Optimizing and modelling of a partially acidulated phosphate rock (PAPR) fertilizer pilot-plant using Chilembwe phosphate rock
Date
2012-09-24
Authors
Chileshe, Fredrick
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Improving food security by increasing crop production is currently Zambia's primary objective. Poor soil fertility and inadequate utilization of chemical fertilizers have often resulted in poor crop yields and that has exacerbated the problems of poverty and hunger in the country. Phosphorus is the most limiting nutrient after nitrogen in most of the Zambian soils. Currently, phosphate fertilizers are imported and are expensive for most of the low-resource farmers in the country. However, local phosphate deposits exist in Zambia but they have remained un-exploited and these include Chilembwe, Mumbwa, Nkombwa Hill, and Kaluwe. If exploited, indigenous phosphate resources could provide less expensive phosphate fertilizers that could help farmers increase crop production. In addition, indigenous phosphate resources could reduce the amount of foreign exchange currently spent on the importation of phosphate fertilizers. Previous research has revealed that Partially Acidulated Phosphate Rock (PAPR) produced using Chilembwe phosphate rock is just as agronomically effective as the imported superphosphate fertilizers on food crops in Zambia.Partial acidulation of phosphate rock involves the reaction of finely ground phosphate rock with only a fraction (usually 50%) of the stoichiometric quantity of acid required to fully convert the apatite in the rock to a soluble form, mono-calcium phosphate. The process of producing PAPR is therefore less expensive ^in terms of acid consumption. The agronomic effectiveness of PAPR largely depends on the percentage of soluble phosphate (available P2O5) content in the PAPR. During the processing of phosphate rock into PAPR the percentage of available P2O5 in the product depends largely upon the various process variables. These process variables include reactiveness of phosphate rock, particle size of phosphate rock, degree of acidulation, acid concentration, process temperature profile, residence time of phosphate rock in the granuiator and drying temperature of PAPR. These variables differ from one phosphate rock to the other. Therefore, it was necessary to optimize the process variables in the pilot-plant during the production of PAPR using Chilembwe phosphate rock in order to improve the product quality and process efficiency.
The main units in the process (i.e. the ball mill in the comminution circuit and the granulator in the granulation circuit) were modelled in order to adopt the types of operation that gave optimum performance of the units. To model a system, a tracer was introduced in the feed stream as an instantaneous pulse while the product stream leaving the system was sampled and analyzed for the tracer concentration at discrete time intervals. The data for the tracer concentration with time was used to calculate the experimental residence time distribution curve which was simulated using the selective recycle model to obtain the number of cycles distribution curve. The number of cycles distribution curve showed the mass of tracer leaving the system after a given number of cycles and'this was used in assessing the performance of the unit. It was assumed that flow in the imain units (i.e. the ball mill and the granulator) was perfectly mixed whilst the classifying units (i.e. the hydrocyclone and the screens) were time delay units. In curve fitting, the generalized number of cycles distribution (NCD) and the inverse of the Laplace transform for a perfectly mixed reactor and a pure time delay unit were used.
Preliminary work involved characterization of a Chilembwe phosphate rock sample in order to determine the constituent mineral phases, the type of apatite, the grain size of apatite, and the chemical composition. Acidulation tests were carried out in the PAPR pilot-plant consisting of a rotary drum granulator, a rotary drum dryer, and a deck of screens. Prior to acidulation tests, phosphate rock sample containing an average of 26.4% CaO, 1.05% MgO, 0.75% K2O, 0.08% NaaC 1.19% AI2O3, 1.60% Fe203, 22.5% P2O5. 1.79% F and 1.00% S was crushed and ground to a desirable particle size distribution.Acidulation tests involved mixing the ground phosphate rock, with water and acid in the rotary drum granulator where the reactions occurred giving a granular PAPR product that was dried in the rotary drum drier. The product from the drier was sampled and analyzed for water-soluble P2O5 and neutral ammonium citrate (NAC)-soluble P2O5. The sum of water-soluble P2O5 and NAC-soluble P2O5 is the available P2O5 for uptake by plants when the fertilizer is applied in the field. Several process variables were studied during the acidulation tests aimed at optimizing the PAPR pilot-plant when using Chilembwe phosphate rock. These included particle size of phospfiate rock, degree of acidulation, acid concentration, initial temperature of phosphate rock, water and acid mixture.residence time of phosptiate rocl< in ttie granulator, and drying temperature of PAPR.
An economic evaluation for a 30 metric tons per hour PAPR project for an initial period of 15 years was finally carried out in order to assess the viability of commercial-scale production of PAPR. Methods employed in the analysis included the minimum rate of return, the Net Present Value, and Benefit-Cost ratio at various prices of the commodity.Results of the powder X-ray diffraction studies revealed that the apatite in Chilembwe phosphate rock exists in the form of calcium nrianganese fluoride phosphate, Ca9.3Mno.7F2 (PO4) suggesting the alteration of the ideal fluorapatite Caio(P04)6F2 by substitution of 0.7 Ca*^ with Mn*^ ions. The apatite grains are in the size range 400)im - SOO^m as revealed by thin sections observation under the microscope. Results from size-sorting assays have shown that the bulk of the apatite could be liberated from the rock matrix by grinding to minus 106^m.
Acidulation test results showed that in the particle size range 42% - 79% minus 75|.im the percentage of water-soluble P2O5 and Neutral Ammonium Citrate (NAC)-soluble P2O5 in PAPR remained constant. The percentage of available 1 P2O5 in PAPR increased with the increase in the degree of acidulation (i.e. 40%, 50%, 60%, 70%, 80%, and 100%). Low acid concentrations, 50% to 70% H2SO4 gave high levels of water-soluble P2O5 and NAC-soluble P2O5 in PAPR. Acid concentration below 40% and above 70% lowered the percentage of water-soluble P2O5 and NAC-soluble P2O5 in PAPR. The initial temperature of the rock, water and acid before the reaction did not have a remarkable effect on the percentage of water-soluble P2O5 and NAC-soluble P2O5 in PAPR in the range 25°C to 110°C. The effect of residence time of phosphate rock in the granulator was that the percentage of water-soluble P2O5 and NAC-soluble P2O5 increased with the increase in residence time up to 3 minutes. After 3 minutes there was no significant change in the level of available P2O5 in the product. The drying temperature in the range of 90°C - 150°C did not affect the percentage of available P2O5 in PAPR. However, water-soluble P2O5 decreased with the increase in drying temperature whilst NAC-soluble P2O5 increased with the increase in drying temperature keeping the available P2O5 constant. The selective recycle model for a generalized NCD was effectively used to simulate the RTD's in the comminution of phosphate rock and in the acidulation and granulation processes. This resulted in the optimization of the operations of the main processing units (i.e. the ball mill in the comminution of phosphate rock and the granulator in the processing of phosphate rock into PAPR). The PAPR produced at optimum parameters contained an average of 9% available P2O5 and 18% total P2O5 at 50% degree of acidulation with H2SO4 at 50% concentration, 3 minutes residence time of phosphate rock in the granulator, and 120°C drying temperature, using phosphate rock containing about 23% P2O5 and 1 particle size 42% minus 75^m.Field tests were conducted by the School of Agricultural Sciences on maize, soya beans, sunflower, and groundnuts and results showed that PAPR was just as agronomically effective as the imported Mono-Ammonium Phosphate (MAP) in providing phosphorus to the plants. For soil re-captalization with phosphorus, PAPR was also found to be more suitable than MAP as it did not depress maize yields at higher application rates.
Economic evaluation of a 30 metric tons per hour PAPR project showed that the project could be profitable and the price of PAPR could be as low as K40,000 per 50kg bag. This price is about half that of compound fertilizers that are currently on the market.This study has shown that Chilembwe phosphate deposit could indeed become an inexpensive source of phosphate that could partially substitute the imported phosphate and greatly benefit low-resource farmers in the enhancement of agricultural crop production in Zambia.
Description
Keywords
Phosphate rock -- Zambia , Phosphatic fertilizers -- Zambia