult detachment of the layers for analyzing their individual properties. As presented in Fig. 8, the layers' contact was mostly mechanical, while the upper layer followed the surface irregularities of the subjacent. No voids or detachments were observed in the contact zones, indicating the special care taken during manufacture. To provide this firm adhesion of the layers, many factors could attribute, with the main one to be the very good compaction. As described by Vitruvius (1960), the ground was carefully stabilized before the construction of the floor mosaic substrate. The role of statumen was probably to level the subground, by flattening inclinations and ground irregularities, as well as to enhance the stability of the structure by undertaking external loads and vibrations. All mortar layers (rudus, nucleus, supra nucleus) were very well compacted with wooden rammers ensuring their cohesion to each other.
5. Conclusions Diachronically floor mosaics had a two-fold role: apart from enhancing the architectural value of constructions as decorative elements, through their stable substructure they served as durable constructional elements (floors). In all cases, the multilayer substrate played an important role for their durability and preservation, in accordance with the care taken to the ground on which they were applied. According to the study, the substrate of Hellenistic (4th cent. BCE) and Roman (3rd cent. CE) floor mosaics located in monuments of Northern Greece, presented similar characteristics, such as: - Stratigraphy of four very well compacted layers, whose thickness was decreased to the surface: statumen (substructure layer of natural pebbles or unshaped stone pieces), rudus (lower coarse mortar layer), nucleus (intermediate fine mortar layer) and supra nucleus (upper fine mortar layer that undertook tesserae). - The good adhesion of the mortar layers was a significant parameter taken into account that ensured the stability of the substrate. To this direction, a synergy of factors contributed, such as the good compaction of each layer (with wooden rammers), the increase of the surface contact area (through scoring of the fresh material of the subjacent layer), as well as the individual characteristics of each layer (i.e. decrease of the aggregates' maximum size and increase of the B/A ratio towards the upper layer). - The binding system was based on lime and natural pozzolan, while brick dust was added in the upper layers (nucleus, supra nucleus) in an effort to enhance the hydraulic character of the mortars and have a colored result. - The B/A ratio was increased towards the surface layer, while porosity and aggregates' gradation were decreased. Natural aggregates of river origin were used, whereas crushed brick was also added. 690
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
- Physical and mechanical properties such as porosity, apparent specific gravity and compressive strength were generally beyond the values found in other types of historic mortars (i.e. structural mortars). Porosity was low (10–16% for Aiges Palace floor mosaic), while compressive strength was generally high (2–4 MPa). According to literature, the corresponding properties of historic structural mortars range around 20–30% for porosity and 0.8–2 MPa for compressive strength.
S.F., Grifa, C., Guarino, V., Mercurio, M., Morra, V., Langella, A., 2016. The art of building in the Roman period (89 B.C–79 A.D.): mortars, plasters and mosaic floors from ancient Stabiae (Naples, Italy). Constr. Build. Mater. 117, 129–143. Leone, G., De Vita, A., Magnani, A., Rossi, C., 2016. Characterization of archaeological mortars from Herculaneum. Thermochim. Acta 624, 86–94. Lysandrou, V., Cerra, D., Agapiou, A., Charalambous, E., Hadjimitsis, D.G., 2017. Towards a spectral library of Roman to Early Christian Cypriot floor mosaics. J. Archaeol. Sci. Rep. 14, 782–791. Pachta, V., 2011. Study of the Technological Evolution of Mortars (Dissertation). Dept of Civil Engineering, Aristotle University of Thessaloniki, Thessaloniki. Pachta, V., Papayianni, I., 2016. Design and application of a data system for the comparative study of historic mortars. In: Ioannides, M. (Ed.), Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 10058 LNCS. Springer Verlag, Cyprus, pp. 701–710. Palivou, K., 1999. Akrotiri of Thera: the constructional technology. In: Library of the Archaeological Society at Athens, Nr 183, Athens, (in Greek). Pantermalis, D., 1987. The Tiling of the Palace of Vergina, Amitos: Honorary Volume for Prof. M. Andronikos, Thessaloniki. (in Greek). Pantermalis, D., 1997. Dion: The Archaeological Site and Museum. Adam Editions, Athens. Papayianni, I., Pachta, V., Stefanidou, Μ., 2012. Analysis of ancient mortars and design of compatible repair mortars: the case study of Odion of the archaeological site of Dion. Constr. Build. Mater. 40, 84–92. Pliny, 1971. Natural History, Book XXXVI. W, Heinemann Ltd., London. Puertas, F., Blanco-Varela, M.T., Palomo, A., Ortega-Calvo, J.J., Saiz-Jimenez, C., 1994. Decay of Roman and repair mortars in mosaics from Italica, Spain. Sci. Total Environ. 153, 123–131. Robinson, D.M., 1930. Excavations at Olynthus, part II, architecture and sculpture: houses and other buildings. In: The Johns Hopkins University, Studies in Archeology No. 9. Rossi, F., 1970. Mosaics. A Survey of Their History and Techniques. Pall Mall Press, London. Siganidou, M., Lilimpaki-Akamati, M., 1996. Pella - The Capital of Macedonians. Ministry of Culture, Athens. Stefanidou, M., Papayianni, I., 2006. The role of aggregates on the structure and properties of lime mortars. Cem. Concr. Compos. 27, 914–919. Stefanidou, M., Papayianni, I., Pachta, V., 2012. Evaluation of inclusions in mortars of different historic periods from Greek monuments. Archaeometry 54, 737–751. Stefanidou, M., Pachta, V., Konopissi, S., Karkadelidou, F., Papayianni, I., 2013. Analysis and characterization of hydraulic mortars from ancient cisterns and baths in Greece. Mater. Struct. 47 (4), 571–580. Stefanidou, M., Kesikidou, F., Antoniadis, K., 2017. Application techniques for high performance traditional renders. Procedia Environ Sci 38, 242–247. Stolz, C.M., Masuero, A.B., Pagnussat, D.T., Kirchheim, A.P., 2016. Influence of substrate texture on the tensile and shear bond strength of rendering mortars. Constr. Build. Mater. 128, 298–307. Thamboo, J.A., Dhanasekar, M., 2015. Characterisation of thin layer polymer cement mortared concrete masonry bond. Constr. Build. Mater. 82, 71–98. Theodoridou, M., Ioannou, I., Philokyprou, M., 2013. New evidence of early use of artificial pozzolanic material in mortars. J. Archaeol. Sci. 40, 3263–3269. Vitruvius, 1960. The Ten Books of Architecture, Translated by Morgan M.H. Dover Publications Inc., New York.
These observations lead to the conclusion that ancient floor mosaic substrates were probably constructed according to diachronic principles, following specific requirements. The raw materials used were carefully selected and applied in each mortar layer, while special attention was given to the layering techniques. As a result, durable substrates were achieved that guaranteed the longevity of these masterpieces, preserved until nowadays. References Alberti, L.B., 1965. In: Rykwert, J., Tiranti, A. (Eds.), Ten Books on Architecture, (J. Leoni (translation), London). Andronikos, M., 1984. Vergina - the Royal Tombs and the Ancient City. Ekdotike Athenon, Athens. Athanasiou, F., Malama, V., Miza, M., Sarantidou, M., 2015. The Restoration of the Remnants of the Galerius Complex in Thessaloniki (1994–2014), Volume A', Thessaloniki. Ephorate of Antiquities of the City of Thessaloniki, Thessaloniki. Baraldi, P., Bracci, S., Cristoferi, E., Fiorentino, S., Vandini, M., Venturi, E., 2016. Pigment characterization of drawings and painted layers under 5th–7th centuries wall mosaics from Ravenna (Italy). J. Cult. Herit. 21, 802–808. Botas, S., Veiga, R., Velosa, A., 2017. Air lime mortars for conservation of historic tiles: bond strength of new mortars to old tiles. Constr. Build. Mater. 145, 426–434. Bruneau, P., 1987. La mosaïque antique, Lectures en Sorbonne, collection dirigée. L' Université de Paris-Sorbonne, Paris. Dumbabin, K.M.D., 1999. Mosaics of the Greek and Roman World. Cambridge University Press, New York. Getty Conservation Institute, 1991. The Conservation of the Orpheus Mosaic at Paphos, Cyprus, California. Getty Conservation Institute, 2003. Israel Antiquities Authority, Illustrated Glossary: Mosaics in Situ Project: Definitions of Terms Used for the Graphic Documentation of in Situ Floor Mosaics, Los Angeles. Goodwin, J.F., West, H.W.H., 1980. A review of the literature on mortar/brick bond. In: The British Ceramic Research Association, Techn. Note No. 308. Groot, C., 1993. Effects of Water on Mortar-brick Bond. Delft University, Delft. Hamarneh, C., Abu-Jaber, N., 2017. Mosaic pavement mortar production in Gerasa in the Byzantine period. Archaeol. Res. Asia 9, 22–33. ICCROM, 1980. Mosaics No2, Safeguard, Perigueux, Rome. Izzo, F., Arizzi, A., Cappelletti, P., Cultrone, G., De Bonis, A., Germinario, C., Graziano,
691
Contents lists available at ScienceDirect
Journal of Archaeological Science: Reports journal homepage: www.elsevier.com/locate/jasrep
Technology of multilayer mortars applied in ancient floor mosaic substrates ⁎
T
Vasiliki Pachta , Maria Stefanidou Lab. of Building Materials, Dept. of Civil Engineering, Aristotle University of Thessaloniki, PO Box: 482, 54124 Thessaloniki, Greece
A R T I C LE I N FO
A B S T R A C T
Keywords: Multilayer mortars Floor mosaics Substrate Historic mortars
Ancient floor mosaics started to be used from the 2nd millennium BCE, while their structure was established during the Classic Era. From the 4th century BCE they were wide spread and systemized around the Mediterranean Basin, as a mean to decorate public buildings, private houses or open spaces (i.e. atriums). Their substrate, as well as the successiveness of the layers played an important role on their structure, enhancing their durability to environmental factors and external loading. In the present study four multilayer mortar samples from Hellenistic and Roman floor mosaic substrates of monuments from Northern Greece, were analyzed. The methodology concerned in situ, macroscopic and microstructure observation, as well as determination of the physical, mechanical and chemical properties of each layer but at the same time of the structure as a whole. The evaluation of the test results led to the identification of the characteristics of each mortar layer, such as the binding system type, the Binder/Aggregate ratio and the application technique followed. According to the study, all multilayer samples presented similar characteristics, such as: stratigraphy of four very well compacted layers, whose thickness was decreased towards the surface layer, binding system based on lime and natural pozzolan, increase of the Binder/Aggregate ratio upwards and decrease of aggregates' maximum size and porosity. A significant aspect was the good adhesion of the mortar layers, which was attributed to the layers' compaction, as well as to different technological factors.
1. Introduction Floor mosaics were used from the 2nd millennium BCE as a mean to enhance the aesthetic and architectural value of constructions (Bruneau, 1987; Dumbabin, 1999; Rossi, 1970). An early example was found in Akrotiri, Thera (1700 BCE), decorated with colored, natural pebbles, inserted in a lime-based mortar layer (Palivou, 1999). From the 4th century BCE and on the building materials and techniques applied in floor mosaics were evolved and their construction started to follow specific criteria. Tesserae comprised of natural pebbles or shaped small pieces of stone of various colors (glass or brick tesserae was also used) depicting mythological scenes, animal, vegetal or geometric patterns (Lysandrou et al., 2017; Pachta, 2011; Robinson, 1930). At the same time, famous mosaic artists were acknowledged, such as Gnossis (4th–3rd cent. BCE), Sossos (3rd–2nd cent. BCE), Asklipiadis (2nd–1st cent. BCE), Dioskouridis (1st cent. BCE) and Aspasios (2nd–1st cent. BCE) (Pachta, 2011). The art of floor mosaics was wide spread during Roman times and started to decline from the 6th century CE (Early Christian times), when the technique of wall mosaics started to be evolved (Pachta, 2011). Diachronically the substrate of floor mosaics was a significant part of their structure, since it guaranteed their durability. Vitruvius (1960)
⁎
Corresponding author. E-mail addresses: [email protected] (V. Pachta), [email protected] (M. Stefanidou).
https://doi.org/10.1016/j.jasrep.2018.06.018 Received 13 April 2018; Received in revised form 31 May 2018; Accepted 14 June 2018 2352-409X/ © 2018 Published by Elsevier Ltd.
and Pliny (1971) were the first ancient authors (1st cent. CE), who described floor mosaics' substrate and proposed technical guidelines for their construction (materials, application). Vitruvius (1960) suggested three mortar layers (supra nucleus, nucleus and rudus) layered on a subground of pebbles (statumen), above a very well compacted subground. Supra nucleus, in which the tesserae were inserted, consisted of marble dust and fine sand, while nucleus contained crushed brick and rudus coarse aggregates. The thickness of the layers was decreased towards the surface, as well as the aggregates' maximum size. In all cases, lime was mixed with natural pozzolan (rudus, nucleus) and brick dust (nucleus, supra nucleus). Pliny on the other hand, proposed two layers applied in well-compacted earth (Pliny, 1971). The first consisted of pebbles or crushed brick and the second of lime, sand, charcoal and ash (thickness 15 cm). Later (15th century), Alberti (1965) suggested four layers: a layer of river pebbles (dimension less than a palm), a mortar layer with crushed brick (5–8 cm thick), a layer of tiles mixed with linseed oil and finally a well compacted mortar layer made of lime, pozzolan, sand and brick dust. The multilayer substrate, as suggested by the ancient authors, has been identified in various floor mosaics located around the Mediterranean basin, dated from the Classic era (beginning of the 4th
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 1. Studied floor mosaics substrates. a, b: Aiges Palace, c, d: Archaeological site of Pella, e, f: Archaeological site of Dion, g, h: Galerius complex.
creating random or ordered lines). Generally the increase of the effective contact area of the layers improves their adhesion. - the pore system of the subjacent layer (rudus), in order for the covering layer (nucleus) to be able to penetrate, as well as the rheological characteristics of the covering mortar - the layering (compaction) technique applied.
cent. BCE) until the Early Christian period (6th cent. CE). Some examples were found at Olynthos, Greece (Classic, beginning of the 4th cent. BCE, four layers) (Robinson, 1930), Pella, Greece (Hellenistic, 3rd cent. BCE, four layers) (Pachta, 2011), Herculaneum and Stabiae, Italy (Roman, 1st cent. CE, three layers) (Leone et al., 2016; Izzo et al., 2016), Histonium, Spain (Roman, 2nd cent. CE, three layers) (Puertas et al., 1994), Pafos, Cyprus (Roman, 2nd cent. CE, six layers) (Getty Conservation Institute, 1991), Jarash, Jordan (Early Christian, 6th cent CE, three to four layers) (Hamarneh and Abu-Jaber, 2017). The importance of the mortar layers' good compaction, in order to enhance the durability of these masterpieces, has been highlighted by the ancient authors (ICCROM, 1980; Vitruvius, 1960). On the other hand, there are many contemporary research studies regarding the adhesion of multilayer systems that focus on various parameters (i.e. mechanical strength, physical properties, chemical reaction), resuming that it is both a result of inter-layer adhesion and mechanical cohesion of the mortar layers (Botas et al., 2017; Goodwin and West, 1980; Groot, 1993; Stefanidou et al., 2017; Stolz et al., 2016; Thamboo and Dhanasekar, 2015). In most cases, it is pointed out that the enhancement of the cohesion of two or more mortar layers depends on a synergy of factors, such as (Groot, 1993; Stefanidou et al., 2017; Stolz et al., 2016):
In the present study, the analysis results of four multilayer mortar samples from Hellenistic and Roman floor mosaics' substrates of Northern Greece are presented and comparatively evaluated. The aim was to identify the materials used and the constructional techniques applied in these special technological artifacts and determine whether their construction followed the diachronic principles given by the ancient authors' instructions. For the better understanding of the results, the analyzed layers were named after Vitruvius, as statumen (substructure layer of pebbles), rudus (lower coarse mortar layer), nucleus (intermediate fine mortar layer) and supra nucleus (upper fine mortar layer that undertook tesserae). The current state of the substratum's was identified, as correlated to the layers' structural and physico-mechanical properties. The results showed the significant role that the substrate played for the maintenance of floor mosaics that whereas subjected to extreme environmental conditions (i.e. weathering, extreme humidity, loading) are still preserved.
- the consistency of the mortar layers - the high ratio of the effective contact area to the total surface (through scoring of the fresh mortar of the subjacent layer and 684
Roman
Aiges Palace, Room M2 (350–340 BCE)
Hellenistic
685
Galerius Complex, Basilica (3rd cent. CE)
Archaeological site of Dion, Polygon (3rd cent. CE)
Archaeological site of Pella, Helens abduction Villa (325–300 BCE)
Monument
Historic period
Table 1 Stratigraphy and characteristics of the substrates.
2.5YR 7/3 light red 7.5YR 7/4 pink 2.5R 8/1 white 2.5Y 7/1 light gray 10R 8/1 white 2.5R 7/2 light gray 2.5Y 6/2 light brownish gray 2.5Y 8/1 white 10R 7/2 pale red 10R 8/2 pinkish white 5YR 8/1 white 10YR 7/1 light gray 10R 6/4 pale red 10R 7/2 pale red 10YR 7/3 very pale brown 7.5YR 7/1 light gray
Rudus (8 cm) Statumen (8 cm) Supra nucleus (0.8 cm) Nucleus (2 cm) Rudus (3–4 cm) Statumen (6 cm) Supra nucleus (0.3 cm) Nucleus (2.5 cm) Rudus (4 cm) Statumen (7 cm) Supra nucleus (1 cm) Nucleus (5 cm) Rudus (7 cm) Statumen (8 cm)
Color (Munsel chart)
Supra nucleus (0.15 cm) Nucleus (4 cm)
Substrate layer/ thickness
Well compacted with low proportion of fine, natural, siliceous aggregates and crushed brick (0–1 mm) Very well compacted mortar with high proportion of natural, siliceous aggregates (0–8 mm). Very good adhesion of paste-aggregates Very well compacted mortar with high proportion of coarse, natural, siliceous aggregates (0–16 mm). Very good adhesion of paste-aggregates Coarse pebbles of calcitic origin, dimensioned around 8 × 10 cm Well-compacted mortar with low proportion of angular, fine aggregates of siliceous origin (0–4 mm). Good adhesion of paste-aggregates Very well compacted mortar with high proportion of angular aggregates of siliceous origin (0–6 mm). Very good adhesion of paste-aggregates Very well compacted mortar with high proportion of angular aggregates of siliceous origin (0–8 mm). Very good adhesion of paste-aggregates Coarse pebbles of calcitic origin, dimensioned around 6 × 8 cm Very well compacted mortar with low proportion of natural, siliceous aggregates and crushed brick (0–4 mm). Very good adhesion of paste-aggregates Very well compacted mortar with high proportion of natural, calcitic aggregates and crushed brick (0–6 mm). Very good adhesion of paste-aggregates Well compacted mortar with high proportion of natural aggregates of calcitic origin (0–8 mm). Unshaped stones of marble, dimensioned around 6 × 8 cm Very well compacted mortar with low proportion of fine, natural, siliceous aggregates and crushed brick (0–4 mm). Well compacted mortar with high proportion of natural, siliceous aggregates and crushed brick (0–8 mm). Well compacted mortar with high proportion of coarse, natural aggregates of siliceous origin (0–16 mm). Unshaped pieces of marble dimensioned around 4 × 8 cm.
Microscopic observation
V. Pachta, M. Stefanidou
Journal of Archaeological Science: Reports 20 (2018) 683–691
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 2. Preservation state of floor mosaics. a, b. Galerius Palace, Basilica, c, d. Archaeological site of Dion, Polygon.
Fig. 3. Sampling, macroscopic and microscopic photos of the Aiges Palace floor mosaic substrate (d., g. graphical scale in cm).
686
Journal of Archaeological Science: Reports 20 (2018) 683–691
25.96 35.17
-
1/2 0–16 2.2
-*: Not measured due to the low quantity of the sample.
33.2 Rudus (10 cm)
1.412
0–8 2.8 1.444 31.6
0–16 0–2 – -* 1.650 -* 30.4 -* Galerius Complex, Basilica
Rudus Supra nucleus (1–1.5 cm) Nucleus (7 cm)
0–8 1.684 26.2 Nucleus
Roman
Archaeological site of Dion, Polygon
Archaeological site of Pella, Helens abduction Villa
3.41
0–8 0–16 0–1 0–4 0–8 0–2 4.04 2.96 -* -* 1.45 -* 11.7 16.1 -* 18.6 25.8 20.2 Nucleus Rudus Supra nucleus Nucleus Rudus Supra nucleus
1.979 1.838 -* 1.725 1.689 1.701
0–2 -* 1.873 10.8 Supra nucleus Aiges Palace, Room M2 Hellenistic
Aiges Palace, Vergina (Room M2) Archaeological site of Pella (Dionysus villa) Archaeological site of Dion (Polygon Building) Galerius Complex, Thessaloniki (Basilica)
Regarding the monuments, the ancient city of Aiges was firstly inhabited during the 3rd millennium BCE and was the first capital of the Macedonian Kingdom. During Hellenistic times (4th cent. BCE) remained as the strategic landmark of the Macedonian Empire (Andronikos, 1984; Pantermalis, 1987). The Palace of the Emperor Filippos B′ (350–340 BCE), was located at Aiges, covering a total area of 9.300 m2. The building was square and organized around the central atrium with several rooms decorated with floor mosaics (Andronikos, 1984; Pantermalis, 1987). The analyzed substratum was sampled from the mosaic of Room M2, which was covered with angular limestone pieces dimensioned around 9 cm (Fig. 1a–b). The ancient city of Pella (located 50 km far from Aiges) was the capital of the Macedonian Empire from the end of the 5th century BCE (Siganidou and Lilimpaki-Akamati, 1996). It was constructed according to the Hippodamian system, consisting of the large agora, the palaces and various private houses (villas) (Siganidou and Lilimpaki-Akamati, 1996). The villa named after ‘Helen's abduction’ had a total area of 2.350 m2 and was organized around the central atrium. In the northern part of the villa there were three symposium halls (‘andronas’) decorated with elaborate floor mosaics. The ‘Helen's abduction’ mosaic, from where the mortars were sampled, covered the larger hall (325–300 BCE) and depicted the mythological scene of Helen being abducted by Theseus (Fig. 1c–d) (Siganidou and Lilimpaki-Akamati, 1996). It was attributed to the artist Gnossis, through an inscription (‘ΓΝΩΣΙΣ ΕΠΟΗΣΕΝ’) (Siganidou and Lilimpaki-Akamati, 1996). It was decorated with small sized tesserae of various colors (black, gray, white, red). The ancient city of Dion, located at the foothills of Mountain Olympus and sacred to Zeus, was known from the end of the 5th century BCE and was rebuilt during the 2nd century BCE (Pantermalis, 1997). It comprised of the large agora, the ancient theatre, city walls, multiple sanctuaries, baths, private houses and public buildings (Pantermalis, 1997). The Polygon building (3rd cent. CE) covered an area of 1300 m2 and was organized around a central atrium. The floor was decorated with a central scene of multicolored square tesserae dimensioned around 0.8 cm and perimetrically with zones of geometric and animal patterns with angular stone pieces (around 5 cm), from where the mortars were sampled (Fig. 1e–f). The Galerius Complex was built in the center of Thessaloniki during the 3rd century CE and concerned many monuments such as the palace of Galerius and the Basilica (Athanasiou et al., 2015). Basilica was the reception building of the complex, decorated with floor mosaics (Athanasiou et al., 2015). The mosaic, from where the mortar substratum was sampled, depicted geometric patterns constructed with multicolored (black, gray, white, red) tesserae dimensioned around 0.8 cm (Fig. 1g–h).
L:P
37.42 29.63 1/1.5
L:P:B
25.52 31.81 37.2 33.54 1/2.5 1/1
L:P L:P:B
43.55 24.52 L:P:B 1/2
27.54 31.69 27.24 34.03 27.33 27.07 33.44 38.20 36.60 33.30 35.80 21.55 L:P L:P L:P L:P L:P L:P:B 1/2 1/3 1/1 1/1.5 1/2.5 1/1
31.69 38.23
The analyzed mortars were sampled from four floor mosaic substrates of Hellenistic and Roman monuments of Northern Greece. The monuments concerned:
Siliceous, crushed brick Siliceous Siliceous Siliceous Siliceous Siliceous Calcitic, crushed brick Calcitic, crushed brick Calcitic Natural, crushed brick Natural, crushed brick Natural
1/1
L:P:B
SiO2 + Al2O3 + Fe2O3 Origin Gradation (mm)
Aggregates Comp. strength (MPa) Ap. spec. gravity Porosity (%) Layer Monument Hist. period
Table 2 Physical, mechanical and chemical and characteristics of the multilayer mortar samples.
2. Archaeological sites
CaO
B/A ratio (by weight)
Binding system L: lime P: pozzolan B: brick dust
Total oxides (% w/w)
V. Pachta, M. Stefanidou
3. Materials and methods Four multilayer substrates (one from each floor mosaic) were analyzed, leading to totally twelve mortar samples, including supra nucleus, nucleus and rudus (Table 1). All materials were carefully selected and sampled, in order to be authentic and not of subsequent repair works (relevant information was given by the sites' responsible). On site inspection assisted to understand the stratigraphy of the mosaic 687
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 4. Brick dust and crushed brick presence in nucleus of the Galerius Palace floor mosaic substrate.
Fig. 5. Aggregates and inclusions (piece of wood) found in rudus of Aiges Palace floor mosaic substrate.
(Pachta and Papayianni, 2016; Papayianni et al., 2012; Stefanidou et al., 2013). These remarks were based upon the comparative evaluation of the physico-chemical analysis results (microstructure analysis, aggregates' gradation, chemical analysis). All results were comparatively evaluated in order to identify the structural characteristics of the substrates, as well as the aspects that were taken into account during their construction (Pachta and Papayianni, 2016; Papayianni et al., 2012).
substrates, the general characteristics of each layer (thickness, color), as well as their state of conservation. A series of laboratory tests were conducted in order to determine their physical, mechanical and chemical properties (Papayianni et al., 2012; Stefanidou et al., 2013). The tests were performed in each mortar layer, which was separated carefully (in order not to create additional cracks and voids) with lancet. Microstructure observation was performed with stereoscope (Leica Wild M10) assisted by image analysis (ProgRes). Aggregates gradation was tested by hand grinding and sieving, according to EN1015-1:1998. Porosity and apparent specific gravity were performed according to RILEM CPC 11.3, while compressive strength was tested in cube-shaped samples of each layer, dimensioned around 4 × 4 × 4 cm (Papayianni et al., 2012). Finally, wet chemical analysis assisted by Atomic Absorption was performed in a fine fraction of the sample (< 75 μm), coming from the finest sieve of the aggregates' gradation testing. Additional comparative observations among samples were made, regarding their structure and constituents, such as binding system type, Binder/Aggregate (B/A) ratio and texture
4. Results and discussion According to the in situ survey, all mosaics were situated in a highly leveled ground, while special attention seemed to have been taken to the surface inclinations in order to avoid water accumulation. Taking into account the direct contact to humidity and the exposition to environmental factors, their current state was generally good. Detachment of the substrates' layers was mainly recorded on the edges of the mosaics, while loosening of tesserae was also partially detected (Fig. 2a–d). 688
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
4–8 cm thick, nucleus 2–5 cm and supra nucleus 0.5–1 cm), with the total thickness of the substrates to vary from 12 cm to 22 cm. In Table 2 the physical, mechanical and chemical characteristics of the mortar samples are presented, as well as their correlating properties, such as B/A ratio and binding system type. Fig. 3 illustrates the macroscopic and microscopic characteristics of the mortar layers of the substrate from the Aiges Palace mosaic. According to the evaluation of the results, all mortar layers were based on lime and pozzolan in a proportion ranging from 1:0.5 to 1:1. The ratio of pozzolan was usually increased in the intermediate mortar layer (nucleus). The systematic addition of pozzolan in these mortars probably emerged from the need to be resistant to intense humidity, as well as durable to loading (Stefanidou et al., 2013; Theodoridou et al., 2013). In some cases and only in the external layers (supra nucleus, nucleus), brick dust was added probably in an effort to enhance the hydraulic properties of the mortars and increase their resistance to humidity (Fig. 4) (Stefanidou et al., 2013). However, brick dust also played an aesthetic role in mosaics, since the reddish hue of the tesserae joints, made the colors and patterns of the mosaic decoration to be more vivid and realistic (Fig. 2a–d). Relevant research has been assessed in wall mosaics' case studies (Baraldi et al., 2016). Aggregates were natural, coming usually from nearby river deposits and were of siliceous and calcitic origin. Their shape was mostly rounded, while angular aggregates were also detected, enhancing the paste-aggregates' bonding (Fig. 3e, i) (Stefanidou and Papayianni, 2006). Crushed brick was added mainly in the upper layers (supra nucleus, nucleus), also in order to enhance mortars' hydraulic character (Fig. 4) (Stefanidou et al., 2013). The aggregates' maximum size (Table 2) was generally decreased towards the surface and their gradation ranged from 0 to 8 mm to 0–16 mm for rudus, 0–4 mm to 0–8 mm for nucleus and 0–1 mm to 0–2 mm for supra nucleus (Fig. 5). Additionally, the B/A ratio was increased towards the upper layers, ranging from 1/2 to 1/3 for rudus, 1/1.5 to 1/2 for nucleus and 1/1 for
Fig. 6. Aggregates' gradation of rudus and nucleus from the Aiges Palace floor mosaic substrate.
All substrates consisted of four layers (Table 1, Fig. 1b, d, f, h), reflecting the typical floor mosaic substrate (Baraldi et al., 2016; Getty Conservation Institute, 2003; ICCROM, 1980; Vitruvius, 1960). In Table 1 the stratigraphy of the substrates and the structural characteristics of each layer are presented. In the Hellenistic substrates (Aiges Palace, Pella), statumen was made of coarse, natural pebbles, while in the Roman ones (Dion, Galerius Complex) angular pieces of marble were used. The diameter of the stone pieces used in the statumen of both period monuments was in the range of 6–10 mm. Regarding the mortar layers (Fig. 3a, b, d, e, g, h), their thickness was decreased towards the surface (rudus was
Fig. 7. Correlation of the porosity and apparent specific gravity of the mortar layers, according to their stratigraphy. 689
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
Fig. 8. Microscopic photo of the contact area of the mortar layers (nucleus-rudus) of Aiges Palace (left) and archaeological site of Pella (right) mosaic substrates.
Other parameters that could influence the adherence between layers, as defined by literature (Stefanidou et al., 2017; Stolz et al., 2016) could be the increase of the effective contact area through the deliberate roughness of the subjacent mortar layer, its higher porosity and the rheological characteristics of the covering layer. According to the study, the porosity of the mortar layers was generally decreased to the surface, while in many cases extreme roughness of the contact surface between layers was observed. Resuming the evaluation of the test results it can be concluded that for > 7 centuries (4th cent. BCE to 3rd cent. CE) ancient floor mosaics presented similar characteristics regarding their substrate. Four, very well compacted layers were usually observed, according to the ancient authors descriptions (statumen, rudus, nucleus, supra nucleus), while several parameters were taken into account for the construction of the multilayer mortar systems (i.e. binding system type, B/A ratio, aggregates' type and gradation, application techniques).
supra nucleus. Taking into account the decrease of the upper layers' thickness, the binder-rich mortars in combination with the decrease of their aggregates' size was necessary for their efficient application, their good compaction and probably the increase of their cohesion to the subjacent layers. In some cases, inclusions were added in the matrix, such as pieces of wood, straw and reed, carefully selected and of specific geometry, probably as a mean to increase the stability of the layers and decrease their cracking tendency (Fig. 6) (Stefanidou et al., 2012). Generally, inclusions were inspected in the lower mortar layer (rudus), which was thicker and probably more prone to cracking. Regarding physical properties (Table 2), porosity was generally decreased towards the upper layer, probably as a result of the aggregates' size decrease and also the increase of the B/A ratio (Stefanidou and Papayianni, 2006) (Fig. 7). The values of porosity were generally low, compared to the ones found in historic mortars of other applications (i.e. structural mortars) (Pachta and Papayianni, 2016; Papayianni et al., 2012; Stefanidou et al., 2013), depicting probably the high compaction of the layers during their application. Apparent specific gravity varied according to porosity, B/A ratio and type - gradation of aggregates, but was generally high. Compressive strength on the other hand ranged from 1.5 MPa to 4 MPa (Table 2). These values were higher than the strength recorded in historic lime-pozzolan mortars of these periods and of other applications (i.e. structural mortars) which were usually around 0.8–2 MPa (Pachta and Papayianni, 2016; Papayianni et al., 2012; Stefanidou et al., 2013). Keeping in mind the continuous contact with the ground and the loads that floor mosaics suffered, the high values of compressive strength could be closely related with the special attention given to the selection of the raw materials and the application techniques followed. A significant parameter that should be taken into account when studying multilayer mortar systems as defined by literature (Botas et al., 2017; Stefanidou et al., 2017) is the contact area between the layers. In the studied mosaic substrates, the interface between the layers was in most cases extremely stable, which led to a very difficult detachment of the layers for analyzing their individual properties. As presented in Fig. 8, the layers' contact was mostly mechanical, while the upper layer followed the surface irregularities of the subjacent. No voids or detachments were observed in the contact zones, indicating the special care taken during manufacture. To provide this firm adhesion of the layers, many factors could attribute, with the main one to be the very good compaction. As described by Vitruvius (1960), the ground was carefully stabilized before the construction of the floor mosaic substrate. The role of statumen was probably to level the subground, by flattening inclinations and ground irregularities, as well as to enhance the stability of the structure by undertaking external loads and vibrations. All mortar layers (rudus, nucleus, supra nucleus) were very well compacted with wooden rammers ensuring their cohesion to each other.
5. Conclusions Diachronically floor mosaics had a two-fold role: apart from enhancing the architectural value of constructions as decorative elements, through their stable substructure they served as durable constructional elements (floors). In all cases, the multilayer substrate played an important role for their durability and preservation, in accordance with the care taken to the ground on which they were applied. According to the study, the substrate of Hellenistic (4th cent. BCE) and Roman (3rd cent. CE) floor mosaics located in monuments of Northern Greece, presented similar characteristics, such as: - Stratigraphy of four very well compacted layers, whose thickness was decreased to the surface: statumen (substructure layer of natural pebbles or unshaped stone pieces), rudus (lower coarse mortar layer), nucleus (intermediate fine mortar layer) and supra nucleus (upper fine mortar layer that undertook tesserae). - The good adhesion of the mortar layers was a significant parameter taken into account that ensured the stability of the substrate. To this direction, a synergy of factors contributed, such as the good compaction of each layer (with wooden rammers), the increase of the surface contact area (through scoring of the fresh material of the subjacent layer), as well as the individual characteristics of each layer (i.e. decrease of the aggregates' maximum size and increase of the B/A ratio towards the upper layer). - The binding system was based on lime and natural pozzolan, while brick dust was added in the upper layers (nucleus, supra nucleus) in an effort to enhance the hydraulic character of the mortars and have a colored result. - The B/A ratio was increased towards the surface layer, while porosity and aggregates' gradation were decreased. Natural aggregates of river origin were used, whereas crushed brick was also added. 690
Journal of Archaeological Science: Reports 20 (2018) 683–691
V. Pachta, M. Stefanidou
- Physical and mechanical properties such as porosity, apparent specific gravity and compressive strength were generally beyond the values found in other types of historic mortars (i.e. structural mortars). Porosity was low (10–16% for Aiges Palace floor mosaic), while compressive strength was generally high (2–4 MPa). According to literature, the corresponding properties of historic structural mortars range around 20–30% for porosity and 0.8–2 MPa for compressive strength.
S.F., Grifa, C., Guarino, V., Mercurio, M., Morra, V., Langella, A., 2016. The art of building in the Roman period (89 B.C–79 A.D.): mortars, plasters and mosaic floors from ancient Stabiae (Naples, Italy). Constr. Build. Mater. 117, 129–143. Leone, G., De Vita, A., Magnani, A., Rossi, C., 2016. Characterization of archaeological mortars from Herculaneum. Thermochim. Acta 624, 86–94. Lysandrou, V., Cerra, D., Agapiou, A., Charalambous, E., Hadjimitsis, D.G., 2017. Towards a spectral library of Roman to Early Christian Cypriot floor mosaics. J. Archaeol. Sci. Rep. 14, 782–791. Pachta, V., 2011. Study of the Technological Evolution of Mortars (Dissertation). Dept of Civil Engineering, Aristotle University of Thessaloniki, Thessaloniki. Pachta, V., Papayianni, I., 2016. Design and application of a data system for the comparative study of historic mortars. In: Ioannides, M. (Ed.), Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 10058 LNCS. Springer Verlag, Cyprus, pp. 701–710. Palivou, K., 1999. Akrotiri of Thera: the constructional technology. In: Library of the Archaeological Society at Athens, Nr 183, Athens, (in Greek). Pantermalis, D., 1987. The Tiling of the Palace of Vergina, Amitos: Honorary Volume for Prof. M. Andronikos, Thessaloniki. (in Greek). Pantermalis, D., 1997. Dion: The Archaeological Site and Museum. Adam Editions, Athens. Papayianni, I., Pachta, V., Stefanidou, Μ., 2012. Analysis of ancient mortars and design of compatible repair mortars: the case study of Odion of the archaeological site of Dion. Constr. Build. Mater. 40, 84–92. Pliny, 1971. Natural History, Book XXXVI. W, Heinemann Ltd., London. Puertas, F., Blanco-Varela, M.T., Palomo, A., Ortega-Calvo, J.J., Saiz-Jimenez, C., 1994. Decay of Roman and repair mortars in mosaics from Italica, Spain. Sci. Total Environ. 153, 123–131. Robinson, D.M., 1930. Excavations at Olynthus, part II, architecture and sculpture: houses and other buildings. In: The Johns Hopkins University, Studies in Archeology No. 9. Rossi, F., 1970. Mosaics. A Survey of Their History and Techniques. Pall Mall Press, London. Siganidou, M., Lilimpaki-Akamati, M., 1996. Pella - The Capital of Macedonians. Ministry of Culture, Athens. Stefanidou, M., Papayianni, I., 2006. The role of aggregates on the structure and properties of lime mortars. Cem. Concr. Compos. 27, 914–919. Stefanidou, M., Papayianni, I., Pachta, V., 2012. Evaluation of inclusions in mortars of different historic periods from Greek monuments. Archaeometry 54, 737–751. Stefanidou, M., Pachta, V., Konopissi, S., Karkadelidou, F., Papayianni, I., 2013. Analysis and characterization of hydraulic mortars from ancient cisterns and baths in Greece. Mater. Struct. 47 (4), 571–580. Stefanidou, M., Kesikidou, F., Antoniadis, K., 2017. Application techniques for high performance traditional renders. Procedia Environ Sci 38, 242–247. Stolz, C.M., Masuero, A.B., Pagnussat, D.T., Kirchheim, A.P., 2016. Influence of substrate texture on the tensile and shear bond strength of rendering mortars. Constr. Build. Mater. 128, 298–307. Thamboo, J.A., Dhanasekar, M., 2015. Characterisation of thin layer polymer cement mortared concrete masonry bond. Constr. Build. Mater. 82, 71–98. Theodoridou, M., Ioannou, I., Philokyprou, M., 2013. New evidence of early use of artificial pozzolanic material in mortars. J. Archaeol. Sci. 40, 3263–3269. Vitruvius, 1960. The Ten Books of Architecture, Translated by Morgan M.H. Dover Publications Inc., New York.
These observations lead to the conclusion that ancient floor mosaic substrates were probably constructed according to diachronic principles, following specific requirements. The raw materials used were carefully selected and applied in each mortar layer, while special attention was given to the layering techniques. As a result, durable substrates were achieved that guaranteed the longevity of these masterpieces, preserved until nowadays. References Alberti, L.B., 1965. In: Rykwert, J., Tiranti, A. (Eds.), Ten Books on Architecture, (J. Leoni (translation), London). Andronikos, M., 1984. Vergina - the Royal Tombs and the Ancient City. Ekdotike Athenon, Athens. Athanasiou, F., Malama, V., Miza, M., Sarantidou, M., 2015. The Restoration of the Remnants of the Galerius Complex in Thessaloniki (1994–2014), Volume A', Thessaloniki. Ephorate of Antiquities of the City of Thessaloniki, Thessaloniki. Baraldi, P., Bracci, S., Cristoferi, E., Fiorentino, S., Vandini, M., Venturi, E., 2016. Pigment characterization of drawings and painted layers under 5th–7th centuries wall mosaics from Ravenna (Italy). J. Cult. Herit. 21, 802–808. Botas, S., Veiga, R., Velosa, A., 2017. Air lime mortars for conservation of historic tiles: bond strength of new mortars to old tiles. Constr. Build. Mater. 145, 426–434. Bruneau, P., 1987. La mosaïque antique, Lectures en Sorbonne, collection dirigée. L' Université de Paris-Sorbonne, Paris. Dumbabin, K.M.D., 1999. Mosaics of the Greek and Roman World. Cambridge University Press, New York. Getty Conservation Institute, 1991. The Conservation of the Orpheus Mosaic at Paphos, Cyprus, California. Getty Conservation Institute, 2003. Israel Antiquities Authority, Illustrated Glossary: Mosaics in Situ Project: Definitions of Terms Used for the Graphic Documentation of in Situ Floor Mosaics, Los Angeles. Goodwin, J.F., West, H.W.H., 1980. A review of the literature on mortar/brick bond. In: The British Ceramic Research Association, Techn. Note No. 308. Groot, C., 1993. Effects of Water on Mortar-brick Bond. Delft University, Delft. Hamarneh, C., Abu-Jaber, N., 2017. Mosaic pavement mortar production in Gerasa in the Byzantine period. Archaeol. Res. Asia 9, 22–33. ICCROM, 1980. Mosaics No2, Safeguard, Perigueux, Rome. Izzo, F., Arizzi, A., Cappelletti, P., Cultrone, G., De Bonis, A., Germinario, C., Graziano,
691