Analyses of uranium and fluoride in diammonium phosphate fertilizers marketed in India during 2021-2022

Lokeshkumar P Ramteke¹, Raghav S Soni¹, Babulal Rebary², Ashwin W Patwardhan^1*, Dilip D Sarode¹ and Pushpito K Ghosh^1*
*Correspondence: Ashwin W Patwardhan, Department of Chemical Engineering, Institute of Chemical Technology, Mumbai, India, Email: Pushpito K Ghosh, Department of Chemical Engineering, Institute of Chemical Technology, Mumbai, India, Email:

Received: 18-Apr-2022, Manuscript No. IJMF-22-61096; Editor assigned: 20-Apr-2022, Pre QC No. IJMF-22-61096 (PQ); Reviewed: 06-May-2022, QC No. IJMF-22-61096; Revised: 13-May-2022, Manuscript No. IJMF-22-61096 (R); Published: 21-May-2022

Keywords

Phosphate fertilizers, DAP, uranium, fluoride, analysis, farmland pollution

Introduction

Rock Phosphate (RP) is the primary feedstock for production of phosphate fertilizers. It is increasingly being viewed as a strategic material (Gilbert, 2009). 270 Million Metric Tonnes (MMT) of RP was mined in 2018, with China, Morocco, United States and Russia accounting for the bulk of the production (Blackwell et al., 2019). The P₂O₅ content in beneficiated rock varies from 32% to 39% w/w (Table 1) (EFMA, 2000). RP exists as fluorapatite [Ca₁₀(PO₄)₆(F,OH)₂] and francolite [Ca₁₀(PO₄)_6–x(CO₃)_x(F,OH)_2+x] in igneous and sedimentary deposits, respectively, the fluoride content varying in the range of 2% to 4% w/w (Van Kauwenbergh, 2010). Popular phosphate fertilizers produced from RP include: Single Super Phosphate (SSP) [3Ca(H₂PO₄)₂. 7CaSO₄. 2H₂O], Triple Super Phosphate (TSP) [Ca(H₂PO₄)₂] and Di Ammonium Phosphate (DAP) [(NH₄)₂HPO₄]. Theoretical P values as per the compositions are: 11.1%, 26.7% and 23.5%, respectively; the corresponding figures expressed in terms of P₂O₅ content are: 22.4%, 60.7% and 53.8%. SSP and TSP are obtained from the reaction of RP with H₂SO₄ (Equation (1)) and H₃PO₄ (Equation (2)), respectively, whereas DAP is obtained from the reaction of H₃PO₄ with NH₃ (Equation (3)).

H₃PO₄ can be produced from RP by the ‘thermal route’ or through the reaction of Equation (4), commonly referred to as the ‘wet process’. The latter process has been adopted in fertilizer production for reasons of cost-effectiveness.

2Ca₅(PO₄)₃(F)+7H₂SO₄ → [3Ca(H₂PO₄)₂+7CaSO₄. 2H₂O] ↓+2HF ↑ (1)

Ca₅(PO₄)₃(F)+7H₃PO₄ → 5Ca(H₂PO₄)₂+HF ↑ (2)

H₃PO₄+2NH₃ → (NH₄)₂HPO₄ (3)

Ca₅(PO₄)3(F)+5H₂S O₄ → 3H₃PO₄+5CaSO₄. 2H₂O ↓+HF ↑ (4)

The fluoride content in RP varies in the range of 2% to 4% w/w (Table 1). Some of the fluoride is released as HF/H₂SiF₆/ SiF₄ in the course of fertilizer and H₃PO₄ production (Eq. 1, 2, 4). The extent of separation depends on process parameters but it is never complete. Commercial processes are therefore operated to further eliminate fluoride from H₃PO₄ (Samrane, 2011). There are several reports of fluoride pollution emanating from phosphate fertilizer plants (Dartan, 2017; Dolar, 2011).

RP contains other harmful constituents also, uranium being among these. In one of the earliest studies reported, the uranium content in 316 RP samples collected from different regions of the world were analysed on a scintillation counter. The concentrations varied by more than two orders of magnitude – from 3 mg [U].kg^-1 in Chile and Ecuador to 399 mg [U]. kg^-1 in South Carolina, USA – the median value being 59 mg [U].kg^-1 (Menzel, 1968). Uranium concentrations are also reported for commercially important RP sources (Table 1). The concentrations range from 9.3-156.9 mg [U].kg^-1 (EFMA, 2000). Advanced analytical techniques such as X-ray fluorescence spectroscopy (XRF), Laser Induced Breakdown Spectroscopy (LIBS) and Low Background Counting (LBC) have been compared recently for uranium estimation in RP (Al-Eshaikh, et al., 2016). Employing high energy resolution fluorescence detected X-ray absorption near-edge structure spectroscopy (HERFD-XANES), it has been further shown that 40% to 60% of uranium in RP exists as U (IV) and the rest as U (VI) (Vogel, 2020). Phosphogypsum waste obtained in the course of phosphoric acid production from RP reportedly contains uranium [ca. 14% of the uranium coprecipitates with gypsum (Eq. 4) and the rest finds its way into the acid (Zohuri et al., 2020)] and other radioactive constituents, leading to significant pollution in the vicinity of phosphate fertilizer plants (EPA, 2015).

The problem of fluoride and uranium emanating from processing of RP is not confined to the factory premises and adjoining areas alone. These hazardous constituents can end up in fertilizers if not fully removed during the manufacturing process. Their application in agricultural fields may, in that case, spread the contaminants far and wide and cause nonpoint source pollution. It has been reported, for example, that the fluorine concentration in agricultural soils in New Zealand has increased with time. The build-up was attributed to the widespread application of phosphate fertilizer year on year (Geretharan et al., 2020). India has a large agricultural base and with the growth in population, demand for fertilizers has been rising steadily. A study of fluoride concentrations in four SSP samples, one Ammonium Nitro Phosphate sample (ANP) and two DAP samples by Ion Chromatography (IC) and Ion Selective Electrode (ISE) was conducted in 2018. It can be deduced from the data that the average concentration of fluoride was 42.7 g.kg_-1 P₂O₅ (Ramteke et al., 2018). The figure was higher in SSP samples (53.44 g.kg^-1 P₂O₅ on average) and lower (13.15 g.kg^-1 P₂O₅; average of two samples) in DAP samples. Four additional DAP samples collected during 2019 revealed that the problem had remained largely unchecked, with average fluoride concentration of 19.95 g.kg^-1 P₂O₅ (Ramteke et al., 2019).

Presence of uranium in phosphate fertilizers has also been reported by various laboratories (Roessler et al., 1979; Lal et al., 1985; Yamazaki and Geraldo, 2003; Schnug and Haneklaus, 2015; Vogel, et al., 2020). The adverse impacts arising from application of such fertilizers on water bodies and soil have also been documented (Spalding and Sackett, 1972; Kratz and Schnug, 2006; Hoyer, 2015; Cioroianu et al., 2001; Jiao et al., 2012). The analytical techniques that have been deployed for uranium analysis in fertilizers and water bodies include: Fission Track (Lal et al., 1985; Yamazaki and Geraldo, 2003), Delayed Neutron Technique (Spalding and Sackett. 1972), Gamma Spectrometric Analysis (Roessler et al., 1979; Khater, 2008), Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES) (Souza, 2016), Inductively Coupled Plasma Mass Spectrometry (ICP-MS) (Vogel et al., 2020), and IC (Al-Shawi and Dahl, 1995).

Lal et al. reported over three decades ago that the uranium content in Indian phosphate fertilizers was in the range of 15.9- 35.8 mg.kg-1, with linear dependence on the P content of the fertilizer (Lal et al., 1985). The data were consistent with the relatively low uranium content (12 mg.kg^-1, on average) in Indian RP (Menzel, 1968). There is, however, paucity of recent data, even as phosphate fertilizer consumption has increased from 1.886 MMT P2O5 during 1984-1985 to 6.910 MMT P₂O₅ during 2018-2019 (Fertilizer Association of India, 2021), P₂O₅ consumption was reportedly 9.268 MMT [11.911 MMT of DAP (46% P₂O₅), 4.489 MMT of SSP (16% P₂O₅), and 11.811 MMT of NPKS complex fertilizer (26% P₂O₅)] during 2020-21 (April- March) (Indian Express, 2021). Indigenous production of RP, on the other hand, has been steadily declining – it was only 1.40 million tonnes in 2019-2020 (Indian Minerals Yearbook 2020) – necessitating import of the bulk of the P₂O₅ requirement in the form of RP, phosphoric acid and finished phosphate fertilizers. Import of DAP alone was 6.1 MMT (Yao, 2020).

Given the heavy consumption of phosphate fertilizer in India and the growing dependence on imports from diverse sources, it is important to monitor commercial phosphate fertilizers not only for their nutrient values but also for the presence of hazardous constituents. Further to our previous work, we report herein new data on fluoride concentrations in fifteen new DAP samples and one MAP sample collected from various regions of India during 2021-2022. Analyses were conducted by IC and ISE methods reported previously (Ramteke et al., 2018). We also report uranium data of 27 phosphate fertilizer samples collected during 2016-2022. The first lot of samples were analysed for uranium by four different techniques [Laser Fluorimetry (LF) (Veselsky, 1988), ICP-OES (Li et al., 2021), ICP-MS (Vogel et al., 2020), XRF analysis (Al-Eshaikh et al., 2016)]. Rest of the samples were analyzed for uranium by ICP-MS alone.

Materials and Methods

Materials

SSP samples 1-4, DAP samples 1-6 and ANP 1, were from previous collections (Ramteke et al. 2018, 2019). DAP samples 7-21 and one monoammonium phosphate sample (MAP 1) were collected during 2021-2022. Sample details of 2021-2022 collections are provided in the Table 2. P₂O₅ and N values in the procured fertilizer samples were checked through ICPMS and IC analyses (refer to detailed procedures under Analytical methods below). P₂O₅ and N concentrations in DAP were in the range of 49.9% to 57.9% w/w and 16.6% to 20.3% w/w against the theoretical values of 53.8% and 21.2% w/w.

Analytical methods

SSP 1-4, DAP 1-2 and ANP 1 were analysed for uranium by ICP-OES [National Centre for Compositional Characterisation of Materials (NCCCM), Hyderabad], LF [Bhabha Atomic Research Centre (BARC), Trombay] and ICP-MS [Central Salt and Marine Chemicals Research Institute (CSMCRI), Bhavnagar; Institute of Chemical Technology (ICT), Mumbai]. SSP 2 and DAP 1 were analysed by XRF additionally. MAP 1 and DAP 7-22 were analysed by ICP-MS only. All samples for uranium analyses were sent to the respective laboratories in coded form. There was no exchange of data among the laboratories engaged in the above analyses.

Analyses of uranium by ICP-OES at NCCCM, Hyderabad: Fertilizer sample (100 mg) was dissolved in 100 mL deionised water and analysed by ICP-OES without any further dilution. Analyses were carried out using an ICP- OES (Ultima 2, Horiba Jobin-Yvon, France) equipped with a concentric nebulizer coupled to a cyclonic spray chamber. The Ultima 2 spectrometer comprises a Czerny-Turner monochromator equipped with a 2400 grooves/mm holographic grating. RF power-1100 W, Plasma gas flow-12.1 L min^-1, auxiliary gas flow- 0.9 L min^-1 with a nebulizer gas flow- 0.4 L min^-1; wavelength of uranium: 409.014 nm.

Analyses of uranium by LF at BARC, Trombay: LF analyses were carried out with LF 003 uranium analyser. The instrument specification and other details have been previously reported (Kumar et al., 2008). ETHOS ONE high-performance microwave digestion system (Milestone, Italy) was used for sample dissolution. Milli-Q Plus water purification system (Millipore Corp., Milford, MA) was used for preparation of 18.2 MΩ water. Supra pure grade HNO₃ (65%), HF (40%) and H₃PO₄ (85%) from Merck were used for all the experiments. Samples were handled in a clean laboratory work area of class 100, designated for trace element analysis. All reusable labware were soaked in 10% v/v HNO₃ overnight, followed by rinsing with Milli-Q water and stored in 0.1% HNO₃ till further use. Each of 0.2 g samples were weighed and transferred to a heavy-duty PTFE vessel of microwave digester. 2.5 mL of HNO₃, 0.25 mL of HF and 2.5 mL of water were added into it. Vessels were packed as per protocol and subjected to microwave for dissolution as per the programme in Table S1. The samples were made up to 10 mL in perfluoroalkoxy volumetric flask with ultrapure water. Each of 1 mL samples from the volumetric flask was taken in PTFE vessel and evaporated to dryness on a hot plate to remove HNO₃ and HF completely and made up to 10 mL with 5% H₃PO₄. Samples were diluted within the calibration range of the instrument and analyzed. Three blanks were also run along with the samples. Final results were blank corrected.

Analyses of uranium and phosphorus by ICP-MS at CSMCRI, Bhavnagar: ICP-MS analyses were carried out on Thermofisher, iCAP RQ single quadrupole ICP-MS analyser. The entire workflow was managed by Thermo Scientific™ Qtegra™ Intelligent Scientific Data Solution™ (ISDS) software. Details of the instrument and operation can be found in the Thermofisher website (Thermo Scientific iCAP RQ ICP- MS). The calibration plot for uranium is shown in Figure S1. The quantification limit for uranium in solution was 3 μg.L^-1. 50 mg samples of phosphate fertilizers were weighed. 250 μL HNO₃ and ultrapure water were added into it to a final volume of 50 mL. At this concentration (1 g.L^-1), the quantification limit of uranium in the fertilizers was 3 mg.kg^-1. Solutions were shaken well prior to analysis. The same instrument was used for phosphorus estimation after 10-fold dilution of the above solution. Calibration plot for phosphorus is given in Figure S2.

Analyses of uranium by ICP-MS at ICT, Mumbai: Some of the uranium analyses were also carried out at ICT employing the same instrument and sample preparation procedure described above for analyses by CSMCRI. The calibration plot is shown in Figure S3.

Analyses of uranium by XRF: XRF analyses of SSP 2 and DAP 1 were carried out through the assistance of ONGC Energy Centre. Analysis was conducted on the SPECTRO XEPOS energy dispersive X-ray (Pd/Co alloy anode X-ray tube) fluorescence (ED-XRF) spectrometer, SPECTRO Analytical Instruments GmbH. SSP 2 was analysed as powder and DAP 1 as pressed tablet. Measurements were made in air atmosphere, with voltage and current settings of 45.1 kV and 0.9 mA, respectively. The spectrum resolution was 132.3 eV and overall measurement time was 150 s. Absolute error was in the range of 0.4-0.5 μg.g^-1.

Analyses of fluoride: Fluoride analyses by IC (CSMCRI, ICT) and ISE (ICT) were carried out by the same procedures reported previously (Ramteke et al., 2018). Calibration plots for IC analyses at CSMCRI and ICT are presented in Figures S4 and S5, respectively.

Analyses of ammonium: NH₄⁺ analyses were carried out at ICT on Dionex ICS 5000+ IC instrument employing Dionex IonPac™ CS16 Analytical (5 × 250 mm) column with IonPac CG16 Guard (5 × 50 mm) column; Eluent: 30 mM methanesulf onic acid; Flow Rate: 1 mL.min^-1; Column Temp: 40ºC; Injection Volume: 25 μL; Suppressed Conductivity Detection (Cation Self Regenerative Suppressor (CSRS 300, 4 mm); Auto suppression Recycle mode, 89 mA). The calibration plot is given in Figure S6.

Results

Analyses of uranium in commercial phosphate samples

Preliminary analyses of fertilizer samples for uranium were carried out by ICP-OES at NCCCM, Hyderabad. Only four samples (SSP 3, SSP 4, ANP 1 and DAP 2) were subjected to analyses. Measurements were made at 1g.L^-1 sample concentrations. The selected spectrum peak of 409.014 nm is reported to give the lowest interference (Li et al., 2021). Two samples gave uranium values <50 mg [U].kg^-1 of fertilizer indicating a quantification limit of 50 μg.L^-1 (Table 3). In view of the need for detection with greater sensitivity, experiments were conducted subsequently at BARC employing the technique of LF. Uranium absorbs strongly in the region of 260–350 nm and fluoresces in the region of 450–600 nm with quantum yield of almost unity. Uranium detection sensitivity below 3 μg.L^-1 can be achieved through this technique (Kumar et al., 2008). Interferences due to other species were taken care of by the standard addition method and checked for recoveries in the samples. The recoveries were within 96% to 102%. The detailed procedure is provided under Materials and Methods. Data for SSP 1-4, ANP 1 and DAP 1-2. are presented in Table 3. SSP and ANP samples, which contained 16% to 20% w/w P₂O₅ as per manufacturer specifications, gave absolute uranium concentrations in the range of 35-55 mg [U].kg^-1 whereas the concentrations for the two DAP samples with higher (46%) P₂O₅ content were in the range of 80 - 110 mg [U].kg^-1. The samples were analysed next by ICP-MS (Vogel et al., 2020) at CSMCRI, Bhavnagar and the data are compiled in Table 3. ICP-MS reportedly has low tolerance towards TDS (0.2% w/v max.) (Thermo Fisher Scientific, ICP-OES/ICP-MS) and hence sample concentration was initially restricted to 0.01% (0.1 g.L^-1). Another set of results was generated for SSP 2 and DAP 1 in a private laboratory that used the technique of XRF to analyse the solid fertilizer samples directly (Table 3). This technique has been investigated previously for uranium analysis in phosphate rocks (Al-Eshaikh, et al., 2016) and online uranium analysis in mining and processing industries, typically with a detection limit of 30-50 mg.kg^-1 (Hasikova et al., 2015). Advanced units such as SPECTRO XEPOS energy dispersive XRF analyser reportedly have detection sensitivity of <5 mg.kg^-1 (Heckel et al., Spectro Analytical Instruments). As can be seen from Table 3, the LF, ICP-MS and XRF data were broadly in accord, the average of the three being 50.12 and 111.75 mg [U].kg^-1 for SSP 2 and DAP 1, with standard deviations (σ) of 7.8 and 5.6 mg [U].kg^-1, respectively. All further analyses were carried out by ICP-MS only. Four commercial samples of DAP (DAP 3-DAP 6) were collected during 2019 and analysed during September to October, 2019 at CSMCRI. Analyses were carried out at 0.1 g.L^-1 sample concentration as before. DAP 3, DAP 4 and DAP 6 showed uranium concentrations higher than those of DAP 1 and DAP 2, whereas it was below the quantification limit (30 mg [U].kg^-1) in DAP 5. DAP 4 and DAP 6 were analysed repeatedly (3-4 times), preparing fresh solutions each time, and the values ranged from 256.2-279.0 and 126.3-149.4 mg [U].kg^-1 of fertilizer, respectively. The margin of error was within ± 10%. The samples were analysed once again (at ICT) after two years of storage in plastic bags. Analyses were carried out at 0.1 g.L^-1 and 1.0 g.L^-1 fertilizer concentrations. (Even at this higher concentration, the TDS of the solution was within the recommended TDS limit of 0.2% for ICP-MS analysis (Thermo Fisher Scientific, ICP-OES/ICP-MS)). The quantification limit was 3 mg [U].kg^-1 at 1.0 g.L^-1 concentration. The results are presented in Table 4. It can be seen that the trend of uranium concentrations obtained at ICT was the same as the earlier analyses at CSMCRI in 2019. Absolute values differed considerably, however, especially in the cases of DAP 3 and DAP 4. The differences were likely on account of changes in the sample during the long period of storage. Furthermore, it can be seen from the ICT data that the results were comparable at both concentration levels. The average over the four samples (DAP 3-DAP 6) was 128.3 mg [U].kg^-1 (σ =69.8 mg [U].kg^-1) for the set of results obtained at 1 g.L^-1 concentration. A spiking experiment was also conducted with DAP 4 at 1 g.L^-1 concentration. Deviation in the reading was +3.32% (Table S2).

Table 5 presents uranium concentrations in 15 DAP samples collected during 2021-2022. All the analyses were conducted with 1 g.L^-1 DAP solutions. Uranium concentration was lowest (4.1 mg [U].kg^-1 DAP) in DAP 18. Uranium concentrations were also comparatively low for DAP 7, DAP 13 and DAP 21. The majority of the samples from 2021-2022 collection exhibited uranium concentrations in the range of 100-200 mg [U].kg^-1 DAP, the maximum value of 189.3 mg [U].kg^-1 DAP being recorded for DAP 19. The average for all 15 samples was 110.9 mg [U].kg^-1 DAP. The standard deviation was high (σ =57.4 mg [U].kg^-1 DAP) in view of the wide variation in concentrations – from 4.1 to 189.3 mg [U].kg^-1 DAP. Table 5 also includes uranium concentrations on P₂O₅ basis. The average uranium concentration was 204.2 mg [U].kg^-1 P₂O₅, with σ =105.7 mg. One sample of MAP (MAP 1), which was procured along with the 2021-2022 DAP samples, was also analysed. The uranium concentration was 3.1 mg [U].kg^-1 MAP (4.0 mg [U].kg^-1 P₂O₅).

Analyses of fluoride in 2021-2022 phosphate fertilizer samples

The fluoride content in the DAP fertilizers (DAP 7-21) collected during 2021-2022 were also analysed. Analyses were conducted employing IC and ISE techniques. The application of these methods has been reported previously (Ramteke et al., 2018, 2019) for analyses of SSP 1-4, ANP 1 and DAP 1-6. In the present study, six of the fifteen samples (DAP 7-10, DAP 18-19) were analysed by both the methods. The results were in reasonable agreement, and exhibited similar trends (Table 6). Rest of the samples were analysed by either of the two methods: DAP 11-17 by ISE and DAP 20 and 21 by IC. The data are summarized in Table 6. Concentrations are expressed both with respect to fertilizer weight and also per kg P₂O₅. Only three of the fifteen DAP samples had fluoride concentration ≤ 0.5 % [F] in DAP (≤ 8 g [F].kg^-1 P₂O₅). Of these, the concentration was lowest for DAP 17 (0.18%). Ten out of fifteen samples had more than 1% fluoride w/w of DAP, while DAP 16 and DAP 20 contained fluoride in excess of 2% w/w of DAP (52.9 g and 42.8 g [F].kg^-1 P₂O₅ in DAP 16 and DAP 20, respectively). MAP 1 was also analysed for fluoride. The concentration was below the detection limit in ISE analysis and 0.08% w/w by IC.

Discussion

Considering that DAP fertilizers were sourced from various parts of the Country during 2021-2022 (Table 2), Table 6 and 7 provide realistic assessments of the extent of uranium and fluoride contamination in DAP fertilisers applied to Indian agricultural crops. Both sets of data have been compiled in graphical form in Figure 1. The average uranium concentration of 204.2 mg.kg^-1 P₂O₅ was comparable to the average of 243 mg.kg^-1 P₂O₅ reported previously for 303 phosphate fertilizer samples (Schnug and Haneklaus, 2015). Data pertaining to older samples going back to 2016 (Tables 3 and 4) revealed that DAP and other phosphate fertilizers with high uranium contamination have been applied in farmlands for at least seven years now. There is no mention of uranium in Indian phosphate fertilizer specifications (Bureau of Indian Standards, 2018). This is probably the case in many other countries also. Some attempts have been made, however, to specify limits: The German Commission for Soil Protection is reported to have suggested the mandatory labelling of phosphate fertilizers exceeding 20 mg [U].kg^-1 P₂O₅ (Schnug and Haneklaus, 2015). A non-negotiable limit of 50 mg [U].kg^-1 P₂O₅ was additionally proposed. It can be seen that the average uranium concentration in the DAP fertilizers from 2020-2021 was four times the proposed ceiling of 50 mg [U].kg-1 P₂O₅. Only two (DAP 13, DAP 18) out of the fifteen samples were within this limit, of which the uranium content in one (DAP 18) was low enough that it would not have required any labelling either. DAP 5 from the 2019 collection would have also fallen within this range. DAP 5 and DAP 18 were procured in different years but both were from the same company (C8). Whether these fertilizers were manufactured from low- uranium RP (Table 1) or some kind of treatment was provided is not known. In some instances there were inconsistenies in the quality of DAP sold by the same company. DAP 13 and DAP 17 were both from company C5, but whereas the uranium level in the former sample was 32.3 mg [U].kg^-1 P₂O₅, the value was 187.0 mg [U]. kg^-1 P₂O₅ in the latter, i.e., a factor of 7 difference in the relative concentrations. Similarly, whereas all the DAP samples from company C1 showed high uranium concentrations, MAP from the same company had negligible uranium (3.1 mg [U].kg^-1 MAP). Considering that Indian DAP consumption was 11.911 MMT during 2020-2021, nonpoint source uranium pollution from DAP application alone was 1,320 tonnes.annum^-1. Bioaccumulation of uranium in crops such as sunflower and Indian mustard would be of concern (Meng et al., 2018). Uranium presence in groundwater has also increased in many parts of the Country (Babu et al., 2008; Coyte et al., 2018; Uranium occurrence in shallow aquifers in India, 2021), but this has not been ascribed specifically to phosphate fertilizer use. If uranium is separated out from phosphoric acid prior to DAP manufacture, it can serve as feedstock for generation of 8.8 GW of nuclear power (@150 MT [U].annum^-1.GW^-1 (Strategy for growth of electricity in India, 2021) [present generation of nuclear power in India is ca. 6.9 GW (World-Nuclear Association, 2021)], besides the environmental gain that would accrue. Mature technologies exist for this purpose and 20% of the uranium demand in the US in the 1980s was met through its recovery from wet process H₃PO₄ (Steiner et al., 2020). Turning to fluoride, compared to the earlier reported data of fluoride content in DAP 1-6 (Ramteke et al., 2018, 2019) – wherein the average concentration was 17.7 g [F].kg^-1 P₂O₅ – a worsening trend was observed in the 2021-2022 DAP samples, with the average concentration rising to 24.3 g [F].kg^-1 P₂O₅ (Table 6, Figure 1). Phosphate fertilizer specifications brought out by the Bureau of Indian Standards provide a test method for fluoride but do not specify a limit as such (Bureau of Indian Standards, 2018). The incremental fluoride load in farmlands from DAP use alone was estimated to be 1,54,830 MT.annum^-1. It is reported that consumption of fluorite rock (CaF₂) in India was 3,24,704 MT during 2019-2020 (Indian Minerals Yearbook, 2020), amounting to fluoride equivalent of 1,58,189 MT. Thus, fluoride pollution in farmlands from DAP application is equivalent to India’s entire annual requirement of fluoride for manufacture of fluorochemicals. Fluoride level in groundwater has been steadily on the rise (Ali et al., 2019). Seepage of fluoride from farmlands may be a contributing factor. Bioaccumulation of fluoride in certain crops such as tea would also be of concern (Ing et al., 2021). Commercial processes exist for treatment of phosphoric acid with reactive silica to liberate residual fluoride as H₂SiF₆ (Samrane, 2011) and also for the further conversion of H₂SiF₆ into anhydrous HF required in bulk by the fluorochemicals industry (Dahlke et al., 2016). Given that the presence of fluoride in RP is ubiquitous, some of the low-fluoride DAP fertilizers (DAP 14, DAP 17) were presumably obtained by adopting such defluoridation measures. As evident from Figure 1, the majority of the DAP fertilizers in the current study were heavily contaminated in both constituents. Thus implementation of treatment processes to affect their removal does not appear to be commonplace. The first step towards effecting a transformation would be to set standards for fluoride and uranium in Indian DAP specifications. Considering the average fluoride and uranium concentrations in DAP reported above, and assuming limits of 5 g [F].kg^-1 P₂O₅ and 50 mg [U].kg^-1 P₂O₅ are enforced, fluoride and uranium pollution can be reduced by 75%. None of the DAP fertilizers procured in 2021-2022 complied with both limits. DAP 17 was the only sample that showed a fluoride value (3.2 g [F].kg^-1 P₂O₅) within the above limit but its uranium concentration was high (187.0 mg [U].kg^-1 P₂O₅). Conversely, DAP 18 and DAP 13 analyzed for uranium within the above limit (8.2 and 32.3 mg [U].kg^-1 P₂O₅, respectively) but their fluoride concentrations were high (16.4 and 25.9 g [F].kg^-1 P₂O₅, respectively). The published literature on fluoride and uranium separation from phosphoric acid generally focuses either on fluoride removal or uranium removal.

There is a need to devise processes that would enable fluoride and uranium to be separated out simultaneously or sequentially from intermediate phosphoric acid when both these constituents are present in high concentrations. The separated contaminants would have potential value as feedstock after due purification and transformation to desired forms. Indeed, this must be the priority of all countries, before tapping into other resources such as fluorite for fluorochemicals and uranium ore for power generation. There is also a need to explore processes for separation of uranium and fluoride from the fertilizers given that all phosphate fertilizers are not produced through the phosphoric acid route. Development of such a technology would also be useful to deal with ammonium phosphate fertilizers such as DAP and MAP that do not conform to stipulated standards. With the above in view, laboratory experiments were performed recently to eliminate fluoride from phosphate fertilizers through a thermal treatment protocol (Ramteke et al., 2021). This was followed by removal of uranium (Ghosh et al., 2021). It remains to be seen whether the processes are amenable to scale-up.

Conclusion

DAP was selected for the present study in view of its commanding presence among phosphate fertilizers. Uranium and fluoride in 2021-2022 DAP samples were analysed by ICP-MS and IC/ISE, respectively, after validation studies. The two pollutants were detected in concentrations of 8–360 mg[U].kg^-1 P₂O₅ and 3.2–52.9 g [F]. kg^-1 P₂O₅, the average values being 204.2 mg and 24.3 g, respectively. Uranium and fluoride discharge in Indian farmlands from DAP application were computed to be 1,320 MT and 1,54,830 MT per year, respectively. A margin of error of ± 10% is assumed conservatively. Actual uranium load in farmlands would be higher considering that DAP is not the only phosphate fertilizer in use. There is sizable offtake of SSP and NPK fertilizers also, and some of these fertilizers were found to be even more polluting than DAP. The most effective mitigation measure for ammonium phosphate fertilisers such as DAP is the separation of uranium and fluoride from phosphoric acid intermediate, for which well-established technologies exist already. That many of the DAP fertilizers were heavily contaminanted in fluoride and uranium indicates that these processes are not implemented by the majority of producers. Formulation of specifications for fluoride and uranium would ensure better compliance and serve as an important step towards environmental sustainability. At the same time, the treatment costs must be affordable so that fertilizer prices do not spiral out of control. Furthermore, the waste must be converted into value for a holistic solution. This would provide the necessary incentive to producers to undertake mitigation measures. Processes that remove uranium and fluoride from finished fertilizers should be developed in parallel to deal with ‘off-spec’ products. Studies also need to be conducted on the fate of the uranium and fluoride in farmlands. Although the current study focused on phosphate fertilisers sold in India, other nations too may be confronted with similar problems. In addition to fluoride and uranium, rock phosphate is known to contain several other harmful ingredients such as arsenic, cadmium, chromium and rare earth elements. Their presence was not investigated in the present study.

Acknowledgements

The following persons are acknowledged for assisting with uranium analyses: Dr. S. Sounderajan, Dr. G. Kiran Kumar and Dr. Sangita Kumar for Laser Fluorimetry data; Dr. A. C. Sahayam and Dr. U. Rambabu for ICP-OES analyses; Dr. R. Patidar for the first set of ICP-MS data; Dr. D. Parvatalu for arranging the XRF analyses; and Dr. K. V. Marathe and Mr. Keyur Moradiya for fluoride analyses by ISE method. All those who assisted in collecting phosphate fertilizer samples from various parts of the Country are gratefully acknowledged. DST, New Delhi is acknowledged for grant of Research Associateship to LPR under WTI project; Rubamin India Private Ltd. is acknowledged for supporting a PhD fellowship to R. S. Soni; and PKG acknowledges the support of the K. V. Mariwala-J. B. Joshi Distinguished Professor endowment.

References

1. Al-Eshaikh MA, Kadachi AN, Sarfraz MM (2016). Determination of uranium content in phosphate ores using different measurement techniques. J King Saud Univ Eng Sci. 28(1):41-6.

   [Crossref] [Google Scholar]

2. Ali S, Fakhri Y, Golbini M, Thakur SK, Alinejad A, Parseh I, Shekhar S, et al (2019). Concentration of fluoride in groundwater of India: A systematic review, meta-analysis and risk assessment. Groundwater for Sustainable Development. 9:100224.

   [Crossref] [Google Scholar]

3. Al-Shawi AW, Dahl R (1995). Determination of thorium and uranium in nitrophosphate fertilizer solution by ion chromatography. J Chromatogr A. 706(1-2):175-81.

   [Crossref] [Google Scholar]

4. Babu MS, Somashekar RK, Kumar SA, Shivanna K, Krishnamurthy V, Eappen KP (2008). Concentration of uranium levels in groundwater. Int J Environ Sci Technol. 5(2):263-6.

   [Crossref] [Google Scholar]

5. Blackwell MS, Darch T, Haslam RP (2019). Phosphorus use efficiency and fertilizers: future opportunities for improvements. Front Agr Sci Eng. 6(4):332-40.

   [Crossref] [Google Scholar]

6. Bureau of Indian Standards, Diammonium Phosphate Specification. IS6448: 2018.
7. Cioroianu TM, Filip D, Filip Gh (2001). Environmental considerations on uranium and radium from phosphate fertilizers. Proceedings of a technical committee meeting at International Atomic Energy Agency (IAEA): 215–24.

   [Google Scholar]

8. Coyte RM, Jain RC, Srivastava SK, Sharma KC, Khalil A, Ma L, Vengosh A, et al (2018). Large-scale uranium contamination of groundwater resources in India. Environ Sci Technol Lett. 5(6):341-7.

   [Crossref] [Google Scholar]

9. Dahlke T, Ruffiner O, Cant R (2016). Production of HF from H2SiF6. Procedia Engineering. 138:231-9.

   [Crossref] [Google Scholar]

10. Dartan GU, Taspinar FA, Toroz I (2017). Analysis of fluoride pollution from fertilizer industry and phosphogypsum piles in agricultural area. J Ind Pollut Control. 33(1):662-9.

    [Google Scholar]

11. Dolar D, Košutic K, Vucic B (2011). RO/NF treatment of wastewater from fertilizer factory—removal of fluoride and phosphate. Desalination. 265(1-3):237-41.

    [Crossref] [Google Scholar]

12. EFMA, 2000. Best available techniques for pollution prevention and control in the European fertilizer industry, booklet No. 4 of 8: Production of phosphoric acid, European Fertilizer Manufacturers’ Association.
13. EPA (2015). TENORM: Fertilizer and Fertilizer Production Wastes.
14. Fertilizer Association of India, 2021. All-India Consumption of Fertiliser Nutrients - 1950-51 to 2020-21.
15. Geretharan T, Jeyakumar P, Bretherton M, Anderson CW (2020). Fluorine and white clover: Assessing fluorine's impact on Rhizobium leguminosarum. J Environ Qual. 49(4): 987-9.

    [Crossref] [Google Scholar]

16. Ghosh PK, Patwardhan AW, Ramteke LP, Hinge S, Soni RS, Sarode DD, Process of uranium removal from contaminated ammonium phosphate fertilizers. Indian Patent Application No: TEMP/E-1/36365/2021-MUM.
17. Gilbert N (2009). Environment: The disappearing nutrient. Nature. 461: 716–8.

    [Crossref] [Google Scholar]

18. Hasikova E, Sokolov A, Titov V (2015). On-line X-Ray fluorescence analysis of uranium and thorium materials in mining and processing industry. In: Uranium-Past and Future Challenges Springer, Cham.793-804.

    [Crossref] [Google Scholar]

19. Heckel J, Seyfarth A, van Driessche M. Surprising new capabilities of ED-XRF technology. Spectro Analytical Instruments.
20. Hoyer M (2015). Uranium contamination of soil and groundwater by phosphate fertilizer application. In Uranium-Past and Future Challenges Springer, Cham. 707-16.

    [Crossref] [Google Scholar]

21. Damodaran, H (2021). Amid DAP shortages, sales of complex fertilisers soar, Indian Express.
22. Determination of uranium content in phosphate ores using different measurement techniques
23. Jiao W, Chen W, Chang AC, Page AL (2012). Environmental risks of trace elements associated with long-term phosphate fertilizers applications: A review. Environmental pollution. 168:44-53.

    [Crossref] [Google Scholar]

24. Khater AE (2008). Uranium and heavy metals in phosphate fertilizers. In: Uranium, mining and hydrogeology. Springer, Berlin, Heidelberg. 193-8.

    [Crossref] [Google Scholar]

25. Kratz S, Schnug E (2006). Rock phosphates and P fertilizers as sources of U contamination in agricultural soils. In: Uranium in the environment. Springer, Berlin, Heidelberg. 57-67.

    [Crossref] [Google Scholar]

26. Kumar SA, Shenoy NS, Pandey S, Sounderajan S, Venkateswaran G (2008). Direct determination of uranium in seawater by laser fluorimetry. Talanta. 77(1):422-6.

    [Crossref] [Google Scholar]

27. Lal N, Sharma PK, Sharma YP, Nagpaul KK (1985). Determination of uranium in fertilisers using fission track method. Fertilizer research. 6(1):85-9.

    [Crossref] [Google Scholar]

28. Li X, Xiong C, Sun K, Fang F, Zhang Q (2021). Optimization of ICP-OES’S parameters for uranium analysis of rock samples. J Korean Phys Soc. 78(9):737-42.

    [Crossref] [Google Scholar]

29. Ing ME, Magnuson BE, Frantz DL (2021). Fluoride Content in Asian Produced Green Teas. J Can Dent Assoc. 87(l1):1488-2159.

    [Google Scholar]

30. Meng F, Jin D, Guo K, Larson SL, Ballard JH, Chen L, Arslan Z, et al (2018). Influences of U sources and forms on its bioaccumulation in Indian Mustard and Sunflower. Water, Air, & Soil Pollution. 229(11):1-1.

    [Crossref] [Google Scholar]

31. Menzel RG (1968). Uranium, radium, and thorium content in phosphate rocks and their possible radiation hazard. J Agric Food Chem. 16(2):231-4.

    [Crossref] [Google Scholar]

32. Ramteke LP, Sahayam AC, Ghosh A, Rambabu U, Reddy MR, Popat KM, Rebary B, et al (2018). Study of fluoride content in some commercial phosphate fertilizers. J Fluorine Chem. 210:149-55.

    [Crossref] [Google Scholar]

33. Ramteke LP, Shah AA, Ghosh PK (2019). Fluoride contaminant in phosphate fertilizer is an unchecked problem in India. Chemical Weekly. 182-4.

    [Google Scholar]

34. Ramteke LP, Sarode DD, Marathe YS, Ghosh PK (2021). Removal of fluoride contaminant in phosphate fertilizers through solid State thermal treatment. J Fluorine Chem. 241:109693.

    [Crossref] [Google Scholar]

35. Roessler CE, Smith ZA, Bolch WE, Prince RJ (1979). Uranium and radium-226 in Florida phosphate materials. Health physics. 37(3):269-77.

    [Crossref] [Google Scholar]

36. Samrane K, Al-hjouj M, Doudin F, Kossir A (2011). Fluosilisic acid: Recovery system and aluminum fluoride production. Arab Fertil Assoc Publ. 11:1-20.

    [Google Scholar]

37. Souza SO, Ávila DV, Monteiro AS, Garcia CA, Alves JP, Maranhão TA, Araujo RG et al (2016). Simultaneous determination of thorium and uranium in mineral fertilizers by inductively coupled plasma optical emission spectrometry. J Braz Chem Soc. 27: 799-806.

    [Crossref] [Google Scholar]

38. Schnug E, Haneklaus N (2015). Uranium in phosphate fertilizers – review and outlook. In: Merkel B., Arab A. (eds) Uranium - Past and Future Challenges. Springer, Cham.
39. Spalding RF, Sackett WM (1972). Uranium in runoff from the Gulf of Mexico distributive province: anomalous concentrations. Science. 175(4022):629-31.

    [Crossref] [Google Scholar]

40. Steiner G, Geissler B, Haneklaus N, (2020). Making uranium recovery from phosphates great again? Environ Sci Technol. 54: 1287–9.

    [Crossref] [Google Scholar]

41. Strategy for growth of electricity in India.
42. Thermo Scientific iCAP RQ ICP-MS.
43. Comparison of ICP-OES and ICP-MS for Trace Element Analysis. Thermo Fisher Scientific.
44. Uranium Occurrence in Shallow Aquifers in India (2020). Central Ground Water Board, Department of Water Resources, Government of India.
45. van Kauwenbergh S (2010). World phosphate rock reserves and resources, IFDC Fertilizer Outlook and Technology Conference.
46. Veselsky JC, Kwiecinska B, Wehrstein E, Suschny O (1988). Determination of uranium in minerals by laser fluorimetry. Analyst. 113(3):451-5.

    [Crossref] [Google Scholar]

47. Vogel C, Hoffmann MC, Taube MC, Krüger O, Baran R, Adam C (2020). Uranium and thorium species in phosphate rock and sewage sludge ash based phosphorus fertilizers. J Hazard Mater. 382:121100.

    [Crossref] [Google Scholar]

48. World Nuclear Association (2021). Nuclear power in India – Indian Nuclear Energy.
49. Yamazaki IM, Geraldo LP (2003). Uranium content in phosphate fertilizers commercially produced in Brazil. Appli Radiat Isot. 59(2-3):133-6.

    [Crossref] [Google Scholar]

50. Yao H (2020). India’s DAP import  to rise in 2020 ease in 2021. Argusmedia.
51. Zohuri B (2020). Nuclear fuel cycle and decommissioning. In: Khan SU, Nakhabov A (eds), nuclear reactor technology development and utilization. Wood head Publishing Series in Energy, 61-120.