Introduction

Coal plays an important role in electric power generation in many countries at the global scene and in Republic of Serbia as well. Thermal power plant “Nikola Tesla” (TPP NT) is a power plant complex, located on the river Sava, approximately 40 km from Belgrade. It is the largest one in Serbia, and its production covers majority of Serbia’s needs for electricity. TPP NT uses lignite mined from the mining basin Kolubara as fuel—a total of 29.2 million tons in the year 2014 [1].

Coal, as the fossil fuel, contains numerous natural radionuclides from uranium and thorium decay chains. Uranium daughter products such as 210Pb and 210Po are very important due to their volatility properties during the combustion process. These radionuclides are partly released in the flue gases and can escape to the atmosphere, and considered to be main contributors to the internal radiation dose of humans [2, 3].

According to the International Atomic Energy Agency (IAEA), coal and products of its combustion (slag and ash) are labeled as naturally occurring radioactive material (NORM). The specific activities of natural radionuclides in the coal are in wide range and strongly depend on the geological formation of coal. Slag and ash, which have specific activities are up to 10 times higher, are often released into the environment causing an increase and/or redistribution of natural radionuclide content.

Majority of radionuclides in coal remains in the ash but some of the ash from coal-fired power plants escapes into the atmosphere. Pollutions from air could contaminate soil, water, vegetation, the whole ecosystem and human health in the vicinity of a coal fired thermal power plant [3].

Thus, the monitoring of the environment should be regularly performed in the vicinity of TPP, together with comprehensive analysis of radiological characteristics of coal, slug and ash samples in order to estimate the radiological risk for health of inhabitants.

Theory

210Po and 210Pb in the environment

Recent studies showed that the activity concentration of 210Po and 210Pb have increased in the environment as a result of human activities such as burning fossil fuels in thermal power plants, the industry of fertilizers, and by vehicles through exhaust fumes [4].

210Pb is a natural radionuclide from uranium decay chain emitting beta and gamma radiation, with a relatively long half-life (22.3 years). It originates from lithogenic minerals in the subsoil [4].

210Pb can be release into the environment from areas of sediment deposition, such as ash landfills from TPP. Once released in environment 210Pb could find its way into ground water and can result in the contamination of drinking water.

The natural distribution of 210Po in environment is mainly determined by its progenitor 210Pb distribution. Also, it is influenced by possible exhalation of its gaseous precursor 222Rn. Namely, if 222Rn escapes from soil or rock matrix, 210Pb is produced elsewhere, which leads to redistribution of this radionuclide and therefore to the temporarily disequilibrium between 210Pb and 210Po in the environment. Furthermore, 210Pb usually is not in secular radioactive equilibrium with its precursors in uranium series in environmental samples [5].

210Po is an alpha emitter with a relatively short half-life (138.4 days) and represents one of the most toxic substances due to its intense specific radioactivity, (1 μg of 210Po has an activity of 1.66 × 108 Bq) [6]. 210Pb is mainly accumulated in skeleton in human body while 210Po accumulates in soft tissues such as muscles and the liver in organisms [4, 7]. After entering the human body a relatively rapid rise of 210Pb daughter product 210Po occurs. Consequently, ingestion/inhalation of 210Pb leads to significant increase in internal exposure to radiation. Therefore, it is necessary to determine 210Pb accurately in different environmental samples, which could be difficult due to its physical and chemical properties.

Methods for 210Pb determination

Activity of 210Pb can be determined by gamma—ray spectrometry via photo peak with relatively low emission probability, 4.252 %, at energy 46.639 keV. Coaxial HPGe (p-type) spectrometers, which are usually used in laboratories for routine measurements, generally have poor detection efficiency at lower energies. Thus, mentioned photo peak commonly have poor statistic, and is not well shaped, so the use of such spectrometers can lead to unreliable determination of 210Pb activity concentration, especially for samples with low activity.

In order to obtain reliable and accurate results, in some cases radiochemical determination of 210Pb activity is necessary. Specific activity of 210Pb can be determined by measuring the beta activity of its daughter product 210Bi (E βmax = 1.17 MeV) or the alpha activity of its granddaughter 210Po (E α = 5.32 MeV). In previously mentioned two methods, 210Pb activity is determined by measurement of its progenies, and it is necessary to know the degree of equilibrium that exists between the progeny and its parent 210Pb. For 210Po a period of more than 4.5 months is required to reach 50 % and 18 months to reach 88 % in-growth. For 210Bi, with half-life 5.01 days, 98 % of equilibrium is reached in 30 days [2]. Changes of relative activities for 210Pb, 210Bi and 210Po in time in an isolated system are presented in Fig. 1.

Fig. 1
figure 1

Changes of relative activities for 210Pb, 210Bi and 210Po in time in an isolated system

Full size image

Alpha spectroscopic measurement of the in-growth 210Po can achieve an ultra low detection limit, often 100–1000 times lower than for gamma-ray spectrometry, and represents the most sensitive method for indirect determination of 210Pb. For this kind of determination it is necessary to carry out a polonium extraction from the examined sample after a minimum 5 half-life periods (almost 2 years) from its collection. Two years period is very long, and certainly is not acceptable for routine measurements. Nevertheless, comparison could be performed after the shorter period of time, before the equilibrium is established, but in this case, in order to determine expected value of \(A_{{{}^{210}{\text{Po,}}\upalpha }} /A_{{{}^{210}{\text{Pb,}}\upgamma }} *\) ratio, Bateman’s equations should be applied. Regarding the quality control in testing laboratory such analysis should be performed in order to conduct internal comparison of results for 210Pb activity obtained by gamma-ray spectrometry.

Since 210Po is a pure alpha emitter, its determination can be carried out only by alpha-particle spectrometry. Passivated implanted planar silicon (PIPS) detectors are commonly used for that purpose. Spectroscopically thin alpha sources are usually prepared by spontaneous deposition of polonium ions onto less noble metal surfaces, such as nickel, copper or silver.

In case of 210Pb determination through its daughter product 210Bi flow-gas proportional counter is commonly used.

In this study, the activity concentration of 210Pb were determined in coal, slag and fly ash samples (29 in total) originated from TPP NT by gamma-ray spectrometry and via in-growth of 210Po and 210Bi by alpha spectrometry and by beta counting, respectively.

Gamma-ray spectrometry measurement of 210Pb was performed during the first half of 2013th, while alpha-particle spectrometry measurement of 210Po was conducted in second half of August and first half of September 2015th. For 9 of these samples (three of each kind—coal, slag and ash) 210Pb was determined by beta counting of in-growth 210Bi in October 2015th.

Experiment

Sample collection and pre-treatment

For the purposes of this study, samples of analyzed coal, slag and ash (29 in total) were collected during the periodical emission tests in November and December 2011th at unit A1 of TPP “Nikola Tesla” and also in September 2012th at unit A3 of TPP “Nikola Tesla”. Coal sampling was performed at feeders, before grinding process and drying coal in the mill. Slag was sampled during the transportation below the boiler, whereas ash was sampled in electrostatic precipitator pyramid hoppers, Fig. 2.

Fig. 2
figure 2

Scheme of coal, slag and ash sampling places at TPP unit

Full size image

Coal, slag and ash samples were prepared for analysis (milled, dried, sieved and weighed) by appropriate standard methods, and ultimate and proximate coal analysis, proximate analysis of slag and ash were performed in accordance with ISO 17025.

Determination of 210Pb by gamma-ray spectrometry

For gamma-ray measurement prepared samples of coal, slag and ash were placed in 125 ml PVC cylindrical boxes. The mass of the samples ranged from 64.85 to 97.99 g, depending on the sample density. All samples were measured by means of coaxial germanium semiconductor HPGe detector (ORTEC GEM-30, 37 % relative efficiency and 1.8 keV resolution for 60Co at 1332 keV line). The detector was calibrated using the coal and soil matrix in identical cylindrical boxes spiked with common mixture of gamma-ray emitters (241Am, 109Cd, 139Ce, 57Co, 60Co, 137Cs, 113Sn, 85Sr and 88Y) certified by CMI (Czech Metrological Institute) [8]. Procedure of preparation of calibration standards is described in details in previously published paper [9], and efficiency curves used in this study have been verified through regular participation in interlaboratory comparisons and PT (proficiency test) schemes. In order to obtain valid curve, net peak areas of 60Co and 88Y were corrected for the coincidence summing effect applying the calculation method of Debertin and Schötzing [10]. The analytical expression of obtained efficiency curves was ε = e−P(lnE), where ε is the detection efficiency, E is the energy and P(lnE) is the polynomial function of the fifth order. Uncertainty of the efficiency calibration includes uncertainty of radionuclides activities in standard, statistical uncertainty and uncertainty of efficiency curve fitting. Measurements of the prepared standards were performed in a close-to-detector geometry.

The obtained spectra were recorded using Canberra’s Genie 2000 software. Recorded spectra were analyzed using Genie 2000, Aptec and Anges software, and all calculations were performed with the Mathematica 5.2 software (Wolfram Research Inc.). In order to achieve acceptable statistics, the samples were measured in a close-to-detector geometry from 66,000 to 250,000 s.

The activity concentrations of 210Pb in coal, slag and ash samples were determined directly by analyzing full-energy peak at 46.54 keV.

Uncertainty of measurement results was determined by applying the general law of propagation of uncertainty. The largest contribution to the total uncertainty was due to the statistical uncertainty (up to 30 %) and efficiency calibration (5 %), whereas uncertainty due to measured mass of the sample and decay parameters could be neglected.

Determination of 210Pb via in-growth of 210Po by alpha-particle spectrometry

About 0.5 to 1 g of samples (coal, slag, and ash) was taken from PVC cylindrical boxes which was already used for gamma measurement and put into Teflon beaker. For recovery determination approximately 0.1, 0.2 and 0.3 g of 209Po tracer was added into weighted coal, slag and ash samples, respectively. Specific activity of the tracer was (0.3373 ± 0.0064) Bq/g, on 01.07.2010. Further, beaker was placed onto plate on 200 °C for wet ashing. Wet ashing was carried out by adding 10 ml of concentrated HNO3 and H2O2 in order to destroy organic matter. After evaporation to dryness, 10 ml of concentrated HNO3, HClO4 and HF were added successfully in that order, and left to evaporate to dryness. Dry residue was converted to chloride form with adding 5 ml of concentrated HCl and again evaporated to dryness. The dry residue was dissolved in 2 ml of concentrated HCl and diluted with distilled water to 100 ml volume. Then, 1 g of ascorbic acid was added into prepared solution in order to reduce iron ions (II and III) and eliminate interference from the silver disc during the deposition process [4]. The spontaneous deposition of polonium on a 19 mm diameter silver disk was carried out at 90 °C for 4 h. After the deposition was finished, the disk was rinsed with deionized water and dried on air. The disk was subsequently counted on alpha spectrometer described below, for activity determination and recovery calculation [2, 11].

For alpha spectrometric measurements of 209Po and 210Po an alpha spectrometer (Alpha Analyst, Canberra, USA), with passivated implanted planar silicon (PIPS) semiconductor detectors (active area of 450 mm2 and efficiency of 28 % for diameter disk of 25 mm) was used. For calibration of the detector standard radionuclide source with following radionuclides: 238U, 234U, 239Pu and 241Am (code: 67978-121) was used. Standard radionuclides source was procured from Analytics, Inc. Data acquisition analysis was done in the Genie 2000 spectroscopy system software. Depending on the radionuclide activities in the samples the counting time varied from 1 to 7 days. Radiochemical recovery was determined by 209Po activity measurement.

Determination of 210Pb via in-growth of 210Bi by proportional counting

In 9 out of 29 samples simultaneous determination of 210Pb and 210Po was performed in all three kinds of samples. At the beginning of the radiochemical procedure, besides 209Po tracer, 25 mg Pb2+ of lead carrier were added to the samples. The dissolution of the samples was done as described in the previous chapter. The residue after total dissolution was dissolved in 2 M HCl and loaded on a Sr resin column (Eichrom Industries Inc.). The analytical method is based on selective separation of lead and polonium by extraction chromatography with bis-4,4′(5′)-t-butyl-cyclohexano-18,6-crown ether [12]. The non-retained ions were washed from the column with 2 M HCl. Polonium was stripped with 6 M HNO3 while lead was removed with 6 M HCl solution. Obtained fractions were evaporated to dryness. Polonium source was prepared by self—deposition on a silver disk from HCl solution (pH = 1–2) with addition of 1 g of ascorbic acid to prevent deposition of interfering elements, as described in previous chapter. Polonium sources were measured by alpha-particle spectrometry.

Source for 210Pb determination was prepared by precipitation of PbSO4. Dry lead fraction residue was dissolved in water and concentrated H2SO4 was added. The precipitate was transferred to a counting planchet and its daughter radionuclide 210Bi was measured at equilibrium on a low background gas flow proportional counter. For measurements a Canberra TENNELEC LB4100-W low background gas-flow proportional counter was used. Measurement system has eight proportional detectors in two drawers. Mixture of argon (90 %) and methane (10 %) was used as a counting gas. To allow corrections because of in-growth of 210Bi and different self-absorption for samples with different radiochemical recovery measurement system was calibrated.

Results and discussion

210Po and 210Pb activities, measured by alpha-particle and gamma-ray spectrometry are presented in Table 1. Recovery was calculated through known quantity/activity of added tracer—209Po, and ranged between 25 and 90 % while in the case of determination of 210Pb by beta counting recovery was determined gravimetrically and varied between 35 and 64 %. From Table 1 it could be seen that results for 210Pb specific activity obtained by semiconductor gamma spectrometry in 2013 are ranged from (16.5 to 116.0) Bq/kg, (20.1–60.8) Bq/kg and (84.3–153.0) Bq/kg, for coal, slag and ash, respectively. Obtained values were corrected for radioactive decay (\(A_{{{}^{210}{\text{Pb,}}\upgamma }} *\)) to enable comparison of results.

Table 1 Specific activities of 210Po and 210Pb, measured by alpha and gamma spectrometry, respectively
Full size table

Activities of 210Pb measured by in-growth of 210Po by alpha spectrometry in 2015 are also given in Table 1 and ranged from (11.9 to 160.5) Bq/kg, (21.1–49.7) Bq/kg, (75.6–151.0) Bq/kg for coal, slag and fly ash, respectively. The obtained values in the analyzed samples are in compliance with results obtained for specific activity of 210Pb worldwide [13, 14]. The specific activities are the lowest in the coal samples, and the highest for the ash samples, as it was expected.

Simultaneous representation of results obtained by alpha-particle spectrometry of 210Po, gamma-ray spectrometry of 210Pb and beta proportional counting of 210Bi for nine samples are presented in Table 2.

Table 2 Specific activities of 210Po, 210Pb and 210Bi, measured by alpha and gamma spectrometry, and beta counting, respectively
Full size table

Additionally, correlations between used techniques are presented in Figs. 3, 4 and 5. Strong correlation between alpha and gamma measurements could be noticed from Fig. 3 (adjusted R 2 = 0.976).

Fig. 3
figure 3

Correlation between alpha-particle and gamma-ray spectrometry

Full size image
Fig. 4
figure 4

Correlation between gas flow proportional counter and gamma-ray spectrometry

Full size image
Fig. 5
figure 5

Correlation between gas flow proportional counter and gamma-ray spectrometry neglecting value for sample Coal-142

Full size image

The highest discrepancy is for Coal-142 sample, but due to large number of samples/points, this does not affect good correlation.

On the other hand, correlations between gamma-ray measurements and beta counting is very weak (adjusted R 2 = 0.361) (Fig. 4).

Again, the highest discrepancy was obtained for sample Coal-142. Unlike the previous case, this point affects the correlation, but neglecting this value, correlation became much better (adjusted R 2 = 0.868) (Fig. 5), and somewise acceptable.

Reason for weaker correlation could be a smaller number of points compared to the correlation between gamma and alpha spectrometry and the fact that the beta counting measurement technique is less sensitive than alpha spectrometry. Also, should be taken into consideration that something is not right with sample Coal-142 (pre-treatment homogeneity etc.) and it should be examined.

Statistical analysis of the measurement results

In order to test statistical difference between two data sets—results of activity measurements performed for the same sample in two detector systems (alpha and gamma, gamma and beta), and to check if obtained results differed from the statistical point of view, Z score test was used for confidence interval of α = 5 % (Critical value \(\left| {{Z_{0.05}}} \right| < 1.96\)) [15].

For all measured samples, results obtained by gamma and alpha spectrometry were not statistically different. For results of activity measurements obtained by gamma-ray spectrometry and gas flow proportional counter for 9 out of 29 samples Z score test showed that these results were not statistically different, except for sample Coal-142.

Also, to check the concordance of the two different methods used a simple test—combination of Z score and χ 2 tests—was used. For Z scores that are normally distributed the sum of the squared Z scores should follow a χ 2 distribution [16]. Thus, the next null hypothesis is:

H 0

There is no difference between two measurement techniques: alpha-particle and gamma-ray spectrometry for the determination of 210Pb.

Starting from H0 it was proved that (Z i ) (i = 1,…, n; where n corresponds to the number of degrees of freedom) are normally distributed. The significance level (type I—error) is set to α = 5 % [17].

Accordingly, if alpha-particle and gamma-ray spectrometry differs by more than 5 %, H0 should be rejected, otherwise, one can say that there is no statistical difference between the two techniques.

The results of the χ 2 test based on the Z γα score values for testing the H0 hypothesis is presented in Table 3. It is statistically proved that two techniques are not significantly different at the confidence level of 5 % for the determination of 210Pb.

Table 3 Statistical comparison based on the χ 2 test between alpha and gamma-spectrometry determination at the 5 % confidence level
Full size table

For checking of the concordance between gamma spectrometry and gas flow proportional counting in this study, application of χ 2 test is not reliable due to insufficient number of measured samples.

Conclusion

In the present study three techniques (α and γ spectrometry, beta counting) for activity determination of 210Pb are described and compared. The obtained results for three different techniques showed that in the most cases results are in good agreement, regardless their different sensitivity and they are within measurement uncertainties for all techniques used. A statistical comparison of results obtained by three different techniques was performed using Z score test. Z score values were lower than critical value for all compared results, except for one sample.

A statistical comparison of activities measured by alpha-particle spectrometry and gamma-ray spectrometry demonstrated that gamma spectrometry is a valid alternative to time-consuming alpha spectrometry for low level activity measurements.

Numerous factors can affect the reliability of the measurement results. Reliability of gamma spectrometry results is strongly affected by: poor detection efficiency at low energy region, background variation, performances of applied software for spectrum analysis, as well as difference in effective atomic number of measured sample and standard used for detector calibration. Different software packages differently calculate net area of the photo peaks and for poorly shaped peak could give different values for net peak areas and its uncertainty. Therefore, calculated radioactivity could be different. In this study, software Anges has proven to be the most reliable in terms of accurate determination of not-well defined net peak area. Difference between sample and calibration standard matrices is included in uncertainty budget through uncertainty of efficiency calibration.

On the other hand, alpha-particle spectrometry and beta counting are reliable, sensitive, and more suitable for samples with very low activity concentration but can be time consuming. Additionally, deposition time, applied temperature, type of backing material for cathode self-deposition disc and others, affect the quality of the source.

Besides above mentioned reasons, discrepancy between alpha-particle and gamma-ray measurement could be caused by the fact that, compared with sample mass for gamma spectrometry, very small quantity of the sample (~1 g) is used for alpha measurements. Small amount of the sample could be not representative for the hole bulk, due to possible presence of “hot-spots” in the bulk sample. This fact could be explanation for high value of discrepancies for sample Coal 142. For the same sample Z γβ score value was higher than critical one, which is in agreement with the assumption that sample Coal 142 is perhaps not homogeneous but it should be investigated.

For all these reasons determination of 210Pb via in-growth of 210Po by alpha spectrometry could be a very useful tool for gamma spectroscopy reliability testing. For a similar, statistically reliable conclusion regarding beta counting, it is necessary to perform measurements with a larger number of samples.