2.1. Emitters of α and β^- in natural water

Three natural radioactivity series and more than 180 radionuclides exist in nature; the natural radioactivity in water resources is primarily from nuclides belonging to the ²³⁸U, ²³⁵U, and ²³²Th series as well as ⁴⁰K [12].

²³⁸U (T_1/2 = 4.468 × 10⁹ y) is the most widely distributed isotope of uranium (99.3%) on the surface; it has 15 decay daughters and stable states, where ²⁰⁶Pb is the terminal member of the decay series. The ²³⁸U series comprises 11 α emitters, 7 β^- emitters, and 10 gamma emitters. Gamma emitters with the highest relative intensity of α and β^- emission are ²²⁶Ra (11%), ²²²Rn (12.8%), ²¹⁴Pb (11.5%), and ²¹⁴Bi (27.6%). Meanwhile, thorium comprises six isotopes, of which ²³²Th (T_1/2 = 1.405 × 10¹⁰ y) is the most representative (99.8%). The ²³²Th series comprises seven α emitters, five β^- emitters, and eight gamma emitters. The decay chain ends at the stable state of ²⁰⁸Pb. The ²³⁵U series, which is the long half-life isotope of uranium (approximately 0.7%), comprises nine α emitters, four β^- emitters, and nine gamma emitters. Radionuclides belonging to the ²³⁵U series are always accompanied by the ²³⁸U series [13]. Because the content of ²³⁵U is much lower than that of ²³⁸U, the contribution of the ²³⁵U series for radioactivity in water is not considered. Radionuclides belonging to the ²³⁸U and ²³²Th decay series are shown in Tables 1 and 2, respectively.

Except for these three decay series, many radionuclides in nature become stable nuclides after only one decay. Potassium is widely distributed in the Earth’s crust and is a major element in many typical minerals; furthermore, it is the main natural radioactive source in natural water [14]. ⁴⁰K (T_1/2 = 1.26 × 10⁹ y) is the only radioisotope of potassium. Although the content of ⁴⁰K is only 0.012% in natural water, it contributes the most to β^- radioactivity [10]. ⁴⁰K stabilizes ⁴⁰Ca through β-decay (89%) and ⁴⁰Ar through electron capture (11%). When electron capture occurs, gamma rays are emitted with an energy of 1460.83 keV [15].

Nuclear activities such as global atmospheric nuclear tests, nuclear accidents, and nuclear waste recycling have generated numerous artificial radioactive materials [16]. Among them, artificial radionuclides spread in the environment along with their unstable isotopes, which have high radioactivity levels. Most of the artificial radionuclides decay to stable states through several decays (Table 3).

²⁴¹Am (T_1/2 = 432.2 y) is a transuranic nuclide that releases α particles by decay with 59.5 keV (78%) of gamma ray emitted, and its disintegration product is ²³⁷Np [15]. ¹³⁷Cs (T_1/2 = 30.2 y) typically appears in the wastewater of nuclear reprocessing plants and releases β-particle by decay with 661.657 keV (94.4%) of gamma ray emitted; its disintegration product is ¹³⁷Ba [15]. ¹³¹I (T_1/2 = 8.02 d) in the environment primarily originates from nuclear industries, nuclear accidents, and nuclear tests. It releases β^- particles by decay with 364.489 keV (83.6%) of gamma ray emitted; its disintegration product is ¹³¹Xe [17]. ⁶⁰Co (T_1/2 = 5.272 y) is a β^- emitter generated in a nuclear reactor; it emits 1173.228 (99.85%) and 1332.492 (99.98%) keV of gamma rays after decay, and its disintegration product is ⁶⁰Ni [15]. ⁹⁰Sr (T_1/2=28.9 y) is one of the fission products of uranium; it is a β^- emitter without gamma ray emission in nuclear waste, and its disintegration product is ⁹⁰Y [15, 18].

2.2. Formula of gross α and β^- activity concentrations

The activity concentration of radionuclides in water can be determined via gamma analysis. Because the emitting ratio of α/β^- particles and gamma rays of radionuclides is fixed, the gross α and β^- activity concentrations can be calculated by the intensity of the characteristic gamma. The equation for determining the gross α and β activity concentrations is shown in formula 1.

where A_G(α or β^-) is the gross α or β^- activity concentration of n radionuclides (in Bq/ L), A_i the activity of the i^th radionuclide in the sample, V the sample volume (117 L), and S_i the number of α or β^- particles by a single decay of the i^th radionuclide. The equation to calculate A_i is shown in formula 2.

where N_i is the pure count of a single peak of the i^th radionuclide; ε_i is the device detection efficiency of the selected peak of the i^th radionuclide, which is obtained via Monte-Carlo simulations and calibration experiments; P_i is the emission probability of the selected peak of the i^th radionuclide; T is the measurement period.

As shown in formula 1, the activity concentration of each radionuclide is key for determining the gross α and β^- activity concentrations. Therefore, the main α and β emitters in water with a high emission probability of gamma rays must be obtained. In the gamma spectrum, the nuclides’ options are extremely limited and peaks associated with α or β^- decay are isolated.

In addition to the radionuclides identified in the gamma analysis, some short half-life α and β^- emitters belonging to natural decay series in water without gamma radioactivity exist, such as ²¹⁸Po, ²¹⁴Po, ²¹⁶Po, ²¹²Po, and ²¹⁰Tl. Assuming that the half-life of the parents of these radionuclides is much longer than that of the daughter radionuclides, and that only the parent radionuclides exist at the initial time, the number of daughter radionuclides after time t can be calculated as follows:

where N_d is the number of daughter radionuclides at time t; N_op is the initial number of parent radionuclides;λ_p andλ_d are the decay constants of the parent and daughter, respectively [19]. After 10 times the half-life of the daughter, the number of parent radionuclides does not decrease, and the parent and daughter radionuclides are in radioactive equilibrium with the same activity. Hence, the activities of these radionuclides can be estimated based on the activities of the parents obtained via gamma analysis.

3.1 Online HPGe gamma measurement system

The instrument used in the experiment is an online gamma monitoring system, which can be used for the real-time continuous sampling and measurement of gamma radioactivity in water without pretreatment. The main structure of this system includes a low-background gamma spectrum measurement device (low-background lead chamber and HPGe detector), multichannel analysis unit, data storage system, communication unit (industrial computer and control system), and continuous water sampling device (Fig. 1).

Fig. 1Full-screen View

(Color online) Structure of online gamma monitoring system. Arrows inside dotted box represent direction of the water flow in operating state.

The HPGe detector was assembled in the lower background lead chamber. The entire measurement process was performed in the lead chamber, which blocked approximately 99.7% of external natural gamma rays. The HPGe detector used in the experiment was manufactured by ORTEC^@ of the United States. The measured energy range was 20 keV−10 MeV, the relative detection efficiency was 40% (1.332 MeV, ⁶⁰Co), the energy resolution (full width at half maximum) was 1.85keV (1.332MeV, ⁶⁰Co) and 870 eV (112 keV, ⁵⁷Co), and the peak-to-Compton ratio was 64:1 (⁶⁰Co).

The measurement process of this system is outlined as follows:

1) Two sampling pumps draw the river water into the sedimentation tank. After most of the solid residue in water has settled, the sample is pumped into the low background lead chamber, and the gamma radioactivity is measured using the HPGe detector.

2) The multichannel analyzer and computer display the gamma spectrum and analysis results on the screen.

3) If one or more radionuclides in the sample exceed the standard, then the sample is preserved in the pollution tank for more accurate measurements and analyses; otherwise, the sample is discharged back into the river through the valve.

3.2. Calibrations

3.2.1. Energy calibration

For energy calibration, ¹³⁷Cs (661.657 keV), ⁴⁰K (1460.83 keV), and ²⁰⁸Tl (2614.533 keV) were used as standard sources. Fig. 2 shows the energy (MeV) and channel diagram: E= 0.00098+0.0004∙Ch, where E is the energy (MeV) and Ch is the channel number of the multichannel analyzer.

Fig. 2.Full-screen View

Energy calibration curve of gamma spectroscopy system.

3.2.2 Detection efficiency of device

Owing to the large sample volume (117 L), the attenuation of gamma rays in water samples is a non-negligible problem in the measurement. The attenuation intensity of gamma rays with an energy of 0.661 MeV reached 92.5% after traversing 30 cm in water [20]. Considering the complexity of gamma characteristics in natural water, a method combining Monte-Carlo simulations and calibration experiments was used to determine the detection efficiency of the device.

3.2.2.1 Monte-Carlo simulation

The Monte-Carlo method is a numerical simulation method that is also known as random sampling or statistical experiment. It can accurately describe the physical experiment process and resolve problems that are difficult to solve using numerical methods.

The Monte-Carlo N Particle Transport Code (MCNP) is a universal neutron, photon, and electron transport program [21]. In this study, the detection efficiency of the gamma ray of each energy was simulated using the MCNP5 program. The simulation model of the detection device is shown in Fig. 3; the thickness of the lead chamber was 10 cm, the outside of the lead chamber was a 1-cm-thick stainless steel, and the inner layer was a 2-mm-thick oxygen-free copper. The outer and inner diameters of the lead chamber measured 72 cm × 87.5 cm and 50 cm × 60 cm, respectively. The HPGe detector, which was located in the center of lead chamber, was separated from the water in the lead chamber by polymethyl methacrylate.

Fig. 3.Full-screen View

(Color online) Simulation model of detection device.

The simulation detection efficiency of a single peak of the device (${\epsilon }_{\text{s}}(E_i)$) is defined as follows:

${\epsilon }_{\text{s}}(E_i)=C_i/N_i,$ (4)

where E_i is the energy of the i^th photoelectric peak in the gamma spectrum, C_i the pure area of the energy peak E_i, and N_i the gross number of gamma rays generated in water.

The detection efficiencies of ²²Na, ⁴⁰K, ⁵⁷Co, ²⁰⁸Tl, ²⁴¹Am, ¹³³Ba, ²¹⁴Bi, ⁵⁷Co, ¹⁵²Eu, ¹³¹I, ¹³⁷Cs, ¹⁹²Ir, ²²Na, ²²⁶Ra, ²³⁵U, and ²³⁸U were simulated and calculated, separately. The simulation results are shown in Table 4.

Table 4.Full-screen View

Detection efficiency of device generated via Monte Carlo simulation

Radionuclide	Energy(keV)	Detection efficiency(×10-3)	Radionuclide	Energy(keV)	Detection efficiency (×10-3)
22Na	511	1.549	152Eu	121.78	2.357
22Na	1274.5	1.130	152Eu	344.28	1.799
40K	1461	1.075	152Eu	778.9	1.335
57Co	122.06	2.357	152Eu	964.08	1.244
57Co	136.48	2.358	152Eu	1085.87	1.198
60Co	1170	1.169	152Eu	1112.07	1.183
60Co	1330	1.113	192Ir	295.96	1.915
131I	284.3	1.945	192Ir	308.46	1.882
131I	364.48	1.764	192Ir	316.51	1.862
131I	636.97	1.432	192Ir	468.07	1.599
137Cs	661.66	1.414	208Tl	510.84	1.547
133Ba	80.997	2.321	208Tl	583.84	1.475
133Ba	276.4	1.969	208Tl	860.37	1.291
133Ba	302.85	1.902	208Tl	2614.7	0.790
133Ba	356.01	1.780	235U	143.76	2.347
226Ra	186.21	2.245	235U	185.37	2.244
241Am	59.537	2.121	238U	66.38	2.218

Show more

3.2.2.2 Experiment verification

In the calibration experiment, the peak efficiency (${\epsilon }_{\text{sp}}(E_i)$) of the device is defined as follows [20]:

${\epsilon }_{\text{sp}}(E_i)=N_i/(A_i\cdot P_i\cdot T),$ (5)

where N_i is the gross count of the peak during the measurement period, also known as the peak area. B_i is the background count therein, A_i the activity of the radioactive source used in the calibration experiment, P_i the emission probability of E_i energy gamma rays by a single decay, and T the measurement period.

In the experiment, standard source solutions with different activity concentrations of ¹³⁷Cs (661.657 keV) and ⁴⁰K (1460.83keV) were prepared to verify the simulation results. In the preparation, certain CsCl and KCl powder concentrations were dissolved in purified water and diluted stepwise to 100, 10, 1, 0.8, 0.3, and 0.2 Bq/L. Table 5 presents the pure areas and detection efficiencies of ¹³⁷Cs (661.657 keV) and ⁴⁰K (1460.83 keV) standard source solutions.

Table 5.Full-screen View

Results of ¹³⁷Cs (661.657 keV) and ⁴⁰K (1460.83keV) detection efficiency

Activity concentration (Bq/L)	Photopeak net count		Detection efficiency (×10-3)
	40K (1460.83 keV)	137Cs (661.657 keV)	40K (1460.83 keV)	137Cs (661.657 keV)
0.2	102	1325	1.009	1.337
0.3	163	2073	1.075	1.395
0.8	438	5444	1.083	1.374
1	550	7349	1.088	1.484
10	5493	71041	1.087	1.434
100	54385	702679	1.076	1.419

Show more

As shown in Table 5, because of the radioactivity statistical variation, the detection efficiencies of each activity concentration differed, but the standard error was extremely low, indicating the stability of the instrument. The detection efficiencies for ¹³⁷Cs (661.657 keV) and ⁴⁰K (1460.83 keV) were $\text{1}{\text{.407×10}}^{\text{-3}}$ and $\text{1}{\text{.076×10}}^{\text{-3}}$, respectively, whereas the percentage error between the experimental detection efficiency and simulated values of ¹³⁷Cs (661.657 keV) and ⁴⁰K (1460.83 keV) was 0.71% and 0.52%, respectively, indicating consistency between the simulation and experimental results. Figure 4 shows the detection efficiency curve in the fitted full energy range (0–3 MeV).

Fig. 4.Full-screen View

Detection efficiency curve of spectrometric system. Detection efficiency of ¹³⁷Cs (661.657 keV) is $\text{1}{\text{.414 × 10}}^{\text{-3}}$ and that of ⁴⁰K (1460.83 keV) is $\text{1}{\text{.075 × 10}}^{\text{-3}}$.

As shown in Fig.4, the detection efficiency of the device first decreased and then increased as the energy of gamma rays reduced at a maximum energy of 140 keV; this was due to the severe attenuation of the low-energy gamma rays to the input window thickness, dead layer, and water. The detection efficiency of the high-energy gamma rays was limited to the sensitive volume of the detector.

4.1. Measurement in Zouma River

The Zouma River is the most important drinking water resource for residents in Chengdu; furthermore, it is the main source of irrigation water for agricultural production in the Chengdu Plain. As many nuclear facilities exist in the upper reaches of the Zouma River, radioactive pollution has garnered the attention of public and environmental organizations. To effectively determine the radiological characteristic and level of natural water in Chengdu city, in this study, two monitoring stations were established in the tributary of the Zouma River (Fig. 5).

Fig. 5.Full-screen View

(Color online) Locations of the Zouma River Basin and monitoring station in Chengdu, Sichuan, China.

The first monitoring station is located in Pixian County, the Northwest of Chengdu, which is at the upper reaches of the Chengdu section of the Zouma River. The Chengdu No.6 waterworks are located downstream of this monitoring station, which provides more than 1,053,000 m³ of drinking water to Chengdu residents daily [22]. In both the rainy and dry seasons, this monitoring station has sufficient and smooth water flows, facilitating sampling and analysis during monitoring.

The second monitoring station is located in the Dongfeng Canal in Chenghua District, Chengdu, which is one of the main irrigation water sources in Chengdu. Owing to the dense population and the large number of surrounding factories, the radioactive contamination level of natural water in the urban area of Chengdu can be determined rapidly and effectively by sampling and monitoring water in the Dongfeng Canal.

4.2. Measurement Results

Figure 6 shows the gamma spectrum of a water sample measured at the Zouma River monitoring station with a measurement period of 72 h. The isolated and intense peaks were marked and calibrated. In the spectrum, peak 1 (²²⁴Ra, 186.211 keV), peak 2 (²²⁶Ra, 240.986 keV), peak 3 (²¹⁴Pb, 351.932 keV), peak 5 (²⁰⁸Tl, 583.191 keV), peak 6 (²¹⁴Bi, 609.312 keV), and peak 7 (⁴⁰K, 1460.83 keV) exceeded the critical level of detection, and the radionuclides at these peaks generated α or β^- radioactivity. It is noteworthy that the peak count with an energy of 511 keV was the highest; however, the gamma rays might generate the electron interior effect in the water sample. The inner wall of the lead chamber and the cladding layer outside the HPGe detector produced annihilation photons with an energy of 511 keV, which might increase the counts of this peak. Hence, it cannot be used to calculate the gross α and β^- activity concentrations. Peak 3 was generated by ²¹⁴Pb (351.932 keV). Because the half-life of ²¹⁴Pb is short (26.8 min) and the lead chamber has been used for a long time, this peak is generated from ²¹⁴Pb in natural water instead of the lead chamber or lead dissolved in water. Combining formulas (1) and (2), ²²⁴Ra (186.211 keV), ²²⁶Ra, (240.986 keV), ²¹⁴Pb (351.932 keV), ²⁰⁸Tl (583.191 keV), ²¹⁴Bi (609.312 keV), and ⁴⁰K (1460.83 keV) were selected in this study to calculate the gross α and β^- activities.

Fig. 6.Full-screen View

γ-rays spectrum of water sample from Zouma River. The measurement period was 72 h. Peak 1 (²²⁴Ra, 186.211 keV), peak 2 (²²⁶Ra, 240.986 keV), peak 3 (²¹⁴Pb, 352.932 keV), peak 4 (511 keV), peak 5 and 9 (²⁰⁸Tl, 583.191 keV and 2614.533 keV, respectively), peaks 6 and 8 (²¹⁴Bi, 609.312 keV and 1764.464 keV, respectively), peak 7 (⁴⁰K, 1460.83 keV).

To illustrate the method to determine the activity in more detail, the spectral analysis results of two water samples from different monitoring stations are presented in Table 6.

The pure area of each peak was obtained as follows:

where N_p is the pure area of a single peak, N_G the gross count therein, B_n the count of natural background therein, and B_s the count of scattering background. Two physical gamma ray processes can result in a high scattering background on the spectrum. One is the self-absorption of the volume source (water), where the gamma rays present a continuous distribution of energy before entering the detector [20, 23]. The other is Compton scattering in the detector.

In addition, ²²⁰Rn (54.5s) and ²¹⁴Po (1.64 × 10^-4 s) are the short half-life daughters of ²²⁶Ra and ²¹⁴Bi, respectively; however, the normal state of ²²⁰Rn is gaseous, which is uncertain in the activity evaluation. ²¹⁴Po can be used as the foundation for activity determination. The α and β^- activity concentrations of each radionuclide of these two water samples were obtained, and the calculation results are shown in Table 7.

As shown in Table 7, the radionuclide with the highest activity concentration in both samples was ⁴⁰K. The activity concentration of each radionuclide in Pixian County ranged from 0.001 to 0.0362 Bq/L, and that in Dongfeng Canal ranged from 0.0016 to 0.0383 Bq/L.

In this study, the gamma radioactivity in the Pixian County monitoring station was monitored continuously from March 2018 to August 2018. Figure 7 shows the activity concentration curve during the measurement period. Table 8 shows the monthly average value of the calculation results of α and β^- activity concentrations of the water samples from the Pixian County monitoring station from March 2018 to August 2018.

Fig. 7.Full-screen View

α and β^- activity concentration curves from March 2018 to August 2018 (114 d) in Pixian County monitoring station.

As shown in Fig. 8, the average α and β^- activity concentrations were the highest in March and declined gradually thereafter. The maximum values of α and β^- activity concentrations were 0.0411 and 0.0854 Bq/L, respectively, whereas the minimum values were 0.0128 and 0.0322 Bq/L, respectively. The standard errors of the α and β^- activity concentrations ranged from 0.0013 to 0.0031 Bq/L and from 0.0064 to 0.0136 Bq/L, respectively. It was clear that the β^- activity concentration was approximately twice that of the α activity concentration.

Fig. 8.Full-screen View

Monthly average activity concentration of Zouma River in Pixian County monitoring station from March 2018 to August 2018 (114 d).

4.3. Testing

For this study, the Sichuan Management & Monitoring Center Station of Radioactive Environment (SMMCR) was commissioned to directly measure the α and β^- activities of the samples at the two monitoring stations to validate the method. Table 9 shows a comparison of the measured and calculated activity concentrations.

As shown in Table 9, both the measured and calculated results were less than the WHO guideline limits [3]. The absolute errors were less than 0.03 Bq/L, and the absolute errors of the gross β^- activity concentration were less than the α activity concentration. Nonetheless, the percentage error range was 9.90% to 89.32% because the water samples were directly sampled from drinking water resources without treatment. The samples without radioactive contamination indicated a lower radioactivity level. In addition, the comparison results of only two samples contain uncertainties [24, 25]. Therefore, the calculation results of 114 d were compared with the test data. Table 10 shows the historical monitoring data of the SMMCR.

As shown in Tables 10 and 11, the gross α and β^- activity concentrations of the Zouma River did not change significantly within four years (years 2015–2018). In March 2018, the absolute error of the gross α activity concentration between the monitoring data and calculation result was 0.0015 Bq/L (percentage error of 3.79%), whereas that of the β-activity concentration was 0.0124 Bq/L (percentage error of 17.04%). In September 2018, the absolute error between the monitoring data and the calculation result of the gross α activity concentration was 0.0189 Bq/L (percentage error of 52.94%), whereas that of the gross β^- activity concentration was 0.0160 Bq/L (percentage error of 33.20%). The historical data indicate that the α and β^- activity concentrations of the Zouma River in March were higher than those in September every year, consistent with the activity calculation curve. Meanwhile, the percentage errors of the data obtained in March 2018 were lower than those in September. Hence, when the radioactivity level increases, the percentage error between the calculated and test results will decrease.

However, the measured and calculated results differed to some extent. Some α ^- and β-emitters such as ⁹⁰Sr that could not be determined using this method appeared in natural water without gamma radioactivity [18]. Moreover, the backscattering peaks generated by ²¹⁴Bi (1112 keV) and ⁴⁰K (1460 keV) might be superimposed on the peaks of ²²⁶Ra (186 keV) and ²²⁴Ra (241 keV) in the spectrum, thereby affecting the ultimate results of α activity [26]. Regarding the test, because the measurement period (24 h) was much longer than the half-life of ²²⁰Rn, radon might escape during the measurement pretreatment, resulting in systematic errors in the measurement results [27]. Therefore, these influencing factors must be further investigated.