当前位置:文档之家› Effect of antiscalants on precipitation of an RO concentrate_ Metals precipitated and ...

Effect of antiscalants on precipitation of an RO concentrate_ Metals precipitated and ...

Effect of antiscalants on precipitation of an RO concentrate:Metals precipitated and particle characteristics for several water compositions

Lauren F.Greenlee b ,Fabrice Testa c ,Desmond https://www.doczj.com/doc/8918071302.html,wler a ,*,Benny D.Freeman b ,Philippe Moulin c

a

The University of Texas at Austin,Department of Civil,Architectural and Environmental Engineering,1University Station C1786Austin,TX 78712,USA b

The University of Texas at Austin,Department of Chemical Engineering,1University Station C0400Austin,TX 78712,USA c

Universite

′Paul Ce ′zanne Aix Marseille,Laboratoire Me ′canique,Mode ′lisation et Proce ′de ′s Propres (CNRS –UMR 6181–M2P2),Europo ?le de l’Arbois-Pavillon Lae ¨nnec BP80,13545Aix en Provence Cedex 4,France

a r t i c l e i n f o

Article history:

Received 22November 2009Received in revised form 22January 2010

Accepted 26January 2010Available online 4February 2010Keywords:Desalination Antiscalants Reverse osmosis Concentrate treatment Precipitation Brackish water

a b s t r a c t

Inland brackish water reverse osmosis (RO)is economically and technically limited by the large volume of salty waste (concentrate)produced.The use of a controlled precipitation step,followed by solid/liquid separation (?ltration),has emerged as a promising side-stream treatment process to treat reverse osmosis concentrate and increase overall system recovery.The addition of antiscalants to the RO feed prevents precipitation within the membrane system but might have a deleterious effect on a concentrate treatment process that uses precipitation to remove problematic precipitates.The effects of anti-scalant type and concentration on salt precipitation and precipitate particle morphology were evaluated for several water compositions.The primary precipitate for the synthetic brackish waters tested was calcium carbonate;the presence of magnesium,sulfate,minor ions,and antiscalant compounds affected the amount of calcium precipitated,as well as the phases of calcium carbonate formed during precipitation.Addition of antiscalant decreased calcium precipitation but increased incorporation of magnesium and sulfate into precipitating calcium carbonate.Antiscalants prevented the growth of nucleated precipitates,resulting in the formation of small (100–200nm diameter)particles,as well as larger (6–10m m)particles.Elemental analysis revealed changes in composition and calcium carbonate polymorph with antiscalant addition and antiscalant type.Results indicate that the presence of antiscalants does reduce the extent of calcium precipitation and can worsen subsequent ?ltration performance.

a2010Elsevier Ltd.All rights reserved.

1.Introduction

Desalination of brackish water,which contains 1–10g/L total dissolved solids (TDS),has emerged as a key alternative water treatment technology.Traditional fresh water resources are

dwindling,and groundwater aquifers are becoming increas-ingly saline (Service,2006).In many countries,large natural brackish water aquifers are essentially untapped water resources (Sandia,2003).Reverse osmosis (RO)membranes are the primary choice in desalination technology;RO

*Corresponding author .Tel.:t151********;fax:t151********.E-mail address:dlawler@https://www.doczj.com/doc/8918071302.html, (D.F.

Lawler).

A v a i l a b l e a t w w w.s c i e n c e d i re ct.c o m

j o u r n a l h o m e p a g e :w w w.e l s e v i e r.c o m /l o c a t e /w a t r e s

w a t e r r e s e a r c h 44(2010)2672–2684

0043-1354/$–see front matter a2010Elsevier Ltd.All rights reserved.doi:10.1016/j.watres.2010.01.034

desalination requires less energy than thermal desalination (evaporation)(1.5–2.5kWh/m3for RO versus15–25kWh/m3 for evaporation)(Service,2006),and improvements in membranes and energy recovery have dropped the cost of RO desalination signi?cantly.

A key limitation of inland RO desalination is the volume of the waste stream,or concentrate,produced.The recovery (ratio of the product volume to the feed volume)of a brackish water RO system is typically limited to60–90%.For compar-ison,typical fresh water treatment plants have recoveries above99%.The large waste volume results in high disposal costs and is environmentally undesirable.Brackish water RO recovery is limited by sparingly soluble salts(CaCO3,CaSO4, BaSO4,SrSO4,silica)in the feed water that become supersat-urated during RO desalination and can precipitate on the membrane surface.Such membrane fouling,called scaling, can sometimes be removed by chemical cleaning processes, but often the membranes are permanently fouled and require replacement.

Synthetic chemicals called antiscalants are added to the RO feed stream to inhibit precipitation.As crystals nucleate, antiscalants adsorb onto growth sites and prevent further growth and precipitation.Antiscalants may also prevent precipitation through particle dispersion.However,anti-scalants(and pH control)only enable RO systems to achieve the recovery range stated above;to further increase RO system recovery,alternative methods must be used.

Controlled precipitation,where the concentrate is fed to a reactor(separate from the membrane system)and treated with chemicals such as base and carbonate to induce precip-itation,followed by solid/liquid separation(typically sedi-mentation and?ltration),is a promising method to treat the RO brackish water concentrate.Concentrate treatment removes problematic sparingly soluble salts,thereby enabling a secondary RO system(following concentrate treatment)to operate at high recovery.This arrangement increases the overall system recovery.Previous research has shown that most(90–95%)of the calcium can be removed,but greater calcium removal was achieved in synthetic RO concentrate (95%)than in the?eld water sample tested(90%)(Rahardianto et al.,2007).The authors hypothesized that natural organic matter(NOM)and/or the antiscalant present(30mg/L)were mostly likely responsible for the decrease in calcium precipitation in the?eld sample.Minimal residual calcium in the concentrate is desirable to limit the potential for precipi-tation when the concentrate is treated by a secondary RO or nano?ltration(NF)system to increase water recovery.With minimal residual calcium,precipitation in a secondary RO system can be largely controlled with pH reduction,anti-scalant addition,and CO2degasi?cation.

This paper focuses on the hypothesis that antiscalants are responsible for some of the reduced precipitation.The objec-tive was to determine how antiscalant compounds in synthetic brackish water RO concentrate in?uence precipita-tion and subsequent solid/liquid separation.Four antiscalants were investigated,and the role of several water components was evaluated using four water compositions of increasing complexity.For the precipitation step,the extent of calcium precipitated was measured,and the precipitated particles were evaluated for differences in size distribution and particle morphology.For the solid/liquid separation step,the?ux through micro?ltration membranes(0.1m m pore size)was evaluated.

Extensive work has been reported on the crystallization and morphology of calcium carbonate,both as a pure,satu-rated solution and with other additives,such as various cations and organic compounds(Brecˇevic′et al.,1996;Falini et al.,2009;Lam et al.,2007;Nancollas and Sawada,1982; Nebel et al.,2008;Reddy and Wang,1980;Sawada,1997; Wada et al.,1995;Westin and Rasmuson,2005;Yang et al., 2001).The thermodynamically stable form of calcium carbonate is calcite,with aragonite and vaterite considered to be metastable forms that eventually form calcite.Calcite, aragonite,and vaterite are all anhydrous forms of calcium carbonate;three hydrated forms of calcium carbonate may also form during precipitation,including monohydrocalcite (CaCO3$H2O),ikaite(CaCO3$6H2O),and amorphous calcium carbonate(ACC)(Brecˇevic′et al.,1996;Lam et al.,2007).

ACC has been shown to have a variable composition,and recent work has indicated that the internal structure can have the crystalline structure of the subsequent anhydrous calcium carbonate phase(Nebel et al.,2008).ACC becomes more stable when magnesium,phosphate,or organic compounds are present.The in?uence of additives is thought to affect not only the solid crystalline forms of calcium carbonate but also the initially formed amorphous

forms w a t e r r e s e a r c h44(2010)2672–26842673

(Lam et al.,2007).Based upon a molecular dynamics study of the solvation of dissolved calcium carbonate in water,the ?rst hydration shell around an ionic bonded calcium carbonate pair is similar in structure to the polymorph ikaite (Bruneval et al.,2007).Lam et al.(2007)found that the formation of anhydrous phases results from the expulsion of water from the amorphous phase rather than dramatic changes in crystal structure.

At higher supersaturation conditions(log SI value greater than2.5for calcite),such as those found in this study,the?rst calcium carbonate phase formed is ACC,while at lower supersaturation conditions(and in the absence of additives), calcite forms.When ACC is formed,subsequent forms can include ikaite(at temperatures below25 C)and vaterite–calcite mixtures.Ikaite can also form when phosphate or organic additives are present(Clarkson et al.,1992).

The presence of magnesium changes the rhombohedral morphology of calcium carbonate,while sulfate can cause calcite particles to form aggregates(Falini et al.,2009).When

magnesium is present in high concentrations relative to calcium(greater than4:1molar ratio)(Meldrum and Hyde, 2001),aragonite is formed instead of calcite(Falini et al., 2009);at lower magnesium concentrations,calcite is the primary precipitate with a variable amount of incorporated magnesium(10–30mole%).Organic compounds,including humic acids,polyacrylates,and organophosphorus compounds,have also been shown to change the distribution of calcium carbonate phases,delay or prevent precipitation, and in?uence the incorporation of ions such as magnesium into precipitating particles(Falini et al.,2009;Tang et al.,2008; Westin and Rasmuson,2005;Yang et al.,2001).

2.Experimental methods

2.1.Antiscalants

Phosphonate antiscalant samples were obtained from Dequest Water Management Additives,a subsidiary of Ther-mophos.The antiscalants included the penta-sodium salt of aminotri(methylene phosphonic acid),or ATMP;the hexa-potassium salt of hexamethylenediamine tetra(methylene-phosphonic acid),or HDTMP;and the hepta-sodium salt diethylenetriamine penta(methylene phosphonic acid),or DTPMP.Dequest refers to ATMP as DQ2006,to HDTMP as DQ2054,and to DTPMP as DQ2066,and these commercial names are used throughout this article.A polymer antiscalant was obtained from Coatex S.A.(France)and was determined by proton nuclear magnetic resonance(NMR)spectroscopy to be a copolymer containing acrylic acid,methacrylic acid,and itaconic acid.Total organic carbon(TOC)and total solids analysis were used to determine the mass and organic carbon concentrations of the antiscalant solutions.

2.2.Water data

The synthetic concentrates used for precipitation experi-ments were based on the chemical composition of a brackish groundwater in Maricopa County,Arizona,USA(Jurenka and Chapman-Wilbert,1996).Four different water compositions,shown in Table1,were tested to determine the effect of major ions such as magnesium and sulfate on calcium precipitation and antiscalant performance.To isolate the impacts of magnesium and sulfate on precipitation,each was added to the Simpli?ed Maricopa water composition.The data for the ‘‘complete’’Maricopa water shown in Table1were deter-mined based on a theoretical80%recovery and100%rejection of all ions.The simpli?cation of assuming100%rejection results in synthetic RO concentrate that is?ve times as concentrated as the feed.In an operating reverse osmosis system,membranes have rejections greater than99%for most ions.The initial pH of the synthetic concentrates was7.8.The individual ion concentrations were used to calculate the saturation index(SI)of a particular salt precipitate.The SI is the ratio of the ion activity product of the participating ions in solution to the solubility constant of the speci?c precipitate.If the SI is greater than1,the system is supersaturated and precipitation may occur.

2.3.Experimental design

The thermodynamic equilibrium software,PHREEQC(version 2.15.0.2697),was used to calculate saturation indices.This software has a database that uses the Pitzer equations for activity calculations(Pitzer,1991);these equations are considered more accurate than others for high ionic strength waters.Furthermore,PHREEQC software simulates chemical reactions,and in doing so,takes into account component complexation and competing ion effects,both of which are not considered in a manual comparison of the ion activity product and the solubility constant.Results from software calculations are described in the Supplementary material for the four water compositions in Table1.

The precipitation experiments were performed as500mL batch experiments in a jar test apparatus(Fisherbrand model 10008or Phipps&Bird Stirrer model7790–400).To start each precipitation experiment,synthetic concentrate was made in the laboratory by?rst adding antiscalant and then adding individual salts from stock solutions.Results were indepen-dent of the order of addition of antiscalant and salts.Before

w a t e r r e s e a r c h44(2010)2672–2684 2674

precipitation,the total carbonate was increased from16mM to42mM HCO3à(2560mg/L)to prevent the availability of carbonate from limiting calcium precipitation and to stabilize the pH.The pH was increased to10.5with6M NaOH. Increasing the pH to10.5allowed signi?cant precipitation even in the presence of antiscalants.

The separation step was performed using0.1m m pore size Millipore nitrocellulose membranes in either a dead-end pressurized(0.5bar)cell with a stir bar or using a Millipore glass?lter holder assembly(47mm diameter,300mL?lter holder)under vacuum.The dead-end?ltration cell was used with a digital mass balance to measure?ltrate(permeate)?ux. Samples?ltered with the vacuum assembly were analyzed for dissolved ions.

2.4.Analytical methods

Inductively coupled plasma atomic emission spectroscopy (ICP-AES)was used to analyze cation concentrations.

A Spectro Ciros CCD Model(Spectro AI GmbH)was used with Smart Analyzer data acquisition software(version3.2,1995–2000).Samples were analyzed for calcium,magnesium, barium,and iron.Standards were made with appropriate sodium chloride additions to avoid ion effects on ICP concentration results.Samples were prepared in15mL screw-cap polypropylene centrifuge tubes with concentrated nitric acid added for a?nal concentration of1.5%(v/v).If necessary,samples were stored at4 C for no longer than2 weeks before analysis.Some calcium and magnesium measurements were made using standard titrations for calcium and hardness with ethylenediaminetetraacetic acid (EDTA)(Eaton et al.,2005).

All pH measurements were taken with a Thermo Electron Corp.pH meter(Orion720At),calibrated with three buffers (pH4,7,and10standard buffers).The pH of a solution changes with ionic strength(Baumann,1973;Wiesner et al.,2006). Therefore,0.14M sodium chloride was added to each pH buffer to account for experimental solution ionic strength. Based on previous work(Wiesner et al.,2006),0.14M NaCl causes a decrease in pH of no more than0.1pH units in the standard buffers,and therefore no recalculation of buffer pH was performed.pH values are reported as recorded based on pH meter calibration with the salted buffers.

Particle size distributions were obtained using a laser granulometer Mastersizer S(Malvern Instruments).The Mastersizer S is a static laser light scattering instrument with associated computer software to convert the data to a relative volume distribution based on equivalent spherical diameter.

A polydisperse deconvolution algorithm and the Fraunhofer theory translated the detected light scattering data(diffrac-tion intensity as a function of diffraction angle)into a best-?t particle size distribution.

An ion chromatography system(Metrohm700series, column Metrosep A Supp5,150/4.0mm)was used to measure sulfate concentration.Some sulfate measurements were made with a Hach Ratio/XR turbidimeter(Standard Method 4500-SO42àE)(Eaton et al.,2005).

Two different scanning electron microscopes(SEM),a LEO 1530and a Hitachi S-5500,were used to obtain images of the precipitates.Both SEMs were equipped with energy dispersive X-ray(EDX)elemental analysis.Samples were mounted on adhesive carbon tabs;precipitates were placed directly onto the carbon tab or were on a nitrocellulose micro?lter that was placed on the carbon tab.All samples were sputter coated with silver.Samples used for SEM analysis were taken from a set of repeat precipitation experiments performed under identical conditions as those performed to obtain particle size distribution measurements.

3.Results and discussion

3.1.SEM imaging and EDX analysis of precipitated particles without antiscalants

Scanning electron microscopy(SEM)images of the precipi-tates formed in the absence of antiscalants are shown in Fig.1. The morphology of the precipitated particles is similar for the Simpli?ed Maricopa and the Simpli?ed MaricopatNa2SO4 water compositions(Fig.1a&b).The precipitates from the Simpli?ed MaricopatMgCl2and Complete Maricopa water compositions(Fig.1c&d)are also quite similar to each other but considerably different from the?rst two.The addition of sulfate to the Simpli?ed Maricopa water did not appear to signi?cantly affect particle morphology,but the addition of magnesium changed and controlled morphology for precipi-tates of both the Simpli?ed MaricopatMgCl2and the Complete Maricopa waters.

In this study,no additional clumping or aggregation of particles was observed when sulfate was present(Fig.1b), while the addition of magnesium dramatically changed the particle morphology(Fig.1c&d).The absence of particle aggregation can be rationalized by the lower concentration of sulfate(10mM)in our studies than in the studies of Falini et al. (2009)(33–330mM).A positive correlation between increased sulfate concentration and increased calcium carbonate particle aggregation has been reported(Kralj et al.,2004). Spherical vaterite and rhombohedral calcite are both observed for precipitated solutions of the Simpli?ed Maricopa and the Simpli?ed MaricopatNa2SO4water compositions.For the Simpli?ed MaricopatMgCl2and Complete Maricopa water compositions,the molar ratio of magnesium to calcium is approximately0.6:1,and calcite with incorporated magne-sium,known as magnesian calcite,is predicted to form rather than aragonite(Falini et al.,2009;Meldrum and Hyde,2001). Both magnesium-containing waters display rough,spheri-cally shaped particles as well as polycrystalline particles having two or more branches emitting from a central point and radial symmetry.Previous studies report similar magnesium-calcite particle morphologies(Loste et al.,2003; Meldrum and Hyde,2001).

Energy dispersive X-ray(EDX)analysis was performed on each of the precipitated water composition samples.Both weight percent and atomic percent values were recorded for each element.The Simpli?ed Maricopa water only contained calcium carbonate as a precipitate;calcium,oxygen,and carbon were included in the analysis.Analysis of the cubic or rhombohedral particles indicated that the elemental compo-sition was similar to that of anhydrous calcium carbonate (CaCO3).Elemental weight percentage values re?ected the

w a t e r r e s e a r c h44(2010)2672–26842675

expected anhydrous composition of 40wt%calcium,12wt%carbon,and 48wt%oxygen;a comparison of the expected elemental composition for different calcium carbonate poly-morphs is shown in Table 2.An analysis of the amorphous spheres indicated the elemental composition of mono-hydrocalcite with some anhydrous calcium carbonate.When magnesium was added to the Simpli?ed Maricopa water,magnesium concentrations in the particles were between 1.5

and 9.5%(wt),and EDX data indicated calcium carbonate polymorphs of anhydrous calcium carbonate and mono-hydrocalcite.Similar incorporated magnesium concentra-tions by weight were reported by Meldrum and Hyde (2001)and Loste et al.(2003).

Elemental analysis of the precipitates in Simpli?ed Maricopa tNa 2SO 4water revealed several calcium carbonate phases;the large amorphous spheres had the elemental composition of ikaite,while the small smooth spheres appeared to be monohydrocalcite,and the rhomboids had the elemental composition of anhydrous calcium carbonate.The sulfate anion is not known to stabilize metastable phases of anhydrous calcium carbonate (vaterite or aragonite),and calcite is typically formed (Falini et al.,2009);therefore,the rhombohedral particles are predicted to be calcite.Even though sulfate was not predicted to precipitate (as gypsum),a small amount of sulfur (0.21–0.43wt%or 0.15–0.25mole%)was present in all particles https://www.doczj.com/doc/8918071302.html,pared to magne-sium,the addition of sulfate caused only minor changes to particle morphology.

An analysis of the precipitates of the Complete Maricopa water revealed both anhydrous and amorphous calcium carbonate phases.The larger spheres yielded elemental compositions for both phases when measurements were taken at different points on the same particle,while smaller

Fig.1–SEM images of precipitates from four different water compositions tested without antiscalant present;(a)Simpli?ed Maricopa,(b)Simpli?ed Maricopa D Na 2SO 4,(c)Simpli?ed Maricopa D MgCl 2,and (d)Complete Maricopa.Precipitation was performed at pH 10.5for 1h.The image in (c)has a magni?cation of 10,000,and the other three images have a magni?cation of 6,000.

w a t e r r e s e a r c h 44(2010)2672–2684

2676

spheres appeared to be primarily anhydrous calcium carbonate.Some particles with a rough rhombohedral shape were observed,and these particles were anhydrous calcium carbonate.The magnesium content ranged between 0.34and 3.44wt%;the small spheres had a magnesium content of 0.34–0.48wt%,while the bigger amorphous spheres and rhomboids had a magnesium content of 1.67–3.44wt%.All measurements showed small amounts of sulfur,barium,and ferric iron in the particles;sulfur content ranged from 0.10to 0.46wt%(0.06–0.33mole%),barium from 0.76to 0.97wt%(0.08–0.16mole%),and ferric iron from 0.48to 0.63wt%(0.17–0.26mole%).

3.2.SEM imaging and EDX analysis of precipitated particles in the presence of antiscalants

SEM images obtained for precipitates of the Simpli?ed Maricopa tNa 2SO 4with several different antiscalant types and concentrations are shown in Fig.2;results obtained for precipitates of the Simpli?ed Maricopa water were similar.The higher concentrations of DQ2066(56mg/L)and DQ2006(85mg/L)caused many small particles to form while most of the antiscalant types and concentrations tested did not signi?cantly change the spherical and rhombohedral geome-tries observed for the antiscalant-free solution pictured in Fig.1b.No effect of antiscalant addition on particle morphology was observed for Coatex (0.8&50mg/L),9mg/L

DQ2006,5mg/L DQ2066,or 4mg/L DQ2054,while a slight distortion of the rhombohedral geometry was observed for 43mg/L DQ2054(Fig.2d).

The ?rst stage of precipitation is nucleation,followed by particle growth;the small particles represent the ?rst stage of crystallizing nuclei.Under the conditions considered in this paper,where the precipitation control of antiscalants is overcome,precipitation does occur but is largely arrested during the ?rst stage as antiscalants adsorb onto crystal growth sites and prevent particle growth.Therefore,the number of large particles decreases,and the number of small particles increases dramatically,as observed for DQ2066and DQ2006in Fig.2a–c.In the case of DQ2066,the small particles appear to attach to the larger particles,creating a rough surface on the normally smooth spheres.Such attachment was not observed for DQ2006(Fig.2c).While the formation of larger particles was not observed when antiscalant was added to the Simpli?ed Maricopa water (data not shown),the result is different for the three other waters,as is evident in Fig.2c for DQ2006and DQ2066;both small and larger particles are observed for several antiscalant types and concentrations.

EDX analysis of precipitated particles of Simpli?ed Maricopa with 85mg/L DQ2006indicated that calcium carbonate precipitated primarily as monohydrocalcite,with an average composition of 32.6wt%calcium,11.7wt%carbon,and 55.5wt%oxygen and a phosphorus content of 0.60wt%(0.37mole%).Similar results were obtained for

precipitated

Fig.2–SEM images of precipitated solutions of the Simpli?ed Maricopa D Na 2SO 4water composition with different

antiscalants;(a,b)56mg/L DQ2066at magni?cations of 6,000and 53,800,respectively,(c)85mg/L DQ2006,and (d)43mg/L DQ2054.Precipitation was performed at pH 10.5for 1h.

w a t e r r e s e a r c h 44(2010)2672–2684

2677

particles of Simpli?ed MaricopatNa2SO4;the primary precipitate phase was monohydrocalcite,with a decreased carbon content(average was10.2wt%)due to incorporation of some sulfate anions into the calcium carbonate crystals (0.29–0.95wt%).Results for precipitates in the samples treated with56mg/L DQ2066were different;the calcium content was slightly higher than that expected for the anhydrous phase, while the carbon content decreased;the average elemental composition was46.2wt%calcium, 6.6wt%carbon,and 47.2wt%oxygen.There are several possible explanations for the change in elemental composition;while the change in precipitate morphology(shown in Fig.S2in the Supplementary material)suggests the precipitate is different, EDX elemental analysis is only semi-quantitative,and the results do not provide an absolute basis for determining the crystal structure.Previous work has shown that the anionic phosphate groups of phosphonate antiscalants can replace carbonate groups in the crystal lattice if the distance between phosphate groups within the phosphonate molecule is similar to that of the carbonate groups(Nygren et al.,1998).The distance between phosphate groups in DQ2066may be similar to that of carbonate molecules in anhydrous calcium carbonate,while the phosphate groups of DQ2006may be too close together to be incorporated into the crystal lattice.Such differences in antiscalant structure result in different calcium carbonate phases during precipitation.The calcium phos-phate precipitate,hydroxyapatite,typically has a hexagonal crystal structure,while anhydrous calcium carbonate tends to form a trigonal rhombohedral structure.Changes in precipi-tate composition may have occurred due to changes in crys-tral structure and exclusion of carbonate groups in favor of an increased calcium content.

Another possible explanation for the increase in calcium content and decrease in carbon content could be in the EDX analysis itself;SEM images(shown in Fig.S2in the Supplementary material)revealed small,non-dense,aggre-gated particles on the surfaces of the larger calcium carbonate particles.These particles may be calcium-antiscalant hydroxyapatite-like precipitates or co-precipitating anti-scalant and calcium carbonate.During EDX analysis,the X-rays penetrate several micrometers into the sample,and in this case,the particles on the surface may have made up the majority of the X-ray penetration depth.This may have skewed the EDX results to show an elemental composition with more calcium and less carbon.

Finally,the formation of basic calcium carbonate(BCC) (CaCO3$Ca(OH)2)may have occurred;the elemental compo-sition of this metastable calcium carbonate polymorph is 46wt%calcium,6.9wt%carbon and46wt%oxygen.Basic calcium carbonate is typically formed through carbonation of a Ca(OH)2slurry and decomposition yields Ca(OH)2, calcite and aragonite(Ahn et al.,2003;Matsushita et al., 1993).While BCC is not usually observed under precipita-tion conditions similar to those used,the presence of DQ2066may have allowed the polymorph to form and remain stable.

SEM images of precipitates of the Complete Maricopa water composition are presented in Fig.3.For the Complete Maricopa water,9mg/L DQ2006,56mg/L DQ2066,and50mg/L Coatex resulted in the formation of many particles approximately 100–200nm in diameter,as well as larger polycrystalline symmetrical crosses and rods.No change in particle morphology was observed for DQ2054(4&43mg/L),5mg/L DQ2066,and0.8mg/L Coatex.For85mg/L DQ2006,only small particles were observed.Similar images were obtained for particles of the Simpli?ed MaricopatMgCl2water composi-tion;no effect of antiscalant on particle morphology was observed for4and43mg/L DQ2054or0.8mg/L Coatex,but all other antiscalant types and concentrations resulted in formation of many small particles(100–200nm)and a small number of larger polycrystalline particles.

EDX analysis was performed on several particles of the Simpli?ed MaricopatMgCl2water composition with56mg/L DQ2066.Both small and large particles were analyzed,and the elemental composition was signi?cantly different in the two types of particles.The composition distribution for each particle type was averaged from four separate measurements taken on each type of particle.The addition of antiscalant to the Simpli?ed MaricopatMgCl2water did not increase magnesium precipitation.The small particles contained1.8% (wt)magnesium,29.3%calcium,60.4%oxygen,and8.4% carbon,while the larger particles contained3.4%magnesium, 41.2%calcium,49.2%oxygen,and6.2%carbon.The atomic ratio of calcium to oxygen of the small particles indicates that the small particles are ACC with at least one unit of hydration, while the ratio for the larger particles indicated that the particles are primarily anhydrous calcium carbonate.The magnesium content of the anhydrous calcium carbonate (larger particles)was higher,indicating a greater inclusion of magnesium as the ACC was transformed into a crystalline anhydrous form.Similarly,results from a sample of precipi-tated Simpli?ed MaricopatMgCl2with85mg/L DQ2006 showed both anhydrous and amorphous calcium carbonate. The DQ2006sample contained a slightly higher average magnesium content(5.5wt%).Phosphorus weight content ranged between0.47and0.62wt%;all phosphorus measure-ments for all samples and antiscalants were similar and within this range.The only source of phosphorus in the synthetic solutions was antiscalant.Phosphonate anti-scalants have been shown to adsorb onto and co-precipitate with anhydrous calcium carbonate(Jonasson et al.,1996; Kan et al.,2005;Nygren et al.,1998).Antiscalant precipita-tion was observed for all four water compositions tested.

EDX analysis was also performed on precipitated particles of the Complete Maricopa water with85mg/L DQ2006.Similar to the results obtained for the Simpli?ed Maricopa and Simpli?ed MaricopatNa2SO4water compositions,the primary phase of calcium carbonate was monohydrocalcite. Both magnesium and sulfur content increased over that observed for the antiscalant-free Complete Maricopa sample; the average particle composition for the DQ2006sample was 4.9wt%magnesium,34.2wt%calcium,51.2wt%oxygen, 6.6wt%carbon,0.8wt%sulfur,0.48wt%barium,and0.35wt% ferric iron.The average particle composition for Complete Maricopa with56mg/L DQ2066had a similar distribution.In contrast to the increased magnesium and sulfate precipita-tion,the barium and ferric iron content decreased slightly when antiscalant was added;the DQ2006sample contained 0.05–0.12mole%barium and0.10–0.22mole%ferric iron(down from0.08–0.16mole%and0.17–0.26mole%).

w a t e r r e s e a r c h44(2010)2672–2684 2678

3.3.Analysis of remaining dissolved ions after precipitation

The effects of antiscalant type and concentration on calcium precipitation were quite different than those observed for magnesium and sulfate precipitation.In general,the addition of antiscalant before precipitation caused a reduction in calcium precipitation,with higher antiscalant concentrations causing a greater decrease in calcium precipitation.Several concentrations of DQ2006(4,20,and 85mg/L)and Coatex (2,10,and 50mg/L)were tested with the Complete Maricopa water data set for a precipitation time of 60minutes,and the results for ?nal dissolved calcium are shown in Fig.4.The remaining dissolved calcium increased with increased anti-scalant concentration for both antiscalants.The results in Fig.4illustrate the differences in precipitation control among antiscalant products (Plottu-Pecheux et al.,2002;Semiat et al.,2003;Shih et al.,2006);antiscalant concentrations typically added to an RO feed are between 0.5and 3mg/L,resulting in concentrate antiscalant concentrations between 2.5and 15mg/L (for an 80%recovery).

Additional experiments were performed for a 60min precipitation time for all four antiscalants and water compo-sitions,and the results for ?nal dissolved calcium are recorded in Table 3.Calculated values are shown for the log saturation index (log SI)of calcite after precipitation;the calculations

were performed in PHREEQC.The ?rst row of data displays the effect of sulfate and magnesium ions on calcium carbonate precipitation;the addition of these two ions to the Simpli?ed Maricopa water composition caused a decrease in

calcium

Fig.3–SEM images of precipitated solutions of the Complete Maricopa water composition with different antiscalants;(a,b)9mg/L DQ2006for magni?cations of 28,000and 68,800,respectively,(c)56mg/L DQ2066,and (d)50mg/L Coatex.Precipitation pH was

10.5.

Fig.4–Increase in ?nal dissolved calcium with increasing antiscalant concentration for antiscalants DQ2006and Coatex in the Complete Maricopa water.Batch

precipitation experiments performed at pH 10.5for 60min with excess total carbonate.

w a t e r r e s e a r c h 44(2010)2672–2684

2679

precipitation.The effect appears to be additive as calcium precipitation decreased further when both ions were present in the Complete Maricopa water.Falini et al.(2009)showed that the addition of sulfate,along with other ions such as sodium and potassium,decreases calcium precipitation.Magnesium has been shown to signi?cantly in?uence calcium carbonate precipitation by changing lattice structure and adsorbing onto crystal growth sites (Chakraborty et al.,1994).However,minor ions present in the Complete Maricopa water,such as barium and ferric iron,might have contributed to the reduction in calcium precipitation,as they are also known to block calcium carbonate growth sites.The decrease in calcium precipitation may also stem from the formation of aqueous complexes such as calcium sulfate (complex represented 5.5%of aqueous calcium in PHREEQC calculations)and magnesium carbonate (complex represented 20.2%of aqueous carbonate);these complexes reduce the free calcium ion activity and therefore increase the calcium solubility,resulting in a decrease in predicted and measured calcium precipitation.

The log SI increased consistently from left to right in most of the rows in Table 3.In some cases,the log SI decreased slightly for the Simpli?ed Maricopa tMgCl 2even though the calcium concentration increased;this result stems from the effect of magnesium on the calculated activity coef?cients for calcium and carbonate.The presence of magnesium increases calcium carbonate solubility and thus decreases the activity coef?cients of the ionic components in solution.

The subsequent rows in Table 3show the effect of the four different antiscalants on calcium precipitation.In general,the remaining dissolved calcium increased with increasing water complexity,following the trend for solu-tions with no antiscalant.However,the concentration of remaining dissolved calcium varied signi?cantly among the four antiscalants,and two antiscalants (DQ2066and Coatex)performed worse (i.e.,DQ2066and Coatex allowed more calcium precipitation)in the Complete Maricopa water.The presence of antiscalants DQ2054and DQ2006increased the log SI for calcium carbonate from 0.63to 1.23–1.98for the Simpli?ed Maricopa water;similar increases were observed for the other three water compositions.For anti-scalant DQ2006,the calcite ion activity product (the product of the activity of calcium and the activity of carbonate in solution)in the Simpli?ed Maricopa water increased by a factor of 20over that of the antiscalant-free solution.The phosphonate antiscalant DQ2006appeared to outperform the other antiscalants in preventing calcium precipitation for both the Simpli?ed Maricopa and Complete Maricopa water compositions,while Coatex prevented the most calcium precipitation for the other two water compositions.DQ2054appears to have a minor effect on calcium precipi-tation,while DQ2066was not effective at preventing precipitation in the Complete Maricopa water.

An analysis of ions remaining in solution 30min after initiating precipitation of the Complete Maricopa water data set revealed an increase in magnesium precipitation with the addition of antiscalant;antiscalants DQ2006(20and 85mg/L)and DQ2054(43mg/L)were tested.Magnesium precipitation increased from 24%of the original solution concentration in the control (no antiscalant)solution to 30%for both DQ2006concentrations.For precipitation experiments allowed to proceed for 60minutes,solutions containing antiscalant DQ2006,DQ2066,or Coatex resulted in a similar increase in magnesium precipitation.The presence of DQ2054resulted in the same amount of magnesium precipitation as the control solution,and in the Simpli?ed Maricopa tMgCl 2water,magnesium precipitation was equivalent for antiscalant-dosed and antiscalant-free solutions.A similar trend was observed for sulfate;all antiscalant-dosed precip-itated solutions resulted in lower ?nal dissolved sulfate concentrations than the antiscalant-free solution.For the Maricopa tNa 2SO 4water,the addition of antiscalant caused the sulfate precipitation to increase from 2%to 5%,and for the Complete Maricopa water,sulfate precipitation increased from 6%to 9%.The trends observed for magnesium and sulfate precipitation with antiscalant addition were consis-tent with EDX data.

Kan et al.(2005)reported that the relationship between dissolved antiscalant concentration and extent of anti-scalant co-precipitation depended on the concentration range of the antiscalant.For antiscalant concentrations less than approximately 0.1–0.3mM,antiscalants adsorb onto calcite following a Langmuir-type isotherm (Kan et al.,2005).The range of antiscalants tested in this research was 0.01–0.3mM.Nygren et al.(1998)reported that phosphonate antiscalants adsorb onto calcite surfaces,particularly at steps in the crystal structure,by replacing two carbonate molecules.In addition to the co-precipitation of antiscalant molecules,magnesium ions are known to incorporate into anhydrous calcium carbonate,and sulfate is a common co-precipitant with calcium carbonate (Falini et al.,2009).

w a t e r r e s e a r c h 44(2010)2672–2684

2680

However,a study on co-precipitation of calcium carbonate and calcium sulfate indicated that the solubility of a co-precipitating mixture follows the solubility of calcium sulfate rather than that of calcium carbonate(Sudmalis and Sheikholeslami,2000).Additionally,Meldrum and Hyde (2001)found no correlation between increased magnesium incorporation and organic additive addition for malic and citric acid.

The results presented here are contradictory to some of the previous work mentioned above.The antiscalant concentration range tested lies within the range expected to have a Langmuir adsorption behavior.However,the precipi-tation of calcium did not follow the solubility of calcium sulfate,which is not predicted to precipitate,but followed the supersaturation level of calcium carbonate.This result may indicate incorporation of sulfate anions into the already existing calcium carbonate precipitating crystals,instead of co-precipitation of individual crystals of calcium carbonate and gypsum.The SEM images for the Simpli?ed Mar-icopatNa2SO4and Complete Maricopa water compositions support this conclusion;no needle-like particles were observed in antiscalant-dosed precipitated solutions(Figs.2 and3),while Sudmalis and Sheikholeslami(2000)reported the presence of both needle-like CaSO4and rhombohedral CaCO3.The hydration shell of a magnesium ion must be removed before magnesium can be incorporated into the calcium carbonate crystal lattice and,therefore,represents an energy barrier to precipitation.The adsorption of a phos-phonate antiscalant,which typically coordinates more strongly to divalent cations than similar compounds containing carboxylic acid moieties(such as malic and citric acid),may allow more magnesium and sulfate to be incor-porated into the precipitating calcium carbonate.The anti-scalant may coordinate with magnesium,thereby lowering the energy barrier to removal of the hydration shell. The phosphonate antiscalant may also coordinate with the calcium ions(in place of carbonate anions),disrupting the crystal structure and delaying the transformation of amor-phous calcium carbonate to anhydrous calcium carbonate, which could allow increased inclusion of sulfate anions into the forming crystal lattice.3.4.Particle size distributions of precipitated solutions

A comparison of measured particle size distributions for two water compositions,Simpli?ed MaricopatNa2SO4and Complete Maricopa,and the four antiscalants is shown in Fig.5.Most of the curves obtained were bimodal,with some trimodal distributions.The relationship between antiscalant type and particle size distribution was the same for the Simpli?ed Maricopa and the Simpli?ed MaricopatMgCl2 water compositions and for the Simpli?ed MaricopatNa2SO4 and Complete Maricopa water compositions.For both groups of water types,antiscalants DQ2054and Coatex had particle size distributions similar to the antiscalant-free precipitated solution,while antiscalants DQ2006and DQ2066caused the particle size distributions to shift.For the Simpli?ed MaricopatNa2SO4water(Fig.5a),DQ2006and DQ2066caused a decrease in modal particle diameter for the mode with the largest relative volume(%).In addition,the relative volume of small particles(i.e.,those between0.1and1m m)increased for DQ2006,indicating a large increase in the number of small particles formed.As con?rmed through SEM imaging,the addition of DQ2006or DQ2066to the Simpli?ed MaricopatNa2SO4water caused the formation of particles that were primarily between100and200nm in diameter,with some larger particles(w10–15m m in diameter)observed.The modal particle diameter for the largest mode of the DQ2006 and DQ2066curves in Fig.5a is slightly greater than that observed by SEM.This discrepancy might be due to smaller particles adhering to the larger particles.Another possible interpretation is that the light scattering theory is based on spherical particles,and the SEM images clearly show nonspherical particles.The presence of DQ2054or Coatex did not cause the formation of these small,nanometer scale particles,so the particle size distributions for these two anti-scalants were quite similar to the case with no antiscalant.

Antiscalants DQ2006and DQ2066caused an increase in the modal particle diameter in the Complete Maricopa water (Fig.5b),and antiscalants DQ2006,DQ2066,and Coatex caused the formation of small particles(100–200nm).The curve for the antiscalant-free sample also indicated the presence of small particles,but SEM images showed no particles in

the

Fig.5–Variation in particle size distribution with antiscalant type for(a)Simpli?ed Maricopa D Na2SO4water composition and(b)Complete Maricopa water composition.Precipitation pH was10.5.

w a t e r r e s e a r c h44(2010)2672–26842681

sub-micron range.This discrepancy is most likely due to the algorithm used to deconvolute the static light scattering data,and the particles are best represented by a monodisperse particle size distribution.No shift in the particle size distri-bution was observed for Coatex,although the relative volume of small particles (?rst mode)increased above that observed for DQ2054.The increase in modal particle diameter for DQ2006and DQ2066appears to be caused by the attachment of many small particles to the larger rod-or cross-shaped symmetrical particles.While small particles are observed in the Coatex sample,the surfaces of the larger particles are smooth and relatively no small particle attachment was observed.This lack of small particle attachment in the Coatex sample may also explain the increase in relative volume of the small particle mode,which is not observed for DQ2006or DQ2066;more small particles were actually measured during analysis of the Coatex sample than for the other two anti-scalants.The addition of DQ2054resulted in no change in the modal particle diameter.

The effect of antiscalant concentration on particle size diameter is shown in Fig.6for antiscalant DQ2006in the Complete Maricopa water composition.All particle size distributions were bimodal;as antiscalant concentration increased,the modal particle diameter of the second mode increased,and the relative volume of the ?rst mode decreased.The sample with no antiscalant did not show two distinct modes but rather two modal diameters within a con-nected particle size distribution.SEM imaging of the antiscalant-free Complete Maricopa water (Fig.1d)showed a range of particle sizes and shapes,while the addition of antiscalant DQ2006resulted in many small spherical particles and some larger particles shaped as a symmetrical ?gure eight.An increase in antiscalant concentration caused an increase in the formation of small particles,which attach to the larger particles and increase the measured modal particle diameter of the second mode.The decrease in the relative volume of the ?rst mode as antiscalant concentration increases is most likely due to an increased attachment of smaller particles to the larger particles,reducing the volume of small particles measured.

3.5.Micro?ltration of precipitated solutions

Micro?ltration (0.1m m)was used to separate the precipi-tated salts from the remaining dissolved ions in solution;a comparison of the ?ux decline for the four water compo-sitions is shown in Fig.7for Coatex and DQ2066.The antiscalant-free solutions of Simpli?ed Maricopa and Simpli?ed Maricopa tNa 2SO 4resulted in approximately 20%?ux decline over seven minutes,while the ?ux decreased approximately 10%for Simpli?ed Mar-icopa tMgCl 2and Complete Maricopa waters (data not shown).The presence of antiscalant during precipitation caused at least a small increase in ?ux decline for all samples.A greater loss of ?ux was observed for the Simpli?ed Maricopa and the Simpli?ed Maricopa tNa 2SO 4water compositions for all four antiscalants.The presence of the Coatex antiscalant caused the greatest ?ux decline for the Simpli?ed Maricopa tNa 2SO 4water;approximately 70%of the ?ux was lost within 6.7min.A smaller loss in

?ux

Fig.6–Effect of DQ2006antiscalant concentration on

particle size distribution for the Complete Maricopa water composition.Precipitation pH was

10.5.

Fig.7–Normalized micro?ltration (0.1m m pore size)?ux (J /J o )of precipitated solutions containing antiscalants (a)50mg/L Coatex and (b)56mg/L DQ2066.

w a t e r r e s e a r c h 44(2010)2672–2684

2682

occurred for the three other water compositions.For the DQ2066antiscalant(Fig.7b),the?ux decline for the Simpli?ed Maricopa and the Simpli?ed MaricopatNa2SO4 water compositions was identical,as was the?ux decline for two the magnesium-containing water compositions.The ?ux decline for the former group was approximately45%, while the?ux decline for the latter was20%.These results mimic those of the particle size distributions,indicating that the presence of substantial concentrations of small particles is a principal determinant of the?ux decline.The addition of antiscalant to deionized water did not cause a decrease in water?ux;therefore,the?ux decline observed for precipitated solutions with antiscalant was not caused by adsorption of the antiscalant to the membrane,but by the changes in particle size distribution,particle and cake morphology,and particle coadhesion caused by the anti-scalant.Micro?ltration results for DQ2054(data not shown) showed no signi?cant difference in?ux decline between the Simpli?ed Maricopa,Simpli?ed MaricopatNa2SO4,and Complete Maricopa water compositions;the?ux decline for the Simpli?ed MaricopatMgCl2water composition was slightly less than the other three waters.The results for DQ2006were similar to those of DQ2066and are described in Fig.S3of the Supplementary material.

4.Conclusions

Antiscalants DQ2006and Coatex displayed the best precip-itation control for the four water compositions studied.In the synthetic water compositions tested,the primary precipitate was calcium carbonate,and incorporation of magnesium and sulfate into precipitating calcium carbonate was observed.Antiscalant addition caused an increase in magnesium and sulfate precipitation and a decrease in calcium precipitation.

In addition to affecting the amount of precipitation,the antiscalants changed the precipitate particle size distribu-tion,particle morphology,and calcium carbonate phases formed.SEM imaging revealed the formation of small (100–200nm)particles with certain antiscalant types and concentrations,and EDX analysis showed antiscalant type-dependent calcium carbonate phase formation.Micro-?ltration performance was highly dependent on water composition,and the Complete Maricopa water resulted in the smallest?ux decline;antiscalant-free precipitated solu-tions had the same or better?ux performance than antiscalant-dosed samples.

Based upon these results,the presence of antiscalant in an RO concentrate could signi?cantly affect a concentrate treatment process based on precipitation and solid/liquid separation(?ltration).The antiscalant would be concentrated along with the dissolved salts and could reduce calcium removal ef?ciency,as well as affect the solid/liquid separa-tion process.Even small antiscalant concentrations can reduce the amount of calcium precipitated and will adsorb onto precipitating particles.The development of an RO concentrate treatment process must include tests with anti-scalants to determine the effect of the antiscalant on the treatment process;for the concentrate treatment process and water composition considered,removal of the antiscalants, through a process such as coagulation or oxidation,prior to salt precipitation would allow optimal precipitation and ?ltration of a concentrate.

Acknowledgements

The authors would like to thank Elise Barbot for her help with obtaining the particle size measurements and Mr.Daniel Dreyer and Professor Christopher Bielawski for their help in obtaining NMR data to determine the composition of the polymer antiscalant.The authors thank the U.S.National Science Foundation Graduate Research Fellowship and International Research and Education in Engineering programs for funding(CBET0553957).The authors acknowl-edge the Water Research Foundation(Project#4061).The views herein are those of the authors and not of the sup-porting agencies.

Supplementary data

Supplementary data associated with this article can be found, in the online version,at doi:10.1016/j.watres.2010.01.034.

r e f e r e n c e s

Ahn,J.W.,Joo,S.M.,Kim,H.S.,Kim,D.H.,Kim,J.P.,Kim,H.,2003.

In?uence of?nal crystal with synthesis of Basic Calcium

Carbonate by carbonation process.In:Kim,H.S.,Park,S.Y., Lee,S.W.(Eds.),Eco-Materials Processing&Design,vol.439, pp.244–253.

Baumann,E.W.,1973.Determination of pH in concentrated salt solutions.Analytica Chimica Acta64,284–288.

Brecˇevic′,L.,No¨thig-Laslo,V.,Kralj,D.,Popovic′,S.,1996.Effect of divalent cations on the formation and structure of calcium carbonate polymorphs.Journal of the Chemical Society-

Faraday Transactions92(6),1017–1022.

Bruneval,F.,Donadio,D.,Parrinello,M.,2007.Molecular dynamics study of the solvation of calcium carbonate in

water.Journal of Physical Chemistry B111(42),12219–12227. Chakraborty,D.,Agarwal,V.K.,Bhatia,S.K.,Bellare,J.,1994.

Steady-state transitions and polymorph transformations in continuous precipitation of calcium carbonate.Industrial& Engineering Chemistry Research33(9),2187–2197. Clarkson,J.R.,Price,T.J.,Adams,C.J.,1992.Role of metastable phases in the spontaneous precipitation of calcium carbonate.

Journal of the Chemical Society-Faraday Transactions88(2), 243–249.

Eaton,A.D.,Clesceri,L.S.,Rice,E.W.,Greenberg,A.E.(Eds.),2005.

Standard Methods for The Examination of Water and

Wastewater.American Public Health Association,American Water Works Association,Water Environment Federation,

Washington,D.C.

Falini,G.,Fermani,S.,Tosi,G.,Dinelli,E.,2009.Calcium carbonate morphology and structure in the presence of

seawater ions and humic acids.Crystal Growth&Design9

(5),2065–2072.

Jonasson,R.G.,Rispler,K.,Wiwchar,B.,Gunter,W.D.,1996.Effect of phosphonate inhibitors on calcite nucleation kinetics as

w a t e r r e s e a r c h44(2010)2672–26842683

a function of temperature using light scattering in an

autoclave.Chemical Geology132(1–4),215–225.

Jurenka,R.A.,Chapman-Wilbert,M.,1996.Maricopa groundwater treatment study.Water Treatment Technology Program

Report No.15.U.S.Department of the Interior,Bureau of

Reclamation.https://www.doczj.com/doc/8918071302.html,/pmts/water/media/pdfs/ report015.pdf.

Kan,A.T.,Fu,G.M.,Tomson,M.B.,2005.Adsorption and precipitation of an aminoalkylphosphonate onto calcite.

Journal of Colloid and Interface Science281(2),275–284. Kralj,D.,Kontrec,J.,Brecˇevic′,L.,Falini,G.,No¨thig-Laslo,V.,2004.

Effect of inorganic anions on the morphology and structure of magnesium calcite.Chemistry–A European Journal10(7), 1647–1656.

Lam,R.S.K.,Charnock,J.M.,Lennie,A.,Meldrum,F.C.,2007.

Synthesis-dependant structural variations in amorphous

calcium carbonate.CrystEngComm9(12),1226–1236. Loste,E.,Wilson,R.M.,Seshadri,R.,Meldrum,F.C.,2003.The role of magnesium in stabilising amorphous calcium carbonate and controlling calcite morphologies.Journal of Crystal

Growth254(1,2),206–218.

Matsushita,I.,Hamada,Y.,Suzuki,T.,Nomura,Y.,Moriga,T., Ashida,T.,Nakabayashi,I.,1993.Crystal structure of basic calcium carbonate and its decomposition process in water.

Nippon Seramikkusu Kyokai Gakujutsu Ronbunshi–Journal of the Ceramic Society of Japan101(12),1335–1339.

Meldrum,F.C.,Hyde,S.T.,2001.Morphological in?uence of magnesium and organic additives on the precipitation of

calcite.Journal of Crystal Growth231(4),544–558. Nancollas,G.H.,Sawada,K.,1982.Formation of scales of calcium carbonate polymorphs–the in?uence of

magnesium ion and inhibitors.Journal of Petroleum

Technology34(3),645–652.

Nebel,H.,Neumann,M.,Mayer,C.,Epple,M.,2008.On the structure of amorphous calcium carbonate–a detailed study by solid-state NMR spectroscopy.Inorganic Chemistry47(17), 7874–7879.

Nygren,M.A.,Gay,D.H.,Catlow,C.R.A.,Wilson,M.P.,Rohl,A.L., 1998.Incorporation of growth-inhibiting diphosphates into steps on the calcite cleavage plane surface.Journal of the

Chemical Society-Faraday Transactions94(24),3685–3693. Pitzer,K.S.,1991.Activity Coef?cients in Electrolyte Solutions.

CRC Press,Inc.,Boca Raton,FL.

Plottu-Pecheux,A.,Houssais,B.,Democrate,C.,Gatel,D., Parron,C.,Cavard,J.,https://www.doczj.com/doc/8918071302.html,parison of three antiscalants, as applied to the treatment of water from the River Oise.

Desalination145,273–280.Rahardianto,A.,Gao,J.,Gabelich,C.J.,Williams,M.D.,Cohen,Y., 2007.High recovery membrane desalting of low-salinity

brackish water:integration of accelerated precipitation

softening with membrane RO.Journal of Membrane Science 289,123–137.

Reddy,M.M.,Wang,K.K.,1980.Crystallization of calcium carbonate in the presence of metal-ions.1.Inhibition by

magnesium-ion at Ph8.8and25 C.Journal of Crystal Growth 50(2),470–480.

Sandia,2003.Desalination and water puri?cation roadmap-A report of the executive committee.DWPR Program Report#95.

U.S.Department of the Interior,Bureau of Reclamation and Sandia National Laboratories.https://www.doczj.com/doc/8918071302.html,/tbndrc/ roadmapreport.pdf(May25,2008).

Sawada,K.,1997.The mechanisms of crystallization and transformation of calcium carbonates.Pure and Applied

Chemistry69(5),921–928.

Semiat,R.,Sutzkover,I.,Hasson,D.,2003.Characterization of the effectiveness of silica anti-scalants.Desalination159, 11–19.

Service,R.F.,2006.Desalination freshens up.Science313,1088–1090. Shih,W.Y.,Gao,J.,Rahardianto,A.,Glater,J.,Cohen,Y., Gabelich,C.J.,2006.Ranking of antiscalant performance for gypsum scale suppression in the presence of residual

aluminum.Desalination196,280–292.

Sudmalis,M.,Sheikholeslami,R.,2000.Coprecipitation of CaCO3 and CaSO4.Canadian Journal of Chemical Engineering78(1), 21–31.

Tang,Y.M.,Yang,W.Z.,Yin,X.S.,Liu,Y.,Yin,P.W.,Wang,J.T., 2008.Investigation of CaCO3scale inhibition by PAA,ATMP and PAPEMP.Desalination228(1–3),55–60.

Wada,N.,Yamashita,K.,Umegaki,T.,1995.Effects of divalent cations upon nucleation,growth and transformation of

calcium carbonate polymorphs under conditions of double diffusion.Journal of Crystal Growth148(3),297.

Westin,K.J.,Rasmuson,A.C.,2005.Nucleation of calcium carbonate in presence of citric acid,DTPA,EDTA and

pyromellitic acid.Journal of Colloid and Interface Science282

(2),370–379.

Wiesner,A.,Katz,L.,Chen,C.C.,2006.The impact of ionic strength and background electrolyte on pH measurements in metal ion adsorption experiments.Journal of Colloid and

Interface Science301(1),329–332.

Yang,Q.F.,Liu,Y.Q.,Gu,A.H.,Ding,J.,Shen,Z.Q.,2001.

Investigation of calcium carbonate scaling inhibition and scale morphology by AFM.Journal of Colloid and Interface Science 240(2),608–621.

w a t e r r e s e a r c h44(2010)2672–2684 2684

相关主题
相关文档 最新文档