Introduction
Cerium oxide makes up about 46% of the rare earth oxides (REO) in a typical carbonatite deposit and about 33% of the REO in the average igneous or hydrothermal deposit. Ionic clay deposits typically have a similar in situ cerium content, but Ce is usually present as Ce(IV) in cerianite—(Ce, Th)O2—resulting from natural oxidation processes. Cerianite is insoluble in the mild leach systems used, so the product from most ionic clay deposits is very low in cerium, typically well under 5%.
The cost for the reagents and utilities needed to process material through a TBP (tributyl phosphate)-nitrate or a P507-chloride separation plant is roughly the same for each of the light REE . So, elimination of Ce from the plant feed substantially reduces the annual operating costs of a separation plant—by roughly 50%.
A quick analysis indicated that the cost for reagents and utilities used to separate the LREE is about $4/kg of REO separated. Cerium oxide is in an oversupply situation and prices in early 2018 were about US$2/kg of 99.95% CeO2. If the separation plant must expend $4/kg to make a product that only sells for $2/kg, there is an obvious incentive to remove Ce before it ever gets to the separation plant.
Cerium Chemistry
Ce in solution is normally trivalent and therefore behaves much as the other REE. However, Ce(III) can be oxidized to the quadrivalent state, Ce(IV), after which its behavior is totally different to the other REE and Ce is readily separated from the other REE through selective dissolution, precipitation or solvent extraction (SX).
Pourbaix diagram for Ce [1]
Standard redox potentials of potential Ce oxidants
Ion | Reaction | Oxidation Potential SHE/V | ||
|---|---|---|---|---|
Perchlorate | ClO4− + 2H+ + 2e− | ⇌ | ClO3− + H2O | 1.2 |
Oxygen | O2(g) + 4H+ + 4e− | ⇌ | 2H2O | 1.23 |
Chlorine | Cl2(g) + 2e− | ⇌ | 2Cl− | 1.36 |
Chlorate | 2ClO3− + 12H+ + 10e− | ⇌ | Cl2(g) + 6H2O | 1.49 |
Permanganate (acid) | MnO4− + 8H+ + 5e− | ⇌ | Mn2+ + 4H2O | 1.51 |
Cerium (HNO3)1 | Ce4+ + e− | ⇌ | Ce3+ | 1.61 |
Hypochlorous acid | 2 HClO(aq) + 2H+ + 2e− | ⇌ | Cl2(g) + 2H2O | 1.63 |
Permanganate (neutral) | MnO4− + 2H2O + 3e− | ⇌ | MnO2(s) + 4OH− | 1.69 |
Permanganate | MnO4− + 4H+ + 3e− | ⇌ | MnO2(s) + 2H2O | 1.7 |
Hydrogen peroxide | H2O2(aq) + 2H+ + 2e− | ⇌ | 2H2O | 1.78 |
Persulphate | S2O82− + e− | ⇌ | SO42− | 2.1 |
Ozone | O3(g) + 2H+ + 2e− | ⇌ | O2(g) + H2O | 2.2 |
Fluorine | F2(g) + 2e− | ⇌ | 2F− | 2.87 |
Ce(IV) precipitates at a pH of about 2.7 whereas Ce(III) and the other trivalent REE precipitate in the pH range of 6 to 8. Thus, if Ce(III) is oxidized at a pH greater than about 2.7, the Ce(IV) produced will likely be precipitated.
Ce(IV) is preferentially extracted by solvents away from other REE. TBP and Cyanex923 are effective from nitrate solutions; carboxylic and organophosphate solvents from sulphate or chloride solutions [3–5].
Cerium Separation Methods
Oxidation of Cerium in Solution
Ozone. Bauer and Lindstrom [6] reported on the separation of Ce by ozone oxidation from sulphate, nitrate, and chloride solutions containing REE in the proportions of 50% CeO2, 36% La2O3, 10% Nd2O3, and 4% Pr6O11 and others. In preliminary experiments, Bauer and Lindstrom reported on oxidation with air or oxygen. Cerium oxidation only started to be seen at 85 ºC and a pH of 6.5, but co-precipitation of the tri-valent REE was significant for example, 23% of the La was precipitated.
High recovery of high purity Ce precipitate using ozone was obtained from sulphate, chloride , and nitrate solutions at ambient temperature with operation at pH 6.5 giving the fastest and most complete reaction. At 85 ºC, the same pH trend was observed, but reaction times were much shortened. Taking Ce and non-Ce REE purity into account, Bauer and Lindstrom recommended a pH range of 4–5. Under these conditions, 98% recovery of Ce at 98% purity was obtained. The authors investigated a two-step ozonation process to maximize Ce and La purities in which Ce and HREE were first precipitated to make a pure La filtrate followed by redissolution of the Ce and its re-ozonation and precipitation to make a pure Ce.
Hypochlorite. When Molycorp upgraded the Mountain Pass plant, it replaced the bastnaesite roasting operation, previously used to oxidize Ce(III), with a process that dissolved all the REE in HCl, precipitated impurities by hydrolysis and sulphide precipitation , and then oxidized Ce(III) to Ce(IV) upon which the Ce precipitated and was removed from solution. Data are not given, but it is reasonable to assume that sodium hypochlorite, generated by the in-plant chlor-alkali facility, was used as the oxidant [7, 8]. The pH is not stated, but if the oxidation was done after impurity removal , it would be about pH 4 and the Ce(IV) would be precipitated.
Ho et al. [9] researched the effect of dose, temperature , and pH on the oxidation of Ce using hypochlorite. In a mixed REE solution, 99% oxidation was obtained at 2.5 times stoichiometric hypochlorite addition. A pH of 4 appeared to give best separation of Ce from other REE with poor selectivity at higher pH values. A temperature of 60 ºC was optimal. Significant co-precipitation of other REE, especially the heavier REE occurred at all levels investigated. For example, with 1.9 times stoichiometric NaOCl addition, 60 ºC, and pH 4, levels of co-precipitation were about 30% HREE, 20% medium REE, 12% Nd, 9% Pr, and 5% La.
Permanganate. Abreu and Morais [10] processed monazite through a series of steps to arrive at a chloride solution. The Ce was then oxidized to Ce(IV) using potassium permanganate. The optimum pH was found to be about 3 where recovery and Ce purity regarding REE were both about 99%. About 30% excess permanganate was needed.
MnO2 was co-precipitated with the CeO2 such that the product contained 69% CeO2 and 14% MnO2. This was re-dissolved in HCl and the Ce re-precipitated as hydroxide or oxalate yielding CeO2 with purities of 99% and 99.5%, respectively.
Hydrogen peroxide . Nechaev et al. [11] examined an oxidation process in which hydrogen peroxide was added to a nitrate solution along with ammonia to hold the pH at 5.5–6. Optimum conditions were found to include a pH of 5.8 and 130% stoichiometric H2O2 leading to a product containing 88.6% CeO2 with Ce precipitation of 100%. Re-pulping and washing the precipitate led to a 97% yield of cerium at 99.5% purity while the non-Ce REE contained about 0.1% Ce on a REE basis. It is not clear if the process has moved beyond the laboratory.
Electrolytic oxidation . Electrolysis of chloride solutions leads to the formation of hypochlorous acid (HClO), chlorite ion (ClO–) and chlorate ion (ClO3–). Vasudevan et al. [12] oxidized Ce(III) through the in-situ electrolytic generation of such oxidants. A temperature of about 30 ºC, a pH of about 6.7, and concentration of 130 g/L CeCl3 were found to be optimum. The cathode of the electrolytic cell was rotated to prevent coating formation. Using a solution containing a typical LREE mixture, cell electrochemical efficiency was reported as 60% and the purity of the precipitated Ce(OH)4 was 95%. The impurities appear to be other REE, which implies significant losses.
Martin and Rollat [13] patented a system in which Ce(III) in a mixed REE nitrate solution was electrolytically oxidized to Ce(IV) in the anode compartment. Ce(IV) was then selectively extracted away from the other REE by, preferably, TBP. The loaded organic was then stripped by placing it, with nitric acid, in the cathode compartment of the same electrolytic cell where it was reduced to Ce(III) and simultaneously stripped from the solvent. In an alternative stripping process, nitric acid was reduced in the cathode compartment to form NO and NO2 gases, which were mixed with the loaded solvent to form Ce(III), which was then water stripped. It is understood that this system was operated on an industrial scale.
Electrolytic oxidation of Ce(III) in nitrate solution is employed in the 2,000 t/a REO separation plant operated in East Kazakhstan by Irtysh Rare Earths Company Ltd. (IRESCO). The original anodes were made from platinum mesh and were expensive and not overly effective. It appears that improved cell designs have been developed using diaphragms and improved anode materials. Discussing new cell designs, Gasanov et al. [14] reported 98% current efficiency and an energy demand of 0.6 kWh/kg of Ce(IV) produced.
Yarosova et al. [15] reported on the electrolytic oxidation of Ce(III) in a mixed REE nitrate solution using a cell with four platinized titanium anodes and a titanium cathode contained in a ceramic diaphragm. More than 99% Ce(III) oxidation was obtained and energy demand less than 0.8 kWh/kg of CeO2. The oxidized solution, containing 55 g/L CeO2 and 70 g/L HNO3, was processed in a series of centrifugal contactors with 30% TBP in kerosene to make a 99.5% pure Ce product. Interestingly, the β value for Ce(IV)/La and for Ce(IV)/Pr were reported as 364 and 121, respectively, showing how readily the cerium is purified once it has been oxidized.
Although several papers on the production of Ce(IV), usually in nitrate media , are aimed at refining Ce, several papers and patents discuss the production of Ce(IV) for use as an oxidant in organic syntheses or destruction of pollutants. Some of these discuss operations from a sulphate medium. There are several papers and patents, many Russian, that discuss the construction of electrolytic cells for the production of Ce(IV).
Photochemical oxidation . Donahue [16] explored the use of lasers for the photooxidation of Ce(III), preferably using ultra-violet (250 nm) light, followed by precipitation of Ce(IV) as the iodate Ce(IO3)4—although separation by SX of Ce(IV) is discussed.
Workers in Japan [17] used ultra-violet light irradiation and a low concentration (0.05 mol/L) of KBrO3 to oxidize Ce(III) in a mixture of La, Ce, and Nd followed by SX of the treated solution using di-2-ethylhexyl phosphoric acid. Separation factors between Ce and adjacent REE were about 400.
Oxidation of Dried Solids
Concentrates. The classic example of the oxidation of solid, REE-bearing material to allow early separation of cerium is Molycorp’s practice from the 1960s through to about 2000. Molycorp dried its bastnaesite concentrate before roasting at 650 ºC in a multiple hearth roaster which drove of fluorine and oxidized Ce(III) to Ce(IV). A cold dilute HCl leach solubilized all the REE except for the Ce(IV), which reported to the leach residue. The leach residue was sold as a glass polishing powder although a portion was purified to give a carbonate of 96% CeO2 purity. Leach recovery of the tri-valent REE was not complete and the Ce(IV) product, which contained about 60% Ce, did carry about 10% non-Ce REE. The Ce-depleted HCl leach solution went on to impurity removal and the separation of individual REE [7, 8].
Oxidation of cerium by roasting the bastnaesite concentrates derived from Bayan Obo and Sichuan ore has featured in several Chinese papers including, as an example, a 2013 paper by Wang et al. [18]. This paper additionally describes the dissolution of the oxidized concentrate in sulphuric acid and use of P507 to preferentially extract Ce(IV), Th(IV) and F(I) followed by sequential stripping of F, Ce, and Th using Al(III), H2O2, and H2SO4, respectively.
Peak Resources proposes [19] to roast its bastnaesite for one hour at 700 ºC to oxidize Ce, but with soda ash to allow subsequent fluorine removal by water leaching . The roasted and fluorine-free calcine would then be leached with 1% HCl solution to selectively extract non-Ce REE away from the Ce in a process similar to that practised earlier at Mountain Pass. Detailed deportment data are not provided.
Precipitates. Oxidation of Ce through drying of hydroxide precipitates has been examined as a means of separating Ce from the mixed REE and Th hydroxides produced by caustic cracking monazite . Using hydroxide containing about equal proportions of Ce and other REE , Swaminathan et al. [20] showed that a drying temperature of 160 ºC was optimal with higher temperatures leading to difficulty in subsequent HCl leaching . The Ce in a 5 mm thick layer of hydroxide attained 95% oxidation to Ce(IV) after 6 h and 99% after 12 h. The dried and oxidized hydroxide was leached with HCl at 50% solids at pH 3 and produced a solution containing 95% of the trivalent REE and CeO2/REO of about 7%. The CeO2 in the leach residue was selectively dissolved away from the Th and gangue materials using HCl at pH 1.5 with a sulphite reductant leading to a product containing up to 85% CeO2.
US Patent 3,111,375 [21] covers the same subject and additionally discusses thorium elimination from the pregnant solution and methods of handling and separating the mixed ceric and thorium hydroxides.
Great Western Minerals proposed [22] to acid bake a monazite concentrate at its Steenkampskraal operation, precipitate a double sulphate, metathesize that to the hydroxide using NaOH, then dry the hydroxide to effect oxidation of Ce(III) to Ce(IV). Dilute HCl leaching was then to be used to leach 99% of the non-Ce REE, less than 1% of the Th and, it appears, 12% of the Ce.
Oxidation of Moist Solids
Although not detailed in the work by Zou, treatment of the oxidized hydroxides in a weak acid would lead to dissolution of the REE except for Ce(OH)4. Probably there would be some residual non-Ce hydroxides left with the Ce(OH)4 leading to some loss of the other REE —and some residual Ce would remain with the tri-valent REE. Such is the experience of Molycorp and other researchers. The purity of the Ce and non-Ce fractions was not discussed in the paper.
Discussion
Processing Costs
To put the oxidation options into context, we can consider a hypothetical plant producing 4,000 t/a of REO of which 50% is cerium oxide. Assuming 8,000 h/a operating time, the production rate is 500 kg/h of REO containing 250 kg/h of CeO2. Some comments on economics follow.
Ozone. Data from Bauer [6] suggest an ozone demand of about 0.3 kg/kg of CeO2 or 75 kg/h of O3 in our hypothetical plant. This is a large amount of ozone. Ozonia [24] is an important producer of industrial-scale ozone generators and only its largest unit would have the required capacity and only when fed with pure oxygen. An informal quotation indicated a capital cost of M$5 and a power consumption of 8 kWh/kg [25].
At a power cost of $0.05/kWh, the cost for power to oxidize Ce using ozone would be about $0.10/kg Ce. Additionally, there will be the cost for oxygen (probably VPSA generated) which might equate to an additional $0.05/kg Ce for a total of about $0.15/kg Ce.
The operating cost for oxidation with ozone seem reasonable and the selectivity is attractive. However, the high capital cost of the ozone generator is a concern. If the M$5 capital cost is spread over 3 years Ce production (6,000 t), it equates to about $1.00 per kg of Ce, which is substantial.
Chemical oxidation . Sodium hypochlorite, potassium permanganate, and hydrogen peroxide can variously be used for oxidation . Using stoichiometric factors from the literature of 2.5, 1.3, and 1.3, respectively, we calculated the reagent demands, expressed on a 100% basis, as 0.66, 0.49, and 0.16 kg/kg Ce(IV) produced, respectively. The cost of each reagent is the same at about $1/kg. Thus, the costs of the reagents for chemical oxidation is approximately $0.66, $0.49, and $0.16/kg of Ce oxidized, respectively.
Electrolytic oxidation . Vasudevan [12] reported an energy consumption of 3.5 kWh/kg of CeO2 or 4.3 kWh/kg Ce. Russian workers [14] have general discussed values around 1 kWh/kg. If we assume that the total power demand is 4 kWh/kg of Ce oxidized, then we have a cost of $0.20/kWh which is lower than the cost for ozone oxidation . The Irtysh plant in Kazakhstan is processing 2,000 t/a of REO using electrolytic oxidation , so our hypothetical 4,000 t/a REO plant does not require a large scale-up factor.
Although there is some data on energy demand, there is little on the capital cost of the complex cells required to oxidize Ce(III) to Ce(IV). Our hypothetical plant processing about 200 kg/h of Ce and using 4 kWh/kg Ce will need cells with a capacity of about 1,000 kW. If we suppose that such cells would cost $2,000/kW, we have a capital cost of M$2 which, over a 3-year period (6,000 t of Ce) equates to about $0.33/kg of Ce oxidized. Added to the cost for power of about $0.20/kg Ce, we have a total oxidation cost of $0.53/kg Ce.
Oxidation after drying. If we assume that a precipitated REE hydroxide contains 50% moisture and the solids contain 50% Ce, then to oxidize Ce by drying requires that 2 kg water must be evaporated for every 1 kg of Ce. At a steam cost of $15/t, evaporation costs equate to $0.30/kg Ce. The cost to blow air over the dried material would be very low.
Slurry oxidation . The costs to blow air through a slurry of hydroxides, as described by Zou et al., might be in the order of $0.10/kg of Ce.
Other costs. As well as the basic oxidation costs outlined above, there will be other costs such as those for the precipitation of hydroxides or other species prior to the oxidation of dried or slurry hydroxides, costs for re-dissolving, etc. The extent of these costs will depend on the prior steps in the process—and might not be assignable to the cerium oxidation process.
General. The very high-level estimates of the owner’s costs, calculated as the sum of operating costs and the capital cost spread over three years Ce production, is in the range of $0.10/kg to just over $1.00/kg with slurry oxidation being lowest cost and ozone oxidation being the highest.
The chemical oxidation methods sometimes involve chemicals that require special handling and in some cases introduce unattractive chemical species into the process environment .
Product Purity
Another aspect of the cerium removal process requiring consideration is the cleanness of the separation and the recovery of the more valuable REE. The drying and chemical oxidation processes generally seem to yield products that are not well separated with some Ce remaining in the “low” Ce stream and significant losses of non-Ce REE to the Ce fraction. The literature information suggests that ozone and the electrochemical processes, and possibly the photooxidation and oxidation-SX approaches are better regarding the purity of the two product streams.
Conclusions
There are several ways to oxidize Ce(III) to Ce(IV) and thereby facilitate its early removal from a bulk REE mixture. Certain methods, including the air oxidation of dried precipitates and electrolytic oxidation , have been practised at an industrial scale.
The conceptual costs of the various options cover a wide range from simple systems, such as oxidation of neutralized slurries at the lower end to ozone systems which appear to be low cost to operate, but require a high capital expenditure.
Product purities and recoveries are, of course, inter-twinned and the several separation methods offer a wide range of combinations with the ozone and electrolytic options appearing to offer the best separation performance.
All told, we suspect that the electrolytic oxidation option offers the best combination of capital expenditure, operating cost, recovery , and product purities. Perhaps that is why it has enjoyed significant industrial usage. That said, each potential cerium pre-separation case will depend on the basic flowsheet producing the mixed REE material, local reagent and utility costs, and the objective of the exercise. Testwork and studies will be required for each case.