Smirmis Nsirimoghddm,Ali Mohei,*,Mohsen Krimi,M.Rez Yrhmdi

a Department of Chemical Engineering,Faculty of Engineering,Shahid Bahonar University of Kerman,Kerman,Iran

b Department of Chemistry and Chemical Engineering,Chalmers University of Technology,SE-41296 Gothenburg,Sweden

c Sarcheshmeh Copper Complex,Kerman,Iran

Keywords:Column flotation Nanoparticles Cu recovery Response surface methodology

ABSTRACT Nanoparticles (NPs)can promote the column flotation process in mining industry.Nanoparticles’ effects on column flotation process (copper recovery,grade and flotation rate constant)are assessed in Sarcheshmeh Copper Complex,Iran,through response surface methodology (RSM)optimization technique.The γ-Al2O3,α-Fe2O3,SiO2,and TiO2nanoparticles are selected for these experiments.A flotation rate constant is chosen as a response to assess the effect of nanoparticles on flotation in its kinetic sense.The process pH and nanoparticle dosage are selected as the influential parameters.Results obtained from RSM indicated that the maximum percentage of Cu recovery and grade is obtained at pH of 12 and nanoparticle dosage of 6 kg/t,through α-Fe2O3and γ-Al2O3nanoparticles,respectively.Applying nanoparticles in particular γ-Al2O3and α-Fe2O3increases the Cu recovery by 8-10% together with the grade by 3-6% in a significant manner.It is revealed that nanoparticles could effectively be applied in enhancing the flotation performance.

1.Introduction

Flotation is one of the most effective techniques applied widely in separating copper ore based on their differences in physical and chemical properties.In this process,after preparing the pulp with chemical reagents,named collectors,it is revealed that some of them are hydrophobic and some hydrophilic.In the selected separation processes,copper minerals attach to air bubbles and move to the surface of the flotation cell and form a stable froth layer containing 20-30% copper.The column flotation is one of the latest flotation methods with a high focus in mineral processing.In general,due to its simpler structure,higher selectivity,low cost,better product recovery and higher grade,it is more efficient than the conventional flotation cells [1-6].

It is found that the flotation performance (i.e.recovery and grade)strongly depends on the froth stability,bubble distribution,bubble mobility and froth properties on which there exit many studies,e.g.[7-10].If the froth is too stable,then detachment of affixed material from the froth is difficult,otherwise,the floated material is lost due to the bubble coalescence in the froth.The run assessments indicate that in addition to the type and concentration of the frother,the hydrophobicity,particle size,type and concentration of present particles affect the froth stability and cause an increase in the hydrophobic particles,leading to an increase in froth stability [11-17].

To improve the froth stability in mineral products in the related industries,it is necessary to apply new technologies.Nanotechnology is one of these growing and important technologies with widespread applications in mineral processing worldwide.Because of nanoparticles’distinct properties and their high surface to volume ratio,they are able to promote the froth stability,and enhance the flotation performance.In addition,nanoparticles(NPs)are demonstrated to play an important role as the effective collector in flotation process [18-22].Nanoparticles are of great potential applications in flotation field and are applied as a stabilising agent,like surfactant to enhance the froth stability by retarding bubble coalescence.The ability of the nanoparticles to stabilize froth on its own and without any other surfactant is the subject of many studies not related to flotation processing.These studies indicate that the nanoparticles increase the froth stability by forming a barrier to coalescence and decelerating the process of thinning the film though preventing the water flow at the bubble surface[19,23].Nanoparticles are recyclable,thus profitable [24-26].

Despite some studies run on nanoparticle-stabilized froth,studies on nanoparticles in the three-phase column flotation with an industrial feed are limited and yet the nanotechnology can be applied in the mineral separation which is a new challenge[27,28].The simultaneous effects of different parameters including process environment pH and NP dosage on copper recovery,grade and flotation rate constant are not assessed enough.Reviews run by the authors here indicate that only a single mineral in the flotation process has been subject of study.To study the simultaneous effects of important parameters,one of the statistical techniques,response surface methodology (RSM),is usually applied [29,30].

The RSM is applied for analyzing experiments where one or more independent variables-as responses-are affected by many variables with the objective to optimize the responses [31,32].One of the advantages of applying RSM in the Design Expert software environment,in addition to reducing the number of experiments,is the possibility of devising equations for independent and dependent variables.The RSM includes different design models like Box-Behnken (BB),factorial,D-optimal,and central composite design (CCD)[30,33,34].This method is successfully applied in optimizing different processes to reduce the cost of analysis methods like computational fluid dynamics (CFD)and their related numerical uncertainties [35-38].

Here,the usability of nanoparticle stabilized froth is assessed through RSM technique based on CCD approach in a flotation process.To materialize this objective here,the simultaneous effects of critical parameters like the process pH and NP dosage on the flotation performance responses are assessed.In this study,assessing the flotation rate constant (k)is the main subject of the flotation kinetics because it is presented as a function of the flotation dynamics and the properties of minerals.In addition to the recovery and grade,assessing the flotation rate constant is of essence[39].Assessing the effect of these nanoparticles on the recovery,grade and flotation rate constant for an industrial feed is new.To assure the size of selected nanoparticles,the dynamic light scattering (DLS)is applied.

2.Materials and methods

The method adopted in this study is experimental.The flotation column is devised in the Sarcheshmeh Copper Complex laboratory.Efforts are made to assess the possibility of applying nanoparticles as a platform for the progress of column flotation both in their quantitative (recovery)and qualitative (grade)sense in real ore flotation.

2.1.Experimental system

Batch flotation experiments are designed through RSM to assess the effects of critical factors on the flotation performance.The experiments are run on a laboratory-scale flotation column setup(Fig.1).

Fig.1.Schematic of the laboratory flotation column of the Sarcheshmeh Copper Complex.

Based on the aspect ratios of the pilot plant and industrial flotation columns,a 0.007 m3flotation column is designed and built.This column is cylindrical and to see the pulp level easily,and is made of transparent Plexiglass.The dimensions of the column are 10 cm diameter and 100 cm height.Water is added to maintain the level of the pulp in the column and the tests are run to reduce the froth depth.A porous polyethylene cylinder is installed at the bottom of the column to distribute air over the column crosssection in a uniform manner with a filter cloth sparger.

The subject copper ore samples are processed in the flotation column with constantly adding reagent at any process.Here,first,the nanoparticle is dispersed by the agitator into 7.5% constant solid pulp.Next,the collector dosages (sodium isopropyl xanthate and butyl sodium dithiphosphate)are added into the pulp to enhance the hydrophobicity of copper particles and to allow 2 min for conditioning.Then,the frother dosages (MIBC and Nasfroth)are added and conditioned for a further 1 min.These flotation conditions are determined by applying the results of the preliminary flotation tests performed without the nanoparticles.Finally,air is injected through the flowmeter with the superficial gas velocity of 1.5 cm/s at the bottom of the column and the concentrates’collection is accomplished at 0.5,1.5,4 and 8 min cumulative times into separate containers with tailings that are filtered,dried and weighed.Here the process pH of tests is adjusted by adding CaO.The Laser Particle Sizer (FRITSCH ANALYSETTE 22 Nano-Tec)is applied to measure the mineral particle size.In this study,the froth depth is kept constant at 4 cm in all runs.

The flotation rate constant (k)as the indicator of flotation performance,is subject to both the dynamics of flotation columns and the chemical properties of the minerals,which are determined by fitting the following first-order rate equation to the recoverytime curves obtained for each particle size fraction [40]:

where Rmaxis the predicted maximum recovery.

The four nanoparticles types of γ-Al2O3,α-Fe2O3,SiO2,and TiO2were purchased from US Research Nanoparticles,Inc.,USA and applied in the designed flotation experiments.Their properties are tabulated in Table 1.

2.2.Experiment design

In this experiment,the flotation experiments are optimized through a three-level(-1,0,+1)and two-parameter CCD approach in Design Expert V 10.0.3 S/W.Selected parameters and their levels are determined from the obtained results of preliminary tests.The parameters considered here consist of the process pH (X1)and NP dosage (X2)as the independent parameters,being the most influenced on the flotation process.The available results indicate that the copper recoveries remain relatively high in the alkaline pH range and the highest recovery and grade are yielded with pH in the range of 10-12 [2];thus,this range was selected for flotation experiments.The applied industrial samples are of pyrite minerals that are naturally depressed at alkaline pH ＞10,indicating that selective separation of copper from samples can be achieved atpH ＞10,thus causing an increase in the grade.In our experiment design,the changes in NP dosage are assessed within the range of 0.5-10 kg/t.The results(Fig.2)showed that recovery variations from dosage of 0.5-2 kg/t are negligible.An increase in NP dosage(with the range of 4-10 kg/t)increases the recovery percentage,but in the range of 6-10 kg/t these percentages are close to each other;therefore here,a 4-6 kg/t range is selected for the four applied NPs.

Table 1 Nanoparticles applied in this study.

The copper recovery (Y1),copper grade (Y2)and flotation rate constant (Y3)are applied as the response variables.The levels of the independent and dependent selected variables are tabulated in Tables 2 and 3,respectively.For each NP,a total of 17 experiments consisting of five replicates at the centre points,one axial point and two full factorial points with three responses are designed and implemented in one block in a random manner.

The reliable approach in evaluating the accuracy of the fitted model is assessed by the coefficient of determination (p-value and R2),the lack of fit tests and the Fisher test value (F value)obtained from ANOVA.The graphical analysis is run to illustrate the optimum condition for column flotation performance.The objective of this design is to assess the optimal pH and NP dosage for improving the column flotation performance.

3.Results and discussion

3.1.Dynamic light scattering (DLS)

The hydrodynamic diameter distribution of the four NP types in water at 90°fixed scattering angle and 25°C constant temperature prepared by the dynamic light scattering (DLS-HORIBA L-550)technique is applied to confirm the hydrodynamic NPs size.The size distributions for the subject NPs are nano-sized,Fig.3,where,as observed it is confirmed that the distribution range of the subject NPs in Table 1 is narrow.

Fig.2.Effect of NP dosage on recovery.

Table 2 Levels of independent variables in the CCD experimental design.

Table 3 Selected dependent variables in the CCD experimental design.

3.2.Design of experiment

The experiments are run according to CCD approach of the RSM to optimize the nanoparticles’ conditions in obtaining maximum column flotation performance through nanoparticles in 17 batch runs for every nanoparticle.Details of the nanoparticles in the design of experiment are tabulated in Tables 4-7.

The obtained results through Design Expert software are represented as ten quadratic polynomial and two linear equations of Y3model for γ-Al2O3and TiO2for predicting all three responses:

For γ-Al2O3NPs:

where X1is the process pH;X2the NP dosage;and X1X2the interaction of these two parameters.

The X1X2is the response changes when two independent variables (X1and X2)are changed simultaneously.The positive and negative signs of coefficient indicate the positive and negative effect on Y1,Y2and Y3.An increase in the pH with the range of 10-12 decreases the flotation rate constant in the chalcopyrite samples;X1has negative impacts on the response Y3,for every nanoparticle.This may be due to the factors like nanoparticle aggregation,changing the pulp viscosity or longer bubble resistance time.

Fig.3.DLS graphs of the four NPs.

Table 4 Experimental design of the column flotation and obtained results for γ-Al2O3.

Table 5 Experimental design of the column flotation and obtained results for α-Fe2O3.

Table 6 Experimental design of the column flotation and obtained results for SiO2.

Table 7 Experimental design of the column flotation and obtained results for TiO2.

To evaluate the accuracy of these models for each nanoparticle,the criteria of R2and adjusted R2coefficient,SD,CV and AP values are applied.The results reveal that these models can estimate the recovery,grade and flotation rate constant of copper,in a sufficient manner.

3.3.ANOVA analysis

Results of the fitted models of responses obtained through ANOVA analysis are tabulated in Tables 8-11 for γ-Al2O3,α-Fe2O3,SiO2,and TiO2NPs,respectively,where,these suggested models are significant at 95% confidence level (p-value ＜0.05)in predicting copper recovery,grade and flotation rate constant.

As shown in Tables 8-11,all model F-values’ amounts for copper recovery,copper grade and flotation rate constant parameters are high for all subject NPs,indicating that the models are highly significant and there is only a 0.01% chance for the model F-value distribution,which is due to noise.The Prob ＞F values less than 0.05 indicate that the model terms are significant for all NPs,and the highly significant models are observed at the 95% confidence level.

The low amounts of lack of fit of suggested models,that describe the variation of the data around the presented models,are not significant,indicating the fitness of the models.The results indicate that R2＞0.9 for all suggested models of the four NPs is of appropriate fit to the models to the experimental data.The adjusted R2values for all response models are higher than 0.9,except Y2for α-Fe2O3which is about 0.88.These findings confirm that these models are appropriate for predicting the responses in all four NPs.

The AP values measure the signal to noise ratio,where AP ＞4 is appropriate,and here,the AP ＞4 holds true in all presented models for all NPs.This value suggests that these models can be applied in navigating the design space.For these models to be reproduced the CV values must not exceed 10% [41].Accordingly,the CV values of these three responses are within the range of 1.06-6.11% for γ-Al2O3,2.16-7.81% for α-Fe2O3,1.46-4.63% for SiO2and 1.42-5.92% for TiO2NPs,indicating a reliable precision and reliability of these models.

3.4.Response surface analysis

To assess the interaction and the simultaneous effects of different independent parameters on the responses,the 3D response surface graphs are plotted in Fig.4a-d,Fig.5a-d and Fig.6a-d for Cu recovery,Cu grade and Cu flotation rate constant,with different NPs,respectively.

Table 8 ANOVA results for Cu recovery (Y1),grade (Y2)and flotation rate constant (Y3)for γ-Al2O3.

Table 9 ANOVA results for Cu recovery (Y1),grade (Y2)and flotation rate constant (Y3)for α-Fe2O3.

Table 10 ANOVA results for Cu recovery (Y1),grade (Y2)and flotation rate constant (Y3)for SiO2.

3.4.1.Effect of pH and NP dosage on the Cu recovery

As to Cu recovery,the 3D simultaneous effects of process pH and NP dosage for γ-Al2O3,α-Fe2O3,SiO2and TiO2NPs are shown in Fig.4a-d,respectively,whereas observed,this recovery depends more on the pH than the NP dosage,thus making pH the most effective parameter.Recovery increases by an increase in both the parameters for all NPs (except for TiO2).This dependency on pH can be attributed to its performance on NPs.Because NPs stabilize the froth with their large surface area;an increase in the NPs concentration produce smaller bubble and bigger surface area,thus causing an increase in recovery.Accordingly,any amount of NP dosage indicates a positive effect on recovery for all NPs.As observed in Fig.4d,as to TiO2the effect of pH and NP dosage on the Cu recovery are positive until the pH pulp reaches 11,after which a decreasing trend is followed as the pulp pH increase.This phenomenon indicates that by increasing pH (with the range of 11-12)the bubble coalescence in the froth on the surface decrease in presence of TiO2which is directly related to drainage of theliquid available in the bubbles which ultimately reduce recovery.As observed in Fig.4d,the presence of TiO2has a positive effect and by an increase in NP dosage increases recovery.

Fig.4.Response surface plots showing the effects of pH and NP dosage on Cu recovery for different NPs.

Fig.5.Response surface plots showing the effects of pH and NP dosage on Cu grade for different NPs.

Fig.6.Response surface plots showing the effects of pH and NP dosage on Cu flotation rate constant for different NPs.

3.4.2.Effect of pH and NP dosage on the Cu grade

The 3D surface plots of the two parameters’ effects on the Cu grade for γ-Al2O3,α-Fe2O3,SiO2and TiO2NPs are shown in Fig.5a-d,respectively.Fig.5a-d reveal the synergistic effect for both the parameters while pH has the greatest influence on Cu grade and a slight effect of the NP dosage is observed on the α-Fe2O3grade.The copper grade gradually increases with an increase in the pH for all NPs.The percentage grade increases with an increase in γ-Al2O3,SiO2and TiO2NP dosage.This finding is in agreement with the p-values in Tables 8-11 obtained from ANOVA analysis.

3.4.3.Effect of pH and NP dosage on flotation rate constant

The combined effects of the two independent parameters on flotation rate constant for γ-Al2O3,α-Fe2O3,SiO2and TiO2NPs are presented in Fig.6a-d,respectively.The drown plots reveal that pH has a greater influence on flotation rate constant than dosage.Both the parameters exhibit a similar impact on the flotation rate constant for SiO2(Fig.6c)and TiO2(Fig.6d)where a linear decrease in flotation rate constant is apparent with an increase in pH and NP dosage.This can be related to the reduction of selectivity with increasing NP dosage.A similar result for the influence of pH on flotation rate constant for γ-Al2O3(Fig.6a)and α-Fe2O3(Fig.6b)is observed,while by increasing the amount of NPs,the flotation rate constant increases for γ-Al2O3,α-Fe2O3NPs as well.This increase can be attributed to high surface area and accessibility of attachment sites by increasing NPs amount.The reduction in flotation rate constant is linear in the Eqs.(4)and (13)and the same holds true in Fig.6a and d.

3.5.Optimizing of RSM results

To study the flotation kinetic behaviour,the effects of different pHs and NP dosages,on the flotation rate constant were assessed,while Cu recovery (Y1)and Cu grade (Y2)remain most important objectives in optimization studies;therefore,only Y1and Y2are of concern.It should be noted that the optimized values of flotation rate constant for the NPs are approximately within the range of 0.82-1.05,such that the maximum value is obtained 1.05 for γ-Al2O3NP.In our study,because the optimal value of the flotation rate constant is not the required response(the maximum recovery and grade values are the most important),and because with considering the optimal value,the desirability function value decreases;therefore,the effect of flotation rate constant was not considered.

The objective of RSM is to optimize the Y1and Y2values for both the independent parameters at different selected ranges in a simultaneous manner.Here,the pH and NP dosage values are maintained within the selected range of experimental condition.The optimal responses are obtained through the desirability function method,most widely applied in optimization of multiple response process in Design Expert software environment [42,43].The optimum pHs and NP dosages in finding the maximum copper recovery and grade with the highest efficiency for these four NPs are tabulated in Table 12.

The confirmatory experiments are run through the recommended optimum conditions for all the NPs to check the model validity.The obtained results indicate that the copper recovery and grade are the highest for α-Fe2O3and γ-Al2O3NPs subject to optimum conditions pH=12 and NP dosage=6 kg/t,respectively.Because recovery is of high importance and also the grade is close to maximum value for α-Fe2O3NP;therefore,this NP is recommended as having the best ability to improve column flotation performance.

Finally,it is important in this case to highlight that nanoparticles can greatly affect the froth phase.The experimental observations indicate that applying nanoparticles in the mineral flotationlead to generation of many small bubbles in the froth phase and by preventing bubbles coalescence or burst,the froth is become more stable.It can be deduced that nanoparticles are highly contributive in having uniform bubble diameter and the recovery and flotation performance strongly depend on froth stability[27,28].

Table 12 Most effective independent parameters applied in finding the optimum responses Y1and Y2for NPs.

4.Conclusions

Here,the laboratory column flotation cell located at Sarcheshmeh Copper Complex is applied in assessing the effect of nanoparticles on the flotation performance through RSM technique.For this purpose,γ-Al2O3,α-Fe2O3,SiO2and TiO2nanoparticles due to their large possessing surface to volume ratio are selected as the froth stabilizers.The α-Fe2O3NPs have the higher copper recovery,71.30% at optimum experimental conditions,obtained through CCD method among its counterparts.The pH is a more effective parameter on all three responses,copper recovery,grade and flotation rate constant than NP dosage.For copper grade at optimum conditions,γ-Al2O3NPs reveal optimal value of 29.81.However,α-Fe2O3nanoparticle is recommended as having the best ability to improve column flotation performance.Assessing the flotation rate constant is run for evaluating the nanoparticle impact on flotation kinetic.The agreement between the actual and predicted values are acceptable,that is,their applicability in the mining industry.The results obtained here at optimum conditions indicate that these nanoparticles can be applied efficiently for mineral separation in industrial samples in general,but here,in specific,could increase copper recovery and grade.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors would like to extend their appreciation to the Sarcheshmeh Copper Complex management for providing the pilot plant Concentrator and R&D of this Complex and their financial support for this research.