Phosphate Plant Yield Comparisons Presented at the 33 rd Meeting of the Central Florida Chapter of the American Institute of Chemical Engineers By: Richard D. Harrison Senior Engineer PegasusTSI Tampa, Florida Copyright 2009 Richard D. Harrison Page 1
Outline Introduction - PegasusTSI Core Business Activities Estimated Range of P 2 O 5 Losses to Gypsum Stacks Process Changes to Reduce Phosphate Losses: Eliminate off-site discharges Reduce phosphate flow to gypsum stack Identify opportunities to recycle process water Redirect fresh water make-up to gypsum stack Process Alternatives 1. Site selection and climate conditions - typical plants, Florida and N. Africa 2. Open Circuit Processes no recovery of pore moisture P 2 O 5 3. Closed Circuit Processes some recovery of pore moisture P 2 O 5 4. Process Water Recycling 5. High Strength Fluoride Recycle 6. Improved Washing and/or Double Filtration 7. Non-Contact Evaporator Condenser Cooling 8. Calcination References Introduction - Pegasus TSI Core Business Activities Process Studies Identify & Compare Alternatives Capital Cost Estimates (+/-) 50%, 30% & 10% Project Management Front End Engineering Detailed Engineering Procurement Construction Management Copyright 2009 Richard D. Harrison Page 2
Estimated Range of P 2 O 5 Loss to Gypsum Stacks Manufacture of phosphoric acid produces from 5 to 6 tons of phosphogypsum for each ton of P 2 O 5 produced. Additionally, each ton of gypsum that is sent to the impoundment is only 67 weight % solids initially, with the remaining 33% made up of process water that occupies the capillary space between the gypsum crystals (pore moisture). The older, lower strata dewater as the pore moisture is expressed by the weight of the newer deposits above. The following graph shows the weight percent solids for a typical gypsum stack for different strata. 85% 80% Gypsum Solids (wt %) 75% 70% 65% y = 0.0014x + 0.6625 60% 0 20 40 60 80 100 Depth (feet) A note of caution regarding the above chart is that the average gypsum solids for a 100 foot tall gypsum stack is not the simple average of the initial 67% and the value at 100 foot depth of 80%, or 73.5% - because the water is disproportionately present in the upper strata. An average gypsum stack gypsum solids value of 70% was used for the calculation on the Schematic Diagram found on page 5. (See also Reference 2) Estimation of water soluble phosphate losses to the gypsum stack depend on three variables: 1. Water soluble P 2 O 5 concentration in the pore moisture solution 2. Quantity of water acting as pore moisture 3. Quantity of phosphogypsum and other solids produced per ton of P 2 O 5 produced Copyright 2009 Richard D. Harrison Page 3
The typical range of these three variables is: 1. P 2 O 5 concentration in gypsum slurry water is between 1% and 2.5% 2. Water acting as pore moisture is initially 0.5 times gypsum weight (33%/67%) 3. Gypsum weight per ton P 2 O 5 produced is between 5 to 6 tons - depending on ore Water soluble P 2 O 5 lost to the gypsum stack initial pore moisture is therefore: Low Range = 5 t gyp./t P 2 O 5 x 0.5 pore moisture/gyp. solids x 1% P 2 O 5 = 2.5% Mid Range = 5.5 t gyp./t P 2 O 5 x 0.5 pore moisture/gyp. solids x 1.75% P 2 O 5 = 4.8% High Range = 6 t gyp./t P 2 O 5 x 0.5 pore moisture/gyp. solids x 2.5% P 2 O 5 = 7.5% The water soluble P 2 O 5 in the gypsum stack can be accounted for in two different ways. If the facility is near the end of its life, then water discharges from the gypsum stack may have to be neutralized at significant expense for surface discharge. Alternatively, if the facility a going concern for the foreseeable future, any water discharges from expressed pore moisture can be recycled and the P 2 O 5 value recovered. The limiting ideal case is where all future P 2 O 5 flow to the gypsum stack is eliminated, and all pore moisture from past gypsum production is recovered creating the possibility for P 2 O 5 production to exceed P 2 O 5 input for the facility while pore moisture from the prior gypsum stack is being reclaimed. For the purposes of this paper Phosphate Facility Yield will be defined as tons P 2 O 5 sold as products / tons ore P 2 O 5 brought on site. Using this definition it will be possible to achieve yields above 100% if future P 2 O 5 losses to the gypsum stack are significantly reduced and P 2 O 5 from the existing gypsum stack is reclaimed. Gypsum must be washed and transported to the gypsum stack with fresh water (0% P 2 O 5 ) in order to minimize water soluble P 2 O 5 losses to the gypsum stack. Due to water balance constraints, this ideal is not approachable without the installation of a second stage gypsum filtration system and significant process water reclaim. The PegasusTSI team is fortunate to have been able to design and build such a system for an application for a confidential client in the Middle East. Please refer to the following process schematic diagram on page 5 for a representation of a typical 1 MM ton P 2 O 5 / year facility (See also References 3 and 4). Copyright 2009 Richard D. Harrison Page 4
Process Alternatives: 1. Site Selection and Water Balance Site selection has an important bearing on phosphogypsum disposal techniques. Two general classes of plant sites are dry locations receiving less than 15 inches of precipitation per year, and wet locations that receive 40 to 60 inches of precipitation per year. Typical dry locations include plants in Morocco, Tunisia, Jordan, Saudi Arabia, and Australia. These plants do not usually employ gypsum slurry transport for gypsum disposal because evaporation is much greater than precipitation. The above graphic from NOAA uses purple to show areas with less than <15 precipitation per year, and blue-green to display areas with between 43 to 57 per year. Please consult the online version of this paper at www.aiche-cf.org for a color version of the graphics in this paper. Copyright 2009 Richard D. Harrison Page 6
Most wet locations that receive over 40 inches precipitation per year practice slurry transport for gypsum stacking. Despite the water from rainfall, these wet locations also require significant input of either well water or river water to maintain the water balance because the evaporative process ponds are sized to require make-up water. 2. Open Gypsum Circuit (no recovery of gypsum effluent P 2 O 5 ) Open Gypsum Circuit discharge practices include: slurry with sea water for discharge to the ocean, conveyor transport to rail cars or barges for transport to uses such as agricultural application, mine backfill, or ocean discharge. Additionally, some sites use belt conveyors for construction of gypsum fields. Please refer to the attached satellite photos of Plant A (below) and Plant B (next page) for examples of open circuit gypsum disposal. The above photo shows Plant A that historically used seawater for process cooling and for gypsum slurry transport to the ocean. The approximate scale is 5 miles (8 kilometers) across the bottom of the photo. Copyright 2009 Richard D. Harrison Page 7
The above photo shows Plant B that uses belt conveyors to transport gypsum to railcars, or to a gypsum field adjacent to the Mediterranean. A close-up of the same plant s gypsum conveyor and gypsum field is attached below. Copyright 2009 Richard D. Harrison Page 8
3. Closed Gypsum Circuits (lined stack systems most wet locations) Plants that operate a closed gypsum circuit usually use process water to slurry gypsum discharge from the filter for transport to the phosphogypsum landfill. Every ton of product P 2 O 5 is accompanied with the generation of approximately 5 to 6 tons of byproduct gypsum solids. Accompanying these gypsum solids are 2 to 3 tons of process water occupying the space between the gypsum crystals after filtration and stacking. Most plants operate with process water between 1% P 2 O 5 and 2% P 2 O 5, resulting in a 2% to 6% loss of water soluble P 2 O 5 with the pore water accumulating in the gypsum stack. The next picture shows Plant C as an example of a plant operating with gypsum stacks (scale is 2 miles across bottom). Copyright 2009 Richard D. Harrison Page 9
The next photo shows Plant D of approximately 1MM tons P 2 O 5 /year capacity also with a closed circuit gypsum stack (scale 2 miles across bottom). 4. Process Water Recovery (neutralized or as-is) The concentration of P 2 O 5 in the process water can be reduced in part by maximizing the quantity of process water consumed in the process. The two largest consumptive uses include filter wash water and ball mill make-up water for wet rock grinding. These two applications account for approximately 2,100 gpm and 600 gpm respectively for a plant with 1 MM tons P 2 O 5 / year capacity operating on 68% solids rock slurry. Please refer to Reference 9 for additional review of potential improvements. Two alternatives exist for using process water in the ore grinding area. The preferred solution is to specify materials of construction compatible with low ph (2 ph to 3 ph) operation. This allows water soluble P 2 O 5 and accompanying fluosilicic acid to react with carbonates present in the phosphate ore to liberate CO 2 gas. Typical phosphate ore contains 3% to 5% CO 2. Copyright 2009 Richard D. Harrison Page 10
The least preferred method of enabling process water use for mill water addition is to add a neutralizing agent to the process water. The cost for ongoing purchase and transport of a neutralizing agent, along with the reduction in P 2 O 5 capacity for the facility make the economics for this process inferior. 5. High Strength Fluoride Recycle Processes Fluosilicic acid can be recovered for introduction into the ground ore storage area to further dissolve carbonates and liberate additional CO 2 gas. Implementation of both process water addition to mill feed and FSA addition to rock slurry storage allows increased P 2 O 5 production for a facility with a limited sulfuric acid production capacity by neutralizing some of the carbonates in the incoming ore (Reference 5). PegasusTSI is pleased to be able to offer the most commercially successful fluoride recovery technology currently available, with over a dozen successful installations in the past 3 years. Fluosilicic acid (FSA) recovery of 0.07 ton F per ton P 2 O 5 is typical (Reference 1). A 1 MM P 2 O 5 ton per year facility can produce 192 tons F per day. PegasusTSI FSA Recovery Process Evaporator with Cyclonic Entrainment Separator FSA Recovery Vessel Barometric Condenser Cooling Water To Vacuum System Liquid FSA Spray Nozzles To Hotwell FSA Product Make-up Water FSA Recirculation Tank Copyright 2009 Richard D. Harrison Page 11
An additional benefit of recovering FSA is that the barometric condenser water can then be isolated for final filter wash and gypsum slurry make-up water. 6. Improved Washing and/or Double Filtration One opportunity to reduce P 2 O 5 losses is to replace the process water that proceeds with the gypsum to it s destination with fresh water. A 1 MM ton P 2 O 5 / year plant will require about 1,100 gpm of water to report as gypsum stack pore moisture. Reducing the P 2 O 5 content of the gypsum slurry water by 1% will reduce P 2 O 5 losses by 65.7 tons/day or 24,000 tons/year. The value of recovering this P 2 O 5 is on the order of $700/t * 24,000 t/year = $16.8 MM/year. Additional to the immediate savings from reducing P 2 O 5 flow to pore moisture, some clients also may benefit by reduced escrow funds required to cover future closure costs. The technology to permit fresh water to displace process water as pore moisture make-up is site specific. PegasusTSI will be pleased to facilitate the implementation of appropriate water processing technologies for your facility to enable the plant water balance to be controlled while allowing process water to be replaced with fresh water for pore moisture service. Our team has installed double filtration and other technologies for clients that are still in successful operation after more than 10 years of service in the phosphate industry. Another opportunity to improve gypsum cake washing is to feed unground wet rock to the digester instead of 68% solids rock slurry. This will allow either an additional 600 gpm of filter wash for a 1 MM t/y P 2 O 5 capacity plant, or the installation of sulfuric acid dilution cooling system. The viability of this alternative is ore and digester specific, and is not practical for all facilities. One additional technique to improve gypsum crystal size, and consequently filterability is to operate with higher sulfate concentrations in the digester. The following graph from Reference 6 illustrates the significantly adverse impact that low sulfate digester conditions can have on gypsum crystal size and resulting wash rates. Copyright 2009 Richard D. Harrison Page 12
7. Evaporator Condenser Secondary Heat Exchanger Evaporator condenser secondary heat exchangers are another technology that our team has successfully implemented in world-class phosphate fertilizer facilities. This technology can be used to separate fresh evaporative cooling water from process condensate formed in a contact barometric condenser, and has been demonstrated in ongoing phosphate plant use for more than 10 years. Condenser Secondary Heat Exchangers offer an alternative to allow sea water use for cooling while eliminating the transfer of phosphate and fluoride into the ocean in dry climates where fresh water availability is limited. 8. Calcination Some ores contain high concentrations of organic matter that dilutes chemical analysis. Calcination of these ores assists improving P 2 O 5 yield in the following four ways: 1. Removes combustible organic matter and some carbonate CO 2 2. Reduces the tons of gypsum solids (and pore moisture) per ton of P 2 O 5 3. Produces a dry rock that allows more filter wash water to be introduced 4. Larger gypsum crystals can be formed if sulfuric dilution coolers are used Copyright 2009 Richard D. Harrison Page 13
References 1. Doug Belle - Innovations in fluosilicic acid recovery technology - F.I. 224-2008 2. John Cameron - Pollution Control In Fertilizer Production - 1994 3. Pierre Becker - Phosphates and Phosphoric Acid 2 nd Edition - 1989 4. Hossein Sepehri-Nik - Fertilizer Technical Data Book 4 th Edition - 1997 5. Lloyd Banning - Fluosilicic Acid Acidulation of Phosphate Rock - 1975 6. B. Moudgil/Univ. of Florida/FIPR - Enhanced Filtration of Phosphogypsum - 1995 7. John Van Wazer - Phosphorus and its Compounds - 1958 8. Charles Felice - Raytheon - 21 st Century Phosphoric Acid Plant Designs - 1998 9. May/Charlot/Edwards - Process Water Management in Today s Phosphate Industry - 1996 Copyright 2009 Richard D. Harrison Page 14