sign in sign up
<<
Home , Magnesium citrate, Magnesium hydroxide, Magnesium sulfate

Article Differential Precipitation of Mg(OH)2 from CaSO4·2H2O Using Citrate as Inhibitor—A Promising Concept for Reagent Recovery from MgSO4 Waste Streams

Szilveszter Ziegenheim 1,2,Márton Szabados 2,3, Zoltán Kónya 4,5 , Ákos Kukovecz 4, István Pálinkó 2,3 and Pál Sipos 1,2,*

1 Department of Inorganic and Analytical Chemistry, University of Szeged, Dóm tér 7, H-6720 Szeged, Hungary; [email protected] 2 Material and Solution Structure Research Group, Institute of Chemistry, University of Szeged, Aradi Vértanúk tere 1, H-6720 Szeged, Hungary; [email protected] (M.S.); [email protected] (I.P.) 3 Department of Organic Chemistry, University of Szeged, Dóm tér 8, H-6720 Szeged, Hungary 4 Department of Applied and Environmental Chemistry, University of Szeged, Rerrich Béla tér 1, H-6720 Szeged, Hungary; [email protected] (Z.K.); [email protected] (Á.K.) 5 MTA-SZTE Reaction Kinetics and Surface Chemistry Research Group, Rerrich B. tér 1, H-6720 Szeged, Hungary * Correspondence: [email protected]; Tel.: +36-62-54-4045

 Academic Editor: Nikolay Gerasimchuk  Received: 8 September 2020; Accepted: 27 October 2020; Published: 29 October 2020

Abstract: In hydrometallurgical processing and acidic wastewater treatment, one of the neutralizing agents employed is MgO or Mg(OH)2. At the end of this process, the resulting solution, which is rich 2 2+ in SO4 − and Mg is treated with lime to remove (or minimize the amount) of these via the precipitation of Mg(OH) and CaSO 2H O (gypsum). In our work, an attempt was made to separate 2 4· 2 the two solids by increasing the induction time of the gypsum precipitation, thus regenerating relatively pure Mg(OH)2 which could be reused in wastewater treatments or hydrometallurgical processing circuits, and in this way, significantly enhancing the economic viability of the process. During our experiments, the reaction of an MgSO4 solution with milk of lime prepared from quicklime was studied. The effects of a range of organic additives, which can slow down the precipitation of gypsum have been assessed. The process was optimized for the most promising inhibiting agent—that is, the citrate . The reactions were continuously monitored in situ by conductometric measurements with parallel monitoring of solution pH and temperature. ICP-OES measurements were also carried out on samples taken from the reaction slurry. The composition of the precipitating solids at different reaction times was established by powder XRD and their morphology by SEM. Finally, experiments were carried out to locate the additive after the completion of the precipitation reaction to get information about its potential reuse.

Keywords: gypsum; ; inhibition; citric ; precipitation separation

1. Introduction Solid precipitation from supersaturated solutions is a common phenomenon; however, in industrial processes it can cause several nuisances—e.g., via scaling or as a co-precipitating side product. Scaling can reduce the efficiency of heat-exchanging surfaces ( cooling systems), clog membranes and tubes (reverse osmosis units, water injection systems in oil and gas production), while co-precipitating

Molecules 2020, 25, 5012; doi:10.3390/molecules25215012 www.mdpi.com/journal/molecules Molecules 2019, 24, x FOR PEER REVIEW 2 of 16 Molecules 2020, 25, 5012 2 of 16 membranes and tubes (reverse osmosis units, water injection systems in oil and gas production), while co-precipitating solids are present as impurities in the final products. One of the most solidscommonly are present precipitating as impurities solids inis theCaSO final4·2H products.2O, gypsum; One oftherefore, the most the commonly detailed precipitatingunderstanding solids of its is CaSO 2H O, gypsum; therefore, the detailed understanding of its is necessary in a crystallization4· 2 is necessary in a variety of processes, which are adversely affected by its precipitation. varietyAn of important processes, field which where are adverselygypsum is a ffpresentected by in itshigh precipitation. quantities is acidic wastewater treatment. Normally,An important these acidic field where contain gypsum heavy is present metal in and high quantities ions, which is acidic are wastewater precipitated treatment. as metal Normally, these and acidic waters sulfate contain using lime heavy for metal neutralization and sulfate [1]. ions, This which way, arethe precipitatedremaining water as metal can hydroxidesbe released to and natural calcium waters. sulfate However, using lime the for sludge neutralization, which forms, [1]. This requires way, further the remaining treatment water and can must be releasedbe disposed. to natural Using waters.magnesium However, oxide/hydroxide the sludge, whichas the forms,first step requires of a two-st furtherep treatmentneutralization and mustprocess be disposed.can provide Using an alternative, magnesium and oxide this way,/hydroxide gypsum as can the be first precipitated step of a two-stepseparately neutralization from the heavy process metal canhydroxides provide [2,3], an alternative, as shown andin Scheme this way, 1. gypsumThe proce canss could be precipitated also be useful separately in the fromdesalination the heavy process metal hydroxides[4]. [2,3], as shown in Scheme1. The process could also be useful in the desalination process [ 4].

Scheme 1. Two-step neutralization of acidic mine drainagedrainage using magnesium oxideoxide/hydroxide/hydroxide for separate precipitation of metal hydroxides.

In thisthis work, work, an an attempt attempt is made is made to improve to improv the lattere the method latter furthermethod by separatingfurther by theseparating precipitation the ofprecipitation magnesium of hydroxide magnesium and gypsumhydroxide in solutionand gypsumby using in additives.solution by This using has the additives. potential This of enhancing has the thepotential economic of enhancing and environmental the economic viability and environmental of the process, viability and the resultsof the process, can be extended and the toresults separate can gypsumbe extended precipitation to separate in timegypsum from precipitation rapidly precipitating in time from solids rapidly other than precipitating Mg(OH)2. solids other than 11 Mg(OH)The2 . of Mg(OH)2 (Ksp is around 1.38 10− ) is significantly lower than that of CaSO4 5 × (Ksp Theis around solubility 2.5 of 10Mg(OH))[5,62]. (K Thesp is nucleationaround 1.38 of ×Mg(OH) 10−11) is significantlyis also highly lower dependent than that on of theCaSO ion4 × − 2 concentrations,(Ksp is around and2.5 is× very10−5) fast[5,6]. (practically The nucleation instantaneous) of Mg(OH) at high2 is supersaturation also highly dependent [7–10]. The on inhibition the ion ofconcentrations, the crystallization and ofis Mg(OH)very fast2 was(practically found to in bestantaneous) rather difficult at whenhigh supersaturation conventional additives [7–10]. wereThe usedinhibition [10–13 of]. the crystallization of Mg(OH)2 was found to be rather difficult when conventional additivesThe kineticswere used of [10–13]. gypsum precipitation, which is significantly slower than that of Mg(OH)2, was investigatedThe kinetics of in gypsum several studiesprecipitation, [6,14– 18which]. As is itsignificantly is a common slower scaling than and that precipitating of Mg(OH)2 solid,, was theinvestigated inhibition in of several its crystallization studies [6,14–18]. has been theAs focusit is a of common various studiesscaling since and theprecipitating 1950s [19]. solid, The eff theect ofinhibition various of additives its crystallization has been has scrutinized been the overfocus recent of various decades. studies Various since the metal 1950s ions [19]. were The found effect toof influencevarious additives the precipitation has been of gypsum,scrutinized aff ectingover rece bothnt decades. growth Various and morphologymetal ions [were20–23 found], even to in theinfluence presence the of precipitation certain amino of acidsgypsum, [24]. affecting Organic phosphonatesboth crystal growth were also and demonstrated morphology to[20–23], be effi evencient inhibitorsin the presence in gypsum of certain precipitation amino [acids25–28 [24].]. Polymeric Organic compounds phosphonates with were carboxylic also demonstrated groups also proved to be toefficient be effective inhibitors in the in inhibition gypsum ofprecipitation gypsum precipitation—the [25–28]. Polymeric most compounds popular is perhapswith carboxylic polyacrylic groups acid; however,also proved some to be other effective “green” in the alternatives inhibition were of gypsum also effi precipitation—thecient [29–32]. Looking most into popular small- is perhaps simplepolyacrylic carboxylic acid; however, , the some performance other “green” of citric altern acidatives seemed were to bealso remarkable efficient [29–32]. [33] and Looking accordingly into itssmall-molecule inhibiting eff ectssimple were carboxylic investigated acids, in the more performance detail [34–36 of]. seemed to be remarkable [33] and accordinglyIt seems plausible its inhibiting that the effects inhibition were of investigated the precipitation in more of gypsum detail [34–36]. is much more promising than that ofIt magnesiumseems plausible hydroxide. that the Accordingly, inhibition of scouting the precipitation experiments of weregypsum performed is much in more a stoichiometric promising Cathan2+ +thatSO of2 reactionmagnesium at pH hydroxide.7 using a numberAccordingl of additivesy, scouting (for experiments details, see Supplementary were performed Materials, in a 4 − ≈ Appendixstoichiometric A. In Ca this2+ + paper, SO42− the reaction optimization at pH of≈ the7 using reaction a number conditions of toadditives separate (for the precipitateddetails, see magnesiumSupplementary hydroxide Materials, from Appendix gypsum A). in time In this using pape oner, the particular optimization inhibitor of the from reaction the aforementioned conditions to additives—citrate ions—is presented.

Molecules 2019, 24, x FOR PEER REVIEW 3 of 16 separate the precipitated from gypsum in time using one particular inhibitor fromMolecules the2020 aforementioned, 25, 5012 additives—citrate ions—is presented. 3 of 16

2. Results 2. Results

2.1. Effect of the Reaction Conditions on the Kinetics of the MgSO4 + Ca(OH)2 + 2H2O → Mg(OH)2 + 2.1. Effect of the Reaction Conditions on the Kinetics of the MgSO4 + Ca(OH)2 + 2H2O Mg(OH)2 + CaSO4·2H2O Reaction CaSO 2H O Reaction → 4· 2 During our experiments, the stoichiometric reaction of MgSO4 + Ca(OH)2 + 2 H2O → Mg(OH)2 + During our experiments, the stoichiometric reaction of MgSO4 + Ca(OH)2 + 2 H2O Mg(OH)2 CaSO4·2H2O was studied using milk of lime (MoL)—prepared from quicklime—as a Ca(OH)→ 2 source. + CaSO4 2H2O was studied using milk of lime (MoL)—prepared from quicklime—as a Ca(OH)2 The effects· of the reaction conditions and added citric acid—as an inhibitor for gypsum source. The effects of the reaction conditions and added citric acid—as an inhibitor for gypsum precipitation—were investigated. The reactions were monitored in situ with conductometry and the precipitation—were investigated. The reactions were monitored in situ with conductometry and the pH and temperature of the reaction mixture were also recorded. The pH of the initial MgSO4 solution pH and temperature of the reaction mixture were also recorded. The pH of the initial MgSO4 solution was found to be neutral without the additive, but it dropped to around pH ≈ 3 upon citric acid was found to be neutral without the additive, but it dropped to around pH 3 upon citric acid addition; addition; however, it raised to pH ≈ 11.1–11.2 when MoL was added to the≈ mixture, which, according however, it raised to pH 11.1–11.2 when MoL was added to the mixture, which, according to the to the literature, coincides≈ with the nucleation pH of Mg(OH)2 [7]. The variation of solution literature, coincides with the nucleation pH of Mg(OH)2 [7]. The variation of solution conductivity vs. conductivity vs. time during the reactions is shown in Figure 1. time during the reactions is shown in Figure1.

12 in absence of citric acid in presence of 1.5 mM citric acid

10

8 (mS/cm)

κ 6

4

2 0 10203040 t (min)

FigureFigure 1. VariationVariation of solution conductivityconductivity during the reactionreaction ofof MgSOMgSO4 ++ Ca(OH)2 + 22 H 2OO → 4 2 2 → Mg(OH)Mg(OH)2 ++ CaSOCaSO4·2H2H2OO with with identical identical (0.1 (0.1 M) M) initial initial reactant reactant concentrations concentrations in in the the absence absence and and in in the the 2 4· 2 presencepresence of 1.5-mM citric acid (T = 2020 °C;◦C; in in di distilledstilled water water media media without without any any supporting supporting ). electrolyte).

UponUpon adding adding MoL MoL to to the of of MgSO 44, ,a a sharp sharp drop drop in in the the conductivity conductivity of of the the reactionreaction mixture instantaneouslyinstantaneously appeared, appeared, both both in in the the presence presence and and absence absence of the of inhibitor.the inhibitor. This dropThis dropis most is probablymost probably the result the ofresult two eofffects—the two effects—the formation formation of the Mg(OH) of the2 Mg(OH)precipitate2 precipitate and, parallel and, to parallelthis, the dissolutionto this, the of dissolution Ca(OH)2. Thenof Ca(OH) a plateau2. Then can be a observed plateau incan both be systems,observed corresponding in both systems, to the correspondinginduction period to priorthe induction to gypsum period precipitation. prior to Thegypsum smooth precipitation. decrease in theThe conductivity smooth decrease corresponds in the to the precipitation of CaSO 2H O. It can be seen that the addition of citric acid significantly increased conductivity corresponds to4 ·the 2precipitation of CaSO4·2H2O. It can be seen that the addition of citric acidthe induction significantly period increased of gypsum the precipitation, induction period under theof specificgypsum conditions precipitation, shown under above, fromthe specific around conditions3 to 10 min. shown above, from around 3 to 10 min. TheThe induction induction times times of of the the reactions reactions were were es estimatedtimated using using the the data data of of the conductivity measurements.measurements. The The induction induction time time was was estimated estimated as as the the time time required required for for the the conductivity conductivity of of the the slurryslurry to to drop drop by by 0.2 0.2 mS/cm, mS/cm, which which is is twice twice the the st standardandard error error of of the the measuring measuring method. method. In In addition, addition, toto quantify thethe reactionreaction rate rate more more exactly exactly and and comparably, comparably, the the measured measured conductivity conductivity data weredata were fitted fittedwith usingwith using the following the following equation: equation: a b y = b + −  t c 1 + i where a is the initial (after the first drop) and b is the final conductivity of the reaction mixture, t is the reaction time, c is the slope factor and i is the half-reaction time (time needed for half of the reactants to Molecules 2019, 24, x FOR PEER REVIEW 4 of 16

𝑎 − 𝑏 𝑦 = 𝑏 + 𝑡 1 + 𝑖 whereMolecules a2020 is the, 25 ,initial 5012 (after the first drop) and b is the final conductivity of the reaction mixture,4 oft 16is the reaction time, c is the slope factor and i is the half-reaction time (time needed for half of the reactants to react). During the fittings, the initial and final conductivities were held constant at the experimentalreact). During values. the fittings, With thethis initialhalf-reaction and final time conductivities the reaction wererates heldare quantitatively constant at the comparable. experimental values.To Withconfirm this the half-reaction above interpretation time the reaction of the conductivity rates are quantitatively vs. time profiles comparable. shown in Figure 1, the concentrationsTo confirm of the Ca2+ above and Mg interpretation2+ during the of experiment the conductivity with soluti vs.ons time containing profiles shown the inhibitor in Figure were1, determined.the concentrations Samples of Cawere2+ andtaken Mg intermittently2+ during the at experiment given times with from solutions the reaction containing slurry. the After inhibitor quick filtrationwere determined. (taking approximately Samples were taken 5–10 intermittentlys with a syringe at given filter) times and from hundred-fold the reaction dilution, slurry. After ICP-OES quick measurementsfiltration (taking yielded approximately the Mg2+ and 5–10 Ca s2+with concentration a syringe of filter) the solution and hundred-fold (Figure 2). dilution, ICP-OES measurements yielded the Mg2+ and Ca2+ concentration of the solution (Figure2).

12 Variation of conductivity 4.5 [Ca2+] [Mg2+] 4.0 10 3.5

3.0 8 2.5

2.0 6 c (g/L)

(mS/cm) 1.5 κ 1.0 4 0.5

2 0.0

-0.5 0 5 10 15 20 25 30 t(min)

Figure 2. Figure 2. ComparisonComparison of of the the results results of in situ cond conductivityuctivity measurement (l (lefteft Y axis) and ICP-OES measurements (right Y axis); latter were performed on samples of filtered mother liquors taken during measurements (right Y axis); latter were performed on samples of filtered mother liquors taken during the stoichiometric reaction of MgSO4 + Ca(OH)2 Mg(OH)2 + CaSO4 2H2O with identical (0.1 M) the stoichiometric reaction of MgSO4 + Ca(OH)2 → Mg(OH)2 + CaSO4·2H· 2O with identical (0.1 M) initial reactant concentrations, containing 1.5-mM citric acid inhibitor in distilled water (initial ionic initial reactant concentrations, containing 1.5-mM citric acid inhibitor in distilled water (initial ionic concentrations were calculated from the weight of the added salts; T = 20 C; in distilled water media concentrations were calculated from the weight of the added salts; T = 20 °C;◦ in distilled water media without any supporting electrolyte). without any supporting electrolyte). The results seem to verify that the magnesium content of the solution practically instantly The results seem to verify that the magnesium content of the solution practically instantly precipitates from the slurry as Mg(OH) (the ICP-OES measurements yielded Mg2+ concentrations in 2 2+ precipitates from the slurry as Mg(OH)2+2 (the ICP-OES measurements2+ yielded2 Mg concentrations in the range of 0.3–0.6% of the initial Mg concentration). The Ca and SO− 4 − ions remain in solution the range of 0.3–0.6% of the initial Mg2+ concentration). The Ca2+ and SO42 ions remain in solution for for a significantly longer period of time before gypsum precipitation commences. It can also be seen a significantly longer period of time before gypsum precipitation commences. It can also be seen that that the Ca2+ variations of the solution concentrations obtained from ICP-OES are superimposable with the Ca2+ variations of the solution concentrations obtained from ICP-OES are superimposable with the conductometric data, confirming the viability of conductometry as in situ means for monitoring the conductometric data, confirming the viability of conductometry as in situ means for monitoring the precipitation of gypsum in this system. the precipitation of gypsum in this system. To get a more detailed picture on the inhibition effect of citric acid, its effect on the precipitation of To get a more detailed picture on the inhibition effect of citric acid, its effect on the precipitation gypsum was also studied in the reaction of Na2SO4 + CaCl2 + 2 H2O 2 NaCl + CaSO4 2H2O with a of gypsum was also studied in the reaction of Na2SO4 + CaCl2 + 2 H2O→ → 2 NaCl + CaSO· 4·2H2O with 0.1 M initial reactant concentration, thus excluding the effect of Mg(OH)2 present in the target system. a 0.1 M initial reactant concentration, thus excluding the effect of Mg(OH)2 present in the target In these reactions, trisodium citrate was added as an inhibitor, to avoid the decrease in the pH and to system. In these reactions, trisodium citrate was added as an inhibitor, to avoid the decrease in the have the inhibitor in the same form as in the target reaction. These reactions were thermostated at 25 C pH and to have the inhibitor in the same form as in the target reaction. These reactions were◦ and carried out in a smooth-surface cylindrical PTFE vessel to avoid any wall effect. The variation of thermostated at 25 °C and carried out in a smooth-surface cylindrical PTFE vessel to avoid any wall the [Ca2+] was calculated from the conductometric measurements using two-point calibration using 2+ the initial and final [Ca ] concentration. The results are shown in Figure3. It can be seen that the citrate significantly retards the crystallization of gypsum. The half-reaction time in absence of citrate was approximately 3.5 min which increased to 28 min in the presence of 1.5-mM citrate inhibitor. Under these circumstances, the inhibition seems to be more efficient than in MoleculesMolecules 2019 2019, ,24 24, ,x x FOR FOR PEER PEER REVIEW REVIEW 55 of of 16 16 effect.effect. The The variation variation of of the the [Ca [Ca2+2+] ]was was calculated calculated from from the the conductometric conductometric measurements measurements using using two- two- pointpoint calibration calibration using using the the initial initial and and final final [Ca [Ca2+2+] ]concentration. concentration. The The results results are are shown shown in in Figure Figure 3. 3. ItIt can can be be seen seen that that the the citrate citrate significantly significantly retards retards the the crystallization crystallization of of gypsum. gypsum. The The half-reaction half-reaction Moleculestimetime in in 2020absence absence, 25, 5012 of of citrate citrate was was approximately approximately 3.5 3.5 min min which which increased increased to to 28 28 min min in in the the presence presence5 of of 16of 1.5-mM1.5-mM citrate citrate inhibitor. inhibitor. Under Under these these circumstances, circumstances, the the inhibition inhibition seems seems to to be be more more efficient efficient than than in in thethe target target reaction reaction under under identical identical conditions, conditions, wh whichich suggests suggests some some interaction interaction between between the the Mg(OH) Mg(OH)2 2 theandand targetthe the additive. additive. reaction under identical conditions, which suggests some interaction between the Mg(OH)2 and the additive.

0.110.11

0.100.10 In In absence absence of of citrate citrate 0.090.09 In In presence presence of of 1.5 1.5 mM mM citrate citrate

0.080.08

0.070.07 ] (M) ] (M) 2+ 2+ 0.060.06 [Ca [Ca 0.050.05

0.040.04

0.030.03

0.020.02 00 204060 204060 t t(min) (min)

2+ 2+ FigureFigure 3. 3.3. The TheThe variation variation variation of of [Caof [Ca [Ca2+] ]during during] during the the reaction reaction the reaction of of Na Na2 ofSO2SO Na4 4+ + CaCl CaClSO 2 2+ + 2 CaCl2 H H2O2O → →+ 2 22 NaCl NaCl H O + + CaSO CaSO2 NaCl4·2H4·2H2O2+O 2 4 2 2 → CaSOwithwith 0.1-M 0.1-M2H Oinitial initial with reactant reactant 0.1-M initial concentration concentration reactant concentration in in absence absence and and in in absencein presence presence and of of in 1.5-mM 1.5-mM presence citrate. citrate. of 1.5-mM citrate. 4· 2

TheThe solidssolids precipitatingprecipitating in in the the modelmodel reactionsreactions werewere studied studied byby usingusing X-rayX-ray didiffractometrydiffractometryffractometry and and scanningscanning electron electron microscopy. microscopy. TheThe The resultsresults results areare are shownshown shown inin in FigureFigure Figure4 4.. 4.

(020)(020) (130)(130) In In absence absence of of citrate citrate In In presence presence of of citrate citrate

(041)(041) (311)(311) (021)(021)

(021)(021) (041)(041) Intensity (a. u.) Intensity (a. u.) (020)(020) (200)(200) (220)(220) (130)(130)

1010 20 20 30 30 40 40 50 50 60 60 22θθ(°)(°)

(A)(A)

(B)(B) (C)(C)

Figure 4. Structural and morphological investigation of gypsum precipitated during the reaction of Na SO + CaCl + 2 H O 2 NaCl + CaSO 2H O. (A): XRD patterns of gypsum precipitated 2 4 2 2 → 4· 2 in the absence and in the presence of citrate (main miller indices were identified using the JCPDS database—CaSO 2H O: # 21-0816). Bottom: SEM images of gypsum precipitated in absence (B) and in 4· 2 presence (C) of citrate. Molecules 2019, 24, x FOR PEER REVIEW 6 of 16

Figure 4. Structural and morphological investigation of gypsum precipitated during the reaction of

Na2SO4 + CaCl2 + 2 H2O → 2 NaCl + CaSO4·2H2O. (A): XRD patterns of gypsum precipitated in the absence and in the presence of citrate (main miller indices were identified using the JCPDS database— MoleculesCaSO20204·2H, 252O:, 5012 # 21-0816). Bottom: SEM images of gypsum precipitated in absence (B) and in presence 6 of 16 (C) of citrate.

FromFrom the the diffractograms, diffractograms, the only detectable crystallinecrystalline solidsolid isis gypsum; gypsum; however, however, the the intensity intensity of ofthe the reflections reflections and and their their ratio ratio to each to each other other change change immensely immensely between between the ones the precipitated ones precipitated in absence in absenceor in the or presence in the presence of citrate. of Thiscitrate. implies This implies that the that morphology the morphology of the two of the solids two is solids significantly is significantly different. different.SEM images SEM support images thissupport assumption. this assumption. While in theWh absenceile in the of absence citrate, theof citrate, gypsum the only gypsum precipitates only precipitatesinto well-shaped, into well-shaped, rod-like ; rod-like in the crystals; solids crystallizedin the solids in crystallized the presence in ofthe citrate, presence only of plates citrate, can onlybe found. plates can Morphological be found. Morphological differences analogous difference tos analogous these were to already these were reported already in thereported literature in the by literatureothers [33 by–39 others]. [33–39]. 2+ TheThe inhibition inhibition effect effect observed—i.e.,observed—i.e., thethe increase increase in in the the induction induction time—may time—may be partlybe partly due due to Ca to Cacomplexation2+ complexation by citrate by citrate in solution in solution (decreasing (decreasing the degree the degree of supersaturation) of supersaturation) [37]. This [37]. is, however,This is, 2+ however,likely to belikely a minor to be eaff minorect due effect to the due low to concentrationthe low concentration of the inhibitor of the inhibitor relative torelative that of to the that Ca of 2 − theand Ca SO2+ 4and− ions. SO42 Theions. surface The surface adsorption, adsorption, which which has been has evidencedbeen evidenced for citrate for citrate as well as aswell for as other for otherorganic organic compounds compounds with with carboxylate carboxylate functionalities functionalit [33ies,35 [33,35,38],,38], is indeed is indeed more more relevant relevant here. here. It was It wasdemonstrated demonstrated [33, 38[33,38]] that that citrate citrate adsorbs adsorbs onto theonto active the active growth growth faces offaces gypsum. of gypsum. This way, This it way, inhibits it inhibitsthe growth the growth in these in directions these directions and results and result in morphologicals in morphological changes changes in the in forming the forming crystals. crystals. In the Inabsence the absence of citrate, of citrate, the (uninhibited) the (uninhibited) growth growth along along the c the axis c resultsaxis results in the in formation the formation of typical, of typical, long, long,needle-like needle-like crystals crystals (Figure (Figure4B) [ 39 4B)]. This [39]. is This inhibited is inhibited in the presence in the presence of citrate, of and citrate, crystal and growth crystal is growthfavored is infavored the a and in the b directions,a and b directions, resulting resulting in the formation in the formation of plate-like of plate-like crystals. crystals. Badens Badens et al. [ 33et ] al.suggested [33] suggested that for that efficient for adsorptionefficient adsorption (or surface (o complexation)r surface complexation) of citrate, theof citrate, distance the between distance the betweensurface calciumthe surface ions calcium and that ions of and the that two of oxygen the two oxygen on theatoms two on carboxylic the two carboxylic groups has groups to match; has tofor match; this, the for faces this, (1the 2 faces 0) and (1 (1 2 10) 1) and appeared (1 1 1) appeared to be the optimal to be the ones optimal [33,39 ones]. In [33,39]. a recent In work, a recent the surfacework, thecomplexation surface complexation of a citrate of monolayer a citrate monolayer was demonstrated was demonstrated by using a by variety using of a experimentalvariety of experimental means and meanscorroborated and corroborated by MD simulations by MD [ 36simulations]. In this work [36]. [36 In], thethis di workfference [36], between the difference the inhibiting between efficiency the inhibitingof citrate andefficiency tartrate of [citrate33] was and also tartrate possible [33] to was be explained. also possible to be explained. To investigate the kinetics of the reaction of MgSO44+ Ca(OH)22 + 2 H2O→ Mg(OH)2 + CaSO44 2H22O To investigate the kinetics of the reaction of MgSO + Ca(OH) + 2 H O → Mg(OH) + CaSO ·2H· O further,further, the the initial initial concentrations concentrations were were systematically systematically increased, increased, while while the the reactant-to-additive reactant-to-additive ratio ratio waswas kept kept constant. constant. The The results results of of these these measurements measurements are are shown shown in in Figure Figure 55..

16 Initial reactant concentration 0.15 M 14 0.125 M 0.1M

12

10

(mS/cm) 8 κ

6

4

0 10203040 t(min)

Figure 5. Effect of increasing reactant concentration on the conductivity vs time profiles of the Figure 5. Effect of increasing reactant concentration on the conductivity vs time profiles of the reaction reaction of MgSO4 + Ca(OH)2 + 2 H2O Mg(OH)2 + CaSO4 2H2O. During the reactions, the of MgSO4 + Ca(OH)2 + 2 H2O → Mg(OH)2 + CaSO→ 4·2H2O. During the reactions,· the additive-to-reactant additive-to-reactant ratio was held constant (the molar ratio of Ca:Mg:citric acid = 0.1:0.1:0.0015), the ratio was held constant (the molar ratio of Ca:Mg:citric acid = 0.1:0.1:0.0015), the initial reactant initial reactant concentrations are shown in the figure (T = 20 ◦C; in distilled water media without any concentrations are shown in the figure (T = 20 °C; in distilled water media without any supporting supporting electrolyte). electrolyte). It can be seen that even as the inhibitor concentration was increased proportionally to the increased reactant concentrations, the rate of the gypsum precipitation reaction increased significantly: the half-reaction time of the reaction with 0.15-M initial reactant concentrations was 8.5 min while with 0.1 M it was found to be 22.8 min. This phenomenon is somewhat expected, as the high agitation speed, Molecules 2019, 24, x FOR PEER REVIEW 7 of 16

It can be seen that even as the inhibitor concentration was increased proportionally to the increasedMolecules 2020 reactant, 25, 5012 concentrations, the rate of the gypsum precipitation reaction increased7 of 16 significantly: the half-reaction time of the reaction with 0.15-M initial reactant concentrations was 8.5 min while with 0.1 M it was found to be 22.8 min. This phenomenon is somewhat expected, as the highthe increasing agitation speed, supersaturation the increasing ratio supersaturation and the bigger amountratio and of the solid bigger (and amount thus potential of solid nucleation (and thus potentialsites) present nucleation in the reaction sites) present mixture in all the facilitate reaction the mixture precipitation all facilitate of gypsum. the precipitation This may be of undesirable gypsum. Thisin practical may be applications,undesirable in and practical must beapplications, somehow takenand must into be account somehow during taken the into secondary account stepduring of theacidic secondary mine drainage step of treatment,acidic mine as drainage the MgSO treatment,4-rich solutions as the MgSO coming4-rich from solutions the first stepcoming may from vary the in firstconcentration. step may vary A potential in concentration. solution for A thispotential can be so dilutionlution for or increasethis can thebe dilution additive or used. increase If it canthe additivebe carried used. out If economically, it can be carried the latterout economically, is probably favored,the latter because is probably it does favored, not make because it necessary it does not to makeincrease it necessary the amount to ofincrease liquid the to be amount treated. of liquid to be treated. To study the eeffectffect of the variation of additive dosage, the reaction with 0.1 M initial reactant concentrations was carried out, varying the amount of citrate added (Figure6 6).).

12 Concentration of citric adcid: 0.75 mM 1.5 mM 3 mM 10

8

(mS/cm) 6 κ

4

2 0 1020304050 t (min)

Figure 6. VariationVariation of conductivity in thethe reactionreaction ofof MgSOMgSO4 ++ Ca(OH)Ca(OH)2 + 2 H2O → Mg(OH)2 + 4 2 2 → 2 CaSO4·2H2H2OO with 0.1-M0.1-M initial initial reactant reactant concentration, concentration, using using different different amounts amounts of citric of acid citric as inhibitoracid as 4· 2 inhibitor(T = 20 ◦C; (T in= 20 distilled °C; in waterdistilled media water without media anywithout supporting any supporting electrolyte). electrolyte).

The added amountamount ofof citriccitric acid acid exerts exerts a a major major eff efectfect on on the the rate rate of gypsumof gypsum precipitation. precipitation. When When the theamount amount used used previously previously (1.5 mM)(1.5 mM) was was doubled, doubled, the half-reaction the half-reaction time time was alsowas almostalso almost doubled: doubled: with with3 mM 3 itmM was it 56 was min 56 compared min compared to 23 min to with 23 min 1.5 mM.with As1.5 the mM. inhibiting As the einhibitingffect is increased effect is significantly increased significantlythis could provide this could a viable provide method a viable to compensate method to the compensate effect of increasing the effect reactant of increasing concentrations. reactant concentrations.However, increasing However, the amount increasing of additive the amount can lower of additive the cost-e canffi ciencylower the of the cost-efficiency process, and of while the process,citric acid and is awhile “green” citric additive acid is a as “green” it is commonly additive as present it is commonly in nature, present it still cannotin nature, be usedit still in cannot large beamounts. used in To large overcome amounts. these To challenges, overcome these its potential challenges, to be its reused potential without to be adding reused a new,without costly adding step toa new,the process costly muststep to be the investigated—this process must be will investigat be discusseded—this later will in thisbe discussed work (see later Section in this 2.3). work (see SectionDuring 2.3). our experiments, the temperature of the reaction mixtures was always monitored. It was observedDuring that our the experiments, initial temperature the temperature of the MgSO of the4 solution reaction did mixtures not change was significantlyalways monitored. upon adding It was observedthe MoL tothat it, the presumably initial temperature because the of the various MgSO heat4 solution effects did roughly not change cancelled significantly each other upon out. adding It was thealso MoL observed to it, thatpresumably the temperature because ofthe the various reaction heat mixture effects remained roughly constantcancelled during each other the precipitation out. It was of the gypsum too—within 0.5 C. Accordingly, the test reactions always commenced at the actual also observed that the temperature± ◦ of the reaction mixture remained constant during the precipitation ofroom the temperaturegypsum too—within and remained ±0.5 °C. approximately Accordingly, constant the test reactions during an always experimental commenced run. Afterat the several actual roomrepeated temperature test reactions, and remained it was observed approximately that the temperatureconstant during of the an reactionexperimental mixture run. also After aff ectsseveral the repeatedrate of reaction. test reactions, The extent it was of thisobserved is qualitatively that the temperature in accord with of the the reaction Arrhenius mixture law. Thealso correlationaffects the ratewith of the reaction. rate of theThe reaction extent of is summedthis is qualitativel up in Tabley in1. accord with the Arrhenius law. The correlation with the rate of the reaction is summed up in Table 1.

Molecules 2020, 25, 5012 8 of 16

Table 1. Correlation between the reaction mixture’s temperature ( 0.5 C) and the rate of the gypsum ± ◦ precipitation reaction.

Temperature (◦C) Estimated Induction Time (min) Half-Reaction Time (min) 20 10.3 22.8 22 9.1 21.5 24 6.8 17.5 26 5.8 15.7

As it can be seen, even relatively small temperature changes result in significant variations in the reaction rate. At a lower temperature the reaction was slower, and it seems natural that the reaction becomes faster with increasing temperature—the temperature dependence of the solubility of gypsum can also play a role in this [40]. Since the temperature can also influence the reaction significantly, it can be seen that the details of the experimental procedure must be considered as a whole to suggest an operational setup for the separation of magnesium hydroxide and gypsum, and these conditions must be considered specifically for the given site, where it is used. It is also worth mentioning that during the treatment of wastewaters from acidic mine drainages the neutralized fluids are likely to contain a range of impurities, such as other metal ions as well as some organic components. Most of the metal ions should be precipitated in the first step of neutralization with /hydroxides in their hydroxide form; however, some studies suggest that they can influence the precipitation of gypsum even in trace amounts [20,22,23]. However, as the pH during the second step can increase up to ~11.2, and it is regulated by the reactants and products, the other residual metal ions are most likely precipitated as hydroxides at this stage. A more relevant factor for modifying the kinetics of the reaction in industrial systems can be assigned to organic impurities. While in this work the solutions were made up using distilled water, the Ca(OH)2 source was made of natural , therefore probably containing some organic impurities. According to the literature and our previous works, additives containing carboxylic acid and phosphonate functional groups can effectively inhibit the precipitation of gypsum [25–39], these types of impurities may, however, even improve the inhibition effect. On the other hand, compounds reacting with these organic molecules or interacting with the precipitants can be responsible for a variety of complications. Therefore, the effects associated with the auxiliary components of the starting solution must always be tested to fit the conditions of the process to the specific requirements.

2.2. Analysis of the Precipitating Solids and the Effect of Washing To characterize the solids precipitating from the reaction mixture (reaction conditions as shown in Figure1), samples were taken from the reaction slurry. To separate the two precipitates, a portion of the slurry was withdrawn at a 1 min reaction time. After a relatively quick filtration (taking ca. 5 min) the separated mother liquor, which is heavily supersaturated with respect to gypsum, was agitated further for 2 h to precipitate the excess in the solution. The solid filtered first mainly contained magnesium hydroxide; however, even with quick filtration, after drying, there was always some gypsum remaining in the solids. The reason for this is probably the remaining mother liquor on the surface of the solid, which still contains gypsum, which precipitates during the various manipulations and the subsequent drying. To eliminate these impurities, the samples were washed. After washing with different amounts of distilled water, we found that the optimal volume is 3–4 times the volume of the withdrawn sample. The separation process is presented in Scheme2. MoleculesMolecules2020 2019, 25,, 501224, x FOR PEER REVIEW 9 of9 16 of 16 Molecules 2019, 24, x FOR PEER REVIEW 9 of 16

Scheme 2. Experimental procedure for separating the precipitating solids during the MgSO4 + 4 SchemeCa(OH) 2.2.Experimental2 Experimental+ 2 H2O → Mg(OH) procedure procedure2 + CaSO for separatingfor4 ·2Hseparating2O reaction. the precipitating the precipitating solids duringsolids theduring MgSO the4 + MgSOCa(OH) 2+ +Ca(OH)2 H O2 + 2Mg(OH) H2O → Mg(OH)+ CaSO2 +2H CaSOO reaction.4·2H2O reaction. 2 → 2 4· 2 To enhance the efficiency of the process the properties of the products were investigated. First, To enhance the efficiency of the process the properties of the products were investigated. First, theTo solids enhance were the analyzed efficiency with of powderthe process XRD; the the proper resultsties are of shown the products in Figure were 7. investigated. First, thethe solidssolids werewere analyzedanalyzed withwith powderpowder XRD;XRD; thethe resultsresults areare shownshown inin FigureFigure7 .7.

⋅ CaSO4 2H2O ⋅ CaSOMg(OH)4 2H2O2 Mg(OH)2

solid precipitated from the solidfiltered precipitated mother fromliquor the after filtered mother liquor after separating Mg(OH)2 separating Mg(OH)2

solid filtered after 1 minute

Intensity (a. u.) solidreaction filtered time, after washed1 minute

Intensity (a. u.) reaction time, washed solid filtered after 1 minute solidreaction filtered time, after unwashed1 minute reaction time, unwashed

10 20 30 40 50 60 10 20 302θ (°) 40 50 60 2θ (°) Figure 7. XRD patterns of the separated, precipitated solids; main reflections were identified using the Figure 7. XRD patterns of the separated, precipitated solids; main reflections were identified using JCPDSFigure database7. XRD patterns (CaSO4 2Hof the2O: separated, #21-0816 and precipitated Mg(OH)2 :so #84-2164)lids; main Washing reflections of Mg(OH) were identified2 with distilled using the JCPDS database · (CaSO4·2H2O: #21-0816 and Mg(OH)2: #84-2164) Washing of Mg(OH)2 with water, four times the volume of taken sample. thedistilled JCPDS water,database four (CaSO times4 ·2Hthe 2volumeO: #21-0816 of taken and sample. Mg(OH) 2: #84-2164) Washing of Mg(OH)2 with distilled water, four times the volume of taken sample. Without sufficient aging, Mg(OH)2 tends to precipitate in poorly defined, most often amorphous Without sufficient aging, Mg(OH)2 tends to precipitate in poorly defined, most often amorphous crystals. The wide reflections appearing on the diffractograms suggest that this happens during the crystals.Without The sufficient wide reflections aging, Mg(OH) appearing2 tends on to the precipitate diffractograms in poorly suggest defined, that most this happensoften amorphous during the reactions in our hands, as the residence time is not sufficient for the formation of well-defined Mg(OH) crystals.reactions The inwide our reflections hands, as appearingthe residence on the time diff isractograms not sufficient suggest for thatthe thisformation happens of duringwell-defined the2 crystals.reactions On in theour other hands, hand, as thethe reflections residence corresponding time is not sufficient to gypsum for were the found formation always of to well-defined be intensive Mg(OH)2 crystals. On the other hand, the reflections corresponding to gypsum were found always to and sharp making it easy to identify even when present in relatively small amounts. It can be seen Mg(OH)be intensive2 crystals. and On sharp the othermaking hand, it easy the reflectionsto identify correspondingeven when present to gypsum in relatively were found small alwaysamounts. to It thatbe intensive the washing and sharp of the making initially it filtered easy to solid identify resulted even in when pure present Mg(OH) in2, relatively and no typical small reflectionamounts. ofIt can be seen that the washing of the initially filtered solid resulted in pure Mg(OH)2, and no typical gypsum is seen on the diffractogram. canreflection be seen thatof gypsum the washing is seen of on the the initially diffractogram. filtered solid resulted in pure Mg(OH)2, and no typical reflectionSince of washing gypsum seemed is seen on to bethe necessary diffractogram. to achieve a reasonable purity of Mg(OH)2, the effect Since washing seemed to be necessary to achieve a reasonable purity of Mg(OH)2, the effect of of washing was studied with systems at various initial concentrations. The optimal time-window washingSince washing was studied seemed with to systemsbe necessary at various to achieve initial a reasonableconcentrations. purity The of optimalMg(OH) time-window2, the effect of of of initial filtration was also tested by withdrawing samples from the reaction slurry at given times. washinginitial filtrationwas studied was withalso testedsystems by atwithdrawing various initial samples concentrations. from the reaction The optimal slurry time-windowat given times. of The The washing procedure was mentioned before—it was carried out with distilled water, four times the initialwashing procedure was also was tested mentioned by withdrawing before—it samples was carried from theout reaction with distilled slurry atwater, given four times. times The the volume of the taken sample. The results are summed up in Table2. washingvolume procedure of the taken was sample. mentioned The results before—it are summed was carried up in out Table with 2. distilled water, four times the volume of the taken sample. The results are summed up in Table 2. Table 2. Summary of gypsum content in the washed precipitated solids taken at given times in the

TableMgSO 2. Summary4 + Ca(OH) of2 gypsum→ Mg(OH) content2 + CaSO in the4·2H washed2O reaction precipitated carried solids out withtaken different at given initialtimes inreactant the MgSO4 + Ca(OH)2 → Mg(OH)2 + CaSO4·2H2O reaction carried out with different initial reactant

Molecules 2020, 25, 5012 10 of 16

Table2. Summary of gypsum content in the washed precipitated solids taken at given times in the MgSO4 + Ca(OH) Mg(OH) + CaSO 2H O reaction carried out with different initial reactant concentrations 2 → 2 4· 2 Moleculesand 2019 keeping, 24, x FOR the reactant–additive PEER REVIEW ratio constant (molar ratio of Ca:Mg:Citric acid = 0.1:0.1:0.0015). 10 of 16 Initial Reactant Concentration Sampleconcentrations Taken after and keeping the reactant–additive ratio constant (molar ratio of Ca:Mg:Citric acid = 0.1:0.1:0.0015). 0.1 M 0.125 M 0.15 M 1 min No gypsum present No gypsum present Gypsum present Initial Reactant Concentration Sample Taken after 5 min -0.1 M Small amount 0.125 of M gypsum present 0.15 Gypsum M present 10 min1 min No Nogypsum gypsum present present No Gypsum gypsum present present Gypsum Gypsum present present 5 min - Small amount of gypsum present Gypsum present 20 min10 min Gypsum No gypsum present present Gyps Gypsumum present present Gypsum present - 20 min Gypsum present Gypsum present -

The time-window,time-window, where where one one can can still still wash wash out theout gypsumthe gypsum from thefrom Mg(OH) the Mg(OH)2, decreases2, decreases quickly withquickly an increasingwith an increasing initial reactant initial concentration.reactant concentration. Besides the Besides faster reactionthe faster rate, reaction the bigger rate, amountsthe bigger of amountssolids in theof solids slurry alsoin the present slurry a also challenge present by makinga challenge the filtrationby making process the filtration longer resulting process in longer more resultinggypsum precipitated.in more gypsum precipitated. To look into the properties of the solids in more detail, SEM images of the washed and unwashed unwashed Mg(OH)2 samplessamples were were recorded. The The pictures pictures are are presented in Figure 88..

(A) (B)

Figure 8. SEM images of the precipitated Mg(OH)2 ((A):): washed sample (does not contain gypsum according to XRD) and (B): unwashed sample (some(some gypsumgypsum presentpresent accordingaccording toto XRD).XRD).

The SEM images of the precipitated magnesium hy hydroxidedroxide attest that the solids did not form regular crystals and that they have no clearclear morphology.morphology. When looking at the samples containing some gypsum impurities, no observable morphology can can be seen; however, the small bumps and spikes hint that gypsum precipitated on the surface and edges ofof thethe Mg(OH)Mg(OH)22. To investigateinvestigate this this phenomenon phenomenon further further and back and up back our earlier up our results, earlier SEM-EDX results, measurements SEM-EDX weremeasurements carried out were on carried the same out solid on the samples. same so Lookinglid samples. at the Looking washed at the Mg(OH) washed2 samples, Mg(OH)2 we samples, found weminute found amounts minute of amounts calcium of in calcium the solids in (thethe Casolids concentration (the Ca concentration was estimated was to estimated be less than to be 1 at%). less Thethan gypsum1 at%). The content gypsum of the content unwashed of the samplesunwashed was samples found was to be found much to higher, be much and higher, the elemental and the elementalmaps of these maps samples of these revealed samples the revealed distribution the dist ofribution calcium of and calcium magnesium and magnesium in the solids in (Figurethe solids9). The(Figure elemental 9). The map elemental of the unwashedmap of the Mg(OH)unwashed2, which Mg(OH) precipitated2, which precipitated proves that magnesiumproves that andmagnesium calcium andare locatedcalcium separately, are located and separately, the gypsum and precipitates the gypsum on theprecipitates surface of on the the Mg(OH) surface2 crystals of the present.Mg(OH)2 crystals present.

Molecules 20202019, 2524, 5012x FOR PEER REVIEW 11 of 16

(A) (B)

Figure 9.9. SEM-EDXSEM-EDX picture picture (A )( andA) and elemental elemental map (Bmap) of a(B representative) of a representative unwashed unwashed Mg(OH)2 precipitate. Mg(OH)2 precipitate. 2.3. The Location of Citrate Ions after the Completion of both Precipitation Reactions 2.3. TheDetermining Location of theCitrate destination Ions after ofthe the Completion added citric of both acid Precipitation is necessary Reactions both from an economic and environmentalDetermining point the of destination view. To determine of the added where citric the additive acid is ends necessary up, initially, both from UV-spectrophotometric an economic and measurementsenvironmental werepoint carried of view. out on To the determine mother liquor wh obtainedere the afteradditive the precipitationends up, initially, of both solids.UV- Thespectrophotometric measurements weremeasurements based on looking were atcarried the absorption out on the band mother of the Cliquor=O double obtained bond after situating the aroundprecipitation 200–230 of both nm. solids. The interference The measurements stemming were from based UV-active on looking impurities, at the absorption most probably bandFe(III), of the madeC=O double a bond interpretation situating ofaround these spectra 200–230 impossible. nm. The Therefore,interference it turned stemming out to from be necessary UV-active to separateimpurities, the most various probably components Fe(III), ofmade the a system sound prior interpretation to analysis. of these spectra impossible. Therefore, it turnedAccordingly, out to be HPLCnecessary measurements to separate werethe various carried components out. The samples of the includedsystem prior (i) Mg(OH) to analysis.2 filtered afterAccordingly, a 1 min reaction HPLC time, measurements which was thenwere washedcarried out. (see The Section samples 2.2); included (ii) the gypsum (i) Mg(OH) precipitated2 filtered fromafter thea 1 mothermin reaction liquor, time, from which which was the abovethen washed Mg(OH) (see2 was Section separated; 2.2); (iii)(ii) the washinggypsum fluidprecipitated used to removefrom the gypsum mother fromliquor, the from Mg(OH) which2 (see the Sectionabove Mg(OH)2.2) and2 (iv) was the separated; mother liquor (iii) the obtained washing after fluid gypsum used precipitation.to remove gypsum The first from two the solid Mg(OH) samples2 (see were Section dissolved 2.2) inand the (iv) eluent the (0.01mother M sulfuric liquor obtained acid). Because after ofgypsum solubility precipitation. issues, the The concentration first two solid of the samples target ion—citrate—approachedwere dissolved in the eluent the (0.01 detection M sulfuric limit ofacid). the HPLCBecause technique, of solubility and issues, an accurate the concentration concentration of determination the target ion—citrate—approached was not possible. The results, the detection however, stronglylimit of the suggested HPLC thattechnique, the citrate and ions an resideaccurate in partconcentration on the surface determination of the Mg(OH) was2 andnot inpossible. part remain The inresults, the mother however, liquor strongly to inhibit suggested the precipitation that the citrat ofe gypsum.ions reside In in the part precipitated on the surface gypsum of the and Mg(OH) in the2 washingand in part fluids, remain no sign in the of citricmother acid liquor was observed.to inhibit the precipitation of gypsum. In the precipitated gypsumTo further and in the verify washing this finding, fluids, no FT-IR sign spectroscopy of citric acid waswas employed.observed. The FT-IR spectra of solids (i) andTo (ii)further were verify recorded. this finding, To obtain FT-IR a reference spectroscopy sample, was the employed. precipitation The FT-IR reaction spectra was carriedof solids out (i) withoutand (ii) addedwere recorded. citric acid, To and obtain the referencea reference solids sample, were obtainedthe precipitation and treated reaction as above. was Duringcarried thisout “uninhibited”without added reaction, citric acid, the and precipitation the reference of the solids gypsum were is muchobtained faster; and therefore, treated as it above. was inevitable During thatthis some“uninhibited” gypsum remainedreaction, the on theprecipitation surface of theof the Mg(OH) gypsu2 meven is much after washing.faster; therefore, The results it was of these inevitable FT-IR measurementsthat some gypsum are shownremained in Figure on the 10 surface. of the Mg(OH)2 even after washing. The results of these FT-IRAs measurements it can be seen are on bothshown gypsum in Figure spectra 10. (Figure 10B), only the characteristic vibration bands of CaSO4 2H2O are present. The two spectra are practically identical, displaying the typical vibrations of · 1 sulfate ions and crystalline water. The bands at the 1100–1200 cm− region are the signals of ν3(SO4), 1 and the broad peaks around 2200 cm− are combinations of the ν1 and ν2 modes of sulfate. The sharp 1 peaks at 1620 and 1680 cm− correspond to the ν2 vibrations of crystal water, and the stretching modes 1 of water peak out of the broad band at the region of 3000–3600 cm− [41]. The spectra of the two (largely) Mg(OH)2 samples are, however, very much different. The spectrum of the solid obtained from the reaction without added citric acid is clearly the combination of the spectra of CaSO4 2H2O 1 · and Mg(OH)2. Besides the earlier described signs of gypsum, at 3700 cm− , one can see the sharp absorption band of OH stretching of the structural OH groups of Mg(OH)2, and also the broad peak in 1 the 1400–1600 cm− region can be assigned to the water bend vibrations of the adsorbed water [41,42],

Molecules 2020, 25, 5012 12 of 16 which is, as noted above, more or less expected. On the other hand, from Figure8A, the sample taken from the reaction with added citric acid did not contain appreciable amounts of gypsum, and even the peaks of Mg(OH)2 are partially masked by some new vibration bands that are not present on the 1 other three spectra. Based on literature data [43,44], the bands in the 1400–2000 cm− region and above 1 3500 cm− correspond to some form of citrate ion, and may well be either magnesium or or some mixed of the calcium-magnesium-citrate type [43,44]. This explains the small amounts of calcium, which were detected by the SEM-EDX measurements; citrate is most probably adsorbed on Molecules 2019, 24, x FOR PEER REVIEW 12 of 16 the surface of very fine Mg(OH)2 precipitate, and the calcium ions are there to compensate its charge.

0,020 A B 0.04

washed Mg(OH)2 precipitated 0,015 in presence of citric acid gypsum precipitated in presence of citric acid

0,010 0.02 gypsum precipitated washed Mg(OH)2 precipitated in absence of citric acid Absorbance in absence of citric acid Absorbance 0,005

0,000 0.00 4000 3500 3000 2500 2000 1500 1000 4000 3500 3000 2500 2000 1500 1000 Wavenumber (cm-1) Wavenumber (cm-1)

Figure 10. FT-IR spectra of the (A): Mg(OH) and (B): CaSO 2H O precipitates obtained from reaction 2 4· 2 mixturesFigure 10. containing FT-IR spectra citrate of the ion (A (upper): Mg(OH) spectra)2 and and (B): thoseCaSO in4·2H absence2O precipitates of the inhibitor obtained (lower from spectra),reaction asmixtures described containing in the text. citrate ion (upper spectra) and those in absence of the inhibitor (lower spectra), as described in the text. The results suggest that Mg(OH)2 binds some of the citric acid, which can explain the decreased effectivenessAs it can ofbe inhibitionseen on both compared gypsum spectra to the reactions(Figure 10B), of NaonlySO the +characteristicCaCl + 2 HvibrationO 2 bands NaCl of+ 2 4 2 2 → CaSO4·2H2H2OO. are Since present. this causes The two a fraction spectra ofare the practically additive identical, to be removed displaying from the solution,typical vibrations only the 4· 2 remainingof sulfate ions citrate and can crystalline inhibit the water. reaction. The bands This must at the also 1100–1200 be considered cm−1 region when working are the signals with an of increased ν3(SO4), −1 initialand the reactant broad peaks concentration, around 2200 since cm a bigger are combinations amount of Mg(OH)of the ν12 andwill ν remove2 modes more of sulfate. citrate The from sharp the solutionpeaks at 1620 decreasing and 1680 the cm effi−1ciency correspond of inhibition. to the ν2 vibrations of crystal water, and the stretching modes of waterEven peak so, these out of results the broad are promising band at forthe the region reusability of 3000–3600 of the citrate, cm−1 [41]. the precipitated The spectra and of the washed two Mg(OH)(largely) 2Mg(OH)containing2 samples part of are, it can however, be reused very in the much first differ step ofent. the The neutralization spectrum of process. the solid The obtained filtrate containingfrom the reaction the rest without can be potentiallyadded citric used acid to is makeclearly up the the combination magnesium of oxide the spectra in a slurry of CaSO and also4·2H for2O theand firstMg(OH) neutralization2. Besides step.the earlier described signs of gypsum, at 3700 cm−1, one can see the sharp absorption band of OH stretching of the structural OH groups of Mg(OH)2, and also the broad peak 3.in Conclusionsthe 1400–1600 cm−1 region can be assigned to the water bend vibrations of the adsorbed water [41,42],During which our is,work, as noted an attemptabove, more was madeor less to expected separate. On the the precipitation other hand, of from Mg(OH) Figureand 8A, CaSO the sample2H O 2 4· 2 intaken situ from by using the citratereaction ions with as anadded additive citric in acid the MgSOdid not+ containCa(OH) appreciable+ 2H O amountsMg(OH) of+ gypsum,CaSO 2H andO 4 2 2 → 2 4· 2 reaction.even the peaks The e ffofects Mg(OH) of citric2 are acid partially concentration, masked by the some initial new reactant vibration concentration bands that and are thenot reactionpresent temperatureon the other werethree investigated. spectra. Based The on reactions literature were data monitored [43,44], the by inbands situ conductometric in the 1400–2000 measurements. cm−1 region Theand conductometricabove 3500 cm−1observations correspond to were some backed form upof citrate by ICP-OES ion, and measurements. may well be Ineither the targetmagnesium reaction, or thecalcium Mg(OH) citrate2 precipitated or some mixed instantaneously salt of the calcium- after themagnesium-citrate two components weretype [4 mixed.3,44]. TheThis precipitationexplains the ofsmall gypsum amounts was of found calcium, to be which significantly were detected slower, even by the in theSEM-EDX absence measurements; of any inhibitor. citrate Adding is citricmost 2 acidprobably to the adsorbed system exerted on the nosurface effect of on very the ratefine ofMg(OH) precipitation precipitate, of Mg(OH) and the2, but calcium increased ions significantly are there to thecompensate induction its time charge. of the gypsum precipitation. Increasing the initial reactant concentration while keepingThe theresults additive suggest ratio that constant Mg(OH) still2 binds resulted some in anof increasethe citricin acid, the which reaction can rate. explain The increasethe decreased in the concentrationeffectiveness of of theinhibition added citriccompared acid also to increasedthe reactions the inductionof Na2SO period4 + CaCl notably.2 + 2 WithH2O an→ increasing2 NaCl + temperature,CaSO4·2H2O. theSince rate this of thecauses precipitation a fraction reactionof the additi was foundve to be to removed increase. from the solution, only the remaining citrate can inhibit the reaction. This must also be considered when working with an increased initial reactant concentration, since a bigger amount of Mg(OH)2 will remove more citrate from the solution decreasing the efficiency of inhibition. Even so, these results are promising for the reusability of the citrate, the precipitated and washed Mg(OH)2 containing part of it can be reused in the first step of the neutralization process. The filtrate containing the rest can be potentially used to make up the magnesium oxide in a slurry and also for the first neutralization step.

Molecules 2020, 25, 5012 13 of 16

Our results show that with careful control over the reaction conditions, the precipitation of magnesium hydroxide can be separated in time from the precipitation of gypsum using relatively small concentrations of citric acid as an additive. The Mg content of the starting solution can be recovered as relatively pure Mg(OH)2 after washing off the inevitable gypsum crystals from its surface. The inhibition effect of citrate is due to the interaction of the anion with the crystal face of the gypsum which is the fastest growing one in the absence of the inhibitor. This effect results in severe morphological changes in the precipitate. The separation of the two precipitation processes can be exploited in different areas—e.g., in the processing of acidic wastewater and mine drainages via neutralization. The observation that the citrate ions partially precipitate on the surface of the Mg(OH)2 is even more helpful, since the product as well as the additive can be reused in the neutralization process reducing the cost of the process.

4. Materials and Methods

4.1. Materials heptahydrate, 99.7% assay (MgSO 7H O), citric acid, 99.9% assay (C H O ), 4· 2 6 8 7 were used in this study and are both products of VWR. Other materials used include: calcium oxide, quicklime made from the limestone of a natural source (CaO), provided to us by Lhoist- lime producer and distilled water.

4.2. Preparation of Milk of Lime (MoL) To get MoL with approximately 20 w/w% solid content, 155 g of quicklime was suspended in 845 g of distilled water in a cylindrical glass reactor (1-L capacity) with vigorous agitation using a shaft stirrer with a screw propeller PTFE shaft (agitation speed was 800 rpm). The quicklime was added to the distilled water in small portions to avoid quick warming. The system was stirred for 2 h, then the insoluble grids were removed with a sieve with 300 µm hole-diameter. The exact solid content was determined by measuring the weight loss of a heated sample—100 g MoL was dried at 120 ◦C for 12 h. To ease the operation, the MoL was diluted to ca. 5 w/w% solid content and the active Ca(OH)2 content was determined by titration with HCl using indicator. The MoL used for the reactions was always freshly prepared—no samples older than 5 days were used, as the aggregation, sedimentation and reaction with airborne CO2 change the properties of MoL significantly after this time period.

4.3. Performing the MgSO + Ca(OH) + 2H O Mg(OH) + CaSO 2H O Reactions 4 2 2 → 2 4· 2 The reactions were carried out in a cylindrical glass reactor also used for the preparation of the MoL. The agitating system was also the same. The agitation speed chosen was 700 rpm to ensure the quick homogenization of the reaction slurry. The reactions were always carried out in 1 L volume. The magnesium sulfate and the citric acid were made up in distilled water and agitated in the reactor; then, MoL—containing stochiometric amount of Ca(OH)2—was added to the system to commence the reaction. MoL with 5 w/w% solid content was used, the volume of the required amount determined the volume of distilled water in which the MgSO4 and citric acid was dissolved. The reactions were followed in situ by conductometric measurements, and the pH and temperature of the reaction were also monitored. The repeatability of the reactions were tested and besides the effect of temperature, the results show that with good control over the experimental conditions the runs of the reactions are well reproducible, as shown in the Supplementary Materials, Appendix B. To analyze the solids, samples were withdrawn at given times from the slurry, and they were filtered with using vacuum filtration as quickly as possible (note that filtration time is strongly dependent on the volume of the taken sample). The Mg(OH)2 samples were washed with distilled water; four times the volume of the sample taken was usually sufficient to wash out gypsum from the samples. Molecules 2020, 25, 5012 14 of 16

4.4. Experimental Methods Conductometric measurements were carried out using a Jenway 3540 pH and conductivity meter (Keison. Products, Essex, UK) and a Jenway 027,013 conductivity cell. The temperature of the reactions was also followed by the built-in sensor of this conductivity cell. The pH of the reaction mixtures was measured with a Sentix 62 pH electrode (WTW, Weilheim, Germany), calibrated to 5 points, namely pH 2, 4, 7, 10, 11.5. For the ICP-OES measurements, samples were taken from the slurry at given times using a syringe, then were quickly filtered with a syringe filter (0.45 µm). The filtrate was diluted one-hundred-fold in 2% nitric acid to avoid further precipitation. The measurements were carried out using a Thermo Scientific iCAP 7400 ICP-OES DUO spectrometer (Waltham, MA, USA). The diffractograms of the solids were obtained using a Rigaku MiniFlex II type Röntgen diffractometer (Tokyo, Japan), in the 2θ = 4◦–60◦ range with 2◦/min scanning speed, using CuKα (λ = 1.5418 Å) radiation. The morphologies of the precipitated solids were studied by scanning electron microscopy, using a Hitachi S-4700 scanning electron microscope (Tokyo, Japan). The elemental mapping was also possible using the coupled Röntec QX2 spectrometer (Berlin, Germany) equipped with Be window. The eluent for the HPLC measurements was 0.01 M ; therefore, two-fold diluted fluid samples were prepared. To make samples from the solid samples with maximum citrate concentration, they were dissolved in 0.01 M sulfuric acid to obtain saturated solution, then the remaining solids were filtered off and the fluid sample was diluted two-fold. Each dilution was needed to avoid solubility issues during the measurements. For the measurements, Agilent 1100 series HPLC (Santa Clara, CA, USA) was used with a Grom-Resin ZH type column. IR measurements were carried out using a JASCO FT/IR-4700 spectrophotometer (Kyoto, 1 1 Japan) with 4 cm− resolution accumulating 256 scans in the 4000–650 cm− wavenumber range. The spectrometer was equipped with a ZnSe ATR accessory and a DTGS detector.

Supplementary Materials: The following are available online, Table S1: The effect of additives on gypsum precipitation in the reaction of Na SO + CaCl + 2 H O 2 NaCl + CaSO 2H O with 0.2 M initial reactant 2 4 2 2 → 4· 2 concentrations; Table S2: The effect of some additives on gypsum precipitation in the reaction of MgSO4 + Ca(OH)2 + 2 H2O Mg(OH)2 + CaSO4 2H2O with 0.2 M initial reactant concentrations; Figure S1: Variation of conductivity during three→ parallel reactions· of MgSO + Ca(OH) + 2 H O Mg(OH) + CaSO 2H O with 0.1 initial reactant 4 2 2 → 2 4· 2 concentration and in presence of 1.5 mM citric acid, at 22 ◦C. Author Contributions: Conceptualization, I.P. and P.S.; methodology, Z.K. and Á.K.; validation, M.S., I.P. and P.S.; investigation, S.Z. and M.S.; resources, P.S., Z.K. and Á.K.; data curation, M.S.; writing—original draft preparation, S.Z.; writing—review and editing, I.P. and P.S.; supervision, I.P. and P.S. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Acknowledgments: We would like to thank the following people for their contribution to this article: Lhoist-Minerals and lime producer for the financial support and supplying chemicals; Ilona Varga Halasiné for her devoted help in the laboratory work, Miklós Csáti for the ICP measurements, Máté Náfrádi for the HPLC measurements, and for every member of the Materials and Structure research group who have helped us during our work. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Zanin, M.; Lambert, H.; du Plessis, C. Lime Use and Functionality in Sulphide Flotation: A Review. Miner. Eng. 2019, 143, 105922. [CrossRef] 2. Ineich, T.; Degreve, C.; Karamoutsos, S.; du Plessis, C. Utilization Efficiency of Lime Consumption during Magnesium Sulfate Precipitation. Hydometallurgy 2017, 173, 241–249. [CrossRef] 3. Nie, J.; Yi, S. Neutralization of Acid Wastewater and Magnesium Hydroxide Slurry from Electrolytic Pretreatment. Adv. Mat. Res. 2015, 1073–1076, 949–954. [CrossRef] Molecules 2020, 25, 5012 15 of 16

4. Gong, M.H.; Johns, M.; Fridjonsson, E. Magnesium recovery from desalination brine. CEED Semin. Proc. 2018, 49–54. 5. Liu, S.T.; Nancollas, G.H. The crystallization of magnesium hydroxide. Desalination 1973, 12, 75–84. 6. Liu, S.T.; Nancollas, G.H. The kinetics of crystal growth of calcium sulfate dihydrate. J. Cryst. Growth 1970, 6, 281–289. [CrossRef] 7. Klein, D.H.; Smith, M.D.; Driy, J.A. Homogeneous nucleation of magnesium hydroxide. Talanta 1967, 14, 937–940. [CrossRef] 8. Carson, R.C.; Simandl, J. Kinetics of magnesium hydroxide precipitation from seawater using slaked . Miner. Eng. 1994, 7, 511–517. [CrossRef] 9. Yuan, Q.; Lu, Z.; Zhang, P.; Luo, X.; Ren, X.; Golden, T.D. Study of the synthesis and crystallization kinetics of magnesium hydroxide. Mater. Chem. Phys. 2015, 162, 734–742. [CrossRef] 10. Chungcheng, C.; Nancollas, G.H. The crystallization of magnesium hydroxide, a constant composition study. Desalination 1982, 42, 209–219. [CrossRef] 11. Harris, A.; Finan, M.A.; Elliot, M.N.; Harris, A.; Finan, M.A.; Elliot, M.N. A theory of the formation of magnesium scales in sea water plants, and means for their prevention. Desalination 1974, 14, 325–340. [CrossRef] 12. Walinsky, S.W.; Morton, B.J. Chemistry of alkaline scale inhibition in seawater desalination by Flocon antiscalant 247. Desalination 1979, 31, 289–298. [CrossRef] 13. Xu, C.; Wang, F.; Liu, D.; Chen, W. Effect of Additive EDTA on Crystallization Process of Magnesium Hydroxide Precipitation. Chin. J. Chem. Eng. 2010, 18, 761–766. [CrossRef] 14. Schierholtz, O.J. The crystallization of calcium sulfate dihydrate. Can. J. Chem. 1958, 36, 1057–1063. [CrossRef] 15. Christoffersen, M.R.; Christoffersen, J.; Weijnen, M.P.C.; van Rosmalen, G.M. Crystal growth of calcium sulphate dihydrate at low supersaturation. J. Cryst. Growth 1982, 58, 585–595. [CrossRef] 16. Witkamp, G.J.; van der Eerden, J.P.; van Rosmalen, G.M. Growth of gypsum I. Kinetics. J. Cryst. Growth 1990, 102, 281–289. [CrossRef] 17. Klepetsanis, P.G.; Koutsoukos, P.G. Spontaneous precipitation of calcium sulfate at conditions of sustained supersaturation. J. Colloid. Interf. Sci. 1991, 143, 299–308. [CrossRef] 18. de Meer, S.; Spiers, C.J.; Peach, C.J. Kinetics of precipitation of gypsum and implications for pressure–solution creep. J. Geol. Soc. Lond. 2000, 157, 269–281. [CrossRef] 19. McCartney, E.R.; Alexander, A.E. The effect of additives upon the process of crystallization I. Crystallization of calcium sulfate. J. Coll. Sci. 1958, 13, 383–396. [CrossRef] 20. Witkamp, G.J.; van Rosmalen, G.M. Growth of gypsum II. Incorporation of cadmium. J. Cryst. Growth 1991, 108, 89–98. [CrossRef] 21. Rashad, M.M.; Mahmoud, M.H.H.; Ibrahim, L.A.; Abdel–Aal, E.A. Crystallization of calcium sulfate dihydrate under simulated conditions of phosphoric acid production in the presence of aluminum and magnesium ions. J. Cryst. Growth 2004, 267, 372–379. [CrossRef] 22. Deng, L.; Zhang, Y.; Chen, F.; Cao, S.; You, S.; Liu, Y.; Zhang, Y. Reactive crystallization of calcium sulfate dihydrate from acidic wastewater and lime. Chinese J. Chem. Eng. 2013, 21, 1303–1312. [CrossRef] 23. Hamdona, S.K.; Al Hadad, U.A. Crystallization of calcium sulfate dihydrate in the presence of some metal ions. J. Cryst. Growth 2007, 299, 146–151. [CrossRef] 24. Hamdona, S.K.; Al Hadad, O.A. Influence of additives on the precipitation of gypsum in chloride solutions. Desalination 2008, 228, 277–286. [CrossRef] 25. Liu, S.T.; Nancollas, G.H. A Kinetic and morphological study of the seeded growth of calcium sulfate dihydrate in the presence of additives. J. Colloid. Interf. Sci. 1975, 52, 593–601. [CrossRef] 26. Klepetsanis, P.G.; Koutsoukos, P.G. Kinetics of calcium sulfate formation in aqueous media: Effect of organophosphorus compounds. J. Cryst. Growth 1998, 193, 156–163. [CrossRef] 27. El–Shall, H.; Rashad, M.M.; Abdel-Aal, E.A. Effect of phosphonate additive on crystallization of gypsum in phosphoric and sulfuric acid medium. Cryst. Res. Technol. 2002, 37, 1264–1273. [CrossRef] 28. Prisciandaro, M.; Olivieri, E.; Lancia, A.; Musmarra, D. Gypsum Precipitation from an Aqueous Solution in the Presence of Nitrilotrimethylenephosphonic Acid. Ind. Eng. Chem. Res. 2006, 45, 2070–2076. [CrossRef] 29. Amjad, Z. Gypsum scale formation on heated metal surfaces: The influence of polymer type and polymer stability on gypsum inhibition. Desalin. Water. Treat. 2013, 51, 4709–4718. [CrossRef] Molecules 2020, 25, 5012 16 of 16

30. Lioliou, M.G.; Paraskeva, C.A.; Koutsoukos, P.G.; Payatakes, A.C. Calcium sulfate precipitation in the presence of water-soluble polymers. J. Colloid. Interf. Sci. 2006, 303, 164–170. [CrossRef] 31. Popov, K.; Rudakova, G.; Larchenko, V.; Tusheva, M.; Kamagurov, S.; Dikareva, J.; Kovaleva, N. A comparative performance evaluation of some novel (green) and traditional antiscalants in calcium sulfate scaling. Adv. Mater. Sci. Eng. 2016, 2016, 1–10. [CrossRef] 32. Rabizadeh, T.; Morgan, D.J.; Peacock, C.L.; Benning, L.G. Effectiveness of green additives vs. poly(acrylic acid) in inhibiting calcium sulfate dihydrate crystallization. Ind. Eng. Chem. Res. 2019, 58, 1561–1569. [CrossRef] 33. Badens, E.; Veesler, S.; Boistelle, R. Crystallization of gypsum from hemihydrate in presence of additives. J. Cryst. Growth 1999, 198–199, 704–709. [CrossRef] 34. Prisciandaro, M.; Lancia, A.; Musmarra, D. The retarding effect of citric acid on calcium sulfate nucleation kinetics. Ind. Eng. Chem. Res. 2003, 42, 6647–6652. [CrossRef] 35. Prisciandaro, M.; Santucci, A.; Lancia, A.; Musmarra, D. Role of citric acid in delaying gypsum precipitation. Can. J. Chem. Eng. 2005, 83, 586–592. 36. Chen, W.; Zhao, W.; Wu, Y.; Wang, Y.; Zhang, B.; Li, F.; Chen, Q.; Qie, Z.; Xu, Z. Origin of gypsum growth habit difference as revealed by molecular conformations of surface-bound citrate and tartrate. Cryst. Eng. Comm. 2018, 20, 3581–3589. [CrossRef] 37. Rabizadeh, T.; Peacock, C.L.; Benning, L.G. Carboxylic acids: Effective inhibitors for calcium sulfate precipitation? Mineral. Mag. 2014, 78, 1465–1472. [CrossRef] 38. Ersen, A.; Smith, A.; Chotard, T. Effect of malic and citric acid on the crystallization of gypsum investigated by coupled acoustic emission and electrical conductivity techniques. J. Mater. Sci. 2006, 41, 7210–7217. [CrossRef] 39. Qu, J.; Peng, J.; Li, B. Effect of citric acid on the crystal morphology of gypsum and its action mechanism. Adv. Mater. Res. 2011, 250, 321–326. 40. Partridge, E.P.; White, A.H. The solubility of calcium sulfate from 0 to 200. J. Am. Chem. Soc. 1929, 51, 360–370. [CrossRef] 41. Bensted, J.; Prakash, S. Investigation of the calcium sulphate–water system by infrared spectroscopy. Nature 1968, 219, 60–61. [CrossRef] 42. Wiley SpectraBase. SpectraBase Compound ID 7kH3z5zFF2C. Available online: https://spectrabase.com/ spectrum/9GqJIqcyzih (accessed on 17 July 2020). 43. Wiley SpectraBase. SpectraBase Compound ID FXpCU14bzDK. Available online: https://spectrabase.com/ spectrum/Kof74u1OB3P (accessed on 17 July 2020). 44. Chanfrau, J.E.R.; Álvarez, L.M.; Crespo, A.B. Evaluation of calcium and from Cuban dolomite. Rev. Cuba. Farm. 2014, 48, 636–645.

Sample Availability: Samples of the compounds upon special request are available from the authors.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 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/).