Insights into the Heterogeneity of the Mercury Isotopic Fingerprint of the Idrija Mine (Slovenia)

: To determine the range of the isotopic composition of mercury (Hg) from the Idrija mine, samples from the mine itself and from the Mercury Heritage Management Centre geological collection were analyzed. Samples from various geological periods, genesis types, ore types, formations, and excavation ﬁelds and levels were analyzed. Both Hg concentration and isotopic composition were measured. The δ 202 Hg ranged from − 1.35‰ to 0.46‰, and the ∆ 199 Hg ranged from − 0.18‰ to 0.16‰. A relatively homogenous ore ﬁngerprint was obtained from one of the excavation ﬁelds; otherwise, the isotopic ﬁngerprint of the Idrija mine seems to be heterogenous. This study presents the ﬁrst statistically robust constraints on the isotopic composition of Hg from the Idrija mine, which may help in further studies of the isotopic composition of similar ore bodies or the potential tracing of Hg from the mine to the environment in the vicinity or downstream of the mine.


Introduction
Tracking the pathways of mercury (Hg), a toxic trace element monitored by the United Nations Environmental Program, is of great importance [1][2][3][4].Part of the terrestrial cycle of Hg [5,6] is represented by legacy mining sites.One such site is located in Idrija, Slovenia, where Hg ore has been mined and processed over the last five centuries.In the last century, the Idrija Hg mine was the subject of intensive research by geologists [7][8][9][10][11][12].Extensive research has shown that the elevated Hg content in this area is due to: (I) outcrops of Hg-bearing geological formations [7,13], (II) mining and processing activities [14,15], (III) the re-emission of mine tailings and ore deposits [15,16], and (IV) the re-emission from soils and biota contaminated by sources I-III [15,17,18].
The ore genesis is likely best summarized by Berce [7].The first Hg impregnation of rock formations occurred during rifting in the Triassic period and was accompanied by the impregnation of rocks with barite, fluor, copper, manganese, strontium, and zinc.The Hg-bearing chlorine vapors came from deep beneath the crust.On their way to the surface, they likely dissolved the pyrite present in the carbonates, which introduced sulphur (S).The vapors also experienced a drop in pressure and temperature.These conditions were favorable for the precipitation of cinnabar (HgS).The impregnation process took a relatively long time and occurred under various conditions, depending on which forms of HgS were formed.
In the so-called Jeklenka ore, with its black steel-like hew, the cinnabar crystals are relatively small, and concentrations of Hg reach up to 78% in some sources [19].In the more reddish-colored Opekovka ore, the crystals are larger, but the Hg concentrations are lower.In some cases, the ore was deposited syngenetically with sedimentary layers, forming banded ore, Opekovka, Jeklenka, and Coral ores, while in others, it was deposited epigenetically, directly into the already-formed rock formation, creating impregnations or substitutions [7,12].Liquid Hg (Hg0 (l) ) is also present and thought to have originated from two sources: (I) the primary direct deposition of Hg during primary impregnation due to a lack of sulfur ions that would bind it to HgS or (II) subsequent remobilization of the primary HgS due to leaching by the pore water, which resulted from the tectonic fracturing in the area [7,11,12].
The emergence of new analytical technologies that enable the determination of the ratios between the stable isotopes of Hg has allowed new possibilities for isotopic fingerprinting and provenience determination [20].This is because Hg isotopes fractionate in both mass-dependent and mass-independent ways.Mass-dependent fractionation (MDF) is mostly the result of reactions such as evaporation, thiol-ligand binding, microbial methylation and reduction, and iron oxide sorption, while mass-independent fractionation (MIF) is mostly the result of photochemical reactions [21,22].MDF is expressed as a ratio between a chosen isotope pair, while MIF is quantified by comparing the measured MDF with the one predicted using the measured MDF value and the kinetic MDF law established through transition state theory.When the variations are minor, typically less than about 10‰, these values can be approximated [21].There are also two types of MIF that are observed.One is the odd-MIF, affecting the odd isotopes of Hg, and the other is even-MIF, affecting the even isotopes of Hg.Odd-MIF stems from the magnetic isotope effect, which takes place in photochemical radical pair reactions, while even-MIF is more cryptic and believed to stem from the nuclear self-shielding effect [21].Ores exhibit a wide range of MDF, sometimes with some odd-MIF and no even-MIF [21,[23][24][25].
The isotopic fingerprinting of Hg likely originating from mining areas has been employed in various settings.For example, the origin of Hg from mines and production facilities has been determined in the Almadén mine in Spain; the New Idrija mine; the New Almadén Hg mine in the US [23,24,[26][27][28][29][30], and the Wanshan mining region in China [31,32].In the Idrija region, few articles have been published on mercury isotopes [18,33], and only two samples of Hg ore have been measured thus far [34].To our knowledge, no study has attempted to significantly constrain the isotopic fingerprint of an excavation field, an ore type, a formation, or a geological period.This information could aid in the further determination of the isotopic fingerprint of Hg excavated during different periods of operation at the Idrija mine.For this purpose, a sampling of the ores from the Idrija mine was carried out.It was expected that potential correlations might be uncovered which would aid in the Hg ore deposition description and the characterization of the isotopic fingerprint of the ores mined at a certain period of the mine's operation.

Sampling
Samples were collected in the summer of 2022 from the still-accessible levels of the former Hg mine in Idrija.Samples were excavated using a hammer and collected in ziplock bags.The Hg (l) , which was in small puddles on the rocks, was scooped from the rocks using a plastic Pasteur pipette.Today, only three of the mine's 15 upper levels are accessible; therefore, most of the samples were taken from the geological collection of the Idrija Mercury Heritage Management Centre.As these were museum exponates, only small pieces could be broken off or sawed off with a circular saw, as was attempted with the more compact pieces.In total, about 5 g of each single sample was collected.The types of samples are given in Table 1.These samples include ore types and formations.The names of the samples were preserved as they are found in the geological collection and are the same as those commonly used in the previous literature about Idrija ore deposits and by the miners during the period of the active Hg excavation.

Carbone shale
Carbone shales are the oldest rocks of the Idrija ore deposit.They are composed of mud, clays, silt, and silicate sand with occasional inclusions of conglomerates, which are interchanged rapidly.Afterwards, the tectonic processes shape the shales into various bodies with the common name of Carbone shales.These rocks are gray to dark gray in color and very soft, which makes them difficult to excavate.Carbone shales include Hg (l) and cinnabar, with the ratio between the two commonly being 1:1.The Hg content reached up to 10% of Hg, but in most cases, it was around 0.3%.

Impregnated cinnabar
Impregnated cinnabar is represented by pyrite (FeS 2 ) and cinnabar (HgS) concretions in mudstone that were impregnated with Hg (l) .The concretion is composed of pyrite, with mostly idiomorphic cryosections substituted with HgS.A detailed description is given by Mlakar and Drovenik [11].
Opekovka ore Opekovka, or brick, ore was named by the miners.Opekovka ore is relatively common in the mine but is usually found in rather small deposits.
Jeklenka ore Jeklenka, or steel, ore is a gaelic cinnabar type of ore composed of a cinnabar rich in organic matter.It has a fluidic structure and a metallic gray hue.
Impregnated dolomite Dolomite CaMg(CO 3 ) 2 impregnated with Hg, mostly as cinnabar, which fills in the pores in dolomite and, in some cases, substitutes for the original rock.

Skonca beds
A geological formation containing both cinnabar and Hg (l) .A detailed description is offered by Čar [19] Hg (l) Elemental Hg 0 in liquid form.

Tufa
Tufa is composed of particles of silicate (SiO 2 ), claystone, and plankton radiolites and some other particles.The matrix is mostly chalcedony (SiO 2 , trigonal) and pyrite.In the polished thin sections, a gradual granularity can be observed.The color ranges from green, in Tufa that is poorer in Hg, to the red, in Tufa that is richer in Hg.

Banded ore
Banded ore presents an interchanging sequence of layers of mudstone with cinnabar-impregnated chalcedony.Many infield fractures with cinnabar are observed.The chalcedony impregnated with cinnabar has the distinct color of Opekovka ore.

Jetrenka ore
Jetrenka, or liver, ore is syngenetically formed Hg in bituminous shale or, in some cases, a bituminous radiolarite.Jetrenka ores are found next to slip-faults where the organic matter mixed with cinnabar got smeared over the fault and produced a distinct liver-like color.
Conglomerate ore Conglomerate ore consists of unsorted, differently sized rounded clasts in a sand-sized dolomite matrix.It contains veins and impregnations of cinnabar, which primarily substitutes for the matrix.

Coral ore
Coral ore is a syngenetic ore represented by black bituminous silicate sandstone that is rich in brachiopod shales.Miners mistook the brachiopod shells for corals, giving this ore its name.It is a part of the Skonca beds.

Karoli ore
Karoli ores are among the most unique ores of the deposit, as they are only found in the deepest part of the mine of the same name.The genesis of this type of ore has not been sufficiently studied.It is, however, known that it is an epigenetic ore with strong pyrite mineralization, which was later crushed and substituted with cinnabar.

Sample Preparation
The samples were crushed in an agate mortar.Rock samples were digested using the Ethos microwave digestion system (Milestone, Sorisole, Italy).Approximately 0.2 g of the sample was weighed into pre-cleaned polytetrafluoroethylene tubes.For digestion, 5 mL of 65% HNO 3 (nitric acid), 2 mL of 40% HF (hydrofluoric acid), 1 mL of 30% HCl (hydrochloric acid), and 1 mL of 30% H 2 O 2 (hydrogen peroxide), all supra-pure grade, were used.The digestion process included 30 min for ramping up to the maximum temperature of 150 • C at a maximum power of 1200 W. This temperature was maintained for an additional 30 min, and cooling to room temperature took 1.5 h.The digestate was quantitatively transferred into 30-mL polyethylene tubes by rinsing them with 10 mL of Milli-Q water.Here, 10 mL of 5% H 2 BO 3 (boric acid) were added to neutralize the HF.The digestion products of the ore samples were then diluted 100-fold.The digestion of the Hg (l) samples was performed using with the same reagent mix but without microwave assistance.It was subsequently diluted 1000-fold.In both cases, the dilutions were carried out with 5% HNO 3 .

Analysis
The Hg concentration was determined using a Model Hg-201 semiautomated mercury analyzer and cold vapor atomic absorption spectroscopy (CV-AAS: Sanso Seisakusho Co., Tokyo, Japan).Quality assurance and control were evaluated using a standard reference material, NIST 2711 Montana Soil II (the closest available analogue to the geological material) [35].The average recovery for Hg was 101.2%, with relative standard deviation of ±5.6% (number of samples (N) was 6).
The isotopic ratios of Hg were analyzed using a Nu Plasma II (Nu Instruments, Wrexham, UK) Multicollector Inductively Coupled Plasma Mass Spectrometer (MC-ICP-MS).The details of the operating conditions are given in Table 2. Briefly, the sample solution was pumped via a peristaltic pump into a T split, where it was mixed with SnCl 2 in solution (3% w/v) to reduce the Hg 2+ in the solution to Hg 0 in the gaseous phase.The mix of solutions was transported to the phase separator, where Ar gas carried the produced Hg 0 to the MC-ICP-MS.The measurements were then performed at Hg concentrations between 1 and 1.5 ng/mL.The signal intensities at an m/z of 202 were between 1 and 5 V, depending on the measurement session.Blank measurements were constantly monitored and presented no more than 1% to 3% of the signal for an individual sample measurement.
Measurements were performed according to the sample-standard-sample bracketing method [35], using the NIST 3133 standard [36].The deviation between the sample and the standard voltage was always less than 10%.The results for isotopic composition were denoted by δ for MDF and ∆ for MIF (Equations ( 1) and ( 2)), where f represents the correction factor (0.2520 for ∆ 199 Hg, 0.5024 for ∆ 200 Hg, 0.7520 for ∆ 201 Hg, and 1.4930 for ∆ 204 Hg) and xxx Hg represents the mass number of a particular mercury isotope [21,37].
Accuracy is reported as the standard deviation of repeated NIST 8610 (UM-Almadén) measurements (Table 3) [38].The standard deviations for the isotope values are given with the expanded factor (k = 2).Additionally, NIST 2711a (Montana soil) standard reference material was used, as it represented the closest match to the samples used here [39].

Results and Discussion
In the mine samples, a range of almost 2‰ was observed for δ 202 Hg (from −1.35‰ to 0.53‰), while ∆ 199 Hg ranged by about 0.4‰ (from −0.16‰ to 0.18‰).The average δ 202 Hg was −0.46‰ (±1.10‰ 2SD), the average ∆ 199 Hg was 0.01 (±0.17‰ 2SD), and the average Hg content (c(Hg)) was 23.9%.No statistically significant even-MIF was observed in these samples, as expected for geological material [21].The sample type, with the collection identification number, when applicable; the individual periods from which particular samples were obtained; the excavation field and level; the isotopic composition and Hg content, are presented in Table 4.
The range of MDF in ore samples (a single Opekovka sample and a single Jeklenka sample) previously reported for Idrija is much more constrained, at 0.49‰ for δ 202 Hg (−0.26‰ to 0.23‰) [34].The range of δ 202 Hg in some other mines was 0.25‰ (−0.09‰ to 0.16‰) for the New Idrija mine (USA) [26], 0.13‰ (−0.57‰ to −0.70‰) for McDermitt cinnabar [23], 1.07‰ (−0.92‰ to 0.15‰) for the Almadén mine (Spain) [24], 1.29‰ (−2.25‰ to −0.96‰) for Monte Amiata (Italy) [25], and 4.09‰ (−2.70‰ to 1.39‰) for the Terlingua District (USA) [23].These high values in the Terlingua District are reportedly due to the fact that various hydrothermally deposited Hg-bearing minerals (montroydrite, kleinite, terlinguaite, metacinnabar, and calomel) were measured.It has been suggested that fractionation in the Terlingua District was caused either by isotopic fractionation during mineral formation due to different formation temperatures, paragenesis or variable redox states or bond strengths between the different minerals [23,40], and, in the case of Hg (l) -containing minerals, by the different proportions between the mineral-bound and liquid phases [23].Similarly, the ∆ 199 Hg ranges of the Terlingua District are more restricted as compared to δ 202 Hg and reach 0.46‰ [24].The results presented here bring the isotopic composition of Idrija close to that of the other well-documented hydrothermal Hg deposits.This indicates that the isotopic composition of Idrija ores cannot be considered homogenous.Table 4. Type of sample, with the museum identification, when applicable; the geologic period from which the sample originates; its Hg isotopic fingerprint, and the Hg content.N.a.stands for not available and was used for any samples for which we had no data in that category.The samples from the museum have the museum identification number written next to them.Plots of δ 202 Hg/∆ 199 Hg and ∆ 199 Hg/∆ 201 Hg are shown in Figure 1.They can be used both in comparison of different trends in data or potential groupings.In general, a trend towards more negative values was observed in ∆ 199 Hg with increasing δ 202 Hg (Figure 1A).This is the opposite of the trend for the minerals from the Terlingua District and Almadén, which show more positive values in both δ 202 Hg and ∆ 199 Hg [23,24].However, the data from Mt. Amiata show a similar trend.The trend at Mt. Amiata was statistically weak and was not discussed further by the authors [25].In addition, results derived from roasted calcines and sediment near the New Idrija mine [29], the Yuba River below the New Idrija mine [27], and contaminated pore waters from Tennessee [41] also show a similar trend to that revealed in this study.Some MIF could be caused by the interactions with the surface [40], which might potentially shift the fingerprint differently in different cases.At this stage, there are not enough analogous results to draw a conclusion.

Type of Sample and Museum Identification
to that revealed in this study.Some MIF could be caused by the interactions with the surface [40], which might potentially shift the fingerprint differently in different cases.At this stage, there are not enough analogous results to draw a conclusion.
∆ 199 Hg/∆ 201 Hg slopes can be a useful tool in Hg isotope studies, as they clearly present any deviation between the two odd-MIF values, which can be present.The slope of the ∆ 199 Hg/∆ 201 Hg presents a regression line of 1.23 (Figure 1B), which is between the slopes of the Terlingua District minerals, with a slope of 1.66 [23], and Mt.Amiata, with a slope of 0.68 and an R 2 of 0.13.It is known from laboratory experiments that the photochemical reduction of Hg 2+ yields residual Hg 2+ , with a ∆ 199 Hg/∆ 201 Hg ratio of 1.00, and the photodegradation of methyl-mercury (MeHg) yields residual MeHg, with a ∆ 199 Hg/∆ 201 Hg ratio of 1.36 [22].∆ 199 Hg/∆ 201 Hg slopes can be a useful tool in Hg isotope studies, as they clearly present any deviation between the two odd-MIF values, which can be present.The slope of the ∆ 199 Hg/∆ 201 Hg presents a regression line of 1.23 (Figure 1B), which is between the slopes of the Terlingua District minerals, with a slope of 1.66 [23], and Mt.Amiata, with a slope of 0.68 and an R 2 of 0.13.It is known from laboratory experiments that the photochemical reduction of Hg 2+ yields residual Hg 2+ , with a ∆ 199 Hg/∆ 201 Hg ratio of 1.00, and the photodegradation of methyl-mercury (MeHg) yields residual MeHg, with a ∆ 199 Hg/∆ 201 Hg ratio of 1.36 [22].
The plots of cHg/δ 202 Hg (Figure 2A) and cHg/∆ 199 Hg (Figure 2B) show that higher concentrations (above 0.1%) show no relationship with the isotopic fingerprint.However, below 0.1%, the isotopic fingerprint shows relatively more positive δ 202 Hg and more negative ∆ 199 Hg values.It was assumed that there could be a potential correlation between either MDF or MIF with cHg.The plots of cHg/δ 202 Hg (Figure 2A) and cHg/∆ 199 Hg (Figure 2B) show that higher concentrations (above 0.1%) show no relationship with the isotopic fingerprint.However, below 0.1%, the isotopic fingerprint shows relatively more positive δ 202 Hg and more negative ∆ 199 Hg values.It was assumed that there could be a potential correlation between either MDF or MIF with cHg.There are no distinct groups formed when we take into account the time periods of the samples.This could be because most of the samples are from the Ladinian period.However, if we group the samples according to type and excavation site, some trends can be identified.The samples of Hg(l), Skonca beds, and Tufa, all of which are derived from the same excavation site (Kropač), are grouped together on the δ 202 Hg/∆ 199 Hg plot (Figure 1A).These samples were taken close to one another (in a single shaft, about 10 m apart).The similarity between the Skonca beds and Tufa with the Hg(l) fingerprint indicates that either there is no major fractionation occurring when Hg(l) is leached or that the original pool of Hg from which both groups are derived had a uniform Hg fingerprint.The pool size, which is unknown in this case, might play an important role as a proportionally small but severely fractionated leachate and would not change the overall fingerprint.
At the other excavation sites, fewer samples were measured, and therefore, they cannot be considered significant, or they had a very broad fingerprint.These include, for example, impregnated cinnabar from the Karbon excavation site, with a range greater than 1‰ (δ 202 Hg), and impregnated dolomite and carbonate shale from the Kreda excavation site, with a range of 2‰ (δ 202 Hg; Figure 1A).
Overall, it appears that the isotopic fingerprint of a single excavation site may be homogenous, but this is by no means true for all excavation sites.The results presented here suggest not only that the isotopic fingerprint of the entire mine cannot be constrained below a few permilles but also that an individual excavation site can have a wide range There are no distinct groups formed when we take into account the time periods of the samples.This could be because most of the samples are from the Ladinian period.However, if we group the samples according to type and excavation site, some trends can be identified.The samples of Hg (l) , Skonca beds, and Tufa, all of which are derived from the same excavation site (Kropač), are grouped together on the δ 202 Hg/∆ 199 Hg plot (Figure 1A).These samples were taken close to one another (in a single shaft, about 10 m apart).The similarity between the Skonca beds and Tufa with the Hg (l) fingerprint indicates that either there is no major fractionation occurring when Hg (l) is leached or that the original pool of Hg from which both groups are derived had a uniform Hg fingerprint.The pool size, which is unknown in this case, might play an important role as a proportionally small but severely fractionated leachate and would not change the overall fingerprint.
At the other excavation sites, fewer samples were measured, and therefore, they cannot be considered significant, or they had a very broad fingerprint.These include, for example, impregnated cinnabar from the Karbon excavation site, with a range greater than 1‰ (δ 202 Hg), and impregnated dolomite and carbonate shale from the Kreda excavation site, with a range of 2‰ (δ 202 Hg; Figure 1A).Overall, it appears that the isotopic fingerprint of a single excavation site may be homogenous, but this is by no means true for all excavation sites.The results presented here suggest not only that the isotopic fingerprint of the entire mine cannot be constrained below a few permilles but also that an individual excavation site can have a wide range of isotopic fingerprints.It appears that the constraints placed on the isotopic fingerprint by sample type and period are also very broad or statistically weak.

Conclusions
Although attempts to constrain the isotopic fingerprint of different periods of Hg deposition have not been informative beyond the mere range of end members, some conclusions with respect to the different sample types or excavation sites can be drawn: 1.
The isotopic fingerprint of the Idrija mine is generally relatively broad as comparable to most other findings, with the MDF fingerprint (δ 202 Hg) ranging from −1.35‰ to 0.46‰.

2.
The MIF ranges are similar to those of other mines, with ∆ 199 Hg ranging from −0.16‰ to 0.18‰.3.
Only one example of a relatively homogenous fingerprint was observed at the Kropač excavation field.4.
The isotopic fingerprints of cinnabar and Hg (l) from the same excavation site statistically indistinguishable.This study's most valuable aspect seems not to be the tracing of Hg in the environment or to the Gulf of Trieste, but the description of a wide variety of samples in the isotopic fingerprint of Hg found in the mine.This could be useful for other researchers, either in designing studies of complex Hg deposits elsewhere around the world or for those specifically looking into the Idrija deposit.

Figure 1 .
Figure 1.The plot of δ 202 Hg versus ∆ 199 Hg with a table of different types of excavation, as well as the labels of different types of ore (A) and ∆ 199 Hg versus ∆ 201 Hg (B).The regression lines for both observed data and the slope of 1 and 1.36 as observed for photoreduction of MeHg[22] for odd-MIF (B).

Figure 1 .
Figure 1.The plot of δ 202 Hg versus ∆ 199 Hg with a table of different types of excavation, as well as the labels of different types of ore (A) and ∆ 199 Hg versus ∆ 201 Hg (B).The regression lines for both observed data and the slope of 1 and 1.36 as observed for photoreduction of MeHg[22] for odd-MIF (B).

Figure 2 .
Figure 2. The plots of δ 202 Hg versus cHg (A) and ∆ 199 Hg versus cHg (B).Note that vertical error bars are much narrower in comparison to the horizontal ones due to the logarithmic scale.

Figure 2 .
Figure 2. The plots of δ 202 Hg versus cHg (A) and ∆ 199 Hg versus cHg (B).Note that vertical error bars are much narrower in comparison to the horizontal ones due to the logarithmic scale.

Table 1 .
Types of samples and descriptions.

Table 2 .
Operating conditions of Nu Plasma II MC-ICP-MS.

Table 3 .
The repetitive measurements of NIST 8610 and NIST 2711a.Standard deviation is abbreviated by SD and where not number was given it is noted by n.a. for not given.