UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Electrochemical aspects of D.C. electroslag remelting Beynon, Gordon Thomas 1971

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1971_A1 B48.pdf [ 9.89MB ]
Metadata
JSON: 831-1.0079201.json
JSON-LD: 831-1.0079201-ld.json
RDF/XML (Pretty): 831-1.0079201-rdf.xml
RDF/JSON: 831-1.0079201-rdf.json
Turtle: 831-1.0079201-turtle.txt
N-Triples: 831-1.0079201-rdf-ntriples.txt
Original Record: 831-1.0079201-source.json
Full Text
831-1.0079201-fulltext.txt
Citation
831-1.0079201.ris

Full Text

THE ELECTROCHEMICAL ASPECTS OF D.C. ELECTROSLAG REMELTING by GORDON BEYNON B.A.Sc, Uni v e r s i t y of B r i t i s h Columbia, 1967 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY I i n the Department of METALLURGY We accept this thesis as conforming to the required standards THE UNIVERSITY OF BRITISH COLUMBIA October, 1971 In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f the requirements f o r an advanced degree at the U n i v e r s i t y o f B r i t i s h Columbia, I agree t h a t the L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree t h a t p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y purposes may be g r a n t e d by the Head o f my Department or by h i s r e p r e s e n t a t i v e s . I t i s understood t h a t c o p y i n g or p u b l i c a t i o n of t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be allowed without my w r i t t e n p e r m i s s i o n . Department The U n i v e r s i t y o f B r i t i s h Columbia Vancouver 8 , Canada Date ABSTRACT It i s predicted from the known i o n i c properties of the slags used i n e l e c t r o s l a g melting, that the D.C. melting process should be accompanied by Faradaic reactions on the slag/ingot and slag/electrode i n t e r f a c e s . In the present work we have determined the magnitude of the overpotentials r e s u l t i n g from concentration p o l a r i z a t i o n at these Interfaces, f o r the case of pure i r o n i n contact with CaF 2 + AljO^ a n < * CaF,, + CaO slags. This has been c a r r i e d out using a galvanos t a t i c pulsing technique i n an e l e c t r o l y t i c c e l l . The p o l a r i z a t i o n overpotential e x i s t i n g on an electrode i n an operating ESR unit has been measured e s s e n t i a l l y by the same technique. I t was found that the pot e n t i a l s observed on melting ESR electrodes agree w e l l with the r e s u l t s from the e l e c t r o l y t i c c e l l . The primary anodic process i s postulated to be the corrosion of i r o n , leading to an 2+ Fe - saturated slag layer on the anode surface at s u f f i c i e n t l y high current d e n s i t i e s . The cathodic process i s suggested to be the 3 + 2 + Faradaic reduction of A l or Ca , to produce a concentration of [ A l ] _ or (Ca) n i n the cathode i n t e r f a c e region. Fe slag The concentration p o l a r i z a t i o n behaviour of other pure electrode materials (Cr, N i , Co) was also investigated i n CaF 2 + A1 20 slags. For these materials, i t has been postulated that the primary anodic process i s the corrosion of the metal which again leads to i n t e r -face saturation by the appropriate metal ions. The concentration p o l a r i z a t i o n behaviour of i r o n a l l o y e l e c t -rodes (Fe-Cr, Fe-C) was also investigated. I t was found that the more e a s i l y o xidizable a l l o y i n g elements are p r e f e r e n t i a l l y removed from i i i these a l l o y s when they are anodically p o l a r i z e d . At very high current d e n s i t i e s both the anodic and cathodic processes may convert to arcs, leading to process i n s t a b i l i t y . The chemical and thermal phenomena associated with D.C. ESR were studied by making ingots from pure electrode materials and i r o n - a l l o y electrode materials. Chemical e f f e c t s which include metal oxidation and possible a l l o y loss were explained i n terms of the Faradaic reaction mechanisms proposed i n the e l e c t r o l y t i c c e l l studies. Thermal e f f e c t s were also explained on the same basi s . It was concluded that electrochemical r e a c t i o n products were responsible for the formation of large numbers of small oxide i n c l u s i o n s i n the pure metals. In the case of a l l o y materials, electrochemically produced oxidation resulted i n s i g n i f i c a n t losses of a l l o y components during melting. TABLE OF CONTENTS Page TITLE PAGE i ABSTRACT i i TABLE OF CONTENTS i v LIST OF TABLES i x LIST OF FIGURES . x ACKNOWLEDGEMENTS x i i i CHAPTER 1. INTRODUCTION 1 1.1 The El e c t r o s l a g Remelting. Process 1 1.2 The problem 2 1.2.1 Chemical e f f e c t s . 3 1.2.2 Mechanical e f f e c t s 4 1.2.3 Thermal e f f e c t s 5 1.3 Background to the problem 5 1.3.1 E l e c t r o n i c conduction 6 1.3.2 Ionic conduction 7 1.3.2.1 Interfaces at equilibrium 10 1.3.2.2 Polarized electrode i n t e r f a c e s 11 1.3.2.3 P o l a r i z a t i o n mechanisms 12 1.3.2.4 Mass transport 15 1.3.2.5 T r a n s i t i o n time 18 1.3.2.6 Ionic chemical and thermal e f f e c t s 23 1.3.3 Arc discharge 24 1.4 Conclusions 26 1.5 Experimental program o u t l i n e , 27 i v V Page CHAPTER 2. SMALL SCALE STUDIES 29 2.1 Experimental considerations . . . . 29 2.2 Small scale p o l a r i z a t i o n apparatus 32 2.3 Electrode materials 41 2.3.1 Ferrovac-E pure i r o n 41 2.3.2 AISI 430 s t a i n l e s s s t e e l 41 2.3.3 Fe - 1.0 wt. % Cr a l l o y 41 2.3.4 Pure chromium 41 2.3.5 Pure n i c k e l 42 2.3.6 Pure cobalt 42 2.3.7 Iron-carbon a l l o y . . 42 2.4 E l e c t r o s l a g remelting p o l a r i z a t i o n experiments 43 2.5 P o l a r i z a t i o n r e s u l t s 47 2.5.1 Ferrovac-E pure i r o n 47 2.5.2 AISI 430 s t a i n l e s s s t e e l 56 2.5.3 Fe - 1.0 wt. % Cr a l l o y 56 2.5.4 Pure chromium 60 2.5.5 Pure n i c k e l 60 2.5.6 Pure cobalt 65 2.5.7 Iron-carbon a l l o y 65 2.6 E l e c t r o s l a g remelting p o l a r i z a t i o n . r e s u l t s . 65 CHAPTER 3. MELT PROGRAM . 72 3.1 Melting procedure 72 3.2 Electrode materials 72 3.2.1 Low carbon mild s t e e l : AISI 1018 73 3.2.2 Ferrovac-E pure iron 73 3.2.3 Armco i r o n 74 3.2.4 F e r r i t i c s t a i n l e s s s t e e l : AISI 430 74 3.2.5 Medium carbon mild s t e e l : AISI 1095 75 v i Page 3.2.6 Pure n i c k e l 75 3.3 Slag materials 76 3.3.1 Calcium f l u o r i d e . 76 3.3.2 Alumina '. . 77 3.3.3 Calcium oxide 77 3.4 Atmospheric control 78 3.5 Melting conditions 78 3.6 Melt records 81 3.7 Melt record c a l c u l a t i o n s 82 3.7.1 Melt rate 82 3.7.2 S p e c i f i c coulombic density (Z) 83 3.7.3 Drop s i z e and surface tension 84 3.7.4 Melt program r e s u l t s 85 3.8 Ingot analysis 85 3.8.1 Oxygen analysis of ingots 85 3.8.2 Aluminum analysis of FVE ingots 87 3.8.3 Analysis of AISI 1095 s t e e l 87 3.8.4 Analysis of AISI 430 s t a i n l e s s s t e e l 87 3.9 Slag cap analysis 87 CHAPTER 4. DISCUSSION OF SMALL SCALE STUDIES 99 4.1 Introduction . . 99 4.2 Previous electrochemical work 99 4.3 Anodic p o l a r i z a t i o n of pure i r o n i n A^O^ slags 101 4.3.1 Apparent t r a n s i t i o n time . 115 4.4 Anodic p o l a r i z a t i o n of pure i r o n i n CaO slags 119 v i i Page 4.5 Cathodic p o l a r i z a t i o n of pure i r o n i n ESR slags 120 4.6 P o l a r i z a t i o n of Fe-Cr a l l o y s and pure chromium . . . . . . . 125 4.7 P o l a r i z a t i o n of pure n i c k e l . 127 4.8 Anodic p o l a r i z a t i o n of pure cobalt 129 4.9 Anodic p o l a r i z a t i o n of an Fe-C a l l o y . 130 4.10 High, current density p o l a r i z a t i o n 130 4.11 E l e c t r o s l a g process p o l a r i z a t i o n 136 CHAPTER 5. DISCUSSION OF MELT PROGRAM RESULTS 137 5.1 Introduction . . . . . 137 5.2 E f f e c t of electrode p o l a r i t y on oxygen content 138 5.3 Inclusion types 142 5.4 E f f e c t of s p e c i f i c coulombic density (Z) on the f i n a l oxygen content 148 5.5 E f f e c t of atmosphere control 148 5.6 D i f f u s i o n of oxygen into an anodic electrode 149 5.7 Drop s i z e and surface tension 153 5.8 Sign i f i c a n c e of i r o n i n the slag caps 154 5.9 Melt rate and heat generation 156 5.10 Calcium oxide slags and ingot porosity 158 5.11 E f f e c t of increasing ingot/electrode diameter r a t i o 158 5.12 E f f e c t of aluminum addition at the electrode during melting. 159 5.13 A l l o y losses during remelting 160 5.13.1 AISI 1095 s t e e l 160 5.13.2 AISI 430 s t a i n l e s s s t e e l 162 y i i i Page 5.14 Remelting of n i c k e l electrodes. . . 163 5.15 A.C. melting of pure i r o n . . . 164 5.16 Electrochemical phenomena i n commercial ESR 164 CHAPTER 6. CONCLUSIONS. . . . . . . . . 166 REFERENCES . . 168 LIST OF TABLES Table Page I. Melt Record Results of AISI 1018 Mild Steel . . . 88 II . Melt Record Results of FVE Ingots 89 II I . Melt Record Results of FVE Ingots . . 90 IV. Melt Record Results of Armco Iron Ingots 92 V. Melt Record Results of Miscellaneous Ingots 93 VI. Drop Size and I n t e r f a c i a l Tension Results for FVE . . . . 94 VII. T o t a l Aluminum Content of FVE Ingots 95 VIII. Composition of AISI 1095 Electrode and Ingots 96 IX. Composition of AISI 430 Electrode and Ingots 97 X. T o t a l Iron Content of FVE Slag Caps 98 XI. C r i t i c a l Current Density Estimation 132 i x LIST OF;FIGURES Figure Page 1. C e l l e l e c t r i c a l c i r c u i t r y 35 2. General c e l l design 36 3. D e t a i l of c e l l assembly 37 4. Method of obtaining overvoltage 39 5. ESR furnace assembly. 44 6. D e t a i l of reference electrode configuration 45 7. P o t e n t i a l decay between the melting electrode and reference electrode observed at current i n t e r r u p t i o n on an ESR anodic electrode 46 8. Anodic p o l a r i z a t i o n curves f o r pure i r o n i n CaF^ + A^O^slags . 48 9. Cathodic p o l a r i z a t i o n curves f o r pure i r o n i n CaF^ + kl^O^ slags 49 10. Anodic p o l a r i z a t i o n curves f o r pure i r o n i n CaF^ + CaO slags. . 50 11. Cathodic p o l a r i z a t i o n curves f o r pure i r o n i n CaF^ + CaO slags. 51 12. Apparent t r a n s i t i o n time i n the anodic p o l a r i z a t i o n of pure i r o n i n CaF 2 + 5 wt. % A l ^ 52 13. High current density p o l a r i z a t i o n curves f o r pure i r o n i n CaF 2 + 5 wt. % A1 20 3. . . 53 14. Sections through a pure i r o n electrode t i p anodically p o l a r i z e d at high current density 55 15. Anodic p o l a r i z a t i o n curves f o r AISI 430 s t a i n l e s s s t e e l i n CaF 2 + A 1 2 0 3 slags 57 16. Cathodic p o l a r i z a t i o n curves for AISI 430 s t a i n l e s s s t e e l i n CaF 2 + A 1 20 3 slags 58 17. Cr concentration gradient produced at anodic surface of AISI 430 s t a i n l e s s s t e e l . . 59 x Figure Page 18. Anodic p o l a r i z a t i o n curves f o r a Fe + 1 wt. % Cr electrode i n a CaF 2 + 1 wt. % A l ^ s l a g 61 19. Anodic p o l a r i z a t i o n curve f o r pure Cr i n a CaF_ + 1 wt. % A l 2 Q 3 s l a g 7 62 20. Single p o l a r i z a t i o n curve f o r an anodic pulse on pure n i c k e l i n a CaF 2 + 2.5 wt. % A l ^ s l a g 63 21. Section through a pure n i c k e l electrode c a t h o d i c a l l y p o l a r i z e d i n a CaF 2 + 8 wt. % A l ^ slag 64 22. Single p o l a r i z a t i o n curve f o r an anodic pulse on pure cobalt i n a CaF 2 + 2.5 wt. % A l ^ slag 66 23. Anodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n CaF 2 + A l 2 0 3 s l a g s 67 24. Cathodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n CaF 2 + A l 2 0 3 s l a g s 68 25. Anodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n < CaF 2 + A 1 2 0 3 slags 69 26. Cathodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n CaF 2 + CaO slags 70 27. Argon fume hood. : 79 28. Argon gas cap. 80 29. Ingot sampling scheme. . 86 30. Schematic representation of the anodic p o l a r i z a t i o n curve f o r pure i r o n i n CaF 2 + A l 2 0 3 s l a g s 102 31. A p p l i c a t i o n of a l i m i t i n g current density law to anodic p o l a r i z a t i o n of pure i r o n i n CaF 2 + A l 2 0 3 s l a g s . 103 32. P a r t i a l phase diagrams of the system CaF 2 - CaO - FeO 108 33. Estimated l i m i t i n g current density for anodic p o l a r i z a t i o n of pure i r o n i n CaF 2 + A 1 2 0 3 slags . I l l 34. IR drop as a function of applied current, I, f o r anodic p o l a r i z a t i o n of pure i r o n i n a CaF„ +2.5 wt. % A1„0„ slag . . .112 Figure Page 35. Anodic pulse on pure i r o n i n CaF 2 + 250 ppm CaO 114 36. A p p l i c a t i o n of t r a n s i t i o n time law to anodic p o l a r i z a t i o n of pure i r o n i n CaF 2 + 2.5 wt. % A1 20 3 slag 117 37. Successive anodic pulses applied to pure i r o n i n a CaF 2 + 2.5 wt. % A1 20 3 slag 118 38. Cathodic p o l a r i z a t i o n of pure i r o n i n a CaF 2 + 1 wt. % A l ^ ^ slag p l o t t e d against the applied current 122 39. Section from i n t e r i o r of FVE electrode negative ingot. . . .143 40. Section from i n t e r i o r of AISI 430 s t a i n l e s s s t e e l ingot. . .144 41. Alumina inc l u s i o n s on the top of a FVE electrode negative ingot 146 42. A l and 0 i n l i q u i d i r o n i n equilibrium with unit a c t i v i t y A 1 20 3 at 1600°C. 147 43. Various electrode t i p s showing oxide present on an anodic i r o n electrode t i p : 155 A C K N O W L E D G E M E N T S The author i s indebted to Dr. A. M i t c h e l l f o r h i s advice and assistance throughout the duration of t h i s work. Thanks are also due to my fellow graduate students and to var-ious f a c u l t y members for innumerable h e l p f u l discussions. The assistance of Mr. A. Thomas and other members of the tech-n i c a l s t a f f i s greatly appreciated. The f i n a n c i a l assistance of Stelco i s g r a t e f u l l y acknowledged. CHAPTER 1 INTRODUCTION 1.1 The E l e c t r o s l a g Remelting Process E l e c t r o s l a g Remelting (ESR) i s a process i n which a metal e l e c -trode i s consumably melted through a slag to s o l i d i f y i n a water-cooled mold forming an ingot. ESR i s one of the few processes s u i t a b l e f o r up-grading the q u a l i t y of the more complex a l l o y s i n use i n s p e c i a l i z e d s t r u c t u r a l a p p l i c a t i o n s . The most widely used process i n North America i s Vacuum Arc Remelting (VAR) which i s the chief a l t e r n a t i v e to ESR. More expensive processes such as Electron Beam Melting are normally used to process only the rarer metals because of the high operating costs and the small melting capacity of those u n i t s . Any r e f i n i n g action during VAR i s a r e s u l t of metal exposure to the vacuum and i s therefore r e s t r i c t e d to pressure s e n s i t i v e reactions. In contrast, i n ESR the melting metal spends a f i n i t e length of time i n contact with the s l a g . For this reason composition changes req u i r i n g s o l i d / l i q u i d or l i q u i d / l i q u i d reactions are possible. ESR appears to be a more s a t i s f a c t o r y secondary r e f i n i n g process than VAR mainly because of the greater f l e x i b i l i t y i n melting conditions With proper u t i l i z a t i o n of slag chemistry and melting parameters, i t should be possible to achieve rigorous composition and structure control hence producing an ingot with better o v e r a l l properties. The rate of production of an ESR un i t i s controlled by the maxi-mum to l e r a b l e f r e e z i n g r a t e . In general a lower melt rate produces a 1 2 f l a t t e r pool p r o f i l e while a higher melt rate produces.a deeper pool pro-f i l e and possibly unfavourable dendrite o r i e n t a t i o n . At present the maximum ESR ingot s i z e i s 150 tons with a diameter of 3 m., and t h i s can be increased to 200 tons with a 3.5 m. diameter without re q u i r i n g any major technological changes. Power consumption during ESR ranges from 120 to 2000 kwh per ton of metal, again depending on which parameters the operator chooses to c o n t r o l , i n contrast to a t y p i c a l 3 0-arc furnace consumption of 550-650 kwh/ton. Both VAR and ESR add approximately the same cost per ton of metal. In this cost structure e l e c t r i c a l power con-sumption accounts for approximately 30-50% of the t o t a l d i r e c t and i n -d i r e c t costs. Optimization of t h i s power consumption would evidently be a major advantage i n furnace operation. 1.2 The problem The majority of ESR units i n operation today use single-phase, l i n e frequency (60 Hz) A.C. power. I t i s de s i r a b l e that the power supply cables have the shortest possible route from transformer to electrode connection and be s i t e d on the same route as the baseplate return l i n e s . Such an arrangement w i l l reduce the inductive loop and maintain trans-mission e f f i c i e n c y . A power factor of between 0.85 and 0.92 i s normally obtained on such equipment without c o r r e c t i o n measures. As larger e l e c -t r o s l a g furnaces are . b u i l t , i t i s necessary to determine whether they w i l l use A.C. or D.C. power. A.C. power factors tend to decrease with increas-ing ingot s i z e , and although these furnaces could operate at low power; fa c t o r s , the supply a u t h o r i t i e s would not permit t h i s because of wave-shape d i s t o r t i o n on the incoming mains. Power f a c t o r c o r r e c t i o n can be 3 accomplished using v a r i a b l e capacitors but the cost i s exceedingly high. D.C. melting, on the other hand, has no e l e c t r o s l a g c i r c u i t power-factor problems and thus has advantages f o r very large ingot s i z e s . S t r i c t l y from power considerations then, D.C. power i s apparently preferable to A.C. power f o r very large e l e c t r o s l a g furnaces, but such a decision must also be made with respect to possible electrochemical, and thermal phenomena which might be encountered during the passage of large amounts of D.C. current through the e l e c t r o s l a g system. One must there-fore attempt to answer the following questions about D.C. e l e c t r o s l a g melting: 1. What electrochemical, chemical, and thermal changes w i l l occur during remelting? 2. Why do they occur? 3. How important w i l l they be i n terms of a f f e c t i n g the f i n a l chemical and mechanical properties of the ingot metal? 1.2.1 Chemical e f f e c t s I t i s a generally accepted f a c t that D.C. melting of a l l o y mater-i a l s produces a higher loss of e a s i l y oxidized a l l o y i n g elements than occurs during A.C. melting, and also that the main concern i s not with the loss of major a l l o y i n g elements, but instead with changes i n minor components such as T i , S i , A l , 0, S, etc. Etienne (1) reported titanium losses from AISI 321 s t a i n l e s s s t e e l (0.5 wt. % T i i n the electrode) of 40% i n the electrode negative mode and losses of 80% i n the electrode 4 p o s i t i v e mode. Kennard (2) reported T i losses of approximately 30% when remelting 18% Ni Maraging s t e e l (0.80 wt. % T i i n electrode) i n the electrode negative mode. Other workers (3, 4) have reported serious losses of S i (~10%) and A l (~15%) from i r o n a l l o y s remelted with D.C. power. Whittaker (5) and Holzgruber (6) found varying l e v e l s of oxygen and sulphur i n ingots made using electrode negative, electrode p o s i t i v e , and A.C. power. As mentioned previously, oxidation losses are higher during D.C. melting. These chemical e f f e c t s must therefore be a r e s u l t of electrochemically produced oxidation. The nature of th i s phenomenon i s unclear due to the lack of d e f i n i t i v e electrochemical studies of metal electrodes i n calcium fluoride-based s l a g s . 1.2.2 Mechanical e f f e c t s General concern for oxidative losses i s not s u r p r i s i n g , since some a l l o y s , where a small concentration of an e s s e n t i a l a l l o y i n g element i s present, are p a r t i c u l a r l y s e n s i t i v e to composition f l u c t u a t i o n s . In Mar-aging s t e e l type 300, for instance, which contains 0.8 wt. % T i , a loss of 0.1 wt. % i n T i w i l l cause the y i e l d strength to drop by 10,000 p s i (3%). It i s apparent, then, that these observed a l l o y losses can be very serious indeed, and one must be aware of such p o t e n t i a l losses during D.C. remelting i n order to decide what mode of melting i s most f e a s i b l e i n large remelting furnaces. E f f e c t i v e loss of a l l o y i n g elements may also occur by formation of oxide i n c l u s i o n s i n the ingot. In th i s case, the elements have not a c t u a l l y been l o s t , but they are no longer a v a i l a b l e to serve t h e i r i n -tended purpose i n the matrix. Consideration must also be given to the d e t r i -mental e f f e c t of inclusions on the mechanical properties of the metal. 5 1.2.3 Thermal e f f e c t s Thermal e f f e c t s i n D.C. E l e c t r o s l a g melting w i l l a r i s e d i r e c t l y i f the mode of current passage at the metal/slag i n t e r f a c e s a l t e r s e i t h e r the rate of heat generation or heat t r a n s f e r . Depending on the type of current conduction i n the s l a g , such changes might a r i s e from a P e l t i e r e f f e c t , from an electrochemical p o l a r i z a t i o n resistance, or from the f o r -mation of an arc. These p o s s i b i l i t i e s w i l l be discussed subsequently. Heat ef f e c t s apparently do e x i s t during e l e c t r o s l a g melting because i t has been noted (7,8) that the three modes of melting (D.C. -ve, D.C. +ve, A.C.) produce differences i n the s p e c i f i c melt-rate (kwh/ton), and, f u r t h e r -more, these findings by d i f f e r e n t i nvestigators are not i n agreement as to which mode i s the most e f f i c i e n t . These differences i n s p e c i f i c melt-rate must a r i s e , i n part, from heat generation e f f e c t s at the electrode, which i n turn must r e s u l t from electrochemical and chemical e f f e c t s associated with the mode of current passage at the slag/metal i n t e r f a c e s . Since i t i s apparent that chemical, mechanical and thermal e f f e c t s are associated with using D.C. power f o r e l e c t r o s l a g remelting, and that the r e s u l t s to date are contradictory regarding these e f f e c t s , i t i s worth-while i n v e s t i g a t i n g these phenomena on both small and large scales. 1.3 Background to the problem Before one can study the occurrence of the e f f e c t s discussed pre-v i o u s l y , i t i s necessary to understand f u l l y a l l the possible methods of current passage i n a l i q u i d metal/fused s a l t system, and i n each case to consider the p o s s i b l e mechanisms which could give r i s e to chemical and thermal e f f e c t s . The three possible modes of current passage between a 6 l i q u i d metal electrode and a fused s a l t are: I. E l e c t r o n i c conduction. I I . Faradaic reactions. I I I . Arc discharge. Each of these have s p e c i f i c c h a r a c t e r i s t i c s which may or may not be able to explain the e f f e c t s being examined. 1.3.1 E l e c t r o n i c conduction If current passage through such a system takes place by e l e c t r o -n i c conduction, i . e . , the slag i s an e l e c t r o n i c conductor, there w i l l be no Faradaic reactions occurring at the slag/metal i n t e r f a c e s . There w i l l therefore be no electrochemically produced material to a f f e c t the chemis-try of the metal and the s l a g . Any chemical changes would r e s u l t only from chemical i n t e r a c t i o n between the slag and the melting metal. Heat gener-ation i n the bulk D f the slag w i l l occur by resistance heating, the amount of heat so generated being equal to the product of the process voltage and current. If any change i n the CaF^ stoichiometry i s to be produced by high temperature reaction i t would be most l i k e l y to lead to metal-excess com-po s i t i o n s . Hence the slag would behave as an n-type semiconductor. It i s possible during e l e c t r o n i c conduction, to generate l o c a l i z e d heating or cooling at an i n t e r f a c e due to the P e l t i e r e f f e c t , which i s important when two materials have d i f f e r i n g electron m o b i l i t i e s . The amount of heat so generated i s given by: q = T T I where ir i s the P e l t i e r c o e f f i c i e n t and I i s the current. The P e l t i e r co-e f f i c i e n t i s given by: 7 TT = CtT where a i s the Seebeck c o e f f i c i e n t and T i s the absolute temperature. The Seebeck c o e f f i c i e n t of one of the best commercial thermoelectric materials (PbTe), i s approximately 400 yV.°C \ but that of a calcium-fluoride s l a g , would l i k e l y be much lower, say approximately 10 yV.°C 1 . At a temper-ature of 2000°K, and a current of 1000 A., the rate of P e l t i e r heating at such a proposed i n t e r f a c e would be 20 cal.sec. \ a n e g l i g i b l e amount i n comparison to the t o t a l process heat generation rate of 6.5 Kcal.sec 1 . Also, such heat generation at one electrode i n t e r f a c e , would be comple-mented by an equivalent amount of heat loss at the other i n t e r f a c e . One can therefore conclude that e l e c t r o n i c conduction during ESR processing would produce no s i g n i f i c a n t chemical changes and that l o c a l i z e d heat generation a r i s i n g from the P e l t i e r e f f e c t would be i n s u f f i c i e n t to a f f e c t the melt rate of the electrode. In view of these deductions, one must, account for any chemical and thermal e f f e c t s observed during e l e c t r o s l a g melting by means other than e l e c t r o n i c conduction, even though the degree of e l e c t r o n i c conduction i n such a metal/fused s a l t system might be quite s u b s t a n t i a l (as w i l l be discussed i n Chapter 5). 1.3.2 Ionic conduction If current passage between l i q u i d metal electrodes i n contact with calcium fluoride-based slags takes place by i o n i c conduction, Faradaic processes must take place at the electrode/slag i n t e r f a c e s . This require-ment has two important consequences. The f i r s t i s that Faradaic r e a c t i o n products are a v a i l a b l e i n the system to react chemically with the metal and the s l a g . The second i s that p o l a r i z a t i o n can occur at the metal/ 8 slag i n t e r f a c e s which could r e s u l t i n l o c a l i z e d excess heat generation and modify i s o - p o t e n t i a l surfaces i n the bulk of the slag. This would a f f e c t the l o c a l dynamic heat balance at the metal/slag i n t e r f a c e s , and would be r e f l e c t e d e i t h e r i n melting e f f i c i e n c y or i n process temperature d i s t r i -butions. Although i t has been stated that calcium fluoride-based slags are completely i o n i c i n nature (9), t h i s i n f a c t means that the degree of e l e c t r o n i c conduction i s extremely small. In pure l i q u i d CaF2» the transport -4 number of electrons i s approximately 10 (10), and arises from a small degree of nonstoichiometry which i s inherent i n a l l compounds at high temperature (11). Calcium f l u o r i d e w i l l have s i g n i f i c a n t e l e c t r o n i c conductivity when calcium i s dissolved i n i t , calcium being completely miscible with calcium f l u o r i d e (12). Although the e f f e c t of dissolved calcium on the e l e c t r i c a l conductivity of calcium f l u o r i d e has not been studied, data are a v a i l a b l e on the e f f e c t of sodium addition to sodium f l u o r i d e and s i m i l a r h a l i d e sys-tems (12). From these data i t appears that a 3O+40 increase i n conductivity f o r 2->-5 mole % addition of calcium to calcium f l u o r i d e i s not unreasonable. This would represent a large degree of nonstoichiometry and therefore a sub-s t a n t i a l degree of n-type e l e c t r o n i c semiconduction. For the moment, calcium f l u o r i d e slags w i l l be presumed to con-duct i o n i c a l l y , and therefore during e l e c t r o s l a g remelting with D.C. power, cathodic and anodic Faradaic reactions take place continuously at the appropriate s i t e s . I t i s d i f f i c u l t to j u s t i f y the presence of substan-t i a l net Faradaic reactions i n an e l e c t r o s l a g furnace i n view of the f a c t that the sla g chemistry i s e s s e n t i a l l y unchanged during melting. The 9 o v e r a l l system r e a c t i o n e f f i c i e n c y must therefore be very low, approxi-mately 1 % , a condition that can a r i s e i n two ways. The f i r s t p o s s i b i l -i t y i s that the i n d i v i d u a l t o t a l anodic and cathodic processes themselves operate at very low current e f f i c i e n c i e s . This would be the case i f there was a s u b s t a n t i a l amount of e l e c t r o n i c conduction. The second p o s s i -b i l i t y i s that, i f the reaction current e f f i c i e n c i e s are high, the i n d i -v i d u a l r e a c t i o n products continuously recombine during remelting such that during melting a small steady state amount of anodic and cathodic products are present i n the s l a g . Observation of the system as a whole would there-fore i n d i c a t e that the reactions behave e f f i c i e n t l y but that the net e l e c -trochemical composition changes i n the slag are small. The lack of s u b s t a n t i a l chemical changes i n the slag implies that the chemical changes which occur i n the metal are also small, although even small losses i n some a l l o y i n g elements can have very serious e f f e c t s on the mechanical properties of the metal. The purpose of t h i s study i s then to formulate Faradaic reaction mechanisms occurring at l i q u i d metal/ fused s a l t anodic and cathodic interfaces which can account for the follow-ing phenomena: I. Chemical changes i n the slag are small. I I . E a s i l y oxidized a l l o y i n g elements are p r e f e r e n t i a l l y removed from the melting metal. I I I . Chemical changes i n the metal are small. IV. Thermal ef f e c t s have been observed and are apparently r e l a t e d to the mode of power used during melting. 10 1.3.2.1 Interfaces at equilibrium It i s u s e f u l at this point to consider the type of metal/slag i n t e r f a c e that would e x i s t between pure i r o n and a calcium f l u o r i d e -based slag at equilibrium, (no net current flow across the i n t e r f a c e ) . If a s o l i d i r o n electrode i s immersed i n a l i q u i d calcium f l u o r i d e -calcium oxide s l a g , equilibrium between the metal and sl a g w i l l be attained by d i s s o l u t i o n of i r o n ions i n the slag at the i n t e r f a c e . A l -though the i r o n surface i s i n equilibrium with the slag, the i n t e r f a c e i s not a s t a t i c one. An exchange current flows, i n which the forward current equals the reverse current. Aqueous systems generally have low -12 -3 -2 exchange current d e n s i t i e s (10 -KL0 A.cm ) but some l i q u i d metal/fused s a l t systems have comparatively high exchange current densities ( i = 210 -2 A.cm. for the Cd(II)-Cd reaction) (13). The exchange current density can be considered to be a measure of the r e v e r s i b i l i t y of an electroche-mical reaction, which i n turns indicates the ease with which an electrode may be p o l a r i z e d . A high exchange current density implies that an electrode i s d i f f i -c u l t to p o l a r i z e . The behaviour of pol a r i z e d electrode i n t e r f a c e s w i l l now be discussed since p o l a r i z a t i o n phenomena can be used to answer some of the questions previously presented with respect to chemical and ther-mal e f f e c t s . At a simple l i q u i d metal/fused s a l t electrode i n t e r f a c e , p o l a r i z a t i o n phenomena have the following consequences. 1. The presence of p o l a r i z a t i o n at such an i n t e r f a c e implies that excess heat w i l l be generated by current passage through the thi n p o l a r i z e d slag layer 11 representing part of the i r r e v e r s i b i l i t y of the re a c t i o n . S i g n i f i c a n t p o l a r i z a t i o n could therefore account f o r any observed thermal e f f e c t s during remelting. 2. P o l a r i z a t i o n w i l l take place by Faradaic reactions, the products of which may lead to small but important composition changes of the metal during melting. 1.3.2.2 Polar i z e d electrode i n t e r f a c e s At equilibrium, the p o t e n t i a l , with respect to a reference e l e c -trode, of an Fe electrode i n contact with a calcium fluoride-calcium 2+ oxide slag i s determined by the concentration of Fe ions at the s l a g / metal i n t e r f a c e . This concentration i s governed by the h a l f - c e l l reac-t i o n : Fe + F e 2 + + 2e~ (1-1) For the case of a simple metal/ion electrode having the electrode reaction valence n = z: Me + Me Z + + z e _ (1-2) the equilibrium electrode p o t e n t i a l , E, i s written i n the form of a Nernst equation: \ o P T a z + E = E + ^ In (1-3) e where E° i s the. standard electrode p o t e n t i a l when a l l reactants and products are at.unit a c t i v i t y . It must be stressed that a l l h a l f - c e l l p o t e n t i a l s are measured with respect to a reference electrode whose p o t e n t i a l i s con-stant f o r a given e l e c t r o l y t e . I d e a l l y , a reference electrode consists of an electrode surface i n contact with e l e c t r o l y t e of a known and in v a r i e n t 12 composition such that a standard reference electrode of known p o t e n t i a l can be used. An example of a standard reference electrode i s the satur-ated calomel electrode. A d i f f i c u l t y that arises with such electrodes i s that i t i s frequently necessary to have two d i f f e r e n t solutions i n contact, which gives r i s e to a l i q u i d junction p o t e n t i a l . The value of the l i q u i d j u nction p o t e n t i a l can be minimized by using a s a l t bridge. In aqueous systems where the s o l u t i o n resistances are generally high i t i s necessary to u t i l i z e t his type of reference electrode i n order that the IR drop between a working electrode and the reference electrode can be minimized by having the reference electrode very close to the working electrode. In fused s a l t systems, the resistances are much lower and i t i s possible to use a remote reference electrode i n contact with the bulk l i q u i d i f the range of l i q u i d compositions studied do not grossly a l t e r the reference electrode p o t e n t i a l . This w i l l be discussed i n d e t a i l i n Chapter 2. The important point at issue i s that a r e v e r s i b l e reference electrode must be used to measure the p o t e n t i a l of the working electrode being studied. If current i s passed through an electrode at equilibrium and the current i s such that the concentration of the electrochemical reactants or products i s a l t e r e d , the electrode i s said to be p o l a r i z e d , and the p o t e n t i a l of the electrode i n t e r f a c e i s altered from the e q u i l i -brium p o t e n t i a l by the p o l a r i z a t i o n overvoltage. 1.3.2.3 P o l a r i z a t i o n mechanisms Vetter (14) defines four types of overvoltage, each one of which i s operative at an i n t e r f a c e i f that p a r t i c u l a r step of the o v e r a l l 13 electrochemical r e a c t i o n i s the slowest or rate-determining process. The types of overvoltage are given below. 1. n t - charge-transfer overvoltage. The transfer of charge c a r r i e r s across the e l e c -t r i c a l double layer i s r a t e - c o n t r o l l i n g . 2. n - r e a c t i o n overvoltage. A slow chemical step i n the o v e r a l l electrode reaction i s rate c o n t r o l l i n g , and the rate constant i s , by d e f i n i t i o n , independent of p o t e n t i a l . 3. ~ d i f f u s i o n overvoltage. Supply of reactants to or removal of reaction products from the electrode i s rate c o n t r o l l i n g . 4. n c - c r y s t a l l i z a t i o n overvoltage. Hindrance of the process by which atoms are incorporated i n t o or removed from the c r y s t a l l a t t i c e leads to c r y s t a l l i z a t i o n overvoltage. In a l i q u i d metal/fused s a l t system at high temperature (~1400°C) i t i s u n l i k e l y that a c t i v a t i o n processes other than d i f f u s i o n would be rate c o n t r o l l i n g . Further discussion w i l l therefore be concerned with d i f f u s i o n a l processes and associated phenomena. D i f f u s i o n overvoltage. D i f f u s i o n overvoltage, r)^, appears when the supply of reactants to the electrode or the removal of r e a c t i o n pro-ducts from the electrode i s rate determining. If a l l chemical processes, including the c r y s t a l l i z a t i o n processes and also the charge-transfer reaction, are i n equilibrium, only d i f f u s i o n overvoltage i s present. 14 The p o t e n t i a l of an electrode experiencing pure d i f f u s i o n p o l a r i z a t i o n can be calculated i n terms of the Nernst equation i n which the concen-t r a t i o n of the components d i r e c t l y at the surface must be used and not those i n the i n t e r i o r of the e l e c t r o l y t e . Therefore, the d i f f u s i o n oyervoltage i s equal to the d i f f e r e n c e : n d = E'-E (1-4) between the equilibrium p o t e n t i a l E and the impressed p o t e n t i a l E' r e -quired f o r current flow. To c o r r e l a t e the d i f f u s i o n overvoltage with concentration changes, a general o v e r a l l electrode reaction w i l l be given. (-v,)S. + (-v„)S.•+ . . .t v.S. + . . . + v S + n e - (1-5) 1 1 1 1 1 1 q q The stoichiometric factors v are p o s i t i v e f or oxidized components and negative for reduced components. At equilibrium the o v e r a l l electrode reaction leads to the-equilibrium electrode p o t e n t i a l E whose concentra-ti o n dependence i s expressed by the Nernst equation: 'f E = E° + EL E v . In a. (1-6) In equation (1-5) a^ are the a c t i v i t i e s a^ = f j c j °f t n e components of the o v e r a l l electrode r e a c t i o n . On current flow the concentrations c_. d i r e c t l y before the surface, are functions of the current density i and of the time t. Hence a^ = a ( i , t) ^ a . Therefore, the p o t e n t i a l E ' ( i , t) i s given i n accordance with the Nernst equation by: RT E ' ( i . t ) = E° + ^ • Zv. • In a . ( i , t ) (1-7) na 2 J and by subs t r a c t i o n of equation (1-6) from (1-7), the d i f f u s i o n over-voltage i s : 1 5 ijrp a ( i , t ) N = E ' - E = — • Zv. « l n -J ( 1 - 8 * 1 d nF i -J a. J D i f f u s i o n p o l a r i z a t i o n therefore r e s u l t s when mass transport of reactants or products to or away from the electrode i n t e r f a c e i s r a t e -c o n t r o l l i n g . This produces concentration changes i n components at the i n t e r f a c e and the r e s u l t i n g d i f f u s i o n overvoltage i s expressed as a Nernst equation i n terms of concentration or a c t i v i t y changes. 1 . 3 . 2 . 4 Mass transport The concept of a d i f f u s i o n boundary layer must be used when d i s -cussing d i f f u s i o n a l transport of species to or away from an electrode i n t e r f a c e . For the simplest case, one assumes a planar electrode sur-face with a d i f f u s i o n layer of thickness 6, which i s assumed to be stationary. Transport of the component S_. having the a c t i v i t y a^ can take place only i f an a c t i v i t y gradient da./dx exists for this compo-nent, x i s the perpendicular distance from the electrode surface. The number of moles N^ . of substance d i f f u s i n g through a cross-section of 2 1 cm. per second i s given by Fick's f i r s t law. For s i m p l i c i t y , the concentration c. w i l l be used i n place of the a c t i v i t y a., and the f l u x J J i s w r i t t e n as: dc. N. = D • ( 1 - 9 ) J j dx A l l components S . of the o v e r a l l electrode reaction ( 1 - 5 ) must J d i f f u s e through t h i s d i f f u s i o n layer. A transfer of one mole of S_. corresponds to a charge amounting to n • F / V j coulombs. Therefore the current density divided by n • F/v_. represents mass transfer i n mole. - 2 - 1 cm. sec. , which passes from the electrode through the layer to the 16 e l e c t r o l y t e . With due consideration to sign i*v. dc. — = -D. • (1-10) nF j dx i s obtained. In the stationary state and i n the absence of homogeneous chemical e q u i l i b r i a through the e n t i r e layer, the flow of the components S. must be constant and a l i n e a r concentration gradient dc./dx = -(c.-c.) J 3 3 3 = constant i s established throughout the e n t i r e l a y e r . Equation (1-10) may now be written as: n • F j 6 In p r i n c i p l e , with the movement of ions through the d i f f u s i o n layer, the influence of an e l e c t r i c f i e l d must also be taken into account. This f i e l d causes an a d d i t i o n a l migration current which i s superimposed on the d i f f u s i o n current. A high excess of an i n d i f f e r e n t e l e c t r o l y t e minimizes this e l e c t r i c f i e l d e f f e c t within the d i f f u s i o n layer to a large extent, since the excess e l e c t r o l y t e causes the transport number t.. of the reacting components to become very small. By assuming that the concentration of ions involved i n the electrochemical r e a c t i o n i s small, the e f f e c t of the e l e c t r i c f i e l d can be neglected. The value of the concentration gradient (c^.-c_.)/6 i s proportional to the current density i > [equation (1-11)], and i n the case of ions d i f f u s i n g to the electrode surface, the concentration gradient w i l l have a maximum value which i s reached when the current density i s such that the i o n i c concentration c. of species S. becomes zero at the e l e c -J J trode surface (c.=0). This maximum concentration gradient i s then c./6 3 3 17 and the corresponding current density i s c a l l e d the l i m i t i n g d i f f u s i o n current density i ^ ..with respect to substance . For each substance i n the o v e r a l l electrode reaction equation (1-5) there i s such a l i m i t i n g d i f f u s i o n current density which i s characterized by the index 3 • D. i , . = - — • F • — £ • c. (1-12) d,j v j 6 j The electrode r e a c t i o n can take place no f a s t e r than the value deter-mined by this l i m i t i n g current density. However, i f a p o t e n t i a l i s im-pressed on the c e l l s u f f i c i e n t to produce a current density i higher than the l i m i t i n g current density i , . , the p o t e n t i a l at the electrode w i l l change to such an extent that a second electrode process may take place at the current density ( i - i ^ ) . The r a t i o of the concentration c./c. i s important for the c a l c u -3 1 l a t i o n of the d i f f u s i o n overvoltage according to equation (1-8). This r a t i o i s e a s i l y derived from equations (1-11) and (1-12), i . e . , c. - c . c. i = _ J 1 = i - _1 1d> j c. c. J J or £- = i _ i ( 1_ 1 3 ) Equation (1-13) i s v a l i d for every component i n the o v e r a l l electrode r e a c t i o n , so that one obtains the general expression f o r the t o t a l d i f f u s i o n overvoltage a f t e r s u b s t i t u t i o n into equation (1-8), i . e . , • * d - § • Z v j • l n C ' b - ^ ( 1 " 1 4 ) 18 The d i f f u s i o n overvoltage r e s u l t i n g from transport of species i s w r i tten i n terms of the current density, i , and the l i m i t i n g current density, i , .. Examination of equations (1-I4)and (1-12) shows that i , Q,J d contains the mass transport v a r i a b l e s inherent i n ' n , . d 1.3.2.5 T r a n s i t i o n time The term chronopotentiometry i s applied to the techniques of e l e c t r o l y s i s at constant current under conditions of l i n e a r d i f f u s i o n con-t r o l . It must be stressed that chronopotentiometry i s a selected form of unsteady state. An important experimental quantity i n chronopotentiometry i s the " t r a n s i t i o n time" x. This time i s the time i n t e r v a l subsequent to the onset of constant current e l e c t r o l y s i s during which the surface concen-t r a t i o n of d i f f u s i n g material undergoing electrode r e a c t i o n reaches zero and a rapid increase of electrode p o t e n t i a l occurs. The value of the t r a n s i t i o n time can be calculated from l i n e a r d i f f u s i o n theory, derived from the follow-ing conditions: 1. Uniform i n i t i a l concentration c^ of d i f f u s i n g species S.. 3 2. Constant concentration gradient at the electrode surface equal to i * v./nFD.. 3 3 The concentration of species S. at the i n t e r f a c e can then be 3 found by solving the p a r t i a l d i f f e r e n t i a l equation for d i f f u s i o n (Fick's Second Law) under the appropriate i n i t i a l and boundary conditions. The i n t e r f a c e concentration c. varies with the distance x and the time t such 3 that c. = c.(x, t) and i s written: .19 9 XT - 1 / 2 2 2 N. t _ x N.x V*' 0 " Zj ~ 1/2 l / 2 , & " 4 Y + D 1 - ' ^ 6 ~ — ( 1 " 1 5 ) D * J y ^ D t where N i s the constant f l u x imposed by the current density i . As stated i n equation (1-10) the fl u x N. i s equivalent to i • v./nF at the J 3 electrode i n t e r f a c e j and th i s can be substituted into equation (1-15) for the boundary condition x = 0 to obtain 2 i v . j ' c.(x = 0,t) = c. - -~^'J^ (1-16) By w r i t i n g 2 i v P = + =•*- • A r (1-17) nF \J TTD : equation (1-16) becomes c.(x = 0,t) = c. - P t 1 / 2 (1-18) 3 3 The expression gives the v a r i a t i o n of concentration of species S^ . at the i n t e r f a c e with time, a f t e r a p p l i c a t i o n of a constant current density, but one seeks also to know the time v a r i a t i o n of the p o t e n t i a l d i f f e r e n c e across the i n t e r f a c e at which the reaction i s occurring. To obtain t h i s r e l a t i o n s h i p , the Nernst equation can be used to r e l a t e the p o t e n t i a l d i f -ference to the i n t e r f a c e concentration. That i s , by s u b s t i t u t i n g (1-18) into the Nernst equation, the electrode p o t e n t i a l measured with respect to a reference electrode i s given by: E'(x = 0,t) = E° + ^  In (c. - P t 1 / 2 ) (1-19) Before one can develop t h i s equation further, consideration must be : given to the material produced by the electrochemical r e a c t i o n . A simple 20 example i s the electro-reduction of f e r r i c ions to form ferrous ions 3+ 2+ (Fe + e -»• Fe ) where the reactant and product are i n s o l u t i o n , and one must take into account the d i f f u s i o n of the electron donor away from the electrode and the v a r i a t i o n of i t s i n t e r f a c i a l concentration with time. I f the reacting species i s and the product species S^, equa-t i o n (1-19) can be written as: RT C a ( x = ° ' t ) E ' ( x - 0 , t ) = E ^ j l n ^ = Q i t ) (1-20) 112 where c (x=0,t) = c - Pt . (1-21) a a If 1 mole of S, i s formed from 1 mole of S and t h e i r d i f f u s i o n b a c o e f f i c i e n t s are the same, one obtains ^ ( x = 0,t) = c b + P t 1 / 2 (1-22) and equation (1-20) becomes: I - P t 1 / 2 E'(x = 0,t) = E° + ^ ~ In -2 -r-pr (1-23) n F c b + P t 1 / 2 If at zero time, the concentration of S, i s n e g l i g i b l e , then c ' = 0 b b and equation (1-23) becomes: •n 1/2 RT °a " P t E'(x = 0,t) = E° + ^ In a ^ 1 / 2 (1-24) Consider the expression (1-24) for the time variant p o t e n t i a l of the working electrode at which a d i f f u s i o n c o ntrolled electro-reduction r e a c t i o n i s stimulated by a constant current switched on at t = 0. The 1/2 - 1/2 product Pt = 0 at t = 0. Hence, c /Pt tends to i n f i n i t y as does 9. 1/2 i t s logarithm. At values of time greater than t = 0 the term Pt i s 21 f i n i t e and at some value of time i s equal to c . This time has been a 1/2 previously defined as T , the t r a n s i t i o n time, and l n ( c - Pt ) tends a to minus i n f i n i t y , and the p o t e n t i a l changes r a p i d l y . In f a c t , i t sinks t i l l i t has become s u f f i c i e n t l y negative so that the electro-reduction of some other i o n i c species can occur. By d e f i n i t i o n , therefore, when 1/2 - 1/2 t = x and (c - Pt ) = 0, c = Px and equation (1-24) can be a a written as: RT P x 1 / 2 - P t 1 / 2 E'(x = 0,t) = E° In ^  ^ (1-25) n F Pt 7 2 where P was defined by equation (1-17) to be and equation (1-25) becomes .1/2 1/2 E'(x = 0,t) = E° + S | In ^ | (1-26) P r i o r to current passage, the equilibrium electrode p o t e n t i a l determined by the concentration of species S i s written as: a ' RT — E = E° + ^ In c (1-27) nF a As defined by equation (1-4), the d i f f u s i o n overvoltage can then be written as> 1/2 _ 1/2 "  E' "  E "  A + §  l n ~ UY-  (1" 28) where the constant A i s given by: RT — A = - ^~ l n c (1-29) nF a 22 By manipulation of the equation 1 / 2 c = P T ^ ( 1 - 3 0 ) a and s u b s t i t u t i o n of equation ( 1 - 1 7 ) , i t can be shown that _ , - 1 / 2 1/2 • 1 / 2 n F C a ff ( 1 - 3 1 ) I T = -2  1/2 The product iT i s seen to be constant f o r d i f f u s i o n c o n t r o l l e d e l e c t r o -l y s i s regardless of the current density. Equation ( 1 - 3 1 ) therefore consti-tutes a simple experimental c r i t e r i o n of d i f f u s i o n c o n t r o l . The preceeding discussion on i o n i c conduction has presented the basic information regarding d i f f u s i o n p o l a r i z a t i o n processes. General expressions were presented to show that the pol a r i z a t i o n -overvoltage can be described i n terms of concentration changes of reactant or product species at the electrode surface, and these changes can be considered i n terms of a Nernst equation: cA±, t) 'd ~ nF ' *wl c. J Using the concept of mass transport at constant current density i n un s t i r r e d s o l u t i o n s , the d e f i n i t i o n of a l i m i t i n g d i f f u s i o n a l current was discussed and i t was shown that the d i f f u s i o n overvoltage can also be expressed i n terms of the current d e n s i t y ( i ) and the l i m i t i n g current density (irf.j) ' n J = • £v. i n -2—— (1-32) 23 The t r a n s i t i o n time, x, was defined as the time a f t e r the on-set of constant current that the concentration of reacting i o n i c species becomes zero at the electrode i n t e r f a c e . The d i f f u s i o n overvoltage can also be expressed i n terms of x and t: . , RT . 1/2 1/2 r\, = A + — In x - t * i 'd nF (1-34) 1.3.2.6 Ionic chemical and thermal e f f e c t s Chemical e f f e c t s . The Faradaic r e a c t i o n products produced at the electrode surfaces are a v a i l a b l e i n the system to i n t e r a c t chemically with the metal and the s l a g . Therefore i o n i c conduction can account for chemical e f f e c t s observed during D.C. e l e c t r o s l a g melting. Thermal e f f e c t s . A polarized electrode i n t e r f a c e through which high current d e n s i t i e s pass w i l l behave i n a quite d i f f e r e n t thermal manner from an unpolarized i n t e r f a c e . Current passage through i o n i c slags produces heat by Joule heating according to the r e l a t i o n s h i p : P = V i (1-35) where P i s the power generation i n watts, and v i s voltage drop i n the sl a g . At an nnpolarized electrode, the voltage gradient i n the slag at the electrode i n t e r f a c e i s the same as that i n the bulk of the s l a g . At a polarized electrode, the voltage gradient i s higher than normal i n the polarized layer and heat generation i n t h i s layer w i l l be according-l y higher. The increased voltage gradient i s produced by the p o l a r i z a -t i o n overvoltage and the excess heat generation i s given by: P = n i (1-36) 24 In this way p o l a r i z a t i o n at l i q u i d m e t a l / l i q u i d slag i n t e r f a c e s can account f o r thermal e f f e c t s observed during D.C. e l e c t r o s l a g melt-ing. I t i s i n t e r e s t i n g to consider the p o s s i b i l i t i e s of chemical and thermal e f f e c t s when melting with A.C. power. I f slow d i f f u s i o n a l processes are responsible for D.C. melting e f f e c t s , and i f the d i f f u -sion times are long i n comparison to the half-wave time of 60 Hz. A.C. power, net chemical e f f e c t s should be zero. This i s to say that reaction products produced on the anodic h a l f - c y c l e should be e n t i r e l y removed by the cathodic h a l f - c y c l e . However,any p o l a r i z a t i o n resistance heating would not behave r e v e r s i b l y . 1.3.3 Arc discharge Consideration w i l l now be given to the case i n which current transfer between a melting electrode and l i q u i d CaF^ based slag takes place by arc discharge. This p o s s i b i l i t y w i l l be discussed i n terms of thermal e f f e c t s and chemical e f f e c t s , and i t i s assumed that the sla g conducts i o n i c a l l y . Very l i t t l e i s known about the mechanisms of current movement through a high current arc, e s p e c i a l l y for arcs between high-temperature sl a g and metal surfaces. In metal electrodes^charge i s conducted by a "degenerate gas" of free electrons, but they remain loosely associated with atoms. With-i n the plasma region of. an arc, electrons and ions e x i s t and move as separate e n t i t i e s but very l i t t l e i s known about the t r a n s i t i o n from one type of conduction to the other at the arc terminals. In the case 25 of arcing during e l e c t r o s l a g melting the conduction sequence i n going from the metal to the slag i s e l e c t r o n i c •> e l e c t r o n i c + i o n i c •> i o n i c . At a m e t a l l i c arc cathode, thermionic emission i s most often proposed as the primary mechanism of electron l i b e r a t i o n , but there i s doubt as to the source of energy needed to sustain emission. Also, the e f f e c t of ion current i n reducing requirements f o r e l e c t r o n i c emission must be recognized. The anodes of high-current m e t a l l i c arcs appear to function p r i -marily as c o l l e c t o r s of electrons and secondarily as contributors of vapours and ions to the plasma. The p r i n c i p a l means of anode heating i s e l ectron bombardment. Hence, i t i s generally concluded that the current at the anode i s almost e n t i r e l y e l e c t r o n i c . With respect to a metal/slag arc, one can propose that current transfer through the arc plasma takes place by electron plus i o n i c con-duction. There w i l l be Faradaic r e a c t i o n products deposited at both the metal/plasma i n t e r f a c e and the plasma/slag i n t e r f a c e . Electrochemical reactions are therefore possible, but the most important aspect of arc phenomena are those of heat generation i n the arc. During e l e c t r o s l a g melting, an arc could be i n i t i a t e d at the electrode i n two ways. 1. By evolution of a Faradaic r e a c t i o n product which i s gaseous or has a very high vapour pressure (e.g., Ca 1, 0 ). This product xrould f a c i l i t a t e electron emission at the i o n i c l i q u i d surface. An arc so formed would be s e l f - s u s t a i n i n g by v i r t u e of continued production of the materials. 26 2. By passage of s u f f i c i e n t current at the electrode to heat the slag at this point to i t s b o i l i n g point and vapourize i t . Both p o s s i b i l i t i e s w i l l be considered i n the discussion, but, at this point, i t i s s u f f i c i e n t to say that the most important aspect of arc discharge w i l l be the high rate of heat production at the electrode. This should produce anomalously high electrode consumption rates. Chemical e f f e c t s would be possible but would be secondary i n importance to the thermal e f f e c t s . 1.4 Conclusions The conclusions of t h i s chapter are as follows: 1. Chemical and thermal e f f e c t s do e x i s t using D.C. e l e c t r o s l a g melting. Chemical e f f e c t s r e s u l t i n o x i -dation loss of important a l l o y i n g elements from the metal, the magnitude of which can be s u f f i c i e n t to a l t e r ingot mechanical properties. Thermal e f f e c t s are apparently manifested as differences i n the s p e c i f i c melt r a t e . 2. E l e c t r o n i c conduction i n the slag cannot account f o r the observed chemical or thermal e f f e c t s . 3. Ionic conduction i n the slag can account for the reported observations, through Faradaic reactions and t h e i r consequent associated p o l a r i z a t i o n e f f e c t s . 27 4. Arc discharge at a melting electrode i s possible and i s important i n regard to thermal e f f e c t s . 5. D e f i n i t i v e small scale and large scale studies are j u s t i f i e d i n order to e s t a b l i s h the chemical and thermal e f f e c t s , and to propose s u i t a b l e electrochemical mechanisms. 1.5 Experimental programme out l i n e I t i s now necessary to decide what experimental r e s u l t s are needed to s a t i s f a c t o r i l y explain the chemical and thermal phenomena observed during D.C. e l e c t r o s l a g remelting. In view of the contradic-tory nature of these phenomena as reported by other workers i n the f i e l d , the following questions must be asked. A. Are the reported e f f e c t s r e a l i n experiments i n which the following parameters are controlled? - atmosphere - process voltage and amperage - a l l o y composition - slag composition B. What i s the magnitude of these eff e c t s ? C. What i s the nature of the electrochemical reactions at l i q u i d metal/calcium-fluoride based slag interfaces? 28 The experimental programme designed to answer these questions consisted of two areas of work, and these are: A. The i n v e s t i g a t i o n of metal/slag i n t e r -faces p o l a r i z e d at r e a l i s t i c current d e n s i t i e s . B. The production of ESR ingot from pure metals and a l l o y s using various modes of power under controlled atmosphere and slag chemistry conditions. CHAPTER 2 SMALL SCALE STUDIES 2.1 Experimental considerations As was outlined i n the preceeding section, one experimental ob-j e c t i v e was the study of electrochemical phenomena occurring on metal electrodes i n CaF 2 slags. As w i l l be shown l a t e r , for experimental rea-sons, c y l i n d r i c a l electrodes were used i n these studies and i t was there-fore necessary to consider the e f f e c t of c y l i n d r i c a l geometry on d i f f u s i o n -a l process occurring i n the s l a g . The d i f f u s i o n problem can be treated i n terms of two v a r i a b l e s , the distance r from the cylinder axis and time elapsed from the beginning of e l e c t r o l y s i s . The f l u x through a cylinder of radius (r + dr) : q(r,t) + [ 8 q ( r , t ) / 9 r ] • dr (2-1) The change i n concentration of the substance being reduced at the electrode i s , i n a s h e l l of thickness dr per unit time: 3 C (r,t) 2TT (r + dr)q(r + dr,t) - 2-rrr q(r,t) ° (2-2) 8 t 2 i r r d r The cylinder surface area i s 2iTr and the cylinder height i s one unit of length. The f l u x of the d i f f u s i n g substance through the surface of a c y l i n -der of radius r i s proportional to the concentration gradient. Thus 3 C ( r , t ) ^ > t ) = J ) o H f e — ( 2 " 3 ) 30 By combining (2-2) and (2-3) one obtains: 8C ( r , t ) 9t = D r 3 2 C Q ( r , t ) ]_ 3C Q ( r , t ) ~2 + r 9lr (2-4) Equation (2-4) w i l l be solved f o r the case i n which the substance i s reduced at the electrode surface as soon as i t reaches i t . The boundary and i n i t i a l conditions are as follows: C (r ,t) = 0 o o where r ^ i s the cyl i n d e r radius. C (r ,o) = C° o o C Q ( r , t ) = C° f o r r °° Integration of equation (2-4) i s rather complicated and only the expansion f o r the current density w i l l be presented i n terms of expanded Bessel functions: i = nFAD C° — o r + 1 _ A A * + 1 2 4 TT 8 (2-5) where <f> represents the dimensionless group D Q t / r Q . When <j> i s s u f f i c i e n t -l y small, a l l but the f i r s t term i n the bracket i n equation (2-5) vanish, and the current i s the same as for l i n e a r d i f f u s i o n . This w i l l be so when the quantity 1/ ( I T 2<j> 2).is smaller than 0.01. The electrodes used i n the small scale studies had r ^ = 0.363 cm. and f o r an average d i f f u s i o n c o e f f i c i e n t -4 2 - 1 X) =10 cm. sec. i n CaF„ slags, a maximum value of t can be calculated o L below which the current w i l l be the same as for l i n e a r d i f f u s i o n (<J> = 0.01). D t o 31 so t = 10 -2 x 1.32 x 10 -1 so t 13.2 seconds. Therefore, i f e l e c t r o l y s i s times l e s s than ~10 seconds are used with the electrodes considered, conditions of l i n e a r d i f f u s i o n can be applied to subsequent analysis. Galvanostatic pulsing. The technique of galvanostatic pulsing was used i n the small scale electrochemical studies rather than a potentio-s t a t i c technique because of the design aspects of the electrochemical c e l l . The reference electrode was well removed from the working electrode and the p o t e n t i a l difference produced by the c e l l resistance when current flowed was s i g n i f i c a n t , and occasionally as large as the measured overvoltage. It was therefore necessary that, for a given slag composition at a given current density, the iR product remain constant which i s the case with a galvano-s t a t i c technique. When using a p o t e n t i o s t a t i c technique, the iR product would decrease continuously as the working electrode p o l a r i z e d and r e l i a b l e overvoltage values would be d i f f i c u l t to obtain. 32 2.2 Small scale p o l a r i z a t i o n apparatus The apparatus used i n these experiments evolved through a ser-i e s of designs i n which many problems were encountered. The aim of the experiments was to study metal/slag interfaces at high temperatures and high current d e n s i t i e s . The slags of i n t e r e s t are very corrosive to most high temperature s t r u c t u r a l materials. The c r u c i b l e s were machined from graphite, and l i n e d with molybdenum f o i l to avoid carbon contamina-t i o n of the s l a g . A disc of molybdenum, 0.25 i n . thick, was placed i n the bottom of the c r u c i b l e to prevent molten metal which had f a l l e n from the electrode corroding the graphite c r u c i b l e and therefore exposing the slag to graphite. Since there i s no material known that w i l l hold l i q u i d i r o n and calcium f l u o r i d e slags isothermally, i t was necessary to hang the working electrode from above so that i t was i n contact only with the slag. The e n t i r e assembly was enclosed i n a s i l i c a tube thermally i n -sulated with carbon f e l t and purged with argon. Induction heating was used i n s p i t e of i t s drawbacks i n giving r i s e to e l e c t r i c a l noise. I t was attempted at a l l times to keep the temperature of the slag below the melting point of the electrode material to ensure that the electrode t i p did not change shape and therefore surface area. It was also necessary to use only slags forming a primary phase with a lower melting point than the electrode m a t e r i a l . This r e s t r i c t i o n arose through the formation of slag skins on the electrode t i p which could not be removed without completely melting the electrode t i p . We have therefore had to con-f i n e a l l our studies to slags with CaF 9 as the primary phase. The 33 graphite c r u c i b l e acted as the counter electrode and was supported on a graphite pedastal mounted on the water cooled bottom plate of the furnace assembly. The top plate on the s i l i c a tube was water cooled and i t served to suspend the working electrode, the slag thermocouple and the reference electrode. The reference electrode was a short graphite rod encased i n a closed end boron-nitride sleeve and supported by a s i n g l e bore alumina tube. A small hole was d r i l l e d i n the side of the boron-nitride tube to expose an area of the graphite to the slag to act as the reference electrode. A molybdenum wire was used as the e l e c t r i c a l contact to the graphite. Before i n i t i a t i n g the galvanostatic pulse, the reference electrode was anodically polarized and the reference p o t e n t i a l was established by the reaction: C + 0 2 - •+ CO + 2e _ (2-6) gr g Reference electrodes of s i m i l a r design have been used i n c r y o l i t e systems, and were found to behave r e v e r s i b l y (15, 15a). I d e a l l y a reference electrode should be i s o l a t e d from the system by an i o n i c membrane such as a boron-nitride filament carrying slag to act as a s a l t bridge, but due to the dimensional l i m i t a t i o n s of our system t h i s was not f e a s i b l e and the absolute reference electrode p o t e n t i a l therefore contains an inherent v a r i a t i o n due to the oxide-ion a c t i v i t y v a r i a t i o n s i n the slags used. However, these v a r i a t i o n s would be small (~30 mv between CaF2 + 0.5 wt. % Al^O^ and 8 wt. % A^O^) compared with the experimental error i n i n d i v i d u a l measurements. In no case did we f i n d that the p o t e n t i a l between the reference electrode and the unpolarized i r o n 34 electrode was outside the range 80 - 110 mv. F i n a l l y , the nature of the high current density experiments d i c -tates that a transient method must be used to avoid gross chemical and thermal changes i n the system. The method used was that of galvanosta-t i c pulsing, e s s e n t i a l l y as used by Gosh and King (16), and as outlined t h e o r e t i c a l l y by Delahay (17) . Since we wanted to investigate high cur-rent d e n s i t i e s we could not use the more usual methods of f a s t - r i s e time constant-current pulsing, and instead used constant-current generators (Hewlett-Packard Models 6269A and 6203B) . These have r i s e times w e l l within the p o l a r i z a t i o n times we investigated, except at very high cur-rents (~20 A), and i n these regions we used the voltage decay rather _2 than r i s e , measurements. At the higher current densities (> 5 A. cm. ), we would expect to f i n d a s i g n i f i c a n t e f f e c t of the e l e c t r i c a l f i e l d superimposed on the chemical d i f f u s i o n gradient. Hence i n these regions also we have used the voltage decay measurements. The voltage-time transients were recorded on a Tektronics storage o s c i l l o s c o p e model 564 B c a l i b r a t e d against a KiethlyModel 153 electrometric potentiometer. The apparatus i s shown i n Figures 1, 2, and 3. A l l the slags used i n these small scale studies were prepared, by mixing weighed quantities of prefused calcium f l u o r i d e ( B r i t i s h Drug House "extra pure") with the required oxide additions of alumina (Nor-ton fused alumina) or r e c r y s t a l l i z e d calcium oxide. The calcium f l u o r i d e was prefused under argon i n the experimental apparatus. The mixed dry slag was then fused i n the apparatus at a high temperature to ensure complete s o l u t i o n of the oxide. The slag temperature was then lowered to the working temperature which was always 20 to 30°C lower than the 35 1 Tektronics Model 564 B CRO Reference electrode -Working electrode Counter electrode, crucible Slog lOk^i ballast resistor Hewlett - Packard Model 6 2 6 9 A , 6 2 0 3 B Figure 1. C e l l e l e c t r i c a l c i r c u i t r y . 36 Working electrode lead Water cooled head Silica tube Carbon felt Induction coil Water cooled base Counter electrode lead 3 R e / 2 5 R e W control thermocouple Figure 2 . General c e l l design. 37 Mo wire A l 2 0 3 sheath 3 Re / 25 Re-W thermocouple Fe electrode Boron nitride Graphite Graphite Mo lifter Slag Graphite 3Re/25Re-W thermocouple Figure 3 . D e t a i l of c e l l assembly. 38 melting temperature of the electrode material being studied. When the temperature was sta b l e , the working electrode was lowered slowly to-wards the slag u n t i l i n i t i a l e l e c t r i c a l contact was made i n the current supply c i r c u i t . This p o s i t i o n was taken as a zero point and the e l e c -trode was then lowered a predetermined distance (normally 1.0 cm.) into the s l a g to present a known surface contact area between the electrode and the s l a g . The temperature of the system was measured and c o n t r o l l e d by two separate W3Re/W25Re thermocouples. At this point, the reference electrode p o t e n t i a l was measured < with respect to the working electrode, and established to be stable and within the range 80 - 110 mv. It was found that a stable reference electrode p o t e n t i a l could be achieved much more quickly by passing a low current (100 ma. f o r 10 sec.) between the two electrodes with the reference electrode being anodic. Once the system was sta b l e , the i n -duction generator was switched o f f , and the current was completed be-tween the counter electrode (crucible) and the working electrode at a pre-determined constant current s e t t i n g . The voltage r i s e between the reference and working electrodes was traced on the os c i l l o s c o p e and followed to i t s constant value, u s u a l l y established i n le s s than 10 s e c , and also followed as i t decayed when the current supply was stopped. A t y p i c a l low current test o s c i l l o s c o p e trace i s shown i n F i g . 4. This shows the gradual buildup of p o t e n t i a l d i f f e r e n c e between the two e l e c -trodes. The schematic below the o s c i l l o s c o p e p i c t u r e sh ows the method of obtaining the overpotential from the trace. The iR p o t e n t i a l drop between the two electrodes i s established very quickly and i s not traced V o l t s Time Figure 4. Method of ob t a i n i n g overvoltage. 40 on the screen. The t o t a l charge passed through the system during one such test i s small as are the bulk chemical changes so produced. Approximately 5 minutes was allowed between tests to allow the work-ing electrode to r e - e q u i l i b r a t e with the l i q u i d slag as the slag was slowly brought back up to temperature. By t e s t i n g at increased applied currents each time, the p o l a r i z a t i o n of a given electrode material i n a given slag can be p l o t t e d against the current density to give a galvan-o s t a t i c p o l a r i z a t i o n curve. It was found that the l a r g e s t part of the experimental error arose from the d i f f i c u l t y of defining a precise and constant e l e c t r o -active surface area on the working electrode. The immersion of the electrode was known but chemical reactions at the i n t e r f a c e gave r i s e to changes i n i n t e r f a c i a l tension which when combined with overheating and p a r t i a l melting of the electrode t i p r e s u l t e d i n a change i n the electrode shape. The second major source of experimental error l i e s i n the nature of the p o l a r i z a t i o n . As w i l l be shown l a t e r , t h i s i s a concentration e f f e c t , and i s therefore modified by both e l e c t r i c a l and liquid-shear f i e l d s . Any error a r i s i n g from e l e c t r i c a l e f f e c t s was eliminated by using decay measurements, but the thermal e f f e c t s produc-ing convection are unknown and cannot be s t a b i l i z e d i n t h i s type of apparatus. The degree of temperature control i n the system was con-sidered to be - 5°C and i t was decided that errors a r i s i n g from i n t r i n -s i c thermal e f f e c t s were small i n comparison to the others. 41 2.3 Electrode materials 2.3.1 Ferrovac-E pure i r o n The Ferrovac-E pure i r o n electrodes were made by machining the 1.25 i n . diameter bar down to 0.50 i n . diameter stock and then swag-ing t h i s down to 0.285 i n . i n diameter. This rod was then cut to 4 i n . lengths and a female thread machined on one end for attachment to the electrode holder. 2.3.2 AISI 430 s t a i n l e s s s t e e l This material was found to be unsuitable f o r swaging i n the as received condition, therefore these electrodes were machined d i r e c t l y from the 1.0 i n . diameter bar. 2.3.3 Fe-1.0 wt. % Cr a l l o y This a l l o y "was produced by melting a piece of Ferrovac-E together with a piece of AISI 410 s t a i n l e s s s t e e l to produce a f i n a l a l l o y of 1 wt. % Cr i n i r o n . The melt was deoxidized with aluminum wire p r i o r to s o l i d i f i c a t i o n . .. The s o l i d slug was then welded to a piece of mild s t e e l rod, 0.75 i n . i n diameter, machined down to the diameter of the rod, and then swaged down to the desired diameter of 0.285 i n . 2.3.4 Pure chromium This material was made by hot extrusion of s t e e l clad chromium metal. The purpose of the cladding was to prevent excessive oxidation 42 of the chromium during the hot working stage. A f t e r extrusion, the cladding was dissolved i n a c i d . The material had an i r r e g u l a r surface but because of i t s b r i t t l e n e s s , machining was not attempted, and i t was used i n the as received condition by s i l v e r soldering i t to a threaded mild s t e e l stub. 2.3.5 Pure n i c k e l These electrodes were made by swaging the as received 0.375" di a . n i c k e l rod down to the desired s i z e , and then cutting to the appropriate length. 2.3.6 Pure Cobalt Because of the d i f f i c u l t y i n obtaining pure cobalt i n the appro-p r i a t e s i z e range, this electrode was made from a deformed s i n g l e c r y s -t a l of cobalt which had a s l i g h t l y i r r e g u l a r surface. The s i n g l e crys-t a l was s i l v e r - s o l d e r e d into a threaded mild s t e e l stub. 2.3.7 Iron-carbon a l l o y These electrode materials were supplied by the Babcock and W i l -cox Company. They were machined to the correct s i z e when received and had only to be threaded before use. Only one experiment was c a r r i e d out with t h i s m a t e r i a l , the purpose of which was to see i f an anodic iron-carbon a l l o y would experience carbon l o s s . Carbon analyses was car r i e d out on the t i p a f t e r e l e c t r o l y s i s had been completed. 43 2.4 E l e c t r o s l a g remelting p o l a r i z a t i o n experiments It was thought, that i t would be meaningful to measure the degree of p o l a r i z a t i o n of an electrode during e l e c t r o s l a g remelting i n order to c o r r e l a t e the small scale experiments with the conditions e x i s t i n g during actual remelting. These experiments were c a r r i e d out with Armco iro n electrodes, and reference electrodes of the same design as used i n the small scale studies. As shown i n Figures 5 and 6, the reference electrode was attached to the electrode by means of i r o n wire and m u l l i t e spacers, such that i t entered the slag a f t e r steady-state melting had taken place f o r approximately 30 minutes. When the reference e l e c -trode entered the slag ( F i g . 6), the process current was switched off and the voltage decay between the reference electrode and the melting electrode was traced on the o s c i l l o s c o p e . An example of t h i s trace i s given i n F i g . 7. The voltage decay times observed were short compared to the time taken f o r the slag to s o l i d i f y i n the electrode region, which i s approximately one minute. I t was e s s e n t i a l that the power be interrupted immediately a f t e r the reference electrode entered the slag. I f this was not the case the p h y s i c a l distance between the reference and melting electrodes was too great and the ohmic p o t e n t i a l drop following current i n t e r r u p t i o n exceeded the f u l l scale defection of the o s c i l l o s c o p e . This prevented any measurement of the p o l a r i z a -t i o n p o t e n t i a l . 44 i Water - cooled copper electrode Colorlith Rubber bellows Reference electrode Fe electrode Gas sea I Cu mold Water - jacket Figure 5 , ESR furnace assembly. 45 1/ A I-IB /- Fe electrode Mo wire Cu mold Mullite spacer Graphite Boron nitride Slag Liquid ingot •Figure 6, D e t a i l of reference electrode configuration, 46 J I I I I I I 1 L 0.5 sec . / div. Figure 7. P o t e n t i a l decay between the melting electrode and the reference electrode observed at current inter-ruption on an ESR anodic electrode. 47 2.5 P o l a r i z a t i o n r e s u l t s 2.5.1 Ferrovac-E pure i r o n The anodic and cathodic p o l a r i z a t i o n curves f o r Ferrovac-E pure i r o n electrodes i n calcium fluoride-alumina slags and calcium f l u o r i d e -calcium oxide slags are given i n Figures 8, 9, 10 and 11. In c e r t a i n _2 slags (5 wt. % kl^O^) at c e r t a i n current d e n s i t i e s (1 A. cm. ) anodi-c a l l y p o l a r i z e d Ferrovac-E electrodes exhibited apparent t r a n s i t i o n times as shown i n F i g . 12. This behaviour was not observed during cathodic p o l a r i z a t i o n of the same electrode m a t e r i a l . The usual electrode current density during melting i n the e l e c t r o -slag unit i s much higher than the current d e n s i t i e s normally applied during the small scale t e s t s . It was therefore thought .to be u s e f u l to carry out small scale tests at normal remelting current d e n s i t i e s . This was done for both anodic and cathodic Ferrovac-E electrodes i n a CaF 2 + 5 wt. % kl^Oy slag using a bank of 12 v o l t lead-acid storage b a t t e r i e s -2 to produce current densities up to several hundred A. cm. . The r e -s u l t s of this work are given i n F i g . 13. At these high current densi-t i e s the electrode melted to some extent during each test due to l o c a l -ized heat production at the i n t e r f a c e as well as resistance heating of the s l a g . These r e s u l t s , although quite inaccurate, appear to agree w e l l with the lower current density r e s u l t s i n the same s l a g . It was noted that, at current densities above points X i n F i g . 13, an arc was established at the electrode. This was seen as a sudden increase i n , r a d i a t i o n when the test current was passed through the c e l l , and was Figure 8 . Anodic p o l a r i z a t i o n curves for pure i r o n i n CaF 2 + A 1 2 0 3 slags. Figure 10 . Anodic p o l a r i z a t i o n curves f o r pure i r o n i n CaF + CaO slags. 51 Figure 11. Cathodic p o l a r i z a t i o n curves for pure i r o n i n CaF 9 + CaO slags. J I I I I I I I L 0.2 sec./div. F i g u r e 12. Apparent t r a n s i t i o n time i n the a n o d i c p o l a r i z a t i o n o f pure i j j o n i n CaF7 + 5 wt.% Al .0 a t i = 1 A.cm 54 often accompanied by excessive electrode melting. In t h i s small scale work on CaF^ + A^O-j s-'-aSs> n o w o r ^ w a s c a r r i e d out s u c c e s s f u l l y at A^O^ concentrations ..greater than 10 wt. %, . which i s the e u t e c t i c composition (18). This arises i n the f a c t that pure A^Og i s the primary p r e c i p i t a t e upon cooling a. hyper eu t e c t i c com-p o s i t i o n , and once a layer of A^O^ had been formed on the electrode t i p , i t could riot be removed without completely melting the electrode t i p . The large experimental s c a t t e r observed i n the "pure" CaF^ measure-ments i n F i g . 8 i s probably due to the v a r i a b l e (100 - 500 ppm) CaO con-tent of t h i s m a t e r i a l . The data points shown are the average of three separate experiments. P o l a r i z a t i o n of Ferrovac-E electrodes i n A^O^ slags r e s u l t s i n s i g n i f i c a n t concentration changes i n the electrode. F i g . 14 shows o p t i -c a l and electronmicroprobe scans of a section through an anodic electrode t i p . This i n i t i a l l y low oxygen (20 ppm) electrode was p o l a r i z e d at high current density i n a CaF^ + 5 wt. % A^O^ slag for several 10 sec. per-iods, and then removed from the s l a g . The o p t i c a l p i c ture shows the r e s u l t i n g " d i r t y " microstructure, and the probe pictures show areas of oxygen-containing ma t e r i a l , probably iron-oxide. The other h a l f of t h i s t i p was analyzed for t o t a l oxygen and found to contain 400 ppm. An equivalent Ferrovac-E electrode was c a t h o d i c a l l y p o l a r i z e d i n a CaF^ + 8 wt. % A^O^ slag for 1 hour at a current of 1 A. A s i g -n i f i c a n t portion of the t i p had melted o f f , but the remaining portion was analyzed f o r A l on the electron microprobe. The concentration of A l i n the matrix was calculated from the probe X-ray i n t e n s i t i e s using the "MAGIC" computer program, and was found to be 3.0 wt. %. In no case did we detect f l u o r i n e or calcium i n c i t h e r electrode t i p . Optical X 2 6 0 AE.I. X 6 5 0 0 X-ray X 650 F i g u r e 14. S e c t i o n s t h r o u g h a pure i r o n e l e c t r o d e t i p anodically p o l a r i z e d at h i g h c u r r e n t d e n s i t y f o r s e v e r a l 10 s e c . p e r i o d s i n a CaF^ + 5 wt.% A^O^ s l a g . 56 2.5.2 AISI 430 s t a i n l e s s s t e e l The p o l a r i z a t ion curves obtained, f o r AISI 430 s t a i n l e s s s t e e l i n CaF 2 + A^O^ slags are given i n Figures 15 and 16. In addition to these experiments, one test was c a r r i e d out f o r the purpose of studying chromium depletion of the a l l o y . An electrode was anodically p o l a r i z e d i n a CaF 2 + 8 wt. % A^O^ slag at a current of 1 A. for 915 sec. The electrode t i p was then cut v e r i t c a l l y f o r examination on the e l e c t r o n microprobe. A h o r i z o n t a l traverse was made from the outside edge towards the center and Cr counts were taken at 20 u steps. These X-ray counts were corrected to give wt. % Cr using the "MAGIC" program, and the r e s u l t s are given i n Figure 17. This f i g u r e shows a d i f f u s i o n p r o f i l e i n which the Cr concentration f a l l s from the bulk concentration of approximately 17 wt. % Cr, to a value of approximately 9 wt. % at the electrode-slag i n t e r f a c e . The purpose of t h i s experiment was to show that e a s i l y o x i -dizable a l l o y i n g elements can be s e l e c t i v e l y oxidized at an anodic i n t e r -face during anodic p o l a r i z a t i o n . The Cr loss so observed indicates a b u i l d -up of C r n + ions i n the s l a g . The current e f f i c i e n c y of Cr removal, assum-3-r ing Cr goes to Cr i n the s l a g , was calculated by assuming the concentra-t i o n p r o f i l e i n F i g . 17 i s l i n e a r between 9 wt. % Cr and 17 wt. % Cr at 300y depth. The t o t a l Cr l o s t during e l e c t r o l y s i s i s therefore approxi-mately 20 mg. which gives a current e f f i c i e n c y for chromium removal of approximately 10%. 2.5.3 Fe - 1 wt. % Cr a l l o y The anodic p o l a r i z a t i o n curves f o r an Fe + 1 wt. % Cr a l l o y i n 58 1000 - 4 0 -2 In i (A.cm. ) 0 + 4 Figure 16. Cathodic p o l a r i z a t i o n curves for AISI 430 s t a i n l e s s s t e e l i n CaF„ + A l 0 slags. 59 Distance in from surface fJL Figure 17. Cr concentration gradient produced at the surface of an AISI 430 s t a i n l e s s s t e e l electrode anodically polarized f o r 915 sec. at i = 360 ma.cm. i n a CaF 2 + 8 wt.% A1 20 3 slag. ° 60 a CaF + 1 wt. % A l 0 slag are given i n F i g . 18. The two curves drawn 2 2 3 correspond to the r i s e overpotential measured when the c e l l c i r c u i t i s closed, and the decay overpotential measured on current i n t e r r u p t i o n . 2.5.4 Pure chromium The anodic p o l a r i z a t i o n curve for a pure Cr electrode i n a CaF 2 + 1 wt. % A1 20 slag i s shown i n F i g . 19. 2.5.5 Pure n i c k e l A pure n i c k e l electrode was anodically p o l a r i z e d i n a CaF^ +. 2.5 wt. % A l ^ ^ i n order to e s t a b l i s h that i t s anodic behaviour was s i m i -l a r to that of a pure i r o n electrode i n A l ^ ^ slags. The behaviour of the n i c k e l electrode i s shown i n F i g . 20. This f i g u r e i s a p i c t u r e of the o s c i l l o s c o p e trace at a current of 3 A. A steady state plateau i s observed, which i s established quickly, and i s followed on current i n t e r -ruption by a r e l a t i v e l y slower p o t e n t i a l decay. A n i c k e l electrode was c a t h o d i c a l l y p o l a r i z e d at a current of 1.25 A. f o r 900 sec. i n a CaF^ + 8 wt. % A^O^ s l a g , and then sectioned and examined i n the electron microprobe f o r Ca and A l . The appropriate electron microprobe scans are given i n F i g . 21. These show a strong con centration of matrix A l and a s i g n i f i c a n t number of Ca containing areas corresponding to the l i g h t e r areas on the AEI. The electrode t i p gave a strong response f o r aluminum X-ray counts, and these r e s u l t s were corrected using the "MAGIC" program to give a matrix concentration of aluminum of 8.9 wt. %. However there was 61 0.2V/div. « i * I ' • » » I sec./div. Figure 20, Single p o l a r i z a t i o n curve for an anodic galvanostatic pulse on pure n i c k e l i n a CaF 2 + 2.5 wt.% A1 20 3 slag. i Q = 1.10 A.cm 64 A.E.I. X 1020 Ca X-ray X 1020 Al X-rayXI020 F i g u r e 21. S e c t i o n t h r o u g h a pure n i c k e l e l e c t r o d e c ^ t h o d i c a l l v p o l a r i z e d f o r 900 s e c . a t i = 450 ma.cm i n a C a F ? + 8 wt.% A 1 2 0 s l a g . ° 65 no response when Ca X-rays were counted i n the matrix, despite the f a c t that there were Ca r i c h areas as shown i n F i g . 21. 2.5.6 Pure cobalt A pure cobalt electrode was polarized anodically i n a CaF^ + 2.5 wt. % Al^O^ s l a g . F i g . 22 shows i t s p o l a r i z a t i o n r i s e and decay at a current of 3 A. The s i m i l a r i t y i s to be noted f o r the case of an anodic n i c k e l electrode ( F i g . 20). 2.5.7 Fe - C a l l o y An Fe + 0.83 wt. % C a l l o y was anodically p o l a r i z e d at a cur-rent of 1 A. f o r 200 sec. i n a CaF^ + 2.5 wt. % Al^O^ slag i n order to study the carbon l o s s . Approximately one-half of the electrode t i p melted o f f during t h i s time. The remaining part of the t i p was cut from i t s base and analyzed f o r t o t a l carbon. The carbon content had dropped to 0.31 wt. % C. 2.6 E l e c t r o s l a g remelting p o l a r i z a t i o n r e s u l t s The e l e c t r o s l a g p o l a r i z a t i o n r e s u l t s are given i n Figures 23->26 for Armco i r o n electrodes being remelted i n A^O^-containing and CaO-con-taining slags. Figures. 23 and 25 are anodic p o l a r i z a t i o n r e s u l t s while Figures 24 and 26 are cathodic r e s u l t s . These r e s u l t s are superimposed on the small s c a l e r e s u l t s given i n Figures 8-KL1, and although some extra-p o l a t i o n was necessary to extend the small scale r e s u l t s i n t o the current density range experienced during remelting, good agreement i s found. The 66 0.2V/div. • • 1 i * • I sec. /div. F i g u r e 22. S i n g l e p o l a r i z a t i o n c u r v e f o r an a n o d i c g a l v a n o s t a t i c p u l s e on pure c o b a l t i n a C a F 2 + 2.5 wt.% A 1 2 0 3 s l a g . i = 1.8 A.cm r 1 1 r wt% Al 0 2 3 / • Furnace Results , 0 1 . 2 3 4 5 -2 In i_ (A.cm. ) Figure 23. Anodic p o l a r i z a t i o n on Armco ir o n E.S.R. electrodes i n CaF 9 + A l 0 slags. 68 1.5 V (V.) wt% A l 2 0 3 • Furnace Results T CD / / / 1.0 0/ / / / / 0.5 0 4 In i (A.cm.2) 0 Figure 24, Cathodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n CaF2 + A^O^ slags. 69 {Figure 25. Anodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n CaF 9 + CaO slags. 70 Figure 26, Cathodic p o l a r i z a t i o n on Armco i r o n ESR electrodes i n CaF 0 + CaO slags. 71 extent of p o l a r i z a t i o n was also measured on AISI 430 s t a i n l e s s s t e e l electrodes and AISI 1095 s t e e l electrodes. The r e s u l t s are too incon-c l u s i v e to present i n graphical form, but w i l l be discussed i n Chapter 5. CHAPTER 3 MELT PROGRAM 3.1 Melting procedure A l l the ingots used i n this i n v e s t i g a t i o n were made on the U.B.C. E l e c t r o s l a g Unit, the design and operation of which have been described by Etienne (4). Each run was started using a compact consisting of metal turnings and calcium-fluoride power, the metal turnings being machined from the electrode material. When the slag was completely molten and the melting conditions were stable, the operators proceeded to record the pertinent process parameters at regular time i n t e r v a l s . In t h i s manner, a complete record of each ingot was taken and used f o r subsequent analysis. 3.2 Electrode materials The electrode materials used i n this program of experiments can be c l a s s i f i e d into two types, the f i r s t type consisting of e s s e n t i a l l y pure metals and the second being iron-base a l l o y s containing other poten-t i a l l y o xidizable elements. The reasons f o r using each material are given together with the analysis of the electrode materials i n the as received condition. A l l compositions are given i n weight per cent. 72 7 3 3 . 2 . 1 Low carbon mild s t e e l : AISI 1 0 1 8 grade (Supplied by Stelco) r KT P S 0 Fe C Mn max max . 1 5 - . 2 0 . 6 0 - . 9 0 . 0 4 0 . 0 5 0 . 0 1 5 5 Bal. This electrode m a t e r i a l , because of i t s low cost, was used i n the i n i t i a l experiments where the melting conditions were uncertain. 3 . 2 . 2 Ferrovac-E: vacuum melted (Supplied by the Crucible S t e e l Company, S o r e l , Quebec) c Mn P S S i . 0 1 0 . 0 0 1 . 0 0 2 . 0 0 4 . 0 0 6 Cr Mo N 0 H < . 0 1 . 0 0 1 . 0 0 0 2 . 0 0 0 9 2 . 0 0 0 0 1 8 Co Cu V W Fe . 0 0 6 . 0 0 6 < . 0 0 2 . 0 2 Bal Ferrovac-E pure i r o n was the i d e a l electrode material on which to study electrochemical phenomena during ESR. Because of i t s very low oxygen content, any oxygen found i n the ingots must have a r i s e n d i r e c t l y as a r e s u l t of remelting the metal whether by chemical or electrochemical means. A l a t e r s e r i e s of ingots were made using a second batch of Ferrovac-E which was excessively high i n oxygen ( 3 1 6 ppm). Apart from t h i s , i t was e s s e n t i a l l y pure i r o n as was the previous batch of Ferrovac-E. Since i t had been found that t h i s high i n i t i a l oxygen content would 74 be l o s t during remelting, and would have very l i t t l e e f f e c t on the f i n a l ingot oxygen content, t h i s material was used i n the same way as was the i n i t i a l batch of Ferrovac-E. Two ingots were made using t h i s second batch of Ferrovac-E with lengths of aluminum wire held i n shallow grooves machined on the electrode surface. The purpose of thi s was to inves t i g a t e whether or not excess aluminum would control the ingot oxygen content i n sp i t e of electrochemical oxidation. The amount of wire added would produce 2000 ppm of A l i n the f i n a l ingot i f none were to be l o s t during melting. 3.2.3 Armco i r o n : (Supplied by Armco Steel Corporation) C Mn P S S i o Fe .012 .017 .005 .025 trace .070 B a l . Armco iron'was used i n place of Ferrovac-E pure i r o n for several runs as i t was found that despite the high oxygen content of the Armco i r o n , ingots made from i t had oxygen contents very close to the 0 contents of Ferrovac-E ingots made under the same conditions. This re s u l t e d i n s u b s t a n t i a l savings i n electrode materials. 3.2.4 F e r r i t i c s t a i n l e s s s t e e l : AISI 430 grade (Supplied by Allegheny Ludlum Industries, Inc.) C Mn P S S i Cr 0 .06 .44- .024 .015 .26 17.35 .0115 It was desired to make ingots from an i r o n a l l o y containing a second major element which could be oxidized during remelting. AISI 430 s t a i n l e s s s t e e l was chosen because i t i s e s s e n t i a l l y an iron-chromium a l l o y containing l i t t l e else which could i n t e r f e r e with Cr oxidative losses. Another advantage of this a l l o y was that i t s melting point i s very close to that of pure i r o n . 3.2.5 Medium carbon mild s t e e l : AISI 1095 grade (Supplied by Stelco, Hamilton, Ontario) C Mn S i P S 0 Fe max max .975 .39 .34 .040 .050 .0020 B a l . Ingots were made from this electrode material • to study the e f f e c t of a sub s t a n t i a l amount of carbon on the f i n a l oxygen content of these ingots. 3.2.6 Pure N i : Supplied by Falconbridge N i c k e l Mines C v S i Mg Fe 0 .01 .17 .18 1.9 .0011 Ingots were made from Falconbridge Nickel to investigate the behaviour of pure n i c k e l during ESR and to compare th i s behaviour to the melting behaviour of pure i r o n . 76 3.3 Slag materials The slags used during this melting program were a l l based on calcium-fluoride with additions of aluminum-oxide (kl^O^) or calcium-oxide (CaO) to a l t e r the melting point, the conductivity, and the a c t i -v i t i e s of the various components. 3.3.1 Calcium f l u o r i d e The calcium-fluoride used i n these slags was a f i n e power pro-duced by a h y d r o - f l u o r i n a t i o n process. Approximately two-thirds of the calcium-fluoride used i n each run was pre-fused i n graphite c r u c i b l e s by induction heating under an argon cover. The slag was held as a l i q u i d f o r only a very short time i n the graphite c r u c i b l e to minimize carbon pick-up from the c r u c i b l e . A f t e r fusion, the cold slag was crushed into small p a r t i c l e s (approx. 1/8" i n diameter) before being mixed with the other slag components. It was found, from experience that i f these calcium-fluoride granules were too large they could bind between the electrode and the mould w a l l thus stopping electrode t r a v e l and causing the current path to be broken during the start-up period of a melt. If the slag was too f i n e , s i n t e r i n g of the s o l i d p a r t i c l e s would occur, again preventing electrode t r a v e l during the start-up. Another reason f o r fusing part of the c a l -cium-fluoride was to d r i v e o f f the surface moisture on the f i n e powder. Moisture i n the slag can a l t e r the slag composition according to the reaction: 77 CaF + H O -> CaO + 2 HF (3-1) to form calcium-oxide i n the s l a g . Calcium-oxide i s the main sl a g im-purity and i s usually found to be at a l e v e l of 500 ppm although i t would vary from melt to melt. This impurity was only important when the slag used during remelting was kept low i n oxide content. For higher oxide content slags (5 wt. % A^O^ or more) the calcium-oxide impurity con-tent i s unimportant i n comparison to the t o t a l oxide content of the sl a g . When i t was desired to make ingots i n low oxide content sl a g s , BDH extra pure calcium-fluoride was prefused i n graphite c r u c i b l e s and used f o r the enti r e calcium-fluoride content of the sl a g . The as r e -ceived analysis of this calcium-fluoride i s given below as maximum im- , purity l i m i t s i n wt. %. Chloride (CI) .005 % Sulphate (SO^) .01 % Iron (Fe) .005 % Lead (Pb) .005 % S i l i c a ( S i 0 2 ) .05 % 3.3.2 Alumina Norton granular alundum (A^O^) of 99.3% p u r i t y was used i n the as received condition for making alumina containing slags. This material i s e l e c t r i c a l l y fused from Bayer process alumina. 3.3.3 Calcium oxide R e c r y s t a l l i z e d calcium oxide (CaO) of 99.5% p u r i t y was used to make calcium-oxide containing slags. This material was supplied by Dyna-mit Nobel, Germany. 78 3.4 Atmospheric control Ingots were made under two types of atmospheric co n t r o l equip-ment. The i n i t i a l 24 ingots, as w e l l as some l a t e r ones were made using a crude type of argon fume hood as shown i n F i g . 27. This fume hood directed an argon flow down towards the slag/atmosphere surface and c o l l e c t e d most of the fumes given off during remelting. A more sophisticated design was used to make the balance of the ingots. As shown i n F i g . 28, i t consisted of a number of rubber bellows, clamped at the lower end to a flanged section of pipe and f i x e d above to a water-cooled copper electrode c a r r i e r , such that the downward t r a v e l of the electrode caused the bellows to collapse. The flanged pipe, which was held to the top of the ingot mould by C-clamps, had two aluminum f o i l blow-out windows. This design provides an atmos-phere the q u a l i t y of which depends almost e x c l u s i v e l y upon the p u r i t y of the argon used. By contrast, chromatographic analysis of the argon fume hood atmosphere indicates that i t may contain up to 1% 0^ with an argon flow rate of 100 1/hr. 3.5 Melting Conditions Most of the ingots produced i n this melt program were made under stable melting conditions which i n turn were determined by such factors as slag composition, electrode p o l a r i t y , and electrode materials. In general, i t was found that melting using e i t h e r A.C. or e l e c -trode negative was quite stable but melting with electrode p o s i t i v e at the same voltage and current as used i n electrode negative was very un-79 Figure 27, Argon fume hood. Figure 28. Argon gas cap, 81 stable, i f not impossible. This i n s t a b i l i t y was the r e s u l t of arcing between the r a d i a l slag segment around the electrode and the mould w a l l and occurred even though the mould was insulated from the base plate (19). The problem of arcing was overcome by painting the inside of the Cu mould with Boron-Nitride paint (type S) which i s an e l e c t r i c a l i n s u l a -tor but has a high thermal conductivity. Using this technique, i t was possible to make ingots using the electrode p o s i t i v e mode at higher operating voltages while s t i l l maintaining the desired electrode-slag-ingot current path. In a number of cases, ingots were made using a l i v e mould. This was done by connecting separate leads d i r e c t l y to the bottom of the mould putting i t i n p a r a l l e l with the base plate thereby allowing the process current to flow through either the mould or the ingot/base plate on the return path. 3.6 Melt records During the melting of an ingot, d e t a i l e d accounts were kept of a l l the important operating parameters. A Sargent Model SR M i l l i v o l t Recorder was used to keep a continuous record of the operating current. When the electrode was melting i n a stable fashion (the slag was complete-l y molten) the following data were recorded i n the log book at 100 sec. i n t e r v a l s : V - process voltage -> D.C. or A.C. A - process current D.C. or A.C. t - time i n seconds with t = 0 the beginning of recording the data 82 -2 C - t o t a l coulombs passed x 2 x 10 P - t o t a l electrode t r a v e l i n mm. from the beginning of the run M.S. - speed of electrode t r a v e l drive motor AT - temperature d i f f e r e n c e i n °C between the i n l e t and o u t l e t temperature of the mould cooling water. Also recorded f or each ingot were any operating p e c u l i a r i t i e s evident during melting. 3.7 Melt record c a l c u l a t i o n s 3.7.1 Melt rate The average melt rate of each ingot was calculated using the following mathematical approach. The s t a r t i n g point f or the c a l c u l a t i o n i s the actual distance traveled by the electrode, this distance being recorded at regular time i n t e r v a l s during a run. The diameter of the electrode i s known (D e) and the average diameter of the ingots (D^) i s measured. To convert the electrode t r a v e l into ingot r i s e as a function of ingot diameter, i t i s assumed that the electrode ingot separation r e -mains constant during a run and that the ingot i s f u l l y dense. The amount of ingot r i s e r e s u l t i n g from 1 mm. of electrode t r a v e l i s given by: f The weight of ingot formed per mm. of electrode t r a v e l i s given by X = [1 + C^ )2] * K^ )2] *?^ £V n _(3-2) Y = X • TrRjVm (3-3) 83 where p m i s the density of the metal. The melt rate i s then calculated according to: M.R. = C ^ - ) (3-4) t sec where P i s the electrode t r a v e l i n mm. and t the time i n seconds over which the t r a v e l occurred. 3.7.2 S p e c i f i c coulombic density (Z) A parameter was required r e f l e c t i n g the rate at which the electro-chemical r e a c t i o n products entered the ingot metal during melting. This parameter could then be compared to the measured f i n a l oxygen content of the ingot. Such a parameter must take into account three terms; the cur-rent at the surface i n question, the area of t h i s surface, and the melt rate (which i s a crude measure of the rate at which this surface i s re-ceiving fresh metal). An appropriate area was assumed to be the area of the ingot top, which was also considered to be f l a t . The rate of forma-t i o n of electrochemical products would be proportional to the process , current (assuming 100% current e f f i c i e n c y of any Faradaic processes) and inversely proportional to the surface area as w e l l as the melt r a t e . This parameter Z could then be calculated as follows: z _ Process current (A.) _^ Ingot top area (cm.2) x melt rate (gm.sec ) coul. sec. 2 -1 cm. x gm. sec Z = coul. 2 cm. x gm. 84 3.7.3 Drop s i z e and surface tension It i s known that oxygen i s surface a c t i v e i n l i q u i d i r o n as are many other elements (B, Se, S) and that the e f f e c t of oxygen i n i r o n i s to reduce i t s surface tension quite markedly. I t was therefore thought to be desirable to measure the s i z e of drops produced at the melting electrode during e l e c t r o s l a g melting and to corr e l a t e the drop s i z e to the surface tension of the drop, hence u l t i m a t e l y to the presence or absence of oxygen i n the drops. Campbell (20) studied droplet formation i n e l e c t r o s l a g remelt-ing and derived the r e l a t i o n given below using a dimensional argument. r 2 = V (3-5) gAp r - drop radius Y - i n t e r f a c i a l tension between two l i q u i d s g - ac c e l e r a t i o n due to gra v i t y Ap - density d i f f . between l i q u i d s at T k - a constant He assumed that the drop radius, r , i s independent of the e l e c -trode radius f o r a s u f f i c i e n t l y large electrode and he states the r e l a t i o n -ship i s v a l i d f o r c o n i c a l electrode t i p s , such as i s the case during the melt program. In order to use t h i s r e l a t i o n s h i p i t i s f i r s t necessary to cal c u l a t e r from the melt records. This i s done by counting the drop rate on the process current recorder chart, each drop coinciding with a peak on the chart. Combining the drop weight with the melt rate, one can f i n d the drop s i z e , then r , and ca l c u l a t e y using the appropriate ph y s i c a l constants. This was done for Ferrovac-E electrodes melted under various conditions. 85 3.7.4 Melt program r e s u l t s The melt record data, the calculated parameters (melt rate and the s p e c i f i c coulombic density [Z]), and the average oxygen content of the ingots are summarized i n Tables I->V. The drop s i z e and i n t e r f a c i a l tension r e s u l t s are given i n Table VI. 3.8 Ingot analysis 3.8.1 Oxygen analysis of ingots A l l of the ingots produced i n t h i s melt program were analyzed for t h e i r average t o t a l oxygen content using a Lecd Rapid Oxygen Analy-ser (No. 734-300). I t must be noted that the oxygen i n these ingots i s i n the form of non-metallic inclusions formed by the r e a c t i o n of dissolved oxygen with either metals dissolved i n the base metal or the base metal i t s e l f . Ingots to be analysed f o r t o t a l oxygen are sectioned as shown i n F i g . 29 giving i n d i v i d u a l 1 g. specimens. An ingot i s cut horizon-t a l l y one-third of the way up from the bottom, and then a v e r t i c a l s l i c e one-quarter of an inch thick i s cut from the top portion. This s l i c e i s then cut h o r i z o n t a l l y into 10 1/4" square s l i c e s and the desired s l i c e can then be cut i n t o 1/4" cubes for a n a l y s i s . Normally, f i v e cubes would be cut from the outside into the center from the second, f i f t h , and tenth s l i c e s to give 1 5 specimens for analysis from, each ingot. Portions of; these s l i c e s were .also used for metallographic and electron microprobe examination of the inclusions i n the ingots. Ingot Top 7j8_ij|LQ - f - r t - T - T - ^ - • 4 - l t-I H m - i - r • | r t - r 4 - M ^ - -10 I J. _ ure 29. Ingot sampling scheme. 87 3.8.2 Aluminum analysis of FVE ingots Five Ferrovac-E ingots were analyzed for t o t a l aluminum using the neutron a c t i v a t i o n method (21). The r e s u l t s of these analyses are given i n Table VII. 3.8.3 Analysis of AISI 1095 s t e e l Electrode material and ingots made from 1095 s t e e l were analysed spectrographically to determine what a l l o y losses took place during r e -melting. The r e s u l t s are given i n Table VIII. 3.8.4 Analysis of AISI 430 s t a i n l e s s s t e e l Electrode material and ingots made from 430 s t a i n l e s s s t e e l were also analysed spectrographically. The r e s u l t s are given i n Table IX. 3.9 Slag cap analysis Several of the slag caps were observed to contain dark green or brown phases, u s u a l l y near the ingot region. These regions were thought to contain a high concentration of iron-oxide. Segments of four slag caps from FVE runs (-, +, +(im) , A.C.) were crushed and analysed f o r t o t a l i r o n by caustic fusion and acid d i s s o l u t i o n to make solutions s u i t a b l e for analysis by the atomic absorption technique. The r e s u l t s are given i n Table X. 88 TABLE I MELT RECORD RESULTS OF AISI 1018 MILD STEEL Atmosphere - Argon fume hood Electrode Dla. = 3.82 cm. Ingot Dia.(av.) = 7.5 cm. Electrode 0 content = 150 ppm. CA calcium aluminate INGOT SLAG ADD'N. ELECTRODE NO. POLARITY VOLTS AMPS MELT RATF. OXYGEN Z _ ^ (wt. %) (gm.sec. ) (ppm) (coul.cm. gm. ) 4 30 CA — 24.6 1160 2.94 481 8.9 5 " 30 CA — 23.8 1200 2.40 500 11.3 11 30 CA . — 23.5 1177 2.97 500 8.9 17 30 CA — 23.2 1025 2.21 475 10.5 18 30 CA — 23.2 1350 3.08 375 9.9 (mo 15 30 CA —- . . vn.a. 0 2000 n.a. i n s e r t ; 14 n i l — 22.2 1181 2.50 205 10.7 6 10 MgF 2 . — 20.5 1150 510 7 10 MgF 2 21.0 1100 505 1 30 CA + 22.1 1020 . 2.45 55 9.4 2 30 CA + 22.5 1015 2.86 50 8.0 3 30 CA ' + 23.Q 1058 3.13 50 7.6 21 30 CA A.C. t r i a l runs 180 22 30 CA A.C. 200 89 TABLE II MELT RECORD RESULTS OF FVE INGOTS Atmosphere - Argon fume hood Electrode Dia. = 3.18 cm. Ingot Dia.(av.) = 5.5 cm. Electrode 0 content = 9 ppm. * •> new FVE 0 content = 316 ppm. CA -»• calcium aluminate INGOT SLAG ADD'N. ELECTRODE NO. POLARITY VOLTS AMPS MELT RATE OXYGEN Z -2 1 (wt. %) (gm.sec ) (ppm) (coul.cm. gm. ) 8 24 CA - 23.5 689 1.28 1075 21.8 9 24 CA - 23.2 862 1.63 788 21.4 10 24 CA . - 23.5 1047 2.06 730 21.4 16 24 CA + 20.5 782 1.32 250 23.2 23 24 A 1 2 ° 3 - 23.3 966 2.33 425 16.7 * 72 25 A 1 2 ° 3 - 22.7 892 1.49 569 22.7 24 24 A 1 2 ° 3 + 20.2 749 1.59 150. 19.0 * 74 25 A 1 2 ° 3 + (im) 22.5 878 1.68 174 17.9 * 73 25 A 1 2 ° 3 A.C. 23.3 558 2.99 165 8.7 19 25 CaO - 22.4 1000 2.20 800 17.8 20 30 CaO - 22.5 1040 1.58 550, 26.6 90 TABLE I I I MELT RECORD RESULTS OF FVE INGOTS Atmosphere - Argon gas cap. Electrode Dia. = 3.18 cm. Ingot Dia.(av.) = 5.5 cm. Electrode 0 content = 9 ppm. * ->• new FVE 0 content = 316 ppm. INGOT ELECTRODE NO. SLAG ADD'N. POLARITY VOLTS AMPS MELT RATE (gm.sec ) OXYGEN Z _2 (wt . %) (coul.cm. 25 25 A 1 2 ° 3 - 22.6 833 1.38 450 27.3 26 25 A 1 2 ° 3 - 22.5 949 1.75 450 22.8 27 25 A 1 2 ° 3 - 23.3 924 1.77 470 22.8 28 25 A 1 2 ° 3 - 22.2 750 1.28 450 26.6 29 25 A 1 2 ° 3 - 24.5 859 1.64 500 23.3 30 25 A 1 2 ° 3 - 24.0 748 1.43 680 22.0 39 25 A 1 2 ° 3 - 23.5 991 1.83 460 23.6 77 25 A 1 2 ° 3 - 26.0 1160 3.55 378 13.7 31 25 A 1 2 ° 3 + 19.0 652 1.30 67 21.2 32 25 A 1 2 ° 3 + 19.1 655 1.11 317 23.9 33 25 A 1 2 ° 3 + 18.9 748 1.28 171 22.9 37 25 A1 20 3 4(±m) 22.0 791 1.77 175 18.1 38 25 A l 2 ° 3 + (im) 22.4 917 1.58 70 24.4 41 25 A 1 2 ° 3 + (im) 22.2 957 2.50 170 16.. 7 35 A 25 A 1 2 ° 3 + (lm) 21.5 857 1.15 690 32.7 83 25 A 1 2 ° 3 + (lm) 22.5 933 1.76 834 22.3 34 25 A1 20 3 A.C. 24.2 655 2.22 225 12.9 91 TABLE I I I (Continued) INGOT ELECTRODE NO. SLAG ADD'N. (wt. %) POLARITY VOLTE 1 AMPS MELT RA^  (gm.sec OXYGEN (ppm) (coul.cm. gm 40 25 M 2 ° 3 A.C. (im) 22.9 819 2.62 200 13.1 36 25 A1 20 3 A.C. (lm) 23.8 810 2.15 225 15.9 * 79 25 CaO - 22.6 1065 2.06 342 20.8 * 78 25 CaO + (im) 21.4 828 2.22 119 15.0 * 82 25 CaO + (lm) 21.5 908 2.11 230 18.1 * 75 25 A1 20 3 - 23.8 1150 3.07 548 8.7 (7.6 large mold * 76 25 A 1 2 ° 3 . +(im) 24.5 1045 2.31 153 10.2 o 3 large mold * 80 25 A1 20 3 - 23.8 779 2.59 548 12.7 K H-W > 81 25 A1 20 3 + (im) 22.0 792 1.75 526 22.1 n 92 TABLE IV MELT RESULTS OF ARMCO IRON INGOTS Atmosphere - Argon gas cap Electrode Dia. = 3.18 cm. Ingot Dia. (av.) = 5.5 cm. Electrode 0 content = 700 ppm. INGOT SLAG ADD'N. ELECTRODE VOLTS AMPS MELT RATE OXYGEN Z NO. POLARITY _ 1 _ 2 (wt. %) (gm.sec. ) (ppm) (coul.cm. gm. ) 42 25 A 1 2 0 3 - 22.4 908 1.49 550 26.5 49 25 A 1 2 0 3 - 22.6 954 2.14 470 18.7 47 10 A1 20 3 - 23.0 886 1.70 820 21.1 65 5 A 1 2 ° 3 22.3 896 2.37 820 15.3 48 1 A l O s - • 22.5 910 1.77 810 22.6 50 n i l - 20.0 536 1.79 660 16.3 54 25 A 1 2 ° 3 + 20.0 732 1.72 300 17.9 57 . 25 A 1 2 ° 3 + 18.5 757 1.80 195 17.0 58 25 A 1 2 ° 3 + 19.0 768 1.80 ; 290 17.3 55 10 A 1 2 ° 3 + 18.5 808 1.53 250 22.1 56 10 A1 20 3 + 18.8 786 1.53 260 21.5 59 5 A 1 2 ° 3 + 19.0 827 1.58 285 21.3 44 25 A1 20 3 + (im) 22.2 899 1.82 255 20.8 43 25 A 1 2 ° 3 A.C. 23.0 710 2.94 260 10.9 60 25 A 1 2 0 3 A.C. 26.4 588 3.06 245 8.1 66 30 CaO - 23.0 973 2.10 505 19.1 52 25 CaO - 23.0 1062 1.91 580 23.4 51 5 CaO — 20.0 1104 0.98 1200 45.7 TABLE V MELT RECORD RESULTS OF MISCELLANEOUS INGOTS Atmosphere - Argon gas cap. Ingot Dia.(av.) = 5.5 cm. MATERIAL ELECTRODE DIA.(cm.) 0 CONTENT (ppm) AISI 430 2.54 115 AISI 1095 2.54 20 Pure Nickel 3.50 11 INGOT ELECTRODE NO. SLAG'ADD'N. POLARITY VOLTS AMPS MELT RAT^ OXYGEN z _ 2 _ x (wt. %) (gm.sec. ) (ppm) (coul.cm. gm. ) 61 25 A1 20 3 22.7 834 1.75 125 20.0 62 25 A 1 2 0 3 - 22.7 877 1.85 115 19.8 63 25 A 1 2 0 3 - 23.1 858 1.79 225 20.1 64 12 A 1 2 0 3 - 23.0 843 1.95 170 18.2 70 25 A1 20 3 + (im) very unstable 91 68 25 A 1 2 ° 3 - 24.0 910 2.15 145 17.8 67 25 A1 20 3 + (im) 22.0 790 2.45 50 13.6 69 10 A 1 2 ° 3 - 23.0 780 1.96 7 15.6 71 10 A1 20 3 + (im) 22.0 628 2.23 362 11.4 TABLE VI DROP SIZE AND INTERFACIAL TENSION RESULTS FOR FVE ELECTRODE POLARITY AND CONDITIONS DROP WT. (gm) y (DYNES cm."2) 2.57 421 + (includes +im 1.11 240 and +lm) A.C. 3.51 517 - (Ingot 80- A l wire) 3.46 513 + (Ingot 81- A l wire) 1.17 250 TABLE VII TOTAL ALUMINUM CONTENT OF FVE INGOTS INGOT NO. MELT CONDITIONS ppm. A l 25 - 522 41 +(im) 55 83 +(lm) 169 80 -{Al wire] 1962 81 +(im)[Al wire]. 2621 96 TABLE VIII COMPOSITION OF AISI 1095 ELECTRODE AND INGOTS ELEMENT ELECTRODE INGOT 68 [ e l . - ve] INGOT 67 [ e l + ve(im)] C .975 .903 .914 Mn .39 .37 .38 Si .34 .27 .37 0 .0020 .0145 .0050 97 TABLE IX COMPOSITION OF AISI 430 ELECTRODE AND INGOTS ELEMENT ELECTRODE INGOT 63 [ e l - ve] INGOT 70 [ e l . + ve(im)] Cr 17.35 17.16 16.90 Mn 0.44 0.47 0.40 S i 0.26 0.15 0.19 C 0.060 0.051 0.060 P 0.024 0.024 0.022 S 0.015 0.006 0.009 0 0.0115 0.0225 0.0091 TABLE X TOTAL IRON CONTENT OF FVE SLAG CAPS INGOT NO. POLARITY (CaF 2 + 25 wt. % A l ^ ) (wt. % Fe ) 39 - 0.52 31 + 1.74 37 + (lm) 1.34 34 A.C. 1.01 CHAPTER 4 DISCUSSION OF SMALL SCALE STUDIES 4.1 Introduction The small s c a l e studies were c a r r i e d out with the purpose of formulating electrochemical r e a c t i o n mechanisms responsible f o r current transfer across l i q u i d m e t a l / l i q u i d slag i n t e r f a c e s , under ESR conditions. The r e a c t i o n mechanisms so proposed must be able to account for chemical and thermal phenomena which e x i s t during D.C. e l e c t r o s l a g melting. It has been stated that CaF^ slags conduct i o n i c a l l y and this was found to be e s s e n t i a l l y true, except when there were s i g n i f i c a n t amounts of Ca and A l metals dissolved i n the s l a g . The mechanism responsible for p o l a r i z a t i o n i n such systems i s the slow d i f f u s i o n of r e a c t i o n products away from the r e a c t i o n i n t e r f a c e into the slag and the metal. In the following section we s h a l l show that the anodic Faradaic r e a c t i o n f o r pure i r o n i n CaF^ - A^O^ slags i s at low current d e n s i t i e s , the anodic corrosion of i r o n leading to FeO saturation i n the slag at the i n t e r f a c e at s u f f i c i e n t l y high current d e n s i t i e s . In addition we s h a l l demonstrate that the cathodic Faradaic reaction involves deposition of A l and/or Ca such that some A l dissolves i n the i r o n and Ca and A l d i s -solve i n the s l a g . 4.2 Previous electrochemical work There have been very few studies i n which the electrochemical phenomena of l i q u i d metal electrodes i n slags at r e l a t i v e l y high tempera-99 100 tures have been investigated. In the case of chloride melts (22), i t has been shown that there e x i s t s clear evidence for electrochemical reac-tions i n which the slow step i s the adsorption of a complex species. This i s an a c t i v a t i o n process and has a very short t r a n s i t i o n time (< 1 p s e c ) . At higher temperatures there are only a few studies i n v o l v i n g the steady state p o l a r i z a t i o n of a carbon-saturated iron/oxide slag i n t e r -face. Shantarin (23) studied the anodic p o l a r i z a t i o n of carbon satura-ted i r o n surfaces i n oxide melts and, although he did not f i n d the nature of the electrochemical reactions, he mentioned that the anodic d i s s o l u t i o n 2+ of Fe to give Fe ions i n the slag was accompanied by considerable current e f f i c i e n c i e s , and was assumed to lead to concentration p o l a r i z a t i o n at the i n t e r f a c e . Other workers (24) studied the cathodic processes of carbon saturated s t e e l s i n various fluoride-oxide slags but they were mainly con-cerned with a l l o y content increases r e s u l t i n g from e l e c t r o l y s i s . Gosh and King (16) studied the discharge k i n e t i c s of oxide ions from lithium s i l i -cate melts on platinum anodes. They employed a galvanostatic pulse tech-nique, and observed electrode p o l a r i z a t i o n phenomena produced by evolution of oxygen gas at the electrode. In another case, the concentration p o l a r i -zation of i r o n and n i c k e l surfaces i n oxide-free and n i c k e l oxide saturated f l u o r i d e melts was studied i n the temperature range 500 - 600°C (25). It was found that an important aspect of the reaction c h a r a c t e r i s t i c s i n the oxide containing melts was the semi-conducting nature of the passive oxide layer formed on the metal electrode surface. Far the case of most low temperature studies, the aim has been to define the electrochemical reac-t i o n steps, while eliminating concentration p o l a r i z a t i o n from the c e l l . 101 This means that a l l these experiments had to be c a r r i e d out at low (<1 A. cm. ) current d e n s i t i e s . However, i n the case where polarography has been studied i n fused s a l t s (26), the system has been found to follow c l o s e l y the r u l i n g equations derived f o r analogous aqueous s i t u a t i o n s under s i m i l a r hydrodynamic c o n t r o l . 4.3 Anodic p o l a r i z a t i o n of pure i r o n i n A^O^ slags The anodic behaviour of pure i r o n i n CaF 2 - A^O^ slags i s shown i n F i g . 8. These p o l a r i z a t i o n curves can be represented schemati-c a l l y i n F i g . 30. The curve has three sections denoted A, B and C. Part A of the anodic p o l a r i z a t i o n curves f i t s an exponential form which, to-gether with the observed long times (.5 ->• 5 sec.) required to e s t a b l i s h steady state p o l a r i z a t i o n , implies that the mechanism i s a d i f f u s i o n con-t r o l l e d process, leading to a l i m i t i n g current density, i n - Values of i ^ were obtained from the curves for 1, 5, and 10 wt. % A^O^ slags by extra-polating the sections A of the curve. The curves were then plotted accord-ing to the equation (1-14): n d = ^ • In ^ — — ) (4-1) as shown i n F i g . 31. Values of n between 1 and 0.1 are obtained which seems to i n d i c a t e that the electrochemical mechanism responsible f o r sec-t i o n A i s indeed a d i f f u s i o n l i m i t e d r e a c t i o n . Such processes generally involve the depletion of a d i l u t e i o n i c species i n the solvent e l e c t r o l y t e near the electrode to such an extent that the rate of d i f f u s i o n of t h i s 102 Figure 31. A p p l i c a t i o n of a l i m i t i n g current density law to anodic p o l a r i z a t i o n of pure i r o n i n CaF + A l 0 slags. 104 species towards the electrode equals the rate of removal of the ions at the electrode. Such a s i t u a t i o n i s u n l i k e l y i n th i s fused s a l t system because the solvent i s being e l e c t r o l y z e d ( t h e r e i s no solute to be deple-ted) . The simplest anodic process i n th i s case i s the reaction Fe g F e 2 + + 2e" (4-2) which w i l l not lead to such a l i m i t i n g law (14). It i s , however, possible 2+ that the corrosion product, Fe , could lead to saturation of the slag at the electrode i n t e r f a c e , and that the measured electrode p o t e n t i a l would therefore reach some maximum value which would be determined by the equi-librium electrode p o t e n t i a l of the Fe/FeO i n t e r f a c e . The values of n obtained from F i g . 31 are then simply a fo r t u i t o u s representation of the 2+ exponential d i f f u s i o n gradient of Fe e x i s t i n g i n the slag at the e l e c -trode i n t e r f a c e at current d e n s i t i e s too low to saturate the i n t e r f a c e . It i s therefore reasonable to expect the values of n to vary with slag oxide content which should not occur i f a s i n g l e d i f f u s i o n l i m i t e d reac-t i o n i n v o l v i n g oxide ions was taking place. When the current density at the electrode i s s u f f i c i e n t l y high, saturation of the i n t e r f a c e i n the anodic corrosion product takes place, and the curves enter section B on the schematic p l o t , F i g . 30. I t i s necessary to stress that the p o t e n t i a l d i f f e r e n c e measured between the working and reference electrodes i s established by the p o t e n t i a l e x i s t i n g at the i r o n electrode surface. Thus a higher current density i n section B of a given curve w i l l only produce a thicker saturated l a y e r , and w i l l not a f f e c t the measured p o t e n t i a l , but this i s true only i f the saturated layer remains i o n i c i n nature, and does not become an e l e c t r o n i c conductor. 105 If the sl a g remains i o n i c , the Faradaic process continues to take place at the electrode surface. If the anodic corrosion of i r o n i s indeed the Faradaic process responsible f o r these r e s u l t s , the rate at which concen-t r a t i o n p o l a r i z a t i o n would a r i s e i n the slag due to the establishment of a steady state concentration gradient, w i l l depend on the electrode current density and on the slag composition (oxide type and content). It i s of i n t e r e s t to c a l c u l a t e the approximate thickness of t h i s proposed FeO saturated layer. This c a l c u l a t i o n r e f e r s to a pure i r o n e l e c -trode i n a Ca F 2 + 2.5 wt. % A^O^ s l a g . Anodic saturation began at a 2 current of 600 mA. on an electrode whose t o t a l exposed area was 2.75 cm . Approximately 3.2 sec. was required to saturate the surface and the coulom--2 b i c density was therefore 0.70 coul. cm. For the reaction 2+ Fe •> Fe + 2e n = 2, and the number of moles of i r o n corroded i s equal to 3.6 x 10 ^ -2 -4 moles, cm. , which i s equivalent to 2.6 x 10 gm. of FeO per sec. The saturation s o l u b i l i t y of FeO i n low oxide CaF 2 slags i s low and f o r th i s slag w i l l be l e s s than 5 wt. %, and an assumed value of 1 wt. % i s used.. _3 The density of pure CaF 2 at 1500°C. i s 2.54 gm. cm. and neglecting the change i n density (increase) due to saturation with FeO, the weight of saturated slag per unit area w i l l then be 2.6 x 10 gm. which has a v o l -- 2 - 3 ume of 1.03 x 10 cm. . This l a t t e r value contains the assumption that a l l the FeO produced remains i n the v i c i n i t y of the electrode surface. Therefore the thickness of the FeO saturated layer at a current density which i s high enough to produce saturation w i l l be approximately 100 mi-crons thick. 106 It i s also of i n t e r e s t to c a l c u l a t e the thickness of a d i f f u -sion boundary layer i n the small scale system. Levich (27) derived an expression for the d i f f u s i o n boundary layer thickness, < 5 , for the case of convective d i f f u s i o n . The d r i v i n g force f o r mass transport i s the density gradient produced by movement of e l e c t r o - a c t i v e species to the electrode surface. The equation so derived should also apply to the case of d i f f u s i o n of Faradaic reaction products away from the electrode surface into the bulk of the slag, and i s : 1/4 6 = x — : — (4-3) where x - distance from the upper edge of the plate Pr - Prandtl No. of the slag g - a c c e l e r a t i o n due to gravity C - concentration of d i f f u s i n g species at the plate surface v - Kinematic v i s c o s i t y of the sla g . This equation applies to natural convection at a planar e l e c -trode surface, however i n the case of c y l i n d r i c a l d i f f u s i o n the divergence of the f l u x would tend to decrease the d i f f u s i o n boundary layer thickness. By assuming that the electrode surface i s planar, <5 can be calculated us-ing the following data: x = 1 cm. Pr = 0.12 which i s the Prandtl No. of l i q u i d NaCl (28) -2 -3 C = 2.54 x 10 gm. cm. for 1 wt. % FeO saturation i n low oxide CaF„ slags 107 v p_ _ 0.6 poise p 2.54 gm.cm. _3 = 0.24 cm. sec. -1 C4-4) and 6 = (0.7K.12) 1/4,980 • 0.0254.1/4 L 4 • 0.058 ; <5 = 0.757 cm. D i f f u s i o n boundary layer thicknesses i n natural convection are found to be s u b s t a n t i a l l y greater than i n forced convection, and the values of the d i f f u s i o n a l fluxes are found to be correspondingly smaller. The calculated value of <5 i s much larger than would be found i n aqueous systems It i s quite l i k e l y that the actual d i f f u s i o n boundary layer thickness i n the small scale system i s smaller than this calculated value, i t s thickness being determined by thermal convection i n the s l a g . There i s no phase diagram information a v a i l a b l e on the CaF -A1„0 -FeO system, but i n analogy with the CaF^ - A^O^ - MnO system (29) and the CaF^ - CaO - FeO system (30) , (the p a r t i a l phase diagrams of which are given i n F i g . 32, and show the FeO s o l u b i l i t y to be strongly composition and temperature dependent), the system i s believed to exhibit a large mia-c i b i l i t y gap. Hence the corrosion rate at which surface saturation i n FeO i s attained would be expected to depend strongly on slag composition and temperature. Once the slag of the anodic electrode surface has become saturated i n FeO, the p o t e n t i a l of the surface may be represented by s i t i o n as i s seen' from the approximately constant plateau values i n F i g . 8. The o v e r a l l p o t e n t i a l d i f f e r e n c e which i s measured between the r e f e r -ence and working electrode w i l l then be represented by the v i r t u a l reac-Fe(s) + 0 -> FeO + 2e (4-5) (slag sat'n.) This p o t e n t i a l w i l l be approximately independent of slag compa-ct 108 CaO Figure 32. P a r t i a l phase diagrams of the system CaF o-Ca0-Fe0, 109 t i o n : Fe + CO t c + (FeO) (4-6) S 8 g r (slag sat'n) which i s obtained by adding reactions (2-6) and (4-5)U; I t i s d i f f i c u l t to check (4-6) w i t h an exact c a l c u l a t i o n because the carbon monoxide pres-sure at the reference electrode i s unknown. However, i f i t i s assumed that the e f f e c t i v e pressure of CO at the reference electrode i s one a t -mosphere, the p o t e n t i a l d i f f e r e n c e e x i s t i n g between the reference and saturated i r o n electrode (a„ _ at the i r o n surface i s unity) can be c a l -FeO J culated using standard thermochemical data (31) at the measured slag temperature of 1480°C (1753°K). C(gr) + - | o 2(g) •+ C0(g) A F° = 63,540 c a l . (1753°K) (4-7) Fe(6) + -|- 0 2(g) + "Fe0"(l) A F° = -36,496 c a l . (1753°K) (4-8) and AF° = + 27,044 c a l (1753°K) = -nFE° so E° = -586 mV. where n = 2 equivalent, mole \ and F = 23,060 c a l . • v o l t equivalent. This calculated value of 586 mV. i s reasonably close to the ob-served value of approximately 500 mV. measured as the beginning of sections B i n F i g . 8. At higher current d e n s i t i e s i n sections B, i t i s thought that the r a t e of electron transfer could be accommodated by increasing the 3 + 2 + r a t i o of Fe to Fe by the reaction* Fe -* Fe J + e (4-9) 1 10 i n the saturated s l a g layer, which w i l l have the e f f e c t of d i s p l a c i n g the measured p o t e n t i a l to a higher value. This i s seen as the slope of the sections B i n Figure 8, and w i l l again be independent of slag composition. As mentioned previously, i t was found to be impossible to carry out small scale studies i n high oxide slags due to A^O^ p r e c i p i t a t i o n on the e l e c -trode and due to the r i s i n g l i q u i d u s temperature at kl^O^ concentrations above the eutectic composition (10 wt. % A^O^) . None the l e s s , a trend can be observed i f we p l o t the extrapolated values of the l i m i t i n g current density (1^) against the mole % of A^O^ i n the slag as shown i n F i g . 33. This seems to i n d i c a t e that as the A^O^ content of the slag i s increased, i p w i l l asymtotically approach a " l i m i t i n g " value. The p r a c t i c a l s i g n i f i -cance of this i s that as long as anodic metal-slag i n t e r f a c e current den-s i t i e s are maintained below the i curve during D.C. e l e c t r o s l a g melting, FeO saturation of the anodic surface w i l l not occur, and oxidation of the metal w i l l be held to a minimum. It was considered necessary to examine the p o l a r i z a t i o n r e s u l t s to see i f there were a c t i v a t i o n mechanisms present which would have very f a s t r i s e and decay time at these temperatures. If such a decay existed, i t would be included i n the p o t e n t i a l d i f f e r e n c e we have at t r i b u t e d to an ohmic p o t e n t i a l gradient seen as the i n i t i a l r i s e on the o s c i l l o s c o p e trace. To check t h i s , the "IR" portion of the r i s e times for a Ferrovac-E anodic electrode i n a CaF^ + 2.5 wt. % Al^O^ slag was plotted against the current I, as shown i n F i g . 34. The s t r a i g h t l i n e behaviour of the p l o t ex-cludes the presence of any a c t i v a t i o n process i n these systems (16). I l l 113 It i s thought that towards the end of sections B i n the anodic p o l a r i z a t i o n curves of F i g . 8, when the rate of electron transfer has 3+ 2+ become too high to be accommodated by increasing the Fe /Fe r a t i o , that gas evolution at the i n t e r f a c e i s achieved by the reactions: 0 2~ + 0* + 2e~ (4-10) 20* + 0 2(g) (4-11) It was shown i n F i g . 14 that anodic p o l a r i z a t i o n at r e l a t i v e l y high current densities produces oxygen i n the bulk of the electrode. F i g . 35 i s the o s c i l l o s c o p e trace of a high current density anodic pulse on a Ferrovac-E electrode i n a low oxide content slag (250 ppm. CaO). The ser-rated behaviour of the steady-state portion of the curve i s c h a r a c t e r i s -t i c of gas evolution on metal electrodes as observed by Gosh and King (16). It was observed that anodic electrodes showed s i g n i f i c a n t d i s s o l u t i o n when polarized at high current density i n section B, but not when polarized i n section A. This behaviour i s expected as a r e s u l t of formation.of l i q u i d FeO on the s o l i d electrode surface, the FeO having a lower melting point (^  1380°C) than the iron'. The oxidation of the electrode i s explained by the following mechanism. During the time when the electrode surface i s saturated with FeO, there w i l l be s i g n i f i c a n t d i f f u s i o n of oxygen into the s o l i d electrode despite the f a c t that the i r o n i t s e l f i s being ano-d i c a l l y corroded. When the oxygen content of the surface layer i s high enough, FeO forms as a l i q u i d which f a l l s away from the s o l i d electrode producing the observed electrode d i s s o l u t i o n . l i l t 0 .5sec /div t ime F i g u r e 35. A n o d i c p u l s e on pure i r o n i n C a F 2 + 250 ppm CaO. The e l e c t r o d e p o l a r i z a t i o n i s seen t o t r a n s f e r from the s a t u r a t i o n c o n d i t i o n t o an a r c , a p o l a r i z a t i o n c o n t i n u e s i = 500 ma.cm r o 115 4.3.1 Apparent t r a n s i t i o n time The type of anodic p o l a r i z a t i o n curves r e s u l t i n g from a s i n g l e pulse t e s t as shown i n F i g . 12 were found at values of current l y i n g i n the middle region of sections B i n F i g . 8. At higher and lower current d e n s i t i e s , the i n d i v i d u a l traces exhibited the usual exponential r i s e and decay behaviour. The shape of the galvanostatic p o l a r i z a t i o n curve shown i n F i g . 12 i s q u a l i t a t i v e l y the same as that expected f o r a t r a n s i t i o n between two slow, d i f f u s i o n c o n t r o l l e d processes as described i n equation (1-28). Therefore analysis of these curves according to the r e l a t i o n s h i p .1/2 1/2 . , RT .. T n, - A + — In d nF t l / 2 should i n d i c a t e whether or not the p o l a r i z a t i o n mechanism being studied was indeed a d i f f u s i o n c o n t r o l l e d r e a c t i o n i n which some i o n i c species i n the slag was being depleted at the electrode surface. The t r a n s i t i o n curves were analysed f o r applied currents of 0.5 A, 2A., and 3A. and the values of the current density ( i _ ) , the t r a n s i t i o n time T, and the pro-1/2 duct i Q T are given below. I(A.) i (A. cm. 2) T (sec.) i x 1 ^ 2 (A. sec."^ 2) o o 0.5 0.18; 8.5 2.92 2.0 0.73 1.3 1.14 3.0 1.09 0.8 0.89 I t i s apparent that the t r a n s i t i o n time does indeed decrease as 1/2 the current density increases, but the product i Q x i s not constant for a given s l a g , a condition whlchmust.be s a t i s f i e d f o r such a mechanism. 116 The d i f f u s i o n overvoltage i s p l o t t e d according to equation (1-28) for the T l / 2 _ 1/2 three applied currents i n F i g . 36 [f(x) = In ] and although the points f o r each current density l i e on s t r a i g h t l i n e s , the slopes increase with the current density. Values of n can be calculated by RT equating the measured slope to and are given below. I(A.) n (gm. equiv../gm. mole) 0.5 5.2 2.0 3.5 3.0 3.0 The values of n so obtained are not unreasonable i n magnitude, but vary over a range which i s much wider than would be expected f o r a correct a n a l y s i s . Because the values of n and, also, the values of the 1/2 product i x are found to be functions of the current density, i t must o be concluded that the observed t r a n s i t i o n curves do not a r i s e from d i f f u -s i o n a l depletion of the slag i n some i o n i c species. This f i n d i n g agrees with the previously proposed mechanism of anodic corrosion of ir o n which i s a s i n g l e step d i f f u s i o n process and would not exhibit a t r a n s i t i o n time behaviour. The apparent t r a n s i t i o n times must therefore be explained i n another way. Such an explanation can be found i f we examine Figures 32 and 37. F i g . 37 i s the o s c i l l o s c o p e trace obtained when r e p e t i t i v e anodic pulses are applied i n the " t r a n s i t i o n " region, at a frequency which main-tains part of the'saturated FeO layer produced by the previous pulse. Here i t i s observed that the i n f l e c t i o n i n the p o l a r i z a t i o n curve i s 117 Figure 36. A p p l i c a t i o n of t r a n s i t i o n time law to anodic p o l a r i z a t i o n of pure i r o n i n CaF + 2.5 wt.% A1 20 3. 2 118 7 0.2v/|_ div. 1 A ~\ J L J I I L J I I I 1 2 sec/div. t ime F i g u r e 37. S u c c e s s i v e a n o d i c p u l s e s a p p l i e d t o p u r e i r o n i n C a F 2 + 2.5 wt.% Al 2°3> showing the d i s -appearance o f the a p p a r e n t t r a n s i t i o n t i m e whe s u r f a c e s a t u r a t i o n i s r e t a i n e d between p u l s e s . 119 gradually removed with, successive pulses. This indicates that the trans-i t i o n curves are the r e s u l t of chemical phenomena rather than e l e c t r o -chemical ones, and can be explained using the s i n g l e electrode r e a c t i o n . 2+ Fe -* Fe + 2e If we examine the p a r t i a l phase diagrams of the CaF^-CaO-FeO system ( F i g . 32), we see a two l i q u i d region i n a composition range through which the slag at the electrode surface must pass i n order to 2+ achieve Fe saturation. Therefore the f i r s t portion of the t r a n s i t i o n curve, shown as l i n e A-A' i n F i g . 32, represents the simple d i s s o l u t i o n 2+ of Fe , while the i n f l e c t i o n point indicates the beginning of the two 2+ l i q u i d slag system as Fe saturation i s approached. 4.4 Anodic p o l a r i z a t i o n of pure i r o n i n CaO slags The anodic p o l a r i z a t i o n behaviour of pure i r o n electrodes i n CaF 2 - CaO slags i s shown i n F i g . 10. Comparisons of Figures 8 and 10 show that the plateau section B i s much more pronounced i n Al^O^ contain-ing slags than i n CaO containing slags where the overpotential at equi-valent current density i s seen to decrease more r a p i d l y with increasing CaO content. This ,-is probably the combined r e s u l t of the lower v i s c o s i t y of CaF 2 - CaO l i q u i d s C32) and t h e i r higher s o l u b i l i t y (30) of F e 2 + , as compared with CaF 2 - A l ^ ^ l i q u i d s at equivalent oxide content. It was also observed that anodic i r o n electrodes melted at much lower current d e n s i t i e s and CaO contents than they did i n an.equivalent A1„0„ content. 120 One can speculate, from the r e s u l t s i n F i g . 10, that an ESR slag con-taining 25 wt. % CaO, at a given current density, would not saturate 2+ a melting i r o n electrode i n Fe , whereas an ESR slag of the same kl^O content would. This might r e s u l t i n the f i n a l metal having a lower oxygen content when melted through the CaO containing slag because of the lower oxidation rate at the anodic surface. 4.5 Cathodic p o l a r i z a t i o n of pure i r o n i n ESR slags In order to provide a Faradaic mechanism at the slag/metal cathodic i n t e r f a c e , we may postulate any combination of the following reactions: C a 2 + + 2e~ -> Ca* (4-12) Ca* + (Ca°) . (4-13) slag Ca* -> Ca°g (4-14) 3+ - * A l + 3e ^ A l (4-15) A l * -> [ A l ] p E (4-16) A l * + (Al°) . (4-17) slag A l * -> A l ° £ (4-18) fo r the appropriate slag cation composition. In the case of the o v e r a l l r e a c t i o n (4-12) and (4-14), the poten-t i a l seen by the measuring c i r c u i t i s represented by the equation: (CaO) + C -y CO + Ca, . (4-19) slag gr g (g) S i m i l a r l y , the p o t e n t i a l seen at an electrode which, i s cathodi-c a l l y p o l a r i z i n g due to reactions (4-15) and (4-16) i s represented by the 121 reaction: (Al.O,) . + 3 C + 3CO + 2[A1]_, (4-2 3 slag gr g JFe giving p o t e n t i a l s : AE°. ., = 800 mV 4-19 and E°. . = 400 mV 4-20 at 1812°K, for Raoultian standard s t a t e s . However, as the observed reac-t i o n must involve s o l u t i o n of A l i n both the metal and the slag and Ca i n the s l a g , we would not expect to see these potentials as the p o l a r i z a -t i o n values. The form of the cathodic r\/ln i curves at the lower current o densities i s again that to be expected from the d i f f u s i o n of r e a c t i o n products away from the cathodic i n t e r f a c e . One might expect to observe a l i m i t i n g p o t e n t i a l at i n t e r f a c e saturation but t h i s w i l l not occur as i t did i n the anodic cases because there i s no electrochemical mechanism by which the i n t e r f a c e can remain saturated as the current density i s increased above the i n i t i a l s a turation current density. At higher cur-rent d e n s i t i e s , i t i s believed that the calcium gas produced forms a s o f t arc which i s very stable and the increase i n measured p o l a r i z a t i o n r e s u l t s from the arc resistance. This concept i s supported by F i g . 38 which i s a p l o t of the cathodic overpotential against the working current of a pure i r o n electrode i n a CaT?^ + 1 wt. % A^O^ s l a g . The curve be--2 comes l i n e a r a f t e r a current of 5 A. CI.8 A. cm. ] which i s c h a r a c t e r i s -t i c of an arc process. In slags w i t h a higher oxide content, i t appears that this very stable s o f t arc does not i n i t i a t e u n t i l higher current d e n s i t i e s are attained. This i s one way i n which the cathodic p o l a r i z a -123 t i o n curves can be reasonably explained. Another possible explanation of the l i n e a r behaviour shown i n F i g . 38 i s that the current density, i Q , and convection boundary layer are r e l a t e d In such a fashion that a l i n e a r overpotential-current density pl o t i s obtained. This i s based on the assumption that the measured over-potentials do indeed r e s u l t from a change i n a c t i v i t y of a deposited spe-cies at the i n t e r f a c e , and that t h i s p o t e n t i a l can be described i n terms of a Nernst equation. This would apply when a steady-state condition i s achieved, with the d i f f u s i o n overvoltage given by where c ( i ) i s the imposed concentration of the d i f f u s i n g species at the i n t e r f a c e , and c i s the concentration before current passage (concentra-t i o n has been substituted for a c t i v i t y ) . If we assume that the i n i t i a l curved portion of the plot given i n F i g . 39, represents the cathodic de-p o s i t i o n of A l i n the i r o n by the r e a ction: and that the l i n e a r portion represents the deposition of Ca which sub-sequently dissolves i n the s l a g , the current density at which the curve 3+ becomes l i n e a r should be the l i m i t i n g d i f f u s i o n current density of A l ions i n the s l a g . The d i f f u s i o n boundary layer can therefore he c a l c u -lated using equation (1-11). _ RT . c ( i ) n., = — In — d nF c A l + 3e (Al) Fe 1 nF = D . Cc - c ) 6 when i = c = 0. 124 -2 For the anodic r e a c t i o n given above, n = 3 and i ^ = 1.8 A. cm. The slag contains 1 wt. % A^O^ (M.Wt. = 102 gm) , and the den--3 - -4 s i t y of the slag i s approximately 2.6 gm.cm. , therefore c = 2.55 x 10 moles, cm. F = 96,500 coul. gm. equiv. -5 2 -1 D = 8.5 x 10 cm. sec. (33) and 6 can be calculated to be 0.347 cm. Therefore the d i f f u s i o n boundary layer for A^O^ deple-t i o n at a cathodic i r o n surface i s approximately 3500 microns thick, a reasonable value for such a system. When the electrode current density 2+ exceeds the l i m i t i n g current density f o r A l deposition, Ca ions must deposit, and the l i n e a r portion of the curve must correspond, i n the ab-sence of a stable s o f t arc, to the reaction: C a 2 + + 2e~ •> [Ca] , slag where the measured overpotential i s determined by the Ca a c t i v i t y at the electrode i n t e r f a c e . For Nernst type behaviour, i s proportional t o . c In c, and from equation (1-11), i i s proportional to -r (c = 0 ) . There-fore i and 6 must be r e l a t e d i n such a way that the concentration, c, of Ca i n the slag at the i n t e r f a c e , v a r i e s to produce the r e l a t i o n s h i p : i a In c and therefore n. a In c. d Although i t i s d i f f i c u l t to conceive of such behaviour during e l e c t r o l y s i s , i t i s f e l t that t h i s explanation i s more r e a l i s t i c that the idea of e s t a b l i s h i n g a soft arc at these low current d e n s i t i e s . 125 In high oxide content slags containing Al^O^, A l and Ca should be deposited together such that some A l w i l l d i s s o l v e i n the Iron e l e c -trode while A l and Ca d i s s o l v e i n the s l a g . The f a c t that A l i s produced at a cathodic surface was shown by- the electron microprobe studies of an cathodic i r o n electrode as discussed i n Section (2-5). . 4.6 P o l a r i z a t i o n of Fe-Cr a l l o y s and pure chromium The anodic and cathodic p o l a r i z a t i o n behaviour of AISI 430 s t a i n l e s s s t e e l electrodes was shown i n Figures 15 and 16. Very l i t t l e information can be drawn from these curves except to say that at an equiva-lent current density and slag composition, the degree of anodic p o l a r i z a -t i o n i s les s on a s t a i n l e s s s t e e l electrode than on a pure i r o n electrode. No B regions, or plateaus were found on the anodic p o l a r i z a t i o n curves of these electrodes. At low current d e n s i t i e s , the anodic r e a c t i o n i s thought to be Cr + C r 3 + + 3e~ (4-21) As the current density i s increased, and the Cr concetnration at the i n t e r -face decreases due to p r e f e r e n t i a l anodic corrosion, i r o n w i l l s t a r t to 3+ 2+ corrode to produce a b u i l d up of Cr and Fe ions at the electrode i n t e r -face. The cathodic p o l a r i z a t i o n curves ( F i g . 16) show a behaviour sim-i l a r to the cathodic curves f o r pure i r o n shown i n F i g . 9. This would i n d i c a t e that the cathodic r e a c t i o n i s not appreciably influenced by the presence of Cr as an a l l o y i n g element i n the Iron matrix. 126 The concentration gradient of Cf at the surface of an AISI 430 s t a i n l e s s s t e e l shown i n F i g . 17 gives ample evidence of the p r e f e r -e n t i a l anodic corrosion of Cr from Fe-Cr a l l o y s . As indicated previous-l y the removal of Cr took place at a current e f f i c i e n c y of only 10%. This means that even at this r e l a t i v e l y low current density, the anodic reaction was accounted f o r i n the main by corrosion of i r o n , not chromium. One might expect to observe an anodic p o l a r i z a t i o n plateau established by FeO saturation i n the slag because of the f a c t that Cr removal accounts for only 10% of the current passage. However, no plateau was observed because the p o t e n t i a l s measured were established by corrosion of both, Cr and Fe such that the e f f e c t of the Cr was to mask the Fe plateau. The anodic p o l a r i z a t i o n curves for the Fe -1 wt. % Cr a l l o y shown i n F i g . 18 can also be explained using the concept of Cr depletion at the electrode surface. P r i o r to current passage, the electrode sur-face has a Cr concentration of 1 wt. %. Upon passage of current, the electrode p o l a r i z e s to produce a plateau at approximately 230 mV., the value of which i s determined by the Cr content at the surface. As cur-rent passage proceeds, the Cr i s quickly depleted at the surface and anodic corrosion of i r o n takes place causing the measured p o t e n t i a l d i f f e r -ence between the reference and working electrodes to increase to the plateau value of an anodic Iron electrode. Upon current i n t e r r u p t i o n , the measured p o l a r i z a t i o n decay i s therefore that of an i r o n electrode as shown by decay p o l a r i z a t i o n curve i n F i g . 18. The anodic p o l a r i z a t i o n curve of pure chromium CFig. 19 ) exhi-b i t s a plateau at a p o t e n t i a l between 200 and 300 mV. This plateau arises 3+ from saturation of the slag at the i n t e r f a c e i n Cr ions and i s therefore 127 comparable to the plateaus found f o r anodic i r o n electrodes i n CaF^ -Al^O^ slags. The p o t e n t i a l d i f f e r e n c e e x i s t i n g between the graphite 3+ reference electrode and the Cr saturated surface i s represented by the equation: 3.C, . + Cr.O. •> 2 Cr, , + 3 CO, > (4-22) Cgr.) 2 3(g) ^s) ^ From the free energies given below (T = 1809°K), 3 Cr, . + -| 0„ -> 3 CO, . AF° = -194,007 c a l . (s) 2 2 ( g ) (g) (4-23) 2 C r r . + | 0 -*• ^ (g) C r 2 ° 3 ( e ) A F ° = - 1 59,481 c a l . (4-24) AF° = -34,526 c a l . rx. and AE° = -250 mV. This calculated value agrees very w e l l with the observed value of the plateau. 4.7 P o l a r i z a t i o n of pure n i c k e l The small scale studies c a r r i e d out on an anodic n i c k e l e l e c -trode i n a CaF 2 + Al^O^ slag were not s u i t a b l e f o r presenting as a p o l a r i -zation curve because at the lower current d e n s i t i e s , the electrode would not p o l a r i z e to a steady s t a t e value i n the time a v a i l a b l e during an i n -d i v i d u a l pulse experiment. Steady state p o l a r i z a t i o n was not achieved. -2 u n t i l the current density was increased to 1.1 A. cm. The o s c i l l o -scope trace of t h i s test i s shown i n F i g . 20, and the observed p o l a r i z a -t i o n p o t e n t i a l i s approximately 700 mV, which should correspond to the reaction: 128 C, . + NiO, . -* N i , . + CO. (4-25) (gr) (s) (s) (g) The free energy change for t h i s r e a c t i o n i s calculated as follows at 1726°K: C ( g r ) + 2 ° 2 ^ C O ^ AF° = -62,996 c a l . (4-26) N i , . + i 0 o -> NiO, . AF° = -21,325 c a l . (4-27) (S) 2  2Qg) (- s^ AF° = -41,671 c a l . rx and AE° - -904 mV. The discrepancy between the observed and calculated values f o r the plateau height might be explained by the r e l a t i v e l y long time needed to e s t a b l i s h the steady-state p o l a r i z a t i o n i n t h i s sytem, a time during which the e n t i r e system i s cooling. If the system temperature was s i g -n i f i c a n t l y lower than the assumed temperature of the melting point of n i c k e l , then a calculated AE° would also be smaller. The electronmicroprobe r e s u l t s obtained upon the examination of a c a t h o d i c a l l y polarized n i c k e l electrode were presented i n F i g . 21. Although calcium i s i n s o l u b l e i n i r o n (34), i t i s soluble i n n i c k e l , but the question arises as to whether or not Ca w i l l be deposited i n the pre-sence of A^C^ containing slags. B e l l (21) considered the i n t e r a c t i o n between CaO + A^O^ i n CaF^ slags and l i q u i d n i c k e l , and under e q u i l i -brium conditions, found that the i n t e r a c t i o n between the oxides and dissolved A l and Ca i n l i q u i d n i c k e l was suck that as long as alumina i s present i n a s l a g , alumina w i l l react with, l i q u i d n i c k e l to a much, greater extent than CaO. He stated that Ca would not d i s s o l v e i n the 129 metal even when the CaO a c t i v i t y i n the slag was 0.56, unless the A^O^ a c t i v i t y was less than 0.00018, which was c e r t a i n l y not the case i n the CaF 2 + 8 wt. % AlyO^ slag used for e l e c t r o l y s i s of the n i c k e l electrode. We must therefore conclude that the Ca containing regions shows i n F i g . 21, were not electrochemically formed, but that this calcium was i n the n i c k s l i n the as received condition. 4.8 Anodic p o l a r i z a t i o n of cobalt The anodic p o l a r i z a t i o n behaviour of cobalt was very s i m i l a r to that of n i c k e l i n that steady-state p o l a r i z a t i o n could not be attained u n t i l the current density was high. F i g . 22 shows the o s c i l l o s c o p e trace -2 obtained at a current density of 2.0 A. cm. This shows a measured p o t e n t i a l d i f f e r e n c e of almost 800 mV. which corresponds to the reaction: C ( g r ) + C O°(s) * C ° ( Y ) + C 0 ( g ) <4-28) The free energy change for t h i s reaction i s calculated at 1766°K, the melting point of cobalt, as follows: C ( g r ) + 2 ° 2 ( g ) + C 0 C g ) AF° = -63,803 c a l . (4-29) Co +4-0 -*• CoO AF° = -26,010 c a l . (4-30) 2 2 ( g ) L S ) AF° = -37,793 c a l . rx and AE° = -819 mV. This f i g u r e of 819 mV. agrees very w e l l with the measured value. If the r i s e times i n Figures 20 and 22 are compared, i t i s apparent that the cobalt r i s e time i s much fas t e r and the degree of system cooling i s therefore kept to a minimum. The f a s t e r r i s e times so observed must be a r e s u l t of a d i f f e r e n c e i n saturation s o l u b i l i t y of CoO and NiO i n CaF 2 slags. 130 In s i l i c a t e melts, i t i s found that the saturation s o l u b i l i t y of oxides of t r a n s i t i o n metals decreases according to the order Cr, Fe, Ni, Co, and i t i s possible that a s i m i l a r trend e x i s t s i n the slags pre-sently studied. 4.9 Anodic p o l a r i z a t i o n of an Fe-C a l l o y As stated i n s e c t i o n (2.5.7) the anodic p o l a r i z a t i o n of an Fe-C electrode s i g n i f i c a n t l y decreased i t s carbon content. This ob-served carbon loss must have taken place by evolution of carbon-monoxide gas despite the f a c t that there i s often ,a bubble nucleation problem associated with carbon-monoxide evolution. I t i s not s u r p r i s i n g , however, that carbon monoxide evolution occurred because, according to D i s t i n et. a l . (35), there w i l l be no nucleation b a r r i e r at an i r o n surface con- 1 taining 0.8 wt. % C when i t i s i n contact with FeO saturated s l a g . From t h i s experiment one can conclude that carbon loss during D.C. ESR could be quite s u b s t a n t i a l . 4.10 High current density p o l a r i z a t i o n The v i s u a l observation of arc e f f e c t s at very high current de n s i t i e s on both anodic and cathodic surfaces i s supported by the behaviour of the p o l a r i z a t i o n curves shown i n F i g . 13. It i s evident that at s u f f i c i e n t l y h i g h current d e n s i t i e s the electrode surface d i s -plays an e s s e n t i a l l y l i n e a r overpotential/current density behaviour which i s unexplainable by a d i f f u s i o n c o ntrolled mechanism. However, i t would be the case i f the observed p o l a r i z a t i o n resulted from the resistance of 131 a gas f i l m around the electrode (36). Such an arc could a r i s e i n either of two ways. F i r s t l y , the electrode r e a c t i o n could evolve gas, e i t h e r Ca, N , 0„ , , or F„, at a rate s u f f i c i e n t to create a steady-state gas (g) 2(g) 2(g) envelope around the electrode. At the high temperature at the t i p , s u f f i c i e n t e l e c t r o n transport would be p o s s i b l e i n t h i s envelope to trans-f e r the electrode reaction to i t s outer (slag/gas) surface, thus r e t a i n i n g the electrode r e a c t i o n , and hence the arc would be s e l f - s u s t a i n i n g . F i g . 35 shows the formation of such a condition i n a CaF^ + 250 ppm. CaO s l a g . The electrode f i r s t p o l a r i z e s , i n t h i s case anodically, then evolves gas at a higher p o t e n t i a l . The anodic r e a c t i o n here would be the i n i t i a l f o r -mation of a saturated layer of FeO at the electrode surface, by r e a c t i o n (4-2), which would be i n s t a n t l y followed by reaction (4-9). When the rate of these reactions i s i n s u f f i c i e n t to sustain the current density imposed by the current source, r e a c t i o n (4-11) w i l l be i n i t i a t e d , evolving a gaseous species to form the arc, which produces a s i g n i f i c a n t increase i n the p o l a r i z a t i o n p o t e n t i a l . This region of arcing i s section C i n F i g . 30. This s i t u a t i o n i s probably the phenomenon referred to i n the l i t e r a -ture as a " s o f t " arc, since although the e f f e c t i s v i s i b l e as a sudden increase i n r a d i a t i o n , i t has no c h a r a c t e r i s t i c arc s p l u t t e r . The i n i t i a -t i o n of such an arc i s very s i m i l a r to the i n i t i a t i o n of the "anode e f f e c t " i n aluminum e l e c t r o l y s i s (37), and as with the anode e f f e c t , the current density required f o r anodic arc formation i s strongly dependent on the oxide content of the s l a g . As i n the case of the t r a n s i t i o n into the FeO saturation condition at lower anodic current d e n s i t i e s , we can only use an approximate extrapolation to extend our findings to higher 132 TABLE XI CRITICAL CURRENT DENSITY ESTIMATION ELECTRODE POLARITY SLAG CaF„ ESTIMATED CURRENT DENSITY FOR ARC INITIATION + wt. % oxide A. cm -2 + + + + + 25 A 1 2 0 3 25 A 1 2 0 3 5 A 1 2 0 3 5 A 1 2 0 3 0 0 5 CaO 5 CaO 25 CaO 25 CaO 250 400 150 200 10 25 300 400 500 500 133 oxide slags. However, since the current density at the point of arc i n -i t i a t i o n marks the upper l i m i t f o r the working of a stable e l e c t r o s l a g process electrode, we f e e l that such, an approximation w i l l provide v a l u -able operating information. The anodic and cathodic current l i m i t a t i o n f igures are shown i n Table XI. The electrode current density i n the U.B.C. e l e c t r o s l a g f u r --2 nace i s approximately 100 A. cm. , while i n a commercial e l e c t r o s l a g , furnace, the current density on an 18 i n . diameter electrode at a current _2 of 15 kA. i s approximately 10 A. cm. Both of these current d e n s i t i e s are below the estimated values of c r i t i c a l current density for arc i n i t i -a t ion i n a 25 wt. % Al^O^ s l a g . On the other hand, the current d e n s i t i e s i n e l e c t r o s l a g welding are very high. For a wire diameter of 1/8 i n . and a process current of 450 A., the electrode current density i s approximate-4 -2 l y 3 x 10 A. cm. This current density i s far greater i.than that r e r quired for arc i n i t i a t i o n and e l e c t r o s l a g welding must therefore operate as a submerged arc process. A second possible mechanism for such arc i n i t i a t i o n i s heat generation by Joule heating i n the p o l a r i z e d slag layer leading to l o c a l b o i l i n g i n the s l a g . This has been considered from a t h e o r e t i c a l stand-point (38) i n analogy with the Leidenfrost phenomenon, and may operate i n the present case. I t i s possible to c a l c u l a t e the required current den-s i t y to b o i l the slag at the electrode t i p using the following model. Consider a " c o n i c a l p o l a r i z e d electrode with a surface area of 2 11.3 cm. (an average electrode tip).. J o s h l (32) calculated the heat transfer c o e f f i c i e n t , h, between slag and electrode to be: 1 3 4 -1 -1 -2 h = 1 cal.°C sec. cm. I f the assumption i s made that a l l of heat generated by current passage through, the polarized slag layer i s transferred to the electrode at t h i s value of h, i t i s possible to c a l c u l a t e the current density needed to r a i s e the slag temperature at the i n t e r f a c e to i t s b o i l i n g point. The heat transfer equation i s q = h • A • AT c a l . sec. (4-31) where A i s surface area of t r a n s f e r , and AT i s the temperature d i f f e r e n c e across the Interface. By assuming that the slag i s at i t s b o i l i n g point and the metal (Fe) i s at i t s melting point. AT = B.P._ _ - M.P._, CaF„ Fe = 2509°C - 1538°C 1 = 971°C Therefore, f o r these assumptions, q = 1 x 11.3 x 971 = 11,000 c a l . s e c . " 1 If the voltage drop i n the p o l a r i z e d slag layer i s assumed to be 10 V. the power d i s s i p a t i o n i s given by: P = 101 watts where I Is the current i n amps. M u l t i p l y i n g P i n kW by the f a c t o r 0.24 converts P to k e a l . sec. 1 Therefore the current can be calculated as follows: 135 11 k c a l . sec. -1 11 .24 kW. = 45.9 kW. Therefore I = 45.9 - = 4.59 kA. 10 = 4590 A. The electrode current density would then be = 400 A. cm. -2 This i s an unreasonably high current density, even i n view of the i n i t i a l assumptions and one can therefore conclude that arc i n i t i a t i o n by t h i s means i s very u n l i k e l y . the measured p o l a r i z a t i o n through i t s e f f e c t on the Fe /Fe r a t i o of the slag at the slag/gas i n t e r f a c e . When gas bubbles are evolved, the i n -crease i n p o l a r i z a t i o n overvoltage observed i s due to the increase i n ohmic resistance i n the region of the working electrode surface. Although the calcium f l u o r i d e used contained only a small quantity of oxide (250 ppm), and extensive precautions were taken to exclude oxygen from the sys-tem, there were no in d i c a t i o n s that f l u o r i d e species were involved i n the p o l a r i z a t i o n reactions i n "pure" calcium f l u o r i d e , as evidenced by changes i n , f o r example, the anodic p o l a r i z a t i o n times. It i s probable that the low. l e v e l s of oxygen present i n the f l u o r i d e were s u f f i c i e n t to provide a saturated layer on the working electrode since oxide s o l u b i l i t y i n "pure" calcium f l u o r i d e i s very small. The observations made on the cathodic process i n d i c a t e that;the arc i n i t i a t i o n condition must be associated with the production of Ca, v . The evolution of gas on the working electrode w i l l only a l t e r 3+ 2+ 136 In CaF^ + CaO slags, t h i s i s evidently the only possible mechanism, and should occur at electrode p o t e n t i a l s of approximately 1 v o l t with, respect to the graphite reference electrode. In CaF^ + Al^O^ slags, the mechanism should remain the same, but the observed p o t e n t i a l s at arc i n i t i a t i o n w i l l be markedly d i f f e r e n t from the CaF^ + CaO case due to the e f f e c t of A^O^ on the CaO a c t i v i t y . 4.11 E l e c t r o s l a g process p o l a r i z a t i o n The values of anodic and cathodic p o l a r i z a t i o n on melting e l e c -trodes i n the melting unit given i n Figures . 23-+26 agree w e l l with the ex-trapolated small scale p o l a r i z a t i o n r e s u l t s . One can therefore conclude that neither the obviously d i f f e r e n t hydrodynamic regime at the ESR e l e c -trode nor the f a c t that the electrode i s continuously melting have a s i g -n i f i c a n t e f f e c t on the p o l a r i z a t i o n r e a c t i o n s . We may also conclude that the s u b s t a n t i a l (8%) 360 Hz r i p p l e present i n the D.C. r e c t i f i e r c i r c u i t has no detectable e f f e c t on the re a c t i o n s . The few anodic and cathodic overpotential values obtained on melting AISI 1095 and AISI 430 s t e e l electrodes are approximately the same as those measured on pure i r o n e l e c -trodes at equivalent current d e n s i t i e s . This indicates that the a l l o y i n g elements had no appreciable depolarizing e f f e c t and that our findings on pure i r o n electrodes apply to a l l iron-based materials. CHAPTER 5 DISCUSSION OF MELT PROGRAM RESULTS 5.1 Introduction The purpose of the melt program was to investigate the chemical and thermal e f f e c t s which occur during D.C. e l e c t r o s l a g melting, and ex-p l a i n these e f f e c t s i n terms of the Faradaic reaction mechanisms proposed i n Chapter 4. The more important findings of the melt program are shown to be: 1. Chemical e f f e c t s e x i s t during D.C. e l e c t r o -slag melting. In the case of pure metals, oxidation occurs at the anodic surface to produce ingots with a f i n a l oxygen content which i s dependent p r i n c i p a l l y on the electrode p o l a r i t y . Deposition of A l occurs at the cathode surface and the Al-0 i n t e r a c t i o n i s p a r t i a l l y responsible f o r oxygen removal from the ingot pool. 2. Thermal e f f e c t s are present during D.C. el e c t r o s l a g melting. They apparently a r i s e from excess heat generation i n the pola r i z e d slag layer at the l i q u i d m e t a l / l i q u i d slag i n t e r f a c e s . These e f f e c t s r e s u l t not i n a higher slag temperature but, i n the case of an anodic electrode, In a higher s p e c i f i c melt rate and, i n the case of an anodic ingot, i n a larger l i q u i d metal volume i n the ingot pool. 137 138 3. In the ,case of D.C. e l e c t r o s l a g melting of a l l o y s , the chemical e f f e c t s r e s u l t i n s i g n i f i c a n t losses of e a s i l y oxidizable a l l o y i n g elements. A l l o y depolariza-t i o n i s small. 4. A l l of these e f f e c t s can be explained i n terms of both Faradaic reaction mechanisms and mass transfer phenomena at the two metal/slag i n t e r f a c e s . 5.2 E f f e c t of electrode p o l a r i t y on oxygen content The oxygen analysis r e s u l t s of the ingots made i n the melt program show that ingots made on the D.C. electrode negative mode have a much higher f i n a l oxygen content than ingots made on the electrode p o s i t i v e mode, inde-pendent of atmospheric i n t e r a c t i o n s . This f a c t i s best i l l u s t r a t e d by ex-amining the melt record r e s u l t s of Ferrovac-E ingots (Table III) i n which the average oxygen content of electrode negative ingots melted through CaF^ + 25 wt. % Al^O^ slag i s 480 ppm. 0, while ingots made on the electrode p o s i t i v e mode i n the same slag have an average oxygen content of 185 ppm. 0. As shown i n the melt records, there are a c t u a l l y three types of electrode p o s i t i v e melting possible. The f i r s t type i s the standard electrode p o s i t i v e mode i n which the maximum rate of power input i s c o n t r o l l e d by the voltage at which arcing to the "in s u l a t e d " mold; occurs (~19 V.). To operate at a higher voltage, i t was necessary to paint the inner surface of the copper mold with boron-nitride paint as described i n Chapter 3. In th i s type of electrode p o s i t i v e melting [+(im)J i t was possible to melt at higher applied -voltages without arcing to the mold (see ingots 37, 38, and 41). These i n -139 gots had an average oxygen content of 140 ppm. 0 which i s only s l i g h t l y lower than that of the normal electrode p o s i t i v e (+) ingots. Two ingots were made i n the t h i r d type of electrode p o s i t i v e mode which i s electrode p o s i t i v e with a l i v e mold, I+(lm)]. In this mode the mold was purposely made to carry current being connected to the power return c i r c u i t with i t s own leads and current shunt. I t was found that i n this mode approximately 90% of the process current passed through the l i v e mold c i r c u i t , and ingots 35 and 83 made t h i s way i n a CaF^ + 25 wt. % k l ^ ) ^ slag had the higher oxygen content (760 ppm. 0). Three ingots made using A.C. power (34, 36 and 40) had an average oxygen content of 215 ppm. 0. As shown i n Table I I I , ingot 36 was made using a l i v e mold connection, while ingot 40 was made i n an insulated mold, but these conditions appeared not to a f f e c t .the f i n a l oxygen content to any extent. This observed dependence of the oxygen content on the electrode p o l a r i t y can be explained using the findings of the small scale studies. Consider f i r s t the case of the ingot being made i n the electrode negative mode. The electrode i s cathodic and the ingot anodic such that A l and Ca w i l l be deposited at the melting electrode t i p . The Ca w i l l dissolve i n the slag and A l w i l l dissolve at a steady-state rate i n t o the l i q u i d i r o n f i l m on the electrode t i p . Therefore, the metal droplets which detach from the t i p w i l l contain dissolved aluminum which w i l l be c a r r i e d to the l i q u i d metal pool ofthe s o l i d i f y i n g ingot. In the U.B.C. e l e c t r o s l a g r i g , the current density at an anodic ingot t i p i s high enough to saturate the slag at the i n t e r f a c e i n FeO, and i r o n oxide w i l l be d i s s o l v i n g continuously into the l i q u i d metal pool. Depending on the r e l a t i v e concentrations of 140 A l and 0 i n the l i q u i d , o x i d e i n c l u s i o n s w i l l nucleate and grow as s o l i d i -f i c a t i o n proceeds. The i n c l u s i o n s that are trapped by the s o l i d i f y i n g i n t e r f a c e represent the f i n a l oxygen content of the ingot metal. This i s a very s i m p l i s t i c view of i n c l u s i o n formation and entrapment i n which no consideration has been given to the problems of i n c l u s i o n nucleation or A l loss from the drops as they pass through the iron-oxide saturated anodic i n t e r f a c e . Ingots made i n the electrode p o s i t i v e [+] or electrode p o s i t i v e with an insulated mold {+(im)] mode w i l l pick up A l and 0 by the same means with the exception that the electrode i s anodic and the ingot cathodic. Why, then, do electrode negative ingots have a higher f i n a l oxygen Content? This can be explained by considering the l o c a t i o n of the anodic reaction s i t e s i n the two cases. The ingot current density i s approximately one-h a l f that of the electrode t i p current density, but the operating current i n the U.B.C. e l e c t r o s l a g r i g i s s u f f i c i e n t l y high to saturate an anodic ingot surface i n FeO. In electrochemical terms an anodic ingot surface i s equivalent to an anodic electrode t i p , and the observed dependence of f i n a l ingot oxygen content on electrode p o l a r i t y must therefore a r i s e from d i f f e r e n -ces i n the rate of metal oxidation at the two possible anodic surfaces. Anodic l i q u i d i r o n surfaces are exposed to FeO saturated slag which repre-sents an unlimited supply of oxidant, and the rate of oxygen d i s s o l u t i o n must be c o n t r o l l e d by the l i q u i d metal flow c h a r a c t e r i s t i c s at these i n t e r -faces. A melting electrode has a l i q u i d f i l m on the surface which i s approx-imately 20.0u-thick (4). As w i l l be shown i n Section (5.6), the amount of oxygen which dissolves i n the melting electrode metal i s equal to 1600 ppm. 141 0 f o r pure i r o n electrodes. It was not possible to ca l c u l a t e an equiva-lent value f o r the amount of oxygen d i s s o l u t i o n at an anodic l i q u i d ingot surface, because too l i t t l e i s known about the f l u i d flow conditions i n the ingot pool. However, due to the larger surface area at the pool, one can speculate that more oxygen w i l l be dissolved at an anodic ingot. One must now consider how t h i s oxygen i s removed from the l i q u i d metal, because the observed f i n a l ingot oxygen contents are much lower than the calculated oxygen content. In the case of the anodic electrode, the metal droplets could conceivably lose some of t h e i r dissolved oxygen during the time the drops f a l l through the slag , e s p e c i a l l y when one considers the f a c t that oxygen i s surface active i n i r o n and the drop surface therefore presents a large amount of the dissolved oxygen to the r e f i n i n g action of the s l a g . Also when the drops reach the cathodic ingot surface, they must pass through a layer of strongly deoxidizing slag containing metal produced by the ca-thode reaction. It i s therefore reasonable to expect that the bulk of the oxygen dissolved i n a melting anodic electrode w i l l be removed from the metal before the metal becomes part of the ingot pool proper. In the case of an anodic ingot pool which i s i n contact with an FeO saturated s l a g , the only supply of deoxidant i s that which i s transported to the pool by the metal drops. Then drops w i l l contain electrochemically produced aluminum which w i l l r e a c t with the dissolved oxygen i n the ingot metal to form i n -clusions, some of which w i l l d i s s o l v e i n the sla g . The dependence of the f i n a l ingot oxygen content on electrode p o l a r i t y can therefore be explained i n terms of differences i n oxygen d i s s o l u t i o n rates at the ingot and e l e c -trode and the chemical i n t e r a c t i o n between the electrochemically produced products. 142 The f a c t that l i v e mold electrode p o s i t i v e ingots have the high-est oxygen content of any mode can be explained i f we again consider the electrochemical r e a c t i o n s i t e s . The anodic surface i s the electrode t i p which behaves i n the same manner as the anodic surface i n any other mode by d i s s o l v i n g oxygen i n the melting metal. In t h i s mode, the bulk of the current flows to the mold and the path of lowest resistance should be a r a d i a l one, orthogonal to the electrode t i p . I t seems unreasonable to propose that the process current passes through the s o l i d i f i e d slag s k i n to the Cu mold. It would be more p l a u s i b l e to have the current transferred through a "soft a rc," (section 4.10) established between the l i q u i d slag and the copper mold. As discussed i n Section (1.3.3) such an arc might con-duct both e l e c t r o n i c a l l y and i o n i c a l l y , with the p r i n c i p a l Faradaic s i t e as the arc/slag i n t e r f a c e . Hence the ingot/slag i n t e r f a c e has l i t t l e cathodic deoxidizing c a p a b i l i t y and thus w i l l not s u b s t a n t i a l l y a l t e r the electrode metal composition. We would expect then that the ingot pool contain large amounts of dissolved oxygen. 5.3 Inclusion types As previously indicated i n Chapter 1, e l e c t r o s l a g processing en-sures that no large inclusions are present i n the ingot metal. Electron microprobe examination of ingot specimens produced i n the present melt program show th i s to be the case. F i g . 39 shows the appropriate electron microprobe pictures obtained upon examination of ingot no. 27 (FVE. el.-ve) while the pictures i n F i g . 40 were obtained from the AISI 430 s t a i n l e s s s t e e l ingot (No. 61). Both ingots contained small round i n c l u s i o n s of high A.E.I. X 1200 F i g u r e 39. S e c t i o n from i n t e r i o r of F e r r o v a c - E e l e c t r o d e n e g a t i v e i n g o t showing A l c o n t a i n i n g i n c l u s i o n s . A.E.I. X 2100 Al X-ray X 2 I 0 0 Figure 40. Section from i n t e r i o r of AISI 430 s t a i n l e s s s t e e l ingot showing A l containing i n c l u s i o n s . 145 aluminum content. The same type of inc l u s i o n s were found i n most of the ingots examined whether they were made from AISI 1095 s t e e l or e s s e n t i a l l y pure n i c k e l . Inclusion formation and composition during e l e c t r o s l a g pro-cessing w i l l not be discussed here, having been adequately covered by B e l l (21). Evidence of i n c l u s i o n removal by f l o t a t i o n was found upon examina-t i o n of the top surface of ingot 30 as shown i n F i g . 41. This shows a large number of alumina inclusions which were about to enter the s l a g . These i n -clusions were la r g e r than those found i n the bulk of the ingot and they ; appear to have formed-by agglomeration of smaller i n c l u s i o n s . This produced i n c l u s i o n s with a larger e f f e c t i v e radius which were able to f l o a t out of the ingot pool. This observation indicates that the measured oxygen content of the ingots i s lower than the electrochemically produced oxygen content. It must be noted that the f i n a l ingot oxygen and aluminum contents i n pure i r o n were f a r higher than would be produced by e q u i l i b r a t i o n of i r o n with alumina. This i s shown to be true i n F i g . 42 which i s a p l o t of exper-imentally determined A l and 0 contents of pure i r o n i n equilibrium with ; alumina of unit a c t i v i t y at 1600°C (21). The r e s i d u a l A l and 0 contents of ingots 25, 41, and 83 are shown to l i e w e l l above the equilibrium con-tents. This supports the argument that the electrochemical re a c t i o n pro-ducts are responsible for the observed ingot i n c l u s i o n s . With respect to removal of inclusions present i n the s t a r t i n g electrode metal, i t has been noted (39) that the electrode t i p i s the s i t e at which these i n c l u s i o n s are most e f f i c i e n t l y removed. As discussed pre-v i o u s l y , t h i s Is not the case with respect to removal of electrochemically formed i n c l u s i o n s . 0 p t i c a l X 9 5 0 F i g u r e 41. Alumina i n c l u s i o n s on the top o f a F e r r o v a c - E e l e c t r o d e n e g a t i v e i n g o t . o Figure 42. A l and 0 i n l i q u i d i r o n i n equilibrium with unit a c t i v i t y A1„0 at 1600°C. Experimental data f a l l s within the hatched area. 148 5.4 E f f e c t of s p e c i f i c coulombic density (Z) on the f i n a l oxygen content As presented i n Chapter 3, i t was postulated that the f i n a l i n -got oxygen content could be correlated to a parameter, Z, which took into account the operating current density- and the melt rate. There may indeed e x i s t such a r e l a t i o n s h i p , but the range of stable operating current densi-t i e s a v a i l a b l e on the U.B.C. e l e c t r o s l a g r i g i s l i m i t e d . Hence, we were un-able to vary Z over a s u f f i c i e n t l y wide range f o r the postulate to be exam-ined with respect to the f i n a l ingot oxygen content. Despite the limited, amount of data, a trend does appear to e x i s t . If one examines the values of Z and oxygen content for ingots 31->33 i n Table III ( e l . +ve.) , i t can be • seen that the lowest oxygen content corresponds to the lowest value of Z, but t h i s may simply be a r e f l e c t i o n of the higher melt rate of ingot 31. Examination of the data for the electrode negative ingots i n Table III shows the same trend i n that ingot 77 has the lowest oxygen content as w e l l as the lowest value of Z and the highest melt rate. However, although the approach appears e n t i r e l y reasonable, the range of accessible experimental data i s i n s u f f i c i e n t to adequately test our i n i t i a l postulate. 5.5 E f f e c t of atmosphere control Two types of atmospheric control were used during the melt pro-gram. The argon fume hood produced a r e l a t i v e l y o x i d i z i n g atmosphere (up to 1% whereas the argon gas cap atmosphere pu r i t y depended, s o l e l y on the p u r i t y of the argon i n use. Etienne (4) states that the atmospheric oxidation r a t e during melting i s dependent upon exchange reactions between slag and atmosphere which i n turn are c o n t r o l l e d by d i f f u s i o n of multiple 2+ 3+ valence ions (Fe /Fe ) i n the s l a g . He also concluded that electrode 149 oxidation above the slag was unimportant as a means of atmosphere oxygen transport to the melting metal. His f i n a l conclusion i s that the p a r t i a l pressure of oxygen i n the atmosphere bears l i t t l e r e l a t i o n to the amount : of atmospheric oxidation except when the actual r a t e of supply of atmos-pheric oxygen i s less than the l i m i t i n g rate of oxygen transfer at the slag/atmosphere i n t e r f a c e . If one compares the oxygen contents of Ferro-vac-E ingots made In the argon fume hood (23, 72, 24, and 74)with those of the Ferrovac-E ingots made i n the argon gas cap \. (25-K30, and 31+33) i t i s obvious that the d i f f e r e n t atmospheres have had very l i t t l e e f f e c t on the oxygen content i n either the electrode negative mode or the electrode p o s i t i v e mode. One must therefore conclude that the ingot oxygen contents are a r e s u l t of metal oxidation by the FeO-saturated slag layer produced by electrochemical reactions at the anode. The o x i d i z i n g power of t h i s anodic slag layer i s much greater than any atmospheric oxidation pro-cesses which might have existed during melting i n the U.B.C. e l e c t r o s l a g r i g . 5.6 D i f f u s i o n of oxygen into an anodic electrode As previously discussed (5.2) a melting anodic i r o n electrode i s exposed to slag saturated i n FeO. Oxygen dissolves i n the l i q u i d electrode f i l m to e s t a b l i s h a steady-state rate of oxygen transfer to the melting metal. Etienne (4) developed a mathematical model by consid-150 irig an i d e a l i z e d s i t u a t i o n i n which, the l i q u i d metal f i l m flows down a con i c a l electrode t i p under the Influence of gr a v i t y , neglecting the ef f e c t s of surface tension and momentum transfer from the slag to the l i q u i d metal f i l m . The t o t a l weight of oxygen Q q transferred per un i t time across the slag/metal i n t e r f a c e i s the same as the flow rate of oxygen at the system e x i t (electrode t i p ) which may be calculated by integr a t i o n of the d i f f u s i o n flow equation over the e n t i r e area of the electrode cone, A. The fl u x of oxygen can then be wirtten as: Q = 2 A[0], / gm. s e c . " 1 (5-1) o I V ift e using A = cos 6 [0]_£ - concentration of 0 i n Fe at slag metal i n t e r f a c e , = 0.2 wt. % (saturation of Fe at 1550°C). -3 = 0.014 gm. cm. d i f f u s i o n c o e f f i c i e n t of 0 i n Fe D o ,m = 3 • 10" 4 cm.2 sec. 1 (40) It has been assumed that the d i f f u s i o n boundary layer i s small compared to the f i l m thickness and the s i t u a t i o n therefore reduces to d i f f u s i o n i n a s e m i - i n f i n i t e medium. The exposure time, t ^ , of a surface element t r a v e l l i n g between the base of a cone and the t i p i s given by: 27rcos9 ,2/3/ . ym \ 1/3 / R \ 5/3 T(5/6) T(l/3) e v 3W ' 1 \p g s i n e / V cos 6 / . r(7/6) ^ l ) m ° where 151 -1 W - melt rate i n gm. sec. m = 1.94 gm. sec. ^  for electrode p o s i t i v e 3 -1 W ' - t o t a l volumetric melt rate i n cm. sec m W m _ 3 —1 = — = 0.277 cm. sec. P p = p - p - e f f e c t i v e density of l i q u i d i r o n due to slag buoyancy -3 = 7.0— 2.6 = 4.4 gm. cm. y - v i s c o s i t y of molten i r o n m = 0.05 poise. R - electrode radius = 1.59 cm. 6 - base angle of cone = 45° (average of electrode t i p s ) g - ac c e l e r a t i o n due to gra v i t y r (5/6) r q/3) w , . - „. •. . . , -( 1 j — - Eulerian functions which evolve from i n t e g r a t i o n of t = 3.259 and t = 0.976 sec. f o r the conditions stated above, e 2 A = 11.234 cm. and Q can now be calculated to be: o -3 -1 Q q = 3.10 x 10 gm. sec. The weigh-t % of oxygen transported Into the melting electrode metal i s then found by d i v i d i n g Q q by W , and Is equal to 0.16 wt. % 0 or 1600 ppm 0. This f i g u r e of 1600 ppm 0 represents the t h e o r e t i c a l amount 152 of oxygen dissolved i n the metal during the melting of pure i r o n i n the electrode p o s i t i v e mode. Comparing t h i s with the observed f i n a l oxy-gen contents of pure i r o n electrode p o s i t i v e ingots (185 ppm 0), i t appears that the bulk of the dissolved oxygen i s l o s t between the drop detaching from the electrode t i p and metal s o l i d i f i c a t i o n i n the ingot. In sections (5.2) and (5.3) i t was shown that oxygen might be l o s t by d i s s o l u t i o n of oxide i n the slag during the time the drops f a l l through the slag and by i n c l u s i o n f l o a t a t i o n from the ingot pool. Ingot 83 (Table III) was made i n the electrode p o s i t i v e l i v e mold mode [+(lm)] and ex-amination of the sla g cap showed that droplets of i r o n were retained i n the slag cap. The larger droplets were analyzed f o r t h e i r t o t a l oxygen content and were found to contain 1750 ppm 0. This f i g u r e i s close to,the t h e o r e t i c a l l y calculated oxygen content of 1600 ppm 0. These drops had been exposed to the slag for a f i n i t e length of time but s t i l l retained a high oxygen content. I t therefore appears that oxide loss during drop f a l l through the slag i s n e g l i g i b l e and that i n c l u s i o n f l o t a t i o n and r e -moval from the ingot pool must account f o r the removal of the e l e c t r o -chemically produced oxygen. In order for aluminum containing inclusions to nucleate homo-geneously i n the ingot pool, there must be high supersaturation of both aluminum and oxygen i n the metal. This Is In contrast with B e l l ' s f i n d -ings (21). He found that ingots made with A.C. power had low supersatur-ations of aluminum and oxygen and incl u s i o n s were nucleated only at the freezi n g ingot i n t e r f a c e . Inclusion f l o t a t i o n i n D.C. ingots i s "short range" and occurs i n the ingot pool near the slag/metal i n t e r f a c e because 153 this i n t e r f a c e i s the electrochemical r e a c t i o n surface at which the high supersaturations are produced. For t h i s reason i n c l u s i o n f l o t a t i o n i s probably independent of convective motion i n the pool. 5.7 Drop s i z e and surface tension The drop s i z e and surface tension data presented i n Table VI, shows a marked dependence of drop s i z e on the electrode p o l a r i t y during melting. These differences must a r i s e from the e f f e c t of oxygen on the i n t e r f a c i a l tension at the slag/metal i n t e r f a c e . Drops produced at e l e c -trode p o s i t i v e electrodes are the smallest because they have a high oxy-gen content a r i s i n g by the mechanism of anodic.saturation of the electrode t i p i n FeO. Drops from cathodic electrode are larger because they have no appreciable oxygen content and i n f a c t should be deoxidized by the d i s s -solved aluminum produced by the cathodic processes taking place. A.C. electrode drops are the l a r g e s t , even larger than electrode negative drops, despite the f a c t that A.C. slags have a higher bulk concentration of FeO (Table X) which should decrease the i n t e r f a c i a l tension at the : drop/slag i n t e r f a c e . This d i s p a r i t y may simply a r i s e from the cathodic electrode containing aluminum produced by Faradaic processes. In general these findings support our i n t e r p r e t a t i o n of the electrochemical phenomena occurring at anodic and cathodic electrode t i p s , and they agree with the findings of Whittaker (5). Our estimated values of i n t e r f a c i a l tension are not unreasonably d i f f e r e n t from the value of approximately 80Q dynes. -2 cm. found by Yokabashvili et. a l (41) for the i n t e r f a c i a l tension be-tween CaF 9 + 25 % Al-O. slag and mild s t e e l . 154 5.8 S i g n i f i c a n c e of i r o n i n the slag caps The r e s u l t s presented i n Table X,show that slags used during e l e c t r o s l a g processing pick up i r o n i n the form of iron-oxide. There i s no detectable i r o n i n these slags before they are used^therefore the observed i r o n contents a r i s e by chemical and electrochemical means during remelting. In the electrode negative mode (ingot 39), the bulk of the slag cap i s very clean i n appearance, with only a very t h i n layer of FeO containing slag at the ingot pool/slag i n t e r f a c e . Both electrode p o s i -t i v e slag caps (ingots 31 and 37) have a s u b s t a n t i a l l y thicker layer of FeO containing slag at the ingot pool/slag i n t e r f a c e . These v i s u a l ob-servations of the slag caps agree with the analysis shown i n Table X. In order to consider the p o s s i b i l i t y of simple chemical forma-t i o n of FeO i n the s l a g , one must study the thermodynamics of the Fe-A^O^ system. Three possible reactions which may produce FeO i n the slag are: 3 F e ( l ) + A 1 2 ° 3 ( s ) * 3 F e ° ( l ) + 2 [ A 1 } F e ( 5 " 3 )  2 F e ( l ) + A 12°3(s) * 2 F e ° ( l ) + A 1 2 ° ( g ) ( 5 " 4 ) F e ( l ) + A 12°3(s) * F e ° ( l ) + 2 A 1 0 ( g ) < ( 5" 5 ) These reactions, however, are a l l highly endothermic and hence would produce no FeO during e l e c t r o s l a g melting. Therefore any FeO occur-r i n g i n the slag must have an electrochemical o r i g i n . The FeO produced at the anode represents only one-half of the o v e r a l l electrochemical reac-t i o n , and there must also be an equivalent amount of electrochemically produced calcium and aluminum dissolved i n the slag during melting. Various electrode t i p s showing o x i d e p r e s e n t on an a n o d i c i r o n e l e c t r o d e t i p . 156 One can envisage a steady-state rate of production of both anodes and cathodic components, the bulk of which d i s s o l v e i n the slag and con-tinuously back-react i n the slag as shown below: (Ca and Al) . + (FeO) , -> (CaO and Al o0„) + Fe, (5-6) slag slag z. 3 1 Thus steady-state amounts of the electrochemical products are present i n the t o t a l slag cap. When the current i s interrupted and e l e c -trochemical reactions cease, the back-reaction w i l l continue presumably u n t i l the slag caps s o l i d i f i e s and a r e s i d u a l FeO content i s retained i n the s l a g . An equivalent amount of r e s i d u a l Ca and A l must also be r e -+ 2+ tained and, i n the case of Ca, w i l l be present as Ca or Ca 2 ions i n the s o l i d i f i e d s l a g . Presently a v a i l a b l e a n a l y t i c a l methods are inade-quate to detect t h i s amount without ambiguities. One must also consider why the FeO containing slag i n an e l e c -trode p o s i t i o n slag cap i s found near the ingot/slag i n t e r f a c e . This i s not immediately c l e a r , but i t must a r i s e from density e f f e c t s i n the slag cap since the FeO was produced i n the electrode region. F i g . 43 shows pictures of clean electrode t i p s produced by melting i n the electrode negative or A.C. mode and a heavily oxidized electrode t i p produced by melting i n the electrode p o s i t i v e mode. This lends support to the pro-posed model requiring FeO saturation at an anodic surface. 5.9' Melt rate and heat generation An important e f f e c t associated with electrode p o l a r i z a t i o n i s heat generation at the electrode t i p . The average s p e c i f i c power consump-t i o n when melting Ferrovac-E electrodes i n CaF 0 + 25 wt. % A l 0, slags i s 157 12.6 kWs. • gm. i n the electrode negative mode and 10.6 kWs. • gm. i n the electrode p o s i t i v e mode. We know from our small scale studies that at the operating electrode and ingot current d e n s i t i e s , an anodic surface i s po l a r i z e d to a greater degree than a cathodic surface. The heat generation due to current passage through these anodic and cathodic p o l a r i z a t i o n resistances i s therefore greater at an anodic electrode, and i t i s evident that the heat t r a n s f e r pattern i n the electrode s l a g / i n t e r -face region i s such that the heat generated i n the polarized slag layers i s transferred e f f i c i e n t l y into the electrode melting process. The e f f e c t i s thus seen i n s p e c i f i c melting e f f i c i e n c y and not d i r e c t l y i n a sla g temperature d i s t r i b u t i o n . S i m i l a r l y , the heat generation at an anodic ingot surface manifests i t s e l f by producing a greater l i q u i d metal volume than i s observed i n a cathodic ingot. It was also noted that A.C. power produced the most e f f i c i e n t -1 melting with a s p e c i f i c power consumption of 7.7 kWs. • gm. If p o l a r i -zation e f f e c t s d i d not e x i s t , then there i s no reason why there should be any differences i n the melting e f f i c i e n c y between A.C. and D.C. melting. This d i s p a r i t y can be explained by using an electrochemical argument. The cathodic products during D.C. melting are calcium and aluminum, both of which are soluble i n CaF2 slags, and s u b s t a n t i a l l y increase the s l a g , conductivity. J oshi (32) found a 35% increase i n slag conductivity when melting with D.C. power as compared to A.C. power. With a higher conduc-t i v i t y s l a g , heat generation i s decreased and the melting e f f i c i e n c y de-creases. This i s i n agreement w i t h the melting e f f i c i e n c i e s observed i n thi s study. 1 58 5.10 Calcium oxide slags and ingot porosity The i n s o l u b i l i t y of calcium i n i r o n was quite apparent when Armco i r o n ingots were made i n CaF^ + CaO slags. The cathodic e l e c t r o -chemical product i s calcium which dissolves i n the slag while, i n the electrode negative mode, the ingot i s e f f i c i e n t l y oxidized. The s o l i d i -f y i ng Ingot metal therefore had a high oxygen content but no dissolved aluminum and the r e s u l t was carbon-monoxide blow hole formation by the. carbon-oxygen r e a c t i o n : [ C ] F e + [ 0 ] F e " C O g ( 5 " 7 ) The carbon content of Armco i r o n i s .012 wt. % and was s u f f i -cient to produce ingots of high porosity. In the case of ingot 52 melted i n the electrode negative mode i n a 25 wt. % CaO s l a g , bubble formation was so d r a s t i c that i r o n bubbles would form on the top of the l i q u i d pool and r i s e up to contact the melting electrode. This i s ad-mittedly an extreme case, but such conditions cannot be tolerated dur-ing commercial e l e c t r o s l a g processing. 5.11 E f f e c t of increasing ingot/electrode diameter r a t i o In an attempt to produce a wider v a r i a t i o n i n Z, ingots were prepared using the electrode/ingot diameter r a t i o as a v a r i a b l e . Ingots 75 and 76 were made using 3.2 cm. diameter Ferrovac-E electrodes but they were melted into a 7.6 cm. d i a . copper mold instead of the usual 5.5 cm. d i a . copper mold. Ingot 76 was made i n the electrode negative 159 mode and had a f i n a l average oxygen content of 548 ppm. 0 which i s some-what greater than the 480 ppm. 0 measured as the average oxygen content of 5.5 cm. ingots. Ingot 76 was made i n the electrode p o s i t i v e mode [+(lm)] and had an average oxygen content of 153 ppm. 0 which l i e s i n s i d e the range of oxygen contents measured f o r 5.5 cm. electrode p o s i t i v e i n -gots. The higher f i n a l oxygen content of ingot 75 must have a r i s e n i n part, because of the lower current density at the anodic ingot surface, a lower current density which s t i l l saturated the i n t e r f a c e i n FeO and produced more e f f i c i e n t net electrochemical oxidation of the i r o n . The higher current density at the cathodic electrode i n t e r f a c e may have r e -sulted i n less e f f i c i e n t aluminum transfer to the melting electrode and therefore less e f f i c i e n t deoxidati on of the ingot pool. These r e s u l t s would indi c a t e that the ingot pool/slag i n t e r f a c e i s the most important electrochemical r e a c t i o n s i t e during D.C. e l e c t r o s l a g melting. 5.12 E f f e c t of aluminum addition at the electrode during melting Experiments were c a r r i e d out i n which aluminum wire was attached to Ferrovac-E electrodes to make ingots 80 and 81. In these ingots, i f a l l the aluminum from the wire was dissolved i n the ir o n during melting, the aluminum content would be 2000 ppm. The aluminum analysis r e s u l t s given i n Table VLT show that there was e f f i c i e n t trans— f e r of aluminum into the ingot metal. These aluminum analysis figures are t o t a l aluminum fi g u r e s because both matrix aluminum and aluminum i n i n c l u s i o n s are counted by the neutron a c t i v a t i o n technique. The lower than normal t o t a l oxygen content of ingot 80 (259 ppm. 0 as compared to an average of 480 ppm. 0) indicates that the added aluminum was respon-160 s i b l e f o r a higher than normal r a t e of i n c l u s i o n formation and removal. The presence of the excess aluminum at the melting cathodic electrode increased the drop s i z e to 3.46 gm. as shown i n Table VI, a drop s i z e comparable to the A.C. drop s i z e . The e f f e c t of aluminum wire on the oxygen content of ingot 81 made i n the electrode p o s i t i v e mode i s un-expected i n view of our other observations. The f i n a l ingot oxygen content (576 ppm. 0) was much greater than the normal electrode p o s i -t i v e ingot oxygen content of 185 ppm. 0. One would have expected the aluminum to prevent excessive oxygen d i s s o l u t i o n i n the melting e l e c -trode metal, but this apparently does not occur even though the aluminum was transferred e f f i c i e n t l y to the ingot which had a f i n a l A l content of 2621 ppm. This agrees with the r e s u l t s given i n Table VI which show the drop s i z e of ingot 81 was not appreciably greater than the drop s i z e of a normal electrode p o s i t i v e ingot. These r e s u l t s could be explained on the basis that both oxygen and aluminum are transferred e f f i c i e n t l y into the ingot pool. The oxide inclusions subsequently p r e c i p i t a t i n g would be higher i n aluminum content than normal electrode p o s i t i v e i n -gots, and might be less e f f i c i e n t l y removed in t o the s l a g . Hence the net e f f e c t of adding aluminum i s to r e t a i n oxygen as ingot oxide p a r t i c l e s . 5.13 A l l o y losses during remelting 5.13.1 AISI 1095 s t e e l The ingot analysis r e s u l t s given i n Table VIII show that there were s i g n i f i c a n t losses i n carbon and s i l i c o n during remelting, hut no manganese l o s s . Ingot 67 melted i n the electrode p o s i t i v e mode l o s t 6% 161 of i t s i n i t i a l carbon but no s i l i c o n , while ingot 68 melted i n the e l e c -trode negative mode l o s t 7% of i t s carbon and 27% of i t s i n i t i a l s i l i c o n content. These a l l o y losses must be associated with the anodically produced layer of FeO containing slag at the anodic i n t e r f a c e whether th i s i n t e r f a c e be at the electrode or the ingot pool. In the case of ingot 67 the carbon loss should occur at the electrode. The lack of s i l i c o n loss i n th i s mode of melting can be explained i f one considers that a continuous supply of high carbon l i q u i d metal i s exposed to the oxid i z i n g i n t e r f a c e at such a rate that the carbon l e v e l w i l l never be depleted s u f f i c i e n t l y to permit s i l i c o n oxidation. Why then i s s i l i c o n loss experienced i n the electrode negative mode? The anodic ingot i s being oxidized during melting to such an extent that the carbon at the i n t e r f a c e i s s u f f i c i e n t l y depleted to allow s i l i c o n to oxidize. By assuming an ingot pool/slag i n t e r f a c e temperature of 1600°C, and neglec-t i n g the i n t e r a c t i o n e f f e c t s between carbon and s i l i c o n , i t can be shown that the equilibrium oxygen content of i r o n with 0.975 wt. % carbon i s approximately 20 ppm. 0, an oxygen l e v e l too low to oxidize s i l i c o n . S i l i c o n w i l l not oxidize at th i s temperature u n t i l the oxygen l e v e l i s above approximately 80 ppm. 0 which could not be reached under equilibrium conditions unless the carbon l e v e l i s depleted to approximately 0.2 wt. % at an i n t e r f a c e . Therefore s i l i c o n oxidation only occurred i n the e l e c -trode negative mode (anodic ingot) because the degree of mixing at the ingot pool/slag i n t e r f a c e was low. Hence the rate of carbon transport to t h i s i n t e r f a c e was i n s u f f i c i e n t to accommodate the rate of oxidation. In the electrode p o s i t i v e case, the rate, of carbon supply to the anodic 162 in t e r f a c e was s u f f i c i e n t l y high and no s i l i c o n losses were observed. The f a c t that no s i l i c o n containing inclusions were found i n these i n -gots tends to substantiate the idea that s i l i c o n losses took place at the ingot pool/slag i n t e r f a c e . The oxygen contents reported i n Table V are higher than both of the C-0 or Si-0 equilibrium values. This supports the proposal that i n c l u s i o n p r e c i p i t a t i o n i s taking place i n the ingot pool surface region and not i n the bulk of the pool. 5.13.2. AISI 430 s t a i n l e s s s t e e l The ingot analyses information given i n Table IX shows that chromium, silicon,- and sulphur were l o s t during the melting of AISI 430 s t a i n l e s s s t e e l , and that the t o t a l a l l o y losses were greatest i n the electrode negative mode. These observations again show that the anodic ingot surface i s capable of more e f f i c i e n t oxidation of a l l o y i n g e l e -ments than i s the anodic electrode t i p . The lower rate of metal trans-port through the ingot surface produces a longer exposure time to the oxi d i z i n g anodic slag i n t e r f a c e than i s encountered at an anodic electrode. A c a l c u l a t i o n following the l i n e s indicated i n Section (5.5) was done for the known melt rate and electrode diameter of t h i s material, and Q , the o -3 f l u x of oxygen i n t o an anodic electrode t i p was found to be 2.0 x 10 gm. sec. ^ I f i t i s assumed that an equivalent amount of chromium i s oxidized 3+ to the Cr state and dissolves i n the sla g , t h i s corresponds to a chrom-ium loss of 0.22 wt. % from the a l l o y , which i s close to the observed chromium loss of 0.19 wt. % i n the electrode p o s i t i v e mode. The s o l u b i l -i t y of oxygen i n pure Fe-Cr a l l o y s i s r e l a t i v e l y high, (~400 ppm at 1600°C i n Fe-20 wt. % Cr) but the presence of 0.26 wt. % S i i n the a l l o y decreases 1 63 the oxygen s o l u b i l i t y to 100 ppm. 0. In the absence of dissolved A l one would expect to f i n d S i containing i n c l u s i o n s i n t h i s m aterial. However t h i s was not observed, and the i n c l u s i o n s found could only have been formed i n a s i l i c o n - d e p l e t e d region. 5.14 Remelting of n i c k e l electrodes The melt record r e s u l t s of the two n i c k e l ingots are given i n Table V. Ingot 69 melted i n the electrode negative mode has a low oxy-gen l e v e l of 7 ppm., while ingot 71 melted i n the electrode p o s i t i v e i n -sulated mold mode has a r e l a t i v e l y high oxygen content of 326 ppm. 0. This i s contrary to the findings for pure i r o n ingots i n which electrode negative ingots had the higher oxygen content. The inclusions i n ingot 71 were found to be aluminum containing, whereas ingot 69 contained a , very large number of carbon-monoxide blow holes. The dependence of f i n a l oxygen content of these n i c k e l ingots appears to have arisen from a com-bination of chemical and electrochemical means. I t i s possible that the rate of aluminum deposition i n the cathodic electrode of ingot 69 was low, and the carbon-oxygen re a c t i o n i n the ingot pool removed most of the anodically produced oxygen. However, i n the electrode p o s i t i v e case, aluminum and possibly calcium were deposited i n the ingot pool more e f f i c i e n t l y to produce a s u f f i c i e n t concentration of deoxidizing elements which consumed the anodic oxygen to form i n c l u s i o n s . Host of these were trapped during ingot s o l i d i f i c a t i o n . The c o r o l l a r y of t h i s i s that i f ingots were made from n i c k e l with a n e g l i g i b l e carbon content, then the electrode negative ingot would have the higher oxygen content. 164 5.15 A.C. melting of pure i r o n The electrochemical phenomena occurring at the two slag/metal i n t e r f a c e s during A.C. melting are extremely d i f f i c u l t to study. The asymmetry of the anodic and cathodic p o l a r i z a t i o n curves on pure i r o n suggests a mechanism by which a degree of electrochemical r e c t i f i c a t i o n could occur. If t h i s was so, then there would be a net chemical e f f e c t . B e l l (21) measured oxygen contents i n pure i r o n A.C. ingots which were i n excess of the t h e o r e t i c a l l y predicted oxygen l e v e l that would be pro-duced by e q u i l i b r a t i o n with Al^O^ containing slags. He also measured a s u b s t a n t i a l D.C. current (5-^ -10 A.) i n the melting unit which did not a r i s e from the A.C. power source. It can therefore be concluded that electroche-mical r e c t i f i c a t i o n does occur during A.C. e l e c t r o s l a g processing of pure i r o n , and that t h i s D.C. current deposits oxygen i n the metal which gives r i s e to non-metallic i n c l u s i o n s . 5.16 Electrochemical phenomena i n commercial ESR It i s necessary to consider the findings of t h i s study i n r e l a t i o n to commercial D.C. ESR p r a c t i c e s . It i s apparent that the current density at the anodic surface, whether i t be the electrode or the ingot, should be maintained below the current density at which i n t e r f a c e saturation i n FeO occurs. In this way, metal oxidation during melting w i l l be kept to a minimum and a l l o y losses w i l l be less extreme. In view of the f a c t that the current densi-t i e s i n commercial ESR furnaces are lower than those used i n U.B.C. f u r -nace., t h i s condition should be s a t i s f i e d when using slags with a high 165 oxide content. The mechanism of drop formation on a commercial electrode i s one of multiple drop formation and i s therefore quite d i f f e r e n t from that i n a smaller furnace i n which drops form one at a time. I t i s not known what ef f e c t t h i s w i l l have on the rate at which the Faradaic r e a c t i o n products enter the melting metal. I t i s also d i f f i c u l t to define the e f f e c t i v e e l e c t r o a c t i v e surface area and any model developed f o r such an electrode t i p w i l l be very complicated. Convective movement i n the l i q u i d ingot pool w i l l be important with respect to i n c l u s i o n nucleation i n the l i q u i d metal and i n c l u s i o n removal by f l o t a t i o n . L i t t l e i s known about the flow patterns i n the l i q u i d pool and i t i s therefore unreasonable to speculate about such phe-nomena . CHAPTER 6 CONCLUSIONS The observations made i n the melt program can be interpreted on the basis of the anodic and cathodic Faradaic r e a c t i o n mechanisms proposed to explain the r e s u l t s of the small scale studies. The pre-dominant anodic r e a c t i o n at pure i r o n surfaces i n CaF^ + Al^O^ slags and CaF 2 + CaO slags i s the anodic corrosion of i r o n according to: 2+ Fe -y Fe + 2e At s u f f i c i e n t l y high current d e n s i t i e s the slag adjacent to the anodic 2+ in t e r f a c e becomes saturated i n Fe ions, which i s formally equivalent to i r o n oxide saturation. This saturation condition leads to oxidation of the electrode metal. The cathodic r e a c t i o n at i r o n surfaces i s pro-posed to be the deposition of aluminum at low current d e n s i t i e s and the codeposition of both aluminum and calcium at higher current d e n s i t i e s . Aluminum dissolves i n i r o n and chemical r e a c t i o n i n the ingot pool be-tween the electrochemically produced oxygen and aluminum produces alumina or aluminum containing i n c l u s i o n s i n the f i n a l ingot metal. These in-' elusions are t y p i c a l l y small (<20 microns) i n diameter. The proposed anodic and cathodic reactions have a high net r e v e r s i b i l i t y even at high current d e n s i t i e s , therefore the slag composition remains e s s e n t i a l l y un-changed during remelting. At equivalent electrode anodic and cathodic current densities i n the U.B.C. e l e c t r o s l a g furnace a higher degree of anodic p o l a r i z a t i o n 166 167 i s observed. The heat generation due to t h i s p o l a r i z a t i o n resistance manifests i t s e l f as a lower s p e c i f i c power consumption i n the electrode p o s i t i v e mode. At excessively high current d e n s i t i e s , an arc i s estab-l i s h e d at the electrode t i p and the melting conditions become extremely unstable due to arc heat generation at the electrode t i p . As has been observed by other workers, a l l o y losses of e a s i l y oxidized elements are very s u b s t a n t i a l during D.C. melting. Results from the present studies on the galvanostatic pulsing of i r o n a l l o y anodes demonstrate that such losses are associated with n e g l i g i b l e a l l o y depol-a r i z a t i o n . Thus the p o l a r i z e d anode behaves e s s e n t i a l l y as i f i t were a pure i r o n surface. I t appears, from t h i s work, that the ingot pool/slag i n t e r -face i s the most important reaction s i t e with respect to electrochemical oxidation and deoxidation reactions. Such reactions occur i n a manner e s s e n t i a l l y independent of the atmosphere. REFERENCES 1. M. Etienne and A. M i t c h e l l : Proc. Sec. Symp. on ESR technology, Part I I , Mellon I n s t i t u t e , Sept. 1969. 2. G.K. Bhat, J.B, Tobias, and R.L. Kennard: Proc. Sec. Symp. on ESR technology, Part I, Mellon I n s t i t u t e , Sept. 1969. 3. B.I. Medovar et a l : E l e c t r o s l a g Remelting, State S c i e n t i f i c and Techn. Publ. House of L i t e r a t u r e on Ferrous and Nonferrous Metallurgy, Moscow, 1963. 4. M. Etienne: Ph.D. Thesis, U.B.C., Oct. 1970. 5. D.A. Whittaker: Ph.D. Thesis, McMaster U n i v e r s i t y , 1968. 6. W. Holzgruber and E. Plockinger: Berg-u Huttenw. Monatshefte, 1968, Vol.113, pp. 83-93. 7. W. Holzgruber, K. Petersen, and P.E, Schnider: Trans. Int. Vac. Met. Conf., p.518, Am. Vac. S o c , 1968. 8. R. Roberts: Proc. Sec. Symp, on ESR technology, Part I I , Mellon Inst., 1970. 9. P.P. Evseev: Avtom. Svarka, 1967, V o l . 176, p.42. 10. R.W. Ure: J . Chem. Phys., 1957, Vol.26, pp. 1363-73. 11. J.S. Anderson: "Nonstoichiometric Compounds," Am. Chem. S o c , Wash. D.C, 1963, pp. 1-22. 12. M. Blander: "Molten Salt Chemistry", Interscience, pp. 367-421, 1964. 13. B.R. Sundheim: "Fused S a l t s " , McGraw-Hill Co., 1964, p.272. 14. K.J. Vetter: "Electrochemical K i n e t i c s " , Academic Press, New York, 1967. 15. B.J. Welch and N.E. Richards: "Extractive Metallurgy of Aluminum" V o l . 2, pp. 15-30, Interscience, New York, 1963. 15a. A.F, Alabyshey, M.F. Lantratov and A.G. Morachevskii: "Reference Electrodes f o r Fused S a l t s , " Sigma Press, 1965. 16. A. Gosh and T.B. King: Trans. TMS-AIME, 1969, Vol. 245, pp.145-52. 17. P. Delahay: "New Instrumental Methods i n Electrochemistry", p.350 f f . , Interscience, New York, 1965. 168 169 18. A. M i t c h e l l and B. Burel: J.I.S.I. 1970, Vol.208, p.407. 19. J . Cameron, M. Etienne, and A. M i t c h e l l : Met. Trans., J u l y 1970, Vol. 1, pp. 1839-1844. 20. J . Campbell: J . Metals, July 1970, Vol. 22, No.7, pp.23-35. 21. M. B e l l ; Master's Thesis, U.B.C., J u l y , 1971. 22. G.P. Smith; "Molten Salt Chemistry", p.427, Interscience, New York, 1964. 23. V.D. Shantarinj O.A. Esin and V.N. Boronenkov: Elektrochlmiya, 1967, Vol. 3(6), pp.775-79. 24. G.A. Toporushev and O.A. Esin: Izv. Vysshikh Uchebnykh Zavedince, Vol.6, 1964, pp.21-27. 25. S. P i z z i n i , R. M o r l o t t i and E. Romer: Euratom Report, EUR 804 e, 1965. 26. H.A. Laitenen, C H . L i u and W,S. Ferguson: Anal. Chem., 1960, Vol. 30, p.698. 27. V.G. Levich: "Physiochemical Hydrodynamics", P r e n t i c e - H a l l , 1962. 28. M. Stewart: Ph.D. Thesis, U.B.C., Dec. 1970. 29. P.N. Smith and M.W. Davies: Trans. Section C, Inst. Min. Met., Vol.80, 1971, pp.C87-92. 30. D.A.R. Kay, A.M i t c h e l l and M. Ram: J.I.S.I., Vol. 208, 1970, pp. 141-46. 31. J.F. E l l i o t , M. G l e i s s e r and V. Ramakrishna:"Thermochemistry for Steel Making", Vol.2, Addison-Wesley. 32. S. Jo s h i : Ph.D. Thesis, U.B.C., Aug. 1971. 33. B. Burel: Master's Thesis, U.B.C. June 1969. 34. D.L. Sponsellor and R.A. F l i n n : Trans. AIME, Vol. 230, June 1964, p. 776. 35. P.A. D i s t i n , G.D. H a l l e t t and F.D. Richardson: J.I.S.I., Vol.206, p.821. 36. E, Manthey and E. Conzelmann: Z, fur das Elektrochemie, Vol. 32(7), 1962, pp. 330-36. 37. H.H. Kell o g : J . Electrochem. S o c , Vol. 97(4), 1950, pp. 131-41. 38. P. Drossbach: Z. fur das Elektrochemie, Vol. 56(1), 1951, pp.38-41. 170 39. M.M. Klyuev and Yu. M. Mironov: Electrometallurgy, V o l . 6, 1967, pp. 480-83. 40. L.S. Darken and R.W. Gurry: "Physical Chemistry of Metals", McGraw H i l l , 1953. 41. S.B. Yakabashvili and I . I . Frumin: Avt. Svarka., No.10, 1962, pp. 41-45. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0079201/manifest

Comment

Related Items