Archaeometric Characterization of the Ferrería Style Pottery (El Carmen de Viboral, Colombia): Raw Materials and Firing Conditions

RESEARCH ARTICLE © ARQUEOLOGÍA IBEROAMERICANA 56 (2025): 207-220. ISSN 1989-4104. <https://www.laiesken.net/arqueologia/>. ARCHAEOMETRIC CHARACTERIZATION OF THE FERRERÍA STYLE POTTERY (EL CARMEN DE VIBORAL, COLOMBIA): RAW MATERIALS AND FIRING CONDITIONS Elizabeth Jiménez,1 Luis Carlos Sánchez,2* Oscar Arnache 1 1 Grupo de Estado Sólido, Instituto de Física, Universidad de Antioquia, Medellín, Colombia 2 Grupo Avanzado de Materiales y Sistemas Complejos (GAMASCO), Universidad de Córdoba, Montería, Colombia * [email protected] (corresponding author) Received: October 29, 2025. Accepted: December 3, 2025. Published: December 16, 2025. Edited & Published by Pascual Izquierdo [P. I. Egea]. Arqueol. Iberoam. Open Access Journal. Creative Commons License (CC BY 4.0). https://n2t.net/ark:49934/400. https://purl.org/aia/5623. Figure 1. Schematic map of the excavation sites found in the Betania rural area of El Carmen de Viboral, Colombia. The sites where excavations were conducted are marked in red circles (https://geoportal.dane.gov.co/). ABSTRACT . A multi-analytical study was conducted on archaeological pottery samples from El Carmen de Viboral, Antioquia, Colombia (5th century BCE–3rd century CE), belonging to the Ferrería style, which is associated with the region’s pre-Hispanic agricultural societies. X-ray diffraction (XRD), scanning electron microscopy (SEM) with energydispersive spectroscopy (EDS), Fourier-transform infrared spectroscopy (FTIR), thermogravimetric analysis (TGA-DTG), Vickers microhardness measurements, and Mössbauer spectroscopy were employed to determine the chemical composition, morphological, and mineralogical characteristics of the pottery sherds. The results indicate that the pottery was primarily made from non-calcareous clays. The presence of mineral phases, such as silicates, is characteristic of a low-temperature firing process. Additionally, the identified silicate phases contain iron in both the divalent (+2) and trivalent (+3) oxidation states, suggesting variable redox conditions during the manufacturing process. KEYWORDS . Ferrería-style pottery, 57Fe Mössbauer spectrometry, XRD, SEM-EDS, FTIR, Colombia. – 208 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 1. INTRODUCTION Pottery is one of humanity’s earliest and most significant technological innovations, a common find in many archaeological contexts. For this reason, it provides a crucial source of empirical data for archaeologists. Studies of variation in pottery production, style, and use have been instrumental in constructing chronologies and interpreting ancient societies. In recent decades, there have been increasing efforts to establish pottery as a primary component in the study of social and cultural groups (Hunt 2017). Traditionally, archaeological work has focused on the search for and classification of pottery artifacts, which has led to the definition of archaeological potential based on the presence or absence of pottery remains (Botero 2001). In Antioquia, Colombia, different pottery styles have been classified based on their style and form, though these have sometimes been categorized without clear distinctions between a “pottery tradition” and a “period of human group occupation” (Obregón 1999). Beyond fulfilling basic needs, the objects produced by potters served as aesthetic, symbolic, and political representations, materialized through patterns recurrent in specific societies. Consequently, pottery production is considered a vital indicator for differentiating human groups. In the eastern region of Antioquia and the Aburrá Valley, archaeologists have identified various cultural phases. These phases are attributed to culturally distinct groups based on their spatial distribution, stratigraphic relationships, absolute dates, and stylistic analyses. Among the human groups identified through their pottery production, the Ferrería-style pottery stands out as one of the oldest known in the region. This style was first identified in the Aburrá Valley at the archaeological site of La Ferrería (Castillo 1995), where distinctive features regarding its decoration techniques, raw material use, and forms were recognized. Furthermore, using absolute dating techniques, the occupation period was established to span from the 5th century BCE to the 3rd century CE. Despite these findings, our understanding of these societies remains limited. Archaeologists currently rely on relative dating methods based on Castillo’s descriptions to associate new discoveries with specific human groups. This is primarily done by analyzing stylistic characteristics such as color, shape, and decoration, as well as stratigraphic relationships. However, this classification method presents significant challenges in defining distinct styles associated with different groups and periods. These challenges arise because 1) broad chronological overlaps have been established for various groups, making it difficult to differentiate them, and 2) thorough studies have not been conducted to determine sudden changes in production technologies. In some cases, the forms and decorations appear to have remained unchanged over time (Botero et al. 2017). To address these limitations, archaeologists in Colombia often use analytical techniques like petrography, granulometry, and hardness testing as a complement to stylistic analyses when making new discoveries. In Colombia, archaeometric research on ancient pottery and metalwork is still in its early stages. There isn’t yet a formal research area dedicated to understanding the microstructural, chemical, and physical characteristics of ancient pottery or other cultural heritage artifacts. Understanding the nature of pottery, including its raw materials and manufacturing processes, can expand our perspective on what is already known about it. Systematizing its characteristics and properties can lead to significant progress in our knowledge of past societies and their manual, technical, and technological developments. To achieve this, it is necessary to perform chemical, physical, and morphological characterization of ancient pottery pieces, specifically of each pottery style, which allows us to delve into the raw materials used for production processes, manufacturing techniques, and technological processes. This includes the conditions of firing, such as temperatures and atmospheres (oxidizing or reducing) of the kilns. This approach can clearly establish differences and transformations between pottery styles. Such knowledge is frequently of interest from the point of view of archaeology or anthropology, as it enables an assessment of the technical skills of ancient potters and, therefore, the cultural achievements of ancient civilizations. As a first step toward understanding the physicochemical properties of ancient pottery from this region, this study presents a new contribution. It provides a detailed chemical and mineralogical characterization of Ferrería-style pottery found in El Carmen de Viboral, Antioquia. The Ferrería style is particularly notable for its ambiguities in form and style, including color and decoration, which it shares with other styles. Furthermore, there is a clear lack of quantitative data and detailed analyses on this pottery, not only in El Carmen de Viboral but throughout the entire region. – 209 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 2. ARCHAEOLOGICAL CONTEXT AND SITE El Carmen de Viboral is located in the Central Andes, in the eastern part of Antioquia, Colombia (6° 4’ 55’’ N, 75° 20’ 3’’ W). The Cimarrona stream is the main water source that flows through the urban area. The rural area of Betania, with an approximate surface area of 0.342 km2, is located on the stream’s right bank, therefore, east of the municipality’s urban center (see Figure 1). The archaeological excavation in Betania was conducted in 2014 under the direction of the archaeologist Gustavo Santos (Santos 2014), covering almost the entire village area. Considering the stylistic and technological features already established for various pottery styles in the Aburrá Valley, the Rionegro Plateau, and the slopes of the Magdalena River in the Central Andes, a historical-cultural classification of the found pieces was carried out. The findings were divided into five main groups with established chronologies: El Oro (3900–2800 BP), Ferrería (3000–1500 BP), Marrón Inciso (2100–1200 BP), Late Pre-Hispanic (1200–500 BP), and Recent (19th and 20th centuries AD). Currently, the pieces found in Betania are part of a significant group of archaeological findings, demonstrating the extensive pottery production that has been taking place in the municipality for over 3000 years. For this research, only fragments classified under the Ferrería style pottery were considered, given its prevalence throughout the Antioquia region and its limited academic attention. This style lacks sufficient information for a more precise conclusion regarding its primary characteristics. The samples were provided by the Museum of Ceramics of the Cultural Institute of El Carmen de Viboral, with the authorization of the Colombian Institute of Anthropology and History (ICANH). This is the highest authority in the regulation and safeguarding of archaeological heritage in Colombia. The fragments selected for the present study were chosen based on their cultural association, as seen in Figure 2. The sherds were selected according to macroscopic characteristics such as color, grain size, and inclusions; however, none of them have evident decorative typologies. For the Ferrería style, their shades vary from light brown to black with medium to fine grains and inclusions ranging from 15 to 20%. For this study, a total of eight pottery samples were selected, specifically those labeled as F7 from excavation site 32; F8 and F10 from excavation site 54; F9, F12, and F14 from excavation site 28; and F11 and F13 from excavation site 30. The excavation sites were selected due to the predominant presence of Ferrería-style pottery. In contrast, other sites exhibited stratigraphic mixing of archaeological evidence (Santos 2014). 3. MATERIALS AND METHODS For the analytical techniques, all the pieces were drycleaned, and a small piece was extracted from each without causing significant damage or contamination. Between 20 and 100 mg per piece were ground to a fine, homogeneous powder using an agate mortar. Additionally, a small piece of a few square millimeters was embedded in polymer resin, and its transverse surface was polished. The experimental techniques employed in this study were utilized to determine morphological characteristics of Ferrería-style pottery, conduct chemical analyses, and identify crystalline phases formed during the manufacturing process. This includes defining more Figure 2. Ferrería-style pottery sherds taken from the Betania rural area at El Carmen de Viboral, Colombia. – 210 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 precisely the technical features of this pottery style and gaining a better understanding of technological changes. For this purpose, scanning electron microscope (SEM) images using backscattered electron mode were recorded using a JEOL JSM-6490LV at an acceleration voltage of 20 kV. A detector INCAPentaFETx-3 was employed, enabling the acquisition of elemental distribution maps and chemical spectra (EDS). An analysis of the crosssectional area of the pottery pieces was conducted. Small sections of approximately 25 mm2 were cut from them, which were then coated with polymer resin to enhance sample stability. The surface to be analyzed was subsequently polished until the sample was fully exposed. The X-ray diffraction (XRD) measurements for all samples were performed using a Malvern-PANalytical Empyrean 2012 model, equipped with a 3D Pixel detector and a Co radiation source (l = 1.790307 Å, operating at 40 kV and 40 mA). Data were collected over the range 10° < 2q < 80°, with a step size of 0.05° and an accumulation time of 50 s/step. Additionally, the Xpert HighScore Plus (version 3.0) software was used for mineral phase identification. Fourier Transform Infrared Spectroscopy (FTIR) was employed to analyze the structural and chemical transitions of phases. The FTIR spectra were acquired using a Shimadzu IRTracer100 spectrophotometer, covering the spectral range 4000–400 cm–1, with powdered samples of approximately 10 mg. Thermogravimetric analysis (TGA) and its derivative curve (DTG) were carried out to measure the weight loss experienced by the ceramic material as a function of temperature, resulting in a weight loss curve for TGA. A more precise analysis can be achieved by deriving the weight loss curve to determine the specific temperature or temperature range at which weight loss occurs (DTG). The thermogravimetric measurements were conducted using a TA Instruments TGA 5500 system, with powdered samples ranging from 5 to 20 mg, under a nitrogen atmosphere, at a heating rate of 10 °C/min, within a temperature range of 30–1000 °C. Mössbauer spectroscopy was carried out at room temperature (RT) using a Mössbauer spectrometer in transmission geometry with a 50 mCi 57Co/Rh radioactive source moving in constant acceleration. The Mössbauer parameters isomer shift (d), magnetic hyperfine field (Bhf), quadrupole splitting (rEQ), quadrupole shift (2e), and absorption area (SA) in different Fe sites were obtained by fitting the Mössbauer spectra with the Recoil program, and the d-values are relative to the RT a-Fe value. Microhardness studies were conducted on the examined pieces using a Shimadzu Micro Vickers Hardness Tester HMV-G 21 series. The entire procedure followed the guidelines of ASTM C1327–15: Standard Test Method for Vickers Indentation Hardness of Advanced Ceramics. The samples were evaluated with loads of 0.05 kgf (0.49 N), 0.1 kgf (0.98 N), 0.2 kgf (1.96 N), and 0.4 kgf (3.92 N) for sample F9, with a load holding time of 15 seconds for all samples. 4. RESULTS AND DISCUSSIONS 4.1 XRD Analysis The X-ray diffraction (XRD) technique is a valuable tool for obtaining information on the crystalline structures of the minerals present in the pottery samples (Feliu et al. 2004). In some cases, it can also provide insights into the firing temperatures used during the production of pottery artifacts, as the presence or absence of certain minerals can be directly correlated with the temperatures reached during firing (Papachristodoulou et al. 2006). The X-ray diffractograms of the Ferrería-style pottery samples are shown in Figure 3. The mineral abbreviations used in this study follow the conventions established by Whitney and Evans (2010). The identification of crystalline phases in all mineral components was performed using the HighScore Plus software with the Panalytical database. The results showed that quartz was identified as the dominant mineral phase in all samples; however, since it remains stable above 1000 °C, it was not considered for estimating firing temperatures. Clay minerals such as illite/muscovite undergo decomposition processes between 700 and 1000 °C (Iordanidis et al. 2009). Given that muscovite and illite are isostructural, their XRD peaks overlap, making it difficult to distinguish between them using this technique (Kostova et al. 2024). Clay minerals are typically found in pottery fired at low temperatures (Kramar et al. 2012); their presence in most of the analyzed fragments suggests that the firing temperature did not exceed 800 °C (Bayazit et al. 2014). Fragment F7 is the only sample that exhibits a kaolinite peak. When kaolinite is heated to temperatures between 400 and 550 °C, its hydroxyl (–OH) groups break down, resulting in the formation of metakaolinite (Mangueira et al. 2011; Ion et al. 2010). This suggests – 211 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 that fragment F7 was fired at temperatures below 600 °C. Fragments F9 and F12, and to a lesser extent F10, exhibit calcite peaks. The thermal decomposition of calcite begins at approximately 600 °C and is completed at around 800–850 °C (Iordanidis et al. 2009). High-temperature mineral phases such as diopside, gehlenite, and wollastonite typically form at temperatures exceeding 950 °C (Martínez et al. 2018; Papachristodoulou et al. 2006; Iordanidis et al. 2009; Kramar et al. 2012). The absence of newly formed crystalline phases in all XRD results suggests that the pottery did not exceed 800 °C during firing (Daghmehchi et al. 2018). The presence of hematite indicates an oxidizing firing atmosphere (Papachristodoulou et al. 2006; Kramar et al. 2012; Rathossi et al. 2010). Additionally, the coexistence of low-temperature phases (quartz, illite/muscovite, calcite, and kaolinite) and the absence of high-temperature phases (diopside, gehlenite) further confirm that firing temperatures did not exceed 800 °C during the production of this pottery (Ricci et al. 2016; Daghmehchi et al. 2018). Except Figure 3. X-ray diffractograms of all samples. Crystalline mineralogical phases are identified and labeled as Tr (tremolite), Qz (quartz), Kln (kaolinite), Cal (calcite), An (anorthite), Hbl (hornblende), Ilt (illite), Ms (muscovite), and Hem (hematite). for sample F7, where the preservation of kaolinite suggests that its firing temperature remained below 600 °C. 4.2 FTIR Analysis FTIR analyses were conducted as a complementary approach to mineralogical and elemental characterization, aiming to identify certain minerals that could not be detected due to the technical limitations of other analytical methods. The results corroborate the findings obtained through XRD while also confirming the presence of hematite (Fe2O3) and traces of organic phases in the samples. Figure 4 presents the FTIR spectra of powdered samples from the Ferrería-style pottery. The vibrational assignments, listed in Table 1, were determined based on reference literature (Gomathy et al. 2021; Ricci et al. 2016; Chukanov 2013; Daghmehchi 2018; Chukanov & Chervonnyi 2016; Oliveira 2020; Benedetto et al. 2002). These spectra are characteristic of ancient pottery studies. The bands at 3622 cm–1 and 3696–3698 cm–1 – 212 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 correspond to the OH stretching vibrations of kaolinite, attributed to internal OH groups and surface OH stretching vibrations, respectively. The broad band centered at 3431–3445 cm–1 is likely associated with the bending (dH–O–H) vibration of H2O in an amorphous clay phase, which forms when high temperatures lead to the decomposition of the clay structure during firing. All samples exhibit small peaks around 2923 cm–1, Figure 4. FTIR spectra of the Ferrería-style pottery. which can be attributed to C–H asymmetric stretching vibrations of CH2, likely due to the presence of organic matter. Additionally, the bands centered at 2377–2385 cm–1 and 2295–2297 cm–1 correspond to atmospheric CO2, which originates from ambient air and therefore does not require further discussion. The pronounced peak at 1634 cm–1 is attributed to the H–O–H bending vibration, indicating the presence of water molecules. These absorbed water molecules would typically be eliminated during the thermal treatment of clay at temperatures above 100 °C. However, since the pottery samples analyzed by FTIR were not subjected to heat treatment, the presence of these bands is likely associated with moisture absorption from the environment both before and after their excavation. The main asymmetric stretching vibration of CO3 2– in calcite is typically found around 1430 cm–1. In this study, several peaks were observed between 1384 and 1448 cm–1, with samples F9 and F12 exhibiting the most pronounced peaks within this range, confirming the presence of calcite, as also detected in the XRD analysis. All spectra display characteristic quartz peaks, specifically the SiO2 symmetric stretching band at ~787–792 cm–1 and the bending vibration at ~462–472 cm–1. However, these peaks are partially masked by the overlapping strong bands of other mineral phases, such Table 1. FTIR vibrational frequency assignments of the pottery sherds: s (strong), w (weak), vw (very weak), m (medium), sh (shoulder). – 213 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 as feldspars (e.g., anorthite at 469 cm–1) and clay minerals (e.g., muscovite at 469 cm–1). The small peaks observed at ~1085 and ~1110 cm–1 on the shoulder of the SiO2 asymmetric stretching vibration also correspond to quartz. The broad peak centered at 1033 cm–1 and the small peak at 910 cm–1 are characteristic of clay minerals, particularly kaolinite. Notably, sample F7 is the only one exhibiting a peak at 910 cm–1, which can be associated with the presence of kaolinite in this sample. This finding is consistent with the XRD results. This band corresponds to Al–OH vibrations within the octahedral layered structure and begins to disappear as temperature increases, vanishing at approximately 500 °C (Velraj et al. 2015). Additionally, the presence of a peak between 521 and 546 cm–1 indicates the presence of hematite. However, hematite was not detected in most samples through XRD analysis. 4.3 SEM-EDS Analysis The morphological and microstructural study was evaluated through targeted SEM-EDS analyses. The microFigure 5. a) SEM image at 200× magnification of the polished surface of pottery fragment F7 (made with non-calcareous clay due to its low Ca content), as indicated by the red box, showing a compact surface; b) EDS spectrum corresponding to the red box and the cross-section of the fragment; c) X-ray element maps illustrating the elemental distribution in the analyzed area, where the other elements detected by EDS in b) are homogeneously distributed throughout the region. graphs facilitated the examination of textural characteristics and elemental composition using maps and energy dispersive X-ray spectroscopy (EDS). SEM micrographs of selected Ferrería samples (see Figures 5 and 6) revealed a consistent morphology characterized by compact surfaces displaying cracks, a blending of components, and the influence of firing temperature. The presence of cracks is likely associated with the thermal processes these samples underwent. Such morphology suggests that this pottery exhibits notable mechanical strength. The elemental composition of the samples was determined using EDS, a valuable technique for studying the raw materials used in pottery production, as the firing process does not significantly alter elemental composition (Seetha & Velraj 2015). The EDS spectra in Figures 5b and 6b revealed prominent peaks of Si, Al, and Fe in all samples. Additionally, smaller amounts of K (except in F7 and F14), Ca, and Mg (except in F9 and F12) were detected; Na was identified in samples F8, F10, and F14; while Mn was present in F8, F9, and F10. The results are summarized in Table 2. – 214 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 The chemical maps in Figures 5c and 6c provide micrographs with color contrast, visually depicting the concentration of each chemical element. The colors represent the distribution of elements within the samples, with silicon being the most predominant. Most of the elements detected by EDS were homogeneously distributed across all analyzed samples (Table 2). After silicon, aluminum was the second most abundant element in all samples. It is suggested that regions with the highest silicon concentrations (cyan color) correspond to quartz crystals (SiO2) embedded in a matrix predominantly composed of aluminosilicates. Regarding the distribution of chemical elements, the mineralogical relationships identified are consistent with previous studies, which found that Ferrería-style pottery pastes contained abundant quartz, plagioclase-type feldspars, and amphiboles derived from a mafic to intermediate geological source (Weber et al. 2020). Furthermore, the clays used in the production of these pottery can be categorized into two types: calcareous clays, with a Ca concentration greater than 6%, and non-calcareous clays, with a Ca concentration below 6% (Seetha & Velraj 2015; Palanivel & Meyvel 2010). Identifying the type of clay is crucial, as it inFigure 6. a) SEM image at 200× magnification of the polished surface of pottery fragment F12 (made with calcareous clay due to its high Ca content), as indicated by the red box, showing a compact surface; b) EDS spectrum corresponding to the red box and the cross-section of the fragment; c) X-ray element maps illustrating the elemental distribution in the analyzed area, where the other elements detected by EDS in b) are homogeneously distributed throughout the region. fluences the transformations occurring during the firing process. For instance, non-calcareous clays produce a more vitrified ceramic body with high density and low fluid permeability, whereas calcareous clays exhibit a characteristic cellular structure with high porosity, and vitrification remains restricted and controlled up to approximately 1150 °C (Shortland et al. 2009). In the case of Ferrería-style pottery, all analyzed samples were made from non-calcareous clays, except for F9 and F12, which contained Ca concentrations above 6%. The Ca and Mg concentrations observed in Figure 6c could indicate the presence of amphibole phases such as tremolite (Ca2Mg5Si8O22(OH)2) or hornblende (Ca2(Mg,Fe,Al)5(Al,Si)8O22(OH)2). Additionally, the presence of Ca is linked to the occurrence of calcite (CaCO3), as identified in the XRD results for samples with high Ca content. The detection of Fe and Ti suggests the possible presence of ilmenite (Fe2+Ti4+O), although this phase was not confirmed by other techniques, possibly due to its low concentration. The prevalence of Fe likely indicates the presence of hematite (Fe2O3). In certain samples, traces of K were detected, which can be attributed to the presence of muscovite (KAl2(AlSi3O10)(OH)2). – 215 – ARQUEOL. IBEROAM. 56 (2025) • ISSN 1989-4104 Table 2. Elemental concentrations EDS (in weight %) of Ferrería-style pot sherds. 4.4 TGA-DTG Analysis Thermogravimetry (TGA) and derivative thermogravimetry (DTG) curves are displayed in Figure 7. All samples exhibit a DTG peak between 0 and 250 °C, associated with the loss of water adsorbed in the pores (dehydration). Large peaks in this temperature range are characteristic of highly porous ceramics fired at low temperatures (Kotryová et al. 2016). A mass loss between 0.511 and 2.862% due to adsorbed water is observed across the samples in this range. A DTG peak in the 300– 400 °C range (Figure 7b) is attributed to the oxidation of organic matter (Bayazit et al. 2014; Maritan et al. 2006), resulting in a weight loss between 4.3 and 8.5%. A mass loss in the 400–600 °C range corresponds to the dehydroxylation of clay minerals, as firing induces the transformation of clay into a pseudo-amorphous meta-clay phase (Daghmehchi et al. 2018; Drebushchak et al. 2011). This peak in the DTG curve is indicative of kaolinite dehydroxylation, occurring between 400 and 550 °C, where metakaolinite is formed (Ion et al. 2010; Mangueira et al. 2011). In this temperature range, only samples F7 and F8 exhibited weight losses of 5.2 and 5.4%, respectively. In the 600–850 °C range, the decomposition of carbonates such as calcite occurs (Shoval & Paz 2013; Fabbri et al. 2014). Samples F9 and F12 showed weight losses of approximately 4.1 and 4.2%, respectively, around 850 °C, which aligns with previous results indicating the presence of calcite peaks exclusively in these high-calcium-content samples. The thermograms presented in Figure 7 are characteristic of non-calcareous or low-Ca clays (Shoval & Paz 2013; Ion et al. 2010), displaying peaks between 450 and 650 °C, where the decomposition of clay minerals occurs without significant weight loss due to calcite decomposition. The only exceptions are samples F9 and F12, which correspond to calcium-rich or calcareous clays (Stratis et al. 2011). The weight loss observed in ancient ceramics is influenced not only by the mineralogical composition but also by the type and amount of temper added by the potter during fabrication. This, in turn, can contribute to an increased weight loss due Figure 7. Thermograms of Ferrería-style ceramics: a) TGA and b) DTG.