Environ. Sci. Technol. 2010, 44, 1806–1812
Conversion of Spent Solid Phosphoric Acid Catalyst to Environmentally Friendly Fertilizer WERNER VAN DER MERWE* Sasol Technology Research and Development
Received September 16, 2009. Revised manuscript received December 7, 2009. Accepted January 26, 2010.
Solid phosphoric acid (SPA) catalysts are widely used in the petroleum industry. Despite a high phosphorus content the spent catalyst is generally not reused. Moreover, due to the limited life spans that are achieved industrially, large quantities of spent catalyst requires disposal, often by landfill. SPA can be readily converted to fertilizer, but the presence of carbonaceous deposits on the catalyst presents a potential environmental hazard. This work demonstrates that these deposits are mostly polyaromatic (amorphous carbon) with smaller amounts of oxygenates and aliphatics. Neither the chemical makeup nor the physical structure of the catalyst or the presence of coke precludes it from use as fertilizer. Subsequently, the spent catalystwasmilled,neutralizedwithlimeandammoniumhydroxide, and then calcined to yield a phosphate-rich fertilizer. Toxicity characteristic leaching tests of the spent catalyst fertilizer showed low levels of metals and organics, establishing that no harmful compounds are likely to be absorbed into plant life or groundwater. A plant growth study of the spent catalyst fertilizer indicated that it is approximately as effective as superphosphate fertilizer when used in alkaline soil. The spent catalyst fertilizer is environmentally benign and economically efficient.
Introduction Solid phosphoric acid (SPA) is a versatile and robust solid acid catalyst widely used in petroleum refineries (1, 2). Applications include oligomerization of light olefins, especially in synthetic fuels refining (2-5), as well as aromatics alkylation (3) and ethylene and propylene hydration (6). SPA is manufactured by adding concentrated phosphoric acid to diatomaceous earth (kieselguhr), followed by extrusion, calcination at high temperature and screening (7). Spent SPA is not regenerated (8), and catalyst lifetime is most often determined by pressure drop constraints (7). A refinery such as that operated by Sasol in Secunda produces on average 1500 tons of spent SPA catalyst annually. Other reports suggest spent catalyst production rates of hundreds of tons per year per crude oil refinery (9, 10). The spent catalyst is considered a hazardous waste that is both corrosive and toxic (11) and is disposed of by landfill. However, spent catalyst is a potential source of phosphorus for other applications, such as conversion to phosphorus-rich fertilizer. Fraps (12) found the phosphorus in spent catalyst to be readily available by dissolution in ammonium citrate. A plant growth study using corn and barley plants treated directly with spent catalyst * Corresponding author phone: +27 16 960 4991; e-mail:
[email protected]. 1806
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 5, 2010
showed that the phosphorus uptake was slightly lower for the spent catalyst than for a superphosphate fertilizer. Later, the same conclusion was reached using spring wheat (13). Today, the sale of spent SPA for conversion to fertilizer is an established commercial practice (9) and information on the process of conversion is readily available (10). Reusing the spent catalyst as fertilizer supports green chemistry (14) and green engineering (15) principles. However, from environmental, health and safety perspectives, there is a concern that the black deposit on the catalyst may contain hazardous volatile or persistent organics and metals. Potentially harmful constituents of a spent catalyst may be classified into two groups (11), those that are present in the fresh catalyst, and those that are added during use. Whereas previous work has focused exclusively on the former, this study focuses mainly on the latter. Research into the nature of the catalyst coke and the physical character of the spent catalyst is therefore motivated here by the objective of establishing the environmental impact of converting the catalyst to fertilizer. The first part of this paper characterizes the spent catalyst in terms of the nature and structure of the phosphorusspecies, the silica support and the coke, using established analytical techniques (16). The second part of this paper evaluates the environmental impact and the efficacy of fertilizer manufactured from spent catalyst that was obtained from the Sasol refinery in Secunda.
Experimental Section Spent Catalyst Characterization. Spent SPA catalyst was obtained from Sasol’s Secunda Fischer-Tropsch-based refinery in South Africa, where C3-C5 olefins are oligomerized to fuel. Samples S1 and S2 correspond to long and short times-on-stream, respectively. Both were unloaded after steaming the reactor with 8 barg steam to remove volatiles. A fresh catalyst was obtained from the manufacturer (Su ¨ dChemie Sasol Catalysts) for comparative purposes. Catalyst samples were characterized according to a procedure that roughly corresponds to that recommended in a coke characterization patent (17) and covers many of the techniques described in a recent review (16), namely thermal (TGA-MS), spectroscopic (Infrared, XRD), microscopic (TEM, SEM, ICP, optical microscopy), and chromatographic (GCMS headspace, pyrolysis GC-MS, solvent extraction with GCMS) analyses, as well as elemental analysis and sorption capacity measurements. Experimental programs are described in Supporting Information (SI) Table S1. Liquid phase 1 H and 13C nuclear magnetic resonance imaging (NMR) on the solvent extracts and solid-state 13C NMR on the spent catalyst powder were also done. For these, the results agreed with those in literature (18, 19) and are not elaborated upon further. Fertilizer Performance. Three spent catalyst-derived fertilizer samples (Samples F1 to F3) were obtained from Solfos, South Africa (the company that processes Sasol Secunda spent catalyst). The three samples correspond to different operating conditions in the Solfos plant. Spent catalyst used to manufacture the fertilizer was used as received since decoking is expensive. TCLP. A 5% extract of Samples F1 to F3 together with S1 were subjected to TCLP analysis, including solution pH, anion concentrations, inorganics, volatile organic compounds (VOC) and semivolatile organic compounds (SVOC) according to the procedures and methods listed in SI Table S2, taken from a U.S. Environmental Protection Agency report (20). 10.1021/es902603d
2010 American Chemical Society
Published on Web 02/10/2010
FIGURE 1. (a) TGA profile with the derived heat flow profile (temperature program in SI Table S1). (b) Corresponding MS trace showing water, CO2 and hydrocarbon fragments. Plant Growth Study. Sample F1 was used in a plant growth study to test its efficacy in alkaline soil. The total breakdown (composition) of the fertilizer was determined, and the phosphorus content of the fertilizer was determined by dissolution in a strong base (NaOH). Subsequently, the fertilizer was dosed in concentrations of 15, 30, and 45 mg P/kg soil into pots containing 4 kg soil (procured from Pretoria, South Africa). Two control experiments were included, namely using a superphosphate dosage equivalent to that of the Solfos fertilizer and a blank (no additional P added). Two pH levels were adopted, namely a set of control experiments at the soil’s inherent pH (5.8) and a target pH of 8.0, for which CaCO3 was added to the soil. Typical dosages of N and K took place during planting and once before flowering. Three germinated corn plants were planted in each pot. After 6 weeks, the dried plant material was weighed and recorded as the yield. In each case, the soil was analyzed for pH, P, Ca, Mg, K, and Zn and the leaves were analyzed for N, P, K, Ca, Mg, Na, Chlorides, S, Cu, Fe, Mn, and B. All experiments were conducted in quadruplet to test reproducibility.
Results and Discussion Spent Catalyst Characterization. TGA-MS. Under inert gas, Sample S1 had a gradual mass loss of 32% (Figure 1). This occurred in two stages: an approximately constant-rate mass loss of 13% (up to 450 °C), followed by a slow first order decay. MS results showed that the first stage mass loss was mostly water (from desorption and catalyst dehydration). The constant rate of water loss implies that the silicon-phos-
phate species undergoing dehydration are distributed throughout the catalyst structure, that is, a three-dimensional water source (16). The second stage is probably due to adsorbed hydrocarbon volatization (hydrocarbon fractions appear in the MS profile). The first order decay indicates that the hydrocarbons have a two-dimensional (surface adsorption) structure (16). Upon switching to air, a rapid mass loss of another 9% occurred, corresponding to a peak in the CO2 (m/e ) 44) MS signal. Note that there is no corresponding peak in the H2O (m/e ) 18) signal, indicating that this is pure carbon burnoff (and not hydrocarbon burning). It is interesting to note that the carbon burnoff is linear at first and then adopts first order decay behavior. This is characteristic of a 2D surface coke (coke that is evenly deposited over the surface of the catalyst). These results are reproducible (not shown). Analogous profiles for the fresh catalyst and Sample S2 are shown in SI Figure S1, with the results summarized in Table 1. The fresh catalyst shows a 14% overall mass loss, all of which are under inert conditions. This is made up of 5-10% water, some hydrocarbon material (see elemental analysis), and P species vaporization. Since the fresh catalyst contains more P than the spent catalyst (demonstrated later), the difference between fresh and spent catalyst is a conservative estimate of coke species content. Profiles for the short timeon-stream spent catalyst (Sample S2) are qualitatively similar to that of Sample S1, although less pronounced. IR Spectroscopy. Infrared spectra of both spent and fresh catalysts (SI Figure S2) could not be used to identify any VOL. 44, NO. 5, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1807
TABLE 1. Comparison of Fresh and Spent Catalysts fresh SPA stage 1: inert (water/CO2) stage 2: inert (hydrocarbon + P) stage 3: air (carbon)
spent SPA (S1)
TGA-MS Results Summary (Mass Loss) 5-10% 13% 9% 18% (coke >9%) 0% 9% (9% carbon)
ortho:pyro ratio average crystallite size (Å)
Crystalline Properties by XRD 330:170 500:0 ortho: 935 pyro: 511 ortho: 640
P Si Al Fe Na K Mg
ICP Results (Mass Basis) 15.9% 19.5% 1.1% 0.5% 1121 ppm 1270 ppm 848 ppm
C H N S
spent SPA (S2) 12% 16% (coke >5%) 3% (3% carbon)
12.9% 9.0% 0.4% 0.2% 297 ppm 558 ppm 259 ppm
Elemental Analysis Results (Mass Basis) 1.06% 18.34% 1.67% 2.48%
