Processing of spent NiW/Al2O3 catalysts

Abstract

Spent oxidized (500 ºC, 5 h) commercial NiW/Al2O3 catalysts were processed using two different routes: a) fusion with NaOH (650 ºC, 1 h), the roasted mass was leached in water; b) leaching with HCl or H2SO4 (70 ºC, 1-3 h). HCl was the best leachant. In both routes, soluble tungsten was extracted at pH 1 with Alamine 336 (10 vol.% in kerosene) and stripped with 2 mol L-1 NH4OH (25 ºC, one stage, aqueous/organic ratio = 1 v/v). Tungsten was isolated as ammonium paratungstate at very high yield (> 97.5%). The elements were better separated using the acidic route.

spent catalysts; metal recovery; Ni-W catalyst

NOTA TÉCNICA

Processing of spent NiW/Al₂O₃ catalysts

Jéssica Frontino Paulino^I; Julio Carlos Afonso^{I, *}; José Waldemar Silva Dias da Cunha^II

^IDepartamento de Química Analítica, Instituto de Química, Universidade Federal do Rio de Janeiro, 21941-909 Rio de Janeiro - RJ, Brasil

^IIDepartamento de Química e Materiais Nucleares, Instituto de Engenharia Nuclear, 21941-906 Rio de Janeiro - RJ, Brasil

ABSTRACT

Spent oxidized (500 ^oC, 5 h) commercial NiW/Al₂O₃ catalysts were processed using two different routes: a) fusion with NaOH (650 ^oC, 1 h), the roasted mass was leached in water; b) leaching with HCl or H₂SO₄ (70 ^oC, 1-3 h). HCl was the best leachant. In both routes, soluble tungsten was extracted at pH 1 with Alamine 336 (10 vol.% in kerosene) and stripped with 2 mol L^-1 NH₄OH (25 ^oC, one stage, aqueous/organic ratio = 1 v/v). Tungsten was isolated as ammonium paratungstate at very high yield (> 97.5%). The elements were better separated using the acidic route.

Keywords: spent catalysts; metal recovery; Ni-W catalyst.

INTRODUCTION

At present, environmental directives impose severe restrictions concerning gas (NO_x, SO_x etc.) and particle emissions from refineries and fuels. More specifically, these directives require improvement of petroleum refining (higher impurities removal). Sulfur removal (hydrodesulfurization) is one of the reactions that take place during hydrotreatment (HDT). The reaction is essential to obtain more acceptable fuels for the environment.¹ The most familiar HDT catalysts are CoMo, NiMo and NiW supported on alumina (Al₂O₃), normally modified by additives that improve activity towards some HDT reactions (such as hydrogenation, cracking) or thermal resistance during its lifetime.^1-5

The increasing severity of environmental directives has prompted the development of a new generation of more efficient catalysts and the proposal of less drastic HDT conditions concerning energy and hydrogen consumption.^2-4 The amount of spent non-reusable HDT catalysts has increased markedly in recent years¹ due to a steady increase in the processing of heavier feedstocks in the petroleum refining industries.

Spent refinery catalyst recycling is rapidly evolving due to the rapid evolution of environmental regulations in many countries and also the constant fluctuation of metal prices.¹ In view of the environmental and economic benefits, increasing attention has been paid to the development of technologies to recover valuable metals from these secondary sources.

There are significantly fewer studies on processing spent NiW catalysts than spent CoMo and NiMo catalysts. It appears that the methods used for the recovery of molybdenum can also be applied to tungsten recovery.¹ Both elements belong to group 6 of the Periodic Table, and their chemistry is very similar. Tungsten can be recovered from NiW catalysts by chlorination (typically at 450 °C for 30 min) as volatile WO₂Cl₂, which is hydrolyzed to WO₃.⁶ Bioleaching of NiW catalysts has previously been studied in recent literature.^7,8

Hydrometallurgical processing of concentrates from tungsten ores with strong acids (HCl, H₂SO₄) tends to keep it in solid residue.⁹ Hydrometallurgical routes using aqueous NaOH or Na₂CO₃ and fusion with NaOH or Na₂CO₃ are also of practical interest.^10,11 The composition of spent NiW catalysts is very different and much simpler than tungsten concentrates. Except for phosphorus, which is normally found as a support additive, the other usual interfering elements in tungsten recovery (Si, Mo, As) are absent in HDT catalysts.¹ However, the amount of aluminum is much higher (as aluminum oxide is the catalyst support), thus posing a challenge for tungsten recovery from these spent catalysts.

Recovery of tungsten via liquid-liquid extraction (LLE) in acidic medium has been gaining importance in recent literature. Aliphatic amines have been widely employed as extractants.^12,13 In general, the efficiency of amine decreases in the order quaternary > tertiary > secondary > primary under the same experimental conditions. Tertiary amines are the most cited in the literature.^14,15 Stripping is normally performed with NH₄OH. Akin to molybdenum,¹⁶ tungsten recovery via LLE presents a critical challenge - its separation from phosphorus - because phosphotungstate ions (PW₁₂O₄₀^3-) are highly stable in acidic medium.¹⁷ They are extracted by amines and also stripped by NH₄OH.^14,18-20 Tungsten can be separated from phosphorus in alkaline medium^9,13 via precipitation of magnesium phosphates in the presence of NH₄OH.

Taking into account that many commercial processes for spent NiMo/Al₂O₃ samples employ alkaline fusion and leaching with strong acids,¹ this work presents a route for processing spent NiW catalysts using these methods. Recovery of components of the active phase (nickel, tungsten) and of the support (aluminum) was performed by a combination of several separation techniques.

EXPERIMENTAL

Catalysts

Spent commercial NiW/Al₂O₃ catalysts were employed (5 mm extruded cylinder, 1.2 mm diameter). Chemical analyses are presented in Table 1. The catalysts were employed for at least 4 years in diesel hydrotreaters in a Brazilian refinery. Samples were milled to particle diameters < 0.106 mm and dried at 110 ^oC for 2 h.

Oxidation of samples

Samples were placed in ceramic crucibles. The height of the catalyst layer was set at 5 mm. The catalyst was placed in a furnace at room temperature and heated at 500 ºC (10 ºC min^-1) for 5 h, according to previous studies in our laboratory.^21,22 The roasted mass was cooled down in the furnace and transferred to a dessicator before running the experiments.

Fusion process

Fusion was carried out using sodium hydroxide. The stoichiometric amount of the reactant was calculated according to the following reactions:

XO₃ + 2 NaOH → Na₂XO₄ + H₂O (X = W, S) (1)

Al₂O₃ + 2 NaOH → 2 NaAlO₂ + H₂O (2)

P₂O₅ + 6 NaOH → 2 Na₃PO₄ + 3H₂O (3)

Based on data from Table 1, the theoretical NaOH/catalyst mass ratio is approximately 0.75:1. A 10 wt.% excess of NaOH was employed. Experiments were carried out using 10 g of sample in nickel crucibles. Fusion was performed at 650 ºC in a furnace for 60 min.²³ The roasted mass was cooled down in the furnace and leached with distilled water (10 mL g^-1 solid, 60 ^oC, 200 rpm, 20 min). The insoluble matter was separated by filtration, washed with water (4 mL g^-1), dried at 150 ºC for 2 h and weighed. The matter was then dissolved in HF + H₂SO₄ (1:2 v/v) at 60 ºC for chemical analysis. The experiments were run in triplicate, and errors for each experiment were consistently below 4%.

Acidic leaching

6 mol L^-1 HCl and 6 mol L^-1 H₂SO₄ were used as leachants since they are the most cited in the literature.¹ The stoichiometric amount of these acids was calculated according to the following reactions:

X₂O₃ + 6 HCl → 2 XCl₃ + 3 H₂O (X = Al, Fe) (4)

X₂O₃ + 3 H₂SO₄→ X₂(SO₄)₃ + 3 H₂O (5)

NiO + 2 HCl → NiCl₂ + H₂O (6)

NiO + H₂SO₄→ NiSO₄ + H₂O (7)

WO₃ is insoluble in both leachants.^13,14 P₂O₅ reacts with water, and natural phosphates (for instance, monazite) are digested with hot strong mineral acids.²⁴ Based on data from Table 1, the theoretical HCl (H₂SO₄)/catalyst ratio is approximately 7.5 (3.75) mL g^-1. A 10 wt.% excess of the leachant was employed. Experiments were carried out using 10 g of sample in glass beakers. The experiments were run at 70 ºC for 1-3 h. The insoluble matter was separated by filtration, washed with water (4 mL g^-1), dried at 150 ºC for 2 h and weighed. The matter was then dissolved in HF + H₂SO₄ (1:2 v/v) at 60 ºC for chemical analysis. The experiments were run in triplicate, and errors for each experiment were consistently below 4%.

Analytical methods

Metal concentrations were determined by atomic absorption spectrometry (Perkin Elmer AAS 3300). The following wavelengths were employed: tungsten, 255.1 nm; nickel, 231.1 nm; aluminum, 396.2 nm. The following detection limits were determined experimentally: 1 mg L^-1 (aluminum), 0.5 mg L^-1 (nickel and tungsten). Phosphate ions were determined by ion-chromatography (Dionex DX-100). Solids were dissolved in a mixture of 2 mol L^-1 HF + 2 mol L^-1 HNO₃ at 60 ^oC prior to the analysis. A LECO analyzer determined carbon and sulfur contents. Acidity of aqueous solutions was determined by potentiometry using an Ag/AgCl reference electrode (Orion 2AI3-JG). Phase identification of the insoluble matter in the leachates was performed by X-ray diffraction (XRD) analysis (Bruker-AXS D5005) by the continuous scanning method at 35 kV and 40 mA, using Co Kα as the radiation source.

RESULTS AND DISCUSSION

After oxidation, the volatile matter corresponds to around 15 wt.% of the spent catalyst. This value agrees with the amounts reported in the literature for spent HDT catalysts^1,8,25 (12-30 wt.%). The volatile matter is basically due to coke burning (CO₂, H₂O). Also, sulfur was oxidized and partially eliminated as SO_x.^21,22

Alkaline fusion and acidic leaching

After fusion with NaOH, most tungsten and aluminum was leached with water, whereas phosphorous partially reacted with NaOH (Table 2). The average tungsten concentration in the leachate is 5 g L^-1, which falls within the range for LLE of the element with amines.^12,13 pH of the leachate is 11.9 ± 0.1, where tungsten is present as WO₄^2-.¹⁴ Iron and nickel remained in the insoluble matter. XRD data indicate that this solid is a poorly crystalline material. Small peaks corresponding to AlPO₄ and NiWO₄ were identified. XRD of the oxidized spent catalyst only contains small peaks corresponding to γ-Al₂O₃ phase.

Data in Table 3 and Figure 1 show that the oxidized spent catalyst was almost completely dissolved in HCl after 3 h at 70 ^oC despite the insolubility of WO₃ in acids.^13,14 The insoluble fraction contains mainly phosphorus (Table 3). XRD data showed that this solid is amorphous. The presence of a considerable amount of soluble phosphorous may explain the solubilization of WO₃ in HCl according to the following reaction:

24 WO₃ + 2 H₃PO₄→ 2 H₃PW₁₂O₄₀ (8)

Taking into account the composition of the oxidized spent catalyst (Table 1) the amount of phosphorus is much greater than that necessary to react with WO₃.

H₂SO₄ was a poorer leachant than HCl (Figure 1, Table 3). The amount of insoluble matter was higher, nickel and aluminum were partially leached and the amount of tungsten solubilized via reaction 8 was also lower. This result is the same as that found when a Brazilian tungsten concentrate was treated with these acids.²⁶ The difference between HCl and H₂SO₄ can be at least partially explained by the concentration of the latter (6 mol L^-1). In general, concentrated H₂SO₄ (> 9 mol L^-1) is used as the leachant for processing tungsten concentrates.⁹

Processing of leachate after fusion with NaOH

Figure 2 illustrates the overall scheme. The first step is of utmost important in order to separate tungsten from phosphorus. The pH of the leachate was adjusted from 11.9 to ~9 by adding 6 mol L^-1 HCl, after which the leachate was heated to 70 ^oC. A mixture of 0.1 mol L^-1 MgCl₂ + 5 mol L^-1 NH₄OH was added dropwise under stirring (200 rpm) until no white precipitate is formed.⁹ This solid was filtered. More than 99% of phosphorus precipitated as MgNH₄PO₄,^9,13 but aluminum (> 99.5%) also co-precipitated. No attempt was made to separate aluminum and phosphorus in this precipitate.

Tungsten was isolated by LLE. Alamine 336 (mixture of trioctyl and tridecylamine) was employed as the extractant (10 vol.% in kerosene). This presented very good results for tungsten recovery from Brazilian wolframite concentrates.²³ Experiments were performed at 25 ^oC, with an aqueous/organic phase ratio (A/O) = 1 (v/v). The equilibrium between phases was reached within around 5 min. The influence of pH (1-3) on tungsten extraction was studied by adding 6 mol L^-1 HCl to adjust the pH. According to Table 4, more than 99% of the element was extracted in one stage at pH 1. At pH 3, efficiency of the amine was greatly reduced.

Stripping (> 99.5%) was easily accomplished in one stage with 2 mol L^-1 NH₄OH (25 ^oC, A/O = 1 v/v). The equilibrium between phases was reached within 5 min. Concentrations of over 4 mol L^-1 led to emulsification (Table 4). The aqueous solution was slowly evaporated (~50 ºC) in a fume hood, yielding crystals of ammonium paratungstate (APT), (NH₄)₁₀W₁₂O₄₁.11H₂O. According to Table 5, its purity grade is very good. Only traces of phosphorous were found. 97.5% of tungsten present in the oxidized spent catalyst was recovered as APT.

The overall process of LLE and stripping of tungsten reported in this work agrees with literature data^15,27 for tungsten alloy scrap and tungsten concentrates^14,23 (which also contain phosphorus) in terms of yield and optimal experimental conditions.

The acidic solution after tungsten extraction is colorless. Neutralization with 2 mol L^-1 H₂SO₄ formed a final colorless and neutralized solution.

Processing of acidic leachate

Figure 3 depicts the overall scheme. The first step was to separate tungsten via LLE using the tertiary amine (Alamine 336) under the same conditions applied to the alkaline leachate except for a different A/O ratio (1/2 v/v) because tungsten concentration is higher (Table 3). Upon acidification, tungsten was converted to PW₁₂O₄₀^3- ions that were extracted^18,19 in one stage (99.5%) under the best experimental conditions shown in the previous section (Table 4). Stripping was also easily performed in one stage with 2 mol L^-1 NH₄OH (25 ^oC, A/O = 1 v/v) without emulsification. The aqueous solution was heated to 70 ^oC after which a mixture of 0.1 mol L^-1 MgCl₂ + 5 mol L^-1 NH₄OH was added⁹ dropwise under stirring (200 rpm) until no white precipitate (MgNH₄PO₄) was formed.^9,13 This solid was filtered. APT was isolated after slow evaporation of the filtrate. Its purity (Table 5) is comparable to APT obtained via the alkaline route. 99% of tungsten present in the oxidized spent catalyst was recovered as APT.

The raffinate was added dropwise to 6 mol L^-1 NaOH at 25^oC (200 rpm). pH was kept at around 12 by adding further NaOH. A green solid precipitated. XRD data showed that this solid is amorphous, but its chemical analysis (Table 6) suggests that this solid is Ni₃(PO₄)₂. The amount of soluble phosphorus is much higher than that necessary to precipitate soluble nickel and iron (Table 3). Nickel phosphate is more insoluble in water than its hydroxide, whereas iron phosphate is more soluble than its hydroxide.^28,29

Aluminum (Al(OH)₄^-) was precipitated as Al(OH)₃ after adjusting pH to 6.5-7 by adding 2 mol L^-1 HCl. The precipitate was filtered and washed with 0.01 mol L^-1 NH₄OH (4 mL g^-1) and water (6 mL g^-1).^28,29 The solid was dried at 150 ºC for 2 h and weighed. This precipitate does not contain nickel and tungsten (Table 6) but iron was co-precipitated. On the basis of phosphorus content, around 1.0 wt% of aluminum was precipitated as AlPO₄.

CONCLUSIONS

Oxidized spent NiW/Al₂O₃ catalysts can be processed in a similar way as oxidized spent NiMo/Al₂O₃ samples. Alkaline fusion (NaOH, 650 ^oC, 1 h) converted 98.5% of tungsten as water-soluble WO₄^2-. Aluminum was also extensively leached (> 90%) whereas nickel and 30% of the phosphorus remained in the insoluble matter. Acidic leaching was best conducted with HCl (70 ^oC, 3 h). Samples were almost fully dissolved but tungsten was present as PW₁₂O₄₀^3- ions.

Processing of the acidic leachate allowed recovery of tungsten (APT) and nickel (Ni₃(PO₄)₂), whereas aluminum was precipitated as Al(OH)₃ with minor amounts of AlPO₄ and iron. Processing of the alkaline leachate allowed recovery of tungsten by LLE without the presence of phosphorus but the latter and magnesium were co-precipitated with aluminum. Metals separation from the acidic leachate was more successful than from the leachate after alkaline fusion. High amounts of aluminum did not interfere in tungsten recovery, but phosphorus made separation steps more difficult, particularly aluminum recovery.

Despite the similarity of the chemistry of molybdenum and tungsten, LLE of the latter with tertiary amines in the presence of phosphorus occurred in a different way. At pH 1-2, while molybdenum is extracted as MoO₂²⁺ (phosphomolybdates are decomposed),²¹ tungsten was extracted as phosphotungstate ions, thus requiring an additional step to recover it as a pure APT.

ACKNOWLEDGMENTS

J. F. Paulino acknowledges CAPES for an M.Sc. fellowship. We are grateful to Centro de Tecnologia Mineral (CETEM/MCT) for XRD analysis.

Recebido em 10/5/12

Aceito em 26/11/12

Publicado na web em 4/4/13

• 1. Marafi, M.; Stanislaus, A.; Res. Conserv. Recycl. 2008, 53, 1.
• 2. Stanislaus, A.; Cooper, B. H.; Catal. Rev-Sci. Eng. 1994, 36, 75.
• 3. Brito, J. L.; Severino, F.; Delgado, N. N.; Laine, I.; Appl. Catal. A 1998, 173, 193.
• 4. Kunisada, N.; Choi, K.; Korai, Y.; Mochida, I.; Nakano, K.; Appl. Catal. A 2004, 269, 43.
• 5. Huges, R.; Catalyst Deactivation, Academic Press: London, 1984.
• 6. Jong, B. B.; Rhoads, S. C.; Stubbs, A. M.; Stoelting, T. R.; Recovery of principal metal value from waste hydroprocessing catalysts, U.S. Bureau of Mines, Report No. 9252, 1989.
• 7. Amiri, F.; Yaghmaei, S.; Mousavi, S. M.; Bioresour. Technol. 2011, 102, 1567.
• 8. Amiri, F.; Yaghmaei, S.; Mousavi, S. M.; Sheiban, S.; Hydrometallurgy 2011, 109, 65.
• 9. Luo, L.; Miyazaki, T.; Shibayama, A.; Yen, W.; Fujita, T.; Miner. Eng 2003, 16, 665.
• 10. Queneau, P. B.; Huggins, D. K.; Beckstead, L. W.; US pat. 4,320,096 1982
• 11. Srinivas, K.; Sreenivas, T.; Natarajan, R.; Padmanabhan, N. P. H.; Hydrometallurgy 2000, 58, 43.
• 12. Smith, B. J.; Patrick, V. A.; Aust. J. Chem. 2000, 53, 965.
• 13. Koutsospyros, A.; Braida, W.; Christodoulatos, C.; Dermatas, D.; Strigul, N.; J. Hazard. Mater. 2006, 136, 1.
• 14. Lassner, E.; Int. J. Refract. Met. Hard Mater. 1995, 13, 35.
• 15. Morís, M. A.; Díez, F. V.; Coca, J.; Sep. Purif. Technol 1999, 17, 173.
• 16. Brasser-Lecarne, C.; Baron, P.; Chevalier, J. L.; Madic, C.; Hydrometallurgy 1997, 47, 57.
• 17. Zhung, Z.; Tain, R.; Rhodes, C.; Can. J. Chem 2003, 81, 1044.
• 18. Zbranek, V.; Zbranek, Z.; Burnham, D. A.; US pat. 4,092,400 1978
• 19. Burwell B.; British pat 925,466 1965
• 20. Goddard, J. B.; US pat 4,346,061 1982
• 21. Valverde Jr., I. M.; Paulino, J. F.; Afonso, J. C.; Quim. Nova 2008, 31, 680.
• 22. Busnardo, R. G.; Busnardo, N. G.; Salvato, G. N.; Afonso, J. C.; J. Hazard Mater 2007, 139, 391.
• 23. Paulino, J. F.; Afonso, J. C.; Mantovano, J. L.; Vianna, C. A.; Cunha, J. W. S. D.; Hydrometallurgy 2012, 127-128, 121.
• 24. Abreu, R. D.; Moraes, C. A.; Miner. Eng 2010, 23, 536.
• 25. Pereira, A. L. S.; Silva, C. N.; Afonso, J. C.; Mantovano, J. L.; Quim. Nova 2011, 34, 145.
• 26. Paulino, J. F.; Afonso, J. C.; Mantovano, J. L.; Vianna, C. A.; Cunha, J. W. S. D.; Quim. Nova 2012, 35, 1854.
• 27. Tamhina, B. Herak, M. J.; J. Inorg. Nucl. Chem 1977, 39, 391.
• 28. Feigl, F.; Spot Tests In Inorganic Analysis, Elsevier: Amsterdam, 1958, chap. 3.
• 29. Vogel, A. I.; Qualitative Inorganic Analysis, 5^th ed., Mestre Jou: São Paulo, 1981.

Publication Dates

• Publication in this collection
  18 July 2013
• Date of issue
  2013

History

• Received
  10 May 2012
• Accepted
  26 Nov 2012