Damage assessment and restoration proposal following the 2023 Türkiye earthquakes: UNESCO World Heritage Site Diyarbakır City Walls, Türkiye

Abstract

Diyarbakır City Walls, one of the longest defensive structures in the world, following the Great Wall of China, the walls of Antakya, and the walls of Istanbul, is a UNESCO World Heritage site since 2015. With a history of approximately 5000 years, the Diyarbakır City Walls have been affected by consecutive earthquakes centered in Kahramanmaraş in 2023, resulting in damages to various sections. Urgent restoration and repair interventions are needed for these sections of the Diyarbakır City Walls due to earthquake-induced damages. Although there are limited studies presenting stone analysis of the Diyarbakır City Walls in the literature, no studies focusing on mortar analysis have been found. The objectives of this study are as follows: (I) to identify the mechanisms and factors of earthquake damages in the Diyarbakır City Walls, (II) to conduct necessary analyses for the selection of mortar materials for post-earthquake repairs, and (III) to provide restoration and strengthening recommendations to ensure the sustainability of the original structure. Observational, petrographic, chemical, and SEM analysis techniques were used, and the findings were interpreted comparatively. The results demonstrate that the most severe damages after the earthquake in the Diyarbakır City Walls were caused by the inadequate adhesion of missing mortar joints and different types of materials used between double-walled structures. Additionally, the presence of clay minerals identified in the mineralogy of the mortar through experimental analysis was defined as an internal issue causing the loss of mortar due to osmotic pressure created by water absorption. Another factor causing the loss of mortar is the presence of chloride-type salts, which were found to be present in a significant amount in all samples and were attributed to the use of Portland cement in previous faulty repairs. It was also determined that recent faulty repointing works contributed to the loss of mortar. Finally, this article presents original restoration and strengthening recommendations to repair the earthquake-induced damages and prevent their reoccurrence in the future.

Introduction

It is essential to transfer authentic technical data of historic buildings to future generations without compromising them after contemporary conservation practices [1,2,3]. Unfortunately, some modern restoration interventions have been the most damaging to these structures [4]. Particularly, engineering solutions that introduce new technologies to enhance structural safety due to reasons such as earthquakes can, in many cases, lead to data loss that deviates from the intended goal of the restoration [5]. Earthquakes, in their aftermath, often alter, diminish, eliminate, or introduce new elements to the original technical information of historical buildings [6,7,8]. Contemporary repair principles emphasize the need for detailed documentation and analysis of structures before post-earthquake interventions, both from technical and historical perspectives [9]. However, it is evident that many repairs carried out in the twentieth century lack a comprehensive analysis of the causes of damages that occurred on historical structures after earthquakes [10].

In the restoration practices and strengthening recommendations of historic buildings, there is a prejudice that assumes the structural behavior of historical structures is similar to modern ones. However, this approach has more disadvantages than benefits in the long run [11]. Since load-bearing elements of masonry buildings do not exhibit linear elastic material behavior, it is essential to examine the structural history for accurate and reliable structural analysis. To examine the structural history of a building, there should be documentation on the types and factors of damages that occurred in similar historical structures after earthquakes. Having knowledge about the past damages in a historic building after an earthquake provides significant advantages in predicting and mitigating future damages during subsequent earthquakes. Knowing the damages that occurred during past earthquakes in a structure guides the reduction of damages in subsequent earthquakes and the development of appropriate seismic strategies. This information helps us determine the correct methods and techniques to enhance the seismic resilience of the structure and anticipate potential risks on the building. Therefore, meticulous documentation and recording of damages that occur during earthquakes in historic structures contribute to the successful planning of future preventive measures against earthquakes [12, 13]. Additionally, conducting structural and experimental analyses to investigate the historical load types and mechanical properties of the structural materials forming the load-bearing system can improve the accuracy of performance predictions during earthquakes for historic buildings [14, 15].

Accurate identification of the types and causes of damage resulting from earthquakes forms the basis for determining the methods to be used in the repair process. Different materials interact differently and exhibit different behaviors during earthquakes [16]. Each material has different mineralogical compositions and physical properties [17]. Therefore, before any intervention related to conservation and restoration is carried out on a structure after an earthquake, the factors causing the damage in the structure should be examined, and the resulting damages should be identified. In this context, observational observations should be made to determine the main factors affecting earthquake damage, followed by experimental and structural analyses to support these observations [18]. This process is crucial in guiding the selection of repair materials and directing the restoration process accurately during post-earthquake repair and conservation of historic buildings, ensuring the transmission of the structure to future generations in its authentic form.

One of the studies is Yalçıner et al. that investigates the resistance factors of structures against earthquake behaviors through various analyses, emphasized the benefits of the study in determining the correct methods and techniques to enhance the earthquake resilience of the structure. The study pointed out that frequent restorations of the Chora Church in the region due to recurring earthquakes led to the loss of the building's authenticity. Moreover, they highlighted that non-destructive diagnostic techniques used could reveal even hidden new historical details by previous restorations [19]. Işık et al. examined the seismic behaviors of tombs in the district of Ahlat through various analyses indicated that Ahlat stone has low durability and a porous, soft structure, leading to mass losses, erosion, and fractures over time under natural conditions. The study found that the secret to the preservation of these tombs lies in the appropriate engineering design of the tomb structures [20, 21]. In a restoration-focused non-destructive structural investigation of the Istanbul Basilica Cistern (Yerebatan Sarnıcı), a site listed in the UNESCO World Heritage List and ranking among the top 10 in Turkey, Aydıngün et al. detected deformations in the waterproofing layers at the base of the cistern. The study suggested that the deterioration causing these abnormalities could be related to past earthquakes and recommended measures in this context [22]. In the literature, the properties of mortar, one of the construction materials used particularly in stone structures, and its importance within stone walls during an earthquake are emphasized. The collective results of these studies indicate that the investigation of mortar properties and their effects on stone walls during an earthquake provides a better understanding and yields significant findings related to factors contributing to earthquake damages [23, 24].

Earthquakes have caused significant damage to historical buildings throughout history, especially in countries situated along earthquake-prone zones like Turkey. The North Anatolian Fault Zone and the East Anatolian Fault Zone are among the highest earthquake risk regions in Turkey, known for earthquakes of high energy and long recurrence intervals. When historical earthquake catalogs are examined, many major earthquakes up to intensity VIII–IX have affected our study area, the Southeast Anatolia Region.The most significant earthquakes that have occurred on the East Anatolian Fault System in the last few centuries include the 1513 Pazarcık earthquake, 1822 Kahramanmaraş earthquake, 1866 Karlıova earthquake, 1872 Antakya earthquake, 1874 Gezin earthquake, 1875 Sivrice earthquake, 1893 Çelikhan earthquake, 1905 Pötürge earthquake, 1971 Bingöl earthquake, 1977 Palu earthquake, June–July 1986 Sürgü earthquakes, and the 08.03.2010 Karakoçan earthquakes. Among these, the ones in proximity significantly affected Diyarbakır and its surrounding areas. Particularly, the 1866 Kulp earthquake and the 1975 Lice earthquake were significant earthquakes that seriously affected Diyarbakır and its surroundings [25].

Finally, on February 6, 2023, two earthquakes occurred in Turkey, centered in Pazarcık, Kahramanmaraş with a magnitude of 7.7, and Elbistan, Kahramanmaraş with a magnitude of 7.6, along the East Anatolian Fault Zone. Even though the earthquakes were more than 300 km away from the focus point, they were strongly felt in the neighboring provinces. These earthquakes resulted in significant damage to modern and historical buildings in Diyarbakır, the focus of our study [26]. Diyarbakır, located in the southeast of Turkey, is a city of significant historical and architectural importance. One of the most prominent features displaying its historical texture is the city's surrounding walls. Declared a UNESCO World Heritage site in 2015 and boasting a history of approximately 5000 years, these walls have been affected by the earthquakes that have occurred and have suffered various damages over time [27, 28]. Finally, the earthquakes on February 6, 2023, resulted in serious damage to certain sections of the Diyarbakır walls. While such damages were not frequently observed in areas where previous repairs and restorations had been carried out, they were more commonly seen in areas where repairs had not yet been made. This situation underscores the urgent need for restoration and repairs due to earthquake-induced damages in the Diyarbakır Walls, which bear the traces of approximately 30 civilizations dominant throughout the historical region. Therefore, a detailed analysis of the causes of earthquake damages should be conducted before restoration, and urgent measures should be taken to prevent the recurrence of the same problems in future earthquakes. Material characterization is also important for the selection of appropriate materials for repairs.

The main construction material used in Diyarbakır Fortress is basalt. While there have been studies in the literature on the characterization of basalt, the primary construction material used in the Diyarbakır Fortress, no research has been conducted on the characterization and deterioration mechanism of the mortars used to fill the joints between the stone materials in the structure [29,30,31]. Even though the properties of stone materials are partially known in the context of restoration practices after earthquakes, there is no information or research available in the literature regarding the properties of mortar materials. However, it is crucial to select the appropriate replacement materials for the mortars before carrying out repair interventions in the urgently needed restoration of the fortress. Unfortunately, there is a lack of literature or resources to guide experts in this matter. The objectives of this study are as follows: (I) to identify the mechanisms and factors of earthquake damages in the Diyarbakır City Walls, (II) to conduct necessary analyses for the selection of mortar materials for post-earthquake repairs, and (III) to provide restoration and strengthening recommendations to ensure the sustainability of the original structure. The main focus of this study is to provide guidance to experts in taking preventive measures for the problems the Diyarbakır City Walls may encounter during earthquakes and to facilitate the implementation of correct practices for post-earthquake repairs, aiming to preserve this cultural and historical heritage.

Study area

Diyarbakır is located in the southeastern region of Turkey, at an elevation of 653 m (approximately 400 and 200 degrees latitude and 380 and 500degrees longitude). Although there is no precise information about the construction date of the Diyarbakır City Walls, it is believed that the inner castle was built on a rock called Fis Kayası, which is 100 m high and adjacent to the eastern part of the city, in the Dicle Valley [32]. The history of the city is thought to date back to the Hittite and Hurrian periods (3500 BC). Recent archaeological research at Ergani Çayönü has shown that the settlement history dates back to 7000 BC. In 349 AD, during the reign of Roman Emperor Constantine II, the city of Diyarbakır was surrounded by walls [33]. Diyarbakır Fortress was inscribed on the UNESCO World Heritage List as the 14th cultural heritage site of Turkey in the 39th session of the UNESCO World Heritage Committee held in 2015 [34].

The Diyarbakır Fortress and Walls, situated 100 m high above the Dicle Valley, consist of complementary inner and outer castles adorned with carvings and reliefs reflecting the architectural features of 30 civilizations that ruled the city. It serves as a collective and documented example of Anatolian history. The Diyarbakır City Walls are one of the grandest and most significant monuments ever built by humans, encapsulating the essential records of that historical era. Along the walls, there are four main entrances: Dagkapı, Urfa Gate, Mardin Gate, and Yeni Gate. The black basalt walls represent an important example of medieval military architecture [33].

The total length around the Diyarbakır city walls is approximately 5200 m. The length of the Inner Castle, which remains within the inner part of the city (Suriçi), is 599 m. The total length of the city walls, including the Inner Castle, is approximately 5800 m. Based on modern measurements, the thickness of the towers and city walls varies between approximately 1.40 to 5.00 m. The thinnest section of the towers is on the eastern side of the city. In this area, the wall thickness is mostly between 1.40 and 2.60 m. The thickest section of the towers is on the north and west sides of the city. The towers in this area have a thickness of about 4.5 to 5 m. The walls of the circular towers have a thicker cross-section. The height of the towers and city walls varies approximately between 7.6 and 22 m [35].

The Diyarbakır City Walls are divided into outer and inner walls. The outer wall is approximately 5 km long and consists of 82 towers that have been restored or built at different times. The main fortress houses 16 towers. The outer walls of the fortress feature monumental towers with inscriptions and reliefs, which have been given specific names by different civilizations of various periods. Some of the notable towers include Evli Beden (Ben-u Sen, Ulu Beden) Tower, Yedi Kardeş Tower, Keçi (Kıci) Tower, Nur (Melik Şah) Tower, Selçuklu Tower, Fındık Tower, Leblebi Kıran Tower, Kral Kızı (Yeni Kapı) Tower, Abbasi Tower, Akrep Tower, and Tek Beden Tower [36] (Fig. 1).

Fig. 1

The geographical location of Diyarbakır province and the fortress walls

The architectural layout of the fortress towers is generally semi-circular, but some are polygonal and use square or rectangular planes. To climb up the fortress, the stairs on the left and right sides near the gate inside the fortress are used. All walls within the fortress have various arrow slits. These arrow-slit trenches were designed to allow armed soldiers to fight and pass through comfortably. The lower floor served as a storage area, while the upper floor was used as barracks for soldiers during times of war [37].

Methods

The study was conducted in four stages, namely observational, petrographic, chemical, and SEM analysis. Firstly, a field survey was conducted to examine the structure on-site and identify the damages caused by the earthquake observationally. Subsequently, samples were collected for material characterization to select the appropriate material for repairing the damages caused by the earthquake. Nine mortar samples were taken from the structure. The mortar samples were selected from towers 41 to 42 under limited permissions due to post-earthquake safety considerations. The experimental analyses were performed at the Istanbul Conservation and Restoration Laboratory (KUDEB). In the third stage, in the laboratory, under the petrographic analysis, the qualities of the materials in terms of binder/aggregate/additive content and ratios were determined using stereomicroscope analysis. Total moisture, organic, and carbonate (${\mathrm{CO}}_{3}^{2-}$) content were determined through calcination analysis, while proteins, fats, and water-soluble salts were identified using spot tests. Lastly, SEM images were taken from different areas of the mortar samples to examine the binders, and EDX analysis was conducted on the marked areas in these images. The steps followed in the study and the findings obtained are detailed below.

Observational damage assessment and sampling

First, the damages that occurred in the structure after the earthquake were examined observationally, and the damaged areas were photographed in detail. In the subsequent stage, samples were collected for the selection of mortar materials to be used in post-earthquake repairs.

The samples collected from the structure were specifically chosen to confirm the damages caused by the earthquake. However, due to the strict protection measures in place after the earthquake, it was not possible to obtain a large number of samples from the walls. Therefore, the experimental analysis was conducted with a limited number of samples. Since there is no existing literature on the material characterization of the mortar used in Diyarbakır City Walls, even the limited characterization carried out in this study will provide a valuable resource for experts in selecting the correct mortar material for post-earthquake restoration applications in such an important structure.

Diyarbakır City Walls consist of 82 towers. Within the limited permissions granted for the structure with 82 towers, samples were only collected from the locations of towers 41–42 (Table 1). The tower identified as Tower 41, known as the Yedikardeş Tower, provides significant information about the traditional urban life in Diyarbakır and the construction technique of the walls, making it one of the most important towers of the historic Diyarbakır City Walls. Similarly, Tower 42, also known as the Yedikardeş Tower, is an important tower due to its construction dating back to earlier periods. Towers 41–42 are located in the southern direction of the city, just south of Ali Paşa Mosque. Diyarbakır City Walls include a total of 36 cylindrical-shaped towers, 29 of which are on the western and northern facades. Tower 41, one of these towers, stands out as the most magnificent and beautiful example of a semi-cylindrical tower plan, while Tower 42 has a broken rectangular plan with rounded corners. There is a height difference of 1.15 m between the inner and outer parts of Tower 41. The samples collected from towers 41 to 42 are described in the table below (Table 1).

Table 1 Description and location of samples in the walls

Petrographic analysis

To prepare thin sections of mortar samples, suitable-sized sections are first obtained. The cut mortar samples are then coated with a material such as epoxy resin. Epoxy resin penetrates the internal structure of the samples, creating a solid matrix that stabilizes the sections. It allows the samples to be accurately cut and subsequently examined under a microscope. The epoxy-coated mortar sections are flattened and prepared for examination using a section grinding and polishing device. This step involves smoothing the surface of the sections to obtain a clear image. The thickness of the sections should be adjusted appropriately for thin sectioning. Thin sections can be examined using a polarizing microscope (cross-polarization). Polarizing microscopes utilize polarized light to study the optical properties of samples. Thin sections are placed in a cross-polarization arrangement, which enhances the visibility and identification of mineral phases. This allows for obtaining information about mineral content and approximate proportions. Mortar samples embedded in epoxy can be examined at different magnification levels using a stereo microscope. Stereo microscopes provide three-dimensional imaging and enable the observation of samples in a wider field. In this way, the mineral content and proportions can be examined in more detail [38].

In the study, the mineral contents and approximate proportions of the prepared thin sections of epoxy-embedded samples were determined by examining them under a polarizing microscope (cross-polarization) and a stereo microscope. The thin sections analyzed for petrographic analysis were prepared from samples obtained by immersing dispersed mortar samples in epoxy resin (Araldite AY103- Hardener HY956, Ciba-Geigy). Minerals were identified in a manner that allowed for qualitative and semi-quantitative evaluation of the mineral content of the samples. These thin sections were used to determine if there were any similarities or differences in mineral content between the samples. Petrological analysis can also assist in identifying acid-soluble particles. After being immersed in epoxy, the samples were cut into thin sections using a low-speed saw. The sample sections were thinned down to 30 microns, and the qualitative and semi-quantitative analysis of minerals was conducted using a Leica MZ6 stereo microscope. The texture and aggregation/bonding characteristics of the samples were obtained from thin/thick sections prepared for analysis using a thin-section light microscope and examined under stereo microscopy and polarized light microscopy [39,40,41].

Chemical analysis

Qualitative and quantitative analyses of water-soluble salts, ignition loss analysis, acid loss analysis, and sieve analysis are used for chemical analysis. For ignition loss analysis, samples were finely ground and weighed in porcelain crucibles (500 mg). The samples in the crucibles were heated in a muffle furnace at temperatures of 105 °C, 550 °C, and 1050 °C for 2 h, 1 h, and 30 min, respectively. After each heat treatment, the samples were cooled in a desiccator and reweighed. Moisture content, loss on ignition at 550 °C, and calcium carbonate content were calculated based on the weight difference of the samples. Qualitative and semi-quantitative analysis of water-soluble salts (chloride, sulfate, nitrate, carbonate, etc.) that may be harmful to the mortar was conducted. Total salt content was evaluated using the measured conductivity values. A stock solution of each mortar sample (0.50 g sample in 50 ml deionized water) was prepared, and the dissolved salts were analyzed by conducting conductivity measurements. In acid loss and sieve analyses, the binder matrix was reacted with 10% HCl acid. At the end of the process, silica aggregates and other insoluble materials were separated, and the acid loss ratio was calculated. After acid loss analysis, a sieve analysis was conducted to determine the size and abundance of the remaining aggregates. The type, size, color, content, and average abundance of these aggregates were measured using a stereoscopic optical microscope. Dried samples weighing 25–50 g were subjected to a reaction with HCl acid to dissolve the binder matrix. The insoluble portion was filtered, washed, and dried again in an oven at 105 °C. Silica aggregates insoluble in acids were passed through sieves of sizes 125, 250, 500, and 1000 microns, 2.4 mm, and 8 mm. Finally, visual analysis was performed under a stereomicroscope on aggregates of various sizes that passed through the sieves [42,43,44,45].

Scanning electron microscope (SEM) analysis

Scanning Electron Microscopy (SEM) is a method used to obtain high-resolution images for analyzing the microstructure, surface morphology, and composition of materials [46]. In the case of mortar samples, SEM images were taken from different areas of the samples to examine the binders, and EDX analysis was performed on the marked areas of these images.

Thick slices (2–4 mm) from the epoxy-embedded samples were polished using diamond polishing powders (3 µ, 1 µ, and 0.25 µ) according to the method. The gold-coated samples were then subjected to SEM and EDX analysis using Leitz rem 1600 T and Philips XL 30 instruments, respectively, to identify the minerals and textures of the samples and determine their contents. All results were compared, and their compatibility was observed. As a result, one SEM image was obtained for each sample in sequence.

Results

Upon the observational examination of the structure, the types of deteriorations present were determined. In the laboratory, mortar samples collected from the field were subjected to petrographic, chemical, and SEM analyses. The results obtained from these analyses are presented in this section.

Observational findings on earthquake damages

Based on the observational analysis conducted in the field, it has been determined that the most significant damage in the structure resulted from out-of-plane failure modes. On the outer facing of the Diyarbakır Castle walls, surface stones were observed to have fallen off, accompanied by the formation of vertical cracks in the affected areas. The reason for such damage is attributed to the different responses of the various types of materials used in the castle walls when subjected to horizontal loads. The walls of Diyarbakır Castle are double-walled. Generally, throughout the walls, ranging from 3 to 5 m in height, high-strength basalt stones were not used, and lower-strength rubble stones, such as unprocessed small stones extracted from quarries, were utilized in the central core of the walls. Despite the reasonable quality of these stones, their flat and round surfaces hindered the development of significant bonding and cohesion, leading to the easy disintegration of the structural walls. While cut stones are visible on the outer faces, the central portions are filled with rubble stones. During an earthquake, the separation and detachment of the cut stones from the outer faces and the rubble stones are frequently observed due to the different responses of cut stones and rubble stones to the horizontal loads during seismic events (Fig. 2a).

Fig. 2

The earthquake damages observed in the walls include: a spalling in the coatings, b shear cracks, c loss of pieces, and d partial collapses and displacements

During the investigations, various instances of diagonal shear damage were also observed in the wall elements of the structure. The primary cause of this condition is the deterioration of mortar in the stone joints over the years (Fig. 2b). The stone masonry of the structure exhibits joint gaps that have formed due to inadequate maintenance prior to the earthquake. This has resulted in insufficient bonding between the stones and the mortar, leading to a decrease in the material's durability. As a consequence of the seismic effects, occurrences of crumbling and falling have been observed in these areas.

Another cause of joint deterioration is the water flowing from the water channels on the walls, which comes into contact with the wall surface. This leads to the softening of the binding mortar and the loss of its cohesive properties. It is believed that during the earthquake-induced shaking, this area is vulnerable to collapse. Additionally, in the city of Diyarbakır, the extreme temperature differences between day and night seasons are thought to have a negative impact and pose a risk. Freezing and thawing of the flowing water due to temperature variations cause expansion and contraction, further exacerbating the situation. In areas where joint gaps exist in the walls and where surface stones have previously fallen off, it was observed that new instances of stone detachment and loss occurred more frequently. The risk of surface stone detachment in these areas is still ongoing (Fig. 2c).

As a result of the behavior of the outer plane of the stone walls, cracks were observed in the areas of highest stress accumulation, and displacements occurred perpendicular to the direction of the earthquake motion. Urgent intervention is required in areas where there is loss of pieces and where the walls are leaning outward. The most significant damage occurred in the section of the wall between bastions 18 and 19. In these areas, extensive falling and collapsing were detected, and it can be seen that the spread of the damage has affected a large area of the upper part of the wall (Fig. 2d).

It has been observed that the rocky cliffs on which the Inner Castle, considered as the original settlement of Diyarbakır, was built, have experienced rock mass failures, and numerous cracks have been observed on the same rocky surfaces (Fig. 2e).

If the damages are listed in general, the following can be observed in the historical walls: partial wall collapses at a total of 30 points between bastions, spalling in the coatings, horizontal-vertical cracks, vertical-horizontal mortar joint gaps, and stone detachments. Urgent intervention is required for the loss of pieces caused by stone detachment and partially collapsed walls. Taking immediate action in these areas is necessary to prevent the loss of structural values within the debris piles and to prevent further losses in these severely damaged areas. Failure to address these issues in a timely manner can lead to even greater losses in these areas.

Petrographic analysis findings

The mineral contents and approximate ratios of samples embedded in epoxy were determined by examining thin sections prepared from them under a polarizing microscope (crossed polarizers) and a stereo microscope, and the results are provided below.

Sample 1: It has been determined that the binder content of Sample 1 is approximately 30% and the mortar is produced using Hydrated Lime with a Blown Cream Lime binder. The aggregates of the sample consist of approximately 35% dolomitic crushed stone and dust, and 35% basaltic sand. The paste structure of the sample contains sporadic black cinder particles. The aggregate size of the sample is below 5 mm (occasionally 8 mm) sieve. The binder-aggregate ratio is approximately 1:2.5–3.

Sample 2: It has been determined that the binder of Sample 2 is a mortar produced by adding 15–20% Blown Cream Lime to the Portland Cement content in the range of 100–150 dosage. The aggregates of the sample consist of a small amount of dolomitic crushed stone and dust, with the remaining aggregates being entirely river sand. The paste structure of the sample contains sporadic black cinder particles. The aggregate size of the sample is below 2 mm (occasionally 4 mm) sieve.

Sample 3: It has been determined that the binder of Sample 3 is a mortar produced by adding 15–20% Blown Cream Lime to the Portland Cement content at a dosage of approximately 150. The aggregates of the sample consist entirely of river sand. The aggregate size of the sample is between 4 and 5 mm (occasionally 6 mm) sieve.

Sample 4: It has been determined that the binder content of Sample 4 is approximately 30–35% and the mortar is produced using Hydrated Lime with a Blown Cream Lime binder. The aggregates of the sample consist of 20% dolomitic crushed stone and dust and 80% river sand. The paste structure of the sample contains approximately 10% ash and impurities in terms of contamination. The aggregate size of the sample is between 3 and 4 sieve. The binder-aggregate ratio is approximately 1:3.

Sample 5: It has been determined that the binder content of Sample 5 is approximately 30–35% and the mortar is produced using Hydrated Lime with a Blown Cream Lime binder. The aggregates of the sample consist of 20% dolomitic crushed stone and dust and 80% river sand. The aggregate size of the sample is below 5 sieve. The binder-aggregate ratio is approximately 1:3.

Sample 6 was determined to be a mortar produced using approximately 30% hydrated lime. The aggregates in this sample consist of 40–45% basaltic sand and 60–65% dolomitic crushed stone and dust. The aggregate size of this sample is below 2 mm (occasionally 4 mm). The binder-to-aggregate ratio is approximately 1:2.5–3.

Sample 7 was determined to be a mortar produced using approximately 30–35% hydrated lime. The aggregates in this sample consist of 20% dolomitic crushed stone and dust and 80% river sand. The mortar structure of this sample contains 2–3% black slag particles and occasional impurities. The aggregate size of this sample is below 4 mm (occasionally 7 mm). The binder-to-aggregate ratio is around 1:3.

Sample 8 was determined to be a mortar produced using approximately 20–25% hydrated lime. The aggregates in this sample consist of 25–30% dolomitic crushed stone and dust, 10% black sand, and the remaining 60–65% is basaltic sand. The mortar structure of this sample contains around 30–35% carbonated lime lumps and occasional black slag particles. The aggregate size of this sample is below 3 mm (occasionally 10 mm). The binder-to-aggregate ratio is around 1:3.

Sample 9 was determined to be a jointing mortar produced using approximately 30–35% hydrated lime. The aggregates in this sample consist of 3–5% black slag particles, 15–20% dolomitic crushed stone and dust, 5–10% black sand, and the remaining 65–70% is basaltic sand. The mortar structure of this sample contains around 35–40% carbonated lime lumps. The aggregate size of this sample is below 2 mm. The binder-to-aggregate ratio is approximately 1:2.5–3 (Figs. 3, 4).

Fig. 3

Petrographic images of samples from stereo mikroskop. (1) The binder area is approximately 25–30%, and the binder-aggregate interface and the binder’s internal phase are good. (2) The binder area is approximately 10–15%, and the binder-aggregate interface and the binder's internal phase are intermittently good and weak. (3) The binder area is approximately 25–30%, and the binder-aggregate interface and the binder’s internal phase are good. (4) The binder area is about 20%, and the binder-aggregate interface and the binder’s internal phase are good. (5) The binder area is approximately 40–45%, and the binder-aggregate interface and the binder’s internal phase are good. (6) The binder area is approximately 20–25%, and the binder-aggregate interface and the binder’s internal phase are good. (7) The binder area is approximately 20–25%, and the binder-aggregate interface and the binder’s internal phase are weak. (8) The binder area is approximately 20–25%, and the binder-aggregate interface and the binder’s internal phase are weak. (9) The binder area is approximately 30–35%, and the binder-aggregate interface and the binder's internal phase are weak

Fig. 4

Petrographic images of samples from polarizan mikroskop. (1) The aggregates of the sample containing approximately 25–30% lime lumps are mainly volcanic rock fragments, with occasional metasandstone fragments and quartz minerals. The aggregates are mostly subrounded to rounded in shape. (2) The aggregates of the sample containing approximately 10–15% lime lumps are minerals (mostly quartz, occasionally alkali feldspar), and there are also rock fragments in the sample (mostly volcanic rock fragments, chloritized rock fragments, sporadic metasandstone and limestone fragments). The aggregates are generally equidimensional and slightly rounded in shape. (3) The aggregates of the sample containing sporadic brick fragments and approximately 40–45% lime lumps are mostly rock fragments (mainly volcanic rock fragments, occasional metasandstone fragments), with occasional minerals such as quartz and amphibole. The aggregates are subrounded to rounded in shape. (4) The aggregates of the sample containing approximately 40–45% lime lumps and spatially 3–5% black slag fragments and brownish masses are quartz and rock fragments. The aggregates are subrounded in shape. (5) The aggregates of the sample containing very few black slag fragments and lime lumps are predominantly sand aggregates (mineral and rock fragments) and brick fragments, almost equally distributed. The rock fragments are of volcanic type, and the minerals are quartz and amphibole. (6) The aggregates of the sample containing approximately 2–3% black slag fragments, approximately 45–50% lime lumps, and around 5% brick fragments are rock fragments (chloritized rock fragments). Additionally, there are sporadic occurrences of quartz and amphibole. The aggregates are rounded in shape. (7) The aggregates of the sample containing approximately 2–3% black slag fragments, a single piece of approximately 5–10% brick fragments, and approximately 20% lime lumps are rock fragments (volcanic type) and quartz mineral. The aggregates are subrounded to rounded in shape. (8) The sample contains approximately 30–35% lime lumps and around 10% quartz, with the remaining being volcanic rock fragments. The aggregates have angular to subangular and slightly rounded forms. (9) The aggregates of the sample containing approximately 3–5% black slag fragments and approximately 40–45% lime lumps are quartz and volcanic rock fragments. The aggregates have angular to subangular and slightly rounded forms

Chemical analysis results

Spot tests were conducted to determine the characteristics and quantities of water-soluble salts (chloride, sulfate, carbonate, and nitrate salts) present in the defined samples, as well as to detect the presence of additives such as saponifiable fats and proteins. The relevant analysis results are provided below.

In the spot anion tests conducted on the mortar samples (Cl⁻, ${\mathrm{SO}}_{4}^{2-}$, ${\mathrm{CO}}_{3}^{2-}$, NO₃⁻), a high level of nitrate and chloride (NO₃⁻ and Cl⁻) salting was detected in the mortar samples, but no sulfate (${\mathrm{CO}}_{3}^{2-}$) salting was found. Due to the presence of lime-based binders, the total water-soluble salt content in the mortars is relatively high. The total salt content in the mortars ranges from 0.89 to 7.82% (average of 3.82%). All samples showed a significant amount of chloride (Cl⁻) salting. While no fat content was found in the mortar samples, protein was detected in all samples (Table 2).

Table 2 Water-soluble salts, conductivity, protein and oil analyzes

The results of the calcination (heat loss) analysis conducted at 105 ± 5 °C, 550 ± 5 °C, and 1050 ± 5 °C for mortar samples are provided in Table 3, which includes the ratio of non-reactive silicate aggregates and their size distribution after the acid treatment. The moisture content of the mortar samples ranges from 1.05 to 7.80% (Table 3). After being treated with acid and fragmented, the silicate aggregates that did not react with the acid were separated according to their sizes through sieve analysis. Subsequently, their visible characteristics were examined under a stereo microscope.

Table 3 Granulometric analysis of aggregates with loss of ignition and acid

In the results of the calcination analysis of the mortars, it can be observed that the total carbonate content of the mortar samples subjected to acid treatment varied between 29 and 72–69.79%. Additionally, the binder/aggregate (B/A) ratio of the mortar samples in the results is found to be 1/2.5–3 (Table 3).

Scanning electron microscope (SEM) micrographs

SEM images were taken from different regions of mortar samples for the examination of their binders, and EDX analyses were conducted in the marked areas on these images. All results were compared, and as a result of the analysis, one SEM image for each sample is provided below, respectively (Fig. 5).

Fig. 5

SEM images and elements detected in the EDX analysis are as follows: (1) It has been determined that the sample contains a higher amount of calcium and silicon compared to other detected elements, and a lesser amount of magnesium, aluminum, iron, sodium, and sulfur. (2) The sample contains a significant amount of calcium, and a lesser amount of silicon, magnesium, aluminum, iron, sodium, potassium, and chlorine. (3) It has been found that the sample contains a higher amount of calcium and silicon compared to other detected elements, and a lesser amount of magnesium, aluminum, iron, potassium, and chlorine. (4) The sample contains a higher amount of calcium and magnesium compared to other detected elements, and a lesser amount of silicon, chlorine, potassium, sodium, aluminum, iron, and sulfur. (5) It has been determined that the sample contains a high amount of calcium, and a lesser amount of silicon, magnesium, aluminum, potassium, and chlorine. (6) The sample contains a high amount of calcium, and a lesser amount of magnesium, silicon, aluminum, sulfur, and chlorine. (7) It has been found that the sample contains a high amount of calcium, and a lesser amount of silicon, magnesium, chlorine, aluminum, sodium, potassium, iron, and sulfur. (8) The sample contains a high amount of calcium, and a lesser amount of magnesium, silicon, aluminum, and iron

Discussion

The aim of the study is as follows: (I) to identify the mechanisms and factors contributing to earthquake damage in the Diyarbakır City Walls, (II) to conduct necessary analyses for selecting mortar materials to be used in post-earthquake repairs, (III) to provide restoration and reinforcement recommendations that ensure the sustainability of the structure's original state. To achieve these objectives, observational, petrographic, chemical, and SEM analysis techniques were employed, and the findings were interpreted comparatively.

Observational analyses have revealed that the major damage in the structure is attributed to out-of-plane deformation. It has been identified that most of the affected areas lack mortar joints. Tomazević [47], based on previous experiences, suggested that the out-of-plane behavior of masonry walls can lead to cracking in regions with the highest stress concentration, resulting in partial losses perpendicular to the earthquake direction. Speranza [48] explained that the out-of-plane behavior of walls and associated partial losses are directly related to the quality and strength of joints and wall elements. When examining the post-earthquake damages in Diyarbakır Castle, they support the findings of these researchers. The majority of the damages are observed to be caused by voids in the mortar joints and inadequate adhesion of different materials used between double-walled structures. The stone elements of the structure have gaps in the joints due to insufficient maintenance prior to the earthquake. This condition has led to poor adhesion between stones and mortar, resulting in decreased material durability of the mortar. The loss of connection between stone materials has significantly contributed to the formation of cross cracks and subsequent collapses, especially in specific wall and junction points. In particular, this situation demonstrates that the void spaces created as a result of gaps observed in the vertical joints of the walls reduce the out-of-plane bending stiffness of the walls during earthquakes, thereby increasing the extent of damage. The findings reveal that inadequate maintenance over the years, leading to improper material connections or material degradation, makes structural elements more vulnerable to horizontal loads induced by earthquakes. Especially in the structure, the void spaces created as a result of gaps observed in the vertical joints of the walls during earthquakes have reduced the out-of-plane bending stiffness of the walls, resulting in increased damage. Characteristic diagonal cracks formed in the walls before a decrease in lateral resistance during the earthquake. It is believed that the water flowing from the gutters before the earthquake reached the wall surface, causing the mortar to soften and lose its binding properties, and that the area collapsed due to the shaking caused by the earthquake.

The Diyarbakır City Walls, including Towers 41 and 42 and the walls between them, as well as the overall structure, were constructed using square and rectangular cut stones (basalt) of different sizes in a technique known as “ashlar masonry” with “continuous joints.” However, field studies have revealed frequent instances of faulty repair practices, particularly in recent repairs. In these applications, the natural joints between the stones are not considered sufficient, leading to the carving of the edges of stone blocks and the creation of continuous joints protruding from the wall surface. These described faulty jointing practices should not be applied in any cultural heritage sites, especially archaeological structures like the Diyarbakır City Walls, where the preservation value is crucial.

The results have revealed that the mortar samples are produced using slaked lime as the binder. The aggregates in the samples consist of varying proportions of dolomitic stone fragments and dust, basaltic sand, and black ash particles. The volcanic origin of the black ash dust contributes to the mineralogical composition and color characteristics of the mortar [49]. Carbonated lime nodules were detected in the paste structure of the samples. These nodules are carbonate minerals formed during lime hydration process, aiming to enhance the durability of the mortar [50]. According to the SEM analysis results, a higher amount of calcium and silicon was found in the mortar samples compared to other elements. The amount of calcium detected in the binder of these samples is significantly higher than silicon, further confirming the use of crushed brick fragments as aggregates [51]. Additionally, clay-sized material dominates the composition of the mortar. Clay can influence properties such as workability, water absorption, and durability of the mortar. The presence and content of clay minerals in the mortar mineralogy pose a significant risk to the integrity of the mortar. Clay minerals contribute to the formation and expansion of cracks due to their swelling properties that can generate pressure in a wet environment. Moreover, fluctuations in moisture on a daily and seasonal basis can lead to an increase or decrease in clay volume. While clay minerals may not behave destructively due to swelling, they can cause deterioration due to osmotic swelling during wet and dry cycles [52].

It is evident that the repairs carried out on the structure in later periods were made by adding a certain amount of Portland cement to the aforementioned mixture. In spot anion tests conducted on all mortar samples, a significant amount of chloride (Cl) salt-induced efflorescence was detected. Dissolved salts are absorbed from the mortar surfaces and accumulate within the structure, particularly in the pores. These salts have been observed to reduce the mechanical strength of the mortar and lead to granular disintegration in the stone due to acidic environments and the crystallization pressure of the salt [53]. All these results indicate the role of Cl salt in the occurrence of occasional melting and depletion in the mortar joints between the masonry observed in the field. Observational findings also revealed the use of cement mortar to fill the empty joint gaps in the walls. Additionally, the spot test results showed a significant presence of Clˉ ions originating from Portland cement. In areas previously repaired with Portland or similar cementitious mortar-plaster-fill materials, condensation occurs during rainless periods, and intense stresses arise due to both condensation and thermal expansion differences during periods with significant temperature changes (day-night). Portland cement-based repair materials, in addition to these physical effects, transfer soluble salts present in their composition to the building material, leading to efflorescence. The low vapor permeability of Portland cement (i.e., its ability to hinder moisture movement in the vapor phase) causes the accumulation of salts present in the water as a result of the wetting–drying cycle, leading to material deterioration [54].

In addition, calcination analysis revealed a relative moisture content in the mortar samples ranging from 1.05 to 7.80%. The observation of occasional erosion and depletion in the mortar joints between the stone masonry on the identified facades, even visible to the naked eye, provides qualitative evidence that the mortar absorbs excessive amounts of water from environmental and climatic sources, leading to increased deterioration, as observed in many previous studies [1,2,3,4,5,6,7].

In the calcination analysis of the mortar, it was observed that the total carbonate content of the acid-treated mortar samples ranged from 29.72 to 69.79%. These percentages indicate a significantly high carbonate content [55]. The high CaCO3 content can be attributed to the use of slaked lime as the binder in conjunction with basaltic sand as the filler in these samples [56, 57]. Additionally, the binder/aggregate (B/A) ratio of the mortar samples was found to be 1:2.5–3 (Table 3). These data indicate the compatibility of the mortar joints and samples' lime/carbonate-rich binder content with traditional practices, which typically exhibit binder: aggregate ratios of 1:2 and 1:3 [47, 58].

Conclusions

Within the scope of the study, structural damage patterns affecting the Diyarbakır City Walls caused by the 2023 earthquakes centered in Kahramanmaraş, Turkey were identified, weaknesses and factors leading to the damage were determined through various analyses, and preventive measures were emphasized. Based on the observational and technical analyses conducted, it was determined that the most significant damage in the structure resulted from out-of-plane deformation. It was observed that the majority of the affected areas lacked proper mortar connections. The factors that exacerbated and contributed to the damage during the earthquake in the affected areas of the Diyarbakır City Walls can be summarized as follows:

• I. Internal Factors (Mortar): The technical analysis revealed various reasons for the issue of voids in the mortar joints. The analysis determined that the mortar samples contained dolomitic stone fragments and powders, along with added basaltic stone dust and lime binder. The loss in the mortar is due to the presence of particles prone to clay formation in the petrographic structure of the mortar, resulting in weak physical and mechanical properties during usage. The presence of clay minerals in the mineralogy of the mortar has contributed to volumetric expansion or contraction of clay and the formation and spread of cracks during wet and dry periods, depending on significant temperature variations in the region and daily and seasonal temperature changes.

• II. Climatic Factors: The geographical region of Diyarbakır, where the walls are located, is influenced by several geoenvironmental conditions, such as temperature fluctuations, strong winds, and humidity. The monument is situated in an environment characterized by hot and arid summers, while experiencing the effects of rain during cold seasons and significant temperature fluctuations throughout the year. Under these climatic conditions, the walls have experienced various internal stresses caused by daily, monthly, and yearly temperature fluctuations, resulting in the fragmentation of the mortar. Under the climate conditions in the region, walls have been found to undergo mortar fragmentation due to internal stresses caused by daily, monthly, and yearly temperature fluctuations.

• III. Traditional Construction Technique (Wall): The traditional construction technique used in the inner and outer layers of the double-layered walls, involving the use of different types of materials, has led to inadequate adhesion during earthquakes and increased damage.

• IV. Faulty Repairs: Subsequent repairs carried out on the structure after its construction appear to have been made by incorporating a certain amount of Portland cement into the mixture. The use of Portland cement in repairs has led to salinity issues in the mortar material. Another reason for the occurrence of voids in the mortar joints that lead to post-earthquake damage in the Diyarbakır City Walls is the high level of salinity in the mortar. The soluble salts present in the water content of the mortar cause a slight dissolution of the mortar during wet-dry cycles, exacerbating the major damage that occurs after earthquakes.

Another finding that contributes to the increased damage during earthquakes in Diyarbakır walls is the improper repair practices observed in recent repairs. This includes the filling of natural stone joints with continuous and uninterrupted mortar joints by carving the edges of stone blocks and creating high-level protruding joints from the wall surface. The use of incorrect joint techniques in mortar repairs has resulted in increased mortar losses.

• V. Neglect: The insufficient protection and incomplete repairs of the structure before earthquakes have led to significant mortar losses and an increase in salinity issues in certain sections of the city walls. The mortar losses and soluble salts present in the mortar have exacerbated the damage during earthquakes by causing slight dissolution of the mortar during wet-dry periods.

The proportional distribution of factors causing damage to the city walls of Diyarbakır is as follows (Fig. 6).

Fig. 6

The proportional distribution of factors causing damage

To repair the damage caused by these factors in the structure and to prevent the recurrence of similar issues in future earthquakes, based on the observational and experimental analysis results conducted in our study, the following practices are recommended:

• Although it is not possible to eliminate environmental factors, the most significant problems causing damage can be mitigated if water ingress into the mortar's composition is prevented. The presence of clay minerals in the mortar's mineralogy, coupled with the significant temperature variations in the region, leads to volumetric expansion or contraction of the clay due to the daily and seasonal temperature changes and contributes to the formation and propagation of cracks during wet and dry cycles. Therefore, the osmotic swelling capacity of the clay in the mortar should be considered during intervention and restoration stages.

• Salts should be removed during cleaning processes involving wetting, covering with paper towels, and brushing to eliminate lichens and other deposits. In the observational and experimental analyses of the structure, it is expected to find higher salt content, particularly chloride (Cl⁻) ions, in areas where repairs have been carried out using Portland cement. Hence, after normal cleaning, the process of removing soluble salts should be repeated using a paper pulp compress [59,60,61]. Based on the analysis results, the recommended mortar mixture for use in the structure would be:

  • 1 part Hydrated Lime

  • 2 parts 5 mm sieve undersize light-colored basaltic sand

  • 1 part light-colored active pozzolan (acidic-andesitic tuff powder)

This mixture is considered suitable for the mortar composition based on the analysis findings and the desired properties for the preservation and performance of the Diyarbakır City Walls. The reason for recommending these ratios is that the original composition of the mortar in the structure typically has a binder-to-aggregate ratio of around 1:2.5–3. This ratio can significantly impact the mechanical strength and performance of the mortar. Higher binder-to-aggregate ratios generally provide higher durability. Factors such as aggregate composition and the presence of carbonated lime nodules contribute to the physical strength of the mortar and enhance its resistance to cracking. Hydrated lime is chemically resistant, which helps ensure that the mortar is resistant to various chemical effects. The binder-to-aggregate ratio and the composition of aggregates have an impact on the mechanical strength of the mortar. Higher binder-to-aggregate ratios and appropriate aggregates enhance the compressive strength and overall stability of the mortar.

• Improper jointing practices should never be applied to heritage structures, especially archaeological sites like Diyarbakır Walls. In repairs, the “adjacent joint” technique should be used in accordance with the minimum intervention principle, which is suitable for the original fabric. Existing gaps between the adjacent natural stone blocks of the structure, where manual application is possible, should be filled without extending the joint material onto the surface but leaving it behind at the wall level to prevent water trapping. The recommended joint mortar should preserve the documentary value, the original fabric, the aesthetic perception, and the authentic character of the structure. The joint mortar for the parts where joints have formed due to stone losses in the “adjacent joint” technique should be as follows. It is recommended to use the following mortar mixture for traditional jointing:

  • 1 part slaked lime

  • 2 parts fine-grained light-colored basaltic sand (sieved below 2 mm)

  • 1/2 part light-colored puzzolan (acidic andesitic tuff powder)

• The correct jointing technique for “adjacent joint” stone masonry walls should be done as follows: the boundaries of the joints should be set back 5 to 10 mm from the stone surfaces on the facade, and the existing gaps between the stones should be filled. It is important to strictly avoid any damage to the stone material or alterations to the stone edges during the jointing process.