Introduction

The construction of nuclear reactors as a source of reliable green energy has gained momentum in many countries of the world in recent years. However there are rising nuclear security concerns about the safety and proliferation of the nuclear materials used in these nuclear power plants [1,2,3] as well as possible diversion for illicit purposes by nuclear terrorists [4, 5]. This has necessitated the characterization of seized nuclear or radioactive material based on chemical and isotopic composition as well as physical parameters which can be used in determining the origin of the interdicted radiological nuclear material. The characteristic difference between lead isotopic composition from the earth’s crust (e.g. uranium ore body) and that from industrial emissions enables the applications of this ratio in nuclear forensics investigations [6].

In South Africa there is a vast uranium ore (uraninite) deposits [7], with a lot of mining and processing activities. It is therefore imperative for each the country to properly collect and compile nuclear forensic signatures in databases and national libraries, that can be used as evidence for attribution of the seized nuclear or radioactive material.

Many instruments are being used to apply various analytical techniques for chronometric analysis of intercepted nuclear materials from a nuclear facility. The LA ICP-MS or laser-ablation micro-sampling (LAM-ICP-MS) are some of the Instruments that have been used for determining the lead isotopic signatures of the material at the front or back end of the nuclear fuel cycle [8,9,10]. Recent study by Fuchs et al., used a LA-ICP-MS to measure trace elements (provenance of uranium) on the samples taken from the Transvaal Supergroup (which included the Carletonville Gold field) in South Africa, and their results showed that the higher U and gold concentrations are embedded in the pyrite rock [7].

The objective of this work was to demonstrate the use of the Perkin Elmer 300Q ICP-MS as a viable tool for resolving nuclear forensic signatures from a uranium mine in South Africa. Results from the first stage in the fuel cycle are described here, and the data presented could form a basis for a South African nuclear forensics library. However, a comprehensive nuclear forensic library can only be developed when all the stages in the fuel cycle have been investigated. This work describes the investigation of Pb isotopic composition and trace elemental analysis to determine respectively the fingerprint lead signatures and the provenance of the uranium in the uraninite ore. Interpretation of the results for possible tracing (attribution) of the origins of South African Pb is presented. We also discus limitations of Pb isotopic fingerprinting technique in this work [11].

Materials and methods

Study area

The study area is indicated in Fig. 1 and the authors cannot give further details due to confidentiality agreements.

Fig. 1
figure 1

Water sampling points in the Carletonville goldfields Area of South Africa. From the top of this Fig.: DSW39/17-Water from DSW36/15 and DSW38/16; DSW36/15-Water from DSW9/14 and DSW42/10); DSW38/16-Discharge water from another mine; DSW42/10-Fissure water straight from underground; DSW9/14-Water from the settling point; DSW7/12-Water coming from the mine shaft and from process plants; DAM3/13-Water after (downstream of) the settling point; DSW43/19-Purified water; DSW40/5 & DSW40/7-Raw water from return water dam; DSW18/3 & DSW112-Sewage water from mine shaft; DSW45/1-water from mine shaft; WV16, WV9, WV13, WV14, WV15 & WV3-Borehole water for public use from West village

Full size image

Geaology of the mining site

The sampling site falls within the Carletonville goldfields which is part of the so called B-Reef or Witwatersrand Supergroup) of South Africa. The uraninite (UO2x) [12] deposits here are unconformity in nature lying on the sedimentary succession of the Witwatersrand Supergroup, with an estimated age of about 2.59 Ga [7]. Gold (and uranium) mining uses the underground mining techniques as the mines are each estimated to be more than 2 km deep. These mines are now over 132 years old, [13] and are still operational.

Sample collection

The Carletonville goldfields have a quartz-pebble conglomerate, intercalated with chert, jasper pebbles and minor shales, bound within siliceous to slightly argillaceous rocks, and in this work, we sampled from mining operations in the Carletonville gold fields [7].

The tailing dams in Fig. 1 are located on the RHS of DSW7/12 (T3 no separate Fig.); below DSW43/19 (T3 no separate Fig.), between DSW40/7 & DSW18/3 (T1 in Fig. 2) and above WV15 (T2 in Fig. 2). Samples used in this study consisted of a total of 12 water samples (mixture of mine water—also called fissure water from underground at the ore deposits—used in cooling the drilling machine; waste water from the processing plant and 24 soil samples from each of the mine tailing dam, T1 & T2. It was observed that the slurry from the plant is deposited on the tailing dam in paddocks whose boundary contours are the edges of the different colors of the slurry, although some cross mixing can be seen. Figure 2 shows an example of a tailing with paddocks at different drying stages. These paddocks are filled from the outside inward and in a sequential manner following the perimeter of the Tailing dam. Hence they dry in that same sequence, each paddock taking as long as more than a month to be reusable again (of course this depend on the size of the dam). The soil and some water samples were collected from the top of the tailing and the discharge canals respectively to represent uranium that has come from the milling and processing stage of the fuel cycle. Fissure water samples were also collected from the mine shafts outlets to represents uranium from the ore body.

Fig. 2
figure 2

Soil (tailing, T1 & T2) sampling points in the Carletonville goldfields area of South Africa, LHS is T1 & RHS is T2

Full size image

Soil sample preparation

Soil, samples were digested using a Multiwave 3000, Anton Paar microwave oven. About 1 g of each sample was weighed into a Microwave digester vial and then digested with the aqua regia acid (3 mL of 55% HNO3, 9 mL of 32% HCl) and 2 mL of 2% H2O2—which enhances the oxidation properties of nitric acid [14]) following standard procedures. The aqua regia extraction is capable of complete recovery for Cd, Cu, Pb (our target element) and Zn [15]. That is, the microwave rotor was transferred to a microwave reactor system for 20–45 min. Then an aliquot (10 mL) of each sample was transferred into a volumetric flask and topped up with Ultra pur distilled water up to 100 mL. They were stored at 4 °C for another 24 h. The reagents used were of Suprapur analytical grade (by Merck Laboratory Supplies Pvt. Ltd). The EPA Method 3052 used in the sample digestion achieves total sample decomposition [14].

Water sample digestion protocol

A whole set of 5 mL each of the water samples were mixed with 5 mL of nitric acid and 1 mL of hydrochloric acid and then left to digest in the Multiwave 3000, Anton Paar microwave oven for 24 h. A 10 mL aliquot of each digested sample was then transferred into volumetric flasks and topped up with Millipore Milli-Q system (resistivity 18 MΩ cm−1) distilled water up to 100 mL and left overnight for ICP-MS analysis.

All digested sampled were then preserved at 4 °C for before analysis in the Laboratory [16].

Instrumentation

The Perkin Elmer, NeXION 300Q, Inductively Coupled Plasma Mass Spectrometer (ICP-MS), was used for all sample analysis in this work. It has a Quadrupole ion deflector that focuses the ion beam to the Dual mode detector. The Isotope-ratio precision of this instrument is defined for the isotope ratio of 107Ag/109Ag internal standard using a 25 μg/L solution, which is achieved by single-point peak hopping with a relative standard deviation (= 100 × SD/AVERAGE, ((0)) of < 0.2% RSD [17]. The optimized operating parameters are summarized in Table 1.

Table 1 NexION 300q ICP-MS instrumental parameters [18, 19]
Full size table

Interference reduction

In the ICP-MS analysis of samples, the following molecular ions are potential sources of interferences; oxides, hydrides, hydroxides, nitrides [20,21,22]. Their effect can however be reduced by using the Perkin Elmer NexION 300Q’s desolvating nebulizer. Also, the Instrument’s Quadrupole ion deflector focuses only the selected isotopes in the ion beam, to the Dual mode detector. The other (interfering) ions are allowed to pass through to the waste [18, 23, 24]. Although the spectral resolution (0.7 amu) for this instrument is inadequate for discrete measurement of elements prone to sample-based spectral interferences, its collision-reaction cell technology render this problem as minor for the trace element determined in this work [25]. This capability of the NexION 300Q, combined with its Universal Cell Technology™, enables significant reduction in most or all the molecular ions in the sample [26, 27]. Also instrument drift is automatically corrected by the Software [28].

Sample run

The samples were loaded on to the auto sampler and initialized using the ICP-MS Instrument Control (Data Acquisition) Software. The instrument was set to Isotopic Ratio Method, operated in the Collision Mode for mass energy discrimination and filtration against interferences [24].

ICP-MS trace calibration for trace element analysis

For analysis of trace elements, the Perkin Elmer, NexION 300Q, (ICP-MS), calibration uses a Dual Detector Calibration Solution as the Atomic Spectrometric Standard, whose specifications are that:

In the Total Quantitative method, the mass calibration stability is measured using a 10 μg/L multi-element standards solution Al, Ba, Ce, Co, Cu, In, Li, Mg, Mn, Ni, Pb, Tb, U and Zn [17]. For every measurement, the Instrument was set to run a blank and a standard check at every ten samples for quality control.

Pb isotopic ratios data for geological CRMs

We have compared our results with those from USGS Reference materials as given in Schudel et al. [28] and the GeoMed Database [29], and Weis et al. [30].

Results and discussion

The results presented here are divided into two categories, viz; soil tailing and water samples. Due to lack of Certified Reference Materials to use in calibration check, the results presented here are approximated.

Soil tailing samples

The provenance of the Uranium, lead and other trace elements in the soil samples taken from tailing dam 1 and 2 are shown in Table 2.

Table 2 Provenance of the uranium, lead and other trace elements in the soil samples taken from tailing dam 1 and 2 (ppm), with an average of n = 3 replications for each entry
Full size table

Tables 3 and 4, shows the Pb–Pb and U–U isotopic ratios (respectively) for the Tailing soil samples. It can be concluded that there is good agreement between the sample ratios and the ULTRASPEC standard used to calibrate the Instrument. We note here that these results were acquired using an Agilent 7500 Series ICP-MS, in which the Multi-element calibration standards (from ULTRASPEC, SOUTH AFRICA), were used to validate the method. It is also noted that the results of the ULTRASPEC STD ratios are in god agreement with those found in the NIST data base. The authors could not use a NIST isotopic standard due to limitation of resources. However the ULTRASPEC standard contained U and Pb elements. This was assumed sufficient for identifying lead and uranium isotopic signatures.

Table 3 Pb-Pb isotopic ratios for tailings soil samples
Full size table
Table 4 U-U isotopic ratios for tailing soil
Full size table

Mine water results

Table 5 shows the atom percent abundances of the U and Pb in the water samples analyzed.

Table 5 Abundances of U and Pb in mine water samples (wt%)
Full size table

The results in Table 5 shows exceptionally high average abundance values of 235U, for samples WV13, WV15, DSW7/12, much above natural uranium isotopic concentrations, probably due to anthropogenic processes. However the water samples are enhanced in 206Pb which is a signature for a radiogenic ore deposit, as also evidenced by the normal Crustal levels for 204Pb. Although sample WV13 is borehole water, the results show that this rock bed is highly depleted in 238U. That would affect all the water samples from WV9, WV14 & WV15 as they are in the same area. There could be pyritic (sulphur containing) [34], hydrothermal underground water infiltration into the borehole waters and thus dissolving and leaching the uranium ore [7]. The bottom table shows a comparison with the NIST and US-DOE standard values for Pb and U respectively, and there is a good agreement.

The results from Table 6, show that all the DWS water samples from this mine have lead isotopic ratios close to the NIST SRM 981 values. These samples contain natural uranium from underground. This table also reveals the following characteristics of the mine. For the purified water (WVs) samples, the isotopic concentrations of the Pb isotopes varied widely from one sample to the other. However most of the samples from fissure water had close isotopic ratios. The concentration of 235U in the fissure water shows that this sample is natural uranium before extraction by chemical processes, while that from the tailing has very low 235U. This in itself is a signature that distinguishes between ore uranium and processed uranium for the South African mine investigated. For this mine too, the concentration of 206Pb is unexpectedly high signifying ore deposits rich in 238U.

Table 6 Lead isotopic ratios for water samples before and after 208Pb/206Pb normalization for mass balance
Full size table

For this geochronological study we use the fingerprint of the two ratios, 207Pb/206Pb and 208Pb/206Pb [11, 26]. The instrument mass bias corrected isotopic ratios of lead normalized to 208Pb/206Pb [35] are shown in Table 7 below. The advantage of using 206Pb as the denominator isotope is that there is significantly less error magnification and much less error correlation between the measured ratios than for ratios to 204Pb [35].

Table 7 Water sample results relative to 204Pb for the mine area [36]
Full size table

Table 7, shows that the uranium ore from which the sample mineralisation was found is a pyrite, with Pb ratios similar to that found by Jopoul et al. [36]. These results also show that the isotopic signatures are less radiogenic (206Pb/204Pb ≤ 20). Also the Pb-Pb plot for these results (see our Fig. 3), confirm that the Carletonville gold fields are pyrite deposits, giving another signature for this mine. The similarity of these plots (data ranges) with those from Ref [36], suggest that the pyrites deposits are of detrital type.

Fig. 3
figure 3

A plot 207Pb/204Pb versus 206Pb/204Pb for mine (fissure) water samples

Full size image

The results of this work are relatively comparable with those from inter-laboratory comparisons although it is worthy pointing out that in those results [35], rigorous sample preparation and isotopic ratio corrections were effected than is done in this work.

In Fig. 3, lead (Pb) isotope ratios for this SA gold mine have 207Pb/204Pb values distinctively ranging from about 13–20 and the 206Pb/204Pb values are ranging from about 16–25. This is distinctively different from lead samples in UK, Germany, Poland, Italy, and Spain [9]. According to Belluci et al., a 206Pb/204Pb ratio greater than 20 indicates that Pb investigated emanates from a uranium ore whereas a less ratio indicates a Pb-rich ore [37]. The U-bearing phases associated with the higher radiogenic Pb content are zircon, monazite and apatite [37]. This study had an extreme Pb 206Pb/204Pb ratio of 16–25 and it differs from that of 17.6–18.0 in Belluci et al., studies which represent a Pb-ore Bunchas mine located in Newfoundland, Canada. As a result it can be used for as a signature for the mine examined in this study conducted here compared to the mine in Canada. The results of Fig. 3 further indicate that this signature from the tailing is more defined (reliable) with less scatter of the data. This could be because in the study area, each tailing dam receives only processing wastes from a particular selected mine shaft(s) which presumably extract from the same uranium ore zone underground. Also the mine (fissure) water from all shafts is pumped to the same discharge pipe where the researchers collected the fissure water at various points in the discharge. The ratios are all within expected ranges [9, 38].

Comparing Figs. 3 and 4, it is evidence that the signature for uranium in the mine (fissure) water, which contains uranium from the ore, is distinctly different in that the last two points in Figs. 3 and 4 are seemingly outliers (DSW9/14-water from the settling point, where all shaft mine waste deposits accumulates). A close look at the isotopes indicate that these two points are actually due to the observed much higher value of 206Pb (implying higher 238U concentration) and even lower value of 204Pb (that is, these two deposits are of radiogenic origin [38]). Results in Fig. 4 indicate the geological origin of the deposit (radiogenic in this case). Here the 208Pb/204Pb values are distinctively ranging from about 32–46.

Fig. 4
figure 4

A plot of 208Pb/204Pb versus 206Pb/204Pb for water samples

Full size image

The characteristic signature of the lead isotope ratios from this mine can be seen in Fig. 5 where most of the data is scattered around 208Pb/206Pb = 2.05 slightly less than the NIST SMR value of 2.07 probably due to lead loss in the ore. In Fig. 6 the data scatter around values of 204Pb/206Pb = 0.059, which is in good agreement with the NIST SRM 981 [31] value of 0.59042.

Fig. 5
figure 5

207Pb/206Pb versus 208Pb/206Pb for mine (fissure) water samples

Full size image
Fig. 6
figure 6

Isotopic composition of the mine water samples relative to 206Pb

Full size image

Discussions

The NexION 300Q ICP-MS has a sensitivity > 40 mg/L and > 20 mg/L (for 24Mg) with a detection limit of < 0.2 ng/L (for 238U) [17]. The instrument Short-term precision is defined as the relative standard deviation (% RSD) for a 1 μg/L multi-element solution, measuring 24Mg, 63Cu, 114Cd and 208Pb, without internal standardization. It is evaluated as < 3% RSD. It has an isotope-ratio precision which is defined for the isotope ratio of 107Ag/109Ag using a 25 μg/L solution. Obtained using single-point peak hopping this value is < 0.2% RSD.

The excellent sensitivity, precision and good accuracy of the ICP-MS operated under collision cell interface, has made it comparable to thermal ionization mass spectrometry (TIMS), for precise isotope ratio measurements. The ICP-MS configuration now allows for the dissociation, thermalizing and neutralizing of the disturbing argon-based molecular ions. The application of the collision cell in ICP-QMS results in a higher ion transmission, improved sensitivity and better precision of isotope ratio [39]. These characteristics formed the basis for using our Perkin Elmer NexION 300Q ICP-MS in this work.

TIMS results shown in Fig. 5b of Weis et al. [30], fall within the range shown in Figs. 3 and 4 of this work.

Sources of uncertainties

The Pb–Pb ratios in these results are relatively lower than the true values. This could imply that instrument optimization alone is not sufficient to correct for mass fractionation although the NexION ICP-MS instrument used here does not use laser ablation (the most common source of mass fractionation [26, 35].

In this work the uncertainties were evaluated as follows:

  • For trace elements—uncertainties are presented as STDEV of all the results in the Tables.

  • For Isotopic ratios—all results are average values of three replications by the instrument.

  • For all graphs—we calculated the STDEV (σ) for each data set plotted and used that to plot the error bars as (± σ) for each point on the graph.

Lead isotopic signatures

The lead isotopes in mineral ores originate either from natural lead or radiogenic lead [40]. In this work, the isotopic composition of lead in the mine ores of Carletonville Gold fields was studied using the ICP-MS isotopic ratio technique as a tool for nuclear forensic investigations. To characterize different materials from the mine shafts, we measured the trace elementary impurities also. The lead isotopic composition of water direct from the mine shaft varied largely from those from the borehole or purified water. Thus this can be applied as a parameter to distinguish ore bodies from different origins.

Trace elements

The concentrations of the trace elements (Pb, Ba, K, La, Mg, Mn, Na, P, Sr, Th and U) reported here indicated that the uranium deposit in this mine is contained in pyrite type ores [7]. The mean values for Pb ranged around 0.090 ± 0.053 to 0.111 ± 0.068 ppm for soil tailing 1 & 2 respectively while U had ranges from 0.534 ± 0.406 in tailing 1 and a mean value of 0.555 ± 0.237 ppm in tailing 2. These amounts then seep into the water systems and could pose a radiochemical and radiological danger to animals and people utilizing the water taken from beyond DSW39/17 in Fig. 1. The concentrations were very high compared to the SA Department of Water Affairs and Forestry (DWAF) values 10 ppb for Pb and 40.5 ppb for U. The trace elemental impurities, show that their concentration in the mine tailing is different from that in the mine water which further support signature of the lead isotope data.

Conclusions

In this work, elemental and lead isotopic ratio analyses of uranium mine ores samples, from the 3 operating mine shafts and three operating tailing dams in South Africa were carried out to determine the lead isotopic signatures to be developed for use in identifying the origin of South African uranium. We conducted elemental analyses of samples from the two tailing dams nearest to the mine shafts using the inductively coupled plasma mass spectrometry (ICP-MS). Lead isotope ratios were measured using ICP-MS isotopic ratio technique. The trace elements concentrations indicated that the uranium and lead had concentrations indicative (signatory) a uranium ore possibly embedded in phosphorite (pyrite) deposit in Apatite and fluorapatite minerals, and is recovered from phosphoric acid leaching followed by lime precipitation to remove impurities and magnesia is used for precipitation of the uranium [41]. The Lead ratios are found to fall within the certified values or standard reference. Also the Pb–Pb plot for these results confirm that the Carletonville gold fields are pyrite deposits, suggest that the pyrites deposits are of uraninite detrital type giving another nuclear forensic signature for this mine.

This work has demonstrated the viability of the ICP-MS Isotopic Ratio method as a techniques for developing nuclear forensics signatures for a South African mine. The nuclear forensic graphs presented here provide enough evidence to confirm that uranium ore signatures are indicative of the geochemical origins of an interdicted sample while those for the processed uranium points to the geographical origin.

Ore (water) samples were shown to contain balanced isotopic ratios of 238U, 235U and 234U, while tailing samples revealed a depletion of these ratios maybe due to efficient modern processing techniques.

There is however, a need to expand this work to at least ten more mines with a wider geographical area in South Africa and also to verify these results with other techniques like TIMS or XRF. These will be most importantly applied in the next stage of the research, from an analytical and metrological point of view.

The researchers therefore will continue this investigation with samples from other mines and then follow each stage of the fuel cycle unto the uranium waste samples (spent fuel from the South African reactors). It is hoped that in 10 years’ time South Africa will have a reliable nuclear forensics library for attribution and prosecution.