Design and Application of Mortars and Grouts for the Restoration of the Byzantine Church of Panaghia Krena in Chios Island, Greece

Abstract

The Church of Panaghia Krena is a very important Byzantine monument situated in the island of Chios, well-known due to the high aesthetic value of the ceramic tile decoration on its facades and of its frescoes. The church suffered severe damage especially due to the 1881 devastating Chios earthquake and different interventions were implemented soon after (1884), consisting mainly of the reconstruction of collapsed areas and the addition of ties. Nevertheless, deterioration of old damages and appearance of new ones was observed. Thus, a restoration program was carried out between 2000 and 2007. This paper presents the basic characteristics of all types of historic materials examined, the main results of the in situ and laboratory experimental program carried out for the selection of the raw materials, and the design of optimum compositions of mortars and grouts to be used for the restoration works. The importance of in situ pilot applications is underlined, as well as of the production of site-specific aggregates for the repointing mortars by crushing a local rock, in order to succeed the adequate reddish color to match with the original mortars and bricks still preserved and achieve an aesthetic harmony with the color hue and texture of the original materials.

1. Introduction

The Church of Panaghia (Virgin Mary) Krena, situated in the surrounding area of the village of Vavyloi in Chios Island, is a very well-known Byzantine monument of exceptional importance [Figure 1]. It comprises the main church and a narthex (or inner narthex), constructed in two different construction phases, both belonging to the late 12th century and the ruins of an exonarthex, added later, during the 16th century [1,2,3,4,5].

Regarding its architectural style, the church belongs to the single-spaced octagonal type, the same as the Katholikon of Nea Moni in Chios Island, that is inscribed in the World Heritage List of UNESCO since 1990. The term “octagonal” refers to the system supporting the central dome, which rests on eight load-bearing pilasters along the four walls of the roughly square central part of the main church. Extended areas of the external façades are constructed with byzantine brick-masonry, according to the so-called “recessed brick” technique and have very rich decorative brick ornamentation (Figure 1 and Figure 2). All the internal surfaces of the walls and vaulted roof are decorated with frescoes, except parts of the narthex that collapsed in 1881 and were reconstructed in 1884 or the early 20th century.

Three-leaf masonry is used for the construction of the walls (Figure 2a), consisting of an external leaf of brick masonry, an internal leaf of rubble stone and/or brick masonry and an inner core of small pieces of stones, bricks and mortar [6,7].

As aforementioned, the majority of the external leaf of the walls (with the exception of the exonarthex) is constructed using the so-called “recessed brick” technique or concealed-course technique (Figure 2), considered as a hallmark of the architecture of Constantinople [4,8]. For further decorative reasons, this technique is also combined with an alternation of stone layers, mainly on the lower part of the walls (Figure 1a,b).

The technique is characterized by the recessing of alternate brick layers from the wall surface. Consequently, as the recessed bricks remain behind the surface (Figure 2a,c,e), they are concealed inside the mortar and thus, the joints appear to be considerably wider than the bricks on the surface (Figure 2b,d–f). As presented in Figure 2b, the width of the joints is approximately 9 cm, while the bricks’ thickness is 3 cm [5]. The Figure 2c,e, illustrate how the recessed brick is located in a depth of a few centimeters from the surface (~2 cm) of the wide joint concealing it. The surface of the joint has a slight inclination downwards, in order to avoid the accumulation of rainwater in the joint (Figure 2d). For the same reason, the brick’s upper edge is coarser, in order to ensure good adhesion (grip) of the mortar.

Careful observation has documented that the pointing mortar, which conceals the recessed brick, is practically the same as the bedding mortar but with less coarse aggregates, as was revealed by visual inspection and further confirmed by analyses.

An extensive network of double horizontal timber laces incorporated in the interior of the masonry was identified along the walls of the main church, connecting them and providing to the entire structure a circumferential confinement and an enhanced capacity to withstand tensile stresses [9,10,11,12]. The timber laces are not visible in the external façades, as they are located behind the external leaf of the three-leaf masonry, consisting either of stone or brick masonry, as schematically presented in Figure 2a. Sometimes the edges of the smaller perpendicular wooden beams connecting the laces were visible. The existence of timber laces along the length of the wall was made visible only in damaged areas, as shown in Figure 2g. Thorough documentation during the design and execution of the restoration works has proven that the timber laces are placed in a systematic way at six levels from the base to the top of the walls. Their existence was initially identified at five levels, by meticulous observation [6,7] and subsequently, by extended in situ investigations using endoscopy, at the following locations: under the windows, at the level of the first and the second cornice (spring of arches and vaults), and at the base and top of the cylindrical drum of the cupola [13,14,15]. The sixth level situated at the level of the rectangular base of the dome was completely invisible and thus, its existence was identified during the execution of the restoration works [15].

The church has suffered serious damage during the centuries due to severe earthquakes. Among them, the major earthquake that hit the island of Chios in 1881 [16,17] led to the total collapse of the west wall of the narthex, of the upper part of its south and north walls and of all its vaulted roof, while in the main church, the hemispherical cupola of the dome and the west side of the central masonry piers of the Sanctuary collapsed, along with partial collapses and heavy cracking of other structural elements. Important restoration works took place at the end of the 19th century and the beginning of the 20th century, including mainly the rebuilding of the collapsed parts of the monument and the installation of iron ties [6,7]. The hemispherical cupola was reconstructed in 1884 and decorated with new frescoes in the same year. The west wall of the narthex is also reported to have been rebuilt in 1884.

In the 1980’s, after the removal of the outer layer of post-byzantine frescoes, created in 1734 by Chios painter Michael Chomatza and exhibited today in the Byzantine Museum of Chios [5], the initial Byzantine frescoes reappeared, revealing a more extended network of cracks in the walls.

Only some urgent local deep repointing was then undertaken mainly on the cracks of the external façades, together with in situ conservation of the frescoes, without, however, carrying out interventions to repair and strengthen the damaged masonry elements, which were in a very bad condition. The damage was especially severe in the upper levels of the main church, which bear the loads of the dome (spherical ring, squinches, niches and arches), and the façade’s arched niches where the iron ties were anchored (Figure 3a and Figure 3b respectively).

Therefore, a first architectural survey was realized in 1998 [18] and served as a base for further detailed architectural and structural studies and extended in situ and laboratory investigations [6,7,13,15]. During these studies and investigations, a detailed survey of the pathology of the church was carried out. It was revealed that one of the main reasons for the documented severe damages was the inherent vulnerabilities of the structural system of the monument, presenting serious weaknesses and particularities in critical areas [7,15]. Moreover, the timber laces were, in most locations, decayed or totally decomposed, and hence, incapable of providing the reinforcement that they were initially designed to offer, provoking further deterioration to the vulnerable masonry [13,15]. These weaknesses, in conjunction with the effect of earthquakes and the ageing and decay of materials, especially due to the action of moisture (washed out surface mortars, deteriorated decoration bricks and tiles, decomposed timber laces, etc.), resulted in the pathology observed. Another important finding revealed during the progress of the works was the particularly careless manner of construction of the inner of the three-leaf masonry, in contrast to the meticulously built external façades; a problem further exaggerated by the additional voids left in the place of the decayed and decomposed timber elements [19].

In this framework, a series of structural and architectural interventions were designed and approved by the Central Archaeological Council and the Competent Authorities of the Hellenic Ministry of Culture. An extended restoration project was realized during the period 2004–2007. A presentation of the pathology of the monument and of the proposed main structural restoration interventions is offered in [7,15,20].

This paper focuses on the methodology followed for the design of the optimal compositions of repointing and deep repointing mortars and hydraulic grouts, based on the documentation of the existing historic materials and the environment of the monument, respecting all performance requirements defined by the restoration design, both in terms of fresh state properties (workability, injectability, etc.) and of hardened state (physico-mechanical, chemical and aesthetic compatibility, durability, etc.).

Emphasis is given to the optimization of the compositions and the application methodology of the repointing mortars, as well as of the new bricks to be used in the restoration works. The main objective of this endeavor was to preserve the historical, architectural and aesthetic value of the façades’ pointing mortars and their rich brick ornamentation, which characterize the appearance and the texture of the masonry of the monument, creating an alternation of light and shade and conferring volume and dimension to the external façades.

Given that, a large percentage of the surface pointing original mortars of the Middle Byzantine “recessed” brick masonry has survived to this day in relatively good condition, inevitably with some slight damage or decay and weathering of materials in some areas, a decision was made (based on the restoration study proposals) to preserve it, treating it as an object of art, in collaboration with the specialized conservators team. However, the application of grouting to strengthen the masonry was necessary. To this end, the installation of grouting tubes was carried out in carefully selected locations and upon completion of the grouting, the tubes were removed, and the holes left were carefully filled with the appropriate repointing mortar, by specialized conservators. For other parts of the monument, various types of interventions were designed, including repointing, deep repointing and grouting, as well as local rebuilding in the parts that were heavily damaged or altered due to previous interventions, using new appropriate mortars and grouts, together with tailored hand-manufactured new bricks. In all cases, both technical and aesthetic compatibility was sought during the design and the implementation of the works [21,22,23,24].

Evidently, in order to rebuild the heavily damaged or altered areas with brick and tiles ornamentations, detailed surveys of the existing geometry, materials and pathology have been made, together with detailed drawings for the geometry, form and materials of the rebuilt areas. The presentation of data relative to this matter is beyond the scope of this paper, which deals mainly with the analysis of existing materials and the compositions of repair mortars and strengthening grouts.

A detailed knowledge of the existing masonry materials and especially of the mortars was considered of paramount importance, both for their structural role having a direct impact on the resistance and durability of the masonry [25,26,27,28] and for their architectural value, regarding the style, appearance and aesthetic of monument’s facades. A lot of research has been undertaken in the last decades on mortars’ characterization and on identifying the role of their application technology and of the environmental conditions on their hardened properties [29,30,31,32,33,34,35,36]. In addition, many efforts have been made to develop a holistic approach to the design of repair mortars and specific recommendations have been established [37,38,39,40,41,42,43,44,45,46,47], while the significance of the way of the application of repointing mortar has also been discussed, and guidance if offered for historic brick masonry [45,47]. Similarly, extended laboratory research and in situ applications carried out over the last four decades have proven that injections, using hydraulic grouts, can be efficient with the prerequisite that they have been adequately designed and implemented, on the basis of specific performance requirements and design criteria developed to this end [48,49,50,51,52,53,54,55,56,57,58].

Hence, samples of all types of construction materials were examined. On the basis of the comparative study of the results of their analysis and the performance requirements set by the restoration study, mortar and grout compositions were designed, by the Directorate for Technical Research on Restoration of the Hellenic Ministry of Culture (DTRR/HMC). Subsequently, in combination with many in situ pilot applications and quality control checks, the adequate compositions were finalized during the works.

In this paper, the optimal final compositions used in the restoration works are presented together with the holistic multidisciplinary methodology adopted. Problems revealed during the realization of the project and the steps followed for their mitigation are also discussed, highlighting the importance of a systematic quality control system with laboratory experimental support, during the execution of restoration works.

Comparative images of selected parts of the façades, five years after the completion of the works are included in the paper, since the evolution of color and texture of pointing interventions over time is very important, in order to evaluate if aesthetic compatibility was achieved.

Considering that, the literature is not very rich regarding case studies presenting the final compositions of restoration mortars and grouts, as adapted during the execution of the works, the paper aims to highlight the importance of worksite investigations and thus enrich our knowledge. The significance of pilot applications and quality control to verify the performance requirements defined by the design, is also emphasized.

Additionally, given the rare type of this unique monument, built according to the “recessed” brick byzantine masonry and decorated externally with mortar and brick ornaments and internally with valuable 13th century frescoes, which are to be preserved and conserved in situ, the content of this paper may be very useful for similar case studies, as the relevant literature is not extensive [59].

2. Materials and Methods

2.1. Sampling

Taking into account the main phases of construction identified in the framework of the relevant literature and restoration studies [2,3,4,6,7], a thorough visual inspection of the monument was carried out, focusing on the building materials and especially on the details of pointing and bedding masonry mortars, in accessible damaged areas.

The meticulous in situ observation comprised not only the mortar’s constituents, texture and color, but also their finish, profile and state of preservation [60,61]. Thus, the locations of sampling points were identified in characteristic areas of the structure, taking into account the construction phases, the type of the mortar (bedding mortar or pointing mortar) and their state of preservation, in relation to possible types and level of decay, weathering or damage. Subsequently, sampling on representative and less weathered building materials (stones, mortars, bricks) was executed, from different points and construction phases of the monument, normally using a hammer and chisel with care to extract, as much as possible, intact samples. In rare cases, samples were removed by hand (especially bricks). The location and methodology of sampling were documented by drawings and photographs. A label was added to all samples comprising of a serial code number, which followed the evolution of sampling, and this label was used in all tests and reports.

Basic information on mortar sampling is presented in

Table 1. Sampling of historic mortars: code number, type of mortar, position of sampling per construction phase or rebuilding or repair time period and color.

No.	Type of Mortar Samples, Position of Sampling	Color
A’ phase of construction (main church late 12th c.)
7	Bedding mortar—Sanctuary’s N wall of the central vault	Reddish brown
11B	Pointing mortar—S façade of the Sanctuary (Diakonikon).	Reddish brown
12	Bedding mortar—Sanctuary’s N wall’s internal face—central part.	Reddish brown
17	Bedding mortar—S wall’s internal face	Reddish brown
18	Bedding mortar—Façade of the central part of the Sanctuary	Reddish brown
19	Bedding mortar—Spherical ring of the dome.	Reddish
23	Bedding mortar—Façade of the central part of the Sanctuary	Reddish
25	Bedding mortar—S wall internal face of the Diakonikon	Reddish brown
26	Bedding mortar—Main church S wall internal face (window).	Reddish
46	Pointing mortar—Middle arch of the main church’s S façade	Reddish
29	Pointing mortar—W arch of S façade of Main Church.	Reddish
30	Pointing mortar—E arch of N façade of Main Church.	Reddish
B’ phase of construction (narthex’s addition, late 12th c.)
4	Bedding mortar—Narthex’s S. arcosolium extrados	Reddish
8	Bedding mortar—E part of narthex’s N wall internal face	Reddish brown
10	Bedding mortar—Narthex’s N. arcosolium intrados (east part)	Reddish brown
22	Pointing mortar—Narthex N façade (main church’s border)	Reddish
C’ phase of construction (exonarthex’s addition, 16th c.)
27A	Bedding mortar -Exonarthex.	Reddish
27B	Pointing mortar—Exonarthex.	Off-white beige
47	Pointing mortar—Internal S. face of exonarthex.	Off-white beige
48	Pointing mortar—External N. façade of exonarthex.	Off-white beige
Rebuilding/repair mortars after the 1881 seismic damage
13/16	Bedding mortar—Narthex’s W façade S part, over the window and at an upper level, respectively. (Rebuilding)	Off-white with red aggregates
15/20	Pointing mortar—S and N Narthex façade, respectively, at the border with the main church. (Rebuilding).	Off-white-beige-red aggregates
24	Pointing mortar—Exonarthex’s SW internal corner. (Repair)	Beige-grey
31	Bedding mortar—Exonarthex Upper part of S wall (Rebuilding).	Beige (clay)
32	Pointing mortar—Narthex’s N arcosolium, E part over the frescoes. (Repair)	Off-white
33	Pointing mortar—East part of the N vault of narthex. (Repair)	Off-white
34	Pointing mortar—Narthex’s W wall arched window. (Repair)	Beige
36/37	Bedding mortar—Narthex’s W part of N vault and N part of W wall, respectively. (Rebuilding)	Off-white with red aggregates
41/42/44	Pointing mortar—Narthex’s W part of S vault and S part of W wall over S arcosolium, respectively. (Repair mortar)	Off-white with red aggregates
45	Bedding mortar—Narthex’s W wall. S part (Rebuilding)	Beige-red (clay)

. Bedding mortars and pointing mortars are included, as well as a sample taken from frescoes substrata. For each sample, the code number, the sampling location, the type and the color of sampled masonry mortars are given in the Table 1, categorized per phase of construction or rebuilding/repair period and color.

As the monument’s external façades are built in the so-called “recessed” brick technique, pointing mortars constitute a crucial component of the façade’s architectural composition and aesthetic. Consequently, sample sizes of such original well-preserved pointing mortars had to be small and sampling locations should not disturb the continuity of intact or slightly damaged areas; thus, the best-preserved areas adjacent to damaged ones were selected for mortar sampling. Likewise, on the interior faces, due to the existence of frescoes on most of the original parts of the monument, sampling was only possible in the areas where the frescoes were missing. In all cases, care was taken to select and sample authentic mortars belonging to the respective phase of construction or reconstruction and not to other less significant local repair works.

Additional sampling of masonry mortars was carried out in some specific areas of the monument, which became accessible only during the works (e.g., the extrados of the vaults, the base of the drum of the dome, etc.) to complete the documentation of the monument; they were kept in the DTRR laboratory, along with all the other samples left over from the analyses, in case further investigations are needed in the future.

2.2. Tests and Equipment

Chemical analysis was performed on more than 24 historic masonry mortars (both bedding mortars and pointing ones) and ceramic samples were taken from different positions of the monument, at the Laboratory of Building Materials of the Aristotle University of Thessaloniki (LBM/AUTH) [62,63,64] The elemental analysis of the examined materials was carried out by Atomic Absorption Spectrometry (AAS) on the fraction passing through 0.85 mm. The loss of ignition was determined after heating the samples at 1000 °C until constant mass.

Thin-section petrographic analysis was carried out on representative mortar samples by means of Optical Microscopy under polarized light and coupled with a digital camera.

Granular analysis of the original mortars, as well as of the aggregates used in the repair mortars, was carried out using standard ASTM sieves series. The mortar samples were dried in an oven at 105 ± 5 °C until a constant mass was obtained. Then the samples were grounded using a mortar and pestle with care to avoid breaking the grains. Consequently, the grounded samples were placed on top of a “nest” of sieves and subjected to shaking action. The sieving was further assisted by applying a soft brush on each sieve in order to separate the binding material from the aggregate grains.

A Siemens D500 powder diffractometer, equipped with α copper source and graphite monochrometer was used for X-ray diffraction (XRD) analysis of binders, aggregates and ceramic samples. The X-ray diffraction was performed in 2θ geometry, scanning between 3° and 70°, with step interval 0.02° and step time 1.0 s.

Selected mortar samples were examined by scanning electron microscopy (JEOL-JSM 5600) coupled with X-ray microanalysis system (SEM/EDX).

Differential thermal and thermogravimetric analysis (DTA-TG) was carried out for the hydraulic lime used as grouting material, by means of a simultaneous Setaram Labsys Evo TG-DTA/DSC thermoanalyser.

Mercury intrusion porosimetry (MIP) was utilized for the evaluation of the open porosity and pore size distribution of both historic and repair bricks Testing was performed on a Quantachrome Autoscan Multiple mercury porosimeter.

For the determination of the flexural strength of the mortars, a three-point bending test was carried out on 40 × 40 × 160 mm specimens, using a Wykeham Farrance (TRITECH 100 KN) testing machine, at a low rate of loading (0.05 mm/min). The compressive strength test was conducted on 40 × 40 × 40 mm cubic specimens resulting from the preceding flexural test, according to EN 1015-11:1999 [65]. The rate of loading for compression was 0.1 mm/min. The measurement of the compressive strength of stones and bricks was conducted on cubic specimens of 5 cm and 3 cm—edge, respectively.

Based on the results of the analyses and testing of the historical building materials, suitable restoration binders, aggregates and new handmade repair bricks were pre-selected after market research. Their compatibility and suitability were examined using the techniques described above and the most suitable raw materials were selected. Mortar mixtures were prepared in the laboratory with a Hobart mixer and specimens (40 × 40 × 160 mm) were molded and cured in the humid chamber to test their mechanical strength over time, according to EN 1015-11:1999 [65]. Moist curing conditions (20 ± 2 °C and relative humidity of 95% ± 5%) were used, as in all mortar compositions the binder selected was either a combination of lime with a natural pozzolan or a hydraulic lime.

Grouts were prepared using a high-turbulence mixer [48] and their penetrability, fluidity and stability characteristics were fully examined for various water/solids ratios, with or without the addition of chemical admixtures. The standardized sand column test method EN 1771 was applied [66] to check the penetrability and fluidity. The time T36 that the grout needs to reach the top of the column was measured, while it was checked that the grouts exit from the column and flow in a steady rate into the special recipient cup of the testing apparatus, as described in the norm EN 1771 and the relevant literature [67].

To characterize the fluidity, the Fluidity Factor Test was carried out [68]. A Marsh Cone of 3 mm nozzle diameter is filled with 1000 mL of grout; the flow time t_f is measured for a flow of Q = 100 mL of grout to pass through. The fluidity factor is calculated Fl = Q/(Axt_f), where “A” denotes the area of the cross section of the nozzle [68]. Moreover, the Flow time test was also carried out. A Marsh cone of 4.75 mm nozzle diameter was used, and 1000 mL of grout was inserted. The time needed for 500 mL of grout to pass through the nozzle was measured [48,68].

Regarding the stability, bleeding was measured after 3 h according to NF P18-359 [69], while segregation was checked by measuring if the thickness of the distinct denser layer sedimented at the bottom of the container after performing the sand column test was >1 mm [70]. Apparent density of the grout was also measured immediately after mixing and one hour later, to check for possible tendency to segregate. To this end, the grout is placed into transparent volumetric tubes (of 1000 to 2000 mL). Apparent density is determined by weighing out a specific grout volume (e.g., 50 mL) collected from the upper third of the height of each tube [48]. Grout specimens 40 × 40 × 160 mm were also molded and cured in the wet chamber (20 ± 2 °C and relative humidity of 95 ± 5%) in order to test their mechanical strength over time [48].

2.3. Methodology for Selecting the Final Mortars and Grouts Compositions

A series of mortars and grouts compositions were designed by the DTRR/HMC, based on the comparative study of the results of the characterization of historic materials sampled and the performance requirements set by the restoration study. The proposed compositions were produced and tested in the Laboratory, prior to their use on the worksite and an initial selection of compositions to be used in the works was made. When the restoration works started, the selected mortar compositions were produced on the worksite (using the raw materials, equipment and workmanship available in situ) and applied in carefully selected areas on the various types of masonry, using different application procedures, regarding mainly the finish of the surface joints. In parallel, specimens were molded and conserved in situ and after a given period transferred to the Laboratory and tested.

This methodology allowed for a series of problems to be identified, both in terms of the quality of the worksite raw materials and the application process, as well as of the finishing of the repointing mortars and their texture and color evolution over time. For example, in order to obtain the appropriate reddish hue for the repointing mortars to match the historic materials preserved in situ and achieve aesthetic harmony with the color shade and texture of the old masonry, it was decided to produce aggregates by crushing a local rock found in an old quarry on the island of Chios.

Moreover, in order to select the optimal compositions of surface repointing mortars, the pilot applications of the candidate compositions were made in selected areas on all the façades of the monument. In some cases, depending on the level of damage, even in two areas of each façade. After specified periods of time, all members of the multidisciplinary team responsible for supervising the works inspected and evaluated the pilot applications to check the quality of the work in terms of placement and finish, as well as of the evolution of their texture and color over time (inevitably affected by humidity and light). Based on the evaluation of the supervising team, all necessary adjustments were implemented until a common decision was reached, regarding the adequacy of each composition. These had to comply with the requirements set for technical and aesthetic compatibility, which concerned not only the historical masonry of the façades but also the interior with the valuable frescoes.

3. Results of the Analysis and Characterization of Existing Materials

3.1. Historic Mortars

The chemical analysis results reported in

Table 2. Mean values of main components found in the chemical analysis of mortar samples.

Type of Historic Mortar	%CaO	%MgO	%SiO2	%Al2O3 + Fe2O3	LOI
Reddish mortars (aver. 9 samples, mainly pointing masonry mortars of A’ and B’ phases of construction)	42.2	1.7	15.6	4.5	33.8
Reddish brown (aver. 8 samples, mainly bedding masonry mortars of A’ and B’ phases of construction)	37.8	1.6	22.1	3.8	32.2
Off-white (+plant fibers) (1 sample) (frescoes substrata)	48.7	1.2	6.08	1.01	42.8
Off-white/Beige (3 samples, rebuilding mortar of narthex)	45.3	0.55	12.0	2.64	38.3
Grayish (1 sample, 20th century repair mortar of exonarthex)	29.1	0.7	34.9	6.73	27.1
Clay mortar (1 sample repair mortar of narthex)	14.0	1.8	47.2	21.3	13.9

concluded that mortar samples examined were lime mortars with low-medium hydraulicity (based on chemical analysis results the modulus of hydraulicity (m) was estimated, by using the equation: m = %(CaO + MgO)/%(SiO₂ + Al₂O₃ + Fe₂O₃). The modulus of hydraulicity (m) expresses the ratio between basic and acid oxides in mortar. If m > 9.0, it is a pure lime binder. The modulus of hydraulicity of hydraulic lime mortars is 1.7–9.0). As noted in the report of LBM/AUTH [62] (See relevant reference in footnote 3), the hydraulicity may be partially attributed to the presence of fine-grained quartzite-stone (reacting with lime to produce hydraulic components).

Most historic mortars are reddish in color (A’ and B’ Middle Byzantine phases of construction), which is mainly attributed to the nature of the aggregates, especially the fine-grained fractions of crushed reddish rock and to the presence of ceramic fragments as well. As is well known, aggregates, and especially the fine fractions of sand, have a significant effect on the color of lime mortars. In most cases, local sand and other aggregates were usually sourced from local sand beds, found near creeks or rivers, from the trimmings of stone on site, or from brick pieces or powder, which is a pozzolanic additive. During the Byzantine era, aggregates of various colors and crushed brick were commonly used in the making of masonry mortars.

The mortars used for rebuilding the collapsed parts or for repointing damaged areas in mitigation of the damages induced by the 1881 earthquake were of “off-white-beige color” with or without coarser red aggregates. A clay mortar taken from a rebuilt area of the narthex was also among the samples examined, as well as a mortar belonging to the substrata of frescoes.

The binder/aggregate ratio usually ranges from 1:1 to 1:2, according to the sieve analysis. The percentage of the various fractions of the aggregates is approximately the following: fraction 0/4 mm 70–90%, fraction 4/8 mm 10–20% and fraction 8/16 mm 10–15%.

Figure 4 presents some characteristic granulometric curves extracted from granulometric analysis, realized in the Laboratory of the DTRR/HMC on selected samples from the bedding masonry mortars of the monument: two samples from the main church (no. 17 and 26 representing the masonry mortars of the A’ and the B’ phase of construction), three samples from the narthex (no. 16, 33, and 36 representing the mortar used for the rebuilding of the collapsed parts of the narthex), and one from the pointing mortar of the exonarthex (no. 48).

Mineralogical analyses were carried out on the aggregates of two representative samples—presented in Figure 5 (Samples 17 and 26)—and on the binder of mortar 26.

Fractions of the binder and aggregates, after separation by sieving, were subjected to X-ray diffraction analysis (XRD), carried out by the Institute of Geological and Mineral Research (IGME). The XRD results, presented in

Table 3. XRD results of binding mortar samples.

	Samp 17 (aggreg.)	Samp. 26 (aggreg.)	Samp. 26 (Binder)
Quartz	45.4%	83.5%	3.9%
Calcite	21.1%	6.6%	92%
Kaolinite	21.2%	10%
Orthoclase	12.3%
Ca2SiO4 (Belite)			4.1%

, show that the main mineralogical phase is quartz, attributed to silicate grains, followed by calcite, attributed to limestone grains, while the presence of orthoclase in sample 17 is apparently due to grains of volcanic rock, which can also be seen by macroscopic observation of the aggregates. Kaolinite, occurring in a quite large percentage, can be attributed both to the binder and to the presence of marly material grains. The increased percentage of calcite can be attributed to binder residues that have remained adhered on aggregates. The composition of the binder (sample 26) is calcitic (product of the carbonation process of calcium hydroxide).

The binder also contains a smaller quantity of quartz, as well as of Ca₂SiO₄ (Belite). This particular compound indicates that the lime probably derived by firing and slaking clay or siliceous limestone.

The observation of thin sections, carried out by (IGME) of mortar samples under the polarizing microscope concluded that the aggregate materials are mainly of angular to sub-angular grains of varying size and color, joined by a dark-colored binding material. The aggregate material is mainly of siliceous composition (forms of quartzite material) of various grain sizes. The maximum size of the quartz grains is about 5 mm. Individual quartz crystals, as well as conglomerate forms, occur in places. Figure 6 presents two characteristic images from sample 16.

The presence of microcrystalline-sparitic calcite crystals (derived from limestone aggregates), as well as of iron oxides (hematite) and hydroxides were also observed. Areas of clay composition are also present. The binding material, which is very well bonded to the aggregate mass, consists of a fine-grained calcite material (cryptocrystalline to microcrystalline), in which there are small pieces of quartz, iron oxides and foliar mica (sericite). The mass ratio of binder/aggregates is approximately 1/1. The binder (powder) was further examined by SEM. The SEM-EDX analysis indicates that the composition is mainly calcitic with a small content of Si and Al (alumino-silicate compounds detected due to the possible existence of hydraulic components).

It has to be noted that the general quality regarding the mechanical characteristics of some original mortars of the main church and the original parts of the narthex (reddish mortars of first and second construction phases) was quite good, taking into account that they were tested after being extracted from the masonry, a process that usually has a negative impact on their characteristics. However, in some cases, it was possible to prepare manually cubic specimens and perform compressive strength tests. The compressive strength of specimens made from the samples of existing mortars was found in the range 0.7–3.0 MPa.

3.2. Historic Building Stones and Bricks

The main categories of the building stones of the monument are as follows:

• Limestones: sedimentary rock (chemical sediment) of high strength; two types were present: a fine-grained hard limestone beige in color and a gray dolomitic one.
• Trachyandesite: igneous rock, semi-crystalline of medium to high density.
• Thimiana stone: it is characterized as sandstone of a reddish-brown color, medium-fine-grained. Stratigraphically is the oldest formation of the Neogene basin of Chios. In the literature is identified as reddish ferruginous clayey marl rich in quartz sand.
• Calcareous sandstone: sedimentary rocks of low/medium strength.

In

Table 4. Characteristics of the building stones of the monument.

	Apparent Density (g/cm3)	Compressive Strength (MPa)
Beige Limestone	2.57	118.0
Gray dolomitic Limestone	2.70	79.2
Trachyandesite	2.33	47.6
Stone of Thimiana	2.25	58.8
Calcareous Sandstone	2.10	26.0

, the apparent density and the compressive strength (measured by the Geoereuna Laboratory O.T.M. EPE and the Laboratory of Reinforced Concrete of the National Technical University of Athens) of the main categories of building stones are reported.

The bricks of the monument are solid, reddish-brown in color, and make up approximately the 30–40% of the total masonry materials of the main church (excluding mortars). There are three different shades of bricks in the monument which were probably due either to their different manufacturing method (i.e., differentiation in terms of clays, inclusions and firing temperature), or the different period in which they were placed in the monument (e.g., during repair interventions).

Representative samples were taken from the three categories of bricks (OR1, OR2 and OR3; presented in Figure 7), as well as a sample (KER-4) from the tiles of the church roof. The physical and mechanical properties (

Table 5. Physical and mechanical properties of ceramic samples.

	Apparent. Density (g/cm3) (MIP)	Real Density (g/cm3) (MIP)	Tot. Introduced Volume cc/cc (MIP)	Water Absorption (%) (Weight/Buoyancy)	Porosity (%) (Weight/Buoyancy.)	Compressive Strength (MPa) *
OR1	2.03	2.53	0.2264	14.0	25.6	19.0
OR2	1.68	2.48	0.3461	25.3	39.1	4.4
OR3	1.55	2.30	0.3482	25.7	39.2	7.3
KER-4	1.93	2.49	0.2399	14.5	27.9	-

* Measured on cubic specimens 3 cm—edge. There was a large deviation in the values of compressive strength due to the fact that some of the bricks found in-situ were not preserved well.

), and the chemical and mineralogical characteristics (

Table 6. Chemical analysis results of the ceramic sample.

%	Al2O3	CaO	SiO2	MgO	Fe2O3	K2O	Na2O	TiO2	LOI
OR1-1	18.3	1.43	63.1	2.7	6.82	3.1	0.69	0.94	1.5
OR2-2	15.3	7.35	56.7	3.49	6.79	1.73	0.67	0.92	5.5
OR3-3	16.0	6.44	57.6	2.95	7.79	1.94	0.45	0.83	4.59
KER-4	17.02	2.52	62.6	2.32	6.3	3.1	0.8	0.83	2.35

and

Table 7. Mineralogical analysis (XRD) results of the ceramic samples.

OR1-1	The major phase of the sample is quartz SiO2. The inclusions of the brick are dolomitic with the presence of the dolomite phases as Ca(Mg0.67Fe0.33)(CO3)2 and as CaMg(CO3)2. There is also the Fe2O3 hematite phase as well as the KAlSi3O8 microcline as a binder.
OR2-2	The major phase of the sample is quartz SiO2. CaO is present as a binder while inclusions are ferrous compounds as Fe2O3 and basalts as anorthoclase or albites (Na,K)AlSi3O3
OR3-3	The major phase of the sample is quartz SiO2. CaO is present as a binder while the inclusions are carbonate as CaCO3 and ferrous as Fe2O3. Clays were also identified as illite in the sample. This is likely to be found either as a binder or as impurity.
KER-4	The major phase of the sample is quartz SiO2. CaO is present as a binder while the inclusions are carbonate as CaCO3 and ferrous as Fe2O3 and albite NaAlSi3O8. Microcline KAlSi3O8 was determined in the sample as a binder.

) of the samples were fully examined (the tests were carried out by (IGME) and the Ceramics and Refractories Technological Development Company (CERECO S.A.)).

The matrix in all the samples consists of microcrystalline material of mainly aluminosilicate composition, while in two of the samples (OR2, OR3), calcareous clay is also contained in a very small percentage. This is documented both by the chemical analysis (Table 5) where the silicon dioxide content is in the order of 60%, as well as the mineralogical analysis with X-ray diffraction. The detection of both aluminosilicate mineral phases, such as microcline, albite, kaliophyllite and muscovite and carbonaceous minerals, such as dolomite and calcite, suggest that the clay has a marl clay origin.

The inclusions in all the samples come mainly from brittle ceramics, but also, to a lesser extent, within microcrystalline ceramic matrix exist grains that appear to come from “volcanic rocks”. This can be justified by the high percentage of SiO₂ observed by both the mineralogical and the chemical analyzes carried out on the samples. Also, some dark grains observed may be of volcanic origin. The presence of anorthoclase which is the main mineral phase of basic volcanic rocks in sample OR2 strengthens the hypothesis.

The porosity of the examined bricks ranges from 25 to 40%, the apparent density is between 1.6 and 2 g/cm³ and the corresponding water absorption is 14–26%. The determined content of salts in all the bricks is low and ranges from 0.01 to 0.3%.

4. Results of the Design, the Optimization of the Composition and the Way of Application of Restoration Mortars and Grouts In-Situ during the Execution of the Works

4.1. Repointing Mortars

Initially, an attempt was made to design three basic compositions for the deep repointing, responding to the construction phases, and the color and texture of the in situ masonry mortars: one for the main church and the original parts of the narthex, both belonging to the A’ and B’ construction phase of the middle Byzantine period (reddish color), one for the narthex’s rebuilt parts constructed at the end of the 19th century to the beginning of the 20th century (off-white with reddish aggregates), and one for the exonarthex constructed at the 16th century (off-white/beige color).

In a first stage, raw materials from the local market of Chios were used, including lime putty and the available sands. However, both the first trial mixes in the laboratory and on site, revealed that the compositions for the main church and the original parts of the narthex, that had a reddish hue, were not satisfactory. Local sands and commercial ceramic powder, which is usually used for byzantine mortars, could not successfully imitate this particular reddish color.

So, it was decided to use aggregates from a reddish stone of siliceous origin which resembled the aggregates found in the surrounding area of the monument. A specific investigation was carried out for this purpose, as it is presented in what follows.

Moreover, the use of the natural pozzolan called “Hfaisteiaki Gaia” of an off-white color, instead of the slight grey color natural pozzolan “Lava Antica” from Milos Island, used for the exonarthex’s mortar, was also proved necessary from the pilot applications.

In all cases, the choice of the type and size distribution of the aggregates, was based on the granulometric gradation curve of the original mortars, on their color shade, as well as on their physical characteristics, such as their mineralogical origin (limestone-silicate) and aesthetic compatibility to the original materials.

A large number of tests and pilot applications, both in the laboratory and on site, was required for the final selection, since the original mortars, especially in the façades of the monument are complex mixtures of aggregates and binders, with a distinct texture and shade which varies from area to area in masonry, depending on various factors, as the width and depth of the joints, the level of damage or wear and the type of adjacent masonry units (stone or bricks).

It should be noted that, according to the restoration study, the patina of time in the original materials (stones, bricks and mortars), should be preserved in situ. Thus, in each area the adjacent materials and their state of preservation played an important role, in the finish of the restoration mortar, sometimes with minor modifications, influencing the final result. As mentioned above, new repointing mortar was only used to repair the defective joints (with crumbling mortar, or extensive erosion) and to rebuilt areas.

Therefore, as mentioned in the introduction, the choice of the most appropriate composition for application on the monument was based on a combination of criteria regarding both physicochemical compatibility and mechanical strength, as well as aesthetic harmony. The latter related to the compatibility of the final color shade and texture of the applied restoration mortar to the original old mortars and bricks conserved in situ. This was extremely important for the surface pointing mortars, which should not “catch the eye” and create contrast with the historic mortars still preserved in situ (Figure 1, Figure 2, Figure 3 and Figure 8), aiming at a certain coherence and harmony of the appearance of the monument as a whole. Figure 8a,c present the NE and the central part of the N façade of the main church, before the restoration intervention; damaged and decayed mortars and bricks can be seen, while some of the recessed bricks under the deteriorated surface mortars are also apparent. Figure 8b,d present the same parts of the monument 5 years after the intervention. The historic materials preserving the patina of time (stones, bricks and mortars), together with the new mortars and bricks can be seen.

The main difficulties faced were the particular reddish color of the mortars and the fact that the width of the joints to be repointed varied from approximately 3 cm for the majority of the joints, to 9 cm in surface joints between two brick courses covering the internally recessed bricks (Figure 2 and Figure 8).

Therefore, special care was taken to ensure an appropriate practical methodology for the application of the repointing in the worksite. To this end, a detailed protocol was drafted in collaboration with the craftsmen, including all the steps to be followed, from the preparation of the joints and the mortar to the placement of the mortar to the joints and the curing of the repointed area. Multiple preparatory applications of mortar in less visible areas, filling the joints in thin layers, repeatedly by hand, using small trowels and specific tools, with stable pressure for compacting them, permitted both finetuning the methodology for the specific needs of this case study, but also selecting the most qualified and experienced craftsmen to execute the repointing works on site. Care was also taken to schedule the works in such a way so that to avoid extreme seasonal conditions (i.e., high and low temperatures), as lime-based mortars are sensitive both to freezing and hot temperatures, as well as to dry and windy conditions.

The procedure of repointing adopted, which is briefly described below, was based on the literature [45,46,47] and on the experience of the Laboratory of DTRR of HMC and the multidisciplinary team supervising the works.

Damaged and eroded mortars were carefully removed from the joints, using hand chisels and mash hammers, until the appropriate depth was reached (at least equal to the width of the joint for surface repointing or approximately two times the width of the joint for deep repointing), in order to ensure adequate bonding. If loose or disintegrated materials were noticed beyond this minimum depth, these were also removed, until a clear rectangular space was achieved. The side and rear surfaces of bricks and stones were meticulously cleaned, as the existence of dust and mortar residue would have an inevitably bad effect on bonding. Thorough pre-wetting of the masonry was also ensured to minimize the loss of the mortar’s water due to absorption by the masonry materials. Pre-wetting started at least one day before the application of the repairing mortar, by spraying water without applying pressure, and continued for several times on the day of the application, until just before the application. The objective was that the joints would be dump at the time of filling with the mortar, but not surface wet (the bricks and stones should not be excessively wet and thus shiny).

During the preparation of the mortar, attention was paid to achieve the desired consistency while keeping the water to a minimum, to avoid shrinkage and to facilitate compaction during the application. The mortar should be workable, but relatively firm and “dryish”. Just after completion of the mixing, the mortar should be protected from water evaporation. Thus, it must be placed in a covered container and avoid direct exposure to windy air and sun. Small amounts should be taken progressively by the craftsman and used in each application area.

The filling of joints with mortar was carried out in layers, using small finger trowels of different widths that fitted within the joints, spatulas and other plastering tools. The filling started from the deeper areas and the thickness of each layer was of the order of 2,0 cm or even less in case of friable old mortar being left in the back of the joints, avoiding the application of too much mortar in each layer, as this could provoke shrinkage cracking. For the same reason, for the deep repointing mortar of thick joints, the addition of coarser aggregates (gravel 8–16 mm, 10% of the weight of solid materials of the corresponding mortar) was prescribed, based also on the granulometry of the original mortars, which contain a fraction of 10–15% of such coarse gravel. Additionally, when repointing mortar was used to fill internal lacunas of larger dimensions, especially during the rebuilding of some collapsed areas, it was also necessary to incorporate small pieces of stones or bricks (the so-called “chips”), by pressing them into the mass of the mortar.

The quality of the compaction of the fresh mortar into the joint is crucial, as it directly affects its bond with the existing materials. Each layer of firm dry mortar was tightly compacted with force into the back of the open joint, packing it well into the back corners. A pressing movement of the pointing tool (e.g., pointing iron) should be used instead of a sweeping one. Subsequently, the surface of each inner layer of mortar was scratched (using various shapes) in order to improve its roughness when it hardens, thus leading to better adhesion of the next layer.

Once the mortar has reached thumb-print hardness, another layer of mortar was applied. Depending on the areas, several layers are needed to fill the joint flush with the outer surface of the masonry. It is important to allow each layer to harden before the next layer is applied. Depending on the area and the period (summer-winter), a curing period of several hours or one or more days is necessary between layers to allow the previous one to stiffen and begin hardening. Since most mortar shrinkage occurs during the hardening process, the objective of layering is to minimize the overall shrinkage; thus, as hardening in lime-pozzolan based mortars is slow, during the curing period between layers, mortars must be inspected very frequently and in case shrinkage cracks appear, re-working of the mortars with adequate tools (small spatulas, etc.) and local firm compaction must be applied to seal the cracks.

When the final layer was thumb-print hard, the surface of the joints was tooled, using adequate means and procedures to reach the appropriate finishing, in order to match to the adjacent historic materials.

During the whole procedure and at least for 14 days after the completion of the repointing, the repointed area was carefully covered with wet burlap fabric, sealed with a plastic sheeting and placed in a short distance from the masonry to avoid possible staining of the mortar. The burlap was systematically sprayed with water to keep the new repointed area constantly damp. Besides avoiding shrinkage cracking, this is very important for hydraulic mortars, to ensure that adequate water for the hydration reactions of lime/pozzolan and hydraulic lime is retained.

4.1.1. Production of Mortar Aggregates by Crushing Local Rock Mined for the Purpose

As aforementioned, the reddish color of the majority of original mortars was achieved by using aggregates from a reddish stone of siliceous origin which resembles that of the surrounding area of the monument. To this end, the DTRR/HMC carried out a specific investigation to find an adequate local quarry with the aim to crush the local stones and produce reddish aggregates in suitable granulometry for the restoration mortars of the project.

Comparative analyzes were carried out at the Institute of Geological and Mineral Research on behalf of DTRR/HMC, between the aggregates of authentic mortars and aggregates collected from the surrounding area of the monument, and the aggregates from two old quarries of Chios, Stenakas and Fyrolakas. The aggregates of Stenakas were selected for use in the construction site. As presented in

Table 8. XRD results of new aggregates samples.

	Site Surrounding Area	Stenakas Quarry	Fyrolakas Quarry
Quartz	81.7%	39.6%	87%
Calcite	5%	43.6%
Kaolinite	1.1%	6.5%	13%
Dolomite	9.4%
Muscovite	2.8%	10.3%

, aggregates collected from the surrounding area have a similar composition to the aggregates of original mortars (compare values of Table 8 with values of Table 3), with the difference that minerals such as dolomite and muscovite were also detected. In general, the samples had a relatively similar mineralogical composition, which indicates their common origin.

Regarding the Fyrolakas and Stenakas quarry samples, X-Ray diffraction analysis (XRD) concluded that the Stenakas sample presents a mineralogical composition that is more similar to the aggregates of the historic mortars, compared to the sample from Fylorakas. The differences lie in both qualitative and quantitative analysis results.

Due to the fact that the quarry of Stenakas was inactive, DTRR/HMC made in 2005 the necessary request to the local Authorities, and a specific permission was issued to quarry the necessary volume of rock for the production of aggregates to be used for the restoration mortars of this unique monument.

Thus, a sufficient quantity of rock was quarried and sent to a crusher to obtain fractions for use in mortar compositions. These fractions were sand 0–4 mm, gravel 4–8 mm and gravel 8–16 mm.

It should be noted that such a special permission was granted only because of the great importance of the monument, as well as the detailed and experimentally documented request.

4.1.2. Repointing Mortars to Be Applied to the Exonarthex, Narthex and Main Church

The first mortar that was designed to have adequate properties for repointing the exonarthex’s North, South and West walls, was of an off-white/beige color. After a series of Laboratory and in situ pilot applications, the composition M1 was selected for implementation. The composition of mortar M1 and all other repairing mortars designed for the restoration works is presented in

Table 9. Mortar compositions applied on the monument.

Raw Materials % w/w	M1	M2	M3	M4	M5	M6	M7–M8	M9
Lime putty (Chios)	15	15
Hydrated Lime (powder)			15	15			10–15	15
NHL3.5					30	30
Natural Pozzolan 1 *	20
Natural Pozzolan 2 **		20	15	20			10	12.5
Ceramic Powder 0–1 mm							2.5–5
River sand light color 0–2 mm							72.5–75	72.5
Crushed sand (Chios 1), 0–10 mm. Light beige	33
Crushed sand (Chios 2), 0–4 mm. Dark beige	10	25
River sand (Kilkis), 0–3.5 mm	10	10	10	10	55	53
Reddish crushed sand (Stenakas Chios), 0–4 mm			10	15
Light-colored river sand 2–5 mm			10	10
Reddish Gravel (Stenakas Chios), 4–8 mm		20	16.25	20	15	10
Gravel (Kilkis), 3.5–8 mm		10	7.5	10
Gravel, dark color 2–8 mm	12		16.25
Reddish Gravel (Stenakas Chios), 8–16 mm						7
Water/binder	0.5–0.7, so that to achieve satisfactory workability, avoiding water excess
Flexural strength 28d (MPa)	0.5	0.7	1.0 k	0.9	1.5	1.5	0.8	1.0
Compressive strength 28d (MPa)	3.5	3.9	4. 5	4.2	5.5	5.5	3.6	4.3

* Light grey color, d_max < 75 μm, 90% of grains < 45 μm. ** White, d_max < 55 μm, 96% < 20 μm.

.

Figure 9a presents the granulometric curve of a representative original mortar from the exonarthex (sample 48) in comparison to the curve of aggregates mixture of pointing mortar M1, which was selected for application on this part of the monument.

Figure 9b presents the granulometric curves of three representative existing mortars which were used in the past to rebuild collapsed parts of the narthex (no. 16, 33, 36), in comparison to the curve of the aggregate’s mixture of the repointing mortar M2, which was selected for these parts of the monument.

A first composition for deep repointing and local reconstructions of the main church and original parts of the narthex was initially designed on the basis of an extended pilot application program. This initial composition was further improved during the course of the works and evolved into composition M3, containing a variety of sands and gravels in order to achieve, in the best possible way, the reddish color of the original mortar.

Regarding the surface repointing mortar, composition M4 was finally designed as the basic mortar. It has to be underlined that in order to achieve the best possible result in surface repointing of the facades, slight modifications were proved necessary to match the specific color and texture existing in each specific area.

Figure 10 illustrates the grading curves of the aggregates of original mortars of the main church (no. 17,26) together with the curves of aggregates mixtures that have been used in the mortar composition M3 for the deep repointing of the main church. As mentioned above, it was recommended that coarser gravel could be added in the composition (max 10% of the total weight of solids), when relatively thick joints or internal big empty spaces had to be filled (e.g., in the areas of local reconstructions of heavily damaged parts of the external façades, etc.). To this end, this quantity was identified per mortar batch, so that the technicians and masons would be able to easily use it in the worksite.

The same graphic plot (Figure 10) presents the grading curves of the aggregates of composition M4, that was designed for surface repointing of the narthex and main church façades. Composition M4 is quite similar to M3, but as expected, it contains finer aggregates than the respective deep repointing composition.

Table 9 summarizes the aforementioned repointing and deep repointing mortar compositions, to be used in the various areas of the monument. The ratio of binder/aggregates ranges between 1/1.85 and 1/2, whereas the water to binder ratio varies from 0.5 to 0.7. The flexural and compressive strength of the mortars applied, at a curing age of 28 days, is also given in Table 9. Of note, the measured strength values of the lime pozzolan mortars are very good, as they reached a compressive strength (28 days) of 3.5–4.2 MPa, values which are in accordance with the relevant literature [47], when a fine and reactive pozzolan and coarse aggregates are used and moist curing conditions are ensured.

Regarding the deep repointing of the extrados of the vaulted roof of the narthex, rebuilt after the 1881 earthquake, two natural hydraulic lime mortar compositions were selected (M5 and M6), which are also included in Table 9. The two compositions are similar; the only difference being that M6 contains coarser aggregates, as it is designed to be used in the case of wider joints or void spaces.

Moreover, mortar compositions M7 and M8, both of reddish color, were designed for the purpose of sealing cracks and installing injection tubes in areas where historic pointing mortars were to be preserved in situ, but grouting should be applied for strengthening the structure. Similarly, composition M9 was proposed for sealing cracks and installing injection tubes in the internal faces of the walls, where an off-white color was preferable over a reddish one.

Figure 11 presents characteristic images of a pilot application of mortars M1 to M3, while Figure 8b,d show mortar M4 applied on the upper part of the South East corner and the central part of the N façade of the main church, respectively. Pictures were taken 5 years after the application of surface repointing and other restoration works.

For the coating of the extrados of the vaulted roof of the exonarthex, a premixed mortar, named Albaria Struttura, was used according to the manufacturer’s instructions. Albaria Struttura is a cement-free pozzolanic lime mortar with natural siliceous aggregates with a maximum diameter of 2 mm. It guarantees a compressive strength >15 MPa and is therefore classifiable as an M15 type masonry mortar according to the European standard EN 998/2 [71]. Specimens of this material taken at the worksite and tested in the laboratory had flexural strength of 4.1 MPa and compressive strength of 18.80 MPa, at the curing age of 9 months.

4.2. Selection of the Repair Bricks

For the production of the handmade repair ceramic bricks and roof tiles, traditional brick and roof tiles manufactures were sought throughout Greece. A manufacturing facility in Kilkis was chosen for its ability to produce a solid brick and a tile type according to the specifications requested by DTRR/HMC.

The properties of the brick are presented in

Table 10. Chemical, mineralogical and physico-mechanical properties of new bricks.

Chemical Analysis %
Al2O3	CaO	SiO2	MgO	MnO	Fe2O3	K2O	Na2O	TiO2	SO3	LOI
16.52	8.57	56.15	4.66	0.01	6.75	2.23	1.82	0.7	0.44	0.45
Mineralogical Analysis (XRD) Results of the New Ceramic Brick
From the mineralogical analysis, it appears that the major phase of the sample is quartz SiO2. The aggregates are carbonates as CaCO3, and as albite NaAlSi3O8. Microcline KAlSi3O8 and Muscovite KAl2Si3AlO10(OH)2 were determined in the sample as binders.
Physico-Mechanical Properties of the New Bricks
Apparent density (g/cm3) (MIP)		Real density (g/cm3) (MIP)		Tot. intr vol. cc/cc (MIP)	Water absorption (%) (weight/buoyancy)		Porosity (%)(weight/buoyancy)		Compressive strength (MPa)
1.86		2.73		0.3302	17.6		33.4		13.0

. This brick, fabricated using specific types of clay, has a predominately aluminosilicate matrix with low concentration of carbonates. The inclusions used, originate from metamorphosed rocks—probably gneisses and quartzites.

Due to the fact that the dimensions of the bricks on the monument vary slightly, it was decided that the desired dimensions of the new bricks to be used for replacement would be 21 × 42 × 2.8 cm.

4.3. Design and Application of Grout Compositions

Taking into account the requirements of the restoration project, the necessary laboratory tests were carried out by DTRR/HMC to determine the compositions of (a) a high injectability grout for filling cracks and voids of a nominal minimum width (W_nom) equal to 0.2 mm, characterizing the monuments masonry [48], and (b) of an injectable mortar, suitable for the filling of large voids left in the masonry due to the disintegrated timber laces. Given the importance of the monument’s decorative elements consisting of various materials, as well as the existence of byzantine frescoes in the internal face of the majority of the walls [48,54,56,58], it was decided, in both cases, to use hydraulic lime-based grouts instead of low cement content ternary compositions. A hydraulic lime (named Calx Romana and described by the manufacturer as hydraulic lime containing pozzolanic material) was used, after having examined its chemical and mineralogical characteristics, summarized in

Table 11. Chemical and mineralogical characteristics of the hydraulic lime used for grouting.

Chemical Analysis %
Al2O3	CaO	SiO2	MgO	Fe2O3	K2O	Na2O	LOI
8	44	26.02	2	3.45	1.09	0.34	12
Thermal Analysis Results (TG-DTA)
This hydraulic lime shows a mass loss of 11.99%. The phenomenon begins at 600° and ends at 800 °C and is mainly due to the loss of residual calcium carbonates CaCO3.
Mineralogical Analysis (XRD) Results
The mineralogical analysis reveals that the major phase of the sample is CaCO3 in the forms of calcite and vaterite. Calcite is attributed to the aggregates. Vaterite, a secondary phase, is formed by the reaction of CaO with atmospheric CO2. CaO and Ca(OH)2, deriving from lime, are also detected. Secondary phases are Ca2SiO4, and Ca2Al2SiO7, which mainly come from the synthetic material, probably pozzolan.

.

Laser grain size analysis (carried out by (CERECO S.A.), using a Malvern Instruments Laser particle analyser) has shown that the diameter at 99% passing (d₉₉) of the hydraulic lime was less than 76 μm (d₉₉ < 76 μm) and the diameter at 85% passing (d₈₅) was less than 32 μm (d₈₅ < 32 μm); thus taking into account the penetrability criteria [48,67], it could be used for preparing grouts able to penetrate in a masonry characterized by an estimated nominal minimum width of voids, W_nom, approximately equal to 200 μm. For the preparation of the injection mortars, silicate aggregates in various particle sizes were added, with a max diameter of grains <0.8 mm or <1.25 mm, alternatively.

As aforementioned, a holistic design of hydraulic lime-based grouts was performed with the aim to ensure high injectability under low pressure, even in cracks of minimum nominal width of two tenths of millimeter (W_nom~200 μm). For this purpose, the penetrability, fluidity and stability characteristics of the suspensions were fully examined for various water/solids ratios, with or without the use of superplasticizers, using the testing methods described above.

The criteria set for the selection of the optimal compositions of the high injectability grout were the following: For the sand column injectability test, a sand of 1.25–2.5 mm was used, simulating a W_nom = 0.2 mm [48,67]. The grout was considered to be injectable if it had the capacity not only to reach the top of the column (360 mm) in a time (T₃₆) less than 50 s, but also when there was a continuous flow of grout from the sand column in the adjacent measuring vessel of at least 20 mL, as requested by EN 1771 [66] and the relevant literature [48,67]. Regarding the fluidity, the Fluidity Factor had to be higher than 0.7 × 10³ mm/s [68] and the flow time (t_d = 4.75 mm) of 500 mL of grout, out of 1000 mL inserted in the Marsh cone with 4.75 mm nozzle diameter, had to be less than 45 s and greater than 20 s [48,68]; a maximum acceptable limit of 5% was set for the bleeding test; no segregation had to be present [48,70].

Table 12. The grout compositions selected to be applied on the monument and their injectability and mechanical characteristics.

Grout Composition	G1	G2 *	G3 *
	% w/w
Hydraulic lime Calx Romana	100	100	100
Silicate sand <0.5 mm		25	12.5
Silicate sand <0.4–0.8 mm			10
Silicate sand <0.8–1.25 mm			2.5
Water	70	60	60
* The percentages of water and sands are % of the hydraulic lime
Penetrability T36 (s): sand column 1.25–2.5 mm (Wnom = 0. 205 mm)	23.2	∞	∞
Segregation; thickness of the distinct denser layer sedimented at the bottom of the recipient after the realization of the sand column test (mm)	<1	<1	<1
Fluidity Factor × 10−3 (mm/s)	1.02	0.9	0.5
Flow time of 500 mL of grout out of 1000 inserted in the March cone with 4.75 mm nozzle diameter (s)	22.5	42.2	35.55
Bleeding NFP 18-359	0.9%	0.3%	0.3%
Flexural strength 28 days/12 months (MPa).	1.3/2.0	1.6	-
Compressive strength 28 days/12 months (MPa).	3.9/5.0	4.9/12.0	4.2/10.0

presents the compositions of the best-performing grouts selected to be applied on the monument, together with their injectability (fluidity, penetrability, stability) and mechanical characteristics. All values measured satisfy the criteria set for such type of grouts as presented in the literature [48].

The grout to fill the empty spaces left after the disintegration of some of the timber laces was applied first, using injection tubes installed for this purpose in two levels along the length of the timber laces. The injection tubes were installed mainly externally, but also internally, in some areas determined together with the specialized Conservators of the frescoes to allow for the grout to be injected and the air to escape without difficulty. The same tubes also served for collecting the overflow of the grout during the injections.

After the completion of the filling of the voids of the timber laces, a second injection campaign was carried out with the aim to fill internal voids and discontinuities of the three-leaf masonry, homogenize the masonry and improve its mechanical characteristics. Due to the rich decorative brick ornamentation of the façades of the walls and the existence of cracked frescoes on most of the internal surfaces, the installation of the appropriate grid of the injection tubes was realized very carefully, in collaboration with the specialized Conservators. The preparation of the masonry and the installation of the grouting tubes and all the necessary works and procedures for the application of grouting followed the methodology applied to the restoration of the Byzantine Katholikon of Daphni Monastery [72,73], and the practical recommendations presented and discussed in the relevant literature [48].

The following criteria were used for selecting the positions of drilling holes for placing the grouting inflow and outflow tubes:

• the distance of consecutive mortar joints, (horizontal and vertical) of the brick and stone masonry),
• the thickness of the wall and its construction characteristics,
• the injectability of the designed grout in relation to the minimum nominal width of voids to be filled,
• the spatial correlation between possible cracks on inner and outer faces of the wall and the locations of severe damage, as well as the locations of timber laces,
• the decoration of most of the internal faces of masonry with frescoes.

As suggested in the literature [48], a dense grid of holes was drilled in the masonry and grouting tubes were installed. Due to the existence of frescoes on the interior faces, the injections should be mainly realized from the exterior façades, thus a denser grid of tubes was installed on them. Moreover, tubes installed in the masonry form the external façades should penetrate depths equal to 1/3, 1/2 and 2/3 of the thickness of masonry, to reach the interfaces of the infill material with the external and internal leaf, as well as the middle of the infill material. Thus, holes were drilled following the closest pattern forming a rhombus of unequal diagonals, with the horizontal one from 20 to 40 cm and the vertical from 40 to 70 cm. The drilling of holes was usually located in selected mortar joints, after careful visual inspection and observation of the masonry, in collaboration with the specialized Conservators, in order to avoid severe or unnecessary damages.

In all cases, it was necessary to install a smaller number of tubes internally, mainly for air exit and grout outflow, but also for the entrance of grout in case of wide cracks. Moreover, the existence of grout outflow tubes on the internal façades, decorated with frescoes, was crucial to monitor the movement of the grout behind the frescoes and to avoid any uncontrollable grout leakage or pressure built up [48,73].

Subsequently, transparent plastic tubes were installed into the drilled holes, externally and internally, having an internal diameter of 10 mm or smaller (6, 4, 1.8 mm), especially in the areas with frescoes (Figure 12).

When repointing is applied, the mortar used for the repointing is also used for the installation of injection tubes. In areas where the tubes are installed in existing historic pointing mortars on the external or the internal façades or in the areas of frescoes, adequate mortar compositions with fine aggregates have been designed, as presented in Table 9, Section 4.1.2.

Figure 12a presents the grouting tubes installed in the north façade of the Church of Panagia Krina, while Figure 12b presents the grouting tubes of an internal face of a wall. The tubes are numbered and marked with different color of tape to indicate their depth in the masonry and are fixed upwards with the help of wires installed for this purpose [48], as shown in Figure 12a,b. In Figure 12b, one can see the tubes installed in areas of frescoes, where the specialized Conservators have been installed, and finer injection tubes with a small diameter, necessary for the evacuation of air and possible outflow of the grout.

After the completion of grouting, all the tubes were removed, and the holes were filled with the repointing mortar corresponding to each area. As can be seen in Figure 13b (in comparison with Figure 13a), but also in Figure 8b,d presenting the monument after the completion of the restoration works, the grouting intervention is invisible. One of the advantages of this technique is that no traces are left after its application.

5. Discussion

The results presented in this paper highlight the importance of combining laboratory analysis and testing with in situ investigations for the survey and characterization of historic materials and the preliminary design of restoration ones, as well as the necessity of pilot applications and on-site quality control, using the worksite available materials and workmanship. This methodology is essential in order to optimize the proposed compositions and select those which are most appropriate in terms of functional and aesthetic compatibility and durability.

Nowadays, after many years of research and efforts, it has been generally accepted by the professionals, that the detailed analysis of the characteristics of the building stones, bricks and historic mortars, including that of the aggregates is undoubtedly absolutely indispensable and constitute the basis for designing suitable restoration mortars, respecting physicochemical and mechanical compatibility.

In the case of old monuments with various brick, mortar and frescoes decorations, this is a matter that cannot be underestimated, especially in relation to the composition and application of repointing mortars and masonry strengthening grouts or the choice of new appropriate stones or bricks to be used in restoration works of the decorated façades. Apart from the physico-mechanical and chemical compatibility and durability, aesthetic compatibility is strongly required, through an adequate harmonization of the repair mortars with the original materials.

This may demand further investigation of local materials and the production in the laboratory of additional candidate compositions using the raw materials to be found in situ. This procedure has not yet been considered as absolutely necessary before the beginning of the restoration works. It has to be admitted, however, that this presents some difficulties, mainly because of the changes in availability of the raw materials (local or even commercial) over time. Some quarries or lime factories may have ceased operation and, in some areas, sand extraction from a river may no longer be allowed, etc.

Thus, the case study presented in this paper highlights the importance of worksite optimization of the compositions, both in terms of the adequacy of raw materials and of the properties of the various mixes and the aesthetic compatibility of the intervention with the existing structure. Of course, this may not be possible in all cases, but when it comes to important monuments, it is highly advisable, in order to avoid jeopardizing the success of the whole restoration project.

Experience acquainted from many case studies proves that finding the adequate raw materials may be a critical matter. To this end, raw materials (different types of lime, pozzolan and aggregates) that are available locally in the area of the project should be analyzed in priority. But in case their characteristics are not satisfactory, one should not hesitate to check other certified materials, since the use of high-quality binders and aggregates is very important.

As proved in the case study presented, the original local aggregates that could give to the repointing mortars a special texture and color hue and thus enhance their aesthetic compatibility with the old materials may no longer be available in the local market, and similar materials cannot be found elsewhere. In such a case, it might be necessary to reproduce them even by quarrying local rocks to obtain enough aggregates for the restoration works.

Similar attention has to be given to the handmade bricks used to replace heavily deteriorated or missing bricks in severely damaged areas of a monument. Once again, based on the detailed analysis of historic bricks, it is advisable to produce several trial specimens, manually manufactured, and bring them to the worksite, in order to choose the more suitable. Their color and texture, porosity and strength should be checked, especially in case of monuments, in which bricks occupy a large part of their façades, as in the case of Panagia Krina, presented herein. It is underlined, that their suitability should also be based on their harmonization with the original materials and the new repointing mortars, and this may only be achieved on the worksite.

In all cases, systematic quality control during the entire project, is indispensable and proved to be very efficient, permitting to take into account all specific local materials peculiarities and demands, thus, ensuring the quality of intervention.

Another important matter to be pointed out is that the final selection of the mortar compositions or the bricks or stones to be used in the project has to be carried out after the evaluation of mature pilot applications of repointing or rebuilding using candidate compositions of mortars or bricks or ways of construction, in various areas of the monument (e.g., in the north or south façade, at upper or lower levels, etc.). The pilot applications constitute a very powerful tool, especially when aesthetic compatibility and harmonic coexistence, in terms of color and texture of new repointing mortars with the historic materials preserved in situ, is sought. The impact of the weather (humidity, light, etc.) should not be underestimated, as well as the fact that a long period of time is necessary for the lime-pozzolan or hydraulic lime-based compositions to set and solidify. In the case study presented, a period of a whole year was devoted in order to select the final compositions for surface repointing.

Nevertheless, in such types of interventions, much more time is necessary, to evaluate the final result, thus, the comparative photographs presented in Figure 8 were taken 5 years after the completion of the works. It is believed that the mortars finally applied constitute a very good solution. The new pointing mortars are discreet and slightly lighter in color than the existing ones, and there is a harmonious coexistence with the old materials, which will improve in the future as the mortars take on a certain patina over time.

6. Conclusions

The works presented in this paper allow for the following conclusions to be made:

• The holistic multidisciplinary approach presented, combining visual inspection, laboratory investigations and in situ pilot applications before and during the execution of the works is of great significance in the case of important monuments. The extensive pilot applications and the experimental verification at the laboratory during the works, allowed for a rational way of optimizing the compositions, complying in all cases with the requirements set for technical and aesthetic compatibility.
• The compositions developed, and the methodology of application used, may constitute a solid base for repointing interventions in monuments built with similar masonry characteristics, especially regarding the so-called “recessed brick” technique and the rich decorative brick ornamentation.
• The satisfactory results presented in this case study and the experience obtained during the whole procedure, from analysis to design and to in-situ application, can be a useful guide for the design and application of mortars and grouts on other monuments.