Accepted Manuscript Factorial experimental design intended for the optimization of the alumina purification conditions Mounaouer Brahmi, Mohamedou Ba, Yassine Hidri, Abdennaceur Hassen PII:
S0022-2860(17)31691-5
DOI:
10.1016/j.molstruc.2017.12.071
Reference:
MOLSTR 24690
To appear in:
Journal of Molecular Structure
Received Date: 10 July 2017 Revised Date:
18 December 2017
Accepted Date: 19 December 2017
Please cite this article as: M. Brahmi, M. Ba, Y. Hidri, A. Hassen, Factorial experimental design intended for the optimization of the alumina purification conditions, Journal of Molecular Structure (2018), doi: 10.1016/j.molstruc.2017.12.071. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 0.002 P2O5 (%) =
Pt S1
Fe2O3 (%) =
(Greece)
S2 (Spain)
SiO2 (%) =
S3 (Germany)
CaO (%) =
0.0 0.002 0.009 0.01 0.009 0.001 0.0 0.006 0.025 0.0 0.064
0.306
Na2O(%) = 0.323 At C a+XRF
alu - T (%) =
17.3 03.9 15.3
st – T (%) =
16.3 04.4
0.0 0.002 0.0 0.01 0.009 0.009 0.0 0.003 0.0 0.064 0.025 0.009
M AN U
P2O5 (%) =
17.6 04.4 15.5
SC
0.259
alu - st (%) = FED
RI PT
C a+XRF
15.8
Fe2O3 (%) =
TE D
SiO2 (%) =
CaO (%) =
AC C
EP
Na2O (%)=
0.264 0.306 0.242
C a: Chemical analysis and X-ray fluorescence analysis (XRF); P t: Prior treatment; F E D: Optimization of the Alumina Purification Conditions by Factorial Experimental Design; At: After treatment; alu – st: quantity of alumina - stirring time; alu – T: quantity of alumina – temperature; st– T: stirring times - temperature
ACCEPTED MANUSCRIPT
Factorial Experimental Design intended for the Optimization of the Alumina Purification Conditions
*1
2,
1
Mounaouer BRAHMI , Mohamedou BA Yassine HIDRI and Abdennaceur HASSEN
1
1
Laboratoire de Traitement et de valorisation des rejets hydriques, Centre des Recherches et des Technologies
des Eaux (CERTE), Technopole Borj-Cédria, P.O. Box 273, 8020 Soliman, Tunisia. Département de Chimie, Faculté des Sciences de Gabes, Cité Erriadh, 6072 Z rig, Gabes, Tunisie.
RI PT
2
*Corresponding author: Tel.: +216; Fax: +216 79 325 802, E-mail:
[email protected] Running title: Optimization of the Alumina Purification
Abstract
SC
The objective of this study was to determine the optimal conditions by using the experimental design
methodology for the removal of some impurities associated with the alumina. So, three alumina qualities of different origins were investigated under the same conditions. The application of full-factorial designs on the samples of different qualities of alumina has followed the removal rates of the sodium oxide. However, a
M AN U
factorial experimental design was developed to describe the elimination of sodium oxide associated with the alumina. The experimental results showed that chemical analyze followed by XRF prior treatment of the samples, provided a primary idea concerning these prevailing impurities. Therefore, it appeared that the sodium oxide constituted the largest amount among all impurities. After the application of experimental design, analysis of the effectors different factors and their interactions showed that to have a better result, we should reduce the alumina quantity investigated and by against increase the stirring time for the first two samples, whereas, it was necessary to increase the alumina quantity in the case of the third sample. To expand and improve this research, we should take into account all existing impurities, since we found during this investigation that the levels of partial impurities increased after the treatment.
AC C
EP
TE D
Keywords: Experimental design; Alumina; Optimization; Chemical analyzes.
1
ACCEPTED MANUSCRIPT
1. Introduction
Aluminum oxide (Al2O3, alumina, corundum) is the most widely used inorganic chemicals produced from the mineral bauxite using the Bayer process [1]. Bauxite is a mixture of hydrated aluminum oxide (AlO (OH)) with iron oxide (Fe2O3), silica (SiO2), and titanium (TiO2) impurities [2]. It results from the decay and weathering of aluminous rocks, often igneous, under tropical conditions [3]. Like kaolin, bauxite occurs as both primary and secondary deposits. The Bayer process produces a nominal 99.5% Al2O3 product [4]. The alumina could be prepared in a range of grades to suit specific applications [5]. The grades differ by the size and shape of the crystals and the impurity content. The dominant impurity, accounting for up to 0.5%, is Na2O. The crystal size could be adjusted to measure among 0.1 and 25 μm [6]. The Bayer process is the principal industrial technique to refine bauxite to pure alumina. There have been
RI PT
limited changes to the basic process since the first plant opened in 1893 [7]. In this process, bauxite containing typically 10-30 wt % Fe2O3, 4-8 wt % SiO2, and 2-5 wt % TiO2 as major impurities [8] is dried, ground, digested in sodium hydroxide under pressure. Impurities are separated by filtration, before an alumina hydroxide is
precipitated and calcined to produce commercial alumina. The relatively high pure alumina produced from the Bayer process is then transferred to the Hall-Heroult electrochemical cell for electrolysis [9]. Natural alumina
has frequently been present as hydrated alumina rocks derived through the metamorphism of alumina-silicate minerals [10]. Thereby, it could exist in both as monohydrate (Al-O-OH) and as trihydrate [Al (OH)3] forms [11].
SC
Furthermore, each hydrated aluminous mineral exhibits polymorphism. The monohydrate has two polymorphic forms as boehmite (ƴ All-O-OH) and diaspore (α All-O-OH), while trihydrate as gibbsite [ƴ Al (OH) 3], bayeriten [α Al (OH) 3] and nordstrondite [Al (OH) 3] [10]. A Crystal system of the monohydrate is orthorhombic, while that of trihydrate is monoclinic. Water molecules play a critical part in their crystal structure [12].
M AN U
Therefore, the most important hydrated aluminous rock, rich in alumina, is bauxite that is not a pure mineral, but consists of two or more hydrated mineral phases such as boehmite, bayerite, diaspore and tohdite [11]. This hydrated alumina is often associated with some specific mineral impurities such as quartz (SiO2), goethite (Fe-O-OH), hematite (Fe2O3), anatase (TiO2), and alkali and alkaline earth oxides [12]. Alumina (Al2O3) is the major component of all aluminous raw materials with a very high melting point, remarkable thermomechanical, and satisfactory electrical and excellent chemical durability properties [13]. Bauxite is an essential high alumina bearing raw material for aluminum metallurgy apart from physical and technological properties of bauxite rock that are largely dependent on the mineralogical composition, particle morphology and associated impurities [14]. Bauxite mineral is the major source of Al2O3 as well as red mud (by-product from the processing
TE D
of bauxite) [15]. In addition, Bauxite is found involved in numerous applications such as the wastewater treatment, the phenols, the heavy metals, the nitrates, the phosphates, and the dye color removal from some raw or natural materials, the manufacture of various ceramic products and pozzolana pigments [16]. The gradual expulsion of water molecules occurs upon thermal treatment of bauxite, and other hydrated alumina minerals leading to the intermediate formation of several metastable transition phases: Chi (χ), Eta (ɳ), Gamma (ƴ), Delta (Δ), Iota (Շ) Theta (ө), Kappa (қ) and Beta, etc., and with progressive variation in
EP
crystallographic nature, specific surface area and grain morphology until the best and the maximum stable
AC C
form of alumina, corundum (α -Al2O3) [13, 17].
2
ACCEPTED MANUSCRIPT
Dehydration involves the loss of physically adsorbed water, interlayer water, followed by the loss of hydroxyl groups from the matrix in the form of water. However, at much higher temperature (> 1000°C), there is no loss in mass. In the endothermic solid-state reaction of aluminous minerals, the dehydration reaction or transformation occurs, concurrently causing complications of the kinetics pathway [18]. Nevertheless, this reaction provides useful information about the mechanism of the interface reaction and the nucleation process. Thus, the interesting subject of the present research work includes the study of the dehydration behavior of aluminous and alumina-silicate minerals while considering that the heat of dehydration represents the most important thermodynamic parameter for the thermal dehydration reaction [19]. Fundamental study of the kinetics of thermal dehydration of alumina bearing minerals revealed that the
RI PT
magnitude of orientation of water molecules is greatly influenced by several factors such as exchangeable ions, mineralogical composition, crystallographic structure, particle size and shape, nature and content of impurities,
and filling density [20]. Therefore, there are wide scopes of research in this field for predicting and expecting
the reaction progress, revealing the mechanism of thermal dehydration and actual energy consumption; by that solid-state dehydration of aluminous minerals may be optimized and controlled in an efficient way [15].
Many investigators extensively studied the non-isothermal transformation kinetics of gibbsite, boehmite and
bayerite to the ultimate corundum phase through series of transition alumina [21-25]. There are several
SC
procedures that have been frequently used for the impurity determination from the alumina precursor, such as
X-ray fluorescence spectrometry [25], secondary ions mass spectrometry (SIMS), spark sources mass spectrometry, ICP-AES [7, 8, 13, 17]. Nevertheless, this last mentioned methods are altogether recognized as high costs techniques, as well it is very difficult to determine separately any eventual single element [26].
M AN U
In this study, we tried to present a new and simple method for the impurity removal from aluminum oxide by using X-ray fluorescence spectrometry at its optimal condition. Thus, we introduced the contribution of the methodology of experimental research (design of experiments) leading to the optimization of the conditions of sodium oxide removal. In a full-factorial experiment, both of the (−1) and (+1) levels of every factor are compared with each other and the effects of each of the factor levels on the response are investigated according to the levels of other factors.
2. Material and methods 2.1. Alumina sample collection
TE D
As previously said, this work is dedicated to the optimization of the conditions for removal of sodium oxide from alumina by using the experimental design method (full-factorial design), and we intended to highlight the influence of definite factors on the alumina purification that it has numerous applications in various fields. So, three samples of alumina of different origins are used in this study. The first is from Greece, the second from Spain and the third from Germany. All of these Alumina samples are under bauxite products that have undergone physicochemical treatment, according to the Bayer process. These samples are referenced as S1, S2,
EP
and S3, respectively.
AC C
2.2. Available alumina determination
3
ACCEPTED MANUSCRIPT
The method for the determination of available alumina (the amount that will be refined to obtain Al2O3 in the Bayer Process), involved bauxite digestion in alkaline medium (NaOH) (2.5 mL of soda solution of about 2 moles/L) under controlled pressure and temperature, by simulating the Bayer process. So, for the determination of available aluminum, a sodium gluconate solution was added to the supernatant to form an aluminum hydroxide gluconate complex. The excess of NaOH, used in the digestion step, was neutralized with the addition of HCl solution (1 mL of hydrochloric acid of 3 moles/L and with the pH between 4 and 9). Therefore, a KF solution was added and back titration was carried away. Afterwards, an excess of standardized HCl solution was titrated with a standardized NaOH solution (NIST SRM 723e; laboratory of Merck KGaA, Darmstadt, Germany, in accordance to DIN EN ISO/IEC 17025). The solid phase, resulting from the stage of the
RI PT
ore’s digestion, was dissolved in a HNO3 solution [27].
2.3. Experimental procedure 2.3.1. Chemical and mineralogical characterization
The alumina samples of various roots were submitted to chemical and mineralogical analysis with the use of: Xray fluorescence, T-60 UV-VIS spectrophotometer, and thermal analysis. During the mineral dressing tests, these samples were subjected to the same characterization techniques of the crude sample.
SC
2.3.1.1. X-ray Fluorescence (XRF)
X-ray fluorescence analysis (XRF) is one of the most significant emission methods that enables a quick and multi-elemental analysis in a very short period, and requires a minimal preparation of the sample [28]. In case, some material is showered with primary high-energy X-rays, it will cause the expulsion of electrons from some
M AN U
of the inner shells (K, L, M) of atoms in that material, which results in the formation of electron holes in one or more atomic orbital closer to the nucleus, by that appropriate atoms get into an excited state [29]. The atom excited in such a way tends to get into a stable state, and therefore the holes in orbits closer to the nucleus are filled up with the electrons from higher orbitals [28]. This transfer is accompanied by the emission of energy in the form of a secondary, i.e. fluorescent X-ray that is characteristic of the given atom [29]. The substance of the X-ray fluorescent analysis measures the intensity of the developed secondary fluorescence radiation. The method is applicable in a wide scope of concentrations. The samples were fused with lithium tetraborate at 1100ºC, in the proportion of 1:6 samples/fluxing agent. The melted bead was analyzed in the X-ray fluorescence spectrometer by dispersive energy (Bruker-AXS model S4-Explorer), equipped with a Rh-anode X-
TE D
ray tube [30]. To obtain the semi quantitative chemical analysis, the sample spectrum was evaluated by the Spectra plus v.1.6 software.
The sample is mixed with a binder, and then crushed to make it a pellet. The measuring of the emission intensities of ray elements was done by means of the X-ray fluorescence spectrometer (Xian Zealchon Electronic Technology Co., Ltd., Zealcho, China), and the conversion of these intensities to concentrations by comparison of the measured intensities with those of the standard transmission.
EP
Analytical measurements of seven elements (Si, AI, Ca, Mg, Fe, Na, and P) were achieved employing X-ray
AC C
Fluorescence (XRF), using a Philips PW 1400 wavelength dispersive spectrometer (Table 1).
4
ACCEPTED MANUSCRIPT
2.3.1.2. T-60 UV-VIS spectrophotometer
The determination of the metal ion impurity content was carried out by the T-60 UV-VIS spectrophotometer (Labindia Instruments Pvt Ltd, India) equipped with 1.0 cm quartz cells and the pH of the solution was controlled with a Thermo, Orion 920A + digital pH meter. The results have been checked by the Perkin-Elmer Optima 5300 DV inductively coupled plasma optical emission spectrometer (Perkin-Elmer, Norwalk, USA). The concentration of an analyte in solution can be determined by the absorbance measure at some wavelength, and the application of the Beer-Lambert Law (Perkin-Elmer, Norwalk, USA).
2.4. Estimation of Fe2O3, Al2O3, TiO2, and Na2O An accurate portion of samples of 0.100 g was weighed and thoroughly mixed with 0.8 g of the fusing agent,
RI PT
namely, oxalic acid, lithium carbonate, lithium tetraborate (1:1:1), in a platinum crucible, which was then covered and placed in a muffle furnace at 925°C for 10 minutes. The crucible lid was then transferred to a 400
mL beaker containing 100 mL of boiling distilled water and 10 mL of concentrated hydrochloric acid. The crucible was cooled by washing its exterior walls with a water flow. It was then put in the same beaker and
covered with a watch glass. The solution was heated until the fusion cake was dissolved, filtered if necessary, and then diluted to 500 mL in a calibrated flask.
This sample solution was reserved for SiO2, Fe2O3, TiO2, Na2O, K2O and Al2O3 determinations. For the
SC
preparation of standards, 0.100 g of a fly ash standard (e.g. ASCRM 010) was mixed with 0.8 g of the fusion
agent, and treated similarly as for the sample, and the solution was diluted to 250 mL in a calibrated flask, and labeled as the stock solution. Standard solutions were prepared by pipetting 25, 35, 50, and 75 mL of the stock solution, separately in a series of 100 mL calibrated flasks, and diluting to the marks with distilled water. These
2.5. Determination of the silica content 2.5.1. Mixed reagents
M AN U
standard solutions were used for the determinations of SiO2, Fe2O3, TiO2, Na2O, K2O, and Al2O3.
Mixed reagents were prepared as indicated follow: The aluminum molybdate solution was prepared by dissolving 10 g of (NH4)
6
MO7O24. 7H2O in 100 mL of water. The ferrous aluminum sulfate solution was
prepared by dissolving 40 g of Fe2SO4 (NH4)2SO4 .6H2O in 100 mL of hot water containing 1 mL of 25% sulfuric acid allowed to cool before use and made a new solution for the use. The Silicon standard solution of 0.01 mg/mL Si was prepared by the use of 10 mL of the standard solution of 1 mg/mL of Si, and then diluted in a
TE D
1000 mL of volumetric flask.
2.5.2. Sample preparation
A mixture was prepared as indicated follow: 1 g of samples, 2 g of sodium tetraborate and 4 g of sodium carbonate are put in a platinum crucible, and then placed in the furnace at 500°C. Next, the temperature was gradually raised to 1000°C for 10 to 30 minutes. When the mix content is melted by stirring the crucible and by
EP
heating for 15 minutes, the crucible was removed and cooled, then filled with water, and gently heated until
AC C
the mix is completely dissolved. Afterward, the mix solution is put in a polyethylene beaker of 250 mL and
5
ACCEPTED MANUSCRIPT
flushed by mechanical stirring and continuous adding of 10 to 20 drops of sulfuric acid 50%. Finally, the mix will be cooled at ambient temperature and analyzed.
2.5.3. Preparation of standard calibration curve By weighing and mixing 2 g of sodium tetraborate and 4 g of sodium carbonate as previously described for the sample fusion, we dissolve the melt transferred into 250 mL of polyethylene beaker. Then, we acidify the bulk until pH = 1.6 and dilute it to the mark.
2.5.4. Standard solutions By using a pipette, we introduce within 1000 mL polyethylene beakers, a 250 mL sample volume and 25 mL of
RI PT
blank solution for the silica determination. We dilute these quantities to a volume nearly 60 mL and adjusted to an exact pH of 1.4 using sulfuric acid 25%. Then, we add 5 mL of ammonium molybdate solution, and we mix
all. After 10 min, we will add 10 mL of 50% sulfuric acid and 5 mL of ferrous ammonium sulfate solutions and we mix all once again. Next, we dilute to an exact volume of 100 mL, and we transfer this mix in the absorption cells of 10 mm and set the spectrophotometer at a wavelength of 608 nm.
mg/mL of Si in 25 mL of the blank solutions. The percentage of SiO2 is given by the following formula:
% SiO2 =
3 DO 480.7 P
DO: Optical density; P: Sample weight (g).
2.6. Determination of the P2O5 content 2.6.1. Stock standard solution
(1)
M AN U
Where,
SC
The calibration curve is obtained by adding respectively, 0.5; 1; 1.5 and 2 mL of the standard solution of 0.01
The mass of 1.9412 g of dry KH2PO4 reagent grade is dissolved in deionized water in a 1-liter volumetric flask. The phosphorus content of this solution expressed as P2O5 is 1000 ppm.
2.6.2. Operational standard solution
From the stock standard solution, a volume of 2.5 mL is sampled and placed into a 50 mL volumetric flask and
TE D
then diluted to mark. From this last solution, we will take a volume of 4 mL and put it into a 50 mL volumetric flask that will be filled to the mark with distilled water. This final resulting solution is of 4 ppm P2O5.
2.6.3. Reducing solution preparation
A mass of 1 g of ammonium molybdate is dissolved in 50 mL of deionized water and 1.76 g of ascorbic acid in 100 mL of deionized water. In parallel, a dilution of 17 mL of concentrated sulfuric acid is made in 200 mL of
EP
deionized water. A solution mix is done into a 250 mL volumetric flask by placing 39 mL of the ammonium molybdate solution, 60 mL of the ascorbic acid solution and at last 125 mL of diluted sulfuric acid. Finally, the
AC C
250 mL volumetric flask is filled with deionized water.
6
ACCEPTED MANUSCRIPT
2.6.4. Sample preparation and measurement
As previously described, samples should be maintained at atmospheric pressure for 24 h before analysis in order to eliminate all gas traces. Then, 0.2 g of the prepared samples is put in a platinum crucible with 0.9 g of sodium carbonate and 0.7 g of boric acid. The mixtures were intimately triturated with a spatula and placed into the furnace with gradually raised temperature to 900°C and for 15 min. After cooling, the remaining residues will be dissolved in 20 mL of HClO4 (5 M) by heating in a sand bath. The recovered volumes will be transferred to a 50 mL volumetric flask and let cooled to room temperature. So, 5 mL of ammonium vanadate, 10 mL of ammonium molybdate were added and completed to a volume of 50 mL with distilled water, shacked goodly, and let coloration develop for 15 minutes. This ready solution will be examined by the
RI PT
spectrophotometer at a wavelength of 400 nm.
2.6.5. Preparation of the different reagents
The perchloric acid (HClO4) 70% was prepared by diluting 214 mL of HClO4 (5 M) with distilled water in a
volumetric flask of 500 mL. As well, the ammonium metavanadate 0.02M solution was prepared by dissolving 1.17 g of ammonium metavanadate in 400 mL of hot distilled water. After cooling, the addition of 17.5 mL of 70% HClO4 and brought to 500 mL with distilled water. The ammonium molybdate 0.2 M was prepared by
dissolving 17.65 g of ammonium molybdate in 500 mL of distilled water. The boric acid, of analytical grades and
SC
anhydrous sodium carbonate of analysis grades were used in this investigation. At last, the standard P2O5 solution of 0.01 mg/mL was prepared by weighting 1.9169 g of KH2PO4 and dissolving in a 1000 mL volumetric
2.6.6. Curve calibration building
M AN U
flask, and then diluted to 1/100 to obtain the solution containing 0.01 mg/mL of P2O5.
The mixture of the following products 4.5 g of Na2CO3, 3.5 g H3BO3, 100 mL of HClO4 (5 M) was put in a 250 mL beaker, and slightly heated on a sand bath for dissolving the H3BO3 and Na2CO3. By the same, we have to prepare a series of four volumetric flasks of 50 mL in which we put in each 20 mL of the above mixture. From the standard P2O5 solution 0.01 mg/mL previously prepared, we put in each flask of 50 mL 0.5; 1; 1.5 and 2 mL, respectively, and then we add in each flak 5 mL of ammonium metavanadate and 10 mL of ammonium molybdate. All these flask preparations will be maintained for 15 min to develop the coloration, and the coloration absorption intensity will be measured at the spectrophotometer at a wavelength of 400 nm
% P2 O5 =
DO 8000 .P
TE D
(Shimadzu Corporation, Japan). The percentage of P2O5 is given by the following formula:
(2)
Where, DO: Optical density; P: Sample weight (g).
2.7. Determination of the Fe2O3 content
EP
2.7.1. Reagents
The hydroxylamine hydrochloride solution was prepared by dissolving 10 g of NH2OH.HCl in 100 mL of distilled
AC C
water. The O-phenanthroline solution was prepared by dissolving 0.25 g in 10 mL of isopropyl alcohol, and
7
ACCEPTED MANUSCRIPT
diluted with distilled water to a volume of 100 mL and stored in the darkness. The Buffer solution of ammonium acetate was prepared by dissolving 231 g of CH3COONH4 in 1000 mL of distilled water. The Iron standard solution 0.01 mg/mL was prepared by dissolving 7.02 g of Fe(NH4)2 (SO4)2.6H2O in about 500 mL water, then added with 10 mL of concentrated HCl and adjusted to the volume of 1000 mL with distilled water. The Iron calibration solution 0.01 mg/mL was prepared by dissolving 7.02 g of Fe(NH4)2(SO4)2.6H2O in about 500 mL of water, adding 10 mL of concentrated HCl and then adjusting to the volume mark of 1000 mL with distilled water. This last Iron calibration solution will be diluted by taking 10 mL of this solution, and then adjusted to the volume mark of 1000 mL with distilled water.
2.7.2. Sample preparation and measurement
RI PT
The mixture of 1 g of samples, 2 g of sodium tetraborate and 4 g of sodium carbonate were done in a platinum crucible and then placed in the oven at 550°C. When the content is melted, the crucible will be agitated and heating will remain for 15 minutes. After, the crucible will be removed and let be cooled at ambient
temperature. The cooled content will be filled with distilled water, gradually heated until the molten mass is completely dissolved. Then this solution will be transferred to a polyethylene beaker of 250 mL with a just rinse of the crucible with 10 to 20 drops of 50% sulfuric acid. This rinse will be added continuously to the liquid in the beaker, and all solution will let cooled and stirred by mechanical means and by a dropwise adding of sulfuric
adjusted to the calibration mark with distilled water.
M AN U
2.7.3. Calibration curve
SC
acid 50% until getting a pH = 1.6. This last preparation will be transferred to a volumetric flask of 250 mL and
The mass of 2 g of sodium tetraborate and 4 g of sodium carbonate will be mixed and dissolved in a 250 mL polyethylene beaker, then acidified to pH = 1.6, and diluted to the calibration mark. By the means of a pipette, we transferred into a series of 100 mL volumetric flasks an amount of 50 mL of samples for the determination of iron and 50 mL of blank solution to determine the blank value. In each 100 mL volumetric flasks, we will add 1 mL of the hydroxylamine hydrochloride solution, and it will be mixed after adding 10 mL of O-phenanthroline solutions and 10 mL of ammonium acetate buffer solution. Each prepared solution will be adjusted to an exact volume of 100 mL, shacked and allowed to stand for an hour until the measure of the optical density at 508 nm and by considering the blank solution in optical zero. The calibration curve is prepared by adding, respectively,
TE D
0; 2; 4; 6 and 8 mL of the standard solution of 0.01 mg Fe/mL to 50 mL. The percentage of iron is determined by the following formula:
% Fe2O3 =
DO 420.7 P
(3)
Where, DO: Optical density; P: Sample weight (g).
EP
The calibration curves of respectively, SiO2 (A), P2O5 (B), and Fe2O3 (C) are shown in Figure 1.
2.8. Experimental statistical design
The statistical optimization technique using the full experimental factorial design is applied to determine the
AC C
best removal conditions of sodium oxide associated with alumina.
8
ACCEPTED MANUSCRIPT 3
For this study, the factorial of the type 2 that consisted of performing 8 experiments was applied. The variables and the levels for the experiment are reported in Table 2. Each factor was studied at both low and high levels. The higher level was designated as (+) and the lower value was designated as (−). The regression equation with three parameters and their interaction with each other was given by the following expression:
Yi = b0 + b1 X 1 + b2 X 2 + b3 X 3 + b12 X 1 X 2 + b13 X 1 X 3 + b23 X 2 X 3 + b123 X 1 X 2 X 3
(4)
Where Yi represented the experimental response, Xi is the coded variable (−1 or +1), bi represented the estimation of the principal effect of the factor i for the response Yi, whereas bij represents the estimation of interaction effect between factor i and j for the response. The coefficient b0 represented the average value of the response of eight assays; b123 represented the interactions between three factors. In this case, the total number of experiments becomes eight. The variables and the levels of the experiments,
RI PT
with actual and coded scales, are shown in Table 3. The higher-level variable was designated as + and the lower level as -.
Following a preliminary study based on the solubility of alumina, the solvent boiling temperature and the
stirring time, we were able to define the experimental field. As given in Table 4, X1, X2 and X3 are the
corresponding values in the coded form. The maximum and the minimum levels are expressed in coded form as +1 and -1, respectively. The coded values are used to convert the absolute quantity into a dimensionless factor that is convenient for handling the experimental data.
Thus, the answer (Yi) is given by the following formula:
[Na 2O ]0 − [Na 2O ]i .100 [Na 2O ]0
(5)
M AN U
Yi =
SC
The experimental response (Yi) is the removal rate of sodium oxide. It is commonly expressed as percentages.
With, [Na2O]0: Concentration of sodium oxide before sample treatment; [Na2O]i : Concentration of sodium oxide after sample treatment for each experiment.
Following a preliminary study based on the solubility of alumina, the solvent boiling temperature and the stirring time, we were able to define the experimental field.
This area is summarized in Table 2. In Table 3, we gave the experimental matrix represented by the normal units. The Matrix of experience includes the experimental matrix, to which are added two lines to clarify the meaning of the levels (Table 4).
TE D
3. RESULTS AND DISCUSSION 3.1. Analytical results of alumina prior to their treatment This study aimed for the determination of the optimum conditions for the removal of sodium oxide from alumina using the experimental design method (or the full-factorial design). Therefore, we planned to highlight the influence of some important factors on the alumina purification. For this purpose, we worked on three alumina samples of different origin: The first originates from Greece, the second from Spain, and the third from
EP
Germany. These three samples are referenced respectively, S1, S2, and S3. Indeed, we have carried out some chemical analysis and fluorescence spectrometry (XRF) of these three
AC C
samples before treatment. This preliminary study permitted to acquire an idea concerning the major existing
9
ACCEPTED MANUSCRIPT
impurities in these alumina samples. Chemical analysis (a) and X-ray fluorescence analysis (XRF) (b) of the alumina sample S1 prior to its treatment are summarized in Figures 2a and b. These Figures 2a and b provided some important information on the content of each prevailing element in the sample S1. In reviewing this alumina quality, we noticed that silica, P2O5, Fe2O3 were in small quantities, and the CaO that indicated the percentage of contamination was of the order of 0.025%. Only the Na2O showed a low content of around 0.306%. On the other side, analytical results of the alumina sample S2 (Figures 2a’, b’) indicated that the sample S2 contained almost no SiO2 and P2O5. In addition, Fe2O3 and CaO contents are likewise very low, except that sodium oxide that presented a relatively high content as compared to the other samples, of around 0.323%. In this same context and based on the analytical results of the alumina sample S3 summarized in Figures 2a” and b”, we noticed that the alumina S3 showed very low contents of P2O5, SiO2 and
RI PT
Fe2O3, of around of 0.01, 0.0025 and 0.0075%, respectively. By the same, the content of sodium oxide is a low and of the order of 0.259 %. It appeared from all these results that the content of sodium oxide was always
important as compared to the other elements in the three-alumina studied samples S1, S2, and S3. In addition, by comparing the contents of this element in the three-alumina samples, we could note that the sample S2 showed the highest content of sodium oxide.
3.2. Application of the experimental design to the alumina processing
SC
In an experiment, we deliberately changed one or more variables (or Factors) process in order to observe the
effect of the changes that will have on one or more variable responses [31]. The statistical design of experiments (DOE) is an efficient procedure for planning experiments, so that the data obtained can be analyzed to return valid and objective conclusions [32]. The objectives of experimental planning were as
M AN U
following: 1) To evaluate these variables that showed a most influential effect on the response; 2) To get an empirical model depending on the independent variables, so that response is almost desired value maximization or minimization; 3) To minimize uncontrolled variables on the output response; and lastly, to find out the optimal process factors. To test the design of experiments for the different alumina samples, we determined at first by XRF the initial content of sodium oxide (Na2O) before treatment. These tests were conducted according to the plan of experiments, and the results were reported in Table 5.
3.2.1. Application of the experimental design to alumina processing in the case of the sample S1 3.2.1.1. Full Factorial Model
TE D
According to preliminary experiments carried out to identify the appropriate parameters and to determine the experimental domain, quantity of alumina as factor 1 (X1), stirring time as factor 2 (X2) and processing temperature as factor 3 (X3) were foreseen as mainly affecting the alumina purification process. The codified 3-
mathematical model for the 2 factorial design is displayed in the equation 6. The analytical expression that quantitatively correlates processing parameters (Quantity of alumina, stirring time and temperature) with Y (the experimental response) is described in the equation 6. After carrying out tests, the obtained mathematical
EP
model is expressed as follows:
AC C
Y1 = 14.32 − 1.02 X 1 + 0.12 X 2 − 0.37 X 3 − 0.37 X 1 X 2 − 0.53 X 1 X 3 − 0.52 X 2 X 3 − 0.52 X 1 X 2 X 3
(6)
10
ACCEPTED MANUSCRIPT
By referring to the equation 2 and to the effects of the diagram (Figure 3), we could deduce that the coefficients determined for the Quantity of alumina (X1), the Stirring time (X2), and the temperature (X3) were around – 1.02, 0.12 and – 0.37, respectively. In the same context, and according to the results advanced by Cornelius et al. [33], the Stirring time coefficient appeared as considerably high as compared to the other, determined coefficients. This result means that the fact of the Stirring time increases was not probably relevant and important. Quantity of alumina, temperature and the combination of both parameters were the major parameters affecting the alumina refining. In the same context, and based on the results reported by Debadatta and Pramanik [34], almost all the effects were negative, indicating that they directed to have a loss of efficiency in the alumina purification. In addition, while considering this diagram of the effects, we noticed that the majority of the factors showed a negative influence on the response; especially in the case of the
RI PT
quantity of alumina that presented the highest effect. By against, the stirring time showed a positive influence on the response. Therefore, we must reduce the quantity of alumina to improve the response. To highlight the influence of different factors, we presented the interactions between the different components.
As coul be seen in the Figure 4 (A) that summarized the interactions between quantity of alumina - stirring time, for only a small quantity of alumina and for a maximum stirring time, the response is of around 15.8% and vice-versa. In this case, it is helpful to work with small quantities of alumina for a high-stirring time. By
against, the interaction quantity of alumina - temperature showed that the best response is of approximately
SC
15.5% (Figure 4.B). This last value is reached when the quantity of alumina is low and the temperature is maximal and vice-versa. According to the interaction diagrams of stirring times - temperature, we could notice that when the temperature was around 25°C and the stirring times of about 60 min, the response reached its
M AN U
maximum value (Figure 4C).
3.2.2. Application of the experimental Design to alumina processing in the case of the sample S2 3.2.2.1. Full Factorial Model
After carrying out tests, the regression analysis was done to fit the response function (purification of alumina efficiency) with the experimental data. After putting values of all coefficients, the codified mathematical model 3
for the 2 -factorial design is displayed in the final regression equation as:
Y2 = 15.42 − 1.60 X 1 − 0.90 X 3 + 0.58 X 1 X 2 − 1.18 X 1 X 3 − 0.901 X 2 X 3
(7)
Equation (7) showed both the individual effects and the interaction effects of variables regarding the alumina purification efficiency. According to this equation (7), the quantity of alumina (X1) and temperature (X3) showed
TE D
a negative effect while the stirring time presented no apparent effect on the alumina purification in aqueous solution and in the variation range of each variable selected in the present study. However, and according to works advanced by Sutar et al. [35], the positive sign of the various coefficients determined indicated a synergistic effect between the parameters and dependent variables, while a negative sign specified an antagonistic response effect.
In addition, the equation (7) showed that the two variables or the three-variable interactions were significant.
EP
The interaction between Quantity of alumina and stirring time was the most influencing interaction. However, the interaction between quantity of alumina and temperature was the lowest influencing interactions. In the
AC C
same way, the equation (7) showed that the three-variable interactions were not at all significant.
11
ACCEPTED MANUSCRIPT
Thus, Figure 3 (S2) representing the effect diagram relative to the sample S2 showed some significant interactions between the different factors influencing the alumina purification. In addition, visual examination of the diagram revealed that all factors have an effect on the process response, except the stirring time and the interaction between stirring time – temperature. Once again, we could observe that the majority of factors ensured and revealed a negative effect, except for the interaction between quantities of alumina - stirring times that presented a positive coefficient sign indicating a synergistic effect between the parameters and a dependent variable. Whereas, the alumina quantity factor showed the greatest, negative impact, indicative of an antagonistic, effective response. In order to minimize this poor effect, the decrease of the alumina quantity leads to the increase of alumina efficiency purification Figure 5 (A) representing the interaction diagrams displayed a significant interaction between quantities of
RI PT
alumina and the stirring time. Indeed, these diagrams showed that the best response was of the order of 17.6%, for a stirring time of 20 min and a quantity of alumina of 0.6 mg/L. In view of these interactions, the quantity of alumina and the stirring time must be reduced to achieving a higher return or yield.
Visual examination of the interaction diagrams showed that when the quantity of alumina was at its minimum value and with a raised temperature, the response reached its maximum value (Figure 5B). The Figure 5C
representing the Interaction diagrams showed a significant interaction between the time of stirring and the temperature. By analyzing the Figure 5C, we could notice that for temperatures of 25 and 80°C, the
SC
experimental response was equivalent to 16.3 and 14.5%, respectively, whatever the stirring time.
In summary, we could deduce that in all three cases, the best response was obtained for an alumina quantity and for the weakest stirring time.
3.2.3.1. Full Factorial Model
M AN U
3.2.3. Application of the experimental Design to alumina processing in the case of the sample S3
The analytical expression that quantitatively correlates processing parameters (Quantity of alumina, stirring time and temperature) with Yi is described in the Equation 4 as follows:
Y3 = 2.22 + 0.67 X 1 − 0.87 X 2 − 0.60 X 3 − 0.67 X 1 X 2 − 0.40 X 1 X 3 + 0.75 X 2 X 3 + 0.40 X 1 X 2 X 3
(8)
The negative sign in front of the terms indicated antagonistic effect, while the positive sign indicated synergistic effect. Indeed, the coefficients for the alumina Quantity (X1), the Stirring time (X2) and the process temperature (X3) were 0.67, -0.87 and -0.60, respectively. Equation (8) and Figure 6 showed the individual variable effect and the interaction purification effects of the alumina Quality. According to this equation (8), the alumina
TE D
quantity, the combination of both parameters such as the Stirring time and the temperature, and the combination of the three interaction variables appeared as the major parameter affecting the pure quality of this alumina. While the stirring time, the temperature, the combination of both parameters such as the quantity of alumina-stirring time and Quantity of alumina temperature showed a negative effect on the purification Quality of alumina in the aqueous solution at the variation range of each variable selected in the present study. However, and likely to the work also confirmed by Cornelius et al. [33]. The positive sign of the
EP
coefficients indicated a synergistic effect between the parameter and a dependent variable, while a negative
AC C
sign indicated an antagonistic effect on response.
12
ACCEPTED MANUSCRIPT
In addition, the equation (8) supposed that the two variables or the three interaction variables were significant. The interaction between the Stirring time and the temperature appeared as the most influencing one. In the same context, and according to Figure 3 that represented the diagram of the relative effects to the sample S3, it is hollowing out that there are three factors that positively influenced the response, namely, the quantity of alumina, the interaction of order 2 (stirring time - temperature) and the interaction of order 3 (quantity of alumina - stirring time - temperature), but the quantity of alumina and the order of interaction 2 (stirring time temperature) showed the most positive effects, while other factors revealed a negative effect, especially stirring time. To improve the response, we should decrease the value of factors that showed negative effects and impacts. The figure 6 (A) representing the interactions between quantities of alumina - stirring time showed that the
RI PT
largest response is of around 4.4%. This last value is reached when the quantity of alumina is maximum and the stirring time is at its minimum. By the same, the figure 6 (B) representing the interactions between quantity of alumina - temperature showed that the best response was of approximately 3.9%. This response attained its maximum value when the quantity of alumina was at its maximum concentration and for a lower temperature.
It appeared from Figure 6. (C), representing interaction between stirring time– temperature, which the best
temperature) are taking their minimum values.
3.3. Analytical results of alumina after treatment
SC
response is of the order of 4.4%. To reach this last value, it is necessary that these two factors (stirring time–
After treatment and for each quality of alumina, we performed a chemical and X-rays Fluorescence (XRF) analyses to determine the high percentage of sodium oxide removal rate. Therefore, Figure 7 (a, b) showed the
M AN U
results of chemical analyzes, and those achieved by XRF of some impurities in the sample S1. Thus, we noticed that there are some elements whose contents remain constant, as compared to those obtained prior to treatment, such as P2O5 and Fe2O3. The SiO2 content showed a slight increase to reach the value of 0.003%. Only the percentage of Na2O and of CaO presented a net decline.
The results of chemical analysis and those realized by XRF to the sample S2 are grouped in the Figure 7 (a’, b’). These analysis results of the alumina quality, before treatment, did not show the presence of SiO2 and P2O5. While the results after treatment indicated that, the P2O5 content showed slight increase and the silica content remained constant. After treatment, the element that showed the largest decrease was the sodium oxide with a value of around 0.264%. By against, the CaO content increased and became around 0.010%. For the quality
TE D
of alumina corresponding to sample S3, the results of the investigation showing the high rate of Na2O are illustrated in Figure 7 (a”, b”).
To sum up, the analytical results that compared the qualities of alumina before and after treatment, allowed mainly noting that the contents of Na2O, SiO2 and P2O5 showed a net decline, whereas the Fe2O3 one remained constant.
As can be seen, it is noticed that there is a gap between the levels of some impurities that we found and those
EP
cited in the literature. The results obtained after treatment showed that SiO2 and Fe2O3 represent nearly the same value in the three qualities of alumina, and these levels are lower than those found in the literature. This
AC C
decrease of the percentage of these two elements, SiO2 and Fe2O3, might be due to the effect of alumina
13
ACCEPTED MANUSCRIPT
washing or to the level of the initial bauxite impurities [36]. Furthermore, and according to the variability of impurity content associated with alumina and in accordance with the results of chemical analysis obtained, we could deduce that the CaO content is variable depending on the quality of alumina examined. Indeed, its content in the first two qualities is pretty meadows close to the values obtained, but the last is of low quality content that is of around 0.009%. This decrease could be explained by the content present in bauxite or the quantity of lime added to decrease the concentration of carbonate or silica [36]. The percentages of Na2O showed a decline, but they remained high in the three qualities of alumina. This increase is due to the difficulty of separating hydrargillite from its crystallization medium, or to the quantity of caustic soda solution added during the step of the alumina extraction [36].
RI PT
4. CONCLUSION The goal of this work was to remove impurities combined to the alumina, using a factorial design method, and so to determine the influence of some activation parameters.
Main results showed that factorial experimental design approach constituted an excellent tool and could successfully be used to develop an empirical equation for the prediction and understanding the methodology of alumina purification. As we observed, the application of full-factorial designs on samples of different alumina qualities allowed following the rate of sodium oxide removal.
SC
Analytical results of samples prior to their treatment acquired by the chemical analysis and the X-ray fluorescence analysis (XRF) respectively, provided a clear idea about the existing impurities. Therefore, it appeared primarily that the sodium oxide content is the highest element among all the existing impurities.
After the application of the factorial experimental design approach, analysis of the different factor effects and
M AN U
their interactions showed that to get a better response; we might reduce the quantity of alumina, and increase the stirring time for the first two samples S1 and S2 tested. In opposite, it was necessary to increase the quantity of alumina in the case of the third sample S3. Therefore, the difficulty of removing sodium oxide seemed linked to the geological origin of the bauxite. To expand and improve the results of this study, we should take into account all the potential existing impurities, since we mainly found during this study that the levels of a few impurities increased after the treatment.
ACKNOWLEDGMENTS
This study is financially supported by the European Union in the framework of the Avicenna Project no
TE D
93AVI054 and by the Tunisian Ministry of Higher Education, Scientific Research, and Technology, Program Contract (2010–2012). The authors express their great appreciation to all the team of the Laboratory of Treatment and Recovery of Hydric Releases, Center of Research and Technologies of Water (CERTE) Borj-Cédria techno-park that contributed to achieving this work, particularly Mr. Aissaoui, Mr. Hachemi and Ms. Mabrouki.
AC C
EP
Note: The authors have declared no conflict of interest.
14
ACCEPTED MANUSCRIPT
REFERENCES
[1] S. Takashi, W. Hideo, F. Masayoshi, T. Minoru, Structural Properties, and Surface Characteristics on Aluminum Oxide Powders, Appl. Chem. 9 (2009) 23-31. [2] T. Shirai, Ch. Ishizaki, K. Ishizaki, “Effect of Manufacture Process on Water Surface Adsorption of High Purity a-Al2O3 Powders”, Interceram. 50 (2001) 176-181. [3] T. Shirai, K. Matsumaru, Ch. Ishizaki, K. Ishizaki, “Analysis of Hydration and Adsorption Layer on Commercial Sub-micron High Purity a-alumina Powders”, J. Japan. Inst. Metals. 68 (2004) 102-105. [4] C. Morterra, G. Magnacca, “A Case Study: Surface Chemistry and Surface Structure of Catalytic Alumina, as Studied by Vibrational Spectroscopy of Adsorbed Species”, Catal. Today. 27(1996) 497-532.
RI PT
[5] J.M. Saniger, “Al-O Infrared Vibrational Frequencies of alumina”, Mater. Lett. 22 (1995) 109-113. [6] B. Rita, P.R. Jean-Pierre, Raman Microspectrometry Applied to the Study of Electrode Materials for Lithium Batteries, Chem. Rev. 110 (2010) 1278–1319.
[7] S. Wang, Y. Boyjoo, A. Choueib, Z.H. Zhu, Removal of dyes from aqueous solutions using fly ash and red mud, Water. Res. 39 (2005) 129-138.
[8] J. Pera, R. Boumaza, A. Ambroise, Development of a pozzolanic pigment from red mud, J. Cement. Concrete. Res. 27 (1997) 1513-1522.
SC
[9] M.A. Rhamdhani, M.A. Dewan, G.A. Brooks, B.J. Monaghan, L. Prentice, Alternative All production
methods: Part 1 – a review of indirect carbo thermal routes. Transactions of the Institutions of Mining and Metallurgy, Section C, Min. Proc. Ext. Met. 122 (2013) 87-104.
[10] D. Smith, Antigorite Peridotite, Metaserpentinite, and other Inclusions within Diatremes on the
M AN U
Colorado Plateau, SW USA: Implications for the Mantle Wedge during Low-angle Subduction. J. Petrol. 0 (2010) 1-25.
[11] E.M. Ginting, N. Bukit, Synthesis and Characterization of Alumina Precursors Derived from Aluminum Metal Through Electrochemical Method, Indones. J. Chem. 2 (2015) 123-129.
[12] S. Ramaswamy, P. Raghavan, Significance of Impurity Mineral Identification in the Value Addition of Kaolin – A Case Study with Reference to an Acidic Kaolin from India, J. Miner. Mater. Charact. Eng. 10 (2011) 1007-1025.
[13] J.T. Kloprogge, H.D. Ruan, R.L. Frost, Thermal decomposition of bauxite minerals: infrared emission spectroscopy of gibbsite, boehmite and diaspore, J. Mater. Sci. 37(2002) 1121-1129.
TE D
[14] T.C. Santini, M.V. Fey, R.J. Gilkes, Experimental Simulation of Long Term Weathering in Alkaline Bauxite Residue Tailings, Metals. 5 (2015) 1241-1261.
[15] Y. Li, R.J. Haynes, Y-F. Zhou, I. Chandrawana, Chemical and Physical Properties of SeawaterNeutralized Bauxite Residue Mud and Sand and the Effects of Leaching, Int. J. Env. Sci. Dev. 7(2016) 273-277. [16] A. Tor, Y. Cengeloglu, M.E. Aydin, M. Ersoz, Removal of phenol from aqueous phase by using neutralized red mud, J. Colloid. Interface. Sci. 300 (2006) 498-503.
EP
[17] S. Bhattacharyya, T.K. Mukhopadhyay, S. Ghatak, Studies on the dehydration kinetics and rehydration
AC C
of YAG precursor powder in the hydroxyhydrogel form, Ceram. Int. 39 (2013) 121-128.
15
ACCEPTED MANUSCRIPT
[18] S. Bhattacharyya, S. Ghosh, S. Ghatak, Physicochemical characteristics of lanthanum hydroxide in hydroxyhydrogel and conventional gel form, Trans. Indian. Ceram. S. 66 (2007) 97-101. [19] E.A. Mohamed, A.Q. Selim, M.K. Seliem, M.R. Abukhadra, Modeling and Optimizations of Phosphate Removal from Aqueous Solutions Using Synthetic Zeolite Na-A, J. Mater. Sci. Chem. Eng. 3 (2015) 15-29. [20] L.B. Williams, D.W. Metge, D.D. Eberl, R.W. Harvey, A.G. Turner, P. Prapaipong, A.T. Poret-Peterson, Evaluation of the medicinal use of clay minerals as antibacterial agents, Environ. Sci.Technol. 8 (2011) 3768– 3773. [21] B. Zhu, B. Fang, X. Li, Dehydration reactions, and kinetic parameters of gibbsite, Ceram. Int. 36 (2010) 2493-2498. [22] H. Li, T. Shao, D. Li, D. Chen, Nonisothermal reaction kinetics of diasporic bauxite, Thermochim. Acta.
RI PT
427 (2005) 9-12. [23] B. Li, Z. Tingan, L. Yan, D. Zhihe, L. Guozhi, W. Xiaomin, M. Jia, J. Xiaoli, The most probable mechanism function and kinetic parameters of gibbsite dissolution in NaOH, Chinese. J. Chem. Eng. 18 (2010) 630-634.
[24] T.K. Parya, D. Ray, G. Pahari, B.K. Sanfui, T.K. Mukhopadhay, Non-isothermal decomposition kinetics of bayeritic bauxite, J. Indian. Chem. Soc. 89 (2012) 1681-1688.
[25] Y. Li, C. Liu, Z. Luan, X. Peng, C. Zhu, Z. Chen, J. Zhang, Z.J. Fan, Jia, Phosphate removal from aqueous
SC
solutions using raw and activated red mud and fly ash, J. Hazard. Mater. 146 (2007) 255-261.
[26] K-E. Khishigbayar, W-S. Tak, Y-G. Moon, K.B. Shim, C.J. Kim, Purification of the alumina precursor and precise impurity measurements with spectrophotometric method, J. Ceram. Process. Res. 15 (2014) 514-518. [27] Y. Cengeloglu, A. Tor, M. Ersoz, G. Arslan, Removal of nitrate from aqueous solution by using red
M AN U
mud, Sep. Purif. Technol. 51 (2006) 374-378.
[28] K. Dragana, L. Dragica, P.S. Jelena, S. Slavica, V. Ljubica, Determination of Alumina Oxide in Bauxites by X-Ray Fluorescence Analysis, Global. J. Inc. (USA). 12 (2012) 1-6.
[29] K. Nakayama, K. Okawa, T. Kakizaki, T. Honma, H. Itoh, T. Inaba, Arabidopsis Cor15am is a chloroplast stromal protein that has cryoprotective activity and form oligomers, Plant. Physiol. 144 (2007) 513–523. [30] J.J. Molgaard, J.D.II. Auxier, A.V. Giminaro, C.J. Oldham, M.T. Cook, S.A. Young, H.L. Hall, Development of synthetic nuclear melt glass for forensic analysis, J. Radioanal. Nucl. Chem. 3 (2015) 1293– 1301.
[31] V.S. Shrirang, Management of Industrial Solid Waste by Utilizing as an Adsorbent for Methylene Blue
TE D
Removal. Master of Technology in Chemical Engineering, National Institute of Technology Rourkela. 2013, 71p. [32] R. Dahima, A. Gupta, D. Rathore, Quality by Design (QbD) Approach for Formulation Development of Hydralazine Hydrochloride Tablets, J. Chem. Pharm. Res. 5 (2016) 336-341. [33] T. Cornelius, P.K. Guillaume, K. Richard, Decreasing Yield and Alumina Content of Red Mud by Optimization of the Bauxite Processing Process, Int. J. Metall. Eng. 6 (2017) 1-9. [34] D. Debadatta, K. Pramanik, A Study on Chemical Leaching of Iron from Red mud using Sulphuric Acid,
EP
Res. J. Chem. Environ. 17 (2013) 50-56.
[35] H. Sutar, S. C. Mishra, S. K. Sahoo, A. P. Chakraverty and H. S. Maharana, Progress of Red Mud
AC C
Utilization: An Overview,Am. Chem. Sci. J. 4 (2014) 255-279.
16
ACCEPTED MANUSCRIPT
[36] H. Kvande, P.A. Drabløs, The Aluminum Smelting Process and Innovative Alternative Technologies, J. Occup. Environ. Med. 5 (2014) S23–S32.
Legends of Figures
Fig. 1. Calibration curves of, respectively, SiO2 (A), P2O5 (B) and Fe2O3 (C).
Fig. 2. Analytical results of samples S1, S2 and S3 prior to their treatment obtained respectively, by chemical
Fig. 3. Chart of effects for the samples referenced respectively, S1, S2, and S3
RI PT
analysis (a, a ', a ") and by X-ray fluorescence analysis (XRF) (b, b', b").
Quantity of alumina – factor 1 (X1), Stirring time – factor 2 (X2), Temperature – factor 3 (X3).
Fig. 4. Diagram of Interactions Between respectively, quantity of alumina - stirring time (A); quantity of alumina - temperature (B) and stirring time - temperature (C) of the sample S1.
- temperature (B) and stirring time - temperature (C) of the sample S2.
SC
Fig. 5. Diagram of Interactions Between respectively, quantity of alumina - stirring time (A); quantity of alumina
M AN U
Fig. 6. Diagram of Interactions Between respectively, quantity of alumina - stirring time (A); quantity of alumina - temperature (B) and stirring time - temperature (C) of the sample S3.
Fig.7. Analytical results of samples S1, S2 and S3 after to their treatment acquired respectively, by chemical
AC C
EP
TE D
analysis (a, a ', a ") and by X-ray fluorescence analysis (XRF) (b, b', b").
17
ACCEPTED MANUSCRIPT Table 1: Different elements analyzed by X-ray fluorescence spectroscopy Elements
Compound
Fe
Contents (%) Maximum
Fe2O3
0.00
0.10
Ca
CaO
0.00
0.25
P
P 2O 5
0.00
0.80
Si
SiO2
0.00
AI
Al2O3
2.00
Na
Na2O
0.10
Mg
MgO
0.00
-
AlF3
66.00
AC C
EP
TE D
M AN U
SC
RI PT
Minimum
1.00
12.50 0.80
0.02
100.00
ACCEPTED MANUSCRIPT Table 2: Operating variables Values used in the 23-factorial design studies. Coded Variable
Operating variables
Low level (-)
High level (+)
X1
Quantity of alumina (mg/l)
0,6
1,2
X2
Stirring time (min)
20
60
X3
temperature (°C)
25
80
AC C
EP
TE D
M AN U
SC
RI PT
(Xi)
ACCEPTED MANUSCRIPT Tableau 3: Design and experimental results of 23 full-factorial design
Stirring time
Temperature
(mg/l)
(min)
(C°)
1
0.6
20
25
2
1.2
20
25
3
0.6
60
4
1.2
60
5
0.6
20
6
0.6
60
7
1.2
20
8
1.2
60
AC C
EP
TE D
M AN U
Experiments
RI PT
Quantity of alumina
SC
Coded experiment matrix
25 25 80 80 80 80
ACCEPTED MANUSCRIPT
Table 4: Factorial design variables and levels, where (–1) is the lowest level; and (+1) the upper one. Coded experiment matrix X1
X2
X3
1
-1
-1
-1
2
+1
-1
3
-1
+1
4
+1
+1
5
-1
-1
6
-1
+1
7
+1
-1
8
+1
+1
Level (-)
0.6 mg
20 min
Level (+)
1.2 mg
-1
-1
-1
+1
SC
+1
M AN U
TE D EP AC C
RI PT
Experiments
60 min
+1 +1 25°C 80°C
ACCEPTED MANUSCRIPT Table 5: Design and experimental results of 23 full-factorial design Measured response Y (%)
X1
X2
X3
Y1
Y2
Y3
1
-1
-1
-1
14.70
18.27
2.32
2
+1
-1
-1
13.40
14.55
6.56
3
-1
+1
-1
15.67
15.17
1.16
4
+1
+1
-1
15.03
5
-1
-1
+1
15.03
6
-1
+1
+1
16.01
7
+1
-1
+1
13.72
8
+1
+1
+1
11.11
AC C
EP
TE D
M AN U
SC
Experiments
RI PT
Coded experiment matrix
17.34
1.16
17.03
1.16
17.65
1.54
12.07
2.32
11.45
1.54
ACCEPTED MANUSCRIPT
(b)
60
50
50
40
40
30
30
20
20
y = 1900x + 5 R² = 0.97
10 0 0,004
0,009
0,014
0,019
70 60 y = 1480x R² = 0.92
0,024
y = 3700x - 12,5 R² = 0.97
40 30 20 10
10 0 0,004
50
RI PT
70
60
(c)
Optical density
70
0,009
0,014
0,019
0,024
0 0,004
0,009
SC
Concentration (mg/L
0,014
0,019
EP
TE D
M AN U
Figure 1. Calibration curves of respectively, SiO2 (A), P2O5 (B) and Fe2O3 (C).
AC C
Optical density
(a)
0,024
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Figure 2. Analytical results of samples S1, S2 and S3 prior to their treatment acquired by
AC C
EP
chemical analysis (a, a ', a ") and by X-ray fluorescence analysis (XRF) (b, b', b"), respectively.
ACCEPTED MANUSCRIPT
S1 X1
X2
X3
X1X2 X1X3 X2X3 X1X2X3
1,5E+00
5,0E-01 0.1 0,0E+00 -5,0E-01
-0.4
-0.4
-1,0E+00 -1
-0.5
RI PT
Values of effects
1,0E+00
-0.5
-0.5
SC
Factors -1,5E+00
1,0E+00
S2 X1X2
X3
X1X3
X2X3 X1X2X3
0.6
5,0E-01
0
0,0E+00 -5,0E-01 -1,0E+00
-0.9
-1,5E+00 -2,0E+00
X2
M AN U
Values of effects
X1
0
-0.9
-1.2
-1.6
TE D
Factors
8,0E-01
X1 0.7
X2
S3 X3
X1X2
X1X3
0.4
4,0E-01
EP
Values of effects
6,0E-01
X2X3 X1X2X3 0.7
2,0E-01
0,0E+00
AC C
-2,0E-01 -4,0E-01
-0.4
-6,0E-01 -0.6
-8,0E-01
-1,0E+00
-0.7 -0.9 Factors
Figure 3. Effect Diagrams established for the samples referenced respectively, S1, S2, and S3. Quantity of alumina – factor 1 (X1), Stirring time – factor 2 (X2), Temperature – factor 3 (X3)
ACCEPTED MANUSCRIPT 14.8
15.8
1,6E+01
(A)
13.5
13
1,4E+01 1,0E+01 Stirring time = 20 min
6,0E+00
Stirring time = 60 min
RI PT
8,0E+00 4,0E+00 2,0E+00 0,0E+00 6,0E-01
1,2E+00
Quantity of alumina (mg)
15.2
15.5
M AN U
12.4
1,4E+01 1,2E+01 Yield (%)
(B)
14.2
1,6E+01
SC
Yield (%)
1,2E+01
1,0E+01 8,0E+00
Temperature = 25 min
6,0E+00
Temperature = 80 min
4,0E+00 2,0E+00
TE D
0,0E+00 6,0E-01
1,2E+00
EP
Quantity of alumina (mg)
(C)
15.3
1,6E+01 1,5E+01 Yield (%)
AC C
1,5E+01
14.3
14
1,4E+01
13.5
Temperature = 25 min Temperature = 80 min
1,4E+01 1,3E+01 1,3E+01
20
60
Stirring time (min)
Figure 4. Interaction Diagrams between, quantity of alumina - stirring time (A); quantity of alumina - temperature (B) and stirring time - temperature (C) respectively, in the case of sample S1
ACCEPTED MANUSCRIPT
17.6 16.4
1,8E+01 1,6E+01
13.3
(A)
14.3
1,4E+01
RI PT
Yield (%)
1,2E+01 1,0E+01 8,0E+00
Stirring time = 20 min
6,0E+00
Stirring time = 60 min
4,0E+00 2,0E+00 0,0E+00 1,2E+00
SC
6,0E-01
Quantity of alumina (mg)
17.3
M AN U
16.7 1,6E+01
11.7
1,4E+01
Yield (%)
1,2E+01 1,0E+01
Tempera ture = 25 min
8,0E+00
Tempera ture = 80 min
6,0E+00 4,0E+00
TE D
2,0E+00 0,0E+00
(B)
15.9
1,8E+01
6,0E-01
1,2E+00
Quantity of alumina (mg)
16.4
EP
1,8E+01 1,6E+01
(C)
16.3 14.5
14.5
1,4E+01 1,0E+01 8,0E+00
Tempera ture = 25 min
6,0E+00
Tempera ture = 80 min
AC C
Yield (%)
1,2E+01
4,0E+00
2,0E+00 0,0E+00
20
60 Stirring time (min)
Fig. 5. Interaction Diagrams between quantity of alumina - stirring time (A); quantity of alumina temperature (B) and stirring time - temperature (C) respectively, in the case of sample S2
ACCEPTED MANUSCRIPT (A)
4.4
5,0E+00
Yield (%)
4,0E+00 3,0E+00 1.7
0,0E+00 1,2E+00
3.9
3,0E+00
M AN U
Yield (%)
4,0E+00
SC
Quantity of alumina (mg)
5,0E+00
1.9
1.7
1.3
2,0E+00 1,0E+00 0,0E+00 6,0E-01
Stirring time = 60 min
RI PT
1,0E+00
6,0E-01
Stirring time = 20 min
1.3
1.3
2,0E+00
(B)
Tempera ture = 25 min Tempera ture = 80 min
1,2E+00
TE D
Quantity of alumina (mg)
(C)
4.4
5,0E+00
EP 3,0E+00
1.7
2,0E+00
AC C
Yield (%)
4,0E+00
1.2
1.5
Tempera ture = 25 min Tempera ture = 80 min
1,0E+00 0,0E+00 20
60
Stirring time (min)
Figure 6. Interaction Diagrams between, quantity of alumina - stirring time (A); quantity of alumina temperature (B) and stirring time - temperature (C) respectively, in the case of the sample S3
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
EP
Figure 7. Results of samples S1, S2 and S3 after to their treatment acquired by chemical analysis (a, a
AC C
', a ") and by X-ray fluorescence analysis (XRF) (b, b', b"), respectively.
ACCEPTED MANUSCRIPT
*Remove impurities aggregated to the alumina, using a factorial design method. *Analytical results of samples gave us an idea about the existing impurities.
AC C
EP
TE D
M AN U
SC
RI PT
*After treatment, reducing quantity of alumina for S1 and S2 and increasing for S3.
