Introduction

Mortar is the essential binding agent of buildings all over the world. It is used for embedding building stones in masonries (mortar), for coating walls internally (plaster) / externally (render) and for coloured coatings of wall paintings and murals. For producing mortars, three different components are utilised: a binder (e.g. lime, gypsum, mud, cement), aggregates (e.g. sand, rock fragments) and additives (e.g. pozzolans, organic materials). Lime, gypsum and mud are the three most common and historically oldest binding materials used for man-made constructions. The existence of lime was probably discovered by accident; many indigenous tribes for example regularly use hot stones for cooking (Elsen 2006; Carran et al. 2012a). The first use of lime as a binder dates back to the Neolithic (9000–6000 BC), where it is documented from archaeological sites in the Near East and Eastern Turkey (e.g. Goren and Goldberg 1991; Goren and Goring-Morris 2008). In the beginning, lime was mostly used for plasters; subsequently, the Greeks and particularly the Romans established the use of lime for mortars. In Europe, historic mortars are mostly lime-based. It was only during the last two centuries that natural cements and nowadays Portland cement replaced lime as the dominant binder (Elsen 2006; Carran et al. 2012a and references therein).

During the calcination process, naturally occurring limestones (CaCO3) are heated to temperatures > 800 °C. The chemical reaction produces quicklime (CaO) and carbon dioxide (CO2) following the formula CaCO3 + °C → CaO + CO2. By slaking the very reactive quicklime with water, calcium hydroxide (= lime) is fabricated (CaO + H2O → Ca(OH)2). In a last step, lime reacts with atmospherically available CO2, resulting in solid CaCO3 (Ca(OH)2 + CO2 → CaCO3 + H2O). The actual temperature that was reached during quicklime production in historic times is still unknown. Possible temperature ranges have been discussed by archaeologists along with the duration of the calcination process (e.g. Boynton 1980; Torraca 1995; Hughes et al. 2002; Goren and Goring-Morris 2008).

Measurements on an experimental pit kiln document a maximum temperature of 870 °C for less than an hour (Goren and Goring-Morris 2008). Roman authors advised for soft burning (low burning temperature and/or shorter burning durations) and constant temperatures throughout the entire process (Cato in Hooper and Ash 1939). Burning experiments on different historic kilns show uneven temperature distributions, resulting in variations of calcination conditions and the presence of unburnt fragments of limestones (Torraca 1995; Hughes et al. 2002). Lime lumps, binder-derived particles acting as aggregates, supply information about the burning temperature. Lumps, showing microfossils and original sedimentary structures, result from low burning temperatures, while overburnt particles indicate the temperature maximum (Elsen 2006). One factor influencing the temperature needed for calcination is the size of the limestones. Large stones require, in contrast to smaller ones, higher temperatures for calcination of the entire block. The usage of large stones can consequently result in underburnt particles with overburnt rims or overburnt lumps (Boynton 1980).

Many naturally occurring carbonate-rich sedimentary rocks (chalks, limestones and marlstones) consist for a large part of calcareous nannofossils, the fossil remains of marine photoautotrophic algae (Bown and Young 1998). Since many artefacts are made of or contain carbonates, calcareous nannofossils have been used for provenance analysis of various archaeological materials like ceramics (Quinn and Day 2007; Quinn 2008), natural building stones of masonries (von Salis 1995; Lübke et al. 2018; Falkenberg et al. 2021) and mortars (von Salis 1995; Falkenberg et al. 2021).

Another application of micro- and nannofossils lies in the field of archaeothermometry. The mineralised or organic walls of fossils are altered or completely destroyed with increasing temperatures. Experimental heating data of microfossils from unaltered raw material may therefore provide insights into the temperatures reached during production. Organic, siliceous and calcareous microfossils were used in this way to better understand the fabrication processes of ceramics (Jansma 1977; Hunt 1996; Quinn and Day 2007; Quinn 2008; Privitera et al. 2015) and stone tools (Brooks and Dorning 1997). Calcareous nannofossils are affected by increasing temperatures, since their calcium carbonate skeletons are progressively altered. The thermal decomposition of calcite at ~ 850 °C finally results in the destruction of the fossils. The exact temperature of decomposition is controlled by factors like the atmosphere during heating (oxidizing/reducing conditions), impurities of the raw material, particle size and variations in open porosity (Webb and Krüger 1970; Quinn and Day 2007). Changes in abundances, preservation and diversity of calcareous nannofossils have been observed in the case of heating Gault Clay samples (Quinn 1999). These experiments were conducted to get a better understanding of ancient ceramic production.

Here, we record changes in abundances, preservation and diversity of calcareous nannofossils related to the calcination procedure of limestones. We aim at a better understanding of how calcareous nannofossils react to increasing temperatures. Which nannofossil species are destroyed at which temperature? How do the different morphological groups react? Are certain species more heat resistant than others? Since different nannofossil species form crystals with variable form and size, we expect to see differences in their thermal behaviour. Furthermore, we want to test whether these preservational patterns change when source rocks with varying carbonate lithologies (chalks, limestones and marlstones) are used. The temperature of calcite decomposition is influenced by factors like impurities and porosity; we therefore anticipate a relationship between nannofossil preservation, temperature and lithology. We have analysed four source rocks with different lithologies, all of them of Late Cretaceous age (100–66 Ma). All four initial samples yield well-preserved and highly diverse nannofossil assemblages. Our results will help to provide a measure of the temperature or temperature range reached during quicklime production. The experimental data will, ideally, result in a better understanding of historic mortar production.

Calcareous nannofossils

The < 30-µm-sized calcareous nannofossils are the remains of single-celled marine photoautotrophic algae, with coccolithoporids (Fig. 1a) being the most diverse group. They evolved in the Late Triassic (approx. 209 Ma; Gardin et al. 2012) and their diversity increased during the Jurassic (201–145 Ma) and Cretaceous (145–66 Ma). The highest diversity with approximately 149 species is known from the Late Cretaceous (100–66 Ma). Calcareous nannofossils occur in many carbonate-rich marine sedimentary rocks like chalks, limestones and marlstones. Nowadays, they are globally distributed with the highest diversities reached in tropical and subtropical oceans. Via photosynthesis, they contribute substantially to the uptake of CO2.

Fig. 1
figure 1

a Cell anatomy of a coccolithophore (modified after Armstrong and Brasier 2005); b Schematic sketches of a coccosphere consisting of individual coccoliths (left) and nannoliths (right); c Schematic sketches of the morphology of the three main groups of nannofossils (hetero-, holococcoliths, nannoliths); d Scanning electron microscope images of typical forms of nannofossils (made using Inkscape)

Full size image

The characteristic feature of coccolithophores is their extracellular shell, the coccosphere (Fig. 1b). This coccosphere is composed of 20–30 individual low-Mg calcite plates (coccoliths) of ~ 5 µm diameter size. Morphologically, it is possible to differentiate hetero- and holococcoliths. Heterococcoliths are composed of a limited number of crystals with varying complex structures like plates, shields and spines. Holococcoliths are built by numerous individual small (typically 0.1 µm) rhombohedral crystals. A third common group are the nannoliths. They are closely related to coccoliths but show a different morphology (Fig. 1c, d).

In geology, calcareous nannofossils are widely used in the fields of biostratigraphy and palaeoceanography/palaeoclimatology. These applications are based on their variable morphology, their high abundance in past and recent oceans, their worldwide distribution and their rapid evolution over short time periods. Another advantage is their minute size, requiring only small sample sizes (< 1 g) for analyses (Haq 1978; Bown and Young 1998; Young and Henriksen 2003; Armstrong and Brasier 2005; Mutterlose et al. 2005; Jordan 2009).

Mortar

The production methods of lime mortars varied considerably throughout history. Prehistoric mortars of Northern Israel (9000–4000 BC) show mixtures of powdered chalk, small-sized stones and clay in lime plasters, some were made of marls and clays. Only small amounts of burnt lime were used in these cases (Goren and Goldberg 1991; Quinn 2008). Early lime kilns (9000 BC) were probably simple pits with a central fireplace and natural limestones at the pit walls. Alternatively, the pits were filled with alternating layers of fuel (wood) and limestones (Stark und Wicht 1998; Goren and Goring-Morris 2008).

More recent types of kilns used until the Industrial Revolution were clamp and flare kilns. In clamp kilns, limestones and fuel were stacked in alternating layers in a circular structure, and the top was either loosely covered or left open. The exterior was made of earth or stones. The lime produced in these kilns was unevenly burnt. Flare kilns had an exterior made of stones with a cylindrical form. Stones and fuel were separated, with the fuel placed on the ground and the stones stacked above on a wooden framework, which rested on a ledge. This technique resulted in even burning (Dix 1982).

The Romans used standardised protocols, where they advised to use “a good stone as white and uniform as possible” (Cato in Hooper and Ash 1939) and “pure carbonate rock; either marble (pure) or white limestone” (Vitruvius in Morgan 1914). They also recommended a long slaking process for at least three years. For hydraulic properties, additives like crushed ceramics or pozzolan were added. Vitruvius also suggested using more porous limestones for plasters and rendering. Cato described specific lime kilns with a fireplace in great depth to avoid exposure to the wind and two openings for fueling the fire and for retrieving the ash. These kilns were mostly of the flare kiln type (Dix 1982).

Following the fall of the Western Roman Empire, the production of mortar and mortar composition became more variable and less well organised. Some mortars had hydraulic properties, others not, and all kinds of additives (e.g. blood, milk, cheese, eggs) were used. Lime lumps are often found in European medieval mortars (Stark and Wicht 1998; Pavía and Caro 2008; Elsen et al. 2011; Carran et al. 2012a).

In the eighteenth century, it was first recognised that the hydraulic properties achieved by the Romans by adding pozzolan could also derive from burning impure limestones (marly limestones) at temperatures between 950 and 1250 °C. During this high-temperature burning process, calcium and aluminum silicates are formed. The produced lime can harden underwater by combining the carbonation of lime and hydration of the calcium silicate, thereby forming calcium silicate hydrates (Boynton 1980; Ingham 2011). The first Portland cement was produced in the nineteenth century by mixing limestones and clay. Lime mortars are nowadays mainly used for repairing historic buildings (Carran et al. 2012a).

Geological setting

The samples analysed in this study are of Late Cretaceous age (100–66 Ma). The Late Cretaceous period was characterised by a major global transgression, a high sea level and greenhouse conditions. The high sea level resulted in oceanic conditions extending onto continental shelves, leading to large epicontinental shelf seas, where pelagic and hemipelagic carbonate sediments accumulated. This major transgression also caused the flooding of northern Germany as part of the Boreal epicontinental shelf sea, also called Chalk Sea, which existed for more than 35 Ma (Fig. 2b–e). In northwest Europe, mostly pelagic chalks, which are fine-grained, soft and porous carbonate rocks, were deposited. The carbonate content is mainly formed by calcareous nannofossils. In some areas, the soft and porous chalks were transformed into consolidated limestones (Hancock and Kauffman 1979; Hay 1995, 2008; Ineson et al. 2005).

Fig. 2
figure 2

Origin of the studied material. a Map of Germany showing the four studied sections (Lägerdorf: circle; Höver: triangle; Wunstorf: square; Castrop-Rauxel: asterisk). PB, Pompeckj Block; LSB, Lower Saxony Basin; MB, Münsterland Basin; b–e Palaeogeographical maps showing continent-ocean configuration for the Cenomanian (b), Coniacian/Santonian (c), Campanian (d) and late Campanian (e) with the studied sample localizations (modified after Linnert et al. 2011 and Voigt et al. 1999) (made using Inkscape)

Full size image

The carbonate-rich sedimentary rocks studied here are all derived from northern Germany, which formed a sedimentary basin during the Late Cretaceous. The study area covers three different subbasins (Fig. 2a), the Pompeckj Block (PB), the Lower Saxony Basin (LSB) and the Münsterland Basin (MB).

The Pompeckj Block consists of Late Cretaceous (Cenomanian–Maastrichtian; sample localization Lägerdorf; Fig. 2a–e) chalks deposited in a pelagic shallow shelf sea (Voigt et al. 2008; Vejbæk et al. 2010).

Late Cretaceous marine limestones (95–100 wt % CaCO3) and marlstones (35–65 wt % CaCO3), deposited in an outer shelf position, crop out in the LSB, south of the PB (sample localization Wunstorf; Fig. 2a, b). A tectonically caused uplift of the LSB started in the late Turonian/Coniacian, leading to the erosion of siliciclastic material (sand, clay) and the subsequent deposition of marlstones. During the mid-early Campanian, marly limestones (85–95 wt % CaCO3) were deposited in rim synclines of salt diapirs (e.g. Hannover-Braunschweig area, sample localization Höver; Fig. 2d).

The Late Cretaceous (Cenomanian–Campanian) sequences of the MB, adjoining the LSB in the south, consist of limestones, marly limestones and glauconitic to sand-rich marlstones. They accumulated on a shallow shelf. Due to the uplift of the LSB in the north and of the Central Netherlands Basin in the west, marlstones were accumulated during the Coniacian and Santonian (sample localization Castrop-Rauxel; Fig. 2c).

Material

A total of four samples coming from the localizations Lägerdorf (PB), Höver (LSB), Wunstorf (LSB) and Castrop-Rauxel (MB) (Fig. 2a) have been collected in 2020–2021. For each sample, ≥ 10 kg of raw material has been gathered from stratigraphically well-constrained levels of the four quarries. All four samples, labelled here as initial samples, yield primarily well-preserved and highly diverse calcareous nannofossils. The samples differ by geological age, the taxonomic composition of the individual nannofossil assemblages and carbonate content (Fig. 3).

Fig. 3
figure 3

Absolute and geological ages and calcareous nannofossil biozonation of the Upper Cretaceous. Absolute ages after Gale et al. (2020); nannofossil biozonation after Burnett (1998) with modification of Lees (2008). Ages of the four studied samples as well as the ranges of the four sections are marked. On the right side: Images of the four quarries from which material was studied here. Bottom right: Classification of carbonate rocks based on their carbonate and clay content after Correns (1939). The four symbols indicate the measured carbonate content of the samples (Lägerdorf sample: circle; Höver sample: triangle; Wunstorf sample: square; Castrop-Rauxel sample: asterisk). Alb, Albian; Sant, Santonian; l, lower; m, middle; u, upper; calc nanno, calcareous nannofossils; limes, limestones (made using Inkscape)

Full size image

Lägerdorf

Lägerdorf is situated about 50 km northwest of Hamburg (northern Germany; Fig. 2a). The Heidestraße chalk quarry (53°53ʹ2.42″ N, 9°33ʹ7.55″ E; source: Google Maps, 2021; Fig. 3) is currently actively mined. Due to salt tectonics, Upper Cretaceous chalks crop out near the surface (Koestler and Ehrmann 1986). A 75-m-thick succession of chalks (85–98 wt % CaCO3), including marl and flint layers (Dägeling and Kronsmoor Formations), of late Campanian age (Fig. 2e) is exposed (Voigt and Schoenfeld 2010). The sample was taken from the west face of the expansion quarry, second working level (late Campanian chalks, nannofossil subzone UC15d; Fig. 3).

Höver

Höver is positioned 1 km southeast of Hannover (northern Germany; Fig. 2a). The Holcim Höver quarry (52°20′29.88″ N, 9°53′45.99″ E; source: Google Maps, 2021; Fig. 3) is composed of a 100-m-thick succession of marlstones (Emscher Formation) and marl-limestone alternations (Misburg Formation, 65–90 wt % CaCO3) of earliest early to early late Campanian age (Fig. 2d). These limestones and marlstones have been quarried extensively for decades by the cement industry (Ernst 1975; Wiese et al., 2013). The sample was taken from the southwest wall of the quarry (early Campanian marl-limestones, nannofossil subzone UC14b; Fig. 3).

Wunstorf

Wunstorf lies about 20 km northwest of Hannover (northern Germany; Fig. 2a). The abandoned Wunstorf quarry (52°24′6.44″ N, 9°29′23.22″ E; source: Google Maps, 2021; Fig. 3) yields a 110-m-thick succession of early Cenomanian to early Turonian aged (Fig. 2b) marl-limestone alternations (Baddeckenstedt, Brochterbeck and Hesseltal Formations; 70–95 wt % CaCO3). These sedimentary rocks are found on two levels: the lower to lower-middle Cenomanian is accessible on the lower level, the upper level exposes the middle to upper Cenomanian as well as several metres of uppermost Cenomanian to lower Turonian deposits (Wilmsen 2003; Wilmsen et al. 2007; Linnert et al. 2010). The sample originates from the northeastern face of the outcrop (late Cenomanian marl-limestones, nannofossil zone UC3-4; Fig. 3).

Castrop-Rauxel

Castrop-Rauxel is located in the Ruhr area (western Germany; Fig. 2a). An 11-m-thick succession of marlstones (Emscher Formation) crops out in the disused Lessmöllmann pit (51°32′6.38″ N, 7°18′42.05″ E; source: Google Maps, 2021; Fig. 3). The carbonate content varies (15–50 wt % CaCO3). The strata are of latest Coniacian to earliest Santonian age (UC10-12; Sorokoletov and Mutterlose 2007; Fig. 2c). The sample was collected from the upper part of the succession of the eastern wall (early Santonian marl-clay, nannofossil subzone UC11c; Fig. 3).

Methods

The carbonate content of the four initial samples was determined by using the ‘Karbonat-Bombe’. For details, see Müller and Gastner (1971). For the heating experiments, 10 kg of raw material from each sample was crushed at the CDN GeoLab in Namêche (Belgium); 400 g (grain size 5–15 mm) was used per calcination run. The samples were heated at the R&D center in Nivelles (Belgium) in a Nabertherm N11/HR furnace to nine different temperatures (100 °C, 300 °C, 500 °C, 600 °C, 700 °C, 750 °C, 800 °C, 850 °C, 900 °C). The samples stayed for 2.5 h at the target temperature. Starting at 100 °C, it took 2 h to reach the individual target temperatures; cooling to 100 °C required 1.5 h for target temperatures < 600 °C and 2.5 h for > 600 °C. Two different rock samples were burnt simultaneously in the furnace to ensure that both samples experienced exactly the same temperature. The Wunstorf and Höver material was heated jointly as were the samples from Lägerdorf and Castrop-Rauxel. The burnt samples were posted in plastic bags, air vacuumed and then kept in a cabinet desiccator at 50 °C to avoid hydration of the produced quicklime. After analysis, the samples were sealed again in an air vacuum. The samples are stored at the Institute of Geology, Mineralogy and Geophysics, Ruhr-Universität Bochum (Germany).

For checking the calcareous nannofossil content, simple smear slides were prepared, following the technique of Perch-Nielsen (1985). A small amount of the material (~ 0.01 g) was scraped onto a glass coverslip. By adding a drop of distilled water, a thick sediment suspension was produced, which was smeared thinly across a coverslip. The coverslip was then dried on a hotplate and glued onto a labelled glass microscope slide with Norland optical adhesive (NOA61). Due to the small size of the calcareous nannofossils, contamination was reduced during the whole procedure by cleaning every used tool. To ensure that the calcareous nannofossils survive the hydration process, the technique described was modified in two different ways. To avoid hydration, a first set was prepared by using mineral oil instead of water (no hydration; quicklime). For the second set, < 1 g material was removed with a scalpel into a beaker and hydrated by adding 1–3 drops of water. The material reacted for half an hour (hydration; lime) and was then smeared across the coverslip. This step is of importance for samples heated to temperatures > 700 °C when the decomposition of calcite and the formation of quicklime starts.

Based on the presence of calcareous nannofossils in both smear slide sets, 40 settling slides were prepared for quantitative analysis using the technique of Geisen et al. (1999). Less than 1 g of material, which was scraped into a beaker, was hydrated with 1–3 drops of distilled water and reacted for half an hour. The sample was then dried and the material pestled. A small amount of material (~ 15–25 mg, depending on the content of nannofossils in the test smear slide) was suspended in ammoniac water (pH: 8.5). After an ultrasonic bath, each beaker was filled up with ammoniac water to a defined volume (500 ml). The suspension was poured into a settling device with three Teflon cylinders, on which coverslips were placed. After 24 h of settling, the water was slowly removed from the devices. The dry coverslips were mounted on microscopic slides using NOA61. This technique allows to calculate the absolute abundances of calcareous nannofossils (coccoliths per gram sediment) by using the sample weight, the volume of ammoniac water as well as the distance between coverslips and the water surface in the settling devices. Another advantage is that the nannofossils are flat lying, evenly distributed and well separated, which simplifies counting.

An Olympus BH-2 cross-polarized immersion light microscope with a magnification of × 1250 was used for identifying and counting the calcareous nannofossils. At least 300 specimens were counted along randomly chosen transects across the slides. In samples exposed to high temperatures, which yield only few nannofossils, at least 150 fields of view were checked for their nannofossil content. Two additional random traverses were scanned in order to spot rare taxa. An Olympus SC100 camera was used for light microscopic pictures. Scanning electron microscopy (= SEM) images were obtained by a Zeiss Gemini 2 Merlin High resolution thermally aided field emission SEM. The smear and settling slides are stored at the Institute of Geology, Mineralogy and Geophysics, Ruhr-Universität Bochum (Germany).

Geological age determination of the samples followed standard approaches. Following Burnett (1998), the first and last occurrences (FO and LO) of calcareous nannofossil marker species were used to subdivide the Upper Cretaceous (Cenomanian-Maastrichtian) into 21 biozones (UC0–20; UC = Upper Cretaceous). A more detailed subdivision resulted in subzones, labelled with small letters (e.g. UC9a), and a total of 51 zones (Fig. 3). To evaluate the nannofossil preservation, the visual criteria introduced by Roth and Thierstein (1972) and Roth (1983) were used and modified: good preservation (minor fragmentation and etching), good–moderate (minor to moderate fragmentation and etching), moderate (moderate fragmentation and etching), moderate–poor (moderate to major fragmentation and etching) and poor (major fragmentation and etching). Diversity was determined by using the Shannon index based on species richness and heterogeneity (Shannon and Weaver 1949). High values of this index stand for highly diverse assemblages, and values closer to 0 represent less diverse assemblages dominated by one species only. The relative abundance of individual species was calculated for each of the nine temperature levels as percentage in relation to all specimens encountered.

Results

Carbonate content and absolute abundance of calcareous nannofossils

Based on the carbonate content, the four samples can be classified as marly chalk (Lägerdorf: 88 wt % CaCO3), marl-limestone (Höver: 83 wt % CaCO3; Wunstorf: 76 wt % CaCO3) and marl-clay (Castrop-Rauxel: 20 wt % CaCO3; Fig. 3).

The absolute abundance of calcareous nannofossils shows, in relation to the temperature, a similar overall trend for the four sample sets (Figs. 4 and 5). Minor modifications have, however, been noted. An abundance decrease from the initial to the 100 °C sample was observed in three sets. It is only in Höver that abundance increases up to the 300 °C sample. From 100 to 600 °C, abundance decreases slightly for Lägerdorf, Höver and Castrop-Rauxel. The Wunstorf set documents a slight increase in absolute abundance from 100 to 500 °C. At temperatures > 600 °C, absolute abundances decrease sharply in all four sets. The 900 °C samples provided very low values for all four localizations (max. 0.35*107 specimens/g sediment, Lägerdorf; min. 0.03*107 specimens/g sediment, Castrop-Rauxel). For Castrop-Rauxel, the absolute abundance is already very low at 700 °C (2.85*107 specimens/g sediment) (Fig. 5); in one case, an increase in absolute abundance was observed from 800 to 850 °C (Lägerdorf; Fig. 4).

Fig. 4
figure 4

Calcareous nannofossil data from the Lägerdorf (top) and Höver (bottom) samples in relation to the heating temperatures. The nannofossil data are plotted on the x-axis, the temperatures on the y-axis. The figure shows temperature-related changes of the absolute abundance, preservation, species richness, Shannon index and relative abundances of selected nannofossil taxa. The initial sample is the original sample taken in the field. Initial sam, initial sample (made using Inkscape)

Full size image
Fig. 5
figure 5

Calcareous nannofossil data from the Wunstorf (top) and Castrop-Rauxel (bottom) samples in relation to the heating temperatures. The nannofossil data are plotted on the x-axis, the temperature on the y-axis. The figure shows temperature-related changes of absolute abundance, preservation, species richness, Shannon index and relative abundances of selected nannofossil taxa are figured. The initial sample is the original sample taken in the field. Initial sam, initial sample (made using Inkscape)

Full size image

Preservation and diversity of calcareous nannofossils

Preservation is good with only minor etching, fragmentation and overgrowth from the initial to the 300 °C sample (Höver, Wunstorf) resp. 500 °C sample (Castrop-Rauxel). For Lägerdorf, a good–moderate preservation was observed up to the 500 °C sample. Preservation deteriorates in all sample sets > 300 °C resp. > 500 °C. Nannofossils are becoming poorly preserved at ≥ 700 °C (Castrop-Rauxel) resp. ≥ 750 °C (Lägerdorf, Höver, Wunstorf; Figs. 4, 5, 6 and 7; Online Resource 1).

Fig. 6
figure 6

Images of the heterococcolith species Watznaueria barnesiae (top) and the holococcolith species Calculites obscurus (bottom) throughout the burning process. The images are cross polarised light microscope images. The initial sample and nine different burning temperatures are plotted on the x-axis, the four studied sample sets (Lägerdorf, Höver, Wunstorf, Castrop-Rauxel) on the y-axis. The initial sample refers to the original sample taken in the field (made using Inkscape)

Full size image
Fig. 7
figure 7

Images of the nannolith species Micula staurophora (top) from Lägerdorf, Höver and Castrop-Rauxel and the nannolith species Eprolithus floralis (top) from Wunstorf throughout the burning process. The images are cross polarised light microcope images. The initial sample and nine different burning temperatures are plotted on the x-axis, the studied sample sets on the y-axis. The initial sample refers to the original sample taken in the field. Bottom: Scanning electron microscope images of the heterococcolith species Watznaueria barnesiae (top) and the nannolith species Micula staurophora (bottom) from the Lägerdorf set throughout the burning process. The initial sample and five selected temperature levels are plotted on the x-axis. The initial sample refers to the original sample taken in the field (made using Inkscape)

Full size image

The species richness is stable in samples from all four localizations up to 500 °C; it decreases at higher temperatures. This decline is most pronounced in the Castrop-Rauxel samples. One exception was noted for Lägerdorf, where the species richness increases slightly from 750 to 800 °C. For the 900 °C, level only 1 to 4 species were observed (Figs. 4 and 5).

A similar trend is seen for the Shannon index. The Shannon index is stable until 600 °C, followed by a decrease to a min. of 0.00 (Wunstorf, Castrop-Rauxel) and a max. of 0.66 (Lägerdorf) at 900 °C (Figs. 4 and 5).

Relative abundances of heterococcoliths

The relative abundances of the various heterococcoliths decrease in general with increasing temperatures (Figs. 4 and 5). Some taxa are, however, relatively enriched at higher temperatures. This deviating pattern has been observed for the genus Watznaueria at temperatures of 500–850 °C (Wunstorf; Fig. 5) and for the Cretarhabdaceae at temperatures of 500–850 °C (Höver)/600–850 °C (Wunstorf)/700–750 °C (Castrop-Rauxel; Figs. 4 and 5). For Castrop-Rauxel, high relative abundances have also been observed for the Arkhangelskiellaceae and Kamptneriaceae at 750 °C and for the genera Chiastozygus, Staurolithites and Eiffellithus at 800 °C, 700 °C and > 500 °C resp. (Fig. 5). A high relative abundance of the genus Zeugrhabdotus was seen at 850 °C (Wunstorf; Fig. 5). The genus Biscutum increases from the initial to the 100 °C (Lägerdorf) resp. 300 °C sample (Höver, Wunstorf, Castrop-Rauxel), followed by a decrease > 100 °C resp. > 300 °C (Figs. 4 and 5).

The various heterococcolith taxa disappear at different temperatures (Figs. 4 and 5). The genera Staurolithites and Helicolithus vanish at low temperatures (min. 500 °C, max. 750 °C; for details, see Figs. 4 and 5; Online Resource 2). Other taxa are disappearing at slightly higher temperatures (min. 600 °C, max. 900 °C). These taxa include the Kamptneriaceae, T. orionatus and the genera Zeugrhabdotus, Chiastozygus and Biscutum (for details see Figs. 4 and 5; Online Resources 1–2). High temperatures (min. 750 °C, max. 900 °C) are tolerated by the Arkhangelskiellaceae, the Cretarhabdaceae and the genera Watznauria, Eiffellithus and Prediscosphaera. At 900 °C, only a few heterococcoliths have been encountered. These include the genera Watznaueria (Höver, Lägerdorf), Prediscosphaera (Castrop-Rauxel) and Eiffellithus (Wunstorf, Lägerdorf; Figs. 4, 5 and 6; Online Resources 1–2).

Relative abundances of holococcoliths

Holococcoliths (Calculites, Lucianorhabdus) have been observed in the Lägerdorf, Höver and Castrop-Rauxel sample sets; they are absent in Wunstorf . Their relative abundances increase consistently in all three sets with higher temperatures (Figs. 4 and 5).

Calculites becomes more abundant at > 300 °C (Höver), > 600 °C (Castrop-Rauxel) and > 700 °C (Lägerdorf); it is still present at 900 °C in two samples (Lägerdorf, Höver; Figs. 4, 5 and 6). Lucianorhabdus increases with temperatures > 600 °C (Höver, Castrop-Rauxel) and > 700 °C (Lägerdorf); Lucianorhabdus is missing at temperatures > 750 °C (Castrop-Rauxel) and at 900 °C (Lägerdorf, Höver; Figs. 4 and 5).

Relative abundances of nannoliths

The relative abundances of nannoliths increase in all four sample sets with higher temperatures. The taxa Micula, Eprolithus and Microrhabdulus decoratus have high relative abundances at temperatures > 600 °C (Figs. 4 and 5). Differences can be seen in their disappearance: M. decoratus is, apart from very rare specimens in the two additionally checked random traverses, absent at temperatures > 700 °C (Wunstorf), > 750 °C (Castrop-Rauxel) and > 800 °C (Höver, Lägerdorf; Figs. 4 and 5; Online Resources 1–2). Micula has a relative abundance of 33% at 900 °C (Lägerdorf); in Höver and Wunstorf , Micula and Eprolithus are both missing at 900 °C. In the Castrop-Rauxel samples, Micula is absent at temperatures > 700 °C, but present in the two random traverses at 750 °C and 850 °C (Figs. 5 and 7; Online Resources 1–2).

Images of calcareous nannofossils species of all four sample sets are documented in Online Resource 1. A list of all counted nannofossils is given in Online Resource 2.

Discussion

Calcareous nannofossils

Absolute abundance, preservation and diversity

In all four sample sets, the data for absolute abundance, preservation, species richness and Shannon index are roughly stable up to 500–600 °C above which they decline dramatically (Figs. 4, 5, 6 and 7; Online Resources 1–2). This decline of the four parameters corresponds to the onset of calcite decomposition at ~ 600 °C, a process that peaks at ~ 900 °C (Kumar et al. 2007). The decomposition starts gradually at ~ 600 °C and increases rapidly with temperatures > 750 °C (Karunadasa et al. 2019). Our data confirm poor preservation for temperatures ≥ 700 °C/ ≥ 750 °C (Figs. 4, 5, 6 and 7; Online Resource 1). Calcareous nannofossils are composed of low-Mg calcite and can incorporate small concentrations of various elements like Mg, Sr, Se, Ni and Fe (Bottini et al. 2020). The decomposition temperatures observed here still fit quite well with those known of pure calcite.

When viewed in more detail, a slight decrease of the absolute abundance has to be noted for the 100–600 °C interval for three localizations (Lägerdorf, Höver, Castrop-Rauxel). Preservation deteriorates at temperatures > 300 °C/ > 500 °C (Figs. 4, 5, 6 and 7; Online Resource 1). The decomposition of calcite is related to a substantial weight loss, which is caused by the release of CO2. A loss of weight is in some cases already observed at lower temperatures before the decomposition (Karunadasa et al. 2019), which explains our findings.

Relative abundances of hetero-, holococcoliths and nannoliths

The three different morphological groups of calcareous nannofossils (hetero-, holococcoliths, nannoliths) studied here provide different relative abundance patterns for the calcination process. Heterococcoliths decrease with high temperatures > 500 °C, while holococcoliths and nannoliths increase in relative abundance (Figs. 4 and 5). Holococcoliths and nannoliths are in general better preserved at high temperatures > 700 °C and are in most cases unfragmented (Figs. 6 and 7; Online Resource 1). These observations suggest that holococcoliths and nannoliths are more heat resistant than most heterococcoliths.

Crystal size is an important factor controlling the calcite decomposition temperature (Webb and Krüger 1970). The observed non-uniform resistance during calcination is, therefore, related to different crystal sizes used by the three groups to build their skeleton. Heterococcoliths are formed by complex crystal structures, and holococcoliths are composed of small rhombohedral crystals. Nannoliths (e.g. Micula, Eprolithus) finally consist of crystals with different shapes, often arranged in laminae (Figs. 1d and 7). Individual smaller crystals are more resistant to heat than larger ones (Carran et al. 2012b). The small size of the holococcolith crystals may thus explain their heat resistance. Small sized heterococcoliths (Biscutum, Zeugrhabdotus erectus) are in contrast destroyed at relatively low temperatures (around 700 °C; Online Resource 1–2).

The different crystal forms of the three nannofossil groups are a second parameter that controls their stability at high temperatures. The rhombohedral form of the holococcolith crystals, in combination with the layered structure (Fig. 1d), apparently leads to a higher heat resistance than the variable complex crystal forms of small heterococcoliths. The observed diverging patterns of the relative abundance of heterococcoliths may also be related to different crystal forms in the various taxa. The high relative abundances of nannoliths can be attributed to their crystal form and their lamina structure.

The specific surface area of the crystals may also influence the heat resistance. Large surface areas can absorb more heat than small ones and are thus more easily affected by a temperature increase. This may explain why the holococcolith genera Micula und Eprolithus, having a compact structure, are more heat resistant than the holococcolith species M. decoratus with a rod shape (Online Resource 1). This phenomen also explains the destruction of elongated crystal structures like the spines of Eiffellithus eximius and Eiffellithus turriseiffelii at low temperatures (> 600 °C) (Online Resource 1).

The described relative abundance patterns, discussed here, may also be related to an observational bias. Certain taxa are easier to recognize from fragments than others. When complex structures like the spines of Eiffellithus eximius and Eiffellithus turriseiffelii are decomposed, it is difficult to determine the exact species (Online Resources 1–2). This bias may explain the early disappearance of small heterococcoliths because it is harder to find and identify their small fragments. The identification of holococcoliths and nannoliths is more straightforward once decomposition has started. Both groups are not fragmented at high temperatures (except for M. decoratus; Fig. 7; Online Resource 1), which may result from their layered structure.

Carbonate content and mineral composition

Absolute/relative abundances and the preservation of calcareous nannofossils clearly correlate with the chemical composition of the raw material (Figs. 4 and 5). The Castrop-Rauxel sample set (20 wt % CaCO3) provided a low absolute abundance at 700 °C. Analogue low abundances were recorded for Wunstorf (76 wt % CaCO3) at 800 °C (Fig. 5), for Höver (83 wt % CaCO3) at 850 °C and for Lägerdorf (88 wt % CaCO3) at 900 °C (Fig. 4). The decline of the species richness, Shannon index and preservation is more pronounced for Castrop-Rauxel than for the other three localizations (Figs. 4 and 5). These observations are here explained by the different carbonate and clay content of the sample sets.

Impure carbonates require lower temperatures for calcination than pure ones (Carran et al. 2012b). The decomposition temperature of calcite during the firing of pottery, where clays are used as dominant raw material, is 650–750 °C (Rice 1987). This is much lower than the temperatures (~ 900 °C) needed for the calcination of purer carbonates. Firing experiments on calcite rich clays also revealed a low calcite decomposition temperature of 700–800 °C (Trindade et al. 2009). In this study, the clay minerals decomposed at lower temperatures than calcite, e.g. kaolinite at 500 °C and illite at 700 °C. The silica and alumina produced during the clay mineral decomposition reacted with free CaO, forming Ca2Al2SiO7 (gehlenite), CaSiO3 (wollastonite) and Ca2SiO4 (larnite) (Trindade et al. 2009). The thermal effects of the reactions of carbonates and other solid phases often influence the decomposition temperature (Webb and Krüger 1970). In addition, even small amounts of impurities (like salts) lower the decomposition temperature of calcite. The presence of organic matter, common in clays, can cause an exothermic oxidation, thereby reducing the decomposition temperature (Webb and Krüger 1970). The burning of organic matter can further increase the porosity, aiding the calcination process (Carran et al. 2012b). The calcite decomposition temperature as well as the formation of silicate phases during clay firing depends on various parametes, e.g. the mineralogical and chemical composition, grain size and heating rate. Thus, an exact calcite decomposition temperature cannot be given for calcite rich clays (Gliozzo 2020).

Comparison to other heating experiments

So far, only few studies have investigated the reaction of calcareous nannofossils and other calcareous microfossils when exposed to increasing temperatures. By comparing the current data set with published results of heating experiments, we want to check whether general patterns are recognisable.

Quinn (1999), who tested the thermal behaviour of calcareous nannfossils of heated Gault Clay samples, observed a progressive decrease in absolute abundance, preservation and diversity with increasing temperature. The same trend was observed in this study. The experiments of the Gault Clay samples further document an increase in the relative abundance of Watznaueria, while the relative abundances of all other taxa decreased. In our study, a positive correlation of the Watznaueria abundances and the increasing temperature has only been observed for the Wunstorf set. The remaining three localizations show decreasing abundances of Watznaueria and increasing ones for holococcoliths and nannoliths. This discrepancy is explained by a different taxonomic composition of the nannofossils of the samples. The assemblages of the raw material of both, the Gault Clay and the Wunstorf samples, are dominated by heterococcoliths (mainly by Watznaueria), while heat-resistant holococcoliths and nannoliths are rare (Crux 1991; Kanungo et al. 2018).

Quinn (1999) observed calcareous nannofossils in an oxidizing atmosphere up to 800 °C, in a reducing setting up to 700 °C. The same temperature (800 °C, oxidizing atmosphere) was noticed in heating experiments of clay samples (Privitera et al. 2015). The Gault Clay Formation has a low carbonate content (8.1–40.9 wt %, mean value: 24.3 wt % CaCO3; Kanungo et al. 2018). In our study, calcareous nannofossils are present up to temperatures of 900 °C in the three samples with high carbonate content. In the low carbonate content samples from Castrop-Rauxel (marl-clay, 20 wt % CaCO3), nannofossils are rare at temperatures > 750 °C. These results, which correspond well with findings of Quinn (1999), suggest that calcareous nannofossils decompose in clay-rich sediments at lower temperatures than in carbonate-rich ones. Low calcite decomposition temperatures observed in calcite-rich clays (Trindade et al. 2009) support this view.

Heating experiments on calcareous foraminifera (Privitera et al. (2015) revealed no relevant morphological changes of the skeletons up to the calcite decomposition temperature of 800–900 °C. In an oxidizing atmosphere, the calcite was decomposed at 900 °C; the foraminifera shells were, however, still recognizable as molds or as pseudo skeletons. No traces of the microfossils have been observed at 1200 °C. In a reducing atmosphere, calcareous foraminifera were still well-preserved at 900 °C (Privitera et al. 2015). These calcite decomposition temperatures closely resemble those encountered in the present study. The nannofossils are, however, not observable as molds like the foraminifera.

Experiments with clays containing unfossilised calcareous mollusk shells reveal progressive changes in the microstructure of the shells, starting at 650 °C. Above 850 °C, the shell fragments are completely decomposed (Maritan et al. 2007). These decomposition temperatures match well with those noted in the present study.

Application to historic mortar production

Based on our new experimental data, four phases characterised by different preservational modes of calcareous nannofossils can be differentiated (Table 1):

  • Phase 1 (≤ 500 °C): This phase is characterised by a high absolute abundance, which is slightly lower than in the initial sample sets. Preservation is as good as in the initial sample. High values for the species richness and Shannon index are typical. The calcareous nannofossil assemblage is overall similar to the assemblage of the initial sample. All three morphological types (hetero-, holococcoliths, nannoliths) including small heterococcoliths are present. The latter are marked by high relative abundances of the genus Biscutum.

  • Phase 2 (> 500 °C and ≤ 700 °C): Moderate absolute abundances were observed for this phase. Preservation is moderate to moderate-poor. The species richness and Shannon index are moderate with slightly lower values than in the initial sample and phase 1. First changes in the calcareous nannofossil assemblage are seen. Heterococcoliths show lower relative abundances than in the initial sample, and small heterococcoliths are present, but rarer. Holococcoliths and nannoliths have slightly higher relative abundances than in the initial sample (peak of M. decoratus).

  • Phase 3 (> 700 °C to ≤ 850 °C): This phase shows low values for absolute abundance, species richness and Shannon index. Preservation is poor. The calcareous nannofossil assemblage is dominated by holocooccoliths, nannoliths and few heterococcoliths (e.g. Watznaueria, Cretarhabdaceae). Small heterococcoliths are absent or very rare.

  • Phase 4 (≥ 900 °C): The absolute abundance, species richness and Shannon index are extremely low; preservation is poor. Only a few calcareous nannofossils have been encountered. These include heterococcoliths (genera Watznaueria, Prediscosphaera, Eiffellithus), holococcoliths (genus Calculites) and nannoliths (genus Micula).

Table 1 Absolute abundance, preservation, diversity and morphogroup patterrns (hetero- and holococcoliths, nannoliths) are documented for four different temperatures ranges. The findings are compared to the results of the initial sample. The four phases can be used for estimating the temperature achieved in historic quicklime production
Full size table

The four phases introduced here provide a means for estimating a burning temperature range of binder and lime lumps of historic buildings. In addition to the four phases, the disappearance of specific taxa, listed in the “Results” section, can be used to further narrow down the burning temperature.

There are additional parameters that need to be considered when reconstructing historic burning temperatures. One important factor is the burning duration. In our study, the samples were burnt for 2.5 h. Quinn (1999) fired the Gault Clay samples at 600 °C for different durations (1, 2, 3, 4 h). A progressive decrease in overall abundance, preservation and diversity was observed with increasing duration, but the effect was less severe than that caused by increasing temperature. The duration of the burning process is therefore of importance and needs to be studied in more detail in the future.

The oxygen content of the atmosphere during the burning process is another factor that contributed to nannofossil preservation. Specimens are destroyed earlier in a reducing atmosphere (> 700 °C) than in an oxidizing one (> 800 °C; Quinn 1999). In contrast, Privitera et al. (2015) and Webb and Krüger (1970) observed better preservations of calcareous microfossils at high temperatures in a reducing atmosphere than in an oxidizing one. This pattern is explained by a postponement of the calcite decomposition in a reducing setting.

A thorough analysis and account of the composition of the initial material (absolute abundance, preservation and nannofossil assemblage) is yet another pre-requisite for obtaining reliable results. The raw materials used here yield high abundant and diverse nannofossil assemblages in all four sample sets. All nannofossil species showed at least a good to moderate preservation. Knowledge of the initial composition is particularly important when studying lime lumps in mortars. The findings of rare and poorly preserved nannofossils can also be attributed to low abundance and poor preservation of nannofossils in the raw material and not to high burning temperatures. The present experiments and those of Quinn (1999) document differing results with respect to the composition of calcareous nannofossils, when exposed to high temperatures. These discrepancies result from a primarily different composition of the nannofossil assemblages in the studied raw material. Ideally, the potential raw material is analysed to account for its unaltered nannofossil content.

Conclusion

The heating experiments on four carbonate-rich sedimentary rocks (marly chalk, marl-limestones, marl-clay) allow the following conclusions with respect to their calcareous nannofossil content:

  1. 1.

    The calcareous nannofossil assemblages are clearly effected by increasing temperatures. Up to a temperature of 500 °C parameters like absolute abundance, preservation and diversity do not change significantly. Decomposition of the carbonates begins at temperatures > 500 °C, resulting in a sharp decrease of the three parameters. This process accelerates at > 750 °C; at this temperature preservation is poor. At 900 °C, no or only very rare specimens have been encountered.

  2. 2.

    The three morphological groups (hetero-, holococcoliths, nannoliths) expose different heat resistances. Holococcoliths and nannoliths have higher relative abundances at high temperature and are thus more heat resistant than heterococcoliths. This is related to their crystal size and form, surface area and their layered structure.

    Individual genera of heterococcoliths show diverging results of the relative abundance. Some taxa (e.g. Watznaueria, Cretarhabdaceae) are more stable and increase in relative abundance with high temperatures. They are more heat resistant than other heterococcolith taxa. Small-sized heterococcoliths (e.g. Biscutum) are less heat resistant and are destroyed at lower temperatures. These differences are attributed to their different crystal forms and surface-to-volume ratio.

  3. 3.

    Our results indicate an influence of the carbonate content on absolute abundance, preservation and diversity. In clay-rich samples, calcareous nannofossils are destroyed at lower temperatures than in carbonate-rich material. An explanation is the carbonate and clay content. Small impurities and organic matter can also lead to lower calcite decomposition temperatures.

  4. 4.

    Based on the nannofossil record, four temperature phases have been recognised. In the future, they can be used to estimate the temperature during quicklime production. These phases provide, however, only a first clue to the burning temperature. Other factors like the burning duration and composition of the atmosphere need to be studied in the future. Furthermore, the preservation and the composition of the original nannofossil assemblage of the source rock have to be considered. Archaeothermometric studies should thus be linked to provenance analyses.