Introduction

Since the prehistoric era, humans have used mineral-based coloring materials (pigments) in paintings, makeup, architectural structures, sculptures, and the ceramic surfaces for the purposes of coloring, decoration, protection, ritual, and ceremonial meaning [1,2,3,4,5]. Among all the natural minerals, iron oxide-based reddish materials, is the most widely used as a pigment in human history around the world [4,5,6,7,8,9]. In general, two major groups of iron-oxides predominate in natural ochre, generated from the weathering environment, such as Fe3+-oxides and -oxyhydroxides, simply termed as “iron oxides” in this paper. Moreover, recent scientific analysis reveals that Neanderthals, who were active before modern humans, also used iron oxide-based red ochre as pigments [10,11,12]. It is because various earthy pigments of yellow, red and brown, commonly called ocher, were generally abundant in surface deposits, so that they were relatively cheap and easily available. Furthermore, the formation of the reddish mineral hematite, well crystalline α-Fe2O3 with weak ferromagnetic properties, in ochre by thermal treatment were common in human history, a process was already known in antiquity [13]. In addition, recent literatures explain the effect of thermal treatment on color changes, along with the particle size, crystallinity, purity, and magnetic property of hematite in iron oxides-based raw materials for ancient red pigments [13,14,15].

In Korea, such iron oxides-based red pigments are also found in archaeological sites and cultural heritages throughout all historical periods. In particular, red paint-ceremony, held in the tumulus, is closely related to social ideas and faith of the ancient Kingdoms, established after the Iron Age. In addition, these traces are considered as ritual action of the ruling class accompanied by an expert group, such as shamans and/or Buddhist monks [4, 16].

Recently, excavation of red pigments in ancient tombs of the Gaya cultural circles, the ancient federation kingdom in the southern region of the Korean Peninsula (1st–6th centuries AD), has been reported; which excavated in various forms, such as stored in vessels, painted on the walls of stone chambers, murals, corpses, and artifacts. Elemental and mineral phase analyses have previously been used to identifications of their raw materials and combinations, resulted in iron oxide-based and cinnabar-based red pigments were distinguished [16,17,18]. However, in the case of iron oxides-based pigments, they have been referred to as simply “high-purity and/or low-purity iron oxide pigments”, despite the extraordinary discovery of pulverized pigment, as well as different fragments of red pigment and a significant number of red-painted archaeological remains at these regions. Because they have never been considered for specific iron oxides phase analysis; their iron-oxides components are often characterized by small crystal size, significantly low concentrations, high foreign element substitutions, and close intergrowths with other associated fine grained mineral particles. With regard to these factors, 57Fe Mössbauer spectroscopy contributes to distinguish and identify individual structural forms, amorphous and superparamagnetic nanostructured iron oxides particles, to analyze magnetic behavior caused by their formation mechanism. Therefore, in combination with chemical and physical methods, 57Fe Mössbauer parameters are can supports to more valuable information for identification and characterization of iron-bearing minerals in natural materials and its artefacts, which are difficult to distinguish by phase analyzes that have been generally performed [8, 19,20,21,22,23,24].

In this context, this study will identify the variety of mineral composition (especially magnetic characteristics of each iron-oxides phases) used as red pigments in four ancient tombs of three renowned Gaya sites in Haman, Gimhae and Hapcheon in Gyeongsangnam-do Province, South Korea (Fig. 1). Furthermore, they were assessed to determine whether the natural minerals were used, or if they were transformed into artificial iron compounds by pre-treatment.

Fig. 1
figure 1

Location and tomb numbers of the pigment collection sites

Full size image

Experimental

Archaeological outline and sample description

Four red pigment samples for each type from the ancient tombs of Gaya cultural circles of these three archaeological regions were considered for analysis: Haman, Gimhae, and Hapcheon (Fig. 1 and Table 1).

Table 1 The collection site, description, Lab CIE color coordinates of the pigment samples
Full size table

Gaya is the ancient federation kingdom in the southern region of the Korean Peninsula (1st–sixth century AD). Among the three regions where the tombs are located in this investigation, Gimhae region was the seat of “Garakguk”, the leader of the Gaya federation. In addition, Haman and Hapcheon were the provenances of small allies such as “Aragaya” and “Daraguk”, respectively. Moreover, their archaeological remains excavated from ancient tombs in these regions reflect the powerful authority of the ruling class and the fact that each small allies grew up on a similar cultural basis.

As described in Table 1, red pigments were preserved in different status in each tumulus, and specimens for analysis were collected in powder form. For specimen analyses, the collected powders were directly used, and additional grinding was performed only for structural analysis.

Instrumentation and measurements

The raw powders of each pigment were observed using a microscope (Axio Imager A2, Zeiss, Germany), and the elemental distribution of each particle was analyzed using a scanning electron microscope with energy dispersive X-ray spectroscopy (SEM–EDS, MIRA3-LMH, Tescan, Czech). Chromaticity determinations were performed by Color Spectrophotometer (S60, X-rite co., USA), and recorded based on the L*a*b* space model proposed by Commission International d’Eclairge (CIE); L* is measure of lightness, a* redness, and b* yellowness.

X-ray diffraction (XRD), Raman and Mössbauer spectroscopy were performed to identify the mineral composition and magnetic properties of the samples. The XRD diffractograms were recorded using a Cu-Kα diffractometer (Empyrean, Malvern Panalytical, BK), which scanned in the range of 5°–60° 2θ at 45 kV and 40 mA with a scanning interval of 0.02° 2θ. Micro-Raman spectra were obtained using an XperRam F2.8 (Nanobase, Republic of Korea), with × 40 focused BX41M-LED microscope (Olympus, Japan). Laser wavelength of 473 nm (2300 gr/mm) was used for excitation. Mössbauer spectra were recorded using 57Co (60 mCi) that was diffused in Rh as a γ-ray source at 295 K. Isomer shifts were referred to as α-Fe at room temperature. Subsequently, the obtained spectra were fitted to Lorentzian shapes and decomposed into quadrupole doublets and sextets using a MossWinn 4.0i program.

Analytical results

Morphology and elemental distribution

The detailed shape and color of the particles in each sample were observed by microscope, SEM, and color spectrophotometer as shown in Fig. 2 and Table 1. Two painting layer samples with the orange and reddish-brown color (type 1 and type 2, respectively), consist inequigranular particles. Overall, the particles are characterized by the coexistence of micro-sized polyhedric particles with sharp borders and platy crystal, and nano-sized dispersed particles that covering the surface of micro-particles. In addition, black colored single particles are partially observed in type 1. As for the red pigment paste sample (type 3), medium to high rounded polyhedron shaped nano-particles were mainly observed, and light-colored coarse-grained mineral particles appear to be mixed. The sample type 4, with the highest redness, has only nano-sized irregular polyhedral particles with rounded edge. Moreover, these nanoparticles of the sample type 4 have clearly higher dimensions, sphericity and rounding than those of the sample type 3.

Fig. 2
figure 2

Microscopic and SEM images of each type of pigments for a, b Type 1, c, d Type 2, e, f Type 3, and g, h Type 4

Full size image

The chemical composition of the particles in each sample were characterized by SEM–EDS. In addition, the EDS mapping of the elements was carried out as shown in Fig. 3, with attention to the indicator elements of the iron oxides and ochres (i.e., iron, silicon and aluminum). The elemental distribution of type 1 shows that iron, results confined to particles of small dimensions compared to silicon and aluminum. In addition, the iron-containing particles are observed in a dispersed pattern on the silicate particles without forming large-sized aggregates, and some Fe-rich single particles are rarely confirmed. In the type 2, most of the iron distribution patterns are similar to type 1, while some particles are observed in an overlapped with silicon and aluminum. In contrast, the EDS elemental mapping of the type 3 and type 4 revealed aggregates of the iron oxides. The type 3 is characterized by the mixture of some silicate and alumino-silicate particles, while the type 4 is the identified as compound of pure iron oxides.

Fig. 3
figure 3

SEM–EDS elemental distribution maps of the pigment samples for a Type 1, b Type 2, c Type 3, and d Type 4

Full size image

Mineral phase identification: bulk powder and reddish particles

The result of bulk powder XRD for each specimen are illustrated in Fig. 4. The type 1 shows as major crystalline phases in the soil, such as quartz, feldspar, mica group minerals, and kaolin. In addition, very weak diffraction patterns of (104) and (110) peaks of hematite are recorded near the 33 and 35.6° 2-theta area; as well as the weak (311) peak of magnetite overlaps with the (110) peak of hematite, which is detected with relatively higher intensity. The diffractogram of type 2 shows the presence of quartz, feldspar, chlorite group minerals, mica group minerals, and more crystalline phase of hematite than that of type 1. In the type 3, quartz and hematite, with a predominance of hematite, are the main mineral phases with weak traces of feldspar and mica. Finally, XRD analysis of type 4 shows the presence of pure hematite.

Fig. 4
figure 4

X-ray diffraction patterns of the pigment samples (Ch: chlorite, F: feldspar, H: hematite, K: kaolin, M: mica, Q: quartz)

Full size image

Figure 5 presents Raman spectra of the red-colored particles in each sample. As for the type 1, characteristic resonant peaks of hematite were not detected, while broad and weak hematite peaks were detected at 222 cm−1, 292 cm−1, 412 cm−1 and 610 cm−1 in the type 2. In addition, for these two samples, the characteristic Raman spectral peaks of quartz appeared at 463 cm−1 and 205 cm−1, 356 cm−1, 463 cm−1, respectively. The Raman spectra of reddish particle in the type 3 and type 4 appear only sharp hematite peak at 223 cm−1, 288 cm−1, 408 cm−1, and 602 cm−1 (cf. for the type 4, another hematite peak at 497 cm−1 was observed).

Fig. 5
figure 5

Raman spectra of reddish particles in the pigment samples (H: hematite, Q: quartz)

Full size image

Characterization of iron compounds: Mössbauer spectroscopy

Further details on the iron compounds were obtained from Mössbauer spectra collected at room temperature (295 K) as shown in Fig. 6, and their Mössbauer parameters were summarized in Table 2. The Mössbauer spectra each consist of several subspectra that are overlapped on each other showing that at least two iron containing phases are present.

Fig. 6
figure 6

Mössbauer spectra of the pigment samples at 295 K

Full size image
Table 2 Mössbauer parameters of the pigment samples at 295 K (δ = isomer shift, EQ = electric quadrupole splitting, Hhf = magnetic hyperfine field)
Full size table

The decomposed spectrum of type 1 shows a superparamagnetic behavior, which dominated by three quadrupole doublet; whose parameters are consistent with superparamagnetic Fe3+ of iron oxides (D1) and Fe3+ in the phyllosilicate structure of fine clay minerals (D2 and D3) [21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36]. In addition, two sextets of hyperfine magnetic splitting were also contained, which associated with magnetite (S1 and S2) [21,22,23,24]. The Mössbauer spectrum of type 2 appears as four distinct doublets (D1, D4, D5, and D6) and a single magnetically split sextet (S3). The computed parameters indicate that for each components, D1 corresponds to Fe3+ in superparamagnetic iron oxides, and D4, D5, and D6 assigned to paramagnetic Fe2+ and Fe3+ in chlorite group minerals (clinochlore), respectively [35, 36]. Moreover, S3 is identified as magnetic phase of hematite [21, 27, 29]. In the hematite rich and deep reddish-brown pigments, type 3, the Mössbauer spectrum appears as the sum of almost fully sextet with relatively sharp lines, attributable to weakly ferromagnetic hematite, and some trace of a central doublet which is associated to paramagnetic iron oxides [21,22,23,24,25,26,27,28,29,30,31]. For the Mössbauer data in type 4, the spectrum consists of a single sextet with sharp lines, and the parameters are fitting with those expected for weakly ferromagnetic hematite [21, 27, 29].

Discussion on pigment preparation

The renewed multi-analysis investigation on four typical red pigments allowed us to improve previous inferences on iron oxides-based painting materials in Gaya cultural circles [16,17,18]. Taking into account micromorphological and chemical evidence, XRD, Raman, and Mössbauer spectroscopic techniques, we are able to further optimize the interpretation of iron oxides composition on various types of red pigment. Moreover, these analyzed results propose how each type of iron-oxides based red pigment were manufactured.

The elemental distribution and mineral phase of red particles in the type 1 and type 2 showed that silicate minerals, such as quartz, alumino-silicate and phyllosilicate (including clay minerals), are the main component. Such abundance of silicates can also be explained by a lower quantity and dispersed iron oxide distribution. These analysis results indicate that the type 1 and type 2 were prepared using natural soil as a raw material. However, these two ochre-based pigments involved remarkably different magnetic behavior of iron oxides. It is clearly indicated that the spectral area of sextet relative to the central doublet appear higher proportion in the brown ochre (type 2) compared to the orange ochre (type 1). These results may correspond to an ultrafine-sized and poorly crystallized hematite and/or Fe3+ ions hidden by the high levels of silicate mineral in reddish particles in the type 1 [7, 21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37]. In contrast, the dominant sextet in the type 2, a fully magnetically split Mössbauer spectra at room temperature, represents a magnetically ordered component, possibly with weakly ferromagnetic properties, and indicating that well-crystalline hematite particles [23, 24, 26, 29, 38]. In addition, we note that the sextets in type 1 is assigned to magnetite, which is observed as black particles in Fig. 2 and iron rich single-domain particle in Fig. 3. The presence of such single-domain sized magnetite indicates that the type 1 used natural red ochre, due to this kind of mineral particles are usually inherited from the parent rocks, and/or occur under soil-forming conditions [39,40,41]. Furthermore, it is known to undergo the phase transition to hematite after heat treatment [42, 43]. Therefore, weakly ferromagnetic hematite in the type 2 indicating that sintering had taken place by thermal process [20, 26]. In addition, the higher ratio of Fe3 + than Fe2 + among the doublets assigned to chlorite group minerals (clinochlore) can also be interpreted as a result of heat treatment in an oxidized atmosphere [36]. These changes in the Mössbauer spectra reflect various phases of decomposition and transformation of the iron oxides.

In contrast to the above ocher-based pigments, the chemical and mineral composition of the type 3 and type 4 are composed of nearly pure hematite. It suggests the possibility that they were manufactured using high-grade massive iron ore or iron-containing materials rather than directly use of natural reddish ocher. In addition, more numerous and sharper shaped peaks in XRD diffractograms and Raman spectra of the samples type 3 and type 4, than those of two ochre-based pigments, correlate with the higher crystallinity of hematite [14, 42,43,44,45]. Moreover, we investigated the rounded polyhedron morphology of the hematite nano-particles, and their weak ferromagnetic properties in the room temperature Mössbauer spectra as shown in Figs. 2, 6. These results indicate that these two hematite-based pigments are synthetic painting materials which were thermally treated on purpose. In human history, this type of pigment is rarely reported before development of artificial ingredients of iron oxides, called Mars colours, nineteenth century [6,7,8, 14].

It is interesting comparing these two hematite-based pigments since they presented a huge variation in particle size and impurities, but the similar shade. The identified mineral composition show that they have been processed from a similar raw material with a high purity of iron components. However, their hematite characteristics indicate that the type 4 was calcined at a higher temperature than type 3, due to former has higher diameter, more rounded particles, and the larger spectral area of sextet related to weakly ferromagnetic properties than those of latter. In addition, another difference identified in the analysis results are the mix of small amount of soil as an extender to augment their volume and weight in the type 3. Considering sample description in Table 1, the two hematitic pigments, type 3 and type 4, were stored in the vessels, and has been identified as a block of red paste and a finely ground red pigment, respectively; in contrast to the two ochre-based pigments (type 1 and type 2) were discovered as painted status. Therefore, we can estimate that they have been stored as powder phase after manufactured by their own processes. Such storage in powder form of hematite can be assumed to preparation for future needs rather than a one-time storage for immediate use [7].

A detailed study of the thermally induced synthesis of hematite and its temperature tracking is an extensive project in itself. Because the experimental results can be flexibly varied depending on the starting raw material, processing method, firing environment, and many other parameters [13,14,15, 20, 26, 27, 42, 43, 46]. Moreover, there is no documentary evidence on high-purity iron materials and additives. Thus, here we have paid attention to those aspects of the morphology, grain size, and magnetic behaviors that are related to hematite crystallinity by thermal treatment.

As is well-known common knowledge, although iron oxides are extremely abundant in the surface soil [47], so far not one of such oxides is suitable for phase analysis as chromophores. Because they are composed of poorly crystalline iron oxides, including amorphous phases and ferrihydrite [9, 48]. Moreover, iron oxides formed in naturally weathered soils consist of less crystalline phases due to a presence of inhibitors of crystallization, especially silica species and organic matter, and hence their color strength is insufficient. Therefore, as low as a few percent of well-crystalline hematite admixtures change the hue of yellowish ochre to brownish or reddish as shown in Table 1 and Fig. 2, because the color strength of hematite is greater than that of amorphous phase of iron oxides with poor crystallinity [7, 8, 15].

In this regard, our results confirm that successful contribution of the characteristic information towards the magnetic behavior of each iron oxides by Mössbauer spectroscopy; which can be a powerful tool for the identification and characterization of iron-bearing minerals in combination with chemical and physical methods. Consequently, multiple-analyzing data in this study are indicate that the color distribution characteristic of the iron oxides contained red painting materials are affected by their purity and crystallinity of hematitic iron oxides. It is clearly show that the hematite-based pigments have higher redness (a*) compared to ochre-based pigments. In addition, a significant decrease in brightness (L*) is observed as the proportion of well crystalline hematite increases in iron oxides-based pigments. Finally, considering our whole data set of four types of samples, it can be explained that the ancient Gaya people made use of these characteristics of iron oxides, and produced, stored, and used various types of iron oxide-based red pigments for each purpose (i.e., such as natural red ochre, thermally induced red ochre, pure hematitic pigment and its mixed paste with extender materials).

Conclusions

In present study, it has been demonstrated that the raw materials and process of the four characteristic iron oxides-based red pigments in the ancient Gaya cultural circles. The analyzing results using multi-characterization techniques (SEM–EDS, XRD, Raman and Mössbauer spectroscopy), indicate that the natural red ochre, thermally induced red ochre, and hematite-based materials (including heat-treated pure synthetic hematite powder, and extender mixed paste) were used as pigment preparation. Moreover, the color distribution characteristics of iron oxides-based red pigments are affected by an increase in redness (a*) and a decrease in lightness (L*), as increase of the concentration, crystallinity, and weakly ferromagnetic properties of hematitic iron oxides by thermal process. These results suggest that the ancient Gaya people produced various types of iron oxides-based pigments according to purpose, by selective application of raw materials and thermal process.

A method for detailed identification of iron oxides-based pigments was successfully improved using the Mössbauer spectroscopy: this tool may be used in the future in the analysis of iron compound-based painting materials, with the aim of uniquely identifying the ancient iron oxides-based pigment, often referred to simply high- and/or low-purity iron oxide pigments.