‘Available online at wwwsciencedirect.com eT eae scence Gomer: liad aes Xo) Cement and Coneete Research 38 (2008) 2261-2270 Immobilization of heavy metals by calcium sulfoaluminate cement S. Peysson, J. Péra*, M. Chabannet Unité de Recherche en Génie Ciit-MATERIAUX,lasttt Nationa des Sciences Appliquées de Lyon, Domaine Scientifique de ta Dow, iiment J Test, 12. Avene des dis, 69 621 Villeurbanne Cade, France Received 28 April 2004; accepted 10 March 2008, Abstract ‘Two types of calcium sulfoaluminate cement containing 20% and 30% phosphogypsum, respectively, were invest hazardous waste stabilization. Fourteen series of pastes were prepared, caci containing the following soluble salt: NasCrOy-4H0;, Na:Cr0,2H.0; CrCly6HsO; PH(NOs)s; Zn(NOs)*6H:0; 2n807H.0; and CdCly-SH20. The level of pollution was 0.069 mol of heavy metal por Kg of cement he study has been carried out by means of X-ray diftaction, thermal analysis, seanning electron microscopy, mercury’ intrusion, orosimetry, electrical conductivity, and leaching tests. Very good retention of lead, cadmium, zine and trivalent chromium is observed. The tention of hexavalent chromium depends upon the nature ofthe binder: the cement containing 20% gypsum develops the best behavious This is explained by the microstructure of the hydrated paste: in the paste containing 30% gypsum, delayed ettringite precipitates and damages the hardened past, © 2005 Elsevier Lid. All rights reserved Keyword: Calcium salfoaluminae erent; Exingite; Louching; Mcrostustr; Sahilizaton 1, Introduetion - A=ALOs, Sulfobelite cements, also called sulfoaluminate-belite cements, contain the phases belite (CS), yeelimite or tetracalcium tialuminate sulfate (CyA3), and gypsum Sulfoaluminate-blite cements have been used to study (C3H.) as their main constituents. When CSA cement the formation of oxyanion-substituted etringites for the hydrates, etingite (CASH) is Formed according to the purpose of fixation of heavy metals [2-5], From a following reactions [1); chemical point of view, trivalent ions as Fe"’, Cr’, and Mn*> [3] can replace AI ions in the crystal structure of CiAs3 + 2CEH; + 34H=°CyAZsHs2 + 2AHS, ettringite. Ca” ions ean be substituted by bivalent ions as, in absence of ealeium hydroxide, Table 1 CiA3¥ + 8CEH: + OCH + 74H-P3C(ARSHoa, Concenration ves of plants Potlane MolKi ‘Quantiy oF hea met! in presence of calcium hydroxide, where: of binder added to the binder (ee) Na.G0,at1,0 0.009 36g of Gre C= C20, 0.069 368 of Cok S=Si0>, 0.069 36g of Coke PoINO. 0.069 143 ¢ of Poke Ze{N0, 640 0.069 45 gof Zukg + Goreponding autor, Te #3347243 8296; fxs 133478949807, 28806 7HLO 0.069 45 go Zn Ke ac 009 173 gof Coke Email address: Jean Peratainsslyonsr (Pea, (0008-884615 - ue front matter © 2005 Elsevier Lid. All rights reserved, o:10.1016j.cmeonres 205, 03.015neo 5 Peysson eta, / Cement and Concrete Research 36 2005) 2261-2270 Sample containing sodium dichromate ‘Sample containing sodium chromate Fig. 1. Damaged samples, CSA2. Me, 212, Mn, Fe’, Co or NP. Moreover S02~ can bye replaced by COF, NOs, SeO#-, Cr0F-, oF B(OH), [6-10], The purpose ofthis tidy was to evaluate the ability of calcium sulfoaluminste cement (CSA) 10 immobilize Giffrent heavy mal ions? cadmium, cbromiam, ead, and zine 2, Experimental Two calcium sulfoaluminate cements (CSA) were used: 'SA1: 80% calcium sulfoaluminate clinker +20% phosphogypsum, CSA2: 70% calcium sulfoaluminate clinker +30% phosphogypsum. The composition of calcium sulfoaluminate clinker, assessed by X-ray diffraction and chemical analysis, was as follows: yeelimite (CaAs2): 53.5%, belite (C28): 21.2%, ferrite phase (C,AF): 16.3%, where F=Fe,0s, perovskite (CT): 9%, where T=TiO,, Phosphogypsum contained 99% pure gypsum. Four heavy metals were investigated under different water-soluble forms: cadmium, chromium, lead, and zine, Cadmium was studied under the chloride form: CaCl;.5H20. Chromium can be found under 3 states, whose valenees are +2, +3, and +6. CP” and Cr are the ‘more frequent states, Cr ® is an acidic form, which forms chromates (CrO,)°~ and dichromates (Cr,07)*~. The tradi- tional technique of stabilization of Cr ™ is initially to reduce it to Cr’ in acid medium using a reducing agent, The most used reducing agents are ferrous sulphate (FeSOs, 7H;0), sodium metabisulfite (Na;S,0.), and sodium hydrosulfite (NayS,0,) [11]. In the present research, CSA was directly used to stabilize cither sodium chromate (Na,C10,-4H20) ‘or sodium dichromate (Na,Cr,0;2H,0), The fixation of Cr’ was also investigated, using chromium chloride (CrCly6H,0), 70% CSA clinker + 30% phosphogypsum 50 « leaching ater 72 hours (meaching ter 48 hours 3 [Seaching ater 24 hours 2» 2 20 10 o = ervyin Gxt in er) chromate chromate ca Po inn anim ZHINO,), 280, 2. 28-day heavy metal leach of CSA2 pastes.S. Pepeon et a. / Cement and Concrete Research 88 (2008) 2261-2270 2263 {0% CSA clinker + 20% phosphogypsum git X 1 Zoi er(vy in cxvyincxlu) ca Pb Znin Zain chromate chromate ZNO} 2n80, Fig, 3 2806um GSAT NeGO.aHO 240742 oa Oa NaCn0y2H0 33° 737.2026. crcysio 206 «92129 PoNO), 204748 TTS ZoNOW6IL0 49 1S 21S 10 23804180 40742 D4 6 cacissm0 2330-3702 CSA2 NesCOe4H0 271 618 29 NaCn0y 210 283 645 95 crchsio 8770 48 PLNO) 29749 16 ZagNOs)s61,0 282 718 65 ZaS0y7h0 12707 69 cachsiz0 245 680 12 cease of CSA1, the quantities of Cr present in the leachate ‘were lower than those observed at 28 days of age, The quantities of Cr collected from the leaching of CSA2 ‘were higher than those observed at 28 days of age: 2.6 times higher for dichromate and 2.3 times higher for chromate, respectively. The behaviour of the two binders is shown in Fig. 6: CSA1 exhibits better fixation of Cr than, SA2, Degree of metal stabilization for each paste is given in Fig. 7, which indicates that well over 99% of the metals ‘cadmium, lead, and trivalent chromium are stabilized in the pastes, Stabilization of Cr°" depends upon the binder used: — itis greater than 98.2% for CSA1, — itis in the range of 65 to 93% for CSA2, 43.2. Microstructural investigation 3.2.1, Results of XRD The phases detected by X-ray diffraction were unreacted yeelimite and gypsum, and ettringite. The Fig, 10, CSA2 doped with sodium chromate intensity of the main peak of ettringite was higher in CSA2 paste, and especially when chromium was intro- duced under either sodium dichromate or chromate forms Fig. 8), 3.2.2, Results of thermal analysis ‘The quantity of ettringite obtained after 28 and 90 days of hydration was assessed by means of thermal analysis, The mass loss obtained from TGA. curves in the range of 100 to 150 °C allowed the calculation of the theoretical quantity of ettringite that was formed in cach paste. The results are presented in Tables 2 and 3 In CSA1, the quantity of eltringite was in the range of 33% to 36% at 28 days and 34% to 38% at 90 days, The differences observed between the different pollutants were small In CSA2, the quantity of ettringite was in the range of 35% to 50% at 28 days and 38% to 52% at 90 days. In this, binder, there was a drastic increase in the ettringite content ‘when sodium dichromate and, especially, sodium chromate ‘were introduced in the paste. Studies conducted by Ogawa and Roy [14] have shown that the presence of sodium accelerated the reactivity of yeelimite and the dissolution of calcium sulfate, thus leading to higher production of cttringite. The higher the presence of gypsum, the higher Fig. 9. CSA2 doped with sodium dichromate Fig 1. CSA2 doped with chomium chloride2268 the precipitation of ettringite. This explains the differences between CSA1 and CSA2. These results were confirmed by. the conductivity study. 3.2.3, Results of mercury intrusion porosimetry ‘The results of mercury intrusion porosimetry are presented in Table 4. CSA2 pastes presented higher porosity than CSAIL pastes. They also showed higher amounts of macropores, especially when they were doped with chromate or dichromate ions. This explains why the quantity of leached Cr" was higher in CSA2 pastes. 3.2.4, Results of SEM investigations SEM investigations showed different morphologies of ettringite. CSA2 paste doped with sodium dichromate presented numerous short hexagonal crystals of ettringite (Fig. 9). In CSA2 paste doped with sodium chromate, crystals of ettringite were larger (Fig. 10), It seems that the growth of needle-like ettringite microcrystals is, hindered in the CSA2 paste doped with sodium dichro- mate or dichromate. Larger erystals appear leading to the 5 Peysson eta, / Cement and Concrete Research 38 (2005) 2261-2270 destruction of the sample. This corresponds to the precipitation of swelling delayed ettringite, as reported by Wang et al. [15] and Kasselouri et al, [16]. When ciltringite precipitates after the matrix has reached some rigidity, it cannot develop according to long fine prismatic crystals. Vacant spaces within the paste are not suffi- ciently available to allow complete development of crystals. Thus, shorter and larger crystals are formed, exerting pressure on the existing microstructure and leading to cracking as shown in Fig. 1 Hollow and more elongated ettringite crystals were present in the CSA2 paste doped with chromium chloride Gig. 1D. 3.3, Conductimetry and solution analysis The conductimetry curves are presented in Fig. 12. The level of conductivity is lower than in Portland coment and. the curves do not present the same shape according to the composition of the CSA cement. In the control cement composed of 70% CSA clinker and 30% phosphogypsum (CSA2), once the maximum con- 70% CSA clinker + 30% phosphogypsum Conductivity (mSiem) {Chromium chloride aa TATE ‘Time (minutes) 180% CSA clinker + 20 % phosphogypsum —£ ‘Sodium chromate Bs. : § ‘Controt : Timo (minutos) Fig. 12, Conducietie curves obtained for suspensions (87/54).5, Peysam et a, / Cement and Concrete Research 35 (2005) 2261-2270 2269 ductivity reached, the level of conductivity remains constant (©8A 80720 Control for 150 min and then decreases, due the precipitation of 2500. citringite, When this cement is polluted, the shape is J zoo. Fes relatively the same but the period of constant conductivity B [ee disappeats and the conductivity decreases after the intial 1500 eak of conductivity. 3 ™ “ £ 1000. 168A 7030 Control 3 —_————— 2 mo Time (minutes) ‘1500 et CSA 80/20 Sodium chromate i — 0 — 5 [Feu & 2000 aa ‘| 5 om g — | Ere ° E1500: Te re) 5 —. {8A 700 Sodium chromate 3 250. err sat) > 7 ae oo ao 000 Y000 2 2000} Time (minutes) Soe. 3] SA 80720 Sodium dichromate ro 2500 F so. ze z i S 2000. 30, 8 s00{ ee | = Ce Pac fret oS E to. Time(s) 5 8A 7000 Sodium dehromate 5 2500 a a ? = Time (minutes) 1500- ‘alt CSA 80/20 Chromium chloride E [rer 20 Eso. z = i 3 a0. i bv E ES Be 5 10 (LE ° 3 Boa ooo a0 Eom. \ Tie (ines) 5 aso 8A 780 Chromium eheride 8 ae ° =e Te 7 aie aloo aot 2000 sec Time (minutes) ‘100 Fig. 1, Ana}yss of oon. CSA 1000 oS —— In the control cement composed of 80% CSA clinker sae and 20% phosphogypsum (CSA1), once the maximum ol conductivity reached, the level of conductivity remains wo 400 ~ao0~—a00=«O Time (minutos) Fig. 13. Analysis of solutions. CSA2. constant for 200 min and suddenly decreases to reach the minimum value after 360 min, due the precipitation of cttringite. Then, the conductivity increases and remainsa0 5. Peysson eta, / Cement and Concrete Revearch 38 (2005) 2261-2270 constant. The addition of chromium chloride and sodium dichromate does not modify the shape of the curve, The audition of sodium chromate retards the precipitation of ettringite and the minimum value of conductivity is reached ater $40 min. The analyses of solutions are shown in Figs. 13 and 14, In the solution of the control cement composed of 70% CSA clinker and 70% phosphogypsum (Fig. 13), the quantity of aluminum and calcium ions deereases slowly versus time, showing progressive precipitation of ettringite in the paste, ‘while that of sulphate ions increases due to the dissolution of gypsum, In presence of sodium chromate, the concentration in sulphate ions is intially higher but decreases after 325 min and chromate, aluminium, and calcium ions concentrations decrease, characterizing the precipitation of ettringite integrating chromate ions. The same phenomenon is observed with sodium dichromate. In presence of chromium chloride, the precipitation of ettringite is well marked after 400 min, The chloride ion concentration remains constant: this ion is not integrated in the structure of ettringite. In this cement, the presence of pollutants does not significantly modify the precip- itation of ttringite, as proved by the shape of the conductimetrie curve (Fig. 12). In the solution of the control cement composed of 80% CSA clinker and 20% phosphogypsum (Fig. 14), impor- tant precipitation of ttringite occurs after 360 min, comesponding to the minimum value of conductivity (Fig. 12). The precipitation of ettringite is delayed in presence of sodium chromate and dichromate: the depletion in sulfate ions only appears after 420. min ‘The presence of chromium chloride has an accelerating effect on the precipitation of etringit. In the polluted solutions, chromium and chloride ions are consumed versus time and thus integrated in the structure of cttringite 4, Conclusions ‘The stabilizing matrix based on caleium sulfoaluminate ‘cement shows a behaviour in leaching tests that can be very different depending on the nature of both the CSA binder ‘used and the pollutant. Regardless of the quantity of gypsum present in the cement, the stabilization of zine nitrate and sulfate, lead nitrate, cadmium chloride, and chromium chloride is excellent. In presence of sodium dichromate or chromate, the lower the quantity of gypsum the better the retention of chromium, (Cx). This is explained on the basis of different micro- structures developed in the cements. Delayed ettringite precipitates in the cement containing 30% phosphogypsum (CSA2), leading to higher porosity and cracking, References (1) 1 Oster, Cements containing calcium sulfoaluminae, Special Ino . sinie Cements, ERFN Spon, London, 2000, pp. [2] V-Albino, 8. Cioffi, M. Manocoli I. Santoro, Ptontialeplication of cerngte generating systems for hazardous waste stabilization, Joural of lazardous Materials S1 (1996) 241-282 [5] R, Berardi, R. Ciofo, L. Saroro, Matix stability and leaching ‘behaviour in etuingite-based stabilization systems doped with heavy metals, Waste Manggemen 19 (8) (1997) 538540, [4] B Classen, B LeRollnd,P.Colombet, Prod inertage de déchets par enrwhage dans un lant hydrauigue e prémix pour la mise eb swuvee de ee proctaé, French patent No FR 2796934, 2001) (i French) [5] R. Cioffi, M. Lavorgra, M. Marroccol L. Santor, Stabilization ofa ralvani sludge by mean of elem slphosluminate cement, Waste Materials in Constuction, Elsevier, Amsterdam, 1997, pp. 823-830. [6] #2. Pomann 1), Kazel, R. Wenda, Solid solitons of eeringts: Par 1 incorporation of OL” and COF- in 3Ca0 ALO, 3Cx80, 32H,0, Cement and Concrete Research 20 (6 (1990) 303322 [7] H Pllmann, 1.3. Kuzel, R. Wonda, Solid solutions of etringts: Pst 1. Incorporation of B(OH}, and CrO}- in 3Ca0 ALO; 3CaS0y 32110, Coment and. Conerete Research 23 (2) (1993) 422-430. [8] P Kumarathaan, GJ, MeCanhy, Dd. Masset, DF, Prughoet Haset, Oxyanion substiuted etrngtes: synthesis and characte ization, and their potential ole in immobilization of As, B, Cr, Se and V,_ Fly Ash and Cosl Conversion By-Products: Characterization and Disposal, vol. 17S, Materials Research Society, Us Pisburgh, 1990, p. 83— 104 [91 J. Bensted, SP. Varma, Esrngte and ts derivatives subtitaton, Seats Industries 23 (195) 315~318, [10] MLD. Goupae, BE. Scheetz, DM. Roy, Blringite and CS. Poland cement phases for wast ion immobilization: eeview, Waste Management 16 1996) 295303, [U1] G. Laforest J. Duchesne, Capacity of chromium ions fixation by blast fume sg, a: RK. Dhit, M.D. Newlands LJ. Ceteni (Ed), Role ‘of Cement Science in Sustunsble Development, Thomas Teor, London, 2003, pp. 335-344, (02) J. Péra, G. Théveain, M. Chabannet, Design of a novel system allowing the selection ofan adequate binder fo solifiction, Cement and Coneete Research 27 (10) (1997) 1833-1542. [03] ©. Vemet, G. Noworya, Conducmetic test for cement admixture systems, "Proceedings of the Sth Intemational Congres on the Chemisty of Cement, New Deli, India, val. 2, 1992, pp. 627-638, [14] K Ogawa, DIM, Roy, C.ASS hydaaton, ecngte formation, and is expansion mechanism: UL ERéet of CaO, NeOH and NaCi oncisions, Cement and Concrete Reseach (12) (1982) 247-256 [5] ¥. Wang, M, Su, R. Yang, B. Lui, A quantatve study of pase microstructures and hydration characters of sulphoshuminae cement, Proceedings ofthe th International Congress on the Chemisty of Cement, New Deli, vol 1V, 1992, pp. 454—460, [15] V- Kassel, Tsakiris, stay the hydration products of « non-expansive sulfehminate cement, Cement and Concrete Rescarch 25 (8) (1995), 1726-1736 chromium