INTRODUCTION
As many studies have by now pointed out (Folk and Valastro Reference Folk and Valastro1976; Hayen et al. Reference Hayen, Van Strydonck, Fontaine, Boudin, Lindroos, Heinemeier and Caroselli2017; Lindroos et al. Reference Lindroos, Ringbom, Heinemeier, Hajdas and Olsen2020), radiocarbon (14C) dating of aerial mortars or plasters may present many issues, which are ascribable to the vast heterogeneity of this kind of material. Despite the relatively simple dating principle (the anthropogenic calcite CaCO3 resulting from the hardening process of the slacked lime can be collected and dated), the possible sources of contamination can be many, such as:
-
aggregates typically added to the slacked lime can contain carbon and can be so fine-grained that their complete separation from the binder can be really challenging;
-
some residues from the unburnt original CaCO3 can be present;
-
dissolution and recrystallization processes can have altered the original calcite content;
-
carbonation can have proceeded over periods longer than typical 14C uncertainties.
A complete characterization of the mortar before 14C dating is thus mandatory to identify the type of mortar and then an efficient procedure that allows us to extract the only carbon deriving from the moment in which the mortar hardens (Toffolo et al. Reference Toffolo, Ricci, Chapoulie, Caneve and Kaplan-Ashiri2020; Cantisani et al. Reference Cantisani, Calandra, Barone, Caciagli, Fedi, Garzonio, Liccioli, Salvadori, Salvatici and Vettori2021; Daugbjerg et al. Reference Daugbjerg, Lindroos, Heinemeier, Ringbom, Barrett, Michalska and Olsen2021). For instance, petrographic analysis gives us information about each component of the mortar (binder, aggregates, and inorganic additives) and about the identification of raw materials and technologies used for their production (Franzini et al. Reference Franzini, Leoni and Lezzerini2000; Riccardi et al. Reference Riccardi, Lezzerini, Carò, Franzini and Messiga2007; Rampazzi et al. Reference Rampazzi, Colombini, Conti, Corti, Lluveras-Tenorio, Sansonetti and Zanaboni2016; Cantisani et al. Reference Cantisani, Fratini and Pecchioni2022). Spectroscopic methods can support us to discriminate between the anthropogenic calcite to be dated and the geogenic calcite (Toffolo et al. Reference Toffolo, Regev, Dubernet, Lefrais and Boaretto2019; Calandra et al. Reference Calandra, Cantisani, Salvadori, Barone, Liccioli, Fedi and Garzonio2022). Other experimental methods focus on the possibility to separate the two components thanks to their different behavior as a consequence of increasing temperature (Barrett et al. Reference Barrett, Keaveney, Lindroos, Donnelly, Schrøder Daugbjergd, Ringbome, Olsen and Reimer2021) or acidification (Lindroos et al. Reference Lindroos, Heinemeier, Ringbom, Braskén and Sveinbjörnsdóttir2017). A possible advantage is represented by the presence of lime lumps, which—when formed during carbonation—can be considered as pure binder and can thus be isolated from the rest of the matrix (Al-Bashaireh Reference Al-Bashaireh2013; Addis et al. Reference Addis, Secco, Marzaioli, Artioli, Arnau, Passariello and Brogiolo2019; Michalska Reference Michalska2019; Sironić Reference Sironić, Borković, Barešić, Bronić, Cherkinsky, Kitanovska and Čukovska2019).
However, mortars and plasters can also contain organic carbon that, in principle, can be collected and dated. Plant remains, such as charcoals and straw fragments, might be actually found among the aggregates as mixed with sand and/or powdered rocks. For example, according to a technique widely used in antiquity (Artioli et al. Reference Artioli, Secco and Addis2019; Maravelaki et al. Reference Maravelaki, Theologitis, Budak Unaler, Kapridaki, Kapetanaki and Wright2021), straw fragments were added to the mortar mixture as inert to improve the mechanical properties of the material, such as its flexibility and lightness.
In particular, straw fragments can be good candidates for 14C dating, since we can expect that the annual plants they originated from had been cut shortly before their use in the mortar mixture. Thanks to this, the straw 14C concentration can be associated to the moment in which the mortar was set. Moreover, such a material is not likely to be affected by the oldwood problem, which, on the contrary, can influence charcoals (Schiffer Reference Schiffer1986; Regev et al. Reference Regev, Eckmeier, Mintz, Weiner and Boaretto2011). The isolation and the following processing of such a material can be nevertheless particularly challenging, especially because of the extremely low masses that one may expect to collect. Anyway, organic remains can really represent the only possibility to date a mortar or a plaster in case any inorganic material is completely unsuitable for the measurement, according to what was mentioned above.
Here we discuss the case of the St. Philip Church (Figure 1), located in the archaeological site of Hierapolis of Phrygia, Denizli, Turkey. Hierapolis is one of the great Hellenistic, Roman and Byzantine cities of southwestern Turkey, protected by UNESCO since 1988. It was founded in the 3rd century BCE, and it has survived for millennia (D’Andria Reference D’Andria2016–2017; D’Andria Reference D’Andria2017), even though the Denizli area has been always characterized by dangerous earthquakes and hydrothermal fluid circulation that clearly can affect the conservation of monuments. In the second half of the 6th century CE, a threenave church was built on the eastern hill of Hierapolis, in the area where the tomb of the Apostle Philip has been recently recognized by D’Andria. The church underwent important transformations during the 9th century CE. Then, in the 10th century CE, a catastrophic earthquake seriously damaged the building, which was no longer rebuilt but continued to be used for residential purposes during the Seljuk empire until the 14th century (Caggia Reference Caggia2016, Reference Caggia2022).
Figure 1 Aerial photo of the St. Philip Church.
The reconstruction of the chronology of the site is not straightforward. Indeed, the presence of a faults system, which is still active today and which has caused—as mentioned—many earthquakes, have contributed to several subsequent collapses of walls and architectural elements over centuries, making the identification of the different archaeological phases of the construction difficult. In such a context, the possibility of dating the construction materials, i.e., mortars and plasters, would be thus important. Anyway, the environment is not favorable to mortar 14C dating. The main critical aspects are the following:
-
main raw materials used to produce mortars, also added as aggregates, were travertine and marble provided from local quarries (Caggia Reference Caggia2018; De Giorgi Reference De Giorgi2018);
-
thermal waters, characterized by high concentration of dissolved CO2, have been responsible for dissolution and corrosion processes and precipitation of various carbonate phases (Vettori et al. Reference Vettori, Cabassi, Cantisani and Riminesi2019a), which may have polluted the mortar mixture with geogenic carbon, thus apparently aging the mortar (Ricci et al. Reference Ricci, Secco, Artioli, Marzaioli, Passariello and Valluzzi2020).
As an alternative to mortar/plaster dating, in the following we will present the 14C measurements on straw found in some of the plaster samples from the church.
MATERIALS AND METHODS
Analyzed Samples
Among all the studied materials from the site, seven samples from plasters of St. Philip Church were selected for the present study because they were characterized by the presence of tiny straw fragments in the matrix (see Figure 2 and left-hand side of Table 1): MSF04 and MSF15 were collected from pillars in the northern nave; MSF06 from west wall of northern nave; MSF33 from west wall of the southern nave; US543, US546 and US547 from west wall of the narthex. MSF-labeled samples are related to structures attributed to the early Byzantine period (second half of the 6th century), even though we are not certain that the plaster revetments belong to the original phase. Instead, the narthex can be dated to the middle Byzantine phase (since 9th century CE onwards).
Figure 2 Plan of the church with localization of the plaster fragments containing straw traces analyzed in this paper.
Table 1 Samples from the plasters of the St. Philip Church where straw fragments were identified: location, XRD data and petrographic characteristics (PLM observation) are reported. Minerals are indicated according to (Whitney and Evans Reference Whitney and Evans2010). In addition to calcite in all samples, considering the aggregate composition, the prevailing presence of quartz and plagioclase and minor amounts of phyllosilicates (mica-like minerals and chlorites) are evidenced. Gypsum is due to alteration phenomena caused by the sulphation of carbonate binder.
* Cal = calcite; gp = gypsum; qz = quartz; arg = aragonite; ms = muscovite; chl = chlorite; pl = plagioclase.
Plasters Characterization and Collection of Samples for Radiocarbon Dating
The plaster samples from the St. Philip Church in Hierapolis were characterized at the Institute of Heritage Science – National Research Council of Italy, Florence (ISPC-CNR). Methods and results have been already published elsewhere (Cantisani et al. Reference Cantisani, Vettori, Ismaelli and Scardozzi2016; Vettori et al. Reference Vettori, Bracci, Cantisani, Conti, Ricci and Caggia2019b) and here only that information that are related to the present discussion is briefly recalled.
The samples were first observed under a stereomicroscope, with a dedicated camera, and then petrographic analyses on thin sections (thickness 30 µm) were performed with a polarized light microscope (PLM). The mineralogical composition of the bulk samples was studied through X-ray powder diffractometry (XRPD) using a powder X-ray diffractometer (Cu anode, Kα line 1.54 Å).
While under optical investigation, straw fragments were identified as aggregates in the plaster matrices. The mechanical separation of these fragments was performed using a stereomicroscope (Figure 3a), as mentioned above. The separation of the fragments was highly difficult: the bulk samples, which the straw fragments were extracted from, were of the order of 2–3 cm long, while the fibers were visible almost exclusively under the microscope (Figure 3b). The collected materials had extremely thin stems and a length of about 1–2 mm. The presence of some mortar granules still in the separated samples makes estimating the straw mass obtained by the mechanical separation process difficult.
Figure 3 The mechanical separation of straw fragments: (a) using the stereomicroscope, (b) selecting straw from the bulk sample. Straw fragments are highlighted using dotted line red circles. (Please see online version for color figures.)
Radiocarbon Measurements: Sample Preparation and AMS
14C-AMS measurements were performed at LABEC in Florence, one of the laboratories of CHNet, the INFN (Istituto Nazionale di Fisica Nucleare) network dedicated to Cultural Heritage.
The first step in sample preparation was clearly focused on pretreatment to remove possible contaminants. In the case of our straw fragments, the most important source of possible exogenous carbon was represented by carbonates coming from the plaster matrix. In addition, considering that we could not have any information about whether the straws might have come into contact with some organic contaminations before being embedded into the plaster, we decided not to completely neglect the possibility of an interference due to such substances. We thus applied an ABA (acid-base-acid) pretreatment, choosing temperature and duration of each of the steps as a balance between the expected cleaning efficiency and the mass preservation of the samples. Straw samples were treated as follows:
-
1 hour bath in 1M HCl at room temperature (instead of at 80°C as in our typical procedure), to remove possible carbonate residues from the fragments;
-
a quick bath in 0.1M NaOH at room temperature, for up to few minutes, i.e., until the complete “whitening” of the fragments. Both lignin from the straw and possible humic acids contaminants are brownish or dark colored and are soluble in alkaline solutions. Thus, we expect that a lighter color of samples can be associated to a good efficiency of the treatment, suggesting us that we removed the possible organic contaminations;
-
1 hour bath in 1M HCl at room temperature (instead of, again, at 80°C as in our typical procedure).
Since the mass of the straw fragments after the pretreatment was clearly smaller than the initial collected mass, in the order of about a few hundreds of micrograms, graphitization was carried out using our so-called Lilliput graphitization line, specifically designed for microgram-sized samples (Fedi et al. Reference Fedi, Barone, Barile, Liccioli, Manetti and Schiavulli2020a, Reference Fedi, Barone, Carraresi, Dominici and Liccioli2020b). In particular, this line is optimized for samples sizes corresponding to about 50 micrograms of carbon. Since the individual masses of the recovered straw samples were actually below the aforementioned limit, we decided to combine some of them together following archaeological evidence, such as chronological and functional criteria. Table 2 indicates how the samples were combined, their total masses before combustion and the extracted CO2. As one can notice, the collected CO2 from MSF04+MSF15 is lower than the carbon dioxide obtained from the combustion of the other samples, the masses being practically equal: this suggests that MSF04+MSF15 might have some components different from the original material, like for instance a possible residual contamination.
Table 2 Straw fragments samples as they were combined after pretreatment: total masses and collected CO2 pressures are shown.
14C AMS measurements were performed at the AMS beam line of the 3MV Tandem accelerator installed at LABEC (Fedi et al. Reference Fedi, Cartocci, Manetti, Taccetti and Mandò2007; Chiari et al. Reference Chiari, Barone, Bombini, Calzolai, Carraresi, Castelli and Mandò2021). Routine procedures provide for both 14C/12C and 13C/12C isotopic ratio measurements during the AMS run. 14C/12C ratios of the unknown samples are corrected for background counts, as evaluated measuring blank samples, and for isotopic fractionation (thanks to measured 13C/12C ratios). Isotopic ratios are also normalized to the ratios measured during the same run in a set of standard samples with a certified 14C concentration (NIST 4990C, Oxalic Acid II). Samples produced from other standard reference material (IAEA C-7) are also measured as a check of accuracy. Normalized and corrected 14C concentrations are expressed in pMC, according to Stenström et al. (Reference Stenström, Skog, Georgiadou, Genberg and Johansson2011). To convert conventional 14C ages to calibrated ages using the software OxCal v 4.4.4 (Bronk Ramsey Reference Bronk Ramsey2009; Bronk Ramsey and Lee Reference Bronk Ramsey and Lee2013), the IntCal20 curve (Reimer et al. Reference Reimer, Austin, Bard, Bayliss, Blackwell, Bronk Ramsey, Butzin, Cheng, Edwards, Friedrich, Grootes, Guilderson, Hajdas, Heaton, Hogg, Hughen, Kromer, Manning, Muscheler, Palmer, Pearson, van der Plicht, Reimer, Richards, Scott, Southon, Turney, Wacker, Adolphi, Büntgen, Capano, Fahrni, Fogtmann-Schulz, Friedrich, Köhler, Kudsk, Miyake, Olsen, Reinig, Sakamoto, Sookdeo and Talamo2020) is employed.
RESULTS AND DISCUSSION
As mentioned above, the complete characterization of the mortar and plaster samples from the St. Philip Church has been already reported in Cantisani et al. (Reference Cantisani, Vettori, Ismaelli and Scardozzi2016). Three different categories of mortars/plasters were identified: (1) air hardening calcic lime with a very abundant amount of carbonate and silicate aggregates (see Figure 4a), (2) air hardening calcic lime with crushed ceramic fragments (cocciopesto) (see Figure 4b), and (3) air hardening calcic lime with straw and very low amount of carbonate aggregates (see Figure 4c).
Figure 4 Microphotographs of thin sections of representative types of mortar and plaster in the St. Philip church: (a) mortar with air hardening calcic lime (lime) with a very abundant amount of aggregate (aggr); (b) mortars with air hardening calcic lime with fragments of crushed ceramic (cc); (c) plasters produced by lime with straw and very low amount of aggregate. (by PLM: (a), (b) crossed nicols, (c) parallel nicols).
Table 1 (on the right-hand side) summarizes the analytical data acquired for those plaster samples belonging to this latter type of material, which were selected in this context because of the presence of sufficient straw fragments (Figure 5). Samples MSF06, MSF33, US546 and US547 are characterized by medium/high porosity and by a binder/aggregate ratio equal to 3/14/1 with a very low amount of aggregates, made of fragments of rocks (schists, quartzite, travertines, and marbles). Aragonite was also found (samples MSF06 and US546): in fact, aragonite was usually found in mortars associated with travertine, present here as aggregate. Samples MSF04, MSF15, and US543 were produced without aggregates and are characterized by very high porosity with shrinkage microcracks and elongated pores due to the loss of straw. The binder is typically not homogeneous and numerous lumps can be recognized, except for US543 that presents homogeneous binder and absence of lumps. The lumps are unburnt fragments and suggest the use of marble as the original stone to produce lime (this was indeed typical of ancient production technologies). As far as 14C dating is concerned, geogenic carbonate aggregates and unburnt fragments of marble are “extremely dangerous” dead carbon bearers. In addition, the medium/high binder porosity indicates the presence of secondary calcite, which is likely to have been caused by dissolution and recrystallization processes, due to rain and hydrothermal percolating waters.
Figure 5 Straw fragments: (a) macroscopic aspect, (b) stereomicroscope image, and (c) microphotograph of thin section of plaster with straw fragments (by PLM: crossed nicols).
All things considered, as already anticipated, the reported data show that these plasters were not suitable for 14C dating and therefore all the attention focused on the recovered straw fragments. Table 3 shows the results of the AMS measurements of the straw samples.
Table 3 Measured 14C concentrations, corresponding conventional 14C ages, and calibrated ages time intervals (when more than one possible range are identified, time intervals are reported considering the lower and the upper extreme, respectively).
The sample MSF04 + MSF15 appears to be older than expected and in fact the calibrated age does not match the archaeological evidence. This, together with the low CO2 collection yield after combustion, suggest that the sample could be still contaminated by residual “old” CaCO3 from the plaster. Probably, some marble traces have not been fully removed during the pretreatment. In addition, complete combustion of marble by elemental analyzer is not easily achievable, thus the possible presence of CaCO3 residues can explain the low carbon dioxide yield.
MSF06 + MSF33 and US543 + US546 + US547 samples are consistent with the expected age based on archaeological evidence (see also Figure 6). MSF06 + MSF33 sample appears to be slightly older than the other dated sample. In fact, considering that the church was built and later modified as the consequence of several earthquakes, two different construction phases can be expected at least in accordance with the time range during which the church was used. However, unfortunately, the experimental uncertainties on the measured 14C concentrations do not allow us to statistically discriminate between them. After all, the higher uncertainties that were achieved in this case are consequence of the extremely small samples that were prepared and measured: in our set-up, smaller masses correspond to lower ion currents extracted and measured into the accelerator and thus to lower count rates.
Figure 6 Calibration graphs of samples MSF06 + MSF33 and US543 + US546 + US547.
CONCLUSIONS
14C dating of mortar is very challenging. Depending on the context, the application of 14C is not feasible, since many factors can interfere with the dating. This is just the case of the St. Philip Church, in the archaeological site of Hierapolis. Mortars and plasters are highly heterogeneous as for the binder composition and the aggregates. Moreover, the presence of carbonate aggregates, secondary calcite and hydraulic binder suggest that direct 14C dating is not ideal. In this situation, the presence of straw fragments as organic inclusions drew our attention: straw fragments represent indeed a very good candidate for 14C dating, since their 14C concentration can be strictly linked to the moment in which the mortar was set.
Here, we have shown that even such small quantities of straw fragments can be identified and isolated from the matrix. Due to the extremely small masses that can be collected and recovered after the mandatory pretreatment, the measurement of the 14C concentration is only possible using a set-up which is specifically optimized for microgram-sized samples. Obtained results of MSF06 + MSF33 and US543 + US546 + US547 are satisfying, since they are consistent at large with the expected archaeological dates. Nevertheless, our data also underline the possible limitation of using this kind of material: a milder pretreatment as the one we chose to preserve as much mass as possible could not be sufficient to remove all the possible contaminants.
Dating straw and, in general, organic inclusions in mortars and plasters can thus represent a valid alternative to the direct dating of the carbonate material, even though the critical aspects connected to sample preparation may suggest that this cannot be applied in routine operations.
Open access
