Without phosphorus we cannot produce food – but at current consumption rates, reserves could be depleted in the next 50 to 100 years. Projections suggest that global phosphorus demand could grow at 2.3% annually just to feed the growing world population, an estimate that was made before the growth of biofuels. The very concept of biofuels as a viable “renewable” source of energy might not hold if one of the fundamental elements is growing more scarce.

Within a few decades, a “peak phosphorus” crunch could seriously threaten agriculture as global reserves of high-quality phosphate rock decline. At present extraction rates (around 40 million t/a P2O5), reserves are unlikely to last much more than 100 years.

In large parts of the developing world, phosphate supply is insufficient, both in terms of total application and imbalance in the N to P to K ratio. If many of the areas being farmed today were to receive sufficient phosphate to prevent mining of soil reserves, this in itself would substantially increase world demand. On the other hand, in some regions of Europe and the USA, there is an oversupply of phosphate to agriculture due to the large combined input of phosphate in the form of fertilisers and organic manures.

Increased imbalances
Part of the answer lies in better husbandry of phosphorus reserves: an effort that may require the creation of an international body to monitor the use and recycling of phosphorus. As animal wastes offer a potentially large source of phosphates for recovery, stricter controls on intensive livestock farming may become the norm, to avoid localised manure surpluses and their disposal.

An assessment for Western Europe shows that the phosphate excreted by livestock in this region could be some 50% more than the amount currently applied as mineral phosphate fertiliser. If a limited quantity of up to 40% of the phosphate contained in manure, mainly from animals in stables, could be collected and spread onto agricultural land, 4-5M tonnes of mineral fertiliser phosphate could be replaced by a more efficient use of manure, mainly in Europe and the USA. Phosphorus capture could be required at all significant sewage treatment works, and its recovery from wastewater will become

economically attractive. The technology of phosphate recovery is straightforward, yielding both the value of the recovered phosphorus, and significant savings in both treatment costs and disposal of the residual sludge.

Realistically, logistical factors, such as the cost of transport and the scale of installations, would make recovery an economic option only in the case of large, geographically concentrated waste streams. In rural areas, agricultural sludge or manure spreading will probably always remain the best option for recycling nutrients. In the UK even a conservative estimate of the potential for phosphorus recovery and recycling (50% recovery applicable to 25% of sewage and to 15% of animal wastes) represents half of industrial phosphate demand. Currently, the excess of nutrients from sewage treatment works, increasingly also from animal wastes, was prevented from reaching surface water through a strategy of phosphorus removal, not recovery. This is where phosphorus is transferred to sludge, either in an organic form as in biological phosphorus removal, or as a chemical precipitate, in iron or aluminium salts.

The majority of phosphorus removal in Europe uses chemical precipitation, often simultaneous with secondary biological treatment, yielding a mix of precipitate with organic sludge which is of limited agronomic value. Because higher concentrations of precipitation chemicals are required than actually combine with the available phosphorus, these methods lead to an increase of sludge production around 40%.

Recovery of phosphorus for recycling, rather than its transfer into sewage sludges, may offer economic and environmental rewards for the water industry. For the phosphate industry it holds out the promise of a significant source of sustainable raw material. These benefits must be compared with the investment and running costs of phosphorus recovery installations.

Routes to recovery
Phosphorus removal by traditional precipitation will generally preclude recovery for recycling, as the resulting iron or aluminium compounds are incompatible with technologies used in the phosphate industry. They either require excessive energy input, to separate the phosphates from the added precipitation chemicals, or interfere with the industrial process.

For either the solvent extraction or electrothermal reduction to yield high purity phosphate compounds, the upper limit is considered 1-2% of iron and aluminium. Incinerated sewage sludge, generated through precipitation, easily contains ten times as much with up to 20% iron (as Fe2O3).

But several technologies are now implemented at both demonstration and full scale to show possible routes for phosphorus recovery. These processes isolate the recovered phosphorus in the form either of a calcium phosphate or magnesium ammonium phosphate (struvite). As a preliminary step to generate a concentrated liquid phosphate stream for recovery, biological phosphate removal looks very promising. In a side steam, the nutrient rich sludge can yield a liquor containing phosphorus in excess of 100mg/l, which would be particularly appropriate for phosphorus recovery.

Full-scale struvite recovery processes were first introduced more than ten years ago in Japan and the Netherlands:

  • The DHV Crystalactor fluid bed process, a the full scale installation at the AVEBE potato processing plant in the Netherlands (150m3/h)
  • The Unitika (Osaka) process was applied at the Ube Industries Sakai plant for industrial wastewaters and commissioned in 1998 at the Shimane Prefecture sewage works, Japan (45,000m3/d)
  • The Geochem Research/Delft University Earth Sciences stirred precipitation process extracts struvite from 700,000t/a of calf manure at Putten, Netherlands in 1998
  • The PHOSPAQ process has been applied since 2006 in Olburgen, Netherlands, treating a combination of potato processing wastewater and centrate from a municipal sludge digester, for 1.2t/d of struvite
  • A second full-scale plant, in Lomm, Netherlands, treats potato processing effluent to produce 800kg/d struvite

Recovered struvite, in the form of easily handled granules or crystals with a high phosphorus and a low heavy metal content, are easily filtered and require no further drying. Therefore, struvite remains an intriguing opportunity, since it forms itself spontaneously in sewage treatment works and therefore is in principle very harvestable.

Whilst struvite undoubtedly has some value as a fertiliser (including its ammonia and magnesium content), it is difficult to transform into other phosphate derivatives using any existing technology at the disposal of the industry. While not usable in a traditional wet acid route, struvite could be processed by electrothermal reduction, but the presence of ammonia requires drastic changes in furnace feed preparation or a radically different process flowsheet.

The same cannot be said of the other option: calcium phosphate is ideal for onward industrial processing, as the recovered material is indistinguishable, in most respects, from mineral phosphate rock. Several full scale processes are already recovering phosphates from wastewater streams through calcium phosphate formation:

  • DHV Crystalactors at the Dutch wastewater plants of Westerbork (12,000 PE), Heemstede (35,000 PE) and (Geesmerambacht 230,000 PE) since 1994
  • Pilot plant at Essex & Suffolk Water in Chelmsford from 1997-1998
  • Reactor tested by Karlsruhe University at Darmstadt Süd, Germany, since 1997
  • Demonstration plant at Warriewood, Australia, by Sydney Water (50,000 PE) since 1995
  • Three plants constructed by Kurita, Japan
  • Fixed bed precipitation at Mercedes factory at Gagenau, Germany, (160m3/h) since 1998

The recovered pellets drain readily to below 5-10% water and contain 5-15% phosphorus. The calcium phosphates deposit by amorphous precipitation around the seed material, rather than true crystallisation, forming compounds with different hydration complexes (calcium hydroxyapatite, dicalcium phosphate dihydrate, octacalcium phosphate, tricalcium phosphate).

The solubility and crystallisation properties of these different molecules vary and the balance between them will modify the overall behaviour of a recovery reactor. Some recovered calcium phosphate has comparatively high residual organics, which would pose no problems for electro-thermal reduction, but for the wet acid recovery would need a simple calcining step, as employed on many natural rocks.

Slick solution in Slough
Thames Water is incorporating a new technology at Slough Sewage Treatment Works, one of the UK’s first biological phosphorus removal plants, commissioned in 1993 with a capacity of 257,000 PE and flow (FFT) of 118.4Ml/d. As a consequence of the rich phosphorus streams, struvite can form scale on the inside of pipes and valves, increasing maintenance.

To reduce the costly maintenance resulting from the damaging build-up of struvite, and simultaneously extract and recover phosphorus and ammonia from its wastewater stream and transform them into a commercial fertiliser, a pilot scale nutrient recovery facility began operating in Slough in March 2010 (see WWT October). After demonstrating the technology’s potential to support efficient operation of the plant’s biological phosphorus removal process, and confirming its full scale viability, this first installation in Europe also marks a unique partnership:

  • The process supplier, Ostara will build and finance the nutrient recovery facility, expected to yield 150t/year of Crystal Green fertiliser
  • Thames Water has agreed to pay a monthly fee for the treatment capacity provided, expected to be completed in mid-2011

Without having to make any capital investment, estimated at £2M, this project will help the operator to efficiently meet nutrient limits, optimise the plant’s efficiency and reduce operational and maintenance cost between £128,260 and £202,000 annually, while recovering a valuable resource. A second facility is being considered in the Netherlands, while others already operate in North America:

  • A reactor in Edmonton, Canada, produces 500kg/d of fertilizer, open since May 2007
  • Tigard, Oregon, is the first US plant, with a capacity of 1t/d, open since May 2009
  • In Suffolk, Virginia, the Nutrient Recovery Facility was launched in May 2010
  • The York, Pennsylvania WWTP implemented a unit in summer 2010

These and other related topics will be presented and discussed at the upcoming WEF IWA Conference on Nutrient Recovery in the US in January: www.nutrient2011.org.

Action inspires action. Stay ahead of the curve with sustainability and energy newsletters from edie