Electrowinning of Iron

Transcription

1 Electrowinning of Iron Geir Martin Haarberg 1 Department of Materials Technology Norwegian University of Science and Technology Trondheim, Norway

2 Background Iron production causes large emissions of CO 2 ULCOS A large EU project supported by the steel industry to develop new methods to produce iron Electrolysis using an inert oxygen evolving anode is an alternative to produce iron without CO 2 emissions 2

3 3 Background CO 2 emissions

4 4 Background CO 2 emissions

5 5 Background CO 2 emissions

6 6 Background CO 2 emissions

7 7 Background CO 2 emissions

8 CO 2 from an average electricity source Average electricity source Coal : 36.5 % Nuclear : 17.4 % Oil : 4.8 % Hydro Power : 19.2 % Natural Gas : 19.1 % Other renewables : 3.0 % CO 2 emission from an average electricity production (g CO 2 )/(kwh el ) = 558 8

9 Standard electrode potentials in aqueous solution at 25 o C Elektrodereaksjon E o /V 1/2 Cl 2 (g) + e - = Cl - 1,36 1/2 O 2 (g) + 2H + + 2e - = H 2 O 1,23 Ag + + e - = Ag(s) 0,80 Fe 3+ + e - = Fe 2+ 0,77 Cu e - = Cu(s) 0,34 AgCl(s) +e - = Ag(s) + Cl - 0,22 H + + e - = 1/2 H 2 (g) 0 Pb e - = Pb(s) -0,13 Sn e - = Sn(s) -0,14 Ni e - = Ni(s) -0,24 Co e - = Co(s) -0,28 Fe e - = Fe(s) -0,44 Zn e - = Zn(s) -0,76 Mn e - = Mn(s) -1,18 Al e - = Al(s) -1,68 Na + + 2e - = Na(s) -2,71 9

10 10 World production of iron and steel

11 Iron ULCOS project Iron smelting by carbon reduction of Fe 2 O 3 CO 2 emissions ULCOS (Ultra Low CO 2 Steelmaking) Purpose: Develop a new process for iron production with reduced CO 2 emissions 11

12 Iron ULCOS project Iron smelting by carbon reduction of Fe 2 O 3 CO 2 emissions ULCOS (Ultra Low CO 2 Steelmaking) Purpose: Develop a new process for iron production with reduced CO 2 emissions 12

13 13 48 organizations

14 13 countries + the EU Norway Sweden Finland Denmark Netherlands Belgium Germany United Kingdom Luxembourg France Austria Portugal Italy Spain 14

15 15 Electrowinning of Iron - Literature

16 Iron Electrowinning process? Possibilities Problems/challenges Aqueous solutions Molten salts Molten oxides Low current efficiency Low current density? Large space required Low Fe 2 O 3 solubility? No inert anode? Corrosive electrolyte Electronic conduction No inert anode 16

17 Iron Electrolysis in Molten Salts - possible approaches Process Traditional dissolved Fe 2 O 3 Problems/challenges Inert anode Fe 2 O 3 solubility High temperature liquid iron product Inert anode Materials Fe 2 O 3 solubility FFC Cambridge deoxidation of solid Inert anode Productivity/current density Fe 2 O 3 in CaCl 2 17

18 Raw material Electrowinning of Iron Principles/requirements - Commercially available Fe 2 O 3 Anode -Inert, O 2 evolving Product - Pure, liquid Fe 18

19 Electrowinning of Iron from Molten Salts Energy and heat ½Fe 2 O 3 (diss) = Fe (s) + ¾ O 2 (g) 800ºC: ΔG o = J/mol, ΔH o = J/mol E rev = V, E iso = V Current density: 0.5 A/cm 2 Cell voltage: 2.2 V Current efficiency: 0.90 Energy consumption: 3.5 kwh/kg Fe 19

20 Estimated space requirements Electrolysis vessel (500 ka ~ 7.5 t Fe/day) 50 cathode in parallel + 51 anodes in parallel Average current through each cathode 10 ka Cathode dimension (H*W ) 0.5mx2 m Average anode cathode distance: 2 cm Average cathode thickness 2.5 cm Anode thickness: 5 cm Volume of molten salt: ~ 1.6 m 3 Mass of molten salt ~ 4 t Area ( footprint ) ~ 10.5 m 2 Area incl. lining, busbar & working space ~ 30 m 2 Total area for 360 cells arrange side by side ~ m m 5m 2.05 m 20

21 Life Cycle Analysis GWP comparison EAF Electrowinning Beneficiation 2 kg CO 2 kg CO BOF BF Beneficiation 1 Iron ore extraction Feed BF EUM BF NUC EW EUM EW CD EW NUC

22 Alkaline electrowinning of iron: concept Cathodic main reaction: Fe 2 O H 2 O + 6 e 2Fe + 6OH Secondary reaction: FeO H 2 O + 2 e Fe + 4 OH Secondary reaction: FeO 2 + e FeO 2 2 Side reaction: 2 H 2 O + 2 e H OH Anodic main reaction: 4OH O H 2 O + 4 e Secondary reaction: FeO 2 2 FeO 2 + e Carbon/iron RDE (cathode) e e Nickel (anode) O2 gas Fe 2 O 3 particle suspension in H 2 O NaOH electrolyte 22

23 Alkaline electrowinning of iron: laboratory cell 2 cm Connected to rotator Stirrer Thermometer Teflon lid Shaft Hg/HgO reference KOH aq. solution (1 M) Bridge Plastic capsule Nickel screen (anode) Carbon rotating disk electrode (cathode) 33 wt %NaOH 33 wt %H 2 O 33 wt %Fe 2 O 3 electrolyte Teflon container Stirrer Rotator Nickel (anode) lead C-RDE (cathode) plug Reference system Heating circulator 23

24 Alkaline electrowinning of iron: cell operation conditio Iron ore: Hematite (Fe 2 O 3 ; average particle size: 10 μm) Electrolyte: 33 wt %H 2 O 33 wt %NaOH 33 wt %Fe 2 O 3 (2149 g) Cathode: Carbon rotating disk electrode (8 mm diameter, 50 mm 2 surface area) Anode: Nickel mesh (317 mm length, 50 mm width) Reference: Hg/HgO in KOH (1 M) Temperature: 110 C Current density:1000, 3000, 6000 A m 2 Anode-cathode distance: 50 mm Rotation rate: 0, 100, 500, 1000, 1200, 1500, 2000, 2200, 2500, 3000 rpm IM6 electrochemical workstation 24

25 Current efficiency, at various rotation rates, current densities 100 j = 1000 A m ω = 1500 rpm Current efficiency, η(%) Current efficiency, η(%) Rotation rate, ω (rpm) Current density, j (A m 2 ) 25 95% was obtained at rotation rates 1200~2000 rpm At high current densities 6000 A m 2, 88% was obtained Energy consumption: 3 kwh/kg (U = 2 V, η = 95%)

26 Electrolytic reduction of Fe 2 O 3 pellet Carbon/iron RDE (cathode) Nickel (anode) O 2 gas Fe 2 O 3 particle suspension in H 2 O NaOH electrolyte Fe 2 O 3 pellet (cathode) e e Nickel (anode) O2 gas wt. % NaOH solution

27 Electrolytic reduction of Fe 2 O 3 pellet Sintered pellet Sintering temp.: 800 C; time: 3 hrs Reduced pellet Current: 300 ma, time: 19.5 h Pt wire embedded in pellet 27 5 mm

28 Electrolytic reduction of Fe 2 O 3 pellet time / Ks Reduced pellet Current: 300 ma, time: 19.5 h current / m A potential / V mm

29 Summary - Alkaline electrowinning of iron 95% current efficiency was obtained at rpm Crystal size decreases with increased j, porosity decreases with increased ω Purity of iron deposit: ~99 wt % Direct electrodeoxidation of solid hematite pellet is also possible 29

30 Middle Temperature Electrolysis Electrowinning of iron from molten salts 30

31 Literature Very little has been found in the literature on the solubility of iron oxide in chloride melts KCl NaCl 10 MgCl 2 => wt % Fe 2 O 3 (720 C) 40 ZnCl 2 60 NaCl => 0.05 wt % Fe 2 O 3 (547 C) 40 ZnCl 2 60 KCl => 0.08 wt % Fe 2 O 3 (547 C) 36 CaCl 2 64 LiCl => 0.12 wt % Fe 2 O 3 (547 C) 55 CaCl 2 45 NaCl => 0.15 wt % Fe 2 O 3 (547 C) Calcium chloride is a candidate electrolyte (very high CaO solubility) 31

32 Experimental Glassy carbon or graphite crucible Techniques Linear sweep cyclic voltammetry Potential step chronoamperometry Bulk electrolysis Electrodes Working electrode: W, C Counter electrode: C Reference electrode: Ag/AgCl or Fe Cathode: C, W, Fe Anode: C, nickel ferrite 32

33 Electrolytes tested: CaCl 2 - NaCl (80-20 and mole %, 800 C) CaCl 2 -NaCl -AlF 3 /AlCl 3 ( mole %, 800 C and mole %, C) CaCl 2 -NaCl -MgCl 2 ( mole %, C) KF NaF (50 50 weight %, 800 C) NaCl KF NaF ( weight %, 800 C) 33

34 CaCl 2 NaCl - AlF 3 ( mol%) Voltammetry, 0.2 V/s, W, 800 C Fe Fe(II) + 2e 0.2 i [Acm -2 ] ,1 mole % Fe2O3-0.3 Fe(II) + 2e Fe E [V] vs. Ag/AgCl reference AlF 3 increases the solubility of Fe 2 O 3 Fe deposition from Fe(II), but poor metal quality 34 Fe 2 O 3 (s) + 2 AlCl 3 (liq) = 2 FeCl 2 (liq) + Al 2 O 3 (s) + Cl 2 (g) G 0 800C = -28,5 kj Fe 2 O 3 (s) + 3 AlCl 3 (liq) = 2 FeCl 2 (liq) + 3 AlOCl (s) + Cl 2 (g) G 0 800C = -177,9 kj

35 Solubility of Fe 2 O 3 (80 CaCl 2-20 NaCl + AlF 3 at 800 C) Fe2O3 solubility [mole %] 4,5 4,0 3,5 3,0 2,5 2,0 1,5 1,0 0,5 0, AlF 3 content [mole %] ICP Titration with KMnO4 35

36 Voltammetry Molten 40NaF-60KF + 0.1Fe 2 O o C, glassy carbon, 0.5 V/s 0.60 Fe Fe(III) + 3e i [Acm -2 ] Fe(III) + 3e - Fe E [V] vs Fe reference 36 Cathode reaction: Fe(III) + 3e - = Fe Diffusion controlled transport of Fe(III)

37 SUMMARY CaCl 2 based melts Enhanced solubility (> 2 mol% Fe 2 O 3 ) upon addition of AlF 3. In melts containing AlF 3 /AlCl 3 : 1. Chemical reaction: (Problem!) Fe 2 O 3 (s) + 2 AlCl 3 (liq) = 2 FeCl 2 (liq) + Al 2 O 3 (s) + Cl 2 (g) 2. Cathode reaction: Fe 2+ (diss) + 2e - = Fe (s) Diffusion coefficient of Fe(II): ~ cm 2 s Bulk electrolysis: Pure iron product (XRD), but poor adherence to cathode Inert anode performance was promising

38 SUMMARY Molten KF NaF Low solubility (~1 mol% Fe 2 O 3 ) Cathode reaction: Fe 3+ (diss) + 3e - = Fe (s) Diffusion controlled deposition Bulk electrolysis: Pure iron product (XRD), but poor adherence to cathode. High cathodic current density (~1 A/cm 2 ) at a rotating cathode. Inert anode performance was promising (Inert anode is produced in laboratory) 38

39 Electrolytes CaCl 2 -NaCl and mole %, 800 C CaCl 2 -NaCl -AlF 3 /AlCl mol%, 800 C mol%, C CaCl 2 -NaCl -MgCl mol%, C KF NaF weight%, 800 C NaCl KF NaF weight %, 800 C LiF - KF - NaF mol%, 800 C CaCl 2 -KF mol%, C CaCl 2 -KF-NaF mol%, C 39

40 Previous results Electrolyte composition / mole% and temperature / C Solubility of Fe 2 O 3 / mol% Comments 80CaCl 2 20NaCl 800 C Very low No Fe deposition 50-80CaCl NaCl 5-15AlF C High solubility Fe 3+ is not stable Fe 3+ + Cl - Fe /2Cl CaCl NaCl 5-15MgCl C High solubility Fe 3+ is not stable Fe 3+ + Cl - Fe /2Cl NaF 60-40KF C 0.27 Relatively low solubility Fe 3+ is stable New studies 40 Fluoride melts with CaCl 2 (CaCl 2 -KF (15-85 mol%), CaCl 2 -KF-NaF ( mol%), CaCl 2 -CaF 2 (80-20 mol%))

41 Experimental cell Ar Thermocouple Rotating working electrode (Fe rod, 260 rpm) Furnace Carbon crucible Electrolyte (CaCl 2 -KF(15-85 mol%), CaCl 2 - KF-NaF( mol%)), CaCl 2 - CaF 2 (80-20 mol%) 41 Counter electrode (Magnetite (Fe 3 O 4 ) or NiFe 2 O 4 -Ni(10 wt%)-cu(10 wt%)) Reference electrode (Pt wire, Fe rod) Working electrode (Mo, W, Ni, Fe)

42 Experimental Electrochemical methods Cyclic voltammetry, chronoamperometry Potentiometric titration Bulk electrolysis Electrodes Working electrode: W, C Counter electrode: C, nickel ferrite Reference electrode: Ag/AgCl or Fe Cathode: C, W, Fe Anode: C, nickel ferrite, magnetite Characterisation/analyses XRD, SEM/EDA and SEM/EPMA ICP and LECO 42

43 Oxygen evolving anode materials Magnetite anode Composition: 95 wt% Fe 3 O 4 5 wt% Fe Sintering: 1100 ºC, in Ar for 2 hours Conductivity 20ºC: 2400 ohm -1 m cm 6 cm Ferrite anode Composition: NiFe 2 O 4 -Ni(10wt%)-Cu(10wt%) Sintering: 1350 ºC in Ar 1 cm 43 5 cm

44 Cyclic voltammetry in molten CaCl 2 -KF Blank After adding Fe 2 O 3 Current density / Acm K or Ca deposition Current density / Acm (a) 0 Fe (b) Ni Reduction of Fe Potential / V vs. Pt Cyclic voltammograms of Ni in CaCl 2 -KF (15-85 mol%) at 867 o C. Scan rate was 0.1 Vsec -1. Potential / V vs. Fe Cyclic voltammograms of (a) Fe and (b) Ni in CaCl 2 - KF (15-85 mol%) at 867 o C. 1.2 mol% Fe 2 O 3 was added. Scan rate was 0.1 Vsec

45 Cyclic voltammetry in molten CaCl 2 -CaF i p c /[Acm-2 ] V/s 0.1V/s 0.2V/s 0.3V/s 0.4V/s 0.5V/s i c p /Acm E[V] vs Fe reference CV s of Mo in molten CaCl 2 -CaF 2 -Fe 2 O 3 ( mol%), 827 C v 1/2 /(V/s) 1/2 Cathodic peak current density vs square root of sweep rate i p c /Acm Reversible cathode reaction Fe (III) + 3 e - Fe (s) Controlled by diffusion Fe(III) towards cathode 45 c Fe2O3 /mol% D Fe(III) = cm 2 s -1 Peak current density vs content of Fe 2 O 3

46 Solubility of Fe 2 O 3 Molten CaF 2 -KF ( mol%) RT E= E + ln C nf Potential / V vs. Pt n ~ 3.0 Solubility ~ 0.33 mol% Fe 2 O 3 mol% OCP change of Fe with addition of Fe 2 O 3 in CaF 2 -KF ( mol%) at 843 o C 46

47 Solubility of Fe 2 O 3 Summary of solubility measurements. KF content / mol% CaCl 2 content / mol% Solubility / mol% Electron number KF-NaF LiF-NaF-KF CaF 2 -KF Fe 2 O 3 solubility / mol% With CaCl 2 Without CaCl 2 CaCl 2 -KF-NaF KF content / mol% CaCl 2 -KF Relationship between Fe 2 O 3 solubility and KF or CaCl 2 content. Fe 3+ is stable in these melts. 47

48 Bulk electrolysis 0.85 Acm -2, molten CaCl 2 -KF, 1144 K Cathode potential / V vs. Pt Cell voltage / V W.E. Fe rod cathode with rotation (260 rpm) C.E. Magnetite (Fe 3 O 4 ) anode R.E. Pt wire 1.5 mol% Fe 2 O 3 addition Time / sec 0.8V is the cathodic limit potential of this melt Changes of cathode potential and cell voltage during galvanostatic electrolysis at 0.85 Acm -2 in CaCl 2 -KF-Fe 2 O 3 (1.5 mol% Fe 2 O 3 added) melts at 871 o C 48

49 Galvanostatic electrolysis 0.85 Acm -2 in CaCl 2 -KF at 1144 K Intensity / cps Fe FeO CaF 2 Pure iron Small amount of impurities 1000 Electrolyte CaF θ (Cu-Kα ) Rinsing the deposit with distilled water FeO XRD pattern of the deposit obtained after galvanostatic electrolysis at 0.85 Acm -2 in CaCl 2 -KF-Fe 2 O 3 (1.5 mol% Fe 2 O 3 added) melt at 1144 K 49

50 Conclusions Pure iron can be deposited from molten salts Fe(III) species are stable in mixed fluoride/chloride melts High current density ( 0.85 Acm -2, in CaCl 2 -KF, rotating cathode ) High current efficiency (> 90 %) Oxygen evolving anode materials show promising behaviour 50

51 Molten oxide electrolytes Electrolyte composition: 45 % SiO 2-18 % MgO - 18 % Al 2 O 3-9 % CaO wt % Fe 2 O 3 51

52 Cyclic voltammetry- Experimental Oxide: 45 % SiO 2, 18 % MgO, 18 % Al 2 O 3, 9 % CaO premelted and crushed after cooling 1-10 wt % Fe 2 O 3 added after premelting Mo crucible Mo working electrode and reference electrode Pt counter electrode Two thermocouples, one above crucible, type B, and one in support under crucible, type S. Ref. + electrode - Sintered alumina sheathing tube 2mm Mo rod (ref.el.) Thermocouple (type B) 2mm Mo cathode rod Pt-anode (braided, spent thermocouple wire) Mo crucible Oxide electrolyte 52

53 Experimental Procedure: The furnace was heated stepwise to 1300ºC, 1350ºC, 1370ºC, 1385ºC, 1400ºC, 1500ºC, 1550ºC, with cyclic voltammetry measurements at 1400ºC, 1500ºC, and 1550ºC and chronopotentiometry during heating from 1400ºC to 1500ºC. The reason for choosing these 3 temperatures is that when we add Fe 2 O 3 as source of Fe-ions, Fe 2 O 3 will decompose to Fe 3 O 4 and oxygen gas at 1450ºC. At 1535ºC, Fe will melt. So, temperatures below and above decomposition of Fe 2 O 3, and above melting of Fe was chosen. Chronopotentiometry was performed during heating to 1500ºC to see if a change in voltage due to the decomposition could be detected 53

54 A constant current (0,05 A) was applied during heating => oxidation at working electrode. The recorded potential decreased as expected with increasing temperature. When the temperature reached 1460 ºC, a new decrease in voltage was detected. This is probably due to the decomposition of hematite that occurs at 1450 ºC: 3 Fe 2 O 3 = 2 Fe 3 O 4 + 1/2 O 2 Chronopot ent iomet ry WE potential [V vs Mo ref.] Started increasing temp. from 1400 t o 1500 Reached 1460 Oxygen evolut ion st art s? Temp.stableat Time (minutes) 54

55 Temperature effect: Cyclic voltammograms obtained at 1400ºC, 1500ºC and 1550ºC 1,0 0,8 0,6 0,1 V/s 1550ºC 1500ºC 0,4 1400ºC I [A] 0,2 0,0-0,2-0,4-0, degrees 1500 degrees 1550 degrees -0,8-0,6-0,4-0,2 0,0 0,2 0,4 0,6 0,8 E [V] vs Mo ref 55

56 I [A] scan 2. scan E [V] versus Mo reference 56

57 Pictures taken after the experiment The Mo- and Pt-wires don t seem to be worn MoO 2 has been formed on radiation shields and can be seen as shiny particles 57

58 The Mo crucible seemed to be intact! Possible to reuse? How to remove oxide slag?? Negligible weight loss after experiment. Crucible: g Oxide: 180,02 g Fe 2 O 3 : 20,00 g Total: 600,14 g After experiment: 599,62 g Weight loss: 0,15 %, probably due to melt removed when raising the electrodes Comment: Negligible amount current passed through the cell. Crucible not polarized 58

59 Summary - Cyclic voltammetry experiments This type of experiments causes minimal wear on the cell components - this in contrast to electrolysis experiments The three electrode arrangement seemed to work well the Mo wire used as reference electrode appeared to be stable in a Mo-crucible So far these measurements has been of a preliminary nature but the voltammograms obtained indicates that iron [as Fe(II) or Fe(III)] is reduced in one step, while oxidation of deposited Fe may occur both via Fe(II) and Fe(III)? The high temperature and the vicious character of the electrolyte will broaden the peaks, making the interpretations more difficult. Alloying with Mo causes some uncertainty. In view of the relative good voltammograms we have obtained we expect that more experiments will help understanding the reaction mechanism in this system. 59

60 Acknowledgments Eirin Kvalheim, Dongju Zhao, Odne Burheim, Hans Martin Jahren, Boyan Yuan - NTNU Sverre Rolseth, Henrik Gudbrandsen, Ana M. Martinez SINTEF Shulan Wang, Tsuyoshi Murakami, Stanislaw Pietrzyk - visitors 60