Sources of Metals in Oils
Crude oils contain various trace elements, especially V and Ni and to a lesser extent Fe, plus a wider range of other metals in trace amounts. The concentration of these metals depends on a range of factors, including the origins and locations of the rock in which the oil is found and the composition and age of the oil [1]. The lowest levels of trace metals occur in light condensate oils (<1 ppb) and their concentration increases with the density of the oil, reaching 2,000 ppm V and 200 ppm Ni in asphaltic oils and up to 5,400 ppm V and 2,600 ppm Ni in bitumens. There are also considerable variations in concentrations of these metals, even in oils from the same geographical area. More typically, the level of V in residual oil is in the range of 30–200 ppm, although fuel oils arising from crude stocks from Venezuela and Mexico in particular can result in higher V levels [2].
The V and Ni are present mostly as various organic complexes, metal sulphides or occasionally arsenides, depending on the age and depositional environment of the oil and on its interaction with the surrounding rocks. Whilst some of the trace metals can be removed by filtration and other physical processes, the organic metal complexes of several key metals, including Ni, V, Fe, Zn and Ca, are soluble in oil [3] and, along with S, become increasingly concentrated in the heavier fractions of oil produced by refinery processes. When these heavy oils are burnt, the soluble organo-metallic complexes decompose to oxides and sulphides and form fine ash particles and soot, where the metals become even more concentrated.
Fuel Ash from Oils
Heavy fuel fly ash particles have a carbonaceous matrix and contain the trace metals in the form of various metal oxides and sulphates [4, 5], with a typical particle size in the range of 0.9–500 microns, depending on the efficiency of the combustion [4, 6]. These particles are of particular interest to users of the oil because the various metal oxides (especially those of Na and V) can combine to form low melting point phases on exposed surfaces, where they can cause fouling and/or erosion of equipment, although this can be alleviated by the inclusion of ash modifying fuel additives [2].
The amount of fly ash produced depends largely on the metals content in the oil. It is reported [6] that 837,000 tonnes of heavy fuel oil were used in Jordan for power production in 2010, giving rise to 420 tonnes of fly ash (0.05 wt%), a figure consistent with the typical distribution of ash contents of marine fuel oils, which shows a peak at around 0.04 wt% ash and with very little exceeding 0.1 wt% ash [2].
There is also considerable variation in the carbon and water content of fly ash, depending on the source and the processes used to generate and collect the particles, so that the true ‘ash’ fraction of the fly ash is typically less than half the overall mass of the fly ash. For example, various authors [4, 6, 7] report a true ash content of typically around 36 wt% in heavy fuel oil fly ash. Consequently, in many cases the fly ash undergoes a further oxidation step to gasify the remaining carbon and residual sulphur to produce a final soot ash material that is largely free of carbon, moisture and sulphur [5, 8]. Not only does this extract the remaining calorific value in the fly ash, but it also reduces the mass and increases the concentration of the valuable metals in the final feed material to the smelting process.
Reported concentrations of V, Ni and Fe in soot ash (wt% of soot ash )
Source | HFOa | HFOa | HFOa | HFOb | VRa | Asphaltinec |
|---|---|---|---|---|---|---|
Fe | 13.7 | 4.0 | Not measured | 1.5 | 5.0 | 2.5–8.9 |
Ni | 8.0 | 7.0 | 8.5–11.3 | 11.1 | 11.0 | 2.4–17.2 |
V | 37.0 | 25.0 | 32.0–39.6 | 59.9 | 35.0 | 29.7–54.7 |
As can be seen from the data in Table 1, the main metals (V, Ni, Fe) comprise typically over half the mass of the soot ash . Since these metals are in the fully oxidised condition, the soot ash consists largely of these three main metal oxides. The main exception to this is Na, which can be typically 2–5 wt% of the ash and in some cases fuel additives as well, which can increase the concentration of MgO and other ash-modifying compounds.
Recovery of Ni and V from Soot Ash
The data in Table 1 demonstrate that the soot ash material arising from the burning of heavier oil fractions, such as heavy fuel oil, bitumen, asphalt, etc. represent potentially highly desirable sources of Ni and V. The demand for these metals grew at average rate of 5.9% per year and 4.6% per year respectively between 1994 and 2013, which represents a doubling of demand roughly every 14 years at a time when primary ore grades are declining and extraction costs are escalating [9]. Since future demand trends will see V and Ni increasing demand for use in catalysts and battery applications, it seems likely this current rate of growth will be maintained.
Ni and V have been recovered from petrochemical-based waste materials using various hydrometallurgical and pyrometallurgical methods. The aim of such methods is to achieve high levels of recovery of the metals from the ash or soot and then to achieve a high degree of separation of the various target metals from each other.
Pyrometallurgical methods include various smelting techniques to produce one or more ferroalloys [5, 7, 8, 10, 11]. Hydrometallurgical methods start by leaching the V and Ni from the material under acidic, neutral or basic conditions, resulting in an aqueous solution containing the recovered portion of the V and Ni, along with a range of other metals that have also been leached. The various metals are then separated from each other by selective precipitation , ion exchange or organic solvent-based techniques [8]. Potential leaching solvents include orthophosphoric acid, sulfuric acid , perchloric acid, nitric acid, ammonium sulphate, sodium, sodium carbonate, EDTA and sodium chloride [4, 12]. It is also possible to conduct a salt roast operation to convert the V into soluble sodium vanadate prior to leaching [13]. The various leaching media have different affinities for the target metals, which allows a degree of selectivity in the initial extraction process, which is then complemented by selective separation methods from the aqueous solution.
Tetronics DC Plasma Smelting Technology
Tetronics’ DC Plasma Arc technology is one of several pyrometallurgical solutions available for the recovery of valuable base metals from secondary sources, the other key ones being shaft furnace and submerged arc furnace smelting . Shaft furnaces and submerged arc furnaces are both typically used at tonnages of more than 10,000 tonnes per month of alloy output [14], which is likely to be an order of magnitude larger than the typical furnace output that could be justified by soot ash availability. Clearly, the amount of soot ash generated in each country or region will depend on many factors, but it seems highly likely that technologies suitable for smaller annual tonnages are likely to be more naturally suited to this application.
Tetronics’ DC plasma smelting technology is ideally suited to operating at these lower annual tonnages as a result of the greater arc stability provided by the plasma. Tetronics’ DC plasma smelting technology has been used commercially to recover Ni, Cr, Mo, Mn and Fe from stainless steel dusts at between 8,000 and 23,000 tonnes per year since 1989 [15, 16], in one case leading to the realisation of $190 million of value from these wastes over a 20-year period [17]. Tetronics’ DC plasma arc technology has also been used for the commercial recovery of platinum group metals (PGM) from automotive and industrial catalysts at between 1,000 and 3,000 tonnes per year for over 30 years. More recently, the technology has been applied commercially to the recovery of precious metals and PGM from printed circuit boards [18].
The benefits of plasma smelting include its ability to operate with a wide range of dusty feed materials without the need for briquetting, tight control of smelting conditions and minimal losses of dust to the exhaust gases as a result of the low gas flows associated typically with DC plasma arc-based processes. The ability to vary the reaction conditions and furnace temperatures also allow the operator to separate the elements in the charge into gaseous, slag and metal fractions, which in turn maximises the value recovered from the waste and minimises the secondary wastes, whilst achieving low ultimate emissions to air and water.
Process Description
- (a)
The recovery of Ni as a Fe–Ni alloy and a V-rich slag via a carbothermic reduction process
- (b)
The recovery of V as a Fe–V alloy from the first stage slag via an aluminothermic reduction process.
The oxides of Ni and Fe are readily reduced by C, but the oxides of V are more stable and require a stronger reductant , such as ferro-silicon or aluminium . This large difference in response to reduction reactions provides a simple and reliable method of extracting Ni and V separately in this two-stage manner that can be easily exploited via a plasma smelting plant. In this case, Al is the preferred reductant for the second stage, because this has been found to result in higher V recovery efficiency and lower levels of Si in the final Fe–V alloy product than when using Si-based reductants [11]. Further processing of the alloys (especially the first stage Fe–Ni alloy) may be required in order to adjust the composition to meet standard ferroalloy specifications, especially for the removal of S and/or P.
Plant Description
Schematic process flow diagram for a typical DC plasma smelting plant
As described above, fly ash from the burning of heavy oil fractions contains significant amounts of residual carbon and sulphur, plus (in some cases) water as well. Materials containing high levels of water can be heated safely and effectively in a typical DC plasma arc smelting furnace , because the process is operated typically in a ‘red top’ mode, i.e. with an open arc and a fairly shallow bed of charge material on the top surface of the slag , which prevents steam pockets building up in the charge layer. However, the energy required to heat water to smelting temperatures is considerable and it usually more economic to dry the feed first.
The presence of a significant and variable carbon content in the feed could prevent the operator selecting the most appropriate level of reduction for the process and increases the difficulty of maintaining a fixed set of operating conditions. Meanwhile, the sulphur content in fly ash risks contaminating any alloy produced and thereby increasing the extent and cost of downstream processing required to bring the final ferroalloy products up to an acceptable quality.
Given these requirements to remove the carbon, water and sulphur from the feed, it is usual to subject the material to a calcining or roasting step [7, 19].
The melting points of the various oxides of vanadium vary widely. In the initial soot ash feed material, V2O5 is the predominant oxide, whereas under the strongly reducing conditions of the second aluminothermic reduction stage, V2O3 is the predominant form. The melting points of vanadium oxides increase with increasing V/O ratio, being 690 and 1940 °C for V2O5 and V2O3 respectively. Furthermore, the soot ash typically contains few slag -forming oxides beyond those of the three target metals (Ni, Fe and V). Therefore, the soot ash is blended with slag -forming additions to create a slag with the most appropriate melting point and viscosity . In common with most ferroalloy practice, a slag based on CaO–FeO–SiO2 is the most obvious choice here. Carbon is also added as a reductant , which is typically in the form of metallurgical coke or anthracite. The charge materials are then fed into the furnace via one or more feed ports in the furnace roof.
Schematic diagram of a plasma furnace
The gases from the outlet of the furnace pass into a secondary combustion chamber, which oxidises flammable species and volatile metals. The combusted off-gas is cooled rapidly to circa 180 °C using water injection to prevent reformation of dioxins and furans and then passes through a filter unit to remove particulates. A small proportion (typically 1–2%) of the blended feed is carried over into the exhaust gas and is collected in the filter for disposal, sale or recycling back in the plasma furnace , depending on its composition. Any acid gases from residual sulphur in the feed are abated by dry sorbent injection upstream of the filter unit. The pressure inside the plasma furnace is carefully controlled to be just below atmospheric pressure by means of a variable speed induced-draught fan downstream of the filter unit, which minimises ingress of air or egress of process gases. The cleaned exhaust gas is then vented to atmosphere via a stack, in accordance with local environmental regulations.
Process Example
Process Inputs
Typical pre-treated and calcined soot ash composition
Species | wt% |
|---|---|
Al | 0.38 |
Ca | 0.32 |
Fe | 13.66 |
Mg | 0.14 |
Mo | 0.44 |
Na | 0.88 |
Ni | 8.00 |
P | 0.03 |
S | 0.29 |
V | 37.0 |
Zn | 0.28 |
Others | <0.1 each |
Total oxides | 99.0 |
Taking a typical consumption of 400,000 barrels of fuel oil per day as the source material for the fuel soot ash and average fuel oil density of circa 900 kg/m3, the total annual fuel oil giving rise to soot ash is circa 20.9 m tonnes per year. With an average ash content of 0.035 wt%, the total amount of soot ash to be processed by a typical plasma plant (after calcining) is circa 7,300 tonnes per year. Assuming this material is being fed into a plasma smelting unit at nominal plant capacity for 75% of the available hours in the year, the mean feed rate into the plasma furnace is therefore 1,111 kg/h. Naturally, the actual feed rate will vary significantly depending on a wide range of circumstances, but this would seem a reasonable assumption for a typical plant size in this application. Note that this same furnace system could also be used for the extraction of Ni, V, Mo and other valuable base metals from petrochemical catalysts [9].
Process inputs—carbothermic reduction stage (7,300 t/year of pre-treated and calcined soot ash )
Feeding materials | Formulation (%) | Feed rate (kg/h) |
|---|---|---|
Soot ash | 62.1 | 1.111 |
Flux , lime (100% CaO) | 13.1 | 233 |
Flux , silica (100% SiO2) | 15.8 | 283 |
Reductant (100% C) | 9.0 | 160 |
Total | 100.0 | 1.801 |
Plasma power | 2.2 MW | |
Process block diagram for the soot ash treatment process
The first stage reduction process gives rise to a ferronickel alloy and a vanadium oxide-rich slag with a low residual level of FeO. Therefore, to create a ferrovanadium alloy from the second stage reduction process, it is necessary to add Fe with the Al in the correct proportions to achieve the appropriate target alloy composition. This can be added as high-quality steel scrap, ferro-aluminium deoxidiser or some other form as appropriate.
Alloy Products
Expected alloy compositions (wt%)
Metal | Carbothermic stage | Aluminothermic stage |
|---|---|---|
Al | 0.0 | 0.01 |
C | 0.14 | 0.0 |
Fe | 59.2 | 24.5 |
Mo | 2.1 | 0.0 |
Ni | 38.0 | 0.34 |
P | 0.144 | 0.0 |
S | 0.129 | 0.0 |
Si | 0.0 | 1.5 |
V | 0.25 | 73.6 |
Data from commercial and pilot plant operation indicates the recovery rates for Ni from steel plant dusts and other related materials via carbothermic reduction in Tetronics’ DC plasma smelting furnaces is typically between 95 and 98% of the input amount [19] and this is expected to be replicated here. Meanwhile, Howard et al [11] report recovery rates of 93% for V from V-rich slags via aluminothermic reduction .
The data in Table 4 show that the overwhelming majority of the P and S in the soot ash are expected to be removed in the plasma smelting stage and that the second stage Fe–V alloy is expected to contain low levels of these contaminants. Since the Fe–Ni alloy is likely to contain higher levels of S and P than are typically required for ferroalloys (<circa 0.05%), it will probably be necessary to apply some form of desulphurisation and/or dephosphorisation process to the Fe–Ni alloy before being suitable for reuse in the steel industry. It can also be seen that any Mo contained in the soot ash will tend to report to the first stage Fe–Ni alloy.
The melting point of Fe–Ni alloys varies by less than 100 °C over the whole composition range from 0 to 100% Ni, so no adjustment in Fe/Ni ratio is required to manipulate the melting point of the alloy. Conversely, the addition of V to Fe lowers the melting point from 1538 °C (100% Fe) to a minimum of circa 1460 °C at around 30–40% V; further additions of V raise the melting point, so that the melting point is circa 1700 °C by circa 78% V. The typical specifications for ferrovanadium alloys lie in the range 35–80% V, so there is flexibility to adjust both the Fe content and the melting point of the final alloy by changing the Fe addition level as required.
Slag Products
Predicted slag compositions (wt%)
Slag | Carbothermic stage | Aluminothermic stage |
|---|---|---|
Al2O3 | 0.68 | 44.6 |
CaO | 20.1 | 23.6 |
FeO | 1.75 | 0.00 |
NiO | 0.19 | 0.02 |
SiO2 | 24.2 | 26.8 |
V2O3 | 23.0 | 3.90 |
V3O5 | 29.64 | 0.44 |
Discussion
It can be seen from the process block diagram (Fig. 3) that the volume of slag being generated form the initial smelting process is several times greater than the volume of metal generated over any given period. This offers the possibility of using a continuous overflow system for at least some of the slag removal , so as to allow the level of metal to build up over a longer period. However, the subsequent aluminothermic reduction process requires the use of liquid slag as a major starting material, which means batch tapping into a secondary treatment ladle is advisable for at least some of the slag . Aluminothermic reduction is highly exothermic and it is possible that a mixture of hot liquid slag and cold slag and/or steel scrap could be used in order to avoid overheating of the slag , metal and ladle during this process. Conversely, as suggested by Howard et al. [11], the heat generated ensures that this process should be possible without the need to add an external source of heat, which allows the aluminothermic stage to be carried out in a ladle or similar, rather than in a separate furnace .
Conclusions
Continuing growth in demand for Ni and V coupled with the growing financial and environmental costs of primary extraction mean the oil soot ash generated by the burning of heavy fuel oils, bitumen and related materials represent potentially valuable secondary sources of these metals. However, extraction of the metals presents both opportunities and challenges for potential extraction methods. For several decades, Tetronics’ DC plasma smelting technology has been used for the recovery of valuable base metals from steel plant dusts and other similar materials in compact, environmentally friendly and efficient plants at the relatively modest scale required by the availability of oil soot ash . A two-stage extraction method, based on generating a ferronickel alloy by carbothermic reduction , followed by aluminothermic of the resulting V-rich slag , makes DC plasma smelting an obvious part of a wider flowsheet for smaller-scale extraction operations based on this important niche secondary source of these key metals.