Article Elimination of the Adverse Effect of Ion on the Flotation Separation of from

Na Luo * ID , Dezhou Wei *, Yanbai Shen, Cong Han and Caie Zhang School of Resources and Civil Engineering, Northeastern University, Shenyang 110819, China; [email protected] (Y.S.); [email protected] (C.H.); [email protected] (C.Z.) * Correspondence: [email protected] (N.L.); [email protected] (D.W.); Tel.: +86-24-8367-3863 (D.W.)

Received: 19 July 2017; Accepted: 11 August 2017; Published: 18 August 2017 Abstract: The separation of magnesite from dolomite was studied by flotation tests, X-ray photoelectron spectroscopy (XPS), and zeta potential measurements in the presence of calcium ion (Ca2+) dissolved from dolomite. Sodium oleate (NaOL) was used as collector, and sodium (Na2CO3) and sodium hexametaphosphate (SH) were used as regulators. The results showed that SH had a good selective inhibition ability in pure flotations of magnesite and dolomite. While in the presence of Ca2+ dissolved from dolomite, magnesite and dolomite were both inhibited by SH. The separation of magnesite from dolomite cannot be realized because Ca2+ can adsorb on the surface of magnesite in the form of CaCO3 and change the surface properties of magnesite. Thus, the magnesite flotation was depressed. When the sequence of reagent 2+ addition was changed to add SH prior to Na2CO3, a complex was made by Ca reacting with SH, which avoided the adsorption of Ca2+ on the magnesite surface and prevented the changing of the magnesite’s surface properties. Then, after adjusting the solution pH with Na2CO3, the flotation separation of magnesite from dolomite could be achieved.

Keywords: magnesite; dolomite; calcium ion; flotation

1. Introduction

Magnesite (MgCO3) is the main source of mineral, and is used in metallurgy, construction, the chemical industry, and other fields [1,2]. As the market demand for mineral is increasing, the amount of high quality magnesite is gradually decreasing. It is necessary to develop a dressing process for low grade magnesite to meet the needs of market. Especially, flotation is one of the most effective methods to increase the recovery of fine-grained and low-grade magnesite. In recent years, mineral processors dealing with magnesite have faced a difficult problem in that the flotation results of magnesite are not ideal in the presence of carbonate minerals, such as and dolomite [3,4]. Dolomite (CaMg(CO3)2) which is one of the most common and typical carbonate gangues, often coexists with rhodochrosite, smithsonite, magnesite, and [5–10]. A great quantity of dolomite enters into magnesite concentrate easily in the flotation process, thus affecting the grade of the concentrate and causing problems for the downstream process. Therefore, it is very important to realize the effective separation of magnesite from dolomite. Both magnesite and dolomite are calcite-group minerals with similar physical and chemical properties and the same structure [11–13]. In addition, magnesite and dolomite, as salt-type minerals, can dissolve and give Mg2+, Ca2+, and other hydrolysis products, which have a crucial effect on the flotation behaviors of magnesite and dolomite in aqueous solution [14]. So, it is difficult to achieve an effective separation of magnesite from dolomite in conventional flotation. Some studies have shown that the separation of valuable minerals from carbonate gangue is extremely complex due to the interaction between minerals and several dissolved metals, leading to

Minerals 2017, 7, 150; doi:10.3390/min7080150 www.mdpi.com/journal/minerals MineralsMinerals2017 2017, 7,, 1507, 150 2 of2 11of 11

sorption processes such as precipitation/coprecipitation, adsorption, and ionic substitution [15,16]. sorptionAlthough processes some reagents, such as precipitation/coprecipitation,such as sodium silicate, starch, adsorption, and carboxymethyl and ionic cellulose, substitution have [15 a, 16good]. Althoughinhibiting some effect reagents, on dolomite such asflotation, sodium there silicate, is no starch, selectivity and carboxymethyl in the flotation cellulose,separation have of magnesite a good inhibitingfrom dolomite effect on in dolomitepractice, which flotation, is probably there is nodue selectivity to the two in minerals’ the flotation similar separation surface ofproperties magnesite and fromdissolved dolomite species in practice, [17,18]. which Therefore, is probably it is very due important to the two to minerals’ study the similar influence surface of dissolved properties mineral and dissolvedspecies on species the flotation [17,18]. separati Therefore,on itof ismagnesite very important and dolomite. to study the influence of dissolved mineral speciesIn on this the investigation, flotation separation the effect of magnesiteof Ca2+ on andthe separation dolomite. of magnesite and dolomite was studied. Additionally,In this investigation, a novel method the effect which of Ca uses2+ on the the comp separationlexation of of magnesite sodium hexametaphosphate and dolomite wasstudied. (SH) was Additionally,applied to aeliminate novel method the adverse which useseffect the of complexationCa2+ on the flotation of sodium ofhexametaphosphate magnesite. Thus, the (SH) flotation was appliedseparation to eliminate of magnesite the adverse and dolomite effect can of Ca be2+ realized.on the flotation of magnesite. Thus, the flotation separation of magnesite and dolomite can be realized. 2. Materials and Methods 2. Materials and Methods 2.1. Minerals and Reagents 2.1. Minerals and Reagents The magnesite and dolomite samples used for all of the experiments were obtained from LiaoningThe magnesite Province, and China. dolomite The samples chemical used analyses for all of themagnesite experiments and dolomite were obtained are listed from Liaoningin Table 1, Province,which indicates China. The that chemical the purities analyses of the of magnesite andand dolomitedolomite areare listed 99.20% in Tableand 98.99%,1, which respectively. indicates thatThe the mineral purities samples of the magnesitewere dry ground and dolomite and screened, are 99.20% and and the 98.99%,−150 μm respectively. fraction was The used mineral for the samples were dry ground and screened, and the −150 µm fraction was used for the flotation flotation tests. NaOL and SH were used as a collector and depressant, respectively. Na2CO3 was used tests.as a NaOLpH regulator. and SH Deionized were used wate as ar was collector used andfor all depressant, of the tests. respectively. Na2CO3 was used as a pH regulator. Deionized water was used for all of the tests. Table 1. The chemical composition of magnesite and dolomite. Table 1. The chemical composition of magnesite and dolomite.

Sample MgO/% CaO/% SiO2/% Al2O3/% Purity/% Sample MgO/% CaO/% SiO2/% Al2O3/% Purity/% Magnesite 47.24 0.17 0.19 / 99.20 Magnesite 47.24 0.17 0.19 / 99.20 DolomiteDolomite 21.5221.52 30.13 30.13 0.21 0.21 0.10 0.10 98.99 98.99

2.2.2.2. Experiments Experiments

2.2.1.2.2.1. Flotation Flotation Tests Tests TheThe flotation flotation of of a singlea single mineral mineral and and the the separation separation of of artificially artificially mixed mixed samples samples were were carried carried outout in in an an XFG XFG laboratory laboratory flotation flotation cell.cell. The mechanical impeller impeller speed speed was was fixed fixed at at 1600 1600 r/min. r/min. The Themineral mineral suspension suspension was was prepared prepared by adding by adding 2.0 g 2.0 of pure g of pureminerals minerals (3.0 g (3.0of artificially g of artificially mixed mixedsamples) samples)to 40 mL to of 40 distilled mL of distilled water. The water. reagent The reagentaddition addition orders and orders the andtime the of timeconditioning of conditioning pulp are pulp shown are in shownFigures in Figures1 and 2.1 Both and 2the. Both froth the products froth products and the taili andngs the were tailings collected, were collected,filtered, and filtered, dried. and The dried.recovery Thewas recovery calculated was on calculated the basis of on weight the basis and of elemental weight and contents elemental of the contents products of obtained. the products obtained.

FigureFigure 1. TheThe common common addition addition order order of reagents.

Minerals 2017, 7, 150 3 of 11 Minerals 2017, 7, 150 3 of 11

FigureFigure 2. The 2. The new new addition addition order order of reagents.of reagents.

2.2.2.2.2.2. Zeta Zeta Potential Potential Measurements Measurements ZetaZeta potentials potentials were were measured measured with awith Nano-ZS90 a Nano-Z zetaS90 potential zeta potential analyzer. Finelyanalyzer. ground Finely minerals ground (20minerals mg) below (20− mg)5 µm below were − added5 μm were to 50 mLadded aqueous to 50 mL solution aqueous and conditionedsolution and by conditioned magnetic stirring by magnetic for 15 minstirring so that for 15 the min suspension so that the was suspension homogenized. was Athomoge each condition,nized. At each the zeta condition, potentials the ofzeta the potentials minerals of werethe measured minerals three were times measured individually three andtimes the indivi averagedually value and was the accredited. average Potassium value was nitrate accredited. was usedPotassium to maintain nitrate the was ionic used strength to maintain at 10−3 themol/L. ionic strength at 10−3 mol/L.

2.2.3.2.2.3. XPS XPS Measurements Measurements XPSXPS measurements measurements were were carried carried out out by by means means of of an an America America Thermo Thermo VG VG ESCALAB ESCALAB 250 250 spectrometerspectrometer using using Al Alα X-raysKa X-rays (1486.6 (1486.6 eV) eV) asa as sputtering a sputtering source source at aat power a power of 150of 150 W W(15 (15 kV kV 10 10 mA) mA).. TheThe binding binding energy energy scale scale was was corrected corrected based based on on a a C1s C1s peak peak from from contaminations contaminations (around(around 284.8284.8 eV) as as the internalinternal bindingbinding energyenergy standardstandard[ 19[19].]. The The prepared prepared samples samples were were dried dried in in a vacuuma vacuum oven oven at at roomroom temperature temperature for for 24 h.24 h.

3. Results3. Results and and Discussion Discussion

3.1.3.1. Flotation Flotation Behaviors Behaviors of Magnesiteof Magnesite and and Dolomite Dolomite FigureFigure3 shows 3 shows the the effect effect of NaOLof NaOL dosage dosage on on the th flotatione flotation recoveries recoveries of magnesiteof magnesite and and dolomite. dolomite. It isIt obviousis obvious that that the the recoveries recoveries of theof the two two minerals minera increasels increase steadily steadily with with the the increasing increasing of NaOLof NaOL dosagedosage at at pH pH 10.5. 10.5. When When the the NaOL NaOL dosagedosage is 120 mg/L, mg/L, both both magnesite magnesite and and dolomite dolomite can can reach reach good goodrecoveries recoveries (>90%). (>90%). The The high high recoveries recoveries are are due due to chemisorptions ofof NaOL NaOL on on the the magnesite magnesite and and dolomitedolomite surface surface [20 [20].]. The The results results in Figurein Figure3 indicate 3 indicate that that NaOL NaOL has has a strong a strong collecting collecting powder powder on on bothboth magnesite magnesite and and dolomite, dolomite, so so it isit difficultis difficult to achieveto achieve the the flotation flotation separation separation of magnesiteof magnesite from from dolomitedolomite without without adding adding any any depressant. depressant. A NaOLA NaOL dosage dosage of 120of 120 mg/L mg/L was was preferred preferred for for all all of theof the otherother flotation flotation tests. tests. The effect of pH on the floatability of magnesite and dolomite in the presence of SH is shown in Figure4. Figure4 illustrates that SH has a significant depression effect on dolomite in the pH range of 7–12. The recovery of dolomite is below 50% in the alkaline pH range. Different from dolomite, the magnesite flotation is not obviously influenced by SH addition at alkaline conditions and the recovery of magnesite is still up to 85%. Therefore, it is possible to separate magnesite from dolomite at alkaline conditions by using SH as the depressant.

Minerals 2017, 7, 150 4 of 11

Minerals 2017, 7, 150 4 of 11 Minerals 2017, 7, 150 4 of 11

Figure 3. Effect of NaOL dosage on the flotation recovery of magnesite and dolomite.

The effect of pH on the floatability of magnesite and dolomite in the presence of SH is shown in Figure 4. Figure 4 illustrates that SH has a significant depression effect on dolomite in the pH range of 7–12. The recovery of dolomite is below 50% in the alkaline pH range. Different from dolomite, the magnesite flotation is not obviously influenced by SH addition at alkaline conditions and the recovery of magnesite is still up to 85%. Therefore, it is possible to separate magnesite from dolomite at alkaline conditionsFigureFigure 3. EffectEffect by using of of NaOL NaOL SH dosage dosageas the ondepressant. the flotation flotation recovery of magnesite and dolomite.

The effect of pH on the floatability of magnesite and dolomite in the presence of SH is shown in Figure 4. Figure 4 illustrates that SH has a significant depression effect on dolomite in the pH range of 7–12. The recovery of dolomite is below 50% in the alkaline pH range. Different from dolomite, the magnesite flotation is not obviously influenced by SH addition at alkaline conditions and the recovery of magnesite is still up to 85%. Therefore, it is possible to separate magnesite from dolomite at alkaline conditions by using SH as the depressant.

Figure 4. Effect of pH on the flotation recovery of magnesite and dolomite using sodium Figure 4. Effect of pH on the flotation recovery of magnesite and dolomite using sodium hexametaphosphate hexametaphosphate (SH) as the depressant (C(NaOL) = 120 mg/L; C(SH) = 20 mg/L). (SH) as the depressant (C(NaOL) = 120 mg/L; C(SH) = 20 mg/L). In order to investigate whether SH is suitable for separating dolomite from magnesite, the samples of magnesiteIn order andto investigate dolomite were whether mixed SH by is a masssuitable ratio for of separating 1:1 and floated dolomite at pH 10.5from using magnesite, NaOL andthe samplesSH as collector of magnesite and depressant, and dolomite respectively, were mixed and theby a results mass areratio presented of 1:1 and in floated Table2. at In pH the 10.5 artificially using NaOL and SH as collector and depressant, respectively, and the results are presented in Table 2. In mixed sample, the original grades of magnesite and dolomite are both 50%. The results show that the thegrades artificially of magnesite mixed sample, and dolomite the original in the grades concentrate of magnesite are only and slightly dolomite different are both from 50%. those The in results mixed Figure 4. Effect of pH on the flotation recovery of magnesite and dolomite using sodium hexametaphosphate showsample. that Therefore, the grades it of is magnesite unfavorable and to dolomite separate magnesitein the concentrate from dolomite are only when slightly SH different is used asfrom the (SH) as the depressant (C(NaOL) = 120 mg/L; C(SH) = 20 mg/L). thosedepressant. in mixed With sample. an increase Therefore, in the it dosage is unfavorable of SH, the to recovery separate of magnesite magnesite from decreases dolomite from when a maximum SH is usedrecovery as the of depressant. 97.39% to 61.35%. With an It indicatesincrease in that the SH dosage has a depressingof SH, the recovery effect on of magnesite magnesite in artificiallydecreases In order to investigate whether SH is suitable for separating dolomite from magnesite, the frommixed a minerals.maximum Additionally, recovery of SH 97.39% also has to a61.35%. good inhibitory It indicates effect that on SH dolomite has a flotation.depressing The effect recovery on samples of magnesite and dolomite were mixed by a mass ratio of 1:1 and floated at pH 10.5 using magnesiteof dolomite in decreases artificially from mixed 97.29% minerals. to 53.07% Additionally, with the dosage SH also of SHhas increasing a good inhibitory from 0 to effect 40 mg/L. on NaOL and SH as collector and depressant, respectively, and the results are presented in Table 2. In dolomiteThe results flotation. are not inThe agreement recovery withof dolomite the results decrea of singleses from mineral 97.29% flotation. to 53.07% Therefore, with the it dosage demonstrates of SH the artificially mixed sample, the original grades of magnesite and dolomite are both 50%. The results increasingthat there arefrom some 0 to other 40 mg/L. factors The that results affect theare flotationnot in agreement separation with of magnesite the results from of dolomitesingle mineral in the show that the grades of magnesite and dolomite in the concentrate are only slightly different from mixed mineral slurry. those in mixed sample. Therefore, it is unfavorable to separate magnesite from dolomite when SH is used as the depressant. With an increase in the dosage of SH, the recovery of magnesite decreases from a maximum recovery of 97.39% to 61.35%. It indicates that SH has a depressing effect on magnesite in artificially mixed minerals. Additionally, SH also has a good inhibitory effect on dolomite flotation. The recovery of dolomite decreases from 97.29% to 53.07% with the dosage of SH increasing from 0 to 40 mg/L. The results are not in agreement with the results of single mineral

Minerals 2017, 7, 150 5 of 11 flotation. Therefore, it demonstrates that there are some other factors that affect the flotation separationMinerals 2017 ,of7, 150magnesite from dolomite in the mixed mineral slurry. 5 of 11

TableTable 2. Results 2. Results of artificial of artificial mixed mixed minerals minerals flotation flotation of magnesitemagnesite from from dolomite dolomite (c(NaOL) (c(NaOL) = 120 = mg/L;120 mg/L; pH = 10.5).pH = 10.5).

Weight MgCO3 SH Dosage MgCO3 CaMg(CO3)2 CaMg(CO3)2 SH Dosage Product WeightRecovery MgCO3 RecoveryMgCO 3 CaMg(CO3)2 CaMg(CO3)2 (mg/L) Product Grade (%) Grade (%) Recovery (%) (mg/L) Recovery(%) (%) Grade (%) Recovery(%) (%) Grade (%) Recovery (%) concentrateconcentrate 97.4597.45 49.85 97.39 97.39 50.15 50.15 97.46 97.46 0 0 tailingtailing 2.552.55 50.19 2.61 2.61 49.81 49.81 2.54 2.54 rawraw ore 100 100 49.88 100 100 50.14 50.14 100 100 concentrateconcentrate 79.1479.14 48.89 75.97 75.97 51.11 51.11 80.79 80.79 2020 tailingtailing 20.8620.86 53.89 24.03 24.03 46.11 46.11 19.21 19.21 rawraw ore ore 100 100 50.93 100 100 50.06 50.06 100 100 concentrateconcentrate 59.1359.13 52.54 61.35 61.35 47.46 47.46 56.85 56.85 4040 tailingtailing 40.8740.87 47.89 38.65 38.65 52.11 52.11 43.15 43.15 rawraw ore ore 100 100 50.64 100 100 49.36 49.36 100 100

2+ 3.2.3.2. Influence Influence of of Ca Ca2+ onon the the Flotation Flotation of of Magnesite Magnesite and and Its Its Elimination Elimination BothBoth magnesite magnesite and and dolomite dolomite are are salt-type salt-type minera minerals,ls, which which are are commonly commonly characterized characterized by by their their highhigh solubility in in water, water, and the extent of of dissolution dissolution in in salt-type salt-type mineral mineral systems systems is is remarkably remarkably higher higher thanthan that in mostmost oxide/silicateoxide/silicate systems. systems. Therefore, Therefore, in in addition addition to to flotation flotation reagents, the dissolved speciesspecies from from minerals will will also affe affectct the flotation flotation behaviors of minerals in flotation flotation pulp. The The above experimentalexperimental results results show show that that the the inhibition inhibition effects effects of SH of SH on onmagnes magnesiteite and and dolomite dolomite are different, are different, but itbut is itdifficult is difficult to realize to realize the theflotation flotation separation separation of ofmagnesite magnesite from from dolomite dolomite in in a amixed mixed slurry slurry system. system. 2+ 2+ The main reason may be thatthat aa certaincertain amountamount ofof MgMg2+ and Ca2+ dissolveddissolved from from the the minerals in in the process of of grinding grinding and and influenced influenced the the flotatio flotationn separation separation of of magnesite magnesite from from dolomite dolomite [21–23]. [21–23 ]. 2+ 2+ TheThe following experiments were carried out to determinedetermine thethe effectseffects ofof MgMg2+ and Ca2+ onon the flotationflotation separation of of magnesite from from dolomite, which which is shown in Figure 55.. TheThe chemicalchemical analysisanalysis 2+ 2+ resultsresults showed that the dissolved concentrationsconcentrations ofof CaCa2+ and Mg2+inin the the pulp pulp of of a a magnesite–dolomite magnesite–dolomite 2+ 2+ mixturemixture were about 2020 mg/L.mg/L. Thus,Thus, the additionaddition quantitiesquantities ofof CaCa2+ andand MgMg2+ areare 20 mg/L in in the followingfollowing experiments. ItIt cancan bebe seenseen fromfrom Figure Figure5 5that that the the floatability floatability of of magnesite magnesite changes changes slightly slightly in 2+ inthe the presence presence of Mg of 2+Mg. The. The recovery recovery of magnesite of magnesite is above is above 85% in 85% the alkalinein the alkaline pH range, pH which range, suggests which 2+ suggeststhat Mg2+ thathas Mg little has effect little on theeffect flotation on the of flotation magnesite. of magn On theesite. contrary, On the the contrary, recovery the of magnesiterecovery of is 2+ 2+ magnesitebelow 55% is in below the alkaline 55% in the pH alkaline range in pH the range presence in the of presence Ca2+. It demonstratesof Ca . It demonstrates that Ca2+ thathas Ca a strong has aprohibitive strong prohibitive impact on impact magnesite on magnesite in the pH in range the pH of range 7–12 using of 7–12 SH using as depressant. SH as depressant.

2+ 2+ FigureFigure 5. Effect 5. of EffectpH on ofthe pH flotation on the of flotationmagnesite of in magnesitethe presence in of the Mg presence2+ and Ca2+ of (C(NaOL) Mg and = 120 Ca mg/L; C(SH)(C(NaOL) = 20 mg/L). = 120 mg/L; C(SH) = 20 mg/L).

Minerals 2017, 7, 150 6 of 11

Minerals 2017, 7, 150 6 of 11 According to relevant research on the mechanism of the influence of metal ions on carbonate minerals’ flotation, the reason that calcium ions influence magnesite flotation may be that calcium ions adsorbingAccording on to a relevant magnesite research surface on in the some mechanism form of ofcalcium the influence compounds of metal to change ions on the carbonate surface propertiesminerals’ flotation, of magnesite, the reason so that that the calcium SH can ions act influence on the magnesite magnesite [24,25]. flotation In may order be thatto eliminate calcium ionsthe adverseadsorbing effect on a of magnesite Ca2+ on the surface flotation in some of formmagnesite, of calcium the adsorption compounds of to Ca change2+ on thea magnesite surface properties surface shouldof magnesite, be prevented. so thatthe SH can act on the magnesite [24,25]. In order to eliminate the adverse effect of Ca2+SHon theis widely flotation used of magnesite,as an inhibitor the adsorption in the carb ofonate Ca2+ onflotation a magnesite process. surface Generally should speaking, be prevented. the mechanismSH is widelyof SH inhibiting used as carbonate an inhibitor flotation in the is carbonate that SH can flotation adsorb on process. the surface Generally of minerals speaking, and enhancethe mechanism the hydrophilic of SH inhibiting property carbonate of minerals. flotation Thus, iscollectors that SH cannot can adsorb act on on the the minerals’ surface of surface minerals to achieveand enhance the effect the hydrophilic of inhibition property [26,27]. ofHowever, minerals. SH Thus, is not collectors only a depressant cannot act on but the also minerals’ a complexing surface agent,to achieve which the can effect react of inhibitionstrongly with [26,27 Ca]. However,2+ and produce SH is nota complex only a depressant with high butwater also solubility a complexing and stability.agent, which On the can basis react of stronglythe complexing with Ca ability2+ and of produce SH, the afollowing complex reaction with high could water occur solubility [28–30]: and stability. On the basis of the complexing ability of SH, the following reaction could occur [28–30]: (NaPO3)6 ⇔ Na4P6O182− + 2Na+ (1)

2− + (NaPO3)6 ⇔ Na4P6O18 + 2Na (1) Na4P6O182− + Ca2 + ⇔ CaNa4P6O18 (2)

2− 2+ 2 + In order to prevent the adsorptionNa4P6O of18 Ca+ on Ca the⇔ surfaceCaNa of4P 6magnesite,O18 the addition order of the(2) flotation test was changed based on the complex characteristic of SH. SH was added before Na2CO3 In order to prevent the adsorption of Ca2+ on the surface of magnesite, the addition order of the to avoid the adsorption of Ca2+ onto the magnesite surface in the form of compounds. The new flotation test was changed based on the complex characteristic of SH. SH was added before Na CO to addition order is shown in Figure 2. Under the new addition order, the effect of dosage of2 SH 3on avoid the adsorption of Ca2+ onto the magnesite surface in the form of compounds. The new addition flotation recovery in the presence of Ca2+ is shown in Figure 6. It is evident from Figure 6 that the order is shown in Figure2. Under the new addition order, the effect of dosage of SH on flotation flotation recovery of magnesite is above 85% over the entire tested dosage of SH, and the recovery of recovery in the presence of Ca2+ is shown in Figure6. It is evident from Figure6 that the flotation dolomite decreases rapidly with the increase of SH dosage. The results illustrate that adding SH recovery of magnesite is above 85% over the entire tested dosage of SH, and the recovery of dolomite before Na2CO3 can eliminate the adverse effect of Ca2+ on magnesite flotation and has no effect on decreases rapidly with the increase of SH dosage. The results illustrate that adding SH before Na CO dolomite flotation. Thus, the new adding order is beneficial to achieve the flotation separation2 of3 can eliminate the adverse effect of Ca2+ on magnesite flotation and has no effect on dolomite flotation. magnesite from dolomite. Thus, the new adding order is beneficial to achieve the flotation separation of magnesite from dolomite.

Figure 6. Effect of dosage of SH on the flotation recovery of minerals in the presence of Ca2+ under the Figurenew 6. Effect addition of dosage order. (C(NaOL)of SH on =the 120 flotation mg/L; C(SH)recovery = 20 of mg/L; minerals pH in = 10.5).the presence of Ca2+ under the new addition order. (C(NaOL) = 120 mg/L; C(SH) = 20 mg/L; pH = 10.5). The pure mineral flotation results suggest that it may be possible to float magnesite away from dolomiteThe pure by adding mineral SH flotation before results Na2CO suggest3. The flotationthat it may separation be possible tests to offloat the magnesite magnesite–dolomite away from dolomitemixture mineralby adding (mass SH ratio before is 1:1)Na2 wereCO3. studiedThe flotation at pH separation 10.5 under tests the newof the addition magnesite–dolomite order and the mixtureresults aremineral shown (mass in Figure ratio 7is. 1:1) It can were be studied seen from at pH Figure 10.57 underthat with the new the SHaddition dosage order increasing and the resultsfrom 0 are to 60 shown mg/L in, theFigure magnesite 7. It can grade be seen increases from Figure from 49.85%7 that with to 85.78% the SH and dosage the recovery increasing decreases from 0 tofrom 60 mg/L, 97.39% the to magnesite 51.57%. Bygrade contrast, increases the from dolomite 49.85% grade to 85.78% decreases and fromthe recovery 50.15% todecreases 13.22% from with 97.39%an increase to 51.57%. in the By dosage contrast, of the SH dolomite and the grade flotation decreases recovery from of 50.15% dolomite to 13.22% decreases with from an increase 97.46%

Minerals 2017, 7, 150 7 of 11 Minerals 2017, 7, 150 7 of 11 in the dosage of SH and the flotation recovery of dolomite decreases from 97.46% to 7.95%. It indicates that, with the change in the adding sequence of SH and Na2CO3, the grades of magnesite and dolomite to 7.95%. It indicates that, with the change in the adding sequence of SH and Na2CO3, the grades of have different change trends and the recovery of magnesite has declined. So, magnesite and dolomite magnesite and dolomite have different change trends and the recovery of magnesite has declined. So, can be selectively separated. According to the indicators of flotation, such as the dosage of SH, grade, magnesite and dolomite can be selectively separated. According to the indicators of flotation, such as and flotation recovery, the optimum dosage of SH is 20 mg/L. the dosage of SH, grade, and flotation recovery, the optimum dosage of SH is 20 mg/L.

FigureFigure 7. Flotation 7. Flotation separation separation results results of the ofmagnesite– the magnesite–dolomitedolomite manual manual mixture mixture mineral mineral under the under new the addingnew order adding (C(NaOL) order (C(NaOL)= 120 mg/L; = 120pH = mg/L; 10.5). pH = 10.5).

2+ 3.3. Mechanism Mechanism of Ca 2+ AffectingAffecting the the Flotation Flotation of of Magnesite Magnesite

3.3.1. XPS XPS Analysis Analysis Previous studies have shown that the dissolved species from one mineral frequently undergoes hydrolysis or chemical reaction with another mine mineralral surface surface in in a a salt-type salt-type mineral mineral flotation flotation system, system, leadingleading to to changes of mineral surface properties. Additionally, Additionally, these these changes changes make make it it difficult difficult to separate such minerals without additional treatment [[31–33].31–33]. InIn order order to to study study the the mechanism mechanism of of calcium calcium ions ions affecting affecting the the flotation flotation of of magnesite magnesite and and different different reagent addition orders influencing influencing the flotation flotation se separationparation of magnesite from dolomite, XPS analysis analysis and zeta potentialpotential measurementsmeasurements were were conducted. conducted. The The XPS XPS analysis analysis ofmagnesite of magnesite samples samples treated treated with withdifferent different addition addition orders orders is shown is shown at Figure at Figure8. According 8. According to the to obtainedthe obtained results, results, XPS XPS is not is not able able to 2+ todetect detect surface surface Ca Ca on theon the natural natural magnesite magnesite sample sample (Line (Line 1). In 1). the In presence the presence of Ca of, Ca the2+ peaks, the peaks of Ca2p of Ca2pand Ca2s and appearCa2s appear on the surfaceon the surface of the magnesite of the magnesite under the under common the reagentcommon addition reagent order, addition suggesting order, suggestingthat Ca species that can Ca precipitate species can on theprecipitate magnesite on surface, the magnesite which confirms surface, the whic earlierh confirms assumption the (Lineearlier 2). When changing the adding order (adding SH before Na CO ), no Ca peaks are observed on the spectra assumption (Line 2). When changing the adding order2 (adding3 SH before Na2CO3), no Ca peaks are 2+ observedof magnesite, on the illustrating spectra of that magnesite, SH can complex illustrating with that Ca SHinto can solution complex and with prevent Ca2+ into the adsorptionsolution and of preventCa species the on adsorption a magnesite of Ca surface species (Line on 3).a magnesite surface (Line 3).

Minerals 2017, 7, 150 8 of 11 Minerals 2017, 7, 150 8 of 11

FigureFigure 8. XPS 8. XPSspectra spectra of ofmagnesite magnesite at atpH pH 10.5 10.5 in thethepresence presence of of Ca Ca2+:2+ (1): (1) natural natural magnesite magnesite sample; sample; (2) magnesite(2) magnesite sample sample under under the common the common reagent reagent addition addition order; order; (3) magnesite (3) magnesite sample sample under under the new the new reagent additionreagent order addition (C(Ca2+) order = 20 mg/L, (C(Ca 2+C(SH)) = 20 = mg/L,20 mg/L). C(SH) = 20 mg/L).

The detaileddetailed valuesvalues of of binding binding energy energy and and offset offs ofet each of each composition composition element element of magnesite of magnesite treated treatedwith different with different adding sequences adding sequences are shown are in shown Table3. in It canTable be 3. observed It can be from observed Table3 thatfrom the Table Ca2p3/2 3 that thepeak Ca2p3/2 and Ca2s peak peak and areCa2s located peak are at located 346.8 eV at and346.8 438 eV eV,and respectively, 438 eV, respectively, attributed attributed to CaCO to3 CaCO[34,35].3 2+ 2+ [34,35].The data The demonstrate data demonstrate that Ca thatadsorbs Ca onadsorbs a magnesite on a magnesite surface in surface the form in ofthe CaCO form3 .of The CaCO formation3. The 2+ 2+2 − 2− formationof CaCO3 mayof CaCO be due3 may tothe be due reaction to the of reaction Ca with of COCa 3 with, which CO3 are, which from are Na2 fromCO3 orNa the2CO ones3 or the exposed ones exposedon a magnesite on a magnesite surface. Whatsurface. is more,What theis more, Mg1s, th C1se Mg1s, and O1sC1s bandsand O1s have bands strong have shifts strong by 0.49,shifts 0.38, by 0.49,and 0.520.38, eV, and respectively, 0.52 eV, respectively which illustrates，which that illustrates CaCO3 mightthat CaCO be strongly3 might(or be chemically)strongly (or adsorbedchemically) on adsorbeda magnesite on surfacea magnesite under surface the common under the adding common order. adding Due toorder. the adsorption Due to the ofadsorption CaCO3, the of CaCO surface3, theproperties surface ofproperties magnesite ofhave magnesite changed. have Therefore, changed. SHTherefore, can depress SH can its floatability.depress its floatability. Compared with the natural magnesite sample, the position of the Mg1s and C1s bands remain unchangedTable 3. underBinding the energy new reagent and offset addition of each compositionorder in the elementpresence of of magnesite Ca2+, and samples the O1s treated band with shifts to a valuedifferent of 0.13 medicine eV, which addition is smaller orders. than the instrument error of 0.3 eV. It illustrates that changing the adding order can prevent the adsorption of Ca species on a magnesite surface, so there is no change of magnesite Samplesurface properties. Binding Energy/eV Atomic orbital Mg1s O1s C1s Ca2s Ca2p Table 3. BindingMagnesite energy and offset of each 1304.9 composition 530.5 element of magnesite 283.1 samples/ treated with different / 2+ medicineMagnesite addition + Ca orders.+ Na2CO 3 + SH 1305.39 530.88 283.62 438.0 346.8 Offset/eV 0.49 0.38 0.52 / / MagnesiteSample 1304.9 530.5Binding 283.1 Energy/eV / / 2+ MagnesiteAtomic + Ca + orbital SH + Na 2CO3 1304.9 Mg1s 530.37 O1s 283.1 C1s Ca2s / Ca2p / Offset/eVMagnesite 1304.9 0 −0.13530.5 283.1 0 / / / Magnesite +Ca2+ + Na2CO3 + SH 1305.39 530.88 283.62 438.0 346.8 ComparedOffset/eV with the natural magnesite0.49 sample, the0.38 position of0.52 the Mg1s and/ C1s bands/ remain unchanged underMagnesite the new reagent addition1304.9 order in the530.5 presence of283.1 Ca2+ , and the/ O1s band shifts/ to a valueMagnesite of 0.13 eV,+Ca which2+ + SH is + smallerNa2CO3 than1304.9 the instrument 530.37 error of 0.3 283.1 eV. It illustrates / that changing / the adding orderOffset/eV can prevent the adsorption of0 Ca species− on0.13 a magnesite0 surface, so / there is no change / of magnesite surface properties. 3.3.2. Zeta Potential Measurements 3.3.2. Zeta Potential Measurements Figure 9 shows the zeta potential of magnesite particles treated with different reagent addition ordersFigure in the9 presenceshows the of zetaCa2+ potential. Line 1 and of line magnesite 2 are the particles zeta pote treatedntials of with magnesite different under reagent the additioncommon addingorders in order the presence before and of Caafter2+. adding Line 1 and NaOL, line 2respectively. are the zeta potentialsLine 3 and of line magnesite 4 are the under zeta potentials the common of magnesiteadding order under before the andnew afteradding adding order NaOL, before respectively.and after adding Line NaOL, 3 and linerespectively. 4 are the zetaIt is clear potentials from Figure 9 that the negative shift value of line 1 is larger than that of line 3. The adsorption of Ca species on the magnesite surface makes SH adsorb on the surface of the magnesite under the common adding

Minerals 2017, 7, 150 9 of 11 of magnesite under the new adding order before and after adding NaOL, respectively. It is clear

Mineralsfrom Figure 2017, 79, 150that the negative shift value of line 1 is larger than that of line 3. The adsorption of9 of Ca 11 species on the magnesite surface makes SH adsorb on the surface of the magnesite under the common 2+ orderadding (Line order 1). (Line When 1). adding When SH adding prior SH to priorNa2CO to3, Na SH2CO can3 ,complex SH can complex Ca2+ into Ca solution,into solution,so it will sonot it adsorbwill not on adsorb the magnesite on the magnesite surface (Line surface 3). After (Line addi 3). Afterng NaOL, adding there NaOL, is almost there no is change almost in no the change zeta potentialin the zeta of potential magnesite of under magnesite the common under the addition common order addition (Line order 1, Line (Line 2). NaOL 1, Line cannot 2). NaOL adsorb cannot on magnesiteadsorb on surface magnesite because surface of the because adsorption of the of adsorption SH. When of changing SH. When the changingadding order, the adding the negative order, shiftthe negative value of shift line value4 is much of line more 4 is muchthan that more of thanline that3. Besides, of line 3.the Besides, stronger the the stronger alkalinity the of alkalinity the pulp, of thethe higher pulp, the the higherdegree the of negative degree of shift. negative It illustra shift.tes It that illustrates NaOL can that adsorb NaOL on can a adsorbmagnesite on asurface magnesite and makesurface magnesite and make float magnesite by adding float SH by before adding Na SH2CO before3. Furthermore, Na2CO3. Furthermore, the larger the the pH larger value, the the pH higher value, thethe floating higher therate floating of magnesite. rate of magnesite.The above Theresults above of zeta results potentials of zeta potentialsare consistent are consistent with the flotation with the testflotation results. test In results. a word, In the a word,new reagent the new additi reagenton order addition can eliminate order can the eliminate adverse the effect adverse of Ca2+ effect on the of flotationCa2+ on theof magnesite flotation of and magnesite promote and the promote adsorption the adsorptionof NaOL on of the NaOL surface on theof magnesite. surface of magnesite.Thus, the flotationThus, the separation flotation separation of magnesite of magnesite from dolomite from can dolomite be realized. can be realized.

2+ FigureFigure 9. Effect 9. Effect of different of different adding adding orders orders on zeta on zetapotential potential of magnesite of magnesite particles particles in the in presence the presence of Ca : (1) MgCO3 +2+ Ca2+ + Na2CO3 + SH; 2+(2) MgCO3 + Ca2+ + Na2CO3 + SH + NaOL;2+ (3) MgCO3 + Ca2+ + SH + Na2CO3; of Ca :(1) MgCO3 + Ca + Na2CO3 + SH; (2) MgCO3 + Ca + Na2CO3 + SH + NaOL; (4) MgCO3 + Ca2+ + SH 2++ Na2CO3 + NaOL (C(Ca2+) = 20 mg/L;2+ C(SH) = 20 mg/L; C(NaOL) =1202+ mg/L). (3) MgCO3 + Ca + SH + Na2CO3;(4) MgCO3 + Ca + SH + Na2CO3 + NaOL (C(Ca ) = 20 mg/L; C(SH) = 20 mg/L; C(NaOL) =120 mg/L). 4. Conclusions 4. ConclusionsFrom the above results, the following conclusions can be drawn: 1.From SH inhibits the above dolomite results, flotation the following and has conclusions little effect can on bethe drawn: flotation of magnesite in pure mineral flotation1. SH tests; inhibits however, dolomite SH flotationhad no selective and has littleinhibition effect in on the the magnesite–dolomite flotation of magnesite manual in pure mixture mineral mineralflotation flotation tests; however, tests. It has SH hada strong no selective inhibition inhibition on both magnesite in the magnesite–dolomite and dolomite. manual mixture 2+ 2+ mineral2. Ca flotation dissolved tests. from It has dolomite a strong makes inhibition magnesite on both be magnesite inhibited and by SH. dolomite. Mg dissolved from the minerals2. Ca has2+ dissolvedno obvious from effect dolomite on the flotation makes magnesite of magnesite. be inhibited by SH. Mg2+ dissolved from the 2+ minerals3. When has nochanging obvious the effect reagent on the addition flotation order, of magnesite. adding SH prior to Na2CO3, the effect of Ca on magnesite flotation can be eliminated and the flotation separation of magnesite from dolomite can2+ be 3. When changing the reagent addition order, adding SH prior to Na2CO3, the effect of Ca on achieved.magnesite flotation can be eliminated and the flotation separation of magnesite from dolomite can 2+ be achieved.4. XPS analysis results and zeta potential measurements show that the adsorption of Ca in the form 4.of XPSCaCO analysis3 on the results magnesite and zetasurface potential makes measurements the magnesite’s show surface that properties the adsorption change. of CaSo,2+ SHin can the adsorb on the surface of magnesite to depress magnesite floating. The new addition order, adding form of CaCO3 on the magnesite surface makes the magnesite’s surface properties change. So, SH can 2+ 2+ SHadsorb prior on to the Na surface2CO3, can of complex magnesite Ca to depressinto flotation magnesite solution floating. and prevent The new Ca addition from adsorbing order, adding on a magnesite surface. Additionally, the flotation2+ separation of magnesite away from2+ dolomite can be SH prior to Na2CO3, can complex Ca into flotation solution and prevent Ca from adsorbing on achieved.

Acknowledgments: This project was supported by the National Natural Science Foundation of China (grant number 51474054); the Fundamental Research Funds for the Central Universities (grant number N130301003); Doctoral Scientific Research Foundation of Liaoning Province, China (grant number 201601031); the Fundamental Research Funds for the Central Universities (grant number N150101001).

Minerals 2017, 7, 150 10 of 11 a magnesite surface. Additionally, the flotation separation of magnesite away from dolomite can be achieved.

Acknowledgments: This project was supported by the National Natural Science Foundation of China (grant number 51474054); the Fundamental Research Funds for the Central Universities (grant number N130301003); Doctoral Scientific Research Foundation of Liaoning Province, China (grant number 201601031); the Fundamental Research Funds for the Central Universities (grant number N150101001). Author Contributions: Na Luo and Dezhou Wei conceived the project and designed the experiments; Na Luo, Caie Zhang, and Cong Han conducted parts of the experiments and analyzed the data. Na Luo and Yanbai Shen wrote the paper. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Karaoglu, H.; Yanmis, D.; Gurkok, S. Magnesite enrichment with pseudomonas oryzihabitans isolated from magnesite ore. Geomicrobiol. J. 2016, 33, 46–51. [CrossRef] 2. Yao, J.; Yin, W.; Gong, E. Depressing effect of fine hydrophilic particles on magnesite reverse flotation. Int. J. Miner. Proces. 2016, 149, 84–93. [CrossRef] 3. Chen, G.; Tao, D. Effect of solution chemistry on flotability of magnesite and dolomite. Int. J. Miner. Process. 2004, 74, 343–357. [CrossRef] 4. Botero, A.E.C.; Torem, M.L.L.; De Mesquita, M.S. Fundamental studies of Rhodococcus opacus as a biocollector of calcite and magnesite. Miner. Eng. 2007, 20, 1026–1032. [CrossRef] 5. Marouf, R.; Marouf-Khelifa, K.; Schott, J.; Khelifa, A. Zeta potential study of thermally treated dolomite samples in electrolyte solutions. Microporous Mesoporous Mater. 2009, 122, 99–104. [CrossRef] 6. Liu, Y.; Liu, Q. Flotation separation of carbonate from sulfide minerals, I: Flotation of single minerals and mineral mixtures. Miner. Eng. 2004, 17, 855–863. [CrossRef] 7. Liu, Y.; Liu, Q. Flotation separation of carbonate from sulfide minerals, II: Mechanisms of flotation depression of sulfide minerals by thioglycollic acid and citric acid. Miner. Eng. 2004, 17, 865–878. [CrossRef] 8. Luo, X.; Yin, W.; Wang, Y.; Sun, C.; Ma, Y.; Liu, J. Effect and mechanism of dolomite with different size fractions on flotation using sodium oleate as collector. J. Cent. South Univ. 2016, 23, 529–534. [CrossRef] 9. Sl´ áczka, A.S.; Paprotny, J. Flocculation of smithsonite and dolomite using polymers containing nitrogen atoms. Int. J. Miner. Process. 1985, 14, 319–325. [CrossRef]

10. Santillán, J.; Williams, Q. A high-pressure infrared and X-ray study of FeCO3 and MnCO3: Comparison with CaMg(CO3)2-dolomite. Phys. Planet. Inter. 2004, 143–144, 291–304. [CrossRef] 11. Gence, N.; Ozbay, N. pH dependence of electrokinetic behavior of dolomite and magnesite in aqueous electrolyte solutions. Appl. Surf. Sci. 2006, 252, 8057–8061. [CrossRef] 12. Gence, N. Wetting behavior of magnesite and dolomite surfaces. Appl. Surf. Sci. 2006, 252, 3744–3750. [CrossRef] 13. Hu, Y.; Chi, R.; Xu, Z. Solution chemistry study of salt-type mineral flotation systems: Role of inorganic dispersants. Ind. Eng. Chem. Res. 2003, 42, 1641–1647. [CrossRef] 14. Pokrovsky, O.S.; Golubev, S.V.; Schott, J.; Castillo, A. Calcite, dolomite and magnesite dissolution kinetics in ◦ aqueous solutions at acid to circumnetral pH, 25 to 150 C and 1 to 55 atm pCO2: New constrains on CO2 sequestration in sedimentary basins. Chem. Geol. 2009, 265, 20–32. [CrossRef] 15. Van Cappellen, P.; Charlet, L.; Stumm, W.; Wersin, P.A. Surface complexation model of the -aqueous solution interface. Geochim. Cosmochim. Acta 1993, 57, 3505–3518. [CrossRef] 16. Vuˇcini´c, D.R.; Radulovi´c, D.S.; Deuši´c, S.Đ. Electrokinetic properties of hydroxyapatite under flotation conditions. J. Colloid Interface Sci. 2010, 343, 239–245. [CrossRef][PubMed] 17. Chen, G.L.; Tao, D. Reverse flotation of magnesite by dodecyl phosphate from dolomite in the presence of sodium silicate. Sep. Sci. Technol. 2004, 34, 377–390. [CrossRef] 18. Luo, X.M.; Wang, Y.F.; Wen, S.M.; Ma, M.Z.; Sun, C.Y. Effect of carbonate minerals on quartz flotation behavior under conditions of reverse anionic flotation of . Int. J. Miner. Process. 2016, 152, 1–6. [CrossRef] Minerals 2017, 7, 150 11 of 11

19. Zhu, Y.; Luo, B.; Sun, C.; Li, Y.; Han, Y. Influence of bromine modification on collecting property of lauric acid. Miner. Eng. 2015, 79, 24–30. [CrossRef] 20. Rao, K.H.; Antti, B.M.; Forssberg, E. Mechanism of oleate interaction on salt-type minerals, part II. Adsorption and electrokinetic studies of apatite in the presence of sodium oleate and sodium metasilicate. Int. J. Miner. Process. 1990, 28, 59–79. [CrossRef] 21. Gautelier, M.; Schott, J.; Oelkers, E.H. An experimental study of dolomite dissolution rates at 80 ◦C as a function of chemical affinity and solution composition. Chem. Geol. 2007, 242, 509–517. [CrossRef] 22. Liu, A.; Ni, W.; Wu, W. Mechanism of separating pyrite and dolomite by flotation. J. Univ. Sci. Technol. B. 2007, 14, 291–296. [CrossRef] 23. Gence, N.; Özdag,ˇ H. Surface properties of magnesite and adsorption mechanism. Int. J. Miner. Process. 1995, 43, 37–47. [CrossRef] 24. Chen, Z.; Ren, Z.; Gao, H.; Liu, J.; Jin, J.; Min, F. The effects of calcium ions on the flotation of sillimanite using dodecylammonium chloride. Minerals 2017, 7, 28. [CrossRef] 25. Shi, Q.; Feng, Q.; Zhang, G.; Deng, H. A novel method to improve depressants actions on calcite flotation. Miner. Eng. 2013, 55, 186–189. [CrossRef] 26. Feng, B.; Lu, Y.; Feng, Q.; Zhang, M.; Gu, Y. –serpentine interactions and implications for talc depression. Miner. Eng. 2012, 32, 68–73. [CrossRef] 27. Nia, X.; Liu, Q. Adsorption behaviour of sodium hexametaphosphate on pyrochlore and calcite. Can. Metall. Quart. 2013, 52, 473–478. [CrossRef] 28. Liu, Y.; Zhang, M.; Feng, Q.; Long, T.; Ou, L.; Zhang, G. Effect of sodium hexametaphosphate on separation of serpentine from pyrite. Trans. Nonferrous Met. Soc. China 2011, 21, 208–213. [CrossRef] 29. Ding, H.; Lin, H.; Deng, Y. Depressing effect of sodium hexametaphosphate on apatite in flotation of rutile. J. Univ. Sci. Technol. B. 2007, 14, 200–203. [CrossRef] 30. Larson, C.E. Use of sodium hexametaphosphate as an anticoagulant. Exp. Biol. Med. 1940, 44, 554–555. [CrossRef] 31. Hu, Y.; Xu, J.; Luo, C.; Yuan, C. Solution chemistry studies on dodecylamine flotation of smithsonite/calcite. J. Cent. South Univ. 1995, 26, 589–594. (In Chinese)

32. Kangal, O.; Sirkeci, A.A.; Güney, A. Flotation behaviour of (Mg3Ca(CO3)4) with anionic collectors. Int. J. Miner. Process. 2005, 75, 31–39. [CrossRef] 33. Nunes, A.P.L.; Peres, A.E.C.; De Armando, A.C.; Valadão, G.E.S. Electrokinetic properties of wavellite and its floatability with cationicand anionic collector. J. Colloid Interface Sci. 2011, 361, 632–638. [CrossRef][PubMed] 34. García-Sánchez, A.; Álvarez-Ayuso, E. Sorption of Zn, Cd and Cr on calcite. Application to purification of industrial wastewaters. Miner. Eng. 2002, 15, 539–547. [CrossRef] 35. Demri, B.; Muster, D. XPS study of some calcium compounds. J. Mater. Process. Technol. 1995, 55, 311–314. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).