External radiation exposure of the Angolan population living in adobe houses

Cabinda

In Angola most of the main cities have the same name as the province they belong to. This is the case of Cabinda which is the main city of the Cabinda province. It is located at 12° 12′ E and 5° 30′ S with a mean altitude of 1 m above sea level (Fig. 1). The province of Cabinda is an enclave located at the North-West part of the country. Its climate is tropical humid.

The geology of the province is characterized by various formations aging from Precambrian to Holocene, including Pleistocene marine deposits at the coastal area.

Huambo

Huambo is the capital city of the Huambo province. It is located 1700 m above the sea and the coordinates are 15° 45′ E and 12° 48′ S (Fig. 1). The province of Huambo is located at the central part of the country. Its climate is tropical highly influenced by the altitude and the cold current of Benguela.

The geology of the province consists of old (Archean to Proterozoic) metamorphic (gneiss, micaschist, metasediments) and igneous (granite, rhyolite to andesites) rocks. At Huambo city Paleocene-Eocene laterites were also mapped.

Menongue

Menongue is the main city of the Cuando Cubango province. It is located at 17° 41′ E and 14° 39′ S with an elevation of 1300 m above sea level (Fig. 1). The province of Cuando Cubango is located at the south west part of the country. Its climate is semi-desert influenced by the desert of Namibe.

Its geology is based on Archean mostly gneiss and Proterozoic rhyolite-andesite and Tertiary–Quaternary Kalahari sediment formations.

In-situ ambient gamma dose rate

The portable device FH 40 G-L10 (Thermo Fisher Scientific Inc.) was used to perform the in situ ambient gamma dose equivalent rate measurements. The detection limit of the mentioned survey meter goes from 10 nSvh⁻¹ to 100 mSvh⁻¹ and from 30 keV to 4.4 MeV regarding the gamma energies. Measurements were done in the living rooms of 45 houses at two or three different points, at 1 m height.

Building material sampling

A total of 60 adobe samples were collected in the three study areas, making 20 per area. The materials were sampled from (i) pieces of remaining adobes after construction, (ii) pieces took directly from houses, and (iii) soil used to make the adobe, in a few cases. The sampling sites were selected randomly within the studied cities (Fig. 1).

Sample preparation

The sample preparation was carried out in the Lithosphere Fluid Research Lab at the Eötvös Loránd University, Budapest. Samples were dried at lab temperature (around 20 °C) at least for a month. Then they were powdered and sieved below 0.5 mm diameter to ensure homogeneity. Afterwards the homogeneous samples were dried in an oven at 40 °C until constant weight is reached.

The dry, homogeneous samples were poured into sample holders made of High-Density Polyethylene (HDPe) which are proven to be radon-tight [15]. There was a need to artificially compact the sample in the sample holder in order to avoid a later self-compaction, because it could have led to an empty space inside the sample holder where ²²²Rn can accumulate. Such inhomogeneity leads to false results [15]. After the filling, samples were left on rest for a minimum of 30 days to ensure the secular equilibrium between ²²⁶Ra, ²²²Rn and its daughters.

Gamma-ray spectroscopy

Gamma-ray spectroscopy measurements were carried out at the Centre for Energy Research, Hungarian Academy of Sciences, Budapest. An n-type HPGe detector (Canberra GR1319) with a relative efficiency of 13%, and a Canberra DSA-2000 acquisition system were used with a low background chamber [15]. The energy resolution (FWHM) is 1.53 and 1.99 keV at 662 and 1332 keV, respectively. The peak-to-Compton ratio is 42.8:1 at 1332 keV. Dead-time losses during the experiments never exceeded 0.05%. The efficiency transfer method was applied for the determination of the full-energy-peak efficiency for the extended samples in close-in geometry. More details on the measurement method are explained by Kis et al. [15].

Samples were measured for 42–120 h (samples with lower radioactivity needed more time) to reach good statistics. The determination of ²²⁶Ra was made from gamma lines of ²²²Rn daughters, ²¹⁴Pb (242, 295 and 352 keV) and ²¹⁴Bi (609, 1120 and 1765 keV), assuming the secular equilibrium inside the sample holder. ²³²Th was evaluated using gamma lines of ²¹²Pb (239 and 300 keV) and ²²⁸Ac (338, 463, 911, 965 and 969 keV) assuming the secular equilibrium in the ²³²Th chain as well. K-40 was evaluated from its own peak (1461 keV).

Hazard indexes calculations

For the qualification and classification of building materials, hazard indexes were determined from the activity concentrations of ²²⁶Ra (C²²⁶Ra), ²³²Th (C²³²Th) and ⁴⁰K (C²³²Th).

Dose estimations

In this work annual effective doses (AED) were determined in two distinct ways: in the first method it was determined from the in situ ambient gamma dose equivalent rate measurements (AED-measured) and in the other it was calculated from the activity concentrations of ²²⁶Ra, ²³²Th and ⁴⁰K of the building material (AED-calculated).

Statistics

Descriptive statistics as mean, average and standard deviation have been determined using the software program “Origin”. The data also have been summarized in box and whisker plots for a comparison of the results obtained at different studied locations.

Pearson correlation coefficients [18] have been determined for the variables using the software program “Statgraphics”. All the statistical tests were performed at the 95% confidence level, in this program.

Measured ²²⁶Ra, ²³²Th and ⁴⁰K activity concentrations

Average and standard deviations of the activity concentrations in Bq kg⁻¹ are 26 ± 7, 36 ± 5, 45 ± 17 in Cabinda, 87 ± 20, 81 ± 21, 82 ± 15 in Huambo and 27 ± 10, 30 ± 10, 73 ± 40 in Menongue for ²²⁶Ra, ²³²Th and ⁴⁰K, respectively (Table 1; Fig. 2a–c).

Table 1

Summary of averages, standard deviations and ranges of activity concentrations, Radium equivalent index (Ra_eq), I index and calculated and measured annual effective doses

Location		Activity concentrations (Bq kg−1)			Raeq (Bq kg−1)	I index	Annual effective dose-measured (mSv year−1)	Annual effective dose-calculated (mSv year−1)
		226Ra	232Th	40K
Cabinda	Average	26 ± 7	36 ± 5	45 ± 17	81 ± 13	0.28 ± 0.05	0.6 ± 0.13	0.4 ± 0.06
	Range	45–15	49–27	92–28	121–56	0.42–0.19	0.86–0.48	0.50–0.22
Huambo	Average	87 ± 15	81 ± 21	82 ± 15	209 ± 45	0.72 ± 0.15	1.5 ± 0.32	0.8 ± 0.2
	Range	116–49	121–39	105–50	273–113	0.95–0.39	2.28–1.12	1.11–0.47
Menongue	Average	27 ± 10	30 ± 10	73 ± 40	76 ± 24	0.26 ± 0.08	1.04 ± 0.16	0.3 ± 0.06
	Range	56–15	57–17	155–21	145–44	0.51–0.16	1.24–0.64	0.62–0.18

Open image in new window

Correlations among ²²⁶Ra, ²³²Th and ⁴⁰K activity concentrations

Correlations between ²²⁶Ra and ²³²Th was found to be 0.66 in Cabinda with a P value of 0.0014 (Table 2; Fig. 3a), 0.55 in Huambo with a P-value of 0.012 (Table 2; Fig. 3b) and 0.80 in Menongue with a P-value of 0.0001 (Table 2; Fig. 3c). According to this result, in all areas the mentioned radionuclides have a statistically significant relationship.

Table 2

Correlations among the three studied radionuclides

Location	226Ra and 232Th correlation	226Ra and 40K correlation	232Th and 40K correlation
Cabinda	R = 0.66	R = 0.62	R = 0.36
	P-value = 0.0014	P-value = 0.0036	P-value = 0.115
Huambo	R = 0.55	R = 0.19	R = 0.5
	P-value = 0.012	P-value = 0.41	P-value = 0.023
Menongue	R = 0.8	R = 0.23	R = 0.48
	P-value = 0.0001	P-value = 0.338	P-value = 0.0312

Open image in new window

Correlations between ²²⁶Ra and ⁴⁰K show the following results: 0.62 with a P-value of 0.0036 in Cabinda (Table 2; Fig. 4a), 0.19 with a P-value of 0.41 in Huambo (Table 2; Fig. 4b) and 0.23 with a P-value of 0.338 in Menongue (Table 2; Fig. 4c). Statistically significant relationship is found in Cabinda, but not in the two other areas.

Open image in new window

Results of correlations between ²³²Th and ⁴⁰K are as follows: 0.36 with a P-value of 0.115 in Cabinda (Table 2; Fig. 5a), 0.50 with a P-value of 0.023 in Huambo (Table 2; Fig. 5b) and 0.48 with a P-value of 0.0312 in Menongue (Table 2; Fig. 5c). Here we have statistically significant relationship in Huambo and Menongue, but not in Cabinda.

Open image in new window

Ra_eq and I hazard indexes

Averages and standard deviations of the Ra_eq index are 81 ± 13 Bq kg⁻¹ in Cabinda, 209 ± 45 Bq kg⁻¹ in Huambo and 76 ± 24 Bq kg⁻¹ in Menongue. For the I index, averages and standard deviations are 0.28 ± 0.05 in Cabinda, 0.72 ± 0.15 in Huambo and 0.26 ± 0.08 in Menongue (Table 1; Fig. 6a, b).

Open image in new window

Doses

Measured in situ ambient gamma dose equivalent rates present the following averages and respective standard deviations: 0.6 ± 0.13, 1.5 ± 0.32 and 1.04 ± 0.16 mSv year⁻¹ in Cabinda, Huambo and Menongue, respectively (Table 1; Fig. 7). Averages of the annual doses calculated from the activity concentration of the radionuclides in building materials with respective standard deviation are: 0.4 ± 0.06, 0.8 ± 0.2 and 0.3 ± 0.06 mSv year⁻¹ in Cabinda, Huambo and Menongue, respectively (Table 1; Fig. 7).

Open image in new window

Influence of geology on natural radiation in Angola

The influence of geology on the spatial distribution of natural radiation has been proven by many studies [19, 20, 21, 22]. Different gamma dose values can relate to different geological backgrounds [23, 24, 25]. It is important to recall that the studied building material (i.e., adobe) is made of the local soil, which develops via weathering process from the surrounded rocks. The composition of the source rock eventually controls the natural radioactivity of the soils of the area [26] and therefore of adobes. Results of present study have shown different levels of radioactivity at the three studied areas of Angola that have different geological backgrounds [14]. As seen in the results, the North part (Cabinda) and the South part (Menongue) show the lowest average activity concentrations: 26 ± 7, 36 ± 5, 45 ± 17 and 27 ± 10, 30 ± 10, 73 ± 40 for ²²⁶Ra, ²³²Th and ⁴⁰K, respectively. Adobe samples collected in Cabinda are mostly made of soils derived from rocks belonging to Pleistocene marine sediment formations consisting mostly of sand deposits. Adobe samples from Menongue are from soils formed from rocks belonging to Eocene-Pliocene sand formations. In contrary, the central part of the country (Huambo) with Archean metamorphic (gneiss and micaschist) and Proterozoic metasediment formations shows the highest average activity concentrations (Fig. 2a–c; [14]). The above described formations are known for their potential to contain elevated concentrations of radioactive minerals [27].

Geochemistry of the studied radionuclides

In order to understand the mutual effects of the three studied radionuclides, ²²⁶Ra, ²³²Th and ⁴⁰K, it is important to analyze their correlations and consequently, their relationships. The strongest correlation between ²²⁶Ra and ²³²Th is found in Menongue (R = 0.8, Fig. 3c). The lithology made up of Eocene–Pliocene sediment formations in Menongue is rather homogeneous [14]. In areas with such a characterization, correlations between ²²⁶Ra and ²³²Th are expected to be high [28]. This strong correlation, found between ²²⁶Ra and ²³²Th, may be explained by the similar geochemical behavior of the parent and progeny nuclides, which infers similar responses to the soil and environmental processes that affect their distribution [26]. Uranium and thorium have the same replacement capacity of a nucleus in the lattice of a mineral because of their radius length. Moreover, despite the fact that uranium tend to be more mobile than thorium during weathering processes, studies have proven the immobility of uranium due to adsorption or co-precipitation by amorphous Fe-oxyhydroxides [27]. Therefore, Dequincey et al. [29] had the same result of uranium and thorium accumulation in the indurated ferruginous cap of African laterites, and of their secondary redistribution into the profile.

Regarding the correlation between ²²⁶Ra and ⁴⁰K, the only area with a statistically significant relationship is Cabinda (Fig. 4). Samples from Cabinda were taken at the littoral part where chemical weathering is not profound, whereas, in the interior part of the country (Huambo, Menongue) not only the rocks but the soils are highly weathered [30, 31]. Both studies by Wilford et al. [32] and Dickson and Scott [33] explain the influence of the weathering process and the topography on the radionuclide distribution [34]. Potassium is more easily weathered and leached out than radium in soils because of its higher solubility and probably hygroscopic feature. Moreover, upper horizons in soils are poor in potassium because of the plant uptake [35].

Contrariwise, to the results from the correlation between ²²⁶Ra and ⁴⁰K, statistically significant relationship exists between ²³²Th and ⁴⁰K in Huambo and Menongue but not in Cabinda (Fig. 5). In some cases for instance when thorium is hydrated [27], the behavior of potassium and thorium is similar during the weathering process. In these cases, both elements are mobile during the weathering process as demonstrated on the study of Dickson and Scott [33]. Consequently, the average potassium and thorium content of soils reflect the average of the rocks content on the same elements from which they are derived. However, the differences in soil radioelement concentrations are relatively small [34].

Overall, the correlation among radionuclides is the strongest between ²²⁶Ra and ²³²Th in all areas which was also found in adobes and soils, for instance, by Szabó et al. [36] and El Afifi et al. [37]. However, as extensively demonstrated by Wilford et al. [32], in the study about different Australian soils, weathering processes and topography might have a big influence on the behavior of the radionuclides, consequently their distribution in the soil.

Measured, calculated and estimated levels in this study