Radiochemical Characterization of Phosphogypsum for Engineering Use

The new phosphogypsum (PG) waste management policy allowed to reduce the negative environmental impact of this residue by finding better alternatives uses with an extremely limited radiological impact. Building material could be one of these alternatives that could lead to the production of final products with good mechanical properties and very limited radionuclides content. The optimization of the radioactive levels in the building materials when PG is used for its production requires the previous knowledge of the content of naturally occurring radionuclides in the PG waste. This article aims the radioactive characterization of two different PG sources (from Spain (Fertiberia S.A., Huelva) and Tunisia (Sfax), before being incorporated in building materials. For this purpose, the natural selected radionuclides content belonging to uranium and thorium decay series and K was determined, by means of two different methods: i) gamma spectrometry with high-purity germanium detectors and ii) laser-induced kinetic phosphorimetry (KPA-11 Chemcheck Instruments Inc., Richland, WA). Also, the semiquantitative chemical composition, the mineralogical study and the morphological aspect of the PG samples were analysed. The results obtained from both techniques show that Ra and Po are the main source of the radioactivity in both studied PG samples. However, PG samples from Tunisia present low natural radionuclide levels (30.7 Bq·kg average value for U, 188 Bq·kg(Ra), 163 Bq·kg(Pb), 12.4 Bq·kg (Th) compared to the level of natural radionuclides in PG samples from Huelva (102 Bq·kg average value for U, 520 Bq·kg(Ra), 881 Bq·kg(Pb) and 8 Bq·kg (Th). Both PG fulfil European Commission Recommendation (ECR) for the maximum activity concentrations of naturally-occurring radionuclides for industrial by product used in building materials in the European Union.


Introduction
Phosphogypsum (PG) is an industrial residue from processing phosphate rock using the "wet acid" process to produce the phosphoric acid (H 3 PO 4 ) in fertilizer plants (Equation ( 1)), which currently accounts for over 90% of phosphoric acid production.Ca 5 F(PO 4 ) 3 + 5H 2 SO 4 + 10H 2 O → 3H 3 PO 4 + 5CaSO 4 •2H 2 O + HF (1) This process is economic however it results in the generation of a large amount of PG (for every ton of phosphoric acid produced, about 5 tons of PG are yielded).The worldwide generation is estimated to be around 100 -280 Mt per year [1,2].PG consists principally of calcium sulphate (CaSO 4 •2H 2 O) but also contains a high level of impurities such as phosphates, fluorides and sulphates, naturally occurring radionuclides, heavy metals, and other trace elements.
Nowadays, PG represents one of the most serious problems faced by the phosphate industry, since commercial uses, in manufacturing gypsum board and Portland cement and in agricultural fertilisers or soil stabilisation amendments, consume less than 15% of the worldwide generation of PG.The remaining 85% is disposed of without any treatment and usually dumped in large stockpiles exposed to weathering processes, occupying considerable land areas and causing serious environmental contamination of soils, water and the atmosphere, particularly in coastal regions [3].The main prob-lem associated with the storage of PG is considered to be the relatively high levels of natural uranium-series radionuclides, naturally present in the phosphate rock and which provoke a negative environmental impact and many restrictions on the use of this residue.Depending on the quality of the phosphate rock source, PG can contain as much as 60 times the levels normally found prior to processing.Previous study performed by Bolivar [4] showed that about 80% of the 226 Ra, 90% of the 210 Po and 20% of the 238 U and 234 U originally present in the phosphate rock remain in PG.Furthermore, the most important source of PG radioactivity is reported to be 226 Ra [5]. 226Ra produces radon gas ( 222 Rn), which has a short half-life of 3.8 days, an intense radiation capacity, and causes significant damage to internal organs [6].Thus the potential problem of PG piles is the emanation of 222 Rn from the alpha-decay of 226 Ra, a radionuclide classified by the USEPA as a Group human carcinogen, whose common presence in PG led to the regulation of PG disposal under the National Emission Standards for Hazardous Air Pollutants (NESHAP) and the National Emission Standards for Radon Emission from PG Stacks [7].The United States Environmental Protection Agency (USEPA) classified PG as a "Technologically Enhanced Naturally Occurring Radioactive Material" (TENORM) [6] and PG exceeding 370 Bq/kg of radioactivity has been banned from all uses by the EPA since 1992.The maximum regulatory limit of 222 Rn exhalation (the flux density of 222 Rn gas entering the atmosphere from the surface of a 226 Ra-bearing material) established by the EPA [8] is 0.74 Bq/m 2 /s.
In Huelva (Spain), PG stacks located on salt marshes contain about 100 Mt of PG (area of approx.1200 ha with average height of 5 m) and are generally not completely watertight or even covered with any inert material, leading to a local gamma radiation level between 5 and 38 times the normal rate (0.74 Bq/m 2 /s) [9].The same situation is observed in Sfax (Tunisia), where PG is accumulated in two enormous warehouses situated at the coastal strip of the urban area, the first one is 12 m high and covers an area of 40 ha, and the other one, 30 m high, covers an area of 60 ha.
The present study was conducted to determine the natural selected radionuclides content belonging to uranium and thorium decay series and 40 K in the two different PG sources mentioned above, using two different methods.

Material and Methods
The PG samples used in this study came from a fertiliser factory in Sfax city, Tunisia and from Fertiberia S.A., Huelva, Spain.
The semiquantitative chemical composition of the PG samples was identified by an X-ray fluorescence analyser (Philips model PW-1404 sequential wavelength dispersion unit).Mineral species were determined by X-ray diffraction (Siemens model D5000, with a Cu tube and LiF monochromator).The morphological aspect of the PG was analysed using scanning electron microscopy (SEM) (Joel model JXA-840) with energy dispersive spectroscopy (EDS).Natural selected radionuclides belonging to uranium and thorium decay series and 40 K present in PG samples have been quantified as shown in the Uranium: the uranium content in the samples was determined using two different methods: 1) Direct measurements by gamma spectrometry with high-purity germanium detectors and 2) Laser-induced kinetic phosphorimetry.The direct measurements were carried out on 700-g aliquots of the samples packed in standard marinelli beakers.The 238 U activity concentration was determined through the photopeaks of its immediate decay product, 234 Th (63 and 92.5 keV), whereas 235 U was measured directly from its 143.8 and 163.4 keV gammaray peaks.Concerning Laser-induced kinetic phosphorimetry technique, 1 g of the sample was completely digested in 15.6 mol•l −1 HNO 3 and the measurements were performed using a kinetic phosphorescence analyzer (KPA-11) (Chemcheck Instruments Inc., Richland, WA) [10].In order to compare the results obtained by both techniques, the total uranium concentration obtained by Laser-induced kinetic phosphorimetry, expressed in µg•g −1 , were then converted to the activity concentration of each uranium isotope.Theoretical values of the isotopic composition of natural uranium (99.3% 238 U, 0.72% 235 U, and 5.5•10 −3 % 234 U) and the specific activities of these isotopes in natural uranium (Bq•g −1 ) were used for this purpose [11].
Polonium: the polonium activity concentration in the both samples was determined by alpha spectrometry, following the 210 Po separation procedure described in the Figure 2 [12].An aliquot of 1 g was digested in a hot plate at a controlled temperature (< 90˚C), using 8 mol•l −1 HNO 3 .Polonium-209 standard dissolution was added to the dissolved samples as a tracer to estimate the recovery of the whole process.The polonium isotopes were self-deposited on silver disks following Flynn's method [13]. 226Ra, 232 Th, 210 Pb and 40 K: the concentrations of these radionuclides were quantified by gamma spectrometry analysis using an HPGe detector.The detector was shielded from external radiation by an iron wall (15 cm thickness).The emission gamma spectrum was analyzed using Genie-2000 application software.To ensure radioactive equilibrium between 226 Ra and its short lived decay products, the samples (700 g aliquot) were packed in standard marinelli beakers, hermetically sealed and stored for about four weeks prior to counting.The concentration of 226 Ra and 232 Th was estimated from their daughters gamma-ray photopeaks, 214 Bi (609 keV) and 228 Ac (911.2 keV, 969.0 keV) respectively. 210Pb and 40 K were measured directly from their gamma-ray emissions at 46.5 keV and 1460.8 keV respectively.The minimum detectable activity limit (DL) was also calculated.

Phosphogypsum Characterization
The chemical composition of both type of PG sample is summarised in Table 1.The data shows that sulphate (expressed as SO 3 ), CaO, SiO 2 and P 2 O 5 are the major elements (50.7, 41.24, 1.38 and 1.2%, respectively) for Tunisian PG, and 52.6, 42.82, 2.72 and 0.7, respectively for Spanish PG.
The The morphological study of PG sample using SEM, illustrated in Figure 4, shows two different sections of the sample.The micrographs reveal a homogeneous and prismatic PG piling arrangement and a well-defined crystalline structure with a majority of orthorhombic shaped crystals [14].Similar results can be observed in the study performed by Miloš and Dragan [15], in which they reported that the marked crystal structure of PG indicates that PG presents a more complex composition than natural gypsum (characterized by a poorly expressed crystalline structure), which may eventually influence its chemical behaviour.

Natural Radionuclide Concentrations
The results of the natural radioactivity concentration analyses of each one of the PG samples by the two techniques employed for this purpose (gamma spectrometry and laser-induced kinetic phosphorimetry) are listed in Tables 2 and 3.The average activity concentration of 238 U, 226 Ra, 210 Po, 232 Th and 40 K in Tunisian PG samples are 30.7,188, 194, 12.4 and 13 Bq•kg −1 , respectively, while in Spanish PG samples are 102, 520, 820, 8 and 39 Bq•kg −1 , respectively.We can deduce that 226 Ra and 210 Po are the main source of the radioactivity in both PG     [18].This different content may be attributed to the nature of the phosphate rock, the depth of sampling [19] and differences in the industrial process applied to obtain phosphoric acid.Natural radioactivity in the different phases of the production system has recently been analysed by Bolivar et al. [20] showing that Pb, Ra and to a certain extent Th isotopes are exclusively supplied by the phosphate rock and remain associated to the PG particles, while uranium decreases according to the number of washings of the PG.
The data showed that the both studied PG samples (Tunisian and Spanish) fulfil European Commission Recommendation for the maximum activity concentrations of naturally-occurring radionuclides in common buildings materials and industrial by-products used for building materials in the EU [21].This finding could make possible the use of the studied PG as building material.
Concerning the measurement techniques used in this study, it found that the use of gamma spectrometry for uranium determination allows the analysis of a more representative aliquot of the whole sample than in the case of the KPA technique.KPA has a lower detection limit (sensitivity) and better uncertainty (6%), but due to the limitations of wet digestion until total dissolution and chemical interferences, only 1 g can be analysed.Nevertheless, the results obtained from both techniques are in good agreement.

Conclusions
The sensitivity of the analytical methods is good enough to detect the radionuclides existing in both types of samples.
For uranium quantification, the gamma spectrometry allows the analysis of a more representative aliquot of the whole sample than in the case of the KPA technique.However; KPA has a lower detection limit (sensitivity) and better uncertainty (6%).

Table 2 . Uranium activity ratio in PG from Spain and Tunisia, expressed in Bq•kg −1 (± 2 s), by means of phosphorimetry and gamma spectrometry techniques.
* DL.Detection Limit