Beneficiation of apatite rock phosphates by calcination: effects on chemical properties and fertiliser effectiveness.

Introduction

Most rock phosphates (RPs) are apatite-bearing rocks that contain enough phosphorus (P) to be utilised for the manufacture of fertilisers, elemental P, and/or phosphoric acid (Gary et al. 1974). Some RPs are calcined during beneficiation to remove impurities, to increase P concentration, or to make Fe- and Al-phosphate impurities more available to plants (Milnes 1976). These RPs may also be used as direct application fertilisers but the effect of calcination on the agronomic effectiveness is likely to be adverse (Khasawneh et al. 1980). However, this effect is not well defined for the various RPs.

The effect of calcination on the properties and agronomic effectiveness of RPs was investigated for 6 apatitic RPs. RPs from Egypt, Morocco, North Carolina, and Sechura contain francolite (carbonate apatite) as the major apatite mineral, while Queensland and Christmas Island A-ore contain fluorapatite. Christmas Island A-ore also contains some Aland Fe-phosphate. All these RPs have previously been evaluated as direct application fertilisers (Bolland and Gilkes 1990).

Methods

The RPs (particle size 46-53 [micro]m) were calcined for 1 h in a muffle furnace. Samples calcined at high temperatures were crushed, as they had sintered. Uncalcined RP and those calcined at 500, 900, and 1100 [degrees] C were investigated further as they represented the various materials produced by calcination. The calcined RPs were subjected to physical, chemical, and biological procedures to evaluate their effectiveness as fertilisers.

Physical analysis

X-ray diffraction (XRD) on a Philips PW 1830 diffractometer with Cu K[Alpha] radiation and a diffracted beam monochromator was carried out to determine mineralogical attributes. Semi-quantitative measurements of crystal size and abundance were determined using XPAS software (Singh and Gilkes 1992). Calculations of apatite unit-cell a and c dimensions were derived from XRD patterns using an iterative least-squares computer software (Novak and Colville 1989). BET- [N.sub.2] (Brunauer et al. 1938) surface area was measured using a Gemini 2360 V4.01 instrument (Micrometrics).

Chemical analysis

Total P was determined using HCl/HCl[O.sub.4]/[HNO.sub.3] acid digestion (AOAC 1991). The extent of dissolution was measured by shaking the RPs in 2% citric acid solution (1:100, solid: solution ratio) for a total of 128 h. Samples of the suspensions were taken from each bottle after 0.5, 1, 2, 4, 8, 16, 32, 64, and 128 h. The amount of P dissolved in the extract was measured using the vanadate-molybdate (yellow) UV spectrophotometric method (Rayment and Higginson 1992).

Biological evaluation

Wheat (Triticum aestivum) was grown in non-drained pots with 1 kg soil treated with basal fertiliser (Palmer 1980) and 0.125, 0.250, or 0.500 g P as the various calcined RPs and the reference fertiliser compound monocalcium phosphate (MCP). MCP was used because it is the major P constituent of superphosphate. The soil type used was the Balkruling lateritic soil, Dy5.81 (Northcote 1971), from Bakers Hill, 73 km east of Perth, Western Australia (Palmer 1980). The soil was first sieved to [is less than]2.00 mm particle size, and has the following properties: clay 8.31%; silt 5.76%; sand 85.93%; pH (1: 5, soil:water w/v) 5.10; and EC (1:5, soil:water w/v) 67 [micro]S/cm

The plants were grown for 30 days before they were harvested, dried, weighed, digested in HCl/HCl[0.sub.4]/ [HNO.sub.3], and analysed for P by the vanadate-molybdate (yellow) UV spectrophotometric method.

Results and discussion

XRD results showed that apatite and quartz persisted to the highest calcination temperature of 1100 [degrees] C. However, dolomite and calcite, which were present in Egypt and North Carolina RPs, were lost between 500 and 900 [degrees] C, presumably altering to oxides or reacting to form complex phosphate or silicate phases. Organic matter present in North Carolina RP was oxidised during calcination. The minor amounts of crandallite and millisite in Christmas Island A-ore became amorphous. XRD patterns have not been included in this paper because the shifts in apatite spacings are so small that they would not be evident.

Unit-cell c dimension (UCD c) of apatite did not change significantly upon calcination for any sample (Fig. 1a), within the error of measurement of about 0.002 [Angstrom]. On the other hand, unit-cell a dimension (UCD a) for francolites changed to the value for fluorapatite (9.37-9.38 [Angstrom] (JCPDS 1983) with increasing calcination temperature due to loss of [CO.sub.2] (Fig. 1b) (Khasawneh et al. 1980). Queensland and Christmas Island A-ore RP are dominantly fluorapatite, so no appreciable change in UCD a occurred. Calcination at 1100 [degrees] C resulted in UCD a for all RPs becoming close to the value for ideal fluorapatite.

[Figure 1 ILLUSTRATION OMITTED]

The average size of crystals or, more accurately, the average size of the coherently diffracting zone estimated from the width of 6 different XRD reflections for each apatite is a measure of crystal perfection, and increased for calcination temperatures above 500 [degrees] C (Fig. 2). This trend results from crystals becoming larger and structure becoming more regular, due to crystal growth, and structural ordering developed by volume diffusion, associated with the expulsion of carbonate, water, and excess fluoride from the apatite structure.

[Figure 2 ILLUSTRATION OMITTED]

Generally, BET-[N.sub.2] specific surface area decreased with increasing calcination temperature for all the RPs (Fig. 3). The reduction in surface area for the francolite RPs was greatest between 500 and 900 [degrees] C, whereas it occurred between 900 and 1100 [degrees] C for Queensland RP. The reduction in specific surface area resulting from calcination is due to both growth of apatite crystals and sintering of crystals, the latter being a consequence of adhesion of particles by expansion of interfacial contact points to form bonded aggregates (Kuczynski 1965).

[Figure 3 ILLUSTRATION OMITTED]

Due to the very different surface areas and internal porosity of samples, equilibrium solubility of RP in 2% citric acid was reached at different times for different samples, but was generally attained within 24 h. There was a large decrease in the P dissolved with increasing calcination temperature, after 128 h of extraction (Fig. 4). Apart from Egyptian RP, the solubility of the apatite RPs became similar for higher calcination temperatures presumably due to fluorapatite replacing the more soluble francolite (Khasawneh and Doll 1978). Egyptian RP became much more soluble after calcination at 1l00 [degrees] C. This presumably reflects chemical reactions with the free carbonate minerals, to form more soluble phosphate(s). The same increase in effectiveness (i.e. solubility) of Egypt RP calcined at 1100 [degrees] C was not reflected in the plant growth experiment because the surface area had been substantially lowered ([is less than] 0.1 [m.sup.2]/g) as a result of sintering.

[Figure 4 ILLUSTRATION OMITTED]

In the plant growth experiment, all applications of phosphate fertilisers increased plant yield, P concentration, and P content of plant shoots. The P content of plant shoots is the clearest indicator of fertiliser effectiveness as it combines responses exhibited by both yield and P concentration (Mackay et al. 1984). Fertiliser relative effectiveness (FRE) (Fig. 5) was measured by comparing the slopes of the linear region of the P content response curves (well before the plateau) for each of the RPs, relative to the slope for monocalcium phosphate (MCP) (Bolland and Gilkes 1990). Uncalcined Queensland and Egypt RPs had the lowest FRE of 0.16 (i.e. only 16% as effective as MCP), whereas, uncalcined Morocco and North Carolina RPs had the highest FRE value of 0.26. Bolland and Gilkes (1990) observed similar results, with Queensland RP being 10% and North Carolina RP 20% as effective as MCP. Thus, FRE increases with amount of carbonate substitution in apatite, with the most highly carbonate substituted apatites (i.e. North Carolina and Morocco RP) being the most effective (White 1971; Bolland and Bowden 1984).

[Figure 5 ILLUSTRATION OMITTED]

Calcination at 1100 [degrees] C reduced the FRE of all RPs to [is less than] 0.02 (Fig. 5). This trend in FRE due to calcination reflects the several changes in mineralogical and physical properties. Conversion of microcrystalline francolite towards larger crystals of fluorapatite in a sintered mass, as indicated by the increase in UCD a, increased crystal size, and reduced surface area, are responsible for the decrease in relative effectiveness. As indicated by the solubility data in Fig. 4, fluorapatite is much less soluble in citric acid than francolite, so that citric acid extraction provides a convenient procedure for evaluating fertiliser effectiveness.

Conclusions

This research has shown that calcination of apatitic RPs may greatly reduce their solubility and fertiliser effectiveness, and should, therefore, be avoided for RPs intended for direct application. The effect is particularly severe for francolite RPs that are quite effective fertilisers without calcination, but are converted to poorly soluble fluorapatite by calcination. Unlike the requirement for the manufacture of chemical fertilisers there is no need to use calcination to remove or deactivate sesquioxide, clay, and organic impurities in RPs to be used for direct application, as these impurities are unlikely to affect the agronomic effectiveness of the RPs.

References

AOAC (1991) `Official methods of analysis.' Vol. 1, 15th edn. (Association of Official Agricultural Chemists: Washington DC)

Bolland MDA, Bowden JW (1984) The initial and residual value for subterranean clover of phosphorus from crandallite rock phosphate, apatite rock phosphate and superphosphate. Fertilizer Research 5, 295-307.

Bolland MDA, Gilkes RJ (1990) The poor performance of rock phosphate fertilisers in Western Australia: Part 2. The soil chemical explanations. Agricultural Science 3, 44-47.

Brunauer S, Emmett PH, Teller E (1938) Adsorption of gasses in multimolecular layers. Journal of the American Chemical Society 60, 309.

Gary M, McAfee R Jr, Wolf CL (Eds) (1974) `Glossary of geology.' p. 535. (American Geological Institute: Washington DC)

JCPDS (1983) `Mineral powder diffraction file: group data book.' (International Centre for Diffraction Data: PA)

Khasawneh FE, Doll EC (1978) The use of phosphate rock for direct application to soils. Advances in Agronomy 30, 159-206.

Khasawneh FE, Sample EC, Kamprath EJ (Eds) (1980) `The Role of phosphorus in agriculture.' (American Society of Agronomy: Madison, WI)

Kuczynski GC (1965) Sintering and related phenomena. In `Proceedings of the Notre Dame Conference'. (Gordon and Breach: NY)

Mackay AD, Syers JK, Gregg PEH (1984) A glasshouse comparison of six phosphate fertilisers. New Zealand Journal of Experimental Agriculture 12, 131-140.

Milnes AR (1976) Beneficiation of Christmas Island plumbogummite ores. In `Proceedings of the International Geological Congress'. Sydney 1976.

Northcote KH (1971) `A factual key for the recognition of Australian soils.' 3rd edn. (Rellim Technical Publications: Glenside, S. Aust.)

Novak GA, Colville AA (1989) A practical interactive least-squares cell-parameter program using an electronic spreadsheet and a personal computer. American Mineralogist 74, 488-490.

Palmer B (1980) A chemical, mineralogical and biological evaluation of calcined Christmas Island c-grade rock phosphate. PhD thesis, The University of Western Australia, Nedlands.

Rayment GE, Higginson FR (1992) `Australian laboratory handbook of soil and water chemical methods.' (Inkata Press: Melbourne)

Singh B, Gilkes RJ (1992) XPAS: An interactive program to analyse X-ray powder diffraction patterns. Powder Diffraction 7, 6-10.

White MS (1971) Calcination of Christmas Island phosphate. New Zealand Journal of Experimental Agriculture 14, 971-992.

Manuscript received 28 January 2000, accepted 9 August 2000

H. H. Lim and R. J. Gilkes

Department of Soil Science and Plant Nutrition, University of Western Australia, Nedlands, WA 6097, Australia.