Zhao 2015

mik, NoFinl

SerpentiniteSulfuric acid leachingCO mineral carbonation

e (Cudiulfchit lee o

95% of iron was recovered from the leachate and leaching residues and valuable Fe-rich substances wereobtained as by-products. After the iron extraction, a ne Mg(OH) -rich powder could be prepared from

effectn captions foean, g

is another option where CO can be stored below ground by inject-

geophysical investigations need to be carefully carried out beforegeological sequestration (Zhang and Song, 2014). After plenty ofevaluations, scientists now believe that mineral sequestration isone of the most promising method for CO2 sequestration to manycountries (Goff et al., 2000). Moreover, it is very unlikely that CO2captured in a thermodynamically and geologically stable material

is extracted from Mg-rich silicate via solidsolid reaction usinglfate at tempera-o et al., 20for magn

extraction at relative low temperature, in which the magnion could be leached out from the ore into solution, and thnesium hydroxide can be easily prepared by adjusting pH vthe leachate (Alexander et al., 2007). Sulfuric acid, hydrochloricacid, nitric acid, formic acid and acetic acid have all been investi-gated as solvents and the work has showed sulfuric acid to bethe most effective reagent in serpentine dissolution (Teir et al.,2007a,b). Some activation pretreatments prior to the leaching pro-cess were demonstrated to improve mineral conversion(Maroto-Valer et al., 2005). Kim and Chung (2002) studied the

Corresponding author at: Mailbox 313, Northeastern University, Wenhua Road,Heping District, Shenyang, China.

E-mail address: [email protected] (Q. Zhao).

Minerals Engineering 79 (2015) 116124

Contents lists availab

Minerals En

els2

ing CO2 into deep geological formations. However, injection sitesrequire proper permeable geological formations thus limiting theapplication area of this approach: Furthermore, comprehensive

recyclable ammonium sulfate or ammonium bisutures of 400500 C (Nduagu et al., 2012b; Rom

An acid leaching process is another optionhttp://dx.doi.org/10.1016/j.mineng.2015.06.0020892-6875/ 2015 Elsevier Ltd. All rights reserved.13).esiumesiume mag-alue ofsequestrations are three candidates of large-scale technologiesfor CCS. In ocean sequestration, CO2 could be directly injected intothe ocean by moving ships, stationary points or by long,bottom-mounted diffusers, but the detrimental effects on marineecosystems, especially in sites with high CO2 concentrations, can-not be ignored (Israelsson et al., 2010). Geological sequestration

pared from serpentine by various methods for subsequent CO2sequestration conducted at high temperature (>500 C) and highpressure (>20 bar) atmosphere (Romo et al., 2013). boAkademi University has been performing intensive research withinthis topic for about a decade and is developing a process routetoward industrial application, in which magnesium in a rst step2

1. Introduction

In recent decades, the greenhousecommon concern worldwide. Carbois considered as one of the main opof CO2 from human activities. Oc2

the Mg-rich solution by precipitation using sodium hydroxide solution. 2015 Elsevier Ltd. All rights reserved.

has become an issue ofture and storage (CCS)r alleviating emissionseological and mineral

(Sanna et al., 2013) would be released to the atmosphere again,making post-storage monitoring unnecessary.

Serpentinite, containing mainly serpentine, is an excellent feed-stock material for CO2 mineral carbonation because of the largequantities available, extensive distribution and relatively low hard-ness (Nduagu et al., 2012a). Magnesium hydroxide could be pre-Keywords:Magnesium hydroxide

and ultrasonication were demonstrated to be effective in controlling the thickness of product layer. AboutPreparation of magnesium hydroxide froleaching for CO2 mineral carbonation

Qing Zhao a,b,, Cheng-jun Liu a, Mao-fa Jiang a, HenraKey Laboratory for Ecological Metallurgy of Multimetallic Ores (Ministry of Education)b Thermal and Flow Engineering Laboratory, bo Akademi University, bo/Turku 20500,

a r t i c l e i n f o

Article history:Received 11 February 2015Revised 30 May 2015Accepted 1 June 2015

a b s t r a c t

Carbon capture and storagtion that has been under stFinnish serpentinite using sstudied. Some details of leathis study. It was found thaproduct layer formed on th

journal homepage: www.serpentinite by sulfuric acid

Saxn b, Ron Zevenhoven b

rtheastern University, Shenyang 110819, Chinaand

CS) by mineral carbonation is a promising way for CO2 emissions mitiga-ed for decades. In this work, the preparation of magnesium hydroxide fromuric acid leaching as the rst step of a CO2 mineral carbonation process wasng behavior of the ore were revealed and a valuable metal was recovered inaching yield of magnesium increased with sulfuric acid dosage, limited by are particles, resulting in incomplete serpentinite decomposition. Agitation

le at ScienceDirect

gineering

evier .com/ locate/mineng

effect of mechanical preprocess on leaching efciency, reportingthat almost all magnesium and iron was leached into the solutionfrom serpentine in only 5 min after 240 min ball-milling pretreat-

(2, 3, 4 and 5 mol L ), and then agitation was started and main-tained during the whole leaching process. After a certain periodof time (10, 20, 30, 40, 50, 60, 90 and 120 min), the leachate wasobtained by ltration and analyzed by ICP-OES. The leaching yieldof metallic elements, expressed as the mass ratio of metallic ele-ments in the leachate and in the raw material, was determined.The experimental set-up for serpentinite powder leaching is shownin Fig. 3(a).

2.2.2. Ore lump experimentsTo study the product layer formed on ore particles, a batch of

experiments was carried out using a serpentinite lump particle.Sulfuric acid (200 mL 4 mol L1) was added into a glass beakerand the lump was immersed into the solution with the smooth atsurface facing upward. Agitation and ultrasonication treatmentwere employed in some tests in an attempt to remove the productlayer. After 2 h of leaching, the lump was taken out of the acid solu-tion and was carefully washed with deionized water to avoiding

Q. Zhao et al. /Minerals Enginement. However, the high energy requirement from this long grind-ing duration works against the original purpose of CO2 emissionsreduction. Moreover, the high cost in recycling the acid is still achallenge for industrialization.

Kodama et al. (2008) conducted metallic elements extractionexperiments from silicates, reporting that a Si-rich phase formedon the surface of the ore particles during the leaching process,which signicantly limited the diffusion of reactive ions.However, a detailed understanding of the role of this product layerand ways to remove it is still lacking.

Researchers admitted that the economy of CO2 sequestration isa key factor for future technology deployment (Olajire, 2013).Besides reducing the high energy requirement by optimizing pro-cess conditions and parameters, it is possible to solve the problemof process economics by producing valuable by-products in thesequestration process. It has, for example, been demonstrated thatthe utilization of various iron compounds from serpentine as rawmaterials for the iron- and steel-making industry could be feasiblesolution to offset CO2 sequestration costs (Nduagu et al., 2012c;Romo et al., 2012). Solvent extraction by organic extractants is awell-known approach for accomplishing the complex separationsof metals from multi-element acid solutions. Many organic extrac-tants can be easily recovered and repeatedly used, controlling thefollow-up input of solvent extraction process. However, little workhas yet been reported on how to extract valuable metals fromserpentine leachate by solvent extraction.

This paper aims at a less energy-intensive process route to pre-pare magnesium hydroxide for CO2 sequestration from Finnish ser-pentinite. Sulfuric acid leaching and solvent extraction wereconducted for serpentinite decomposition and iron extraction.The leaching behavior of metallic elements, the effects of agitationand ultrasonication treatment on this passivating layer, and theoptimization of extraction conditions were investigated in thecurrent work.

2. Experimental

2.1. Materials

Serpentinite used in this study was taken from the stockpileof the Finnish Hitura nickel mine of Belvedere Resources Ltd.(formerly Outokumpu Mining Oy). Inductively coupledplasma-optical emission spectrometry (ICP-OES) analysis was car-ried out by Varian Vista-MPX against suitably calibrated standards(520 ppm) on aqueous extracts from 100 mg samples diluted to500 ml to detect the chemical composition of the serpentinite;the results are given in Table 1. The phase composition of serpen-tinite was analyzed by Philips Xpert X-ray diffraction (XRD) withCu Ka source (k = 1.5418 ) over the range 2h = 1570 at a stepsize of 0.008 and specied by Crystallographica Search-Match(CSM) software with the Powder Diffraction File (PDF) databasesfrom International Centre for Diffraction Data (ICDD), which canbe seen in Fig. 1. The ore was found to contain 83 wt.% serpentine(Mg3Si2O5(OH)4). In the remaining impurities, magnetite (Fe3O4)accounted for the largest fraction (82 wt.%). The Mg/Fe mass ratiois 2.2, BET surface area 26.45 m2 g1 m2/g and pore volume0.0347 cm3 g1.

Table 1Chemical composition of Finnish serpentinite (wt.%).Mg Fe Ca Ni Al Cr Cu Si

21.80 10.10 0.34 0.28 0.02 0.01 0.08 11.60Serpentinite powder with a size smaller than 74 lm wasobtained by grinding and screening, which was smaller than theprevious study (Teir et al., 2007b). This was used for studying theleaching behavior of metallic elements in the ore. A series oftests using a serpentinite lump was also carried out to elucidatethe product layer formation and removal behavior in theleaching process. Thus, the rock was cut and polished to get asmooth at surface. A LEO 1530 Gemini with a scientic ultradry silicon drift detector was employed in scanning electronmicroscopy-energy-dispersive X-ray spectroscopy (SEMEDS)analysis. A photograph and a SEM image of the lump surface areshown in Fig. 2. On the basis of results from EDS connected withSEM, it could be conrmed that plenty of bright phases, with sizesin the range 1030 lm, distributed in the dark silicate matrix weremagnetite. This implies that it would be difcult to effectively sep-arate magnetite by magnetic separation from the Mg-rich silicatesince it has not been liberated. As a result, iron recovery tests bothfor leachate and leaching residues were conducted by solventextraction and magnetic separation, respectively, after the sulfuricacid leaching process.

2.2. Methods

2.2.1. Ore powder experiments10 g of serpentinite was poured into an Erlenmeyer ask with a

certain amount (the ratios of ore powder mass and acid volumewere 0.1, 0.2, 0.4 and 1 g mL1) of sulfuric acid of certain molarity

1

Fig. 1. XRD pattern of Finnish Hitura serpentinite.

ering 79 (2015) 116124 117morphology changes, followed by a drying process at room tem-perature. The apparent morphology of the lump was investigatedusing Olympus 3D measuring laser microscope (3DMLM).

Fig. 2. Photograph (left) and SEM image (right) of Finnish serpentinite lump.

Fig. 3. Schematic illustration of the experimental set-up for leaching the Finnish serpentinite. (a) Leaching Finnish serpentinite powder; (b) Leaching Finnish serpentiniteeaclas

ginelump without treatment; (c) Leaching Finnish serpentinite lump with agitation; (d) LRetort stand; 3 Erlenmeyer ask; 4 Serpentinite powder; 5 Sulfuric acid; 6 G118 Q. Zhao et al. /Minerals EnFig. 3(b)(d) present a schematic illustration of the equipment andprocedure.

2.2.3. Solvent extraction2-Ethylhexyl dihydrogen phosphate (P507) is regarded as an

excellent extractant for Fe3+ in chemical industry. A certainamount of hydrogen peroxide was needed to oxidize all of theFe2+ ions in the leachate to Fe3+ ions before the solvent extraction.After this, some industrial P507 (>95 wt.%) was mixed with sul-fonated kerosene in different ratios to reduce the viscosity ofP507, and then some 10 wt.% sodium hydroxide solution wasadded into the diluted extractant to saponify the P507. The saponi-ed extractant and the serpentinite leachate were mixed in a sep-arating funnel at room temperature, shaken for a few minutes toextract Fe3+ into organic phase followed by a short time of stand-ing, ensuring that a complete separation of organic and aqueoushing Finnish serpentinite lump with ultrasonication. 1 Variable speed blender; 2 s ask; 7 Water; 8 Serpentinite lump; 9 Ultrasonic vibration equipment.ering 79 (2015) 116124phases was achieved. Reactions of P507 saponication and extrac-tion of Fe3+ are as follows:

Fe3+ and Mg2+ contents left in aqueous phase were determinedby ICP-OES to calculate the extraction yield according the massratio of metallic ions in aqueous phase before and after solventextraction.

2.2.4. Product preparationMagnesium hydroxide was precipitated from the aqueous

phase when the pH value reached 10 using sodium hydroxidesolution. 4 mol L1 Hydrochloric acid was employed to strip ironfrom organic phase into aqueous acid solution for valuableFe-containing products preparation. After evaporation of theFe-rich solution and lter cake washing, iron oxide was obtained.Almost all of the P507 was regenerated in the stripping process

gineering 79 (2015) 116124 119Q. Zhao et al. /Minerals Enfor reutilization, and some HCl and vapor were also recovered in aglass condenser pipe during the evaporation phase. Magnetite wascollected from the leaching residues by magnetic separation with amagnetic eld intensity of 50 mT using a DTCXG-ZN50 magneticseparator. A ow sheet of the overall process route proposed in thisstudy is given in Fig. 4.

3. Result and discussion

3.1. Leaching behavior of ore

Both concentration (2, 3, 4 and 5 mol L1) and volume (theratios of ore powder mass and acid volume were 0.1, 0.2, 0.4 and1 g mL1) were simultaneously considered in 2 h-tests to studythe effect of sulfuric acid on extraction yield of magnesium, givingthe results shown in Fig. 5. The acidity and volume of sulfuric acidproved to be signicant factors for the leaching yield of magne-sium. When 2 mol L1 sulfuric acid was used in the leaching pro-cess, only about 70% of the magnesium could be extracted fromthe serpentinite. Higher than 80% of leaching yield was achievedin the experiments using acid with concentration higher than3 mol L1. Furthermore, the acid volume could be reduced if highacidity was employed, but sulfuric acid concentration exceeding

Fig. 4. Flow sheet of t

Fig. 5. Effect of sulfuric acid on leaching yield of magnesium.4 mol L1 showed no further improvement on magnesium leach-ing. Therefore, a reasonable acid dosage was 4 mol L1 sulfuric acidwith a ore mass/acid volume ratio of 0.4 g mL1.

Leaching behavior of the Finnish serpentinite was investigatedusing the optimal sulfuric dosage. A batch of tests was conductedfor different leaching durations, and the concentrations of allmetallic elements in the leachates were determined by ICP-OES.The changes of leaching yield with duration are shown in Fig. 6,in which the error bars was the average value of the results oftwo parallel experiments. It was found that recovery rates of allmetallic elements steadily increased with duration until 30 min,while no notable change occurred from 30 min to 120 min. Theoptimal duration obtained in this study was shorter than itreported by Alexander et al. (2007) who worked on the serpentinefrom the Cedar Hills quarry in SE Pennsylvania using similar theleaching conditions, demonstrating the high reactivity of the

he process route.Finnish serpentinite. Most of the metallic elements except ironand nickel were leached out from the ore, with a leaching yieldfor magnesium of about 86%. A leachate with about 3 mol L1 ofMg2+ was obtained from this process, which could be used to pre-pare magnesium hydroxide for the CO2 capture after iron recovery.Therefore, a leaching time of 30 min is considered long enough formagnesium extraction in 4 mol L1 sulfuric acid.

Fig. 6. Relationship between the leaching yield of metallic elements and theleaching duration.

To clarify the reasons for the low leaching yield of iron and

deposited on the surface of particles and acted as an obstacleimpacting the contact of the reactants. More details were investi-gated and are disclosed in the next section.

3.2. Product layer investigation

A polished serpentinite lump was immersed in 4 mol L1 sulfu-ric acid for 2 h, and was then removed and dried at room temper-ature or 90 C for 2 h. After this the apparent morphology wasstudied using 3DMLM; some results are illustrated in Fig. 8.Before the leaching treatment, a smooth ore lump surface (cf. 1layer) can be seen in the gure. A glassy layer (cf. 2 layer) wasformed and covered the whole lump in room temperature tests.By EDS and ICP-OES the elemental composition of this glassy phasewas detected to be silicon and oxygen (It should be noted thatneither EDS nor ICP-OES can detect hydrogen.). Additionally, nodiffraction peaks of new Si-bearing crystal appeared in anypatterns (cf. Fig. 7), demonstrating the amorphous form of thisSi-rich phase. A hump detected below 20 2-Theta cannot bespecied by any known standard, which was speculated to be afunction of the XRD itself. With reference to an Eh-pH diagram of

Fig. 7. XRD patterns of Finnish serpentinite and residues after different leachingdurations.

120 Q. Zhao et al. /Minerals Engineering 79 (2015) 116124nickel, the phase compositions of residues from 10 min, 30 minand 50 min leaching tests were detected by XRD (Fig. 7) andcompared with the compositions of the original ore. The resultsdemonstrated the major phases of all samples to be Fe3O4 andMg3Si2O5(OH)4. The intensity of diffraction peaks of serpentineweakened signicantly with leaching duration but did not showany notable changes from 30 min to 50 min, coinciding with theresults obtained with ICP-OES. As for the diffraction peaks of mag-netite, they decreased slower than the serpentine peaks becausemagnetite barely dissolves in sulfuric acid solution of this acidity.Therefore, some magnetite was left in residues after the leachingprocess, leading a low leaching yield of iron. He (2010) proposedthat Ni-bearing serpentine is also a refractory phase in acid solu-tion, so nickel may concentrate in residues after the leaching treat-ment. Experimental results of this study showed that the contentof nickel in the leaching residues reached 0.62%, which is consider-ably higher than the original 0.28%. This means that these residuescould be utilized as a secondary nickel resource in nickel extractionindustries. Kodama et al. (2008) suggested that the reason for theincomplete extraction of magnesium is that a Si-rich phaseFig. 8. Apparent morphology of serpentinite lump beforeMgSiOH at 25 C and the glassy state (cf. 2 layer), this phasegenerated in the serpentinite leaching should be the amorphoussilicic acid rather than silica. After a drying at 90 C, the productlayer was transformed into silica particles (cf. 4 layer) becauseof the dehydration, explaining the confusion of Park et al. (2003)and Kodama et al. (2008) on the phase determination of theproduct layer. Moreover, when the leaching process is conductedat high temperature like 160 C, silicate could transform intoamorphous silica directly (Zhao et al., 2014).

Furthermore, it was found that this product layer can beeasily removed, exposing a relatively smooth inner surface ofthe lump (cf. 3 layer). Therefore, in the decomposition of ser-pentinite during the leaching process it could be speculated thathydrogen ions diffused through this layer from the solution tothe surface of the ore particles to react with the inner core.The shrinking process of the particles progressed in a uniformway, so a smooth surface of the inner unreacted core wasobtained. There is no doubt that the removal treatment for thispassivating layer is critical for the leaching rate and for thecompletion of the reactions.and after 2 h immersion in 4 mol L1 sulfuric acid.

nite lump leached with different treatments.

gineSome lump experiments with an agitation at 400 r min1 orwith an ultrasonication at the frequency of 40 kHz were conductedto remove the product layer and compared with the immersiontest. The apparent morphology of all lumps after 2 h leaching in4 mol L1 sulfuric acid was detected by 3DMLM, as shown inFig. 9. The serpentinite surface of the lump could be generallyobserved through/below a glassy phase after the agitation test,showing that the product layer was much thinner than the onewithout any treatment. It can be inferred that the ow of solutioncaused by stirring could effectively control the thickness of theproduct layer in the leaching process, and the diffusion rate ofhydrogen ions through this layer, as a consequence, would notsignicantly decrease with duration. Therefore, the decompositionof ore particles in sulfuric acid may be brought closer to comple-tion. As for the ultrasonication test, a product-layer-free lump with

Fig. 9. Apparent morphology of serpenti

Q. Zhao et al. /Minerals Ena clear surface was obtained. The explanation of this could beattributed to the cavitation that occurs when the solution is sub-jected to rapid changes of pressure in ultrasonication treatment.Cavities are formed where the pressure is relatively low, and thevoids implode when they are subjected to higher pressure, whichcan generate an intense shockwave. The product layer cannot stayattached to the surface of lump under the effect of this cavitation.Furthermore, solution temperature was also elevated by above50 C by the ultrasonication treatment, which favored the leachingprocess as well. However, it is important to note that this treat-ment may have a signicant energy input requirement, whichshould be considered in evaluating the appropriateness of theprocess for overall emission reduction.

A set of batch ore powder tests with different treatments wascarried under the same conditions as the lump tests and someresults are presented in Fig. 10. The leaching yield of magnesiumis seen to improve from about 48% to 86% with the employmentof agitation, and exceeded 90% in an ultrasonication test. Ironleaching was also elevated by these treatments. In addition tothe effect of a removal of the product layer discussed above, oreparticles could now uniformly distribute in the solution by stirringand cavitation, ensuring a full exposure of the particles to the acidsolution. As a result of this, the decomposition of serpentinite wasmore complete. Therefore, keeping in mind the energy inputrequirements, the agitation treatment was concluded to be themost viable approach to remove the product layer in powder testsof this study.ering 79 (2015) 116124 1213.3. Iron recovery

An industrial hydrogen peroxide with a 1.2 times volume thanthe theoretical one was added in the leachate to oxidize Fe2+ toFe3+ ions for a subsequent solvent extraction using the organicextractant P507. P507 is a widely used extractant in acid systemthat exchanges with H+ in POH and becomes POFe in theway of cation exchange mechanism (Wu et al., 2013).

The separation factor of Fe3+ and Mg2+ (bFe3=Mg2 ) observed in

the current work is a signicant index for evaluation of theseparation result of Fe3+ and Mg2+, which dened by

bFe3=Mg2 CFe3O=CFe3ACMg2O=CMg2A

where CFe3O is the concentration of Fe3+ in the organic phase,

CFe3A is the concentration of Fe3+ in the aqueous phase, CMg2O

is the concentration of Mg2+ in the organic phase, and CMg2A isthe concentration of Mg2+ in the aqueous phase.

Some extraction parameters including pH value (1.00, 1.25,1.50, 1.75 and 2.00), concentration of P507 (20, 30, 40 and

Fig. 10. Leaching yield of magnesium in sulfuric acid with different leachingmethods.

diti

on r

gineTable 2Separation factor and extraction yield of Fe3+ and Mg2+ under different extraction con

No. pH value Concentration of P507 (vol.%) Saponicati

1 1.00 40 402 1.25 40 403 1.50 40 404 1.75 40 405 2.00 40 406 1.50 20 407 1.50 30 408 1.50 50 409 1.50 40 0

10 1.50 40 20

122 Q. Zhao et al. /Minerals En50 vol.%), saponication rate (0%, 20%, 40% and 60%) and duration(0, 2, 4, 6 and 8 min) were investigated with a constant phase ratioof 1:1, dened as the ratio between organic volume and aqueousvolume. The separation factor and the extraction yield of bothFe3+ and Mg2+ ions under different extraction conditions are reportin Table 2 and Fig. 11.

The relative separation of the curves gives the possibility for theselective extraction of Fe3+ from Mg-rich solution. The experimen-tal data reveals that the extraction for Fe3+ was higher than that forMg2+ in all tests. From the results with respect to pH, it can be con-cluded that the extraction yield of Fe3+ rose when pH increasedfrom 1.00 to 1.50, while it decreased slightly when the pH valueincreased further. The extraction yield of Mg2+ did not exhibit anotable change with pH in the range of 1.00 to 2.00. Therefore,

11 1.50 40 6012 1.50 40 4013 1.50 40 4014 1.50 40 40

Fig. 11. Effect of extraction condons (wt.%).

ate (%) Duration (min) Fe3+% Mg2+% bFe3=Mg2

6 47.3 4.0 226 76.7 3.2 1086 98.1 6.3 7686 95.2 5.2 3616 94.7 7.1 2526 55.8 1.2 1256 83.1 2.9 1586 99.1 13.7 6086 77.0 3.1 1126 87.9 4.4 183

ering 79 (2015) 116124the highest separated factor of Fe3+ and Mg2+ appeared atpH = 1.50 where about 98% of the Fe3+ was extracted from theaqueous into the organic phase with a minor Mg2+ extraction ofabout 6%. A batch of tests was implemented with different P507dosages in a constant organic volume to investigate the effect ofextractant concentration on the extraction results. It was foundthat the extractions for Fe3+ and Mg2+ were both improved withan increase in concentration of P507 in the range 2060 vol.% while40 vol.% was the optimal value for Fe3+ and Mg2+ separation in thecurrent study. Higher extractant concentration than the optimalled to a better Mg2+ extraction but showed no signicant increasein Fe3+ recovery. Furthermore, the separation efciency stronglydepends on the viscosity of extractant, and an emulsicationmay occur when the viscosity rises too high. Therefore, 40 vol.%

6 97.8 9.0 4902 92.2 5.8 1754 98.2 6.1 7688 98.9 6.4 768

itions on extraction results.

Goff, F., Guthrie, G., Lipin, B., Fite, M., Chipera, S., Counce, D., Kluk, E., Ziock, H., 2000.

gineis here considered an appropriate extractant concentration in thiswork. Fig. 11 also presents the relationship between saponicationrate of P507 and extraction results, showing the maximum valuesof extraction yield of Fe3+ and separation factor of the two metallicelements in the tests at the point where the saponication rate ofP507 was 40%. The Fe3+ recovery did not experience any signicantchanges when more P507 was saponied. Moreover, the pH valueof the aqueous phase rose from its initial value of 1.50 to 3.39when 40 vol.% P507 was employed. This demonstrates that morehydroxide ions were released from hydrolyzate of Fe3+ than hydro-gen ions provided by unsaponied P507, so some hydroxide ionsleft in the aqueous phase led a decrease in the pH value. The dataobtained from duration tests can be also seen in Fig. 11. The resultsshowed that the extraction process of P507 for Fe3+ (and Mg2+) isvery fast: more than 98% of the Fe3+ could be extracted into organicphase by saponied P507 within 4 min. The indices for the solventextraction treatment changed little when extraction durationexceeded 4 min.

Hydrochloric acid was used as the stripping agent as its strip-ping capacity is better than that of sulfuric acid and nitric acidaccording to preliminary experiments. 4 mol/L hydrochloric acidwas used with a phase ratio of 1:1 in a stripping process to recoverFe3+ from the organic phase for subsequent preparation ofFe-containing inorganic products. The experimental procedure forsolvent extraction was reported in Section 2.2.3. This treatmentwas conducted repeatedly for 8 min, each time followed by theICP-OES analysis for Fe3+. The ndings show that about 62% ofFe3+ was stripped into aqueous phase after one stripping period,and 99% of Fe3+ was recovered from organic phase after fourstripping periods. Hardly any Mg2+ was stripped out by thehydrochloric acid in the tests, so a Fe-rich solution was obtained.

When the iron recovery process is applied in industrial scale, itis important to evaluate the possibility of material cyclic utiliza-tion. In this study, P507 was regenerated by the stripping treat-ment because H+ ions replaced the Fe3+ ions in Fe-bearing P507,and the P507 thus recycled in this stage can serve as an extractantagain. HCl was also recovered in this study by evaporating the fer-ric chloride acid solution for hydrochloric acid solution regenera-tion, and then an iron oxide (>99 wt.%) was obtained in residuesas a by-product. The recycle of extractant and hydrochloric acidgives this iron extraction method a high competitiveness in theperspective of economics. But from an environmental point ofview, manufacturing ferric chloride by concentrating the Fe-richsolution is a promising option for utilization of the iron extractedfrom Finnish serpentinite without any HCl gas generation.

To recover the magnetite in the leaching residues, some mag-netic separation tests with a magnetic eld intensity of 50 mT wereconducted and the same magnetic separation was also carried outto ore powder for comparison. The magnetic separation yield wasfound to be about 46% in the ore tests, while about 93% of iron wasrecovered from leaching residues. Moreover, the content of Fe3O4in magnetite concentrate was increased from 73.2% to 92.1% afterthe leaching process. The reason for these improvements is thatplenty of magnetite aked from serpentine during the leachingprocess by acid corrosion and solution stirring, so a higher recoveryrate and purity can be obtained due to a very high phaseseparation.

3.4. Products

After the iron extraction, magnesium hydroxide in micron scalewas precipitated from the Mg-rich aqueous phase by a pH adjust-ment until the pH value reached about 10 (using sodium hydroxide

Q. Zhao et al. /Minerals Ensolution), followed by a drying for 2 h at 110 C. Magnesiumhydroxide and iron oxide prepared in this study were weightedand detected by XRD and ICP-OES analysis, which, as expect, showsEvaluation of Ultramac Deposits in the Eastern United States and Puerto Ricoas Sources of Magnesium for Carbon Dioxide Sequestration. Los AlamosNational Laboratory and U.S. Geological Survey, New Mexico.

He, Z.X., 2010. Study on the comprehensive utilization of Ni-bearing serpentine.MSc. Thesis. Central South University, Changsha (in Chinese).

Israelsson, P.H., Chow, A.C., Adams, E.E., 2010. An updated assessment of the acuteimpacts of ocean carbon sequestration by direct injection. Int. J. Greenh. Gas.Con. 4, 262271.

Kim, D.J., Chung, H.S., 2002. Effect of grinding on the structure and chemicalextraction of metals from serpentine. Part. Sci. Technol. 20, 159168.

Kodama, S., Nishimoto, T., Yamamoto, N., Yogo, K., Yamada, K., 2008. Developmentof a new pH-swing CO2 mineral carbonation process with a recyclable reactionsolution. Energy 33, 776784.the major phases to be Mg(OH)2 and Fe2O3 with few impurities.The results also indicated that the recovery rate of magnesium ofthe whole process, expressed as mass ratio between magnesiumin the magnesium hydroxide and in the original serpentinite ore,was 83.7%. Moreover, solvent extraction and magnetic separationrecovered 95.3% of iron, so iron oxide and magnetite were concen-trated as by-products. The iron oxide purity was higher than 99%,and the content of Fe3O4 in the magnetite concentrate was about92%, as reported in Section 3.3 of this study.

4. Conclusions

The preparation of magnesium hydroxide from Finnish serpen-tinite by sulfuric acid leaching for CO2 mineral carbonation wasinvestigated at ambient temperature and pressure. About 86% ofmagnesium was leached out in 30 min by 4 mol L1 sulfuric acidwith the ore mass/acid volume of 0.4 g mL1, while somemagnetite and Ni-bearing serpentine was left in the leaching resi-dues. Agitation and ultrasonication treatments were demonstratedas effective methods for passivation or removal of product layer,leading to a more complete decomposition of the ore. About 98%of the iron was extracted in 4 min from the leachate by 40 vol.%P507 with a saponication rate of 40%. Almost all of the iron wasstripped from organic phase by 4 mol L1 hydrochloric acid.About 93% of the iron from leaching residues by magnetic separa-tion, and a magnetite concentrate was obtained as a by-product.Mg(OH)2-rich powder was nally prepared for CO2 mineralcarbonation. Recovery rates of magnesium and iron of the wholeprocess were 83.7% and 95.3%, respectively.

Based on the conclusions of this work, future research will focuson a comprehensive utilization of other valuable metallic elements,especially the nickel in the leaching residues, in a cleaner andeconomical way. Another line of future work would be to reduceacid consumption with the nal goal to implement thislow-energy process in industrial CO2 mineral carbonation.

Acknowledgements

The authors gratefully acknowledge supports by ChinaScholarship Council (CSC) for the visit of Qing Zhao to boAkademi University, Finland. Program for New Century ExcellentTalents in University of Ministry of Education of China (No.NCET-11-0077) and the Fundamental Research Funds for theCentral Universities (No. 130402020) are also acknowledged.Funding from TEKES and Finnish metals industry within the SIMPresearch program under the Finnish Metals and EngineeringCompetence Cluster (FIMECC Oy) is gratefully acknowledged.

References

Alexander, G., Maroto-Valer, M.M., Gafarova-Aksoy, P., 2007. Evaluation of reactionvariables in the dissolution of serpentine for mineral carbonation. Fuel 86, 273281.

ering 79 (2015) 116124 123Maroto-Valer, M.M., Fauth, D.J., Kuchta, M.E., Zhang, Y., Andresen, J.M., 2005.Activation of magnesium rich minerals as carbonation feedstock materials forCO2 sequestration. Fuel Process. Technol. 86, 16271645.

Nduagu, E., Bergerson, J., Zevenhoven, R., 2012a. Life cycle assessment of CO2sequestration in magnesium silicate rock-A comparative study. Energy Convers.Manage. 55, 116126.

Nduagu, E., Bjrklf, T., Fagerlund, J., Mkil, E., Salonen, J., Geerlings, H.,Zevenhoven, R., 2012b. Production of magnesium hydroxide from magnesiumsilicate for the purpose of CO2 mineral carbonation-Part 2: Mg extractionmodeling and application to different Mg silicate rocks. Miner. Eng. 30, 8794.

Nduagu, E., Fagerlund, J., Zevenhoven, R., 2012c. Contribution of iron to theenergetics of CO2 sequestration in Mg-silicates-based rock. Energy Convers.Manage. 55, 178186.

Olajire, A.A., 2013. A review of mineral carbonation technology in sequestration ofCO2. J. Petrol. Sci. Eng. 109, 364392.

Park, A.A., Jadhav, R., Fan, L.S., 2003. CO2 mineral sequestration: chemicallyenhanced aqueous carbonation of serpentine. Can. J. Chem. Eng. 81, 885890.

Romo, I.S., Gando-Ferreira, L.M., Zevenhoven, R., 2013. Combined extraction ofmetals and production of Mg(OH)2 for CO2 sequestration from nickel mine oreand overburden. Miner. Eng. 53, 167170.

Romo, I.S., Nduagu, E., Fagerlund, J., Gando-Ferreira, L.M., Zevenhoven, R., 2012.CO2 xation using magnesium silicate minerals. Part 2: Process energy

efciency and integration with iron-and steelmaking. Energy Int. J. 41 (1),203211.

Sanna, A., Wang, X., Lacinska, A., Styles, M., Paulson, T., Maroto-Valer, M.M., 2013.Enhancing Mg extraction from lizardite-rich serpentine for CO2 mineralsequestration. Miner. Eng. 49, 135144.

Teir, S., Kuusik, R., Fogelholm, C.J., Zevenhoven, R., 2007a. Production of magnesiumcarbonates from serpentinite for long-term storage of CO2. Int. J. Miner. Process.85 (13), 115.

Teir, S., Revitzer, H., Eloneva, S., Fogelholm, C.J., Zevenhoven, R., 2007b. Dissolutionof natural serpentinite in mineral and organic acids. Int. J. Miner. Process. 83 (12), 3646.

Wu, W.Y., Zang, F.Y., Bian, X., Xue, S.F., Yin, S.H., Zheng, Q., 2013. Effect of loaded organicphase containing mixtures of silicon and aluminum, single iron on extraction oflanthanum in saponication P507-HCl system. J. Rare Earth 31 (7), 722726.

Zhang, D.X., Song, J., 2014. Mechanisms for geological carbon sequestration. Proc.IUTAM 10, 319327.

Zhao, Q., Liu, C.J., Shi, P.Y., Zhang, B., Jiang, M.F., Zhang, Q.S., Saxn, H., Zevenhoven,R., 2014. Sulfuric acid leaching of South African chromite. Part 1: Study onleaching behavior. Int. J. Miner. Process. 130, 95101.

124 Q. Zhao et al. /Minerals Engineering 79 (2015) 116124

Preparation of magnesium hydroxide from serpentinite by sulfuric acid leaching for CO2 mineral carbonation1 Introduction2 Experimental2.1 Materials2.2 Methods2.2.1 Ore powder experiments2.2.2 Ore lump experiments2.2.3 Solvent extraction2.2.4 Product preparation

3 Result and discussion3.1 Leaching behavior of ore3.2 Product layer investigation3.3 Iron recovery3.4 Products

4 ConclusionsAcknowledgementsReferences