Ultrasound-assisted leaching of vanadium from fly ash using lemon juice organic acids

In this work, vanadium (V) was selectively extracted from fuel-oil fly ash using a leaching process utilizing organic acids extracted from lemon juice with assistance from ultrasound and H2O2. Response Surface Methodology (RSM) was used to optimize the main operating factors. The V recovery was 88.7% at the optimal conditions: 27.9% (v/v) lemon juice, 10% (v/v) hydrogen peroxide (H2O2), solid/liquid (S/L) ratio 0.01% (w/v), ultrasound power 159 W at 20 kHz in 2 h, and initial temperature of 35 °C. The effect of time on the V recovery was examined. The maximum recovery was 100% after 3 h. Furthermore, the individual effects of ultrasound and H2O2 on V recovery were studied, and the results showed that without H2O2 and ultrasound, the V recovery decreased greatly, indicating that both factors were essential in the leaching process. According to the modified shrinking core model, test results indicated that mass diffusion was the controlling step of the overall reaction kinetics. The activation energy of the leaching reaction in the temperature range 25 to 65 °C was found to be 17.1 kJ mol−1.


Introduction
Every year large volumes of bottom and y ashes are generated around the world by burning fuel oil in thermal power plants. 1 These ashes contain valuable metals such as vanadium (V), nickel (Ni), copper (Cu), zinc (Zn), cobalt (Co), etc., which have many industrial applications. 1,2 On the other hand, if there is no treatment of y ashes, they would create many environmental problems, including soil, water and air pollution. [2][3][4] Therefore, the extraction of metals from produced y ashes is very desirable from both environmental and economic aspects. 1 Different chemical and physical processes including pyrometallurgy and hydrometallurgy were used for the extraction of valuable metals from ores and wastes. However, they consume too much energy or create environmental problems. 5,6 Nowadays, bioleaching is an attractive alternative for metal extraction. Bioleaching relies on microorganisms such as Acidithiobacillus thiooxidans, Acidithiobacillus ferrooxidans and Leptospirillum ferrooxidans, Aspergillus niger. [7][8][9] The metal extraction mechanisms are based on the production of leaching agents such as ferric ion, sulfuric acid, and different organic acids including citric acid, maleic acid and ascorbic acid are reported in the literature. 8,10,11 These organic acids dissolve metals by the displacement of metal ions from the solid matrix with hydrogen ions or by the formation of soluble metal complexes and chelates. 10 Although this method is safe and environmentally friendly, its use for industrial and very toxic feeds is limited. Furthermore, the slow kinetics due to slow microbial growth and nutrient diffusion limitation are bottlenecks in bioleaching. 12 An alternative way to get organic acids is to use organic acids from juices of fruits such as lemon, lime, etc., which are rich in citric acid, ascorbic acid and malic acid. These biogenic acids are very effective in the extraction of metals. 13,14 Using acids extracted from agricultural products are not only environmentally friendly but also good at resolving slow kinetics of bioleaching by microorganisms. [15][16][17] Moreover, using lemon juice as a renewable source of organic acids, the cost of leaching wastewater treatment is eliminated.
To assist the effect of acid in the leaching process, different methods including ultrasound, H 2 O 2 and microwave have been used. 18,19 Ultrasound relies on the cavitation of microbubbles which have a strong effect on a solid surface. It can increase the kinetics and recovery of metal leaching processes. In an ultrasound-assisted bioleaching process for the recovery of Zn, Cu, and Co from black shale, one group of researchers achieved optimal leaching efficiencies of 92% Cu and 87% Zn and 71% Co, respectively, at a sonication time of 7 min with 15 d of pregrowth and 36 d of bioleaching. 20 In the bioleaching of printed circuit boards (PCBs), Huang et al. reported that using 300 W ultrasound waves, 93.8% Cu was recovered compared to 90.7% without ultrasound. 21 In the bioleaching of Ni from Sukinda laterite, ultrasound reduced the leaching time from 20 d to 14 d. 22 Ultrasound also increased the removal efficiency of Ni from lateritic nickel ore considerably. 23 There are several published reports about the extraction of valuable metals from power plant ashes using bioleaching in the literature. 1,2,7 Leaching using a fruit juice has not been reported before. This work used organic acids from inexpensive lemon juice to extract V in the y ash from a fuel-oil power plant. Ultrasound and H 2 O 2 were used to assist the leaching process. Response Surface Methodology (RSM) was applied to optimize the operating factors of the process. The kinetic and thermodynamic parameters of the process were also investigated.

Preparation of the y ash
A fuel-oil y ash sample, labeled as power plant residue (PPR), was collected from the Neka Thermal Power Plant, Mazandaran, Iran. It was crushed and screened through 75 mm sieves. The raw PPR was washed with hexane and acetone (1 : 1 mass ratio), and the slurry was mixed with a magnetic stirrer for 120 min at 50 C. Then, a ne powder was obtained by ltering the slurry through a 0.40 mm Whatman lter paper. The powder was dried at 70 C for 0.5 h before use. Aqua regia (hydrochloric acid and nitric acid with a molar ratio of 3 : 1) was prepared for the complete digestion of PPR in 24 h at 50 C. Inductively coupled plasma optical emission spectrometry (ICP-OES) (Vista Pro, Varian, Inc., California, USA) was used to analyze chemical elements in the solid samples. The chemical elements in the raw PPR are shown in Table 1. Fig. 1 shows the X-ray diffraction (XRD) (X'Pert MPD, Philips, Netherlands) patterns and it was determined that the initial PPR sample was mostly Fe 2 O 3 , NaV 6 O 15 and NiV 2 O 6 .

Preparation of lemon juice
Lemon was procured from a local market in Sanandaj, Iran. A kitchen juicer was used to extract lemon juice. In order to remove the pulp, the raw juice was centrifuged for 15 min at a speed of 3000 rpm. High performance liquid chromatography (HPLC) (LC20AD, Shimadzu, Japan) was used to measure the concentrations of organic acids in the lemon juice. The concentration of citric acid in lemon juice was 90 mg g À1 (juice), whereas the concentration of malic acid and ascorbic acid were 0.86 mg g À1 and 1.24 mg g À1 , respectively.

Analytical methods
A portable pH and Eh meter (Metrohm, Switzerland) was used to measure pH and Eh. To determine the concentrations of the metal ions in the leaching solutions, ICP-OES was used. The elemental composition in the V precipitate was determined using X-ray uorescence (XRF) (PW1410, Philips, Netherlands). Surface characteristics of the PPR ash before and aer the leaching process were determined using a eld emission scanning electron microscope (FE-SEM) (TSCAN, Czech Republic).

Experimental design
Design Expert 7.0.0 soware was adopted to optimize the leaching parameters. RSM was applied to design the tests. Response Surface Methodology (RSM) is a statistical method used to improve, analyze, and optimize processes. 2 Central Composite Design (CCD) was applied for studying the effects of four main factors, namely ultrasound power, lemon juice percentage, H 2 O 2 percentage, and solid/liquid (S/L) ratio on V recovery. Table 2 shows the coded values of the variables. A CCD analysis using the four factors resulted in 30 runs which are listed in Table 3. In this study, the percent of V extracted was considered the experimental response. Analysis of variance (ANOVA) was used to determine the signicant models and parameters. The behavior of the system was interpreted using the empirical polynomial model described by the following equation where Y is the model response, b 0 the model constant. b i , b ii and b ij are linear, quadratic and interaction coefficients, respectively. X i represents independent factors, and 3 is error. Finally, the obtained models were validated by comparing the experimental results carried out at optimal conditions with the model predicted results.

Leaching experiment
According to the test conditions in Table 2, certain amounts of the PPR ash, lemon juice and H 2 O 2 were put into a 150 mL glass beaker. An ultrasonic probe (UF55/UN55, Memmert, Germany) was inserted in the solution during leaching. The ultrasound generator was set at varying power in the range from 50 W to 400 W at xed 20 kHz frequency. Aer the leaching was terminated, the spent solid was removed from the liquor using centrifugation for 15 min at a speed of 3000 rpm. The ICP-OES analysis was used to determine the concentrations of metals in the liquor. The V recovery was calculated using the following equation: where R is leaching efficiency or recovery of the metal (%), C 1 and C 2 are the metal concentrations (ppm by volume) in aqua regia and the liquor aer leaching, respectively, M 1 and V 1 are the solid mass and aqua regia volume, respectively, and M 2 and V 2 are the solid mass and liquor volume in the bioacid leaching liquor, respectively. All these leaching tests, except in the thermodynamic study, lasted 2 h with an initial temperature of 35 C. Due to ultrasound energy input, the leaching solution temperature typically rose by 20 C aer 2 h. The variations of Eh and pH during the leaching process were monitored.

V precipitation from liquor
In order to precipitate V selectively from the leaching liquor aer the leaching process, the impurities in the liquor were removed rst by increasing the liquor pH to 9-10 using 1 M sodium carbonate (Na 2 CO 3 ), and then adding 0.1% (w/v) of CaCl 2 . Aerwards, the solution was stirred vigorously for 30 min at 80 C before the solution was ltered using a 0.42 mm Whatman lter paper to precipitate other metal ions from the solution. In the next step, NH 4 Cl was added to the solution to precipitate V in the liquor as ammonium metavanadate (NH 4 VO 3 ) while the solution was stirred for 1 h. The obtained NH 4 VO 3 precipitate was separated from the solution using a 0.42 mm Whatman lter paper and then calcinated at 550 C for 4 h in a muffle furnace to obtain V 2 O 5 . 24

Statistical analysis
The ANOVA outcome of the data and the results are shown in Table 4. A reduced cubic model was used for the V extraction. The correlation between V recovery and the four dened factors is shown as follows: V recovery (%) ¼ 43.65 À 9.71A + 4.16B + 7.06C + 12.54D À 1.60AB + 1.11AD + 0.78BD + 4.57A 2 + 1.35B 2 À 3.52C 2 + 1.95ABC À 1.01ACD + 1.48BCD À 11.28A 2 D + 5.70AB 2 À 10.27B 2 C where A, B, C, and D are S/L ratio, ultrasound power, lemon juice% (v/v), and H 2 O 2 % (v/v), respectively. The relatively high R-squared and adjusted R-squared values in Table 4 indicated that the model was a good t with experimental results, thus validating the suitability of the model. Moreover, results showed that at a 95% condence level, the calculated p-value of the model was less than 0.05 (Table 4), indicating that the model was statistically signicant. The adequate precision, which indicated the signal to noise (S/N) ratio, was 15.4, much greater than the commonly accepted threshold of 4. Fig. 2 shows two-dimensional graphs for V extraction. These graphs show the interactions between different parameters on the V recovery. Fig. 2a shows the interaction between S/L ratio and ultrasound power at constant 20% H 2 O 2 and 60% lemon juice. According to Fig. 2a, a maximum of 75.1% V extraction was obtained at 0.9% S/L ratio and 140 W ultrasound power. Decreasing S/L ratio from 1.5 to 0.9% and increasing ultrasound power from 87 to 143 W led to increased V recovery. Decreasing S/L ratio resulted in decreased viscosity of the mixture and diffusion mass transfer resistance, thus promoting acid penetration into the solid particles. This led to increased V recovery. 18 The positive effect of increased ultrasound power was due to enhanced cavitation, leading to the breaking of chemical bonds. 25 Fig. 2b shows the interaction between H 2 O 2 and ultrasound power at constant 50.3% lemon juice and 0.75% S/L ratio. According to Fig. 3b, the V extraction was enhanced when ultrasound power and H 2 O 2 concentration increased from 87 W to 162 W and 10% to 20%, respectively. Increasing both ultrasound power and H 2 O 2 percentage led to increased solution potential (Eh) which resulted in enhanced leaching efficiency. The positive effect of ultrasound in the presence of H 2 O 2 was due to the breaking of the chemical bond between vanadium and oxygen. 26,27 The following equation shows the reaction between NiV 2 O 6 and citric acid in the presence of H 2 O 2 : This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 1685-1696 | 1689 Fig. 2c shows the interaction between lemon juice concentration and S/L ratio when ultrasound power and H 2 O 2 percentage were xed at 117 W and 20%, respectively. It shows that the V recovery increased from 57% to 71% when lemon juice concentration increased from 20% to 60% at constant 0.9% S/L ratio. The concentration of lemon juice (containing three acids, namely citric acid, malic acid and ascorbic acid) increased the availability of H + ions and acted as the oxidant during the leaching process. The dissociation of H + from citric acid was shown in eqn (5) HCit 2À / Cit 3À + H + pK a 3 ¼ 6.4

Contour plots
Similarly, the dissociation of H + from malic acid are expressed in eqn (8) and (9) Another positive effect of the lemon juice was due to the carboxyl and hydroxyl groups of in the organic acids in the lemon juice. Aer losing a proton, they became chelating agents, which formed a complex with a metal as shown in eqn (10), where M is a metal in the solid matrix and R-(COOH) m organic acid. 28,29 Ash-M + R-(COOH) m / Ash + R-M(COOH) m (10)

Process optimization and model validation
The goal of optimization was to achieve the highest V extraction. At the optimum conditions, the maximum V extraction was 88.7%. The optimal conditions were 28.0% lemon juice, 10.0% H 2 O 2 at 0.01% S/L ratio, and ultrasound power of 159 W. In order to verify the suitability of the model, a test was conducted at the optimal conditions predicted by the model. Results indicated that the experimental optimal V recovery (88.7%) was not far from the model prediction (79.7%). The experimental result fell within the 95% condence interval (C.I.) of the model value.

The roles of H 2 O 2 and ultrasound
The control tests to understand the V recovery change by skipping a leaching agent (ultrasound and H 2 O 2 ) were performed. Results indicated that in the presence of ultrasound and H 2 O 2 under optimal conditions, the V recovery was 88.7%. However, without ultrasound the metal extraction decreased to 37.5%. It is known that ultrasound waves can increase metal recovery by creating cavitation and bubbles in the liquid. The bubbles explode with increasing pressure. As a result of the energy released by the bursting of bubbles at the particle surface, the metal recovery is increased. 30,31 Moreover, Run 4 in Table 4 was done without H 2 O 2 and result showed the V recovery decreased greatly to 8.2%, indicating that H 2 O 2 was also essential in achieving good V recovery.

Particle morphology analysis
In order to show the effects of ultrasound and lemon juice acids, FESEM was used to analyze the particle surfaces of the raw PPR ash (Fig. 3a), the residue aer leaching without ultrasound (Fig. 3b) and the residue aer leaching with ultrasound (Fig. 3c). By comparing Fig. 4a with Fig. 3(b and c), it can be seen that leaching reduced particle sizes. A comparison of Fig. 3b with Fig. 3c also suggests that ultrasound eroded the particle surface considerably which resulted in increased penetration of lemon juice acids into internal particle surfaces.  washing in the presence of ultrasound, and (c) with initial washing and in the presence of ultrasound.

Eh and pH variations
3.6.1. Eh variation. Fig. 4a shows the changes of Eh with time at different conditions tested above. In the rst case without the initial PPR ash washing in the absence of ultrasound, the Eh of the PPR suspension remained almost constant at about 464 mV until 110 min due to the presence of an impermeable oily layer on the dirty particles. Aer 110 min, Eh decreased from 467 mV to 48 mV due to the elimination of the oily layer as the diffusion barrier and the H 2 O 2 reacted with the internal surfaces of the particles.
However, in the second case without the initial PPR ash washing in the presence of ultrasound, due to the existence of the impermeable oily layer around the solid particles, the hydroxyl radicals of H 2 O 2 in the solution were not consumed and then the amount of Eh remained almost constant at about 533 mV for the rst 80 min. The initial potential for a constant Eh range was numerically larger than in the previous case (without ultrasound). This was due to the production of hydroxyl radicals induced by ultrasound. 32 Aer 80 min, due to the breakdown of the diffusion layer and the consumption of hydroxyl radicals, Eh decreased to 140 mV. In this case, the breaking of the oily layer occurred 30 min earlier than the previous case because ultrasound accelerated the breaking.
In the third case with washing and ultrasound, the initial washing removed the oily layer which resulted in consumption of hydroxyl radicals at the beginning of the process and at the same time ultrasound produced hydroxyl radicals. Thus, the consumption and production of hydroxyl radical were simultaneous from the beginning, leading to almost constant Eh without a sharp drop.
3.6.2. pH variation. Fig. 4b indicates the variations of pH during time in different cases. In the rst case without the initial PPR ash washing in the absence of ultrasound, pH decreased till 30 min, which was due to the release of H + from acids in the lemon juice according eqn (5)- (9). Then, until about 60 min, the produced H + ions were consumed to remove the oily layer, causing pH to increase. From 60 to 100 min, pH decreased again due to the degradation of the oily layer as well as the release of H + from more acid molecules.
In the second case without the initial PPR ash washing in the presence of ultrasound, the pH level was lower than in the previous case, which was the result of ultrasound dissolution citric, malic and ascorbic acids that led to pH reduction early on. The pH had fewer uctuations than in the rst case, which was due to the ultrasound action that destroyed the oily layer, as well as its acceleration of the production and consumption of H + .
In the third case with washing and ultrasound, the pH uctuations are much less pronounced than the two previous cases. This was the result of the fast destruction of the oily layer and the fact that the production and consumption of H + were almost simultaneous.

Kinetic study
Experimental results showed that with increasing time, the V recovery at the optimal conditions predicted from RSM increased and the maximum was 100% aer 3 h.
The kinetic parameters of the process were analyzed using the shrinking core model which represents the speed of reactions in penetrating the solid network and conducting a chemical reaction in the particle surface layer. Based on the model, the diffusion-control and the reactioncontrol scenarios are presented in eqn (11) and (12), respectively: 33,34 where X is leaching efficiency (%), K r reaction-controlled rate constant (min À1 ), K d diffusion-controlled reaction rate constant (min À1 ), and t leaching time (min). Fig. 5 shows the plots of eqn (10) and (11). The obtained R 2 values indicated that the data were tted with the diffusion-controlled model slightly better than the reaction-controlled model.

Thermodynamic modeling
The rate equation for the reaction in eqn (4) can be expressed as follow: where in which a-g values are reaction orders. Using the denition of k based on the Arrhenius equation (eqn (15)), Eqn (13) can be rearranged to give eqn (16): where R is universal gas constant (8.314 J mol À1 K À1 ), E a activation energy of reaction and T absolute temperature. The linearized equation of eqn (16) is eqn (17) By plotting ln(r) vs. 1/T (Fig. 6), the activation energy was estimated to be 17.1 kJ mol À1 which is less than 25 kJ mol À1 . This suggests diffusion control in the leaching process.

Separation and purication of V
Aer the bioacid leaching, many ions such as Fe, Ni, and Cu existed in the solution. These impurities had to be removed before V precipitation. It was found the adding CaCl 2 to the solution at pH 9-10 could remove other metal ions with a high efficiency. V ion in this pH range is soluble. NH 4 Cl was added to precipitate V as shown in eqn (18). Finally, ammonium metavanadate (NH 4 VO 3 ) was formed with a low solubility that led to a precipitate as the product. 2 NaVO 3 + NH 4 Cl / NH 4 VO 3 Y + NaCl Fig. 7 shows a summary of the overall leaching-purication process. XRF analysis was applied to determine the composition of the precipitate and the results in Table 5 indicate that V was selectively precipitated with a high purity.

Conclusion
The leaching of V from a PPR ash was carried out using bioacids from lemon juice with assistance of ultrasound and H 2 O 2 . The recovery of V was optimized using RSM. The optimal leaching conditions were found to be 27.9% lemon juice, 10% H 2 O 2 , S/L ratio of 0.01% and ultrasound power of 159 W which resulted in 88.7% V extraction in 2 h with initial temperature 35 C experimentally. Additionally, the effects of skipping one of the two assisting leaching agents (H 2 O 2 and ultrasound) were assessed. It was found that both were essential for good V recovery. The reaction kinetics analysis results showed that the diffusion-control model was favored over the overall reaction-control model. The activation energy of the reaction was found to be 17.1 kJ mol À1 in the temperature range 25 to 65 C. Finally, V was puried from the leaching solution by precipitation as ammonium metavanadate with a high purity.

Conflicts of interest
There are no conicts to declare.