Introduction

John of Görlitz (1370–1396) was a member of the House of Luxembourg and the only Duke of Görlitz (Zgorzelec) from 1377 until his death. He was the third son of Emperor Charles IV (1316–1378), Holy Roman Emperor and Bohemian King, from his fourth marriage with the princess Elizabeth of Pomerania (1347–1393). John of Görlitz was a younger half-brother of Wenceslaus IV, elected King of the Romans (1376–1400) and Charles’ IV successor on the Bohemian throne (1378–1419). The Duke of Görlitz died unexpectedly on March 1, 1396, in Neuzelle Monastery in Lower Lausitz. In the preceding evening he went to his bed, apparently healthy, and next morning was found dead. The cause of his death, which is one of the greatest mysteries of our history, is still unknown. The sudden death triggered speculations and rumours about him being poisoned, which was typical for similar cases in the middle ages. To obtain information about his health status, his ilium bone underwent recently bone histology (cf. Figure 1) [1] accompanied, among others, by histochemical staining reaction for Al using the aluminon technique (cf. Figure 2) [1, 2].

Fig. 1
figure 1

Duke‘s bone tissue shows complete mineralization of trabeculae using von Kossa´s method, magnification 210× [1] (with permission)

Full size image
Fig. 2
figure 2

Bone trabeculae with red stains (shown by arrow) indicating probably the aluminium deposits, magnification 380× [1] (with permission)

Full size image

The microscopic bone examination shows that the basic structure of both cortical and trabecular tissues is not disturbed, although a mild thinning of bone trabeculae can be seen, because not all parts of the trabeculae are connected. More interestingly, the histochemical staining reaction with aluminon (the triammonium salt of aurin tricarboxylic acid) [2] yielded linear red deposits on the surface of the trabeculae localized between their mineralized and unmineralized parts. This finding resembled aluminium osteopathy frequently seen in long-time dialysed patients [1]. Therefore, the primary aim of this study was to employ INAA for verification whether the red deposits are really due to an elevated Al content in the bone or whether they are due to elevated levels of other, interfering elements [2]. We were also looking for a number of other minor and trace elements, especially toxic ones, whose elevated levels could possibly explain Duke’s sudden death.

Experimental

Samples and calibrators

A slice of Duke´s bone sample was removed from a block of poly(methyl methacrylate) glass (PMMA) used for the bone histology examination. The major part of PMMA was removed mechanically, the remaining part by five times repeated dissolution in about 5 mL of chloroform. To remove the possible external contamination, the bone sample was washed for 5 min in warm 2 M HNO3 obtained by dilution of subboiled HNO3 followed by repeated washing with demineralized water till neutral reaction, and dried at 60 °C for 12 h. The bone mass amounted to 96.23 mg. A piece of mechanically removed PMMA after the above described washing procedure was analysed as a process blank. Its mass was 153.02 mg. For quality control purposes, NIST SRM 1486 Bone Meal with mass 84.76 mg was also analysed in the ‘as received’ state. The samples were packed for irradiation into pre-cleaned polyethylene (PE) disk-shaped irradiation capsules with 25 mm and 20 mm diameters for irradiation with thermal and epithermal neutrons, respectively, made by sealing PE foils of 0.2 mm thickness. For relative standardization, multielement calibrators were prepared by pipetting of weighted aliquots of certified element solution (Analytika®, Ltd., Czech Republic) onto a disk of chromatographic paper Whatman 1 with a 16-mm diameter, drying and sealing into the above PE irradiation capsules as the samples.

Irradiation and counting

The samples and multielement calibrators were irradiated in the LVR-15 reactor of the Research Centre Řež [3], where neutron flux densities of thermal, epithermal and fast neutrons of 3.1 × 1013 cm−2 s−1, 1.1 × 1013 cm−2 s−1 and 5.8 × 1012 cm−2 s−1, respectively, are available in a vertical channel for short-time irradiation behind Be reflector, while the above flux densities of 3.6 × 1013, 8.4 × 1012 and 8.6 × 1012 cm−2 s−1, respectively, are available in a vertical channel in a Be reflector for long-time irradiation. The channel for short-time irradiation is connected with the laboratory equipped with gamma-spectrometer with a recently modernized fast pneumatic (rabbit) system with transport time of 3.5 s [4]. For aluminium determination by short-time irradiation, epithermal neutron activation analysis (ENAA) was also used to be able to correct for the 31P(n,α)28Al interfering reaction with fast neutrons. For this purpose, the samples were irradiated with epicadmium neutrons in a cylindrical Cd box with internal diameter of 22 mm, 1 mm thick walls, and height 7 mm, which can be accommodated in the transport rabbit [4].

Table 1 summarizes the employed irradiation, decay and counting times (ti, td, tc, respectively), HPGe coaxial detectors used, and counting geometry as distance of the sample from the top of the detector during the measurement (Geo).

Table 1 ENAA and INAA experimental parameters
Full size table

For short-time irradiation, the samples, calibrators and one blank PE irradiation capsule were irradiated individually in the PE rabbit together with Al-0.1% Au alloy monitors (Certified nuclear reference material IRMM-530a) to check stability of the neutron flux density in time. In the case of long-time irradiation, the samples, calibrators and the blank PE capsule were stacked together to form a column that was inserted into an aluminium irradiation can. Three iron monitors were placed on the top, in the middle and at the bottom of the stack to determine the axial neutron flux density gradient. Delays of several days between the individual SI and LI runs were used to allow for decay of induced activities of short-lived radionuclides in the samples. An optimal irradiation sequence would be SI-ENAA and INAA to assess the 31P(n,α)28Al interfering reaction (cf. first line in Table 1) followed by SI-INAA and LI-INAA to determine minor and trace elements in the samples. Since, however, the facility for short-time irradiation was unable to irradiate samples in Cd capsules during system modernization [4] in 2023, the SI irradiation runs for assessment of the 31P(n,α)28Al interfering reaction were carried out with delay 9 months after the LI-INAA.

Gamma-ray spectrometry measurements of the induced radionuclides were performed using the above specified coaxial HPGe detectors (cf. Table 1) interfaced to a Canberra Genie computer controlled gamma-spectrometry analyser through a chain of associated linear electronics, which included a Canberra 599 loss free counting module to correct the variable count-rate and dead time. Canberra Genie™ 2000 spectrometry software was used for evaluation of gamma-ray spectra. Nuclear parameters of radionuclides measured have already been given earlier [5].

Interference correction in Al determination

The analytical radionuclide 28Al is produced not only by the 27Al(n,γ)28Al reaction with thermal neutrons, but also by the interfering reaction 31P(n,α)28Al with fast (epicadmium) neutrons, which causes a significant increase of an apparent Al content in samples with a high phosphorus content, such as bone. We have chosen an experimental arrangement that allows a simple interference correction calculation as follows. After SI-ENAA, INAA irradiations, the samples were measured using the same decay and counting times in the same counting geometry. In these experimental conditions, the true number of pulses for the calculation of an unbiased Al content, N(28Al)true can be approximated by Eq. (1)

$$ N\left( {^{28} Al} \right)_{{{\text{true}}}} = N\left( {^{28} Al} \right)_{{{\text{whole}}}} - N\left( {^{28} Al} \right)_{{{\text{epi}}}} $$
(1)

where N(28Al)whole is the number of counts measured after sample irradiation with the whole reactor neutron spectrum and N(28Al)epi is the number of counts measured after sample irradiation with epicadmium (fast) neutrons. This approach takes advantage of the drastic reduction of the (n,γ) reaction rate in the low neutron energy region up to the cadmium cut-off energy ECd = 0.55 eV, although the cross section of the (n,γ) reaction at higher neutron energies is not negligible. Thus, this method is applicable for samples with not excessively high Al contents, such as uncontaminated bone. An alternative approach could be based on using an interference correction factor of the 31P(n,α)28Al reaction, which was found to be 10.45 µg g−1 of apparent Al per 1% of P and knowledge of a P content in Duke´s bone. Since the determination of P is inaccessible by INAA, we would need to use the literature value for modern populations given as a range of 7.1–12.5% without any specification whether it concerns bone ash, dry mass or wet mass [6]. Obviously, this would lead to a large uncertainty of the interference correction.

Results and discussion

Table 2 shows that in total thirty-two elements were assayed by INAA in the samples analysed. It is obvious that PMMA contains very low contents of the elements studied, which suggests that no contamination of Duke´s bone would be caused even in the case of incomplete PMMA removal, which we believe did not occur. Our results for NIST SRM 1486 agree within the stated uncertainties with the NIST certified element mass fractions and NIST information values, where available, except for a slightly lower Ca content and slightly higher Na content. Thus, the accuracy of our results has been proven. In Table 2, there is also a comparison of our INAA results for Duke´s bone with literature data [7] for populations from the twentieth century. It can be seen that the Al content in Duke´s bone is within the literature range. This means that the red deposits seen in histochemical staining reaction (cf. Figure 2) are not due to an elevated Al content in the bone. It is known that aluminon is not a specific reagent for the Al detection, but similar red stains can also be due to the interaction of aluminon with various metal species, namely Be(II), Sc(III), V(II), Cr(III), Fe(III), Ga(III), In(III), Tl(III), some alkaline earths and rare earth [2]. Of these, Be, Ga, and In were not determined in this work, other elements were within the reported literature ranges. Thus, the reason for the positive staining reaction remains unknown.

Table 2 INAA results in µg g−1 and their uncertaintiesa, unless otherwise stated (dry mass)
Full size table

Noteworthy, we found significantly higher values for Mn, As, Sb, and especially for Ag in Duke´s bone compared with the literature values [7]. Such high values would suggest adverse effects on Duke´s health (as discussed below), provided that they are due to exposure to harmful substances, probably by ingestion. It should also be pointed out that the high values of the above elements suggest exposure for several years, because the mean turnover for cancellous (trabecular) bone is 17.7% per year, which means that the mean bone lifespan is approximately 5.6 years [9].

The most elevated level was found for silver. Although it is known that the element and its compounds are active agents for purification and disinfection of drinking water and have been used until recently for treatment of skin wounds, including burns [1, 10], there are also adverse effects of silver on animal and human health. One such effect is attributable to long-term exposure to silver and silver compounds, which causes a grey or blue-grey discolouring of the skin (argyria). It may be caused by the photoreduction of silver chloride to metallic silver. The metallic silver is then oxidized by tissue and bound as black silver sulphide [10]. However, there is no report concerning such Duke’s discolouration. Toxic effects of acute exposure to silver are very rare. Digestion of water-soluble silver nitrate can lead to blood hypotension resulting in a collapse of blood circulation. Moreover, silver nitrate can be transformed by bacteria, e.g., at the gastrointestinal tract inflammation, to nitrite, which reacts with hemoglobin to methemoglobin resulting in cyanosis and death due to hypoxia [1, 10, 11]. The acute intoxication with silver thus appears as one possibility how to explain Duke’s sudden death. However, the acute intoxication should be preceded by a long-term exposure to Ag, otherwise content of this element in Duke´s bone would not be that high as found in our study. Concerning the long-term exposure, it can also be speculated whether it could be due to the use of silver tableware. However, dining on silver (or even on gold) was not used by the nobility of the fourteenth century in our part of Europe so frequently. It became more popular in the Renaissance period and later. We also have not found any reports on harmful effects of dining on the above precious metals.

Manganese is an essential trace element for humans that is found at low levels in almost all diets. Manganese ingestion represents the principal route of human non-occupational exposure. Humans generally maintain stable tissue manganese levels due to homeostatic mechanisms [12]. However, high-dose exposures are associated with increased tissue manganese levels, causing mainly neurological disorders with a motor dysfunction syndrome, commonly referred to as manganism or Parkinsonism-like disease [13]. Neuropathological changes are detectable in the basal ganglia of the brain of humans suffering from manganism [14]. It is known that increased manganese levels in drinking water caused a number of neurological disorders, such as those occurring in West Bengal and Bangladesh [15], where there are also severe problems with arsenic contamination in groundwater and consequent drinking contaminated water from deep wells [16]. It has also been reported that chronic intake of drinking water containing elevated levels of manganese caused an increased occurrence of neurological signs in the elderly residents of two small towns in Greece [17]. However, it seems that most adverse health effects of manganese are due chronic intoxication and could hardly be attributed to Duke’s sudden death, unless there was a synergy with exposure to other toxic elements or compounds.

Arsenic exhibits a number of severe adverse health effects in humans. These effects are species dependent. It is well known that inorganic arsenic species (As(III), As(V)) are much more toxic than organic ones (monomethylarsonic acid-MMA, dimethylarsinic acid-DMA, arsenocholin, arsenobetaine, arsenosugars, etc.). The most toxic species (As(III)) is transformed in organisms to MMA and DMA by methylation. The most important exposure route is through the intake of contaminated water and food. There are large populations chronically exposed to arsenic in some countries, mainly in Bangladesh, Chile, India, Mexico, Taiwan, United States of America and Uruguay, where the groundwater contains high amounts of inorganic arsenic above the WHO recommended limit of 10 μg L−1 [18]. Long-term exposure to inorganic arsenic can lead to chronic arsenic poisoning (arsenicosis), which is mostly manifested by skin lesions and several types of cancer. However, only intake of high doses of inorganic arsenic can lead to acute poisoning followed by death [19]. We will not discuss any further the possibility of Duke’s poisoning by arsenic as the reason of his sudden death, because bone analysis cannot provide a proof of acute exposure, and moreover the literature values of arsenic are based on the element content in bone ash [7], which can be negative biased due to volatilization losses.

There are many similarities in some chemical and toxicological properties of arsenic and antimony, although the latter element does not undergo methylation so easily as arsenic. However, there is little information about antimony to evaluate its toxicology and determine its impact on the human health [20]. Human exposure to antimony mostly occurs in workplace settings [21]. It is also known that antimony and its compounds have been used to treat a variety of illnesses over centuries, especially using Paracelsian alchemical medicines [22], which, however, were not commonly used in Duke’s times.

Thus, it appears that the elevated levels of Ag, As, Mn, and Sb in Duke’s bone can hardly explain his sudden death for several reasons. One concerns the comparison with the available literature values, because the authors of the lastly published review stated that ‘reliable chemical compositional data of bones and teeth, particularly for minor and trace elements, are scarce’ [7]. In our study this especially applies to arsenic, the value of which was derived from analysis of bone ash. Another one concerns the inability of bone to detect acute exposure due to the long turnover (remodelling) time of this tissue [9]. Further reason is that post-mortem contamination of Duke’s skeleton with the above elements cannot be excluded. To illustrate possibilities of post-mortem contamination, it can be mentioned that Rasmussen et al. found highly elevated manganese contents in various parts of skeleton of two individuals from medieval Denmark buried (in the ground) in the Franciscan Friary in Svenborg, which were attributed to diagenesis processes [23]. Unfortunately, nothing is known not only about Duke’s health status and life style shortly before his death, neither the circumstances of his funeral and burial are known [1], namely whether his remains were somehow treated (embalmed) and when they arrived in the place of his eternal rest in a tomb at St. Vitus cathedral in Prague, where other members of the House of Luxembourg, and many Bohemian kings and Holy Roman Emperors are buried. Thus, the reason of John’s of Görlitz sudden death remains shrouded in mysteries of the Bohemian history.

Conclusions

This work demonstrated that INAA is a useful tool for elemental characterization of bones of historical personalities in cultural heritage studies. In these cases, the bone elemental composition can help to reveal the health status of individuals, especially in cases where exposure to inorganic harmful substances is suspected. Thus, INAA can be considered as a valuable supplement to traditionally used examinations, such as bone histology and histochemical staining techniques. In the case of suspected elevated aluminium based on positive reaction with aluminon, we have also confirmed that this staining technique needs verification by a reliable analytical method, because its lack the required specificity.

On the hand some difficulties of bone analysis aimed at assessment of the health status and/or exposure to harmful substances of examined subjects should be mentioned. When a decision is to be taken whether element contents in bones are elevated, we have to compare them with literature data for healthy contemporary populations. However, reliable data are missing for some elements so far, as was the case of arsenic in our work. An alternative approach consisting in a comparison with elemental composition of bones obtained from a cohort of living or accidentally deceased individuals is complicated due to medico-legal implications.

Another restriction of bone elemental characterization is due to the bone turnover (remodelling) time, which spans from several years to tens of years, depending on the bone type. This means that the bone composition can be an excellent biomarker of long-term, chronic exposure to harmful substances, but it can hardly detect acute exposure, e.g., due to poisoning.

Possibilities of contamination of fossil and archaeological bones should also be considered in cultural heritage studies. The post-mortem preservation and contamination of the bone tissue depend upon a number of circumstances and processes, a type of burial and the possibility of embalming among others [23, 24]. Contamination of the surface of cortical bone can be removed mechanically easily, e.g., by brushing or drilling. The reverse is true for much more porous trabecular tissue, which may contain diagenetic deposits [25].

Nevertheless, despite of the above complications, INAA of historical bones provides a precious tool for multi-elemental assay to reveal information about the health status and possible long-time exposure to harmful substances, especially in cases when other, possibly more suitable biomarker tissues have already decayed. In the future, the results of the present work can help after all in elucidating the mystery of Duke’s sudden death, provided that the missing data of his anamnesis are discovered.