UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Non-metallic inclusions in electroslag refined ingots Reyes-Carmona, Fidel 1983

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

Item Metadata

Download

Media
831-UBC_1983_A1 R49.pdf [ 23.64MB ]
Metadata
JSON: 831-1.0078752.json
JSON-LD: 831-1.0078752-ld.json
RDF/XML (Pretty): 831-1.0078752-rdf.xml
RDF/JSON: 831-1.0078752-rdf.json
Turtle: 831-1.0078752-turtle.txt
N-Triples: 831-1.0078752-rdf-ntriples.txt
Original Record: 831-1.0078752-source.json
Full Text
831-1.0078752-fulltext.txt
Citation
831-1.0078752.ris

Full Text

NON-METALLIC INCLUSIONS IN ELECTROSLAG REFINED INGOTS by FIDEL REYES-CARMONA Ing. Quim. Met., U n i v e r s i d a d N a c i o n a l Autdnoma de Mdxico, 1976 M.Sc, The U n i v e r s i t y of I l l i n o i s a t Urbana-Champaign, 1978 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES Department of M e t a l l u r g i c a l E n g i n e e r i n g We accept t h i s t h e s i s as conforming to the r e q u i r e d standard THE UNIVERSITY OF BRITISH COLUMBIA January 19 83 © F i d e l Reyes-Carmona, 1983 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. F i d e l Reyes-Carmona Department of M e t a l l u r g i c a l Engineering The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date March 21, 1983 E - 6 (3/81) ABSTRACT The o b j e c t i v e of t h i s r e s e a r c h was t o i n v e s t i g a t e how n o n - m e t a l l i c i n c l u s i o n s ( i n c l u s i o n s ) are p h y s i c a l l y and c h e m i c a l l y transformed, removed and c o n t r o l l e d from e l e c t r o d e s to the f i n a l ESR-product. S e v e r a l 1020, 4340 and r o t o r (Ni-Mo-V) s t e e l e l e c t r o d e s were r e f i n e d by two ESR-units (7.5 mm and 200 mm i n mould diameter) under d i f f e r e n t s l a g systems. R e f i n i n g of these e l e c t r o d e s was done under d i f f e r e n t d e o x i d a t i o n p r a c t i c e s , namely pure A l , C a S i , CaSiAlBa and A I S i a l l o y s . Through t h i s r e s e a r c h i t was found t h a t i n c l u s i o n s i n the e l e c t r o d e are p h y s i c a l l y and c h e m i c a l l y transformed i n the e l e c t r o d e t i p by the thermal g r a d i e n t s . I n c l u s i o n s are c h e m i c a l l y a l t e r e d by the presence of l i q u i d s l a g a t the l i q u i d f i l m and they are e n t i r e l y d i s s o l v e d i n the m a t r i x when the d r o p l e t i s completely formed. No i n g o t i n c l u s i o n s were i d e n t i f i a b l e as of e l e c t r o d e o r i g i n and i t i s concluded t h a t a l l e l e c t r o d e i n c l u s i o n s are e i t h e r d i s s o l v e d or removed by the s l a g . The e f f e c t s o f the s l a g w i t h and without d e o x i d i z e r s on the chemical composition of the l i q u i d p o o l and i n g o t were t r a c e d d u r i n g r e f i n i n g and hence the chemistry of i n c l u s i o n s was determined by e x t r a c t i n g s l a g and l i q u i d metal samples d u r i n g r e f i n i n g . The t o t a l oxygen content was measured by the vacuum f u s i o n technique, chemical analyses i i o f s l a g by s p e c t r o p h o t o m e t r i c techniques, e l e c t r o n micro-a n a l y s i s by SEM and EPMA and x-ray ( c r y s t a l l o g r a p h i c ) a n a l y s i s . The assays were used to formulate and c o r r o b o r a t e the d e o x i d a t i o n and p r e c i p i t a t i o n mechanisms. The chemical composition o f i n c l u s i o n s i n r e f i n e d i n g o t s are more s t r o n g l y i n f l u e n c e d by the d e o x i d a t i o n p r a c t i c e than by the e l e c t r o d e or the s l a g composition i n low S i 0 2 content s l a g s . The p r e c i p i t a t i o n of complex A l - C a - S i i n c l u s i o n s i s p r e d i c t a b l e i n h i g h s i l i c a s l a g s (>10.0 wt%) and the most a p p r o p r i a t e s l a g system t o perform an e f f i c i e n t d e o x i d a t i o n i s the 50 wt% CaF 2, 30 wt% A l 2 0 3 and 20 wt% CaO. The d e o x i d a t i o n i n ESR i n g o t s takes p l a c e by the process of c o o p e r a t i v e r e a c t i o n s between s l a g and d e o x i d i z e r s i n the f o l l o w i n g sequences: 2 [Al] + 3 (FeO) t ( A l ^ ) + Fe [Ca] + (FeO) t (CaO) + Fe (A1 20 3) + [Ca] t 3 (CaO + 2 [Al] The p r e c i p i t a t i o n r e a c t i o n s are c o n t r o l l e d by the oxygen p o t e n t i a l i n the melt, thus the t r a n s i t i o n s t o be expected are : A l 2 0 3 « F e O + MnS I I -»• c t - A l 2 0 3 + MnS I I or I I I -»• x Ca0«y A l ^ + MnS I I I o r (Ca, Mn) S -*• x Ca0«y A l ^ + CaS An e x c e s s i v e d e o x i d a t i o n w i t h Ca r a i s e s the A l content i n the i n g o t a c c o r d i n g t o : X Ca + Y ( A 1 2 0 3 ) * * X CaO* (y - ^) A l 2 0 3 + | X [Al] i i i Radial i n c l u s i o n size d i s t r i b u t i o n as well as dendrite arm spacings i n samples extracted from l i q u i d pool and ingots were determined. It was found that the inc l u s i o n size obeys the normal d i s t r i b u t i o n and there i s a normal v a r i a t i o n of the inc l u s i o n size along r a d i a l distances. Hence the i n c l u s i o n composition and size i s a function of l o c a l s o l i d i f i c a t i o n conditions and also of the l o c a l thermochemical conditions. i v TABLE OF CONTENTS Page Abstract i Table of Contents i v L i s t of Figures i x L i s t of Tables x v i i L i s t of Symbols xix Chapter I INTRODUCTION 1 II LITERATURE REVIEW 4 2.1 Literature Survey on Electrode In-clusions 4 2.2 Literature Survey on Slag-Liquid Metal Reactions and t h e i r Influence on the ESR Ingot Chemistry 12 2.2.1 P r i n c i p l e s of the Reaction Scheme i n the ESR Process 12 2.2.2 On the Nature of the ESR Reaction Scheme 18 2.2.3 Thermodynamic Approach of the ESR Slag Systems 32 2.2.4 Overall View on the Modelling of ESR Reactions 34 2.3 P r e c i p i t a t i o n of Inclusions 36 2.3.1 General * . . . . 36 2.3.2 Nucleation and Growth of In-clusions 39 2.3.2.1 Homogeneous Nucle-ation 39 2.3.2*2 Heterogeneous Nucle-ation 42 2.3.3 Growth of Inclusions 4 3 V Chapter Page 2.3.4 Sulfides . . . „ 47 2.3.5 S p e c i f i c Sulfides 51 2.3.6 Oxisulfides 53 2.3.6.1 The Fe-O-S System 5 3 2.3.6.2 The Fe-O-S-Mn E q u i l i -brium 55 2.3.6.3 The Fe-O-S-Si-Mn Equilibrium 62 2.3.7 Oxides 66 2.3.7.1 Aluminates 66 2.3.7.2 Calcium Aluminates 76 2.3.7.3 Complex Oxides 90 2.4 Inclusions i n ESR-Ingots 94 III NATURE OF THE PROBLEM 106 3.1 Inclusions i n the Electrode 106 3.2 The Chemical Influence of the ESR components on the Composition of Inclusions 108 3.3 The P r e c i p i t a t i o n of Inclusions from Liquid Pool to Ingot I l l 3.4 D i s t r i b u t i o n of Inclusions During S o l i d i f i c a t i o n 113 3.5 Establishment of the Proposal and Objectives Sought Through t h i s Research 115 IV EXPERIMENTAL WORK AND TECHNIQUES 116 4.1 Experimental Procedure 116 4.2 Analysis of Inclusions 118 4.3 Total Oxygen Analysis 121 4.4 Inclusion Extraction Method 122 4.4.1 Apparatus and Experimental Pro-cedure 123 4.5 Crystallographic X-ray Analysis of Extracted Inclusions 126 4.6 Atomic Absorption Analysis (Spectro-photometry) ? 127 v i Chapter Page 4.7 Metallographic Analysis -^8 V RESULTS AND DISCUSSION • 130 5.1 Mechanism by which Electrode Inclusions are Eliminated 130 5.1.1 Behavior of Oxisulfide Inclus-ions i n 1020 M.S. Electrode Tips *. , . . 130 5.1.2 Removal of Oxide and Sulfide Inclusions in 4 340 and Rotor Steels 141 5.1.2.1 Removal of Oxides and Sulfides i n 4340 e l e c -trodes 144 5.1.2.2 Calcium Aluminum S i l i -cates i n a Rotor (Ni-Cr-Mo) Steel 146 5.1.3 F i n a l Remarks About the Re-moval Mechanism 148 5.2 The Chemical Influence of the Elec-trode, Slag and Deoxidizer on the Chemical Composition of Inclusions .. 151 5.2.1 Description of Experimental Findings 151 5.2.1.1 Preliminary Studies on the E f f e c t of the Slag and the Deoxidation 151 5.2.1.2 Intermittent CaSi Addi-tions and the Re-action Scheme 157 5.2.1.3 Refining of 1020 M.S., 200 mm Diameter In-gots Deoxidized Contin-uously with A l 159 5.2.1.4 1020 M.S. Ingots De-oxidized Continuously with a CaSi Alloy ... ..... 164 5.2.1.5 Corroboration and Ex-tension of Previous Findings to a 4340 Steel CaSi (continu-ously) Deoxidized Discussion of Results i n Terms of E l e c t -rode and Slag Composition, Related to the Second Question 5.3.1 The E f f e c t of the Electrode on the Inclusion Composition of ESR Ingots 5.3.2 Elucidation of the E f f e c t of Slag and Deoxidizers (pre-liminary studies) , 5.3.3 Preliminary Discussion on the Deoxidation Mechanism 5.3.4 Comprehensive Discussion on the Deoxidation Mechanism 5.3.5 F i n a l Remarks Findings and Discussion Related to the Third Question 5.4.1 Description of Experimental Results 5.4.1.1 The Inclusion Mean Diameter 5.4.1.2 Findings from Indi-vidual Experiments 5.4.1.3 Complimentary Studies 5.4.1.4 Summary of Experi-mental Findings 5.4.2 P r e c i p i t a t i o n of Inclusions i n the Fe-Al-Ca-O-S (Mn) system 5.4.3 Discussion of Results » 5.4.3.1 Nucleation Growth and Fl o t a t i o n of Inclusions . . . 5.4.3.2 Comparison between Theoretical and Ex-perimental Results...* v i i i Chapter Page VI THE RADIAL DISTRIBUTION OF INCLUSIONS IN CaSi AND A l DEOXIDIZED INGOTS 2 2 3 6.1 Experimental Details and Techniques...... 223 6.2 Experimental Findings 2 2 4 6.3 Discussion of Results 2 2 5 VII CONCLUSIONS 228 VIII SUGGESTIONS FOR FUTURE WORK 2 32 LIST OF REFERENCES 235 FIGURES 251 TABLES 3 5 1 APPENDIX . . . ........ 3 7 4 ix LIST OF FIGURES Figure Page 1. Schematic i l l u s t r a t i o n of an ESR system .... 251 2. Predicted and measured temperature pro-f i l e s for a 1018 MS electrode 25 mm i n diameter 252 3. Manganese content of the metal for uni-variant equilibrium y - i r o n + "MnO" + "MnS" + l i q u i d (1) for Fe-Mn-S-0 sys-tem and univariant equilibrium Y - i r o n + "MnS" + l i q u i d s u l f i d e for Fe-Mn-S system 253 4. Univariant e q u i l i b r i a i n Fe-Mn-S-0 system i n the presence of y-iron and Mn(Fe)0 phases 254 5. Univariant e q u i l i b r i a involving s o l i d metal and Mn(Fe)0 i n the Fe-Mn-S-0 system bonded with ternary Fe-Mn-0 and Fe-S-0 terminal-phase f i e l d s (e) , ,(p) , (f) , (n), (g) and (h) 255 6. Location of the planes i n the quaternary (FeO-MnO-MnS-Si02) system 256 7. Equilibrium phases i n three planes of the FeO-MnO-MnS-SiC^ system. a) MnS-FeO-2MnO«Si02, b) MnS-2FeO«Si0 2-2MnO«Si0 2 and c) MnS-FeO-MnO 256 8. Behavior of inclusions enriched i n Mn and Si as a function of temperature 257 9. Schematic i l l u s t r a t i o n of changes i n i n -clus i o n composition i n a 1020 MS e l e c t -rode produced v i a acid e l e c t r i c furance.... 258 10. Wt. % A l and wt. % Ca and wt. % 0 i n l i q u i d i r o n at unit A l 2 0 3 and CaO a c t i v i t y 259 11. Isothermal Fe-Al-Ca-0 p r e c i p i t a t i o n (Henrian a c t i v i t i e s ) diagram 260 X Figure Page 12. Ternary Al 20 3~(Ca,X)0-Si0 2 i n c l u s i o n diagram 261 13. Slag chemical composition used i n the ,.(34,53,83) , a. • 4.-past and present i n v e s t i g -ation 262 14. Schematic i l l u s t r a t i o n of the ESR arrange-ment used in this investigation 26 3 15. Schematic i l l u s t r a t i o n of the "inclusion extractor" 264 16. Typical inclusions from 1020 MS electrodes (optical microscopy) * 265 17. Deformed inclusions i n 1020 MS electrodes and t h e i r X-ray spectrum analysis (SEM) ... 266 18. Macrostructure of a 1020 MS electrode t i p where l i q u i d f i l m , p a r t i a l l y molten and f u l l y r e c r y s t a l l i z e d areas are shown 267 19. Macrostructures of a 4340 electrode t i p . Droplet, l i q u i d f i l m , p a r t i a l l y molten, f u l l y and p a r t i a l l y r e c r y s t a l l i z e d zones are shown 268 20. Macrostructure of a rotor (Ni-Cr-Mo) s t e e l . Liquid f i l m , and p a r t i a l l y molten areas are shown 269 21. Schematic i l l u s t r a t i o n of 1020 MS electrode t i p (acid e l e c t r i c furnace produced) subjected to ESR-thermal gradients 270 22. Multiphase ( r e l a t i v e l y grown) inclusions i n a 1020 MS electrode t i p 271 23. Single phase inclusions i n p a r t i a l l y and f u l l y molten regions i n a 1020 electrode t i p 272 24. Complex Ca-Al-Si-Mn inclusions located i n the l i q u i d f i l m and droplets 273 25. Spectrum X-ray analysis of Ca-Al-Si-Mn i n -clusions i n a 1020 MS electrode located i n the l i q u i d f i l m and droplets 274 x i Figure Page 26. Changes i n i n c l u s i o n chemical composition i n a 4 340 electrode t i p subjected to ESR thermal gradients 275 27. Changes i n in c l u s i o n chemical composition i n a 4340 electrode t i p with a strong recrys-t a l l i z e d region 276 28. Behavior of oxide inclusions in a e l e c t - . rode t i p of a rotor s t e e l subjected to ESR thermal gradients 277 29. Aluminum s i l i c a t e s i n a 4340 ESR ingot 75 mm i n diamter. Precipitated inclusions i n a l o c a l i z e d region 278 30. Influence of CaSi and FeO intermittent additions on the oxygen content of a 1020 M.S. (RIII-W) 279 31. Changes i n oxygen content i n a 1020 MS ingot as a r e s u l t of CaSi and FeO i n t e r -mittent additions (RII-W) 280 32. Changes i n slag chemical composition i n a 1020 MS (RIII-W) as a r e s u l t of CaSi and FeO intermittent additions 281 33 Changes i n slag chemical composition as a r e s u l t of intermittent additions of CaSi and FeO i n slag during r e f i n i n g (RII-W) 282,283 34. Changes i n ingot chemical composition as a r e s u l t of CaSi and FeO additions i n slag during r e f i n i n g (RIII-W) 284 35. E f f e c t of CaSi and FeO additions i n the slag on the chemical composition of a 1020 MS ingot (RII-W) 285 36. Changes i n in c l u s i o n composition and size i n a 1020 MS ingot (RIII-W) as a r e s u l t of intermittent additions of CaSi and FeO i n the slag 286 37. Chemical analysis of slag samples i n R I - I l . 287 x i i Figure Page 38. Ingot chemical analysis i n R I - l l 288 39. Slag chemical analysis (wt. %) in a con-tinuously A l deoxidized 1020 MS ingot (RII-Il) 289 40. Slag chemical analysis (wt. %) i n R I I - I l . . . 290 41. Inclusion mean diameter and t o t a l oxygen content i n a continuously (Al) deoxidized ingot, (RII-Il) 291 42. Total oxygen content and inc l u s i o n mean diameter i n ingot (RII-I2) 292 43. Inclusion chemical composition (at. %) as a function of continuously increasing deoxidation rates i n RII- I l 293 44. Inclusion chemical composition (at. %) as a function of the ingot height (or continuously increasing deoxidation rates) i n RII-I2 294 45. Ingot chemical analysis i n RII-Il 295 46. Ingot chemical analysis i n RII-I2 296 47. "Alumina galaxies", (EPMA), associated to MnS II i n i n c i p i e n t aluminum deoxidized ingots, (RII-Il and RII-I2) 297 48. "Alumina galaxies" (EPMA) and MnS II i n A l deoxidized ingots 29 8 49. Faceted alumina (a-A^C^) i n samples from l i q u i d pool and ingot deoxidized with A l .. 299 50. Calcium aluminates low i n Ca from highly A l deoxidized ingots. A l , Ca and S composition RII-I2 300 51. Composition dependence of s u l f i d e phases on the Ca:Al r a t i o i n the oxide phase (RII-Il) 301 x i i i Figure Page 52. Composition dependence of s u l f i d e phases on Ca-aluminate inclusions phases i n RII-I2 .. 302 5 3 • Segregated material i n an A l deoxidized ingot ( l i q u i d pool) . . 3 0 3 54. Dependence of "FeO" contents i n slag on the Ai2 0-.:CaO r a t i o i n slag of a continu-ously A l deoxidized ingot (RII-Il) 304 55. Dependence of "FeO" contents i n slag on the A^O^tCaO r a t i o i n slag of a continu-ously A l deoxidized ingot (RII-I2) 305 56. Changes i n t o t a l oxygen content and i n c l u s -ion mean diameter i n a continuously CaSi deoxidized ingot (RIII-Il) 306 57. Inclusion mean diameter and t o t a l oxygen content i n a continuously CaSi deoxidized ingot (RIII-I2) 307 58. Inclusion chemical composition as a function of deoxidation rates i n RI I I - I l 308 59. Inclusion chemical composition as a func-t i o n of deoxidation rates i n RIII-I2 309 60. Changes i n slag chemical composition i n a continuously CaSi deoxidized ingot (RIII-I l ) 310 61. Changes i n slag chemical composition i n a continuously CaSi deoxidized ingot (RIII-12) 311 62. Changes i n A l and Si i n RII I - I l as a con-sequence of continuously increasing CaSi deoxidation rates 312 63. Changes in ingot composition as a con-sequence of continuously increasing CaSi deoxidation rates i n RIII-I2 313 X I V Figure Page 64. Dependence of "FeO" contents i n slag samples on the deoxidation rates i n RI I I - I l • 65. Dependence of "FeO" contents i n slag samples on the deoxidation rates i n RIII-I2 314 315 66. Sulfur (as sulfides) content i n inclusions as a function of the Ca:Al r a t i o i n the Ca-aluminate inclusion phases i n RI I I - I l .. 316 67. Sulfur (as sulfides) content i n inclusions as a function of the Ca:Al r a t i o i n the Ca-aluminate in c l u s i o n phases i n RIII-I2 .. 317 68. Chemical composition of inclusions (as Ca:Al ratios) as a r e s u l t of continuously increasing deoxidation rates i n RIII-I2, (a) and sulfu r (as sulfide) i n inclusions i n Ca-aluminates, (b) . s 318 69. Segregate enriched i n A l , Ca and S i i n a sample extracted from the l i q u i d pool of ingot RI I I - I l 319 70. Slag chemical analysis of a 4340 ingot continuously deoxidized with a CaSi a l l o y , [R-4340 (1)] 320 71. Ingot chemical composition i n R-4340 (1)... 321 72. Inclusion size d i s t r i b u t i o n and t o t a l oxygen content i n R-4340 (1) 322 73. Changes i n "FeO" contents i n the slag as a consequence of the continuously i n -creasing CaSi deoxidation rates i n R-4340(1) 323 74. Inclusion chemical composition (as Ca:Al r a t i o s i n at. %) i n samples of l i q u i d pool and ingot i n R-4340 (1) 324 75. Inclusion composition as Ca:Al r a t i o s and S content i n R-4340 (1) 325 X V Figure 76. 77. 78 . 79. 80. 81. 82. 83. 84. 85. 86. 87. Page Inclusion composition (oxide and s u l f i d e phases i n a rotor s t e e l deoxidized with Si based a l l o y s , namely: Ca-65 wt.% S i , Al-65% Si and "Hypercal". R-RS(I), R-RS(II) and R-RS(III) 326 Segregate enriched i n A l , Ca and S i from the A l S i deoxidized ingot, R-RS(II) 327 Inclusion p r e c i p i t a t i o n sequence i n a st e e l containing two le v e l s of sulfu r 328 S t a t i s t i c a l determination of the mean i n -clusi o n diameter 329 Ce-distribution i n a i n c l u s i o n of a sample extracted from the l i q u i d pool 330 Ce and La d i s t r i b u t i o n i n an inclus i o n of a sample extracted from the l i q u i d pool. La and Ce come from a RE wire located i n the quartz tubing 331 A l , Ca and Zr d i s t r i b u t i o n s i n an inclus i o n of a sample extracted from the l i q u i d pool. Zr was i n the quartz tubing 332 Di s t r i b u t i o n of oxide formers i n an i n -clusion of a sample extracted from the l i q u i d pool , D i s t r i b u t i o n of oxide-sulfide former i n an i n c l u s i o n of a sample extracted from the l i q u i d pool Inclusion d i s t r i b u t i o n i n a dendrit i c structure of 1020 MS samples taken from l i q u i d pool during r e f i n i n g Inclusion d i s t r i b u t i o n i n a dendrit i c structure of a 4 34 0 sample from the l i q u i d pool Isothermal (1823 K) p r e c i p i t a t i o n (Fe, A l , Ca, 0, S) diagram at 0.1 a c t i v i t y of aluminum 333 334 335 336 337 xvi Figure Page 88. E f f e c t of the a c t i v i t y of A l ( h A 1 = 0.001, 0.01 and 0.1) on the " p r e c i p i t a t i o n se-quence" of Ca-aluminates 338 89. E f f e c t of the a c t i v i t y of S (h = 0.01, 0.01 and 0.001) on the " p r e c i p i t a t i o n se-quence" of Ca-aluminates 339 90. Arrangement of aluminates i n r e l a t i v e l y low CaSi deoxidized ESR-ingots 340 91. Typical arrangement of inclusions i n a r e l a t i v e l y low CaSi or high A l deoxidized ingots 341 92. X A1 20_• Y CaO/CaS interface i n an ESR ingot, R-RS (I I I ) , deoxidized with "hypercal". X-ray spectrum analyses are also included 342,343 93. Secondary DAS i n a round (200 mm i n diameter) 1020 MS ESR ingot 344 94. Secondary DAS i n a 1020 MS ESR ingot 345 95. Secondary DAS i n a 4340 ESR ingot 346 96. Radial size d i s t r i b u t i o n of inclusions i n an A l deoxidized ingot, (RII-I2) 347 97. Radial i n c l u s i o n size d i s t r i b u t i o n i n a low CaSi deoxidized ingot (RIII-Il) 348 98. Radial i n c l u s i o n size d i s t r i b u t i o n i n a 200 mm ESR ingot CaSi deoxidized, (RIII-Il) 349 99. Radial i n c l u s i o n size d i s t r i b u t i o n i n a 4340 (ESR) ingot CaSi deoxidized 350 LIST OF TABLES Table I. C o r r e c t i o n f a c t o r to the Stokes' Law I I . Thermochemical data f o r a) I n v a r i a n t e q u i l i b r i a i n Fe-S-0 system and b) Estimated data f o r i n v a r i a n t e q u i l i -b r i a i n Fe-Mn-0, Fe-Mn-S and Mn-S-Q t e r n a r y systems I I I . Estimated data f o r i n v a r i a n t e q u i l i b r i a i n Fe-Mn-S-0 quaternary system IV. C a l c u l a t e d and p u b l i s h e d f r e e energy data f o r Fe-O-Ca-Al system a t 1823 K (1550°C)... V. E q u i l i b r i u m c o n s t a n t s f o r d e o x i d a t i o n r e a c t i o n s VI. Equations f o r i n v a r i a n t e q u i l i b r i a i n the i s o t h e r m a l (Fe-O-Ca-Al) system V I I . Chemical a n a l y s i s of e l e c t r o d e s used i n t h i s r e s e a r c h V I I I . Experiments and t h e i r f e a t u r e s used i n t h i s i n v e s t i g a t i o n IX. Chemical composition of d e o x i d i z e r s used i n the p r e s e n t i n v e s t i g a t i o n X. a) I n c l u s i o n composition as a r e s u l t of r e m e l t i n g 4340 e l e c t r o d e s i n a s m a l l ESR-furnace (75 mm i n diameter) b) S l a g - d e o x i d i z e r e f f e c t on i n c l u s i o n composition c) I n c l u s i o n chemical composition of 4340 e l e c t r o d e s used i n the s m a l l ESR-furnace XI. Chemical e f f e c t of s l a g and e l e c t r o d e sur face p r e p a r a t i o n on i n c l u s i o n composition. (A 1020 MS and a r o t o r (Ni-Cr-Mo) s t e e l were r e f i n e d i n the 200 mm i n diameter ESR-furnace through s e v e r a l s l a g systems ). . x v i i Page 351 353 355 356 357 358 359 360 361 362 362 362 x v i i i Table XII. XIII. XIV. XV. XVI. XVII. XVIII. Page Slag chemical analysis of ESR (Ni-Cr-Mo) ingots deoxidized with Si based deoxi-dizers 364 Chemical analysis of ingots deoxidized with: a) the Al-65% Si a l l o y 365 b) the Ca-65% S i a l l o y and 366 c) the CaSiAlBa (hypercal) a l l o y 367 Typical data recorded from EPMA analysis of i nclusions. a) l i q u i d pool and 368 b) ingot E f f e c t of i n i t i a l number of inclusions on growth during cooling of l i q u i d metal 369 370 Derived equations for invariant (isothermal) e q u i l i b r i a i n the Fe-O-Ca-Al-S system Computed compositions based on data given i n Table XV and e^ 1 = - 25, e 0^ = -62 and e C * = -40 3 7 1 Computed compositions i n the Fe-Ca-Al-0-S system by using information i n Table XV, variable e~r (-535, -400, -300, Al -250 and -200) i n addition to the e Q = -62 and e C s a = -110 3 7 2 373 xix LIST OF SYMBOLS l (A) [A], A A: B a c t i v i t y of the i component 'A' species i n slag 'A' element i n solution i n l i q u i d i ron r a t i o of A to B species A ( g ) ' A ( l ) ' A ( s ) 'A' species i n gaseous, l i q u i d or s o l i d state A l S i deoxidant, composition of which i s Al-65 wt.% S i . (Table IX) A l 2 ° 3 * alumina as a primary deoxidation product at. % atomic percent ion i c species with either a pos i t i v e or negative valence a-Fe a l l o t r o p i c state of iron a - A l 2 0 3 corundum; i t i s also given as 'A' when i t i s referred to as a part of the Ca-aluminates p r e c i p i t a t i o n sequence degrees i n the Celsius (centi-grade) scale C:C» supersaturation r a t i o i n terms of concentrations (1)'(s) l i q u i d or s o l i d CaO and C 3A 2 stoichiometric Ca-aluminates as given by the pseudo binary (CaO-Al 20 3) phase diagram, i . e . , CaO«Al 20 3, CaO-2Al 20 3, CaO«6Al 20 3, 12CaO«7Al 20 3 and 3CaO«2Al 20 3 X X (CaO) * (Ca,Mn)S CaS (CaS)* CaSi CaSiAlBa DAS DAS 1 1 D c 6„ or 6-iron Fe e i EPMA f peripheral phase on a calcium aluminate oxide inclusion double s u l f i d e with Ca and Mn phase heterogeneously p r e c i p i t a t e d on a Ca-aluminate phase peripheral calcium s u l f i d e phase i n equilibrium with the CaO from the Ca-aluminate and oxygen and sulfur i n solution in iron deoxidant, composition of which i s given i n Table (IX) deoxidant ("hypercal"), composition of which i s given i n Table IX dendrite arm spacing secondary dendrite arm spacing c r i t i c a l drag force on a spherical p a r t i c l e a l l o t r o p i c state of iron electron i n t e r a c t i o n c o e f f i c i e n t ; the e l e -ment for which the a c t i v i t y co-e f f i c i e n t i s being calculated i s designated j and the element causing the e f f e c t i s desig-nated i . electron-probe-micro analyses Henrian a c t i v i t y c o e f f i c i e n t FeO" iron oxide, (FexO) xxi Fe(x,y,z)-0-S pseudo ternary i n c l u s i o n phase diagrams with an oxide and a su l f i d e phase;(Hilty and co-,(129) workers) f^ l i q u i d f r a c t i o n GR product of growth rate by thermal gradients Y, k, or 8-Al2C>2 a l l o t r o p i c states of the alumina Y^ a c t i v i t y c o e f f i c i e n t of species A h^ Henrian a c t i v i t y of species A HIC hydrogen induced cracking HSLA high strength low a l l o y s t e e l e x c i t a t i o n voltage i n k i l o v o l t s kV K £1 absolute degrees- in the Kelvin (absolute) scale Kg t o n - 1 deoxidation rate i n kilograms of deoxidant per (metric) tonne of remelted ingot l i q u i d o x i s u l f i d e £^ l i q u i d metal L and L„ l i n e s d i v i d i n g the ternary i n -elusion and slag compositions suggesting the formation of low melting phases, Figures (12) and (13) X wave length i n X-rays m(CaO) • nfA^O.j) m and n are c o e f f i c i e n t s of the Ca-aluminate phases which are equivalent to those i n the CaO-A l 2 0 3 pseudo-binary phase diagram x x i i MnS(I,II,III) y yA ym o, "o", or oxi ppm RII-W, RIII-W RII-I l and RII-I2 RI I I - I l and RIII-I2 R-43040 (1) R-RS(I), R-RS(II) and R-RS(III) P a u V + < manganese s u l f i d e type I, I I , or III v i s c o s i t y specimen current density in (EPMA) microamperes unit length, microns oxide of the type Mn(Fe)0 concentration in parts per m i l l i o n ingots i n which CaSi and FeO were added (200 mm i n diameter) (1020) ESR ingots Al-deoxidized (200 mm i n diameter) (1020) ESR ingots CaSi deoxidized (200 mm i n diameter) (4340) ESR ingot CaSi deoxidized (200 mm i n diameter) Rotor (Ni-Cr-Mo) s t e e l deoxidized with CaSi, A l S i and CaSiAlBa alloys density < i n t e r f a c i a l tension r e s u l t i n g vector v e l o c i t y v e l o c i t y vector degree of accuracy (plus or minus) i n chemical analysis less than -«- or = s or ^ reaction in equilibrium approximately gaseous phase ACKNOWLEDGEMENTS I would l i k e to express my sincere gratitude to my supervisor Dr. Alec M i t c h e l l for his concise advise. I am also thankful to Professors R. Butters and B. Hawbolt for t h e i r contributions and discussions during t h i s work. I also appreciate the technical assistance of A. Lacis, G. S i d l a , R. Cardeno, H. Tump, R. McLeod and M. Mager. I would l i k e to thank the Banco de Mexico, Consejo Nacional de Ciencia y Technologia (CONACyT), Universidad Nacional Autonoma de Mexico (Departamento de Metalurgia de l a UNAM) and the Department of Met a l l u r g i c a l Engineering of the University of B r i t i s h Columbia for the f i n a n c i a l support given during my professional studies. The author i s also g r a t e f u l for the f i n a n c i a l assistance of the American Iron and Steel I n s t i t u t e (Project No. 32-445). This work i s s p e c i a l l y dedicated to my parents, brothers and s i s t e r s and everyone who has contributed to reach my goal. 1 CHAPTER I INTRODUCTION In today's technology where there i s the demand for extra-high-quality materials, there are only three second-ary steelmaking processes capable of f u l f i l l i n g the re-quired stringent standards 1) Electron Beam Melting (EBM), 2) Vacuum Arc Refining (VAR) and 3) El e c t r o s l a g Refining (ESR). The properties r e s u l t i n g from any of the above pro-cesses can be categorized as: c r y s t a l structure, chemical homogeneity, s u l f u r and phosphorus content and in c l u s i o n chemistry and size d i s t r i b u t i o n . The EBM and the VAR pro-cesses o f f e r the lowest gas content. The v e r s a t i l i t y i n control and operation, c r y s t a l structure, r e l a t i v e e l e c t r i -c a l e f f i c i e n c y and r e p r o d u c i b i l i t y are the main features of the ESR process. * Research c a r r i e d out since 1967 has widely demonstrated that the ESR-process o f f e r s d e f i n i t e advantages over con-ventional practice and i n some respects also offers advan-tages over other secondary steelmaking practices. The con-tinuous demand for high q u a l i t y materials has increased since 1967. From 1960 to 1973 the western world increased i t s ESR production from 2 600 to about 120 000 tonne/year and i t i s forecast to increase to 600 000 tonneVyear i n 1985. The Soviet Union's production i s about three times that * F i r s t Int. Symp. on ESR 2 of the Western world. This marked increase i n production indicates the a c c e p t a b i l i t y of the products manufactured by the ESR-technology. Materials produced for p a r t i c u l a r purposes such as rotors used i n thermal and nuclear e l e c t r i c a l plants, land-ing gear used i n a i r c r a f t , crank shafts used i n large vessels, gun barrels, superalloys used i n turbine blades, high q u a l i t y t o o l and b a l l bearing steels , etc. are ex-amples of the wide variety of materials produced by the ESR-process. These components are subject to dr a s t i c temperature and environmental conditions and/or to dy-namic stresses. These two adverse conditions by them-selves generate microcracks which are usually associated with chemical inhomogeneities and/or with inclusions in the metal matrix. Inclusions play a major role i n f a i l u r e s under the described s i t u a t i o n s . Thus t h e i r index of s p h e r i c i t y , degree of cohesion with the metal matrix, i n t e r p a r t i c l e distance, volume f r a c t i o n , size d i s t r i b u t i o n , p l a s t i c i t y , thermal contraction and expansion c o e f f i c i e n t s with respect to the matrix and chemistry are the parameters which determine the service l i f e , i n terms of inclusions, for a given material. The pote n t i a l of the ESR-process, because of i t s ver-s a t i l i t y , can be extended to other types of uses such as 3 Electroslag Casting (ESC) and Electroslag Welding (ESW). It i s important to note that while these processes are very r a r e l y used i n North America, i n the Soviet Union's technology they are widely practiced. The ESR process because of i t s multiple degrees of freedom also o f f e r s a wide range of parameters to be modi-f i e d and improved without modifying the standard furnace. To optimize the mechanical properties which are s t r i c t l y related to inclusions i n a ESR-product, a series of i n t e r -actions between a l l the components of the process should be evaluated, i . e . electrode, slag, deoxidizer and s o l i d i -fying ingot should a l l be considered. If the mechanisms by which inclusions are formed are known, then the ESR-process c a p a b i l i t i e s and r e s t r i c t i o n s i n t h i s respect can be defined. 4 CHAPTER II LITERATURE REVIEW 2.1 Literature Survey on Electrode Inclusions From the thermal point of view the electrode t i p should be considered as one of the sources by which one part of the t o t a l heat produced by the slag i s consumed, Figure (1). This amount of thermal energy which i s trans-ferred from the slag to the electrode plays several r o l e s : 1. It i s primarily converted to sensible heat, thus producing a f i n e l i q u i d f i l m which afterwards w i l l form droplets, and 2. I t also determines the temperature gradients above and below the slag/gas interface. Thus, i t establishes the amount of possible surface oxidation and the d i s s o l u t i o n of c e r t a i n second phase p a r t i c l e s or inclusions i n the elec-trode . (1-5) Theoretical and experimental studies have been per-formed to determine whether inclusions from the electrode are eliminated before or a f t e r the metal droplet i s formed. Heat and mass transfer models have also been developed to predict the maximum in c l u s i o n diameter which can be d i s -solved under given ESR-conditions. (6) M i t c h e l l , Joshi and Cameron have studied the temp-erature d i s t r i b u t i o n above the slag/gas interface i n a laboratory ESR furnace. They have i n d i c a t e d t h a t the r a d i a l temperature g r a d i e n t s i n a l a r g e r e l e c t r o d e - i n g o t c o n f i g u r a t i o n b< came s i g n i f i c a n t . T h e i r r e s u l t s a l s o suggest t h a t d i ; s o l u t i o n of second phase p a r t i c l e s to a v a r y i n g ex-t e n t i s f e a s i b l e . (7) M a u l v a u l t and E l l i o t t who have developed a one dimensional model have taken i n t o account the v e r t i -c a l movement of the e l e c t r o d e . T h e i r computations, which have been based i n an assumed p a r a b o l i c p r o -f i l e , have shown a reasonable agreement with the expe ment a l l y (37 mm diameter e l e c t r o d e ) determined v a l u e s (8) Mendrykowski e t a l . ' s work by u s i n g a s i m p l i -f i e d one-dimensional-heat flow model and c o n s i d e r i n g the e l e c t r o d e p a r t immersed i n the s l a g have a l s o found t h a t w h i l e n e i t h e r r a d i a t i o n nor gas phase con-v e c t i o n p l a y a major r o l e , f o r a g i v e n s e t of c o n d i -t i o n s , the c o n v e c t i v e heat t r a n s f e r from the s l a g to the e l e c t r o d e i s indeed more s i g n i f i c a n t . T h e i r r e -s u l t s suggest t h a t c o n d u c t i o n along the e l e c t r o d e pre dominates as the h e a t - t r a n s f e r - c o n t r o l l i n g mechanism. ( 9 ) Tacke e t a l . along the l i n e s with work per-formed a t U.B.C. (6,10) have coupled two models (one 6 to determine the slag temperature and the other to determine the heat fluxes) to calc u l a t e the electrode temperature, i t s melting p r o f i l e and i t s depth of im-mersion i n the slag. It has been claimed that computed values obtained by th i s two-dimensional flow are in agreement with the experimental findings. The r a d i a l e f f e c t was"also t h e o r e t i c a l l y and experimentally anal-yzed. These re s u l t s i n agreement with M i t c h e l l et (6) a l . show that the temperature gradients become steeper i n the electrode nearer the l i q u i d f i l m . A representative example of t y p i c a l gradients i n an electrode are shown i n Figure (2). It i s important to mention that i n a l l the above models the thermal energy spent for r e c r y s t a l l i z a t i o n or grain growth, as shown by several r e s e a r c h e r s ^ ' 11-14) n Q t j 3 e e n considered. The i n c l u s i o n d i s s o l u t i o n phenomenon has also been approached by several researchers from the mass (12) transfer view point, Kay and Pomfret were the f i r s t researchers to have suggested and modelled the d i s s o l u t i o n of oxide inclusions ( s i l i c a and alumina) i n the electrode f i l m under normal ESR conditions. They claim that 7 although inclusions can be dissolved as a r e s u l t of the l i q u i d f i l m - l i q u i d slag i n t e r a c t i o n during the droplet formation stage, their computed values for d i s s o l u t i o n rate would only require the time that an electrode material spends before i t becomes l i q u i d . Their c a l c u l a t i o n s for d i s s o l u t i o n of alumina and s i l i c a inclusions whose di a -meters were 4 and 20 ym, were performed under the as-sumption that thermodynamic equilibrium at the electrode t i p / s l a g interface was reached at 1800 and 2000°C. M i t c h e l l based on heat transfer c a l c u l a t i o n ^ ' has recalculated the d i s s o l u t i o n of inclusions (12) using the conditions of Pomfret and Kay . M i t c h e l l ' s r e s u l t s show that even by using a two-fold superheating (8) (70° C) above that found by Mendrikowski et a l . no solution i s predicted below 1600° C. Hence the i n c l u s i o n -d i s s o l u t i o n mechanism i n the s o l i d electrode t i p was not considered to strongly influence the o v e r a l l i n c l u s i o n removal. (16) Hajra and Ratnam have also performed mass trans-fer c a l c u l a t i o n s and experimental research i n a laboratory ESR-furnace. Their approach was on the same basis as the previously described works. Their r e s u l t s i n agreement with Mitchell's show that slag/metal reactions play an important r o l e in the electrode-inclusion removal. They also found that oxide p a r t i c l e s c h a r a c t e r i s t i c of the electrode were not traced i n the ingot. 8 Experimental and t h e o r e t i c a l work, so far des-cribed, has only been concerned with laboratory ESR (17) furnaces. Medovar et a l . also reported that, i n 800 - 1200 mm consumable electrodes refined by ESR, the i n c l u s i o n removal occurs i n the molten metal f i l m or i n the process of droplet formation. They have also indicated that inclusions i n droplets d i f f e r e d from those i n the electrode. They claim that the i n c l u s i o n shape, sizes and d i s t r i b u t i o n i n the s o l i d droplets were sim i l a r i n nature to those i n the ESR ingot. (19 20) Research ca r r i e d out i n the Soviet Union ' on a quantitative basis has suggested that inclusions i n the electrode during r e f i n i n g are s p e c i f i c a l l y l o -cated i n the l i q u i d f i l m and they have a well defined size d i s t r i b u t i o n . Based on these findings i t has been claimed that inclusions are mechanically and chemi-c a l l y removed by the slag. (19) On the other hand Roshchin et a l i n agreement with other i n v e s t i g a t o r s ^ 1 ' ' h a v e established that due to the "high temperature heating" manganese su l f i d e s 9 are f i r s t l y spherodized and afterwards dissolved. It has (19) been observed that s i l i c a t e inclusions were sphero-dized, transformed and s l i g h t l y enlarged, i n contradiction to several investigators' r e s e a r c h 1 3 ^ before they reach the l i q u i d f i l m . The same contradiction i s found with respect to second phase p r e c i p i t a t e s . While some researchers believe that d i s s o l u t i o n occurs i n the s o l i d s t a g e ^ others have reported that t h i s takes place in the l i q u i d stage and s t i l l others ^  have suggested that they do not dissolve and serve as nucleating agents i n the r e f i n i n g ingot. Roshchin et a l . ' s work was performed exclusively using o p t i c a l and opti c a l - q u a n t i t a t i v e (inclusion size d i s t r i b u t i o n ) techniques. These researchers claim that simulated heat treated samples (under an i n e r t atmo-sphere) subjected to several periods of time and temp-erature ranges, produced equivalent re s u l t s to those ob-served i n actual ESR-electrodes. (13) Other studies i n l i n e with the previous work, using d i f f e r e n t schedules have agreed with the above f i n d -ings. The major disadvantage of these simulated experi-ments i s that the calculated thermal g r a d i e n t s ^ and the time-temperature schedules are so d i f f e r e n t to that experienced by the electrode t i p s that a self-consistent conclusion cannot be derived, (15) Studies c a r r i e d out on a quantitative basis (total oxygen content and i n c l u s i o n chemical analysis) have deter-mined that inclusions are dissolved i n the l i q u i d f i l m . These a n a l y t i c a l studies however were performed exclusively on. material belonging to the molten family. The idea which supports the existence of a continuous reoxidation due to continuous i n c l u s i o n d i s s o l u t i o n , from the heat affected region to the electrode l i q u i d f i l m , has also been p r o p o s e d 1 3 ^ . The chemical nature of these inclusions, however, was not investigated. Theoretical s t u d i e s ^ ' ^ ' ^ ^ indicate that for a given electrode-mold diameter configuration some superheating i s expected at the electrode t i p , although for a short period of time. On t h i s basis i t has been anticipated that i f (18) inclusions contact the slag or e a r l i e r i f they are s i l i c a type, t h e i r d i s s o l u t i o n rate should be extremely high (21) for a l l common slag - i n c l u s i o n combinations (22) Paton et a l . have also suggested that the i n t r i n s i c l iquidus-solidus length of each a l l o y system and the e l e c -trode-steelmaking practice also play an important r o l e . Their studies were performed on 1200 mm diameter electrodes and a gradual d i s s o l u t i o n of s u l f i d e s was (optically) ob-served. The c r i t i c a l length at which changes i n sulfur con-centrations were observed was about one centimeter above the "fusion l i n e . " (23) Zhengbang et a l . ' s studies based on the con-9 5 centrations of a r t i f i c i a l Zr 0 2 inclusions have shown that the chemistry of inclusions during the ESR process change gradually from the electrode to the ingot. Other research-ers claim that the major reaction s i t e where inclusions from the electrode are eliminated i s at the l i q u i d f i l m -(23 24) (15) slag interface ' . M i t c h e l l who has refined electrodes containing calcium aluminum s i l i c a t e s has re-ported small inclusions i n the l i q u i d f i l m , the composition of which did not correspond to the stoichiometric 2FeO«Si0 2 phase. 2.2 Literature Review on Slag-Liquid Metal Reactions 12 and th e i r Influence on the ESR Ingot Chemistry 2.2.1 P r i n c i p l e s of the Reaction Scheme i n the ESR Process Among the conventional and secondary steelmaking prac-t i c e s the ESR-process represents one of the most complex metallurgical reactors. The degree of d i f f i c u l t y i n i t s study arises because reactions take place at s i t e s (elect-rode-slag, droplet-slag and l i q u i d pool interfaces) which have separate and d i s t i n c t chemical and electrochemical re-(26 27) gimes ' . The droplet, due to i t s size sees no pot e n t i a l difference between i t and the surrounding slag therefore i t reacts under the thermochemical conditions dictated by the slag. The electrode ( l i q u i d film) - slag and the molten l i q u i d pool - slag interfaces react almost e n t i r e l y by imposed electrochemical potentials. Reactions at these s i t e s are controlled by the surface environments and they are not d i r e c t l y influenced by the slag chemistry. It i s worthwhile to mention that the above patterns are mainly applied to DC-ESR and to a given extent to the AC-ESR (26) (2728) operation . Several researchers ' have suggested that even i n t h i s l a t t e r operating mode slag-metal ex-change and surfaces are ruled by the p o l a r i z a t i o n behavior. (28 29 ) " Several studies ' on the current density-poten-t i a l behavior have shown that since there i s not a lin e a r r e l a t i o n s h i p between these parameters for DC and AC ranges used i n ESR, then p o l a r i z a t i o n e x i s t s . Schwerdtfeger 1s (28) studies have also shown that af t e r the l i m i t i n g cur-rent i s approached for a given slag system a plateau i s reached. This l i m i t i n g current can be increased again only i f the p o t e n t i a l i s markedly increased. Thus, the magni-tude of the l i m i t i n g current from t h i s current density-p o t e n t i a l r e l a t i o n s h i p has given a clear i n d i c a t i o n that 2+ a surface saturation i n Fe r e s u l t s i n the separation of an i r o n - r i c h phase which remains fixed on the anode sur-face by i n t e r f a c i a l tension f o r c e s ^ 3 ^ . Therefore, t h i s incomplete i r r e v e r s i b i l i t y leads to a net "FeO" production i n the slag and also to a net solution of aluminum or c a l -cium i n the i r o n . Besides the r e c t i f i c a t i o n of AC current caused by i n t e r -(30 32) f a c i a l e f f e c t s there i s evidence ' i n the l i t e r a t u r e which establishes that r e c t i f i c a t i o n due to current passing 2+ through the slag-skin/mould increases the Fe i n the slag bulk, thus r a i s i n g a l l oxidation rates i n the ESR-reaction (33) scheme. Hawkins et a l . have shown that i f 5% - 30% of the t o t a l current i s passed through the mould with an e f f i -ciency of 2% for the anodic reaction a r i s e of about 40 ppm of oxygen would occur, under normal ESR-conditions. (29) It has been speculated that the mechanism which con-t r o l s t h i s type of r e c t i f i c a t i o n i s the presence of small arc contacts which pass through the slag-skin into the slag. (27) M i t c h e l l has also indicated that due to the high temperature and high current-density slag-metal interface the reaction scheme i s probably not a well-defined 14 Faradaic interface. Thus e l e c t r o l y t i c reactions are pos-s i b l e only to the extent permitted by such p o l a r i z a t i o n phenomena. (28 30) The suggested ' reaction scheme i s as follows: FeU) * F e 2 + + 2e , (1) C a 2 + + 2e * Ca X (2) Ca X t (Ca) slag or [Ca] (3) and at high current d e n s i t i e s : A l 3 + + 3e t [Al] (4) F e 2 + + 2e t . F e ( £ ) (5) It has also been c l a r i f i e d that other reactions with higher decomposition potentials than those allowed by the above scheme w i l l not take place. (26) From the above reaction scheme i t has been envisaged that during the anodic h a l f cycle iron oxides w i l l form adjacent to the metal surface either by discharge of oxy-gen ions or by d i s s o l u t i o n of Fe which w i l l replace Ca-ions l o c a l l y . This leads to a high iron oxide a c t i v i t y ( a F e 0^ W 1 t h consequent d i s s o l u t i o n of oxygen into the iron bulk. Simult-2+ 3+ aneously FeO or Fe (or Fe ) w i l l be transported by the hydrodynamic regime into the bulk of the slag causing a grad-ual increase in the a„ n . The cathodic half cycle w i l l d i s -FeO 3+ 2+ 2+ charge A l , Ca or Fe . Any of these ions will contribute to reduce the a _ in the slag. It can also be established that i f in this 2+ last electrochemical reaction there i s not s u f f i c i e n t Ca or 15 (33) i f Ca i s extensively evaporated a slag deoxidation i s necessary to avoid the electrode s a c r i f i c i a l deoxidation. Up to t h i s point in t h i s review only the reactions as a r e s u l t of the inherent electrochemical nature of the ESR-process have been considered. There are, however, other (38) types of reactions involving the slag atmosphere i n t e r -face which also a f f e c t i t s o v e r a l l reaction patterns. Holz-gruber ( 3 4^ has claimed that i f remelting i s ca r r i e d out under a pure oxygen atmosphere the oxygen content i n ingots ranges from 1.8 to 3.8 times that obtained under an argon atmosphere. In thi s work i t i s also shown that oxygen content i n ESR-ingots i s very dependent on the slag chemis-(1 37) try. Several researchers have established v ' ' that i f remelting i s not c a r r i e d out under an i n e r t atmosphere and i f the slag i s not deoxidized then the oxygen content (and the loss of reactive elements) i n the ingot can only be controlled by the slag (oxygen p o t e n t i a l ) . Miska et a l . ' s r e s u l t s (38) also i n favour of the above theory show that the lowest oxygen content i s controlled s t r i c t l y by the slag chemistry and intermediate oxygen contents are strongly influenced by the slag-electrode chemistry. While some researchers (38) have found that the i n t r o -duction of oxygen into the slag i s a mass transfer controlled process others d ' 33) believe that in addition to t h i s mechanism there i s an "oxygen sink" (at the slag-metal i n t e r -face) which acts as a dr i v i n g force. It i s thought that the d e s u l f u r i z a t i o n reaction i s controlled by t h i s dual mech-It i s generally accepted that there i s a continuous introduction of iron oxide into the melt, due to the continuous oxidation of the electrode surface. M i t c h e l l v has pointed out that the ESR-reaction pattern i s more strongly influenced by t h i s phenomenon than by the oxygen introduced as a r e s u l t of reactions taking place at the slag-atmosphere interface. Other studies have also shown that by r e f i n i n g electrodes with d i f f e r e n t chemistry (36,39) o r different surface preparation under other-wise equivalent ESR-conditions, d i f f e r e n t composition or d i f f e r e n t mechanical properties are observed. This be-havior although i t i s i n d i r e c t , has been attributed to the introduction of various quantities of ir o n oxide into the slag as (electrode) scaling. (40 41) There are reports i n the l i t e r a t u r e ' x ' ' which indicate that the evaporation of gaseous f l u o r i d e com-pounds in certa i n slag systems also affects the reaction pattern. The reaction: largely contributes to s h i f t the Al 20 3:CaO r a t i o . This reaction becomes very important where the calcium oxide anism (1, 12) (A1 20 ) + 3(CaF 2) t 3(CaO) + 2A1F3+ (8) a c t i v i t i e s are less than 10 -2 (42, 43) Mi t c h e l l (36) 17 has pointed out that i n order to trace the actual s h i f t i n g i n the chemical composition p a r t i c u l a r l y i n slags where the CaF« and AlF have about the same vapour pressure, an ^ 3 2- - 2+ 3+ analysis i n the 0 :F r a t i o as well as the Ca : A l should be considered. Complementary reactions are: (CaF 2) + H 2 0 ( g ) t (CaO) + 2HF+ (9) and 2(CaF 2) + (Si0 2) t (CaO) + S i F ^ (10) Chouldhury et a l . K i i i i ) ±n a recent communication have pointed out that by remelting ingots with low frequency current the S i losses from an a c i d i c slag are n e g l i g i b l e . They notice, however, that f o r a slag where the CaO:Si0 2 r a t i o i s greater than 4, S i losses r i s e to 65% during re-melting. This finding i s also supported i n previous i n v e s t i -(40, 45) m . . , , „. ^, . (39) gations . Tobias' and Bhat's work suggests that reaction (9) can be considered as an additional source of oxygen i n the system. In t h i s work although a mechanism i s not c l e a r l y s p e c i f i e d , i t i s considered that moisture as a condensed phase in moulds, water vapour from the atmo-sphere or chemically bonded moisture i n the flux markedly a l t e r s the recovery of titanium and s i l i c o n i n ESR-ingots. 2.2.2 On the Nature of the ESR-reaction Scheme A great deal of attention has been dedicated to i n -vestigate the o r i g i n , sequence and consequence of the re-actions i n the ESR-process. The major objective of these studies have been to control the ingot composition with-out s a c r i f i c i n g i t s chemical i n t e g r i t y . I t i s conventionally accepted that the broadest c l a s s i f i c a t i o n of reactions taking place during r e f i n i n g can be subdivided as follows: Reactions which are controlled by the oxygen pote n t i a l [Me] + [0] * (MeO) (11) and reactions as a r e s u l t of the exchange between two l i q u i d s : [Me] + (MO) t (MeO) + [M] (12) Among the reactions comprised i n the f i r s t category, type (11), are: [Ca] + [0] t (CaO) ( 4 8 ' 6 2 ) (11-i) 2[A1] + 3[0] t (A1 20 3) ( 3 6 ' 4 8 " 5 2 ' 6 2 ) (11-ii) [Si] + 2 [ 0 ] t (Si0 2) (26, 33,35,36,49-51,63,64) ( n _ i i i ) [Ti] + 2[0] t (Ti0 2) ( 5 2 ' 5 8 ) (11-iv) m i r I -*• /i-i ^ \ (26,33,35.36,63,64) . x[Fe] + y[o] (Fe 0 ) ' ' ' • ' (11-v) x y [Ti] + 3(Ti0 2) t 2 ( T i 2 0 3 ) ( 5 2 ) (11-vi) [Ti] + 2(Ti0 2) t ( T i 2 0 3 ) + (TiO) ( 5 2 ) (11-vii) The group of reactions of the kind (12) ± s subdivided i n the so c a l l e d deoxidation reactions, (12a), namely: [Mn]+ (FeO) t- (MnO) + Fe (57,59-61) (12-i) [Si]+ 2(FeO) t (Si0 2) + 2Fe d,57,60,61) (12-ii) [Til + 2 (FeO) t (Ti0 2) + 2Fe (52,58,60) ( 1 2 - i i i ) 2[A1]+ 3(FeO) t (A1 20 3) + 3Fe (35,48,57,60) (12-iv) [Ca] + (FeO) t (CaO) + Fe ( 6 2 ) (12-v) and the exchange reactions which also involve the reactive species, (12-b): 3[Ti] + 2(A1 20 3) t 3(Ti0 2) + 4[A1] (18,39,46-51) (12-vi) 2[Ti] + (Si0 2) t 2(Ti0 2) + 2[Ti] (52,58) (12-vii) 3[Si] + 2(A1,0-) t 3(SiCO + 4 [ Ai]d/44,46,47,49,55) ^ J Z ( 1 2 - v i i i ) 2[Mn] + (Si0 2) t 2(MnO) + [Si] (44/56,57,76) (12-ix) 2[A1] + 3(MnO) J ( A l ^ ) + 3 [ M n ] ( 5 7 ) (12-x) Other reactions included i n t h i s category observed -(^ 1) i n laboratory (ESW) experiments are (12-C): (CuCl 2) + [Al] t Cu + (A1C13) 3(Cu 20) + 2[A1] t Cu + ( A l 2 0 3 ) (Ni0 2) + Fe t Ni + (FeO) (Si0 0) + 2 Fe t Si + 2(Fe O) (MnO) + Fe t 2 Fe + (Si0 9) C + (FexO) t CO ( g ) + Fe 2 0 Although the depicted categorization of the reaction scheme has been presented i n an oversimplified manner i t s t h e o r e t i c a l basis should be enunciated to e s t a b l i s h • the reacting conditions under which i t takes place and the governing reacting mechanism. A wide range of opinions and sometimes apparently controversial r e s u l t s are found in the l i t e r a t u r e i n re-gard to the approach predicting the ESR-reaction sequence. ( 6 3 ) While some investigator's work support the theory that ( 5 8 ) there e x i s t s a state of equilibrium, other studies ( 3 3 ) based i n thermodynamic data and experimental r e s u l t s have found that either a "dynamic equilibrium" or k i n e t i c fac-tors govern the reaction pattern. It i s well recognized^®^ that i f an ESR-furnace i s considered s t r i c t l y as a reactor even i n the absence of electrochemical factors, i t s nature i s ( 6 5 6 6 ) such that true thermal equilibrium ' i s never reached and hence i t should be instead considered as a reactor which operates under three d i f f e r e n t regimes, namely: 1 ) an unsteady state which holds for about three times the ingot diameter from bottom. 2 ) a quasi-steady state for most of the r e f i n i n g time and 3 ) the hot top stage which i s at the end of the r e f i n i n g time. Other p e c u l i a r i t i e s of the ESR process which influence the reaction sequence are i t s hydrodynamic regime ( 3 0 caused by i t s r a d i a l and v e r t i c a l temperature gradients ' 65,67)^ Thus, s t r i c t l y speaking an actual state of e q u i l brium i s not reached because of the changing thermal cond tions, hence influencing i t s chemical nature. From t h i s d escription and from the electrochemical p r i n c i p l e s a l -ready described i t can be seen that the reaction scheme must be considered according to both thermal regime and i t s r e s u l t i n g k i n e t i c factors. These ideas have been sup ported by Kay's s t u d i e s d ' 4 8 ) o n the behavior of the ESR reaction pattern. In t h i s work i t has been suggested that slag-metal composition relationships are governed by (33) k i n e t i c factors. Hawkins et a l . have c l e a r l y shown that t h i s p a r t i c u l a r state of equilibrium i s not unique but i t can be represented as a thermal-parameter denomin-ated " c h a r a c t e r i s t i c temperature." This parameter which was calculated on a thermochemical basis indicates that the system may see simultaneous reactions at th e i r cor-responding thermal regions. Although t h i s "characteris-t i c temperature" as a parameter does not have any physi-c a l meaning i t does represent the unsteady thermal-chemi c a l behavior of the ESR process. (33 61) It has been proposed ' that o v e r a l l (ESR) re-actions are the r e s u l t of a well defined series of steps which involve mass transfer (di f f u s i o n , convec-t i o n and hydrodynamic flow) and chemical factors (reorgan zation of the r e l a t i v e p o s i t i o n of ions, atoms or mole-cules). The k i n e t i c aspect of electroactive interfaces i s ruled by: the a c t i v i t i e s of reacting species on both sides of t h i s s i t e , d i f f u s i o n c o e f f i c i e n t s , temperature and concentration gradients extending from the i n t e r -(33) face to the l i q u i d iron or slag bulk. Hawkins et a l . have also suggested that a threefold-stage reaction se-quence can be envisaged, namely i) Transport of reactants to the slag-metal i n t e r -face. This step i s attained by d i f f u s i o n and convection of the two contributing phases (Slag and l i q u i d metal). i i ) • Electrochemical reactions which involve the ion-electron exchange process, and i i i ) Transport of reacting products away from the interface. This step i s again ruled as stage ( i ) . Experimental (ESW) work ( 6 ^ has shown that the rate of reaction i s strongly dependent upon the reaction products at the e l e c t r o a c t i v e interfaces. It has been found that l i q u i d miscible reaction products r e a d i l y d i f f u s e and are e f f i c i e n t l y transported away from the reacting interface. Gaseous reaction products which are formed at the electro-active interface slow the reaction by i n t e r f e r i n g with the d i f f u s i o n products through the boundary layer. And s o l i d reaction prod-ucts also slow the reaction rate by blocking off the area available for reaction. Patchet (^D has also investigated " i n d u s t r i a l cases" in ESW involving multi-component electrode and slag systems. It i s noted i n these experiments that reactions can take place simul-taneously and they a l t e r the composition of the refined (28 product. Work carried out by several investigators ' 33, 57, 5 8 ) i n agreement with Patchet's findings has suggested that although the process operates under a (chemical and kinetic) dual regime, the s t a r t i n g thermo-dynamic conditions at least can be used to i n i t i a t e the calculations involved i n predicting the "dynamic equilibrium" conditions. On the other hand, Cooper et a l . have reported that during AC-melting; metal drops leaving the electrode i n terms of sulfur, reach chemical equilibirum with the slag and that the extent of the reactions between l i q u i d pool and slag were almost n e g l i g i b l e . They also claim that the only reactions which occurred at t h i s (latter) s i t e were due to minor temperature and compositional differences with those of the electrode. This apparent contra-d i c t i o n i s c l a r i f i e d by M i t c h e l l ( 2 7 ' 3 0 ' 4 1 ? and Hawkins et al.(58) who have established that although an actual thermodynamic equilibrium may not be attained i n i n d u s t r i a l ESR-operation; the k i n e t i c s of the pro-cess, however, are so favorable that a state close to e q u i l i -brium i s reached. It has also been found that low. v i s c o s i t y slags and highly e f f e c t i v e reacting area enable the (49 53 reactions to reach such a state of n e a r - e q u i l i b r i a ' ' 6 3 7 8 ) (168) ' . Several studies on C a F 2 ~ A l 2 0 3 slags have shown that although this slag system has high v i s c o s i t y concentrations of A l , Si and Mn i n ESR-ingots have the tendency to achieve the t h e o r e t i c a l equilibrium. Other researchers ( 6 3 , 6 9 ' 7 0 ^ have also found that regardless of the number of electrochemical or k i n e t i c parameters of the ESR-process, a simple thermo-chemical (equilibrium) approach i s s u f f i c i e n t to pre-(71) d i e t the f i n a l ingot composition. M i t c h e l l who has studied the S i - S i 0 2 reaction, has also sug-gested that i f there i s a constant i r o n oxide source i n the slag i n which y° i s high then i t i s f e a s i b l e that the e f f e c t i v e oxygen p o t e n t i a l w i l l be that of the Fe-FeO equilibrium. Perhaps the most t y p i c a l work which supports the (ESR) equilibrium theory i s Holzgruber and Petersen's (72, 73)_ I n ^h^g W O r k i t has been es-tablished that sulfur removal i s e n t i r e l y dependent upon the slag chemistry. Other studies along the l i n e s of the previous work was also carried out by Miska et (38) a l . • . These researchers have also agreed with Holzgruber et a l . ' s proposal. Miska et a l . ' s findings obtained by remelting low a l l o y s t e e l using A l as a deoxidizer were not i d e n t i c a l to Holzgruber et a l . ' s who used d i f f e r e n t electrode chemistry and Si-deoxidation practices. They have attributed these differences to the deoxidation practice and also to the presence of Mn and A l i n the electrode. The Si-SiC^ reaction has been studied under several (ESR) slag systems. Holz-gruber^ 7 5^, Holzgruber and Plockinger ^7*^ and Miska (38) and Wahlster have found that i n i n d u s t r i a l ESR-furnaces t h i s reaction reaches a state of equilibrium. (75) The influence of several a l l o y s and the influence (38) of the A12C>2 in the slag i s also shown. Kusamichi (77) et a l . ' s findings in agreement with previous work, have also shown that the Si-SiC^ behavior i s l i n e a r l y related to the b a s i c i t y index of the slag. Although there are indications of the v a l i d a t i o n of the thermo-chemical equilibrium achieved during r e f i n i n g i t should be pointed out that this equilibrium i s very temperature-dependent^ 4 8^. Retelsdorf and Winterhager ^ 9 ^ f Boucher and (79) Jager and Kuhnelt have studied the Al-A^O^ re-action and conclude that the Al and the oxygen content from ESR-ingots indeed follow the the o r e t i c a l (thermo-dynamic) equilibrium. (46 47 49-51) Abundant information v ' ' ' exists i n the l i t e r a t u r e which establishes the v a l i d i t y of the (83) equilibirum-reaction theory. Rehak et al . ' s studies on the A l and Si d i s t r i b u t i o n i n ESR-ingots also favour the thermodynamic approach of the ESR-reaction system. They claim that reactions involving these species are ruled by the electrode composition and by the i n d i v i d u a l a c t i v i t y of the components of the slag. (76) Holzgruber and Plockinger's work and Choudhury (44) et a l . ' s are also i n agreement with Holzgruber and (72 73) Petersen's ' . Their thermodynamic approach on slag chemistry as a function of the a c i d i t y - b a s i c i t y concept and A l deoxidation techniques i n i n d u s t r i a l p r a c t i c e , c l e a r l y reveal that indeed equilibrium i s reached. Kamardin et a l . who have refined 0.38 wt. % carbon low a l l o y Cr-Mo steels deoxidized with aluminum have concluded that despite t h e i r approximate method to c a l -culate the a c t i v i t y of the slag components (by means of ternary diagrams of the CaD-A^O^-SiG^-system) , the pre-d i c t i o n of the A l and Si-concentrations i n refined ingots can be estimated using a thermodynamic treatment. The previous description has e s s e n t i a l l y been em-phasized on reaction of the type ( 1 1 ) . The other cate-gory which involves reacting species between two l i q u i d s has also been extensively studied. The importance of t h e i r study i s that they largely contribute to a l t e r the chemistry of i n d u s t r i a l ESR - ingots. The modification of the ingot chemistry ("longitudinal seg-27 regation") becomes more c r i t i c a l where high a c t i v i t y of reactive elements i s present. Because of the important ro l e played by the T i in either superalloys or T i -s t a b i l i z e d s t a i n l e s s s t e e l s , a considerable e f f o r t has been devoted to understand the mechanism by which a homogeneous longitudinal T i - d i s t r i b u t i o n i s reached in refined ingots. Although there are evidences in the 1-4. a. (37, 48, 71, 74) . . . . ,. ^ , l i t e r a t u r e • • which indicate the mechanism, there are few studies which c l e a r l y reveal i t . Pateisky's (46 47) studies ' have shown that the reaction (12-vi) indeed attains equilibrium. It i s also indicated that t h i s equilibrium i s strongly affected by the thermal (52) reaction conditions. Krucinski's studies i n agree-ment with Pateisky's have also indicated that the reaction (12-vi) actually reaches a state of e q u i l i -brium. In addition to the above facts Krucinski has also pointed out that Ti-oxides of lower valency i n the slag should be considered, i . e . reactions (11-vi and 11-vii) Another inte r e s t i n g reaction from the i n d u s t r i a l (47 49 55) viewpoint which has been extensively studied ' ' i s the reaction ( 1 2 - v i i i ) . A l l i b e r t et a l . ' who have remelted ingots under SiC^-slags of lower vola-t i l i t y have shown that t h i s reaction does follow i t s stoichiometric r a t i o . The [Al] vs [Si] p l o t whose slope i s 0.781 indeed corroborates t h i s f a c t and also indicates the r e v e r s i b i l i t y which i s reached as a mani-fe s t a t i o n of a state of equilibrium. Kay's s t u d i e s ^ suggest that t h i s reaction i s not influenced by the oxygen p o t e n t i a l . This conclusion i s also supported (81) (78) by Opravil . Choudhury et a l . who have refined ingots (2300 mm i n diameter) using 2.5 Hz AC. ESR have studied the reaction (12-ix). Their r e s u l t s i n agree-ment with Holzgruber 1 s and Holzgruber and Plockinger 1 s ^ c l e a r l y show that the slag b a s i c i t y plays a s i g n i f i c a n t role i n the Si-Mn d i s t r i b u t i o n i n ESR-ingots. Choudhury 1s re s u l t s also show that the type of current does not i n -fluence the ingot chemistry. I t i s also pointed out that Mn losses take place only when the CaOtSiC^ r a t i o i n the slag i s less than 3. They also claim that despite the continuous Al-addition during r e f i n i n g the loss of S i and Mn are unavoidable. The stoichiometry of t h i s reaction was not investigated. (37) Buzek and Hlineny concerned with the low Mn-recovery rates (50-58%) have used isotopic oxygen 18 {O } above the slag to determine the reaction mechanism. Their findings* indicate that by Al-deoxidation im-proved Mn-recovery i s attained. Although they have not speci-f i c a l l y established i t s thermodynamic behavior they did indicate i t s high effectiveness. They claim that the reduction-oxidation mechanism i s ruled by the reaction (12-x). Regarding the so-called deoxidation reactions, Kay ^ has suggested the use of extreme care in any approach since several reactions may operate simultaneously. It has been pointed out that reaction (12-ii) may i n some instances be controlled by reaction (8). Since the re-action (12-i) has a well-defined equilibrium constant of a value close to unity i n the range of ESR-tempera-tures, Fraser has d e l i b e r a t e l y selected i t to e l u c i -date i t s mechanism. He has noted that t h i s reaction may proceed to the l e f t at the electrode-slag interface and subsequently reverse at the higher temperature of the slag-ingot interface. This proposal has led to the b e l i e f that the steady state conditions which are f r e -quently observed i n reactions of t h i s type are a con-sequence of a dynamic balance created by the thermal difference between the two major electroactive s i t e s . M i t c h e l l ( 3 0 ) has extended t h i s proposal to reactions of the type (12-vi). Krichevec et a l . (^O) h a v e studied the S i , T i , Mn and the A l d i s t r i b u t i o n s by r e f i n i n g ingots under several slag systems have found a " s a t i s f a c t o r y " stab-i l i t y of these elements. They claim that these findings are an i n d i c a t i o n of the " p r a c t i c a l " equilibrium conditions attained during r e f i n i n g . Their studies were performed on "deoxidation reactions" (12-iv). 30 As described i n previous sections the formation of iron oxide during r e f i n i n g i s almost unavoidable, e i -ther as a r e s u l t of electrochemical reactions or as a r e s u l t of the reactions between the slag and the atmo-sphere during r e f i n i n g . • These sources of iron oxide in conjunction with the oxide on the electrode surface formed before and during r e f i n i n g , in the presence of the oxide components of the slag generate a l i q u i d with extensive (1 38 82) im m i s c i b i l i t y ' ' . Hence, the a F e 0 r i s e s rapidly to unity at very low "FeO" concentrations. Studies on binary (CaF2~FeO) ^ , ternary (CaF2-CaO-FeO) ^ 1 *, (CaF 2~ Al 20 3~FeO) ^  and i n quaternary (CaF 2-Al 20 3-CaO-Al 20 3-FeO) ^ systems have shown t h i s behavior. A CaF 2 slag can permit very l i t t l e oxygen before i t becomes oxidizing with respect to ir o n . Consequently, any element which forms an oxide more stable would then be oxidized from the metal into the slag. This f a c t becomes more c r i t i c a l where reactive metals such as A l , S i , Zr, and T i are present. Several techniques have been proposed to overcome such a problem: 1) A complete removal of scale on the electrode surface and the use of an i n e r t atmosphere (74) (He or N 2) , 2) painting the electrode surface after scale i s removed with an A l or a magnesia-alumina spinel (38) paint to prevent oxidation of the electrode , 3) en-(83) r i c h i n g electrodes i n oxidable elements such as Zr, A l , S i , etc., 4) continuous additions of a) strong oxide formers as elements (Al, S i , T i , Zr, e t c . ) , b) ferro a l l o y s (FeAl or FeSi) ^ 8 4^ and c) slag-deoxidizer composites (CaF 2-Ca> ( 8 5' 8 6>. (34) Holzgruber suggests that i n addition to the use of a protective atmosphere a more e f f i c i e n t deoxi-dation i s achieved i f a deoxidizer (Al, S i or Ti) i s added i n a slag system which does not contain i t s oxide. Other . (39 , 46, 55, 60) . _ ^ researchers ' ' ' have proposed that i f an element i s prone to oxidation l i k e T i , S i , Zr, during ESR-process then additions of i t s respective oxide pre-vents i t s losses. Kay's ^ findings indicate that by re-f i n i n g a l l o y s which contain A l , T i and Zr i n free-titanium oxide slags the three elements i n the electrode act as deoxidizers. As the amount of Ti02 i n the slag increases the deoxidation i s only c a r r i e d out by Al and Zr. The Al-increment and the Si-decrement i n the ingot are con-t r o l l e d by the exchange reaction ( 1 2 - v i i i ) . Kay has also found that by adding 0.5% Zr.0 2 to the slag i n i t i a l l y containing 15% T i 0 2 i s s u f f i c i e n t to protect the Zr con-(79 ) tent of the refined a l l o y . Jager et a l . have shown that the e f f e c t of the Al-content from the electrode or as a deoxidizer i s very dependent upon the slag chemistry. The net content of A l transported into the ingot i s almost constant for a given A^O^-content (up to 24 wt. %) in the slag and also for a given CaO: S i 0 2 r a t i o (up to 2). If these two parameters are increased the net Al-content of the ESR-ingot increases d r a s t i c a l l y hence leading to (78) deleterious mechanical properties. Chouldhury et a l . (79) and Jager et a l . have claimed that by using an "im-proved technique" a maximum A l content of 0.01 wt % can be attained i n i n d u s t r i a l 125 tonnes ESR-ingots. In these two communications the deoxidation rates, how-(55) ever, are not given. Kajioka et a l . have studied the e f f e c t of Al-deoxidation under several slag systems, namely CaF2~CaO, C a F 2 - A l 2 0 3 and CaF 2-Al 20 3-CaO-Si0 2. They have found that deoxidation rates between 0.05 to 0.1 wt. % A l produce the best " r e s u l t s " , i . e . an almost steady A l - d i s t r i b u t i o n (34) i n the ingot. On the other hand, Holzgruber proposes a deoxidation rate of 6.2 wt. % A l . 2.2.3 Thermodynamic Approach of the ESR-Slag Systems Another important area of the ESR-slag system i s i t s thermo chemical approach. In t h i s f i e l d although very important, very l i t t l e information has been reported. (45) M i t c h e l l has found that in the common system CaO + A1 20 3 + S i 0 2 + CaF 2 there are only three fluoride-containing compounds other than the in d i v i d u a l f l u o r i d e s , i . e . C 1 1 A 7 F ' C 3 A 3 F a n d C 9 S 3 F ' w n e r e c ' A ' F a n d s stand for 33 CaO, A1 20 3, CaF 2 and S i 0 2 respectively. This observation has been taken as an i n d i c a t i o n that the CaF 2 may be con-sidered as an i n e r t diluent in highly basic areas of these systems, since the acid-base interactions involving 2- - (53 0 would be stronger than those for F . A l l i b e r t et a l . ' 54) who have remelted a l l o y s through 70 wt. % C a F 2 ' 30 wt. % A1 20 3 with 0, 2, 5, 10, and 15 wt. % S i 0 2 res-pectively, have applied Mitchell's c r i t e r i a . Their findings show that the acid-base interactions i n the quat-ernary CaO + A l ^ O ^ + S i 0 2 + CaF 2 are approximately the same as those i n the ternary CaO + A^O^ + SiC^ s Y s t e m f d i l u t e d i n the CaF 2. There are other indications i n the 1 -4. t.. *_ , ^ • , (38, 58, 88 , 89) l i t e r a t u r e which also support t h i s theory (14 53 54) A l l i b e r t et a l . ' ' also suggest an alternate method for reactions related primarily to b a s i c i t y , i . e . the CaO/Si0 2 r a t i o concept which i s strongly supported i n the (34, 49-51, 72-76) M . , . , . (38) German l i t e r a t u r e ' ' . Miska and Whalster who have studied the CaF 2-CaO-Al 20 3-Si0 2-FeO system have found that i f the Al 20 3:CaO r a t i o i s greater than 3.0, the CaF 2 at ESR-temperatures remains i n e r t and the slag composition behaves as i t were i n the CaO-Al 20 3 binary. Chai's and Eagar's s t u d i e s o n CaF 2~metal oxide welding fluxes have concluded that the oxidizing p o t e n t i a l of these types of slags i s reduced only by the d i l u t i o n of these species i n the CaF 2 < The addition of CaF 2 i n these 34 slag systems has almost no e f f e c t on the more stable ox-ides and hence the CaF 2 was v i r t u a l l y considered as a diluent. 2.2.4 Overall View on the Modelling of ESR-Reactions (57) A recent work has c l e a r l y revealed the current understanding of the ESR-reaction system. I t has been indicated above that there are e s s e n t i a l l y two ways to ap-proach the reaction pattern. These are the equilibrium reactor and the "single-stage reactor" concepts. The equilibrium-reactor method i s by far a simpler approach. It i s supported by the idea that an actual state of e q u i l i -brium i s reached during r e f i n i n g , as previously i n d i -cated. Its p r i n c i p l e i s to (thermodynamically) e q u i l i b r a t e a (ESR) slag with an electrode of a given composition This technique i n t r i n s i c a l l y assumes a unique equilibrium temp-erature and a mass transfer flow through a fixed slag v o l -ume. These two major assumptions as already indicated, are not necessarily true. The second approach considers a t r a n s i t i o n a l phase con-tact and a "lumped mass-transfer c o e f f i c i e n t . " I t has been pointed out that although t h i s model resembles the non-equilibrium nature of the process i t r e l i e s on a mass-transfer c o e f f i c i e n t which does not have a t h e o r e t i c a l background i n i t s computation. Another disadvantage of t h i s technique i s that i t i s not able to account for reactions of the type (12). M i t c h e l l has also pointed out 2+ that both techniques re l y on the Fe concentrations as an input datum. This parameter, however, i s unknown and has to be determined experimentally. As described here, these two concepts have both ad-vantages and disadvantages. Nevertheless, i t can be un-mistakeably seen that the main objective of these two a l -ternatives i s to predict and control the ingot chemical homo-geneity . 36 2.3 P r e c i p i t a t i o n of Inclusions 2.3.1 General Since endogeneous inclusions are no longer considered as foreign p a r t i c l e s but "natural" components of s t e e l , today's technology demands from metallurgists s k i l l f u l control of them to y i e l d products which could f u l l y sat-i s f y the required stringent standards. A vast amount of (90-98) research has been devoted to study the e f f e c t s of deoxidizers and deoxidation techniques i n the past. It i s known that since elemental oxygen i s highly soluble in l i q u i d iron (0.168 and 0.20 wt. % at the monotectic and at ordinary steelmaking temperatures r e s p e c t i v e l y ) , then an appropriate deoxidizer can be selected to maintain the oxygen content to a given l e v e l . The f i r s t requirement expected from any deoxidizer i s obviously a high c a p a b i l i t y to react with oxygen and i n p a r t i c u l a r cases simultaneously with sulfur i n the melt. The second important requirement i s i t s a b i l i t y to be removed from the melt once oxidation has taken place. „ . .. .. . ,. (99-106) . S o l i d i f i c a t i o n - p r e c i p i t a t i o n studies have shown that the i n c l u s i o n p r e c i p i t a t i o n sequence, as a measure of the degree of deoxidation i s a series of con-tinuous processes. These comprise the nucleation, growth and elimination of the deoxidation products. The nucle-ation phenomena can be homogeneous or heterogeneous and i t can take place i n the l i q u i d stage or during s o l i d i -f i c a t i o n . To homogeneously nucleate a phase i t i s neces-sary to originate a supersaturation and i t can be reached by undercooling of the species i n solution, by additions (9 of either a deoxidizer or oxygen and during s o l i d i f i c a t i o n If there are some s o l i d p a r t i c l e s i n a melt where deoxid-ation products can di f f u s e to then a heterogeneous nucle-ation (precipitation) occurs. At steelmaking temperature some oxides, o x i s u l f i d e s and su l f i d e s are generally found i n solution with l i q u i d s t e e l . Once s o l i d i f i c a t i o n s t a r t s , segregation (due to incomplete d i f f u s i o n i n the solid) i n the in t e r d e n d r i t i c l i q u i d occurs. As soon as t h i s i n t e r d e n d r i t i c l i q u i d i s saturated pre-c i p i t a t i o n of inclusions takes place. P r e c i p i t a t i o n of i n -clusions continues u n t i l the solidus temperature of the melt i s reached. Under p a r t i c u l a r circumstances as w i l l be further described, p r e c i p i t a t i o n takes place at even lower temperature ranges. Inclusion s i z e , shape, quantity and d i s t r i b u t i o n i s p r i n c i p a l l y given by the s o l i d i f i c a t i o n rate and by the s o l u b i l i t y of ce r t a i n chemical species i n the l i q u i d s t e e l . Maximum s o l u b i l i t y of these species influences the p r e c i p i t a t i o n sequence during s o l i -d i f i c a t i o n . As an a p r i o r i rule, for species which have low s o l u b i l i t y i n l i q u i d s t e e l p r e c i p i t a t i o n starts when i n -c i p i e n t s o l i d i f i c a t i o n i s observed. These types of i n -38 elusions w i l l have longer time for growth. Oxi-sulfides and s u l f i d e s which are p r e c i p i t a t e d as a r e s u l t of monotectic reactions are included i n t h i s cate-gory. On the other hand, i f s o l u b i l i t y i s r e l a t i v e l y high p r e c i p i t a t i o n takes place at the l a s t stage of s o l i d i f i -cation and hence these phases are confined to l o c a l i z e d areas. While in the f i r s t case inclusions are globular, large and have large i n t e r p a r t i c l e distances, i n the sec-ond group inclusions are very small and usually located around grains or dendrites. Single phase or composite films around dendrites are also included i n t h i s second kind. The s o l u b i l i t y of some species i n l i q u i d s t e e l can a l t e r the s u l f i d e chemistry and hence i t s p r e c i p i t a t i o n , . (90,106,107) sequence and size ' 2.3.2 Nucleation and Growth of Inclusions 2.3.2.1 Homogeneous Nucleation According to the homogeneous nucleation theory pro-posed by Volmer and Weber-Becker and D o r i n g ^ ® ^ ' ^ the formation of a new phase from a l i q u i d phase occurs only i f supersaturation e x i s t s i n the parent phase. Sev-, , (99,101,110,111) . i T x. e r a l researchers ' • have revealed that the i n t e r f a c i a l tension (a) and the degree of super-saturation (C:C») of a melt influence the nucleation rate. P o p e l ^ 1 1 ^ has calculated the rates of nucleation for i n t e r f a c i a l tensions ranging from 180 to 100 ergs/cm 2 and supersaturation r a t i o s from 1 to 10. His studies show that for systems (FeO-MnO) where the a = 180 ergs/cm 2, 2 8 he obtained a nucleation rate of about 10 at super satur-ation r a t i o s as low as .1.5. With C:C°° r a t i o s of about 10, 35 the nucleation rates were increased to 10 . On the other hand i n systems (FeO-MnO-Si02) where the i n t e r f a c i a l tension i s about 700 ergs/cm 2 the nucleation rate of -79 inclusions i s extremely small (10 ) for supersaturation 40 r a t i o s of less than 3. It becomes appreciable, however, at C:C°° r a t i o s of about 10. For i n t e r f a c i a l tensions of about 1000 ergs/cm 2 as i n the CaO-A^O^-SiC^ system, homogeneous nucleation of inclusions becomes very d i f f i -c u l t . These conclusions are s i g n i f i c a n t for deoxidation during cooling and s o l i d i f i c a t i o n . During cooling of a well-deoxidized melt which does not contain any inclusions there i s a p o s s i b i l i t y of formation of an oxide with lower i n t e r f a c i a l tension, e.g. FeO'A^O^ and FeO-rich s i l i c a t e s . Turpin and E l l i o t using Volmer and Weber's ^^8) c l a s s i c a l theory obtained an expression for c r i t i c a l super-saturation i n terms of the free energy for homogeneous nucle-ation. They have found that hercynite was formed i n iron melts where thermodynamically alumina should have pre-c i p i t a t e d . Thus, they suggested that supersaturation i s required to produce homogeneously nucleated alumina phases. Chipman ^ X 1 1 ^ has suggested that both A l ^ O ^ and FeO'A^O-j may p r e c i p i t a t e simultaneously i f there i s not complete mixing (112-113) i n the melt. McLean and Ward have indicated that hercynite may also p r e c i p i t a t e af t e r an i n i t i a l alumina pre-41 c i p i t a t i o n when the metal i s depleted in A l . The e q u i l i -brium i n t h i s case is between A^O^, FeO-A^O^ dissolved (99) oxygen and the l i q u i d iron. Turpin*s and E l l i o t ' s work shows that i f the c r i t i c a l free energy to homogeneously nucleate A^O^ were much greater than that for hercynite, the p r e c i p i t a t i o n s of the l a t t e r phase becomes more fea-s i b l e . These researchers have also concluded that at high concentrations of deoxidizer or a l t e r n a t i v e l y of oxygen r e l a t i v e to normal bath compositions in equilibrium with pure oxide i s required to overcome the surface tension e f f e c t and thus to p r e c i p i t a t e phases homogeneously. They have also pointed out that the common method of adding the deoxidizer to a molten pool provides s u f f i c i e n t super-saturation to form very small inclusions. This con-clusion was also reached by Turkdogan ^ ^1) showed that i f the deoxidant i s assumed to be evenly d i s t r i b u t e d i n the melt before nucleation. then the supersaturation at-tained would be an order of magnitude less than that neces-sary for homogeneous nucleation. He therefore postulated that the d i s s o l u t i o n of deoxidizer in l i q u i d s t e e l takes a f i n i t e time during which c e r t a i n regions of the melt are expected to be very r i c h i n solute concentration. In these regions l i q u i d metal attains s u f f i c i e n t l o c a l super-saturation for homogeneous nucleation. Investigations c a r r i e d out by Von Bogdandy^ 1 1 4^' on nucleation of aluminum containing phases showed that when the deoxidant i s c a r e f u l l y introduced into the melt the inclusions form i n a layer at a d e f i n i t e time and location. They have postulated that very high super-saturation r a t i o s 14 (10 ) are required for homogeneous nucleation of these oxides. These values are several orders of magnitude (99) higher than those obtained by Turpin 2.3.2.2 Heterogeneous Nucleation In actual practice large degrees of supersaturation are not required i f the p r e c i p i t a t i o n of inclusions take -i , . , . (115) place on s o l i d surfaces From the above discussion i t i s evident that nucle-ation of inclusions i n s t e e l i s not the rate c o n t r o l l i n g step because high supersaturation i s reached near the regions where the deoxidizers are transported into the l i q u i d pool. These nuclei are uniformly d i s t r i b u t e d due to s t i r r i n g of the melt (during refining) and p r e c i p i t a t e further oxides as cooling and s o l i d i f i c a t i o n takes place. However, the presence of pre-existing inclusions from electrodes ( i f any), make heterogeneous nucleation more favorable where r e f i n i n g conditions are not optimum. 2.3.3 Growth of Inclusions By using Shewman's f o r m u l a t i o n ^ one can see that among the phenomena which contribute to growth of i n -clusions from nuclei (0.1 ym) to sizes found i n ESR i n -gots (2-4 ym), the growth by d i f f u s i o n of solutes (oxy-gen and deoxidizers) and p r e c i p i t a t i o n on the nuclei i s the mechanism which i n the least amount of time (^  1-2 sees.) could generate i n c l u s i o n sizes in the range ordin-a r i l y found i n the ESR process. The above calc u l a t i o n s 7 are very dependent on the number of nuclei present (10 cm - 3). Turkdogan u s e ( j a d i f f e r e n t number of nuclei 5 (10 ) at the s t a r t i n g conditions and hence his c a l c u l a t i o n s showed that a d i f f e r e n t maximum radius of inclusions (23 ym) was reached i n a s l i g h t l y higher period of time (6 sees ). The above information plus some other theor-(117) e t i c a l c a l c u l a t i o n s performed by Lindberg and T o r s e l l i n a similar a p p l i c a t i o n of Wert's and Zener 1s model^ 1 1 8^ c l e a r l y indicate that t h i s growth mechanism i s a very f a s t process. Experimental work performed by several re-(107 117 119 120) searchers ' ' ' agree with previous c a l c u l a t i o n s . During cooling and s o l i d i f i c a t i o n the growth of inclusions becomes more s i g n i f i c a n t as p r e c i p i t a t i o n continues on pre-ex i s t i n g p a r t i c l e s . I y e n g a r ^ 1 2 ^ has formulated a model 44 to predict the growth under the above conditions. His re-su l t s have shown that the time for completion of growth i s dependent on the i n i t i a l size of inclusions. This time de-6 7 creases with increase i n radius. With 10 -10 nuclei/cm 3 the growth was almost n e g l i g i b l e . Thus, these r e s u l t s are c l e a r l y i d e n t i f i e d with r e s u l t s found i n E S R ' " ' ' 2 1 ' 27 30) ' l i q u i d pools and ingots, i . e . growth i s almost en-t i r e l y achieved by d i f f u s i o n of solutes. (117 The second phenomenon which has been observed ' 120-122) to largely contribute to the growth of inclusions i n conventional melting practices i s the c o l l i s i o n co-alescence mechanism. This mechanism i s s t r i c t l y related to the motion of oxide inclusions i n l i q u i d s t e e l . For a given size d i s t r i b u t i o n of inclusions i n a melt there i s a d i s t r i b u t i o n of r i s i n g v e l o c i t i e s ' 2 1 ^ . The larger inclusions w i l l l e v i t a t e more rapidly than the smaller , ,,. . (117,122) ones and as a r e s u l t c o l l i s i o n may occur ' The coalescence of inclusions depends on impact, speed and angle, surface properties such as surface ten-sion, chemical composition and t h e i r physical state ( i . e . l i q u i d or s o l i d ) . It has been observed that the motion of inclusions i n s t e e l f a l l s within the viscous flow regime where the Reynolds number of the p a r t i c l e i s less than one. This 45 includes i n c l u s i o n sizes i n the 0-50 ym range. Under t h i s condition, Stokes 1 Law Ust = (13) i s applicable provided the system i s considered to be d i -lute and p a r t i c l e i n t e r a c t i o n can be neglected. The deriv-ation of Stokes' Law^ 1 2 6^ depends upon the following condi-tions: 1) incompressibility of the medium, 2) i n f i n i t e extent of the medium, 3) very small and constant terminal v e l o c i t y , 4) r i g i d i t y of inclusions, 5) absence of s l i p p i n g at the f l u i d - p a r t i c l e interface and 6) sp h e r i c i t y of p a r t i -c l e s . The parameter involved i n t h i s expression are: Ust = terminal v e l o c i t y according to Stokes' Law, (cm/sec.) p , p = density of the p a r t i c l e and the medium res-s pectively, (g/cm3) r - radius of the sphere, (cm) y = v i s c o s i t y of the medium, (g/cm. sec.) The above conditions are very d i f f i c u l t to s a t i s f y i n r e a l i t y . Corrections, however, can be applied to ac-count for the deviation from i d e a l i t y . Table (I) l i s t s the a v a i l a b l e corrections. Stokes' Law i s a p p l i c a b l e ^ 1 2 6 ^ provided that gravity i s the only external force acting on inclusions. If an inclu s i o n i s i n a melt as i n the ESR 46 process, which i s i t s e l f i n motion, then the drag on a spherical p a r t i c l e i s given by: Dc = -6iryr (v-u) (14) where v = v e l o c i t y (vector) of sphere due to gravity and u = r e s u l t i n g v e l o c i t y (vector) due to movement of the sur-rounding f l u i d . In order to obtain an expression for the v e l o c i t y of the inclusions r e l a t i v e to a fixed coordinate system i t i s necessary to know the r e s u l t i n g v e l o c i t y (u) at every location i n the melt. Iyengar's work' 2 1^ has c l e a r l y revealed through sev-e r a l mathematical approaches that growth of inclusions by any other mechanism i s almost n e g l i g i b l e . 47 2.3.4 Sulfides The idea of "hot shortness" was conceived i n the l a s t century. This topic, however, received more attention (127 128) i n the early 30's ' . The so c a l l e d " s t e e l burning" or "overheating phenomenon" has been studied since then. Research through the years has led to the conclusion that s u l f u r was less detrimental when manganese was present. Since then Mn has been used to modify the s u l f i d e phase. If Mn i s absent or present i n i n s u f f i c i e n t amounts the formation of FeS i s almost i n e v i t a b l e . This phase has a very low melting point (1190° C) and when i t i s combined with i r o n or FeO i t forms a lower melting point eutectic phase (940°C). This eutectic p r e c i p i t a t e s between grains (129-132) or dendrites . I f t h i s s t e e l i s to be heated at a u s t e n i t i c temperatures for further mechanical working then grain detachment o c c u r s ^ ^ " ^ ^ . This problem (133) • (overheating") has been observed even i n ESR-mater-i a l s . Thus Mn has the c a p a b i l i t y to modify the s u l f i d e phase from a l i q u i d "FeS" to a s o l i d "MnS" at hot working temperatures (1900- 1200°C) . This t r a n s i t i o n occurs where (92) the Mn:S r a t i o i s higher than 4 . The i d e a l form for s u l f i d e inclusions i n low carbon steels i s to have a pure MnS-phase which melts at about 1610°C. Three d i f f e r e n t (90 134) kinds of MnS have been usually reported ' , MnS I-III. Their s t a b i l i t y depends on the cooling rate and on the chemistry of the melt (136-139) ^  T ^ e t r a n s - L t i o n f r o m type I to II occurs i f the oxygen content i n the s t e e l i s diminished to l e s s than 100 ppm. The t r a n s i t i o n of type II to III i s obtained by the influence of two parameters, a low oxygen content and a high a c t i v i t y of s u l f u r . It has been reported that the presence of C, S i , P, A l and C a(140 142) p r o m 0 4 - e formation of MnS I I I . Fredriksson and H i l l e r t ' 3 ^ who have studied the Fe-O-S ternary sys-tem have claimed that as a r e s u l t of a cooperative-eutectic reaction where MnS forms as a c r y s t a l l i n e phase together with the Fe-rich phase, a MnS-type IV i s formed. The morphology of these s u l f i d e s has generally been des-cribed as globular,branched rod and idiomorphic for type (93 142) I, II and I I I , respectively ' . The mechanism by which they are formed i s as follows. Sulfide inclusions type I are usually associated with o x i - s u l f i d e s (eutectic type) and are p r e c i p i t a t e d as a r e s u l t of a monotectic reaction. Since p r e c i p i t a t i o n of s u l f i d e s takes place almost simultaneously with the deoxidaton process then an oxide^sulfide and a s u l f i d e (type I) enriched phase w i l l be formed. The phase which p r e c i p i t a t e s f i r s t i s (91 142) richer i n oxygen than the second phase ' . Rimmed or semikilled i n d u s t r i a l ingots (with r e l a t i v e l y high oxygen content and low sulfur s o l u b i l i t y ) contain t h i s 49 type of s u l f i d e . Ingots which have been deoxidized with the t h e o r e t i c a l required amount of aluminum w i l l have low oxygen content and high sul f u r s o l u b i l i t y . This deoxidation practice produces MnS II. This type was i n i t i a l l y thought to be formed as a r e s u l t of a eutectic reaction ^ 1 3 ^ . More recent ideas indicate that i t i s originated by a cooperative monotectic reaction. It i s also believed that t h i s s u l f i d e p r e c i p i t a t e s at i n -(91) c i p i e n t s o l i d i f i c a t i o n stages. K i e s s l i n g has pointed out that since alumina and MnS II are frequently observed together but as d i s t i n c t phases then the alumina phase acts as a substrate for the s u l f i d e (type I I ) . Lower oxygen content and lower sulf u r s o l u b i l i t y are required to p r e c i p i t a t e MnS III i n s t e e l s , than i n MnS II. These conditions are attained because of the excess of Al in solution. This s u l f i d e p r e c i p i t a t e s i n the early stages of s o l i d i f i c a t i o n as a single phase. It has been observed^ 1 4 2^ that the three types are p r e c i p i t a t e d i n between grains or dendrites. Recent communications^ 1 3 8' 1 4 3^ have suggested that MnS IV which presents a "ribbon" pattern i s a modifi-cation of either I - I I I ( 1 3 9 ) or I I ( 1 4 3 ) . Ito et a l . ( 1 3 9 ) who have studied the k i n e t i c s and chemical influence of melts on the MnS-shape claim that type II r e s u l t s from a eutectic reaction and type I and III (type N) are pre-50 c i p i t a t e d from s o l i d s t e e l . A mechanism, however was not given. Steinmetz et a l . d 4 3 ) who have studied the Fe-Mn-O and the Fe-Si-0 systems at 1500° and 1600°C have shown that there are areas of s t a b i l i t y for the three MnS types when the [S] vs. [0] and the [S] vs. [Mn] are plotted. In t h i s work, i t i s also proposed that MnS-morphology i s very dependent upon the l o c a l a c t i v i t y conditions of oxy-gen and the degree of deoxidation (Si or Mn). Thus, i f a succession of d i f f e r e n t conditions i n terms of l o c a l a c t i v i t y or degree of deoxidation of melts i s made then a continuous series of morphological changes i n MnS's would take place, i . e . from spherical (oxisulfide) at high oxy-gen a c t i v i t i e s , r o d l i k e , d e n d r i t i c and "skeleton" shape at medium oxygen a c t i v i t i e s and pseudo and highly crys-t a l l i n e (MnS) at very low oxygen (local) a c t i v i t y . This l a s t " t r a n s i t i o n " (MnS II to MnS III) has been observed when a melt i s deoxidized either by A l , Ca or Ca-bearing (140,142,144,145) a l l ovc: ' ' ' 51 2.3.5 S p e c i f i c Sulfides (91) (92) Kies s l i n g and Lange and Salter and Pickering have established that double s u l f i d e s of the type (Mn, Me)s are frequently found. Me represents any of the following elements T i , V, Ni, Cr, Fe, Zr, Mg and Ca. Salter and Pickering ^ h a v e studied the replacement of Mn by Fe in the double s u l f i d e . They found that i t ranges from 0.5 to 32 wt. % Fe. The maximum s o l u b i l i t y l i m i t , however, (91) disagrees with previous work performed by Kiessling , (41.0). The Cr substitution for Mn i n these s u l f i d e s has also been studied by these researchers. While Salter (92) and Pickering have reported a replacement of Mn by (91) Cr ranging from 0-5 to 25 wt. % (Cr), Ki e s s l i n g who has studied these compounds i n synthetic s u l f i d e s has found a maximum replacement of 26 wt. % Cr. Perhaps the most important double s u l f i d e which has become a focus of attention with the advent of the Ca-injection processes i s the (Ca, Mn)S. There i s abundant information i n the , . ^  . (140, 144, 158) . . . ^ , , . , . . l i t e r a t u r e ' ' which establishes that t h i s com-pound i s pr e c i p i t a t e d on complex Ca-aluminates forming a peripheral envelope. Salter and Pickering d 4^) n a v e found that since CaS i s isomorphous with the MnS (NaCl type of l a t t i c e ) then extensive CaS s o l u b i l i t y i n MnS should be expected. They have also found a lim i t e d solub-i l i t y of FeS (4% Fe) i n the above (CaS-MnS) system. Kiess- , l i n g and Westman' 4^ have determined the existence of a (triangular shaped) m i s c i b i l i t y gap which shows a maxi-mum i m m i s c i b i l i t y at approximately 120 0°C and at 50 Ca/ Ca + Mn, (in a t . % ) . At approximately 1000°C i t i s increased from 2 5 to 8 5 Ca/Ca + Mn. These researchers suggest that t h i s m i s c i b i l i t y gap may divide the (Ca, Mn)S into two types i . e . Ca-rich and Mn r i c h phases. There i s c e r t a i n disagree-ment with respect to the Ca and Mn s o l u b i l i t i e s i n the CaS or MnS phases. Church et al.'^®^ have reported from 3.0 to 6.0 wt.% Mn i n CaS and from 3.0 to 7.0 wt. % Ca i n MnS. Salter and P i c k e r i n g ' 4 * ^ report from 1.0 to 12.0 wt. % and from 1.0 to 19.0 wt. % for Mn i n CaS and Ca i n MnS respect-i v e l y . Eklund s t u d i e s ' ^ ^ ^ supported by K i e s s l i n g and West-man's f i n d i n g s o n the i n i t i a t i o n of corrosion i n i n c l u s -ions, has also i d e n t i f i e d the existence of these phases. In t h i s study i t i s shown that while Ca-sulfide enriched i n Mn remains almost i n e r t the s u l f i d e phase enriched i n Ca was severely attacked. This finding i s i n agreement with Kiess-l i n g and Westman's who have also observed the CaS decompos-i t i o n to hydrogen s u l f i d e and calcium hydroxide i n the pre-sence of water. 53 2.3.6 Oxisulfides 2.3.6.1 The Fe-Q-S System Rosenqvist and Dunicz (161) and Turkdogan et a l . (162) have determined the s o l u b i l i t y of sulfur i n high purity ir o n . I t i s established that the maximum s o l u b i l i t y of sulfur i n d e l t a - i r o n i s 0.18 wt.% at 1365°C. At the same temperature the gamma-iron holds only 0.06 wt.%. The solu-b i l i t y generally decreases with temperature but alpha-i r o n at 913°C holds 0.02 wt.%. The Fe-O-S ternary system has been studied extensively. H i l t y and C r a f t s ^ ^ 3 ^ have determined the liquidus surface based on chemical and metallo-graphic analysis. Their r e s u l t s show that the Fe-FeS-FeO as the most important part of the Fe-O-S system i n the range of p r a c t i c a l i n t e r e s t , i s constituted by a ternary eutectic at 67 wt.% Fe, 24 wt.% S, 9 wt.% 0 and about 920°C. A m i s c i b i l i t y gap which extends into the system from the Fe-0 side to a sulfu r content at 21.5 wt.% with a minimum point ( p l a i t point) at approximately 81.5 wt.% Fe, 16.5 wt.% S, 2 wt.% O and at 1345°C. The s o l u b i l i t y of oxygen i n ir o n at several temperatures has also been deter-mined by H i l t y and C r a f t s ^ ^ 3 ) ^ j t i s indicated that for increasing sulfur concentrations the oxygen content of iro n f i r s t decreases s l i g h t l y up to about 0.1% S and then increases r a p i d l y . The s o l u b i l i t y of sulfur i n iron as a 54 r e s u l t of the oxygen influence has been determined by Turk-dogan et a l . 166)^ Their findings indicate that i n the equilibrium Fe-S-O. saturated with wustite i n the temperature range between 913°C to 1375°C, there i s a pro-nounced expansion on the su l f u r s o l u b i l i t y curve which reaches a maximum of 143 ppm of S at about 1200°C. This value corresponds to almost half of that of the solidus on the Fe-S diagram. Yarwood et a l . ' ^ " ^ have proposed an inc l u s i o n pre-c i p i t a t i o n model for the Fe-FeS-FeO system. Their f i n d -4. - 4 . U 4 - v , u (90,134,135, 163) mgs i n agreement with other researchers ' ' , show that the p r e c i p i t a t i o n sequence as a r e s u l t of the monotectic reaction, i s indeed affected by the %0:%S r a t i o , (0:S), i n the a l l o y . If thi s r a t i o i s greater than 0.05. inclusions w i l l have a wide range of compositions. They also found that while the maximum oxide content i n inclusions was dependent on the i n i t i a l 0:S r a t i o , the minimum content J m i J • (164-166) . -it, , • was not. Kor and Turkdogan's i n q u a l i t a t i v e agree-ment with Yarwood et a l . ' s ' 0 5 ^ indicate that for 0:S ra t i o s greater than 0.12 two l i q u i d s form at a given temp-erature and composition. Their experiments on an al l o y con-taining 0.02% C and an 0:S r a t i o of 0.29, c l e a r l y reveal the presence of the l i q u i d oxysulfide. 55 2.3.6.2 The Fe-O-S-Mn "Equilibrium" H i l t y and C r a f t s ^ 1 2 9 ' i n t h e i r aim to gain a bet-ter understanding of the deoxidation practice and the i n -clusion chemistry have studied the Fe-S-O, Fe-S-Mn, Mn-O-S and the Fe-O-Mn systems to develop the Fe-O-S-Mn system. They have suggested that the Fe-S-0 i s strongly modified i f the Mn content i n the a l l o y i s high enough to enhance the p r e c i p i t a t i o n of the s u l f u r - f r e e oxide and the oxygen-free s u l f i d e s , simultaneously with the s o l i d i f i c a t i o n of the metallic phase. Semi-quantitative work performed on the Fe-O-S system by these researchers indicates that by adding 0.3% Mn the im m i s c i b i l i t y regions from the Fe-O-S and the Fe-S-Mn are encountered and hence a continuous im-miscible region i s formed. They propose that the metal oxide and metal s u l f i d e eutectics are i n i t i a l l y formed close to the iron corner, diagonally intersect the imm i s c i b i l i t y region and meet the "ternary" eutectic. It was also found that the eutectic i n the modified ternary remained almost i n the same location as in the o r i g i n a l Fe-O-S system. Van Vlack et a l . ^ 1 3 ^ also concerned with the "hot shortness" problem have observed that i f the Mn content i s about 0.8% i n the Fe-O-S system, t h i s a l l o y originates during quenching duplex oxide-sulfide inclusions. The s u l f i d e phase enriched i n Mn was a r e s u l t of a primary c r y s t a l l i z a t i o n . It was also noted that i f the Mn content 56 i n the ternary a l l o y (Fe-O-S) was low then FeS and Mn-r i c h oxide phases were pr e c i p i t a t e d . Turkdogan and Kor (164-166) . f , .. . _ m a series of papers have compiled thermodynamic information concerning the Fe-Mn-S-0 system. Oxygen and sulf u r p o t e n t i a l diagrams which are involved i n th i s quater-(93) nary have been developed '. Turkdogan and Kor based on H i l t y and Crafts observations'** 3^ and on t h e o r e t i c a l thermodynamic data developed by Darken and Gurry have pro-posed to represent the s t a b i l i t y phase f i e l d s for the Fe-Mn-S-0 system under several conditions. They have e s t i -mated the coexistence of gamma iron, Fe(Mn)0, FeS, Mn(Fe)S and a l i q u i d oxysulfide, £^ , as the equilibrium (condensed) phases at about 900°C, Figure (3). The a p e S and a M n g were e s t i -mated to be unity and 0.4 respectively and hence the e q u i l i -_3 brium manganese i n gamma ir o n was computed to be lOppm (10 % ) . The four f o l d phase equilibrium involving the gamma iron, Mn(Fe)0, Mn(Fe)S and l i q u i d oxysulfide (£^), considered as perhaps the most important univariant equilibrium i n the quaternary (Fe-O-S-Mn) system has been estimated on the premise of an id e a l l i q u i d oxysulfide solution, i . e . a + a + a _ + a = i . They claim that since i d e a l FeO FeS MnO MnS J mixing behavior i s observed i n the FeO-FeS l i q u i d i n e q u i l i -brium with gamma iron and MnO-MnS and since the FeO-MnO forms an i d e a l solution then the id e a l behavior considered i n the FeO-FeS-MnO-MnS i s a reasonable assumption. The behavior of Mn i n iron under the above conditions i s given i n F i g -ure (3) and Table (II). The reaction scheme used to estab l i s h the phases i n -volved i n such e q u i l i b r i a are: M n 0 ( s ) + F e ( s ) * F e O ( 1 ) + Mn ( s ) (15) M n S ( s ) + F e ( s ) t F e S ( 1 ) + Mn ( s ) (16) MnO ( s ) t MnO ( 1 ) < 1 7) MnS ( s ) t M n S ( i ) ( 1 8 ) It has been elucidated that i f a., _ = a.. _ s 1 (be-MnS MnO cause of th e i r low Fe i n solution and the Mn a c t i v i t i e s or atom fr a c t i o n s i n iron) then an expression i s obtained to represent the Mn content of s o l i d i r o n for t h i s e q u i l i -brium: 1 - N (K, + K_) ( — - — — ) + K-, + K. =1; where N = atom f r a c t i o n . Mn By taking a., „ = 0.4 and a., = 0.5 at the invariant 3 MnS MnO equilibrium located at 900°C and using the above expression, the dotted curve i n Figure (3) i s obtained. The Mn contents of gamma iron i n equilibrium with s o l i d Mn(Fe)S and l i q u i d s u l f i d e , i n curve (k) have been established using the "FeS"-"MnS" phase diagram. Raoult's Law was assumed for the solub-i l i t y of "FeS" i n "MnS". The curves j and k show the strong e f f e c t of oxygen on the melting point of the oxysulfide phase i n the Fe-Mn-O-S (j) and 58 the Fe-Mn-S (k) systems respectively. It i s seen that while a l i q u i d phase (point A) i s formed when the Mn content i s less than 10 % i n the former system ( j ) , the minimum Mn content to suppress the l i q u i d phase on the Fe-Mn-O-S system at 1200°C i s 1%. The invariant equilibrium (j) which i n -volves the gamma iron, Mn(Fe)0, Mn(Fe)S and the l i q u i d oxy-su l f i d e phases i s the most important equilibrium i n the Fe-Mn-O-S system by which the "hot shortness" can be avoided. Turkdogan and Kor have shown that as long as the s t e e l con-tains Mn(Fe)0 and Mn(Fe)S i n equilibrium with the metal, a l i q u i d oxysulfide may form between 900° and 1225°C depending on the concentration of Mn i n solution i n s t e e l . The higher the Mn-content i n solution the higher i s the temperature above which a l i q u i d phase i s present. (165 Experimental work performed by Kor and Turkdogan ' 166) have c l e a r l y shown the above by oxidizing iron (0.34% Mn and 150 ppm S) at about 900°C. They found a de-p l e t i o n of Mn and accumulation of S i n the metal close to the scale-metal interface. These changes bring about the formation of l i q u i d oxysulfide near the surface above 900°C and the p r e c i p i t a t i o n of pyrrhotite below 900°C. Because of low i n t e r f a c i a l tension the l i q u i d oxysulfide has been seen to penetrate into the grain boundaries i n the metal and the scale. They have also observed that most of the l i q u i d phase i s found at the scale-metal interface. These (130, observations have also been reported by other researchers 1 6 9 ) who have studied more complex systems. Turkdogan and Kor who have taken as a basis the i n -formation and technique used to construct figure (4), have extended t h e i r experimental and th e o r e t i c a l data, table (III), to describe the Mn behavior in the presence of other (terminal) phase f i e l d s . Namely the Fe-S-O, Mn-S-0 and Fe-Mn-O. They have suggested that due to the low s o l u b i l i t y of oxygen in iron stable deoxidation prod-ucts i n commercial steels must be present. In the quater-nary Fe-Mn-O-S system the Mn(Fe)0 has been taken as the most stable oxide which i s present at a l l temperatures of int e r e s t . As shown i n Figure (4), the Mn-potential i s given by two invariants at two temperatures, 900 and 1225°C: 1) the invariant given by the in t e r s e c t i o n of the m-n-j univariants which i s constituted by the gamma iron, Mn(Fe)0 as (oxi), FeS, i^, "MnS" and the gaseous phase at about 900°C and 2) the invariant given by the inte r s e c t i o n of the univariants j , p and q. The involved phases i n thi s equilibrium are an Fe/Mn s o l i d phase, "MnS", "MnO", l i q u i d oxide (SL^) and l i q u i d metal {l^) at about 1225°C. The most complete representation of the phase changes i n -cluding the l i q u i d , delta and gamma iron has also been developed by Kor and Turkdogan ^ and i t i s given i n Figure (5). The invariant VII represents the immiscible region i n the Fe-Mn-0 system at 1527°C. It i s important to point out that this invariant was assumed to be equi-valent to the Fe-0 m i s c i b i l i t y gap. The phases i n e q u i l i -brium at VII are delta-iron, Mn(Fe)0, (Oxi), l i q u i d oxide (£ ^ ) and l i q u i d metal " i ^ - • The univariants V and VI were previously described i n Figure (4). The equilibrium phases involved i n the univariant e which are for the Fe-Mn-0 system, are delt a - i r o n "MnO" and l^- Since the "MnO" has a low s o l u b i l i t y product at high Mn-activities (>1% Mn) the equilibrium oxygen i n solution i s so low that the s o l i -dus i n the Fe-Mn-0 i s almost equivalent to the Fe-Mn system. The univariant g was also assumed equivalent to the Fe-Mn binary system. In general i t can be said that the univariants g and f represent the gamma to delta and the delta to l i q u i d i ron transformations respectively. The three new phase f i e l d s in t h i s figure are the delt a - i r o n + Ox + which i s lim i t e d by the e-f univariants, the gamma-iron + Ox + £ 2 which l i e s between f';^ e univariants and the 6 iron + Ox + which i s located between the univariants g and f. The f - f ' univariants correspond to the Fe-Mn-0 system already described i n Figure (4). Crafts and H i l t y ' s w o r k ' 2 9 ' * 6 7) on pseudo-equilibrium i n c l u s i o n p r e c i p i t a t i o n diagrams has also included the representation of the Fe-O-Si-Mn-S system as a pseudo-ternary, Fe(Mn, Si)-0-S. It i s proposed that t h i s system 61 i s integrated by metal-oxide, metal-sulfide and s u l f i d e -oxide m i s c i b i l i t y gaps which int e r s e c t themselves prod-ucing an almost isometric i n t e r n a l t r i p l e l i q u i d region. Equivalent to their proposed Fe(Mn)-0-S pseudo-ternary there are metal-oxide and metal s u l f i d e eutectics which o r i -ginate i n the metal corner and pass through the three-f o l d immiscible region. These binary eutectics continu-ously decrease i n temperature u n t i l they reach the pseudo-t r i p l e e u t e c t i c . It i s also postulated that while the above "ternary" i s useful to represent s i l i c a saturated melts a pseudo ternary which could describe low Si to Mn r a t i o s should be d i f f e r e n t . They claim that since the Mn fluxes s i l i c a a s h i f t of the metal oxide-eutectic towards the oxide corner of the diagram would be expected. I t has also been proposed that since there i s s o l u b i l i t y between "MnS" and MnO-SiC^ then the sulfide-oxide m i s c i b i l i t y gap may disappear and the pseudo-ternary eutectic would move towards the oxide corner. It i s anticipated that more i n t e r -granular s u l f i d e s are expected i n t h i s case than i n the higher s i l i c o n s t e e l s . 62 2.3.6.3 The Fe-Si-O-S-Mn System Other more complex systems have also been generalized i n the modified pseudo-ternaries, namely the Fe(Si)-0-S, Fe(Mn, S i , high Al)-0-S, Fe(Mn, S i , low Al)-0-S and the Fe(Al,Ca)-0-S s y s t e m s ( 1 2 9 , 1 6 7 ) . The extensive e f f o r t dedicated to control the "hot-shortness" problem i s c l e a r l y revealed by Crafts and H i l t y (129) studies . I t i s seen that the search for adequate de-oxidizers which could promote the formation of high melting point phases has been the main aim. The most common feature of these diagrams i s that the l a s t l i q u i d to s o l i d i f y i s a ternary eutectic. These ternary diagrams as t h e i r authors have stated are not true ternaries and hence "the r a t i o n a l i z a t i o n of the problem i s i n t u i t i v e in character and l i a b l e to considerable error, but should be hel p f u l i n the e f f o r t to bring inclusions under control". S i l v e r m a n ^ also concerned with the "hot shortness" problem has studied the Fe-Mn-Si-S-0 system. He has claimed that i f the metallic phase i s r e l a t i v e l y neglected the inc l u s i o n chemistry of the f i v e f o l d system can be reasonably well represented by the MnS-MnO-FeO-Si02 system. This system was " s l i c e d " , as depicted i n Figure (6) i n three ternary planes, shown in Figure (7a-c). Two of these planes are pseudo ternaries the 2FeO«Si0 2-2MnO«Si0 2-MnS and the FeO-MnO-MnS. The t h i r d plane. FeO-MnO•Si02~MnS i s more d i f -63 f i c u l t to analyze because of i t s f i v e primary phase f i e l d s . The "A" area i s present i n the three planes and i t rep-resents the "FeS" s t a b i l i t y f i e l d . The primary c r y s t a l l i z a t i o n product on the 2FeO«SiC>2-2MnO«Si0 2-MnS-plane i s a s o l i d solution consisting of 2FeO«SiC>2 and 2MnO«SiC>2. The second and t h i r d products are the FeS'-phase at "A" and a mixture of two immiscible l i q u i d s at "B". Silverman's observations can be summarized as follows: 1) The MnS-FeO-MnO•Si02 plane shows that l i q u i d s i l i c a t e s enriched i n "FeO" allow more "MnS" i n solution than l i q u i d s i l i c a t e s enriched i n "MnO". 2) This plane also shows that "MnS" i s more soluble i n "FeO" than i n the l i q u i d s i l i c a t e s i n t h i s plane. 3) As the "FeO" content of the l i q u i d s i l i c a t e decreases and the S i 0 2 con-tent increases, the s o l u b i l i t y of "MnS" i n l i q u i d s i l i c a t e decreases. 4) Samples from area "A" i n the MnS-FeO-MnO«Si0 2 plane suggest that the "FeS"-phase s o l i d i f i e s as a eutectic at about 910°C. S i l v e r m a n ^ has concluded that since MnS i s fluxed by s i l i c a t e s and oxides i n the planes studied then inclusions i n th i s system are p a r t i a l l y l i q u i d at r o l l i n g and f i n i s h i n g temperatures. Van Vlack et a l . ' s r e s e a r c h ' 3 ^ i n l i n e with Silverman's work'^ 9^ has semi-quantitatively shown how the s i l i c o n a f f e c t s the in c l u s i o n chemical behavior i n various systems. 64 In t h i s work the metallic phase neglected by Silverman i s now taken into account. Van Vlack et a l . ' s findings can be summarized i n the following points: 1) If s i l i c o n i s added i n "small or moderate" quantities to an Fe-S a l l o y undetectable changes i n i n c l u s i o n shape and composition w i l l be observed. "FeS" was the only phase present. 2) In an Fe-S-0 a l l o y s i l i c o n was also added. In t h i s a l l o y two non-metallic phases were present. One phase was enriched i n "FeS" and the other enriched i n " S i 0 2 " . 3) If s i l i c o n i s added to a ternary Fe-Mn-S a l l o y whose Mn:S r a t i o was 3:1, inclusions remain s o l i d at a u s t e n i t i c temperatures (about 1200°C). Globular inclusions of the type (Mn,Fe)S were observed. 4) If s i l i c o n and oxygen were added to the Fe-Mn-S system (they generate the Fe-Mn-Si-S-0 system) a l i q u i d phase s i l i c e o u s i n character was observed. This l i q u i d phase was associated with saturated MnS. They also observed that i f the Si:0 r a t i o was larger than that required to form Si02 then a glassy type of i n c l u s i o n was formed during cooling. On the other hand, i f t h i s r a t i o was smaller the glassy phase disappeared. Van Vlack et a l . also observed that i f oxygen was i n excess of the s i l i c o n content then the l i q u i d composition was s h i f t e d from the s i l i c e o u s range to a more oxidizing compositions. This l a t t e r phase was similar in nature to that encountered i n the Fe-Mn-O-S-system. This finding was also observed by 65 H i l t y and C r a f t s < 1 2 9 ' l 6 7 ) . Van Vlack et a l . ' s work based on t h e i r own findings and also on Silverman's work (MnS-MnO-FeO-Si02~system) have i l l u s t r a t e d the phase changes i n the l i q u i d phase using Figures (8 a-b) which represent the MnS-MnO-FeS-Si02-system. It has been pointed out that although t h i s system i s q u a l i t a t i v e in character and the metallic phase i s "temporarily" ignored the approach used i n t h e i r experimental work indeed describes phase changes to which the inclusions are subjected. Van Vlack et a l . d 3 ( ^ have proposed that since the l i q u i d phase varies from a s i l i c a enriched composition (A) to a MnO enriched composition (C) then the quaternary (at high temperature ranges) can be represented by a Mn0-Si0 2 b i n -ary equilibrium. Once the s o l u b i l i t y product for S i 0 2 i n the melt i s reached the excess s i l i c o n goes into solution i n the metal and the excess oxygen reacts with Mn to produce MnO (and l i m i t e d l y with iron to produce FeO). These re-action products w i l l generate a composition B i n Figure (8) which fluxes the MnS, as indicated by Silverman's work i n Figure (76). The r e s u l t i n g l i q u i d w i l l be composed of 50 % "MnS" at about 1320°C and hence, during s o l i d i f i c a t i o n a mixture of "MnS" and a s i l i c a t e either tephroite (2MnO*Si02) or Rhodonite (Mn0«Si0 o) depending on the MnO:Si09 r a t i o w i l l p r e c i p i t a t e at low au s t e n i t i c temperature. Van Vlack et a l . ^ 1 3 0 have also observed and q u a l i t a t i v e l y predicted with t h i s s i m p l i f i e d model that the l i q u i d phase becomes a s i l i c e o u s glass i f the l i q u i d contains an excessive amount of s i l i c o n . Under t h i s condition a l i q u i d which i s rapidl y s o l i d i f i e d can remain as a "glassy" i n c l u s i o n at room temperature. Composition C i n Figure (8a) i s attained i f the oxygen content exceeds the s i l i c o n content. Under t h i s circumstance the l i q u i d phase dissolves a considerable amount of "MnS" and the excess oxygen may react with some Mn from the MnS. It i s also believed that some of the iron i s transferred into the l i q u i d to balance the su l f u r . These s h i f t s i n compositions are sketched i n Figure (9). In subsequent s o l i d i f i c a t i o n stages the formation of FeS i s expected. 2.3.7 Oxides 2.3.7.1 Aluminates Experimental and th e o r e t i c a l studies on the Al-0 e q u i l i -brium have been traced i n the l i t e r a t u r e since early i n t h i s century. The d i f f i c u l t i e s presented i n determining the thermodynamic equilibrium are obviously observed i n Figure (10). This summarizes the equilibrium values found by several researchers ( 1 7 ° " 1 8 3 ) at 1600°, 1800° and 1900° C. The l a t e s t equilibrium values given by Gustafsson and Melberg^ 1 7^ .. - , , u (176, 177, are summarized work from several researchers 179—18 2) Gustafsson and Melberg claim that t h i r d order 67 polynomial (regression) parameters should be included i n the determination of the involved a c t i v i t i e s where there e x i s t s strong oxygen-metal (Al or Ca) i n t e r a c t i o n . Their tech-nique unfortunately only works when binary oxygen-aluminum or oxygen-calcium systems are considered. Sims^ 9^ has pointed out that the observed discrepancies can be reconciled on the basis that the oxide phase i n equilibrium with the Fe-O-Al i s not pure A^O^ but instead FeO'A^O^ and hence t h i s spinel phase w i l l always be present. Other studies on t h i s matter have shown that by deoxidizing a melt with aluminum at several lev e l s three general stages can be ob-served: 1) If i n s u f f i c i e n t aluminum i s added to an iron melt ( i . e . there i s an excess of oxygen i n solution) ferrous oxide and hercynite should be p r e c i p i t a t e d as dictated by (112-113) the FeO-Al^O^ equilibrium diagram . 2) At i n t e r -mediate A l - l e v e l s (about 0.4-0.5% Al) a mixture of hercy-* . • - a - 4- ^d83, 184) _. . nite and alumina i s p r e c i p i t a t e d ' . The hercy-nite phase w i l l be the major phase present and 3) At high Al-contents i n an i r o n melt almost pure A^O^ i s f o u n d ' 8 3 ' 1 8 4 ) . According to the deoxidation diagram proposed by H i l t y and F a r r e l l ^^5) a f u n v k i n e < 3 carbon or low a l l o y s t e e l deoxidized with aluminum, the s o l i d i f i c a t i o n of the metal-oxide-sulfide system starts by s o l i d i f y i n g oxide c r y s t a l s (A^O^) instead of metal. As the temperature 68 decreases s o l i d i f i c a t i o n of metal and some oxide takes place simultaneously u n t i l the metal-oxide binary eutectic i s reached. Soli d metal, s o l i d oxide and a phase r i c h i n su l f i d e w i l l p a r t i a l l y p r e c i p i t a t e i n further s o l i d i f i c a t i o n stages. The remaining phase r i c h i n s u l f i d e w i l l f i n a l l y p r e c i p i t a t e , i n the same manner as the Fe-O-S system, as the temperature approaches the ternary eutectic. McLean (112-113) and coworkers who have studied the thermodynamic behavior of the Fe-O-Al system have found that there are clear thermochemical conditions under which either pure A^O^ or hercynite are formed. In thi s work i t has been pointed out that to prevent the formation of hercynite the oxygen a c t i v i t y should be reduced to le v e l s below 0.058 at 1600°C. A c t i v i t i e s of oxygen equal to 0.058 represent the location of the (A^O^-FeO• A^O^) t r a n s i t i o n point. Several researchers have also agreed that k i n e t i c fac-tors are involved in the FeO-FeO'A^O^-A^O^ p r e c i p i t a t i o n sequence. T o r s e l l and O l e t t e ^ 1 2 ^ have observed inclusions i n the submicron sizes one second after aluminum was added. Hammar's'^^ th e o r e t i c a l predictions are i n agreement with T o r s e l l * s and O l e t t e 1 s findings. Hammar1s experimental work, however, did not follow such a behavior. He claims that the f i r s t transformation i s given by FeO'A^O^dT * FeO'A^O-jCs) and i t i s very dependent of the inclu s i o n s i z e . This transformation has been traced 17 seconds afte r the Al-addition. Later deoxidation stages transform the FeO'A^O^s) to almost pure A^O^. The mechanisms proposed to control these transformations are the simultaneous d i f -fusion of oxygen from the p a r t i c l e and d i f f u s i o n of aluminum into the p a r t i c l e . Since, Hammar1s res u l t s were not i n agreement with his theory i t was proposed that inclusions may have a peripheral case of FeO-A^O^Csf which enclose the FeO «A1 20 3 ( i f thus d i f f u s i o n of A l into l i q u i d p r e c i p i t -ates was prevented. This proposal was indeed correct, p n u „ . ,. (184, 186-188) , . . , since EPMA-studies ' showed an enriched Al-case surrounding the FeO'Al^O^ which was o r i g i n a l l y i n l i q u i d state. Another i n t e r e s t i n g observation traced by Hammar was that the FeO-phase was not detected one second aft e r the (18 7) addition of aluminum. Wadby and Salter have noted that a sharp decrease i n oxygen as well as pure alumina inclusions are seen within a period of 30 seconds. Cremer and Driole s t u d i e s ' 8 9 ^ on the influence of electromagnetic s t i r r i n g and the removal of inclusions, have established that a f t e r 20 seconds of the aluminum addition spherical p a r t i c l e s were grown inside c l u s t e r s . Hammar has also ob-served r e l a t i v e l y large FeO-A^O^ inclusions transforming to pure A^O^ i n the period of one to three minutes a f t e r the addition of A l into the molten iro n . * FeO«Al 20 3 i s not intended to indicate stoichiometry and i n fact the FeO«Al 20 3(s) would have a d i f f e r e n t Al 20 3/FeO r a t i o than FeO«Al 20 3(1). 70 Straube and P l o c k i n g e r ^ ' 1 9 1 ^ who have studied the pre-c i p i t a t i o n of alumina i n melts containing some manganese, have stated that the primary deoxidation products are lean i n alumina and contain e s s e n t i a l l y Mn-oxides. The A^O^-content increased very rapidly during the next few seconds u n t i l the composition reached that of the spinel (Fe,Mn)0« A^O-j. It was observed that inclusions were i n a f l u i d or (18 7) p a r t i a l l y f l u i d state. Waudby et a l . have also agreed (190 191) with Straube and Plockinger's findings ' . Waudby (187) et a l . noted that once the spinel type was formed i t reacted with A l i n further stages to produce i r r e g u l a r highly aluminous inclusions which enclosed i r o n . Morgan (188) et a l . who have studied the deoxidation e f f e c t on the i n c l u s i o n chemistry have also observed s i l i c a t e deoxidation products peripherally p r e c i p i t a t e d on A^O^ or FeO'A^O^ phases. Waudby's and Salter's experiments were performed at several Al-deoxidation l e v e l s , namely 0.05, 0.15, 0.3 and 0.5 wt. %. Their samples were quenched and heat treated afterwards for 7 days at 1150°C. Under the above experimental conditions, i t was found that by deoxidizing the iron melt with 0.5% A l i n as cast condition duplex hercynite-alumina inclusions were observed. After the heat t r e a t -ment, however, inclusions dissolved more oxygen u n t i l the hercynite equilibrium composition was reached. Hammar •has also i d e n t i f i e d similar compounds, (Fe,Ni) 'A^O^, 71 ir r e s p e c t i v e of the quenching time. The outstanding crys-(183) tallographic work performed by Watanabe and coworkers (192) along the l i n e s of Sloman's and Evan's work on alum-inum deoxidized melts has c l e a r l y revealed the nature of (183) the transformation i n the Fe-O-Al system. Their work was c a r r i e d out by melting l e v i t a t e d samples under a p u r i -f i e d Ar-atmosphere and s o l i d i f i e d under three d i f f e r e n t cooling conditions. The 0:A1 r a t i o was 1.45 which i s close to the stoichiometric r a t i o found i n the A1 20 3 phase. Ex-tracted inclusions from the met a l l i c matrix were analyzed by X-ray (powder) d i f f r a c t i o n using Cr-Ka ra d i a t i o n . Their r e s u l t s showed that since the FeO«Al 20 3 phase had i n t e r -planar spacing approximately equivalent to the y'-A^O^a') then the y'-A^O^ was though to originate from the spinel phase, as follows: F e O A l 2 0 3 -•• Y ' - A l 2 0 3 ( a ' ) -> y l - A l 2 0 3 ( a ) + Y'-AljOg ( b ) The f i n a l Y , - A 1 2 0 3 i s obtained according to the degree of simultaneous migration of iron and aluminum out and into the o r i g i n a l FeO«A1 20 3 phase. I t was also proposed that since i n the process of K - A 1 2 0 3 formation the Fe0«Al 20 3 phase disappears and Y ' - A 1 2 0 3 (b) shows up, then K - A 1 2 0 3 i s derived from Y ' ~ a 1 2 ° 3 ^ • T n e t h i r d cooling condition i n t h e i r experiments produced Y ' - A 1 2 0 3 (b), 0-Al 2O 3 and a - A l 2 0 3 . The a - A l 2 0 3 phase was the more abundant phase. Hence, the o v e r a l l transformation sequence was proposed to be: FeO-Al 20 3 + y ' - A l 2 0 3 (a') y' A 1 2 ° 3 (a> + ^' ~ A l 2 ° 3 ( b ) "*" *~ A l2°3 -* 6 - A l 2 0 3 a - A l 2 0 3 . Since y' -A12C>3 was found i n a l l the cooling conditions, i t was suggested that the rate of transformation from FeC"Al 20 3 to y '-A12C>3 i s f a s t and that the y '-A^C^ (b) to a - A l 2 0 3 i s r e l a t i v e l y slow. The morphology of Al 2C> 3 type of inclusions has been widely studied. T o r s e l l and O l e t t e ^ 1 2 ^ were the f i r s t researchers who proposed that i n addition to a l o c a l supersaturation, a continuous d i f f u s i o n of aluminum and oxygen into these regions i s required to p r e c i p i t a t e the dendritic type of inclusions. Under these conditions i t has been proposed that d e n d r i t i c alumina was a r e s u l t of a homogeneous nucle-(189) ation. Cremier and Driole have suggested that spherical alumina inclusions are products of heterogeneous nucleation. This type grows i n areas low i n deoxidizer. T o r s e l l and Olette have also proposed that clusters are not formed i n regions enriched i n aluminum. It i s also indicated that c l u s t e r s of alumina which are very commonly observed i n aluminum deoxidized melts are formed due to c o l l i s i o n of single p a r t i c l e s as a r e s u l t of thermal, mechanical or electromagnetic a g i t a t i o n of the molten bath. Cremier and Driole have concluded that the importance of magneto-hydro-dynamic phenomena on the k i n e t i c s of deoxidation of s t e e l i s mainly a physico-chemical process, i . e . shape and type of inclusions i n the Fe-Al-0 system are co n t r o l l e d by the l o c a l flow and a c t i v i t y conditions. Steinmetz et a l . ' 8 4 ^ who have studied the above system under several conditions ( i n d u c t i o n - s t i r r i n g , convection-free and i n the gas phase) have also agreed with that proposal. They found that with an oxygen supply high enough compared to the supply of res-pective elements, l i q u i d phase of high FeO-activity pre-c i p i t a t e . I t i s also suggested that t h e i r growth i s a l -most s t r i c t l y controlled by the flow conditions at which the given region i s subjected. They claim that the spheri-c a l contours become unstable and change to "rosettes" and f i n a l l y to dendrite shape as the "phase-specific concen-t r a t i o n s " or "materials flow" are changed. The degree of l o c a l deoxidation and the type and shape of oxides and s u l -fides have also been shown i n t h i s work. The association of p l a t e - l i k e alumina and the MnS III are given where the l o c a l concentration of aluminum i s higher (2 to 3%). At intermediate deoxidation (0.25-1.0% Al) a mixture of coarse s u l f i d e type II and a c i c u l a r oxide i n the presence of s u l -f i d e are found. Between these two ranges (1.25-1.75% Al) a mixture of MnS II and III i s expected and at very low aluminum contents (0-0.25%) oxysulfides, primary s u l f i d e s , type (184) II and de n d r i t i c oxides may be formed. Steinmetz et a l . have 74 a l s o proposed the mechanism already described for su l f i d e s for the alumina, i . e . 1) dendrites and coarse globular hercynite at high i n i t i a l oxygen contents, 2) i n i t i a l globular to " c o r a l " shape alumina for low oxygen contents under very d r a s t i c cooling conditions. Above 0.019% oxy-gen only d e n d r i t i c A^O^ i s expected. Slower growth i s re-quired for the branched-irregular shape alumina type. They have also proposed that under low i n i t i a l oxygen contents r a d i a t e d - c r y s t a l l i n e to compact n i t r i d e s are grown on the alumina. Braun et a l . ^ l 9 3 ^ , who have studied the influence of s t i r r i n g time, s t i r r i n g rate and i n i t i a l oxygen i n iron melts have c l a s s i f i e d the morphology of the alumina inclusions into f i v e types, namely: de n d r i t i c , faceted, p l a t e - l i k e , spherical and c l u s t e r s . It was found that these types are not exclusively found as a unique type but as a mixture for a given set of experimental conditions. They have also observed that inclusions which integrate the clusters change from de n d r i t i c to p l a t e - l i k e shapes at low oxygen contents to spherical at high oxygen l e v e l s . T o r s e l l and O l e t t e ( 1 2 0 ) , Braun et a l . ( 1 9 3 ) , Okohira et a l . ( 1 9 4 ) , Ooi et a l . ( l 9 5^ and Cremer and D r i o l e ^ 1 8 9 ^ have agreed on the mechanism by which alumina c l u s t e r s are formed, i . e . c o l l i s i o n and coalescence of single p a r t i c l e s as a r e s u l t of f l u i d motion i n the melt. Ooi et a l . 's studies on non-stirred 75 and s t i r r e d melts have shown that under the former condi-t i o n dendritic and alumina clusters are the major types. Spherical inclusions were almost absent. Their second type of experiment produced 1) inclusions larger than 20 ym i n diameter with adhered p a r t i c l e s of about 0.5 -2.0 ym i n diameter and 2) clus t e r s composed of very small spherical inclusions with the maximum diameter of which was about 2 ym. Ooi et a l . ^ 1 9 ~ ^ have corroborated the " c o l l i s i o n coalescence and sin t e r i n g theory" on the formation of alumina cl u s t e r s by measuring the neck growth and assuming volume d i f f u s i o n as the c o n t r o l l i n g mechanism. H i l t y and C r a f t s ^ 1 6 7 ^ have found that 0.5% Mn i n l i q u i d i ron enhances the Al-deoxidation power as much as f i v e times. McLean^ 1 1 3^ suggests that Mn lowers the oxygen and raises the Al concentration for univariant t r a n s i t i o n from (Fe, Mn). A l 2 0 3 to h l ^ O ^ i as suggested by Plockinger ) and Waudby ^ 1 8 7 \ Sims^ 9^ has studied the dual A l - S i deoxidation of melts (0.4% A l and 0.5% S i ) . He noted that the inclusions are c h a r a c t e r i s t i c of those exclusively deoxidized with A l and stronger deoxidation was reached than when the melt i s Al-deoxidized. Sims points out that i f s i l i c o n i s added before or with the aluminum one minute i s s u f f i c i e n t to obtain the maximum cleanliness • Waudby1s and Wilson's s t u d i e s ^ l 9 ^ on progressive deoxidation of iron-oxygen melts by A l - S i a l l o y s (0.3, 0.6 and 0.9% Al-Si) . have 76 found that because of the much more rapid formation of alumina compared with s i l i c a , the alumina content of the inclusions proportionally increase with the A l - S i addition. Hence, they concluded that the dual (Al-Si) deoxidizer behaves i n a complex manner only when a melt i s (Al-Si) deoxidized i n r e l a t i v e l y c r i t i c a l additions. Deoxidation of melts by large quantities of A l - S i behave as conventional additions of the strongest element i n the deoxidizer. As proposed by Sims^ 9^ formation of alumina clust e r s i s expected. G a t e l l i e r et a l . ' 9 ^ i n agree-ment with Waudby's and Wilsons's and Sim's findings have noted that i f aluminum i s f i r s t l y introduced to the metal bath the s i l i c o n remains as a passive deoxidizer. G a t e l l i e r et a l . ' s t h e o r e t i c a l consideration have shown that three d i f f e r e n t behaviors might be observed, at 1600°C: 3/4 1) If the a s i / a A l ° 6 0 ° / alumina should be pr e c i p i t a t e d , 3 /4 2) i f a / a A l " 1400/ pure SiC>2 precipitates and 3) i n the 3/4 intermediate range of these s t a b i l i t y ranges (600 S a c - / aA l i 1400), mullite i s the most stable phase. 2.3.7.2 Calcium aluminates Although the use of calcium as a deoxidizer and de-s u l f u r i z e r i n iron melts has become an a t t r a c t i v e alternative (198) since Sponseller's and Flinn's research i t s use i n the foundry industry, to desulfurize, inoculate and to 77 enhance the s p h e r o i d i z a t i o n o f g r a p h i t e has been used s i n c e e a r l y i n t h i s c entury. S p o n s e l l e r ' s and F l i n n ' s work has been c o n s i d e r e d as one of the most fundamental p i e c e s of r e s e a r c h t h a t has c o n c l u s i v e l y c o n t r i b u t e d to the develop-ment of the Ca-treatment of l i q u i d i r o n . They have a l s o s t u d i e d i t s i n t e r r e l a t e d e f f e c t s w i t h other elements, i . e . A l , C/ N i , S and Au. They found t h a t the s o l u b i l i t y of l i q u i d c a l c i u m i n l i q u i d i r o n under p r e s s u r e i s 0.032% a t 1880°K. At t h i s temperature i t s vapour p r e s s u r e i s 1 . 8 7 ( l 9 8 ) - 1 . 6 4 5 ( 1 9 9 ) atm. and i t b o i l s a t about 1 7 8 0 ° K ( 1 9 9 ) . In l i g h t of S p o n s e l l e r ' s and F l i n n ' s f i n d i n g s i n t e n -s i v e r e s e a r c h has been d e d i c a t e d to improve i t s s o l u b i l i t y , t o overcome the problem of d e n s i t y w i t h r e s p e c t t o l i q u i d i r o n and to d i m i n i s h i t s vapour p r e s s u r e i n the l a s t 20 y e a r s ( 1 4 4 ' 1 5 5 ' 1 9 7 ' 1 9 9 _ 2 0 2 ) . W o r k has been devoted to re d u c i n g i t s vapour p r e s s u r e by a l l o y i n g i t wit h S i , C, A l , Ba or as m u l t i p l e mixtures of v a r i o u s elements, C a A l S i , (9 5) Ca A l S i F e , CaAlBaFe, e t c . P h i l b r o o k who has reviewed the s t a t e o f the a r t of oxygen and i t s r e a c t i o n s with d i f -f e r e n t d e o x i d i z e r s , has d e s c r i b e d the c u r r e n t understanding of Ca-treatment up to 19 77. In June of the same year i n Sweden, the F i r s t I n t e r n a t i o n a l Conference on I n j e c t i o n M e t a l l u r g y took p l a c e . In the proceedings of t h i s meeting^ 9*^ a new d i r e c t i o n on the d e o x i d a t i o n p r a c t i c e i s c l e a r l y seen. 78 Thermodynamic and k i n e t i c theory to support the experi-mental work, f i n a l l y give c r e d i t to the deoxidation capa-b i l i t y of Ca-bearing mixtures. The second conference on (97) i n j e c t i o n processes also held i n Sweden i n 1 9 8 0 , once more, reconfirms the advantages (deoxidizer and desulfur-izer) and disadvantages (low yield) of i t s usage. Holappa^ 9 8^ i n a more recent communication also presents an o v e r a l l view of the Ca-treatment i n the l a d l e . In addition to the previously described advantages of the usage of Ca-bearing mixtures the major att r i b u t e s of t h i s deoxidation practice i n terms of inclusions are: 1) the elimination of c l u s t e r s and angular alumina i n -clusions which otherwise would be formed from the A l -deoxidation p r a c t i c e ' 4 ^ 1 6 8 ) ^ 2) ^he t r a n s i t i o n of MnS type II to MnS III or peripheral calcium s u l f i d e around C a - a l u m i n a t e s ( 1 4 4 ' 1 4 5 ' 1 4 7 ' 1 6 8 ' 1 9 7 ] . The presence of the MnS III has been observed at r e l a t i v e l y low Ca-content (< 20 ppm) and the "CaS" at much higher Ca-t O A , . , ^  ( 1 5 7 , 158, 168) TJ_ . , content (> 20 ppm) i n the melt ' • . I t has also j. * • *.u T * . ( 1 4 6 , 159, 160) _ . been reported i n the l i t e r a t u r e • • that i n addition to the above c h a r a c t e r i s t i c s excellent de-79 (96—98) oxidation, d e s u l f u r i z a t i o n and some dephosphoriz-( 1 5 7 , 1 5 8 , 2 0 4 ) , , , i . . . . ation ' ' are reached by the calcium i n j e c t i o n processes. Gaseous and metallic calcium has been added"to the st e e l stream during tapping ^ 8 , ' p l u n g i n g ' ^ t ( 2 0 4 ) o r s n o o t i n g i t as b u l l e t s i n t o the metal bath. When metallic calcium i s added to the melt i t turns into vapour bubbles which rapidly r i s e to the metal surface and subsequently react v i o l e n t l y with the slag and the oxygen i n the a i r producing considerable f l a r e , splashing and fumes. Burn-off and reactions with the slag and l i n i n g materials reduce the y i e l d of the Ca-treatment to about 1 0 - 5 0 % ( l 9 7 ' 2 0 5 ) m Thus, new means to u t i l i z e calcium as ( 9 7 ) calcium-composite wires or iron tube containing Ca-( 1 5 5 ) alloys have been developed . In addition to these techniques obviously the simultaneous deoxidizer additions (as mixtures of deoxidizers with or without slags) either top or bottom blown into the converter have also been i n d u s t r i a l l y practiced. It i s a general practice in conventional steelmaking, to f i r s t l y deoxidize e f f i c i e n t l y the melt with A l and secondly by the calcium t r e a t m e n t ^ . Regarding the mechanical properties, i t i s generally accepted that globular inclusions improve the anisotropy ( 9 2 9 5 - 9 7 ) of the mechanical properties ' . I t has been pro-posed as a p r i o r i rules that ei t h e r a Ca:S r a t i o greater than 1 . 2 5 ( 1 5 3 ) or more than 2 0 - 3 0 ppm of C a ( l 5 1 ) i s required to 80 achieve the t r a n s i t i o n from the alumina (clusters) to (144) calcium aluminates and the MnS II to at least MnS III or to peripheral calcium s u l f i d e s . G a t e l l i e r et a l . ' 9 7 ^ have found i n experimental melts that a complete elimination of A l 2 0 3 as clust e r s i s attained when the 0:Ca r a t i o (wt %) in inclusions i s about three. This rule was found to be independent of the ingot chemical composition. (209) Faulring et a l . have reported that the nozzle block-age by the aluminate type of inclusions i s eliminated when the Ca:Al r a t i o (in wt.%) i s greater than 0.14. This r a t i o represents compositions which c l o s e l y correspond (133) to the CaO • 2A1 20 3 phase. Boldy et a l . have sug-gested that "burning" which occurs even i n ESR-ingots where the sulfur content i s usually thousandths of a percent might be eliminated only by rare-earth or Ca-treatments. These researchers have reported that the problem was eliminated, i n conventional p r a c t i c e : when manganese su l f i d e s were completely transformed to CaS. (157 158) Japanese researchers ' who have studied the hy-drogen induced cracking (HIC) on pipeline steels have concluded that a Ca:S r a t i o greater than or equal to 1.5 i s required to prevent the reoxidation of Ca during the teeming of the melt. Under these conditions high r e s i s -tant or t o t a l l y insusceptible steels to HIC were developed. It has also been established that Ca-aluminates are also 81 present in steels produced v i a the basic e l e c t r i c arc f u r n a c e ^ 2 ^ ^ . Their presence arises due to the aluminum which i s used as a deoxidizer and the b a s i c i t y of the slag or by the Ca-(Si) treatment. While some work i n the l i t e r a t u r e reports the presence of Ca-aluminates by the Ca-treatment ( 1 4 4' 1 4 7 ' 1 4 8 ' 1 5 1 ' 1 5 3 ' 2 0 7 ) , o t h e r s ( 9 6 ' 9 7< 19 7 20 8) ' have shown that complex calcium-aluminum-silicates (159) are formed. Church et a l . ' s studies on a s t e e l processed under two d i f f e r e n t conditions, a i r melt and deoxidized i n the lad l e and carbon deoxidized i n vacuum have found globular (galaxite) oxides, s u l f i d e s and stringer type of inclusions. In the s t e e l treated under vacuum, inclusions were smaller, fewer and were less complex than i n that treated under a i r . These inclusions consisted of a nucleous of galaxite surrounded by a Ca-Al-Si matrix. Whereas the a i r melted steel contained (Mn, Ca)S around the globular oxides and MnS I with some Ca, Cr and Fe i n solution, i n the s t e e l melted under vacuum only the l a t t e r type was observed. The presence of slag i n some conventional processes and the chemistry of the deoxidizers employed i n the Ca-treatment plus the chemistry of the melt indeed complicate the elucidation of the mechanism(s) by which the i n c l u s i o n chemistry i s controlled. Ja*ger and Holz-(202) gruber have found that a l l o y s of Ca with A l , Mn or S i used as deoxidants i n 18-8 steels after the Al-treatment increase the Ca y i e l d when 0.1 wt.% (Ca + Ba) i s also present. D^type of inclusions consisting of 40-60 wt.% CaO and 60-40 wt.% ^2°3 w i t h peripheral CaS are found under t h i s treatment. Salter and P i c k e r i n g ' 4 0 ^ i n th e i r studies on C-Cr bearing steels dexodized with a CaSi a l l o y and H i l t y and c o w o r k e r s ' 4 4 ' 1 6 8 ^ who used CaSi, CaSiB'a and CaSiBaAl al l o y s to deoxidize an iron melt and CaSiTi to deoxidize some casting melts, have highlighted the p r e c i p i t a t i o n scheme. These researchers have found that inclusions generally obey the sequence of phases given by the pseudo-binary Ca0-Al 20 3 diagram. Although these phases did not necessarily follow a stoichiometric r e l a t i o n s h i p the calciim aluminates i d e n t i f i e d were: CaO*f>Al203 (CAg), CaO-2Al 20 3 (CA 2), C a O - A l ^ (CA) , 12CaO• 7 A l 2 0 3 ( C 1 2 A 7 ) . Salter and Pickering have reported as an exceptional case a Ca-aluminate the composition of which corresponded cl o s e l y to the C^ 2A 7. P i c k e r i n g 1 s ^ studies on de-oxidation i n the ladle, have also found the same sequence of reactions. In these studies the C^A^ was i d e n t i f i e d . Faulring et a l . ^ 2 < ^ 9 ^ and Salter et a l . ' 4 ( ^ and others (201, 207) have agreed that for a given l e v e l of deoxi-dation with calcium a mixture of at least two d i f f e r e n t stoichiometric calcium aluminate phases are found. (144) H i l t y and coworkers have approached the i n -clus i o n formation mechanism as another p a r t i c u l a r case of the general theory to explain the metal-oxide s u l f i d e 83 co-pr e c i p i t a t i o n . These researchers' work coi n c i d e n t a l l y to Salter's and P i c k e r i n g • s ( 1 4 0 J and o t h e r s ( 1 4 7 ' 1 4 8 ' 190, 210) S U g g e s t that since CaO substantially fluxes A^O^ then the CaO decreases the melting point of the " A l 2 0 3 " to produce Ca-aluminates which melt within the range of steelmaking temperatures. H i l t y et a l . ^ ^ 8 ^ propose that the pseudo ternary eutectic i n the i r ternary (metal-oxide-sulfide) diagram should be moved closer to the oxide corner and hence a higher melting point "eutectic" should be expected. The Ca-aluminate p r e c i p i t a t i o n se-quence given by H i l t y and F a r r e l l suggest that t h i s i s modified by the sulfu r content, i . e . while a st e e l con-taining 0.015% S pre c i p i t a t e s (CA 2), a steel with 0.005% S and the same Ca- content ('v, 40 ppm) , i t w i l l p r e c i p i t a t e (CA) . Laboratory and i n d u s t r i a l work performed by Takenouchi (154) and Susuki have agreed with the res u l t s previously described. The shape control of the Ca-aluminates and the disappearance of the manganese su l f i d e s was also obtained. The deoxidizer used was a CaAl a l l o y as wires 4.8 and 7 mm diameter shielded with a s t e e l plate 0.2 mm i n thickness. Emi et a l . ^ X 1 ^ who have used the same deoxidizer and technique, have reported e s s e n t i a l l y the same trend of (144) r e s u l t s as that given by H i l t y and F a r r e l l . The stoichiometric C^A-y phase was c l e a r l y seen at approxi-mately 75-80 ppm of Ca i n the (HSLA) s t e e l . 84 Researchers at (Wakayama works) Sumitomo Metal i n -dustries i n Japan^^3) a i o n g the l i n e s of Takenouchi et a l . ' s and Emi et a l . ' s work (pipeline steels and Ca treatment) have introduced the Ca-alloy by the " A l -b u l l e t shooter" into the Steel contained i n the 160 ton L.D. converter. The CaO to A^O^ r a t i o was equivalent to the above research. The presence of the i n t e r n a l (Ca-aluminates) and the external (A^O^-MnS and CaS) phases have shown a clear dependence on the t o t a l calcium con-tent of the previously (Al)-deoxidized s t e e l . Saxena and coworker's ' research on i n -j e c t i o n of CaO-bearing slags into the (30 kg) melt which was previously deoxidized with aluminum, have shown that by t h i s treatment alumina clu s t e r s into the melt are changed to CaO-A^O^ inclusions and that MnS gradually 'disappears." I t i s also indicated that the inclusions, present aft e r the prefused slag i s injected are spherical calcium (147, aluminates with peripheral s u l f i d e s . Saxena et a l . 1 4 8 J have proposed that as soon as the CaO-bearing slag i s i n contact with the melt two primary reactions take place, namely: mCaO, , . + nAl_0 * t mCaO-'n A1,0,(1) (19) (slag) 2 3 2 3 and 3 C a 0 ( s l a g ) + 2 [ A 1 ] * A l 2 0 3 * ( s ) + 3[Ca] (20-a) 85 where the c o e f f i c i e n t s m and n i n reaction (19) represent stoichiometric factors according to the equilibrium pseudo * binary (CaO-A^O^) phase diagram. A l ^ O ^ represents the primary (Al) deoxidation products. (19 7) (210) G a t e l l i e r et a l . and Holappa also support the deoxidation mechanism given by the reaction (20). Holappa gives an equivalent reaction which comprises both e q u i l i b r i a ; namely Al-deoxidation and Ca-treatment: x[Ca] + y ( A l 2 0 3 ) . n c l u s i o n t xCaO.[y - K r]Al 20 3 + |x.[Al] (20-b) Saxena and c o w o r k e r s ^ 1 4 7 ' 1 4 8 ^ have pointed out that these reactions (19) and (20) or (21) take place insofar as the bath contains s u f f i c i e n t aluminum and hence low oxygen a c t i -v i t y . Since a simultaneous deoxidation and d e s u l f u r i z -ation^ ' ' ' ; i n Ca-injection processes , . , _ (147,148) , , ^  has been observed then Saxena has proposed to represent t h i s equilibrium by the reaction: (CaO)* + [S] = (CaS)* + [0] (21) It i s anticipated that i f the CaO and CaS have unit a c t i -v i t i e s then a^ = 0.0266 a and hence Saxena and co-O s workers predict the p r e c i p i t a t i o n of CaS sol e l y i f the oxygen l e v e l i n the melt i s lower than or equal to 10 ppm. Thus, i f a strong deoxidation i s obtained to reach such oxygen leve l s pure CaS would p r e c i p i t a t e . They propose 86 that since the CaO has also a very high a f f i n i t y for A l 2 0 3 then a series of calcium aluminates would be formed, namely: CaO + 6A1_0_ Z Ca0«6Al_0 o (CA,) (19-a) 2 -i 2 3 6 CaO + 2A1 20 3 t Ca0-2A1 20 3 (CA2) (19-b) CaO + A1 20 3 t CaO-Al 20 3 (CA) (19-c) 12CaO + 7A1 20 3 t 12CaO-7Al 20 3 ( C 1 2 A 7 ) (19-d) 3CaO + 2A1 20 3 t 3Ca0*2Al 20 3 ( C ^ ) (19-e) Although i n these laboratory s t u d i e s ' 4 7 ' * 4 8 ^ oxides en-riched i n calcium were i d e n t i f i e d , reaction (19d) and (19e) which p r e c i p i t a t e the C^ 2A 7 and the C 3A 2 phases were not , , (147,148) ,^ ^ c l e a r l y revealed. Saxena et a l . propose that the CaO-CaF2 slags do not contribute to form CaS on i n -clusions unless the A1 20 3 i s f i r * s t l y transformed into Ca-aluminates. H i l t y et a l . ( 1 6 8 ) , Salter et a l . ( 1 4 0 ) and Nashiwa et a l . ' 5 * ^ have also agreed with Saxena' s proposal. G a t e l l i e r et a l . ( 1 9 7 ) , Saxena et a l . ( 1 4 7 , 1 4 8 ) and Holappa( 2*°) who have studied the deoxidation i n ladles with Ca-treatments have suggested that the in c l u s i o n morphology can be retained i n the f i n a l ingot only i f sources of oxy-gen, which produce reoxidation are r e s t r i c t e d . (159) Church et a l . have proposed that nucleation of Ca-bearing s u l f i d e s take place exclusively on Ca-aluminates. • ^ - . - ^ • , ^  , (147,148) Experimental evidence given by Saxena and Engh shows that as the i n j e c t i o n time of CaO-slags into the melt 87 increases., the amount of s u l f u r i n the (Ca) s u l f i d e phase also increases gradually up to a l e v e l which i s thought to be the maximum sulfur s o l u b i l i t y i n calcium aluminates. As a further corroboration of these observations chemical analyses of samples extracted during the i n j e c t i o n process show a gradual and continuous increment of Ca which reaches a plateau at approximately 20 ppm at l a t e r i n j e c t i o n stages. Saxena's and coworkers previous lab-oratory work on CaO-CaF2 i n j e c t i o n has been extended to i n d u s t r i a l t r i a l s ^ 2 1 4 ) . Although the re s u l t s on a lab-oratory scale have indicated a r e l a t i v e l y high y i e l d , i n terms of transformation of Al 2°3 t o Ca-aluminates and MnS II to Ca-sulfides, the i n d u s t r i a l scale t r i a l s did not show such e f f i c i e n c y . The MnS II was only transformed to duplex-(Ca, Mn)S- s u l f i d e and pure CaS by i t s e l f was not traced. In an apparent disagreement with a l l of the previously , , . (140, 147, 148, 159, 168) . .. described investigations with respect to the required conditions to change the A^O^ and the MnS II morphology, i t has been reported ^ 2 1 ! ^ that "pure" CaS i s formed exclusively a f t e r the "A^O^" content i n the Ca-aluminates i s reduced by Ca to less than about 40.0%, i . e . when the CaO:Al2C>2 r a t i o i s 3.0 or when the 3CaO«Al20.j stoichiometric compound i s formed. I t i s also indicated that once t h i s r a t i o i s reached the CaS i s sharply increased. (209) Faulring and H i l t y have observed CaS i n the pre-88 sence of CaO-A^O^ and CaO^A^O^ as the major and minor compounds respectively. According to the schematic trans-( 9 7 ) formation model proposed by Tahtinen et a l . and sup-( 9 4 ) ported by Holappa i t i s seen that faceted inclusions, probably as the hexagonal Ca-aluminate which corresponds to CaO'GA^O^, represent the i n c i p i e n t t r a n s i t i o n of the a-A^O-j to the Ca-aluminates and the simultaneous tran-(94) s i t i o n of the MnS to (Ca, Mn) ,S. Gustaffson and Melberg (209 217) Faulring et a l . ' have observed these phases i n (15) Ca-treated ingots. M i t c h e l l has also reported these phases i n ESR-ingots. The most comprehensive work which analyzes, on thermo-dynamic p r i n c i p l e s , the p r e c i p i t a t i o n sequence of A^O^ and Ca-aluminates i s that developed by Faulring and Ramalingam^ 2 1^. These researchers have developed a ternary Al-O-Ca equilibrium, isothermal (1550°C, 1823°K) p r e c i p i t a t i o n diagram based on Henrian a c t i v i t i e s . They have established that although diagrams of thi s kind (three components, isothermal and Henrian behavior) are hypo-t h e t i c a l i n nature, these are very h e l p f u l i n the understanding and predicting of the id e n t i t y of inclusions from the chemistry of the bath of vice versa. It i s also emphasized by these researchers, that Henrian a c t i v i t y behavior was assumed due to the incon-sistency found i n the thermodynamic data available for 89 calcium. This three dimensional diagram (h^, ^ c a ' ^ A l ^ was developed by projecting the isothermal Al-O, Ca-O and Ca-Al binary e q u i l i b r i a . Thus, or i g i n a t i n g the sat-urated and unsaturated surfaces which w i l l give the volume of s t a b i l i t y , Figure (11). Thermodynamic data used to construct t h i s diagram i s condensed i n tables (IV, V and VI) . Faulring's and Ramalingam's experimental and t h e o r e t i -c a l findings are condensed i n the following points: 1) The Ca:Al r a t i o determines the i d e n t i t y of the i n -clusion phases. 2) The amount of calcium for a given amount of aluminum varies over a narrow range for each Ca-aluminate phase. 3) If h A ^ > 0.01 i n s t e e l , calcium has a n e g l i g i b l e e f f e c t as a deoxidizer but does a l t e r the composition and thus the morphology of the inclusions and 4) Close control of the Ca:Al r a t i o to obtain a desired Ca-aluminate as the major phase F i n a l l y , Faulring and Ramalingam have determined em-p i r i c a l l y several correction parameters based on the h C a : h A l a n <3 t n e %Ca:%Al r a t i o s from Ca _ Ca • %Ca  h A l ' f A l % A 1 2.3 x 10~ 6 for CA C, CA 0 and CA and when i.e. Ca "Al 90 alumina i s present as one of the phases = 10 x 10 u . r A l 2.3.7.3 Complex Oxides (92) Pickering i n his aim to c l a s s i f y the nature of non-metallic inclusions i n complex s i l i c a t e systems has defined f i v e d i f f e r e n t categories -, namely: 1) Pyroxenes, 2) Olivines, 3) Garnets, 4) Feldspars and 5) Cord-i e r i t e s . At the same time, these categories can be sub-c l a s s i f i e d as follows: 1) Pyroxenes, these are compounds of the type MO«Si0 2. Where M can be Fe, Mn and Mg. Their names are grunerite (FeO«Si0 2), rhodonite (MnO«Si0 2) and enstatite (MgO«Si0 2), respectively. Since there i s extensive solu^ b i l i t y between CaO and MgO i n the presence of S i 0 2 then the diopside (CaO«MgO•2Si02) may be considered as a mix-ture of CaO*Si0 2 and MgO«Si0 2-2) Olivines, t h i s category comprises the same e l e -ments as the previous c l a s s i f i c a t i o n ; t h e i r stoichiometry, however, i s given as: 2 MO«Si0 2. Thus, f a y a l i t e (2FeO-Si0 2), tephroite (2MnO«Si0 2) and f o r s t e r i t e (2MgO«Si0 2) are the main phases of t h i s kind. A compound with Ca i n this category i s not included due to t h i s large ion i c s i z e . It i s anticipated, however, that since the main three phases have complete mutual s o l u b i l i t y they can dissolve up to 50% CaO. 3) Garnets. This series of compounds follows the general stoichiometry given by 3M0«A1 20 3• 3SiC>2 . M i n t h i s case can be Fe, Mn, Mg and Ca. Thus, almandine (3FeO* A1 20 3«3Si0 2), spessartite (3MnO«A1 20 3•3Si0 2), pyrope (3MgO«Al 20 3«3Si0 2) and g l o s s u l a r i t e (3CaO•A1 20 3•3Si0 2) are the s u b c l a s s i f i e d phases in t h i s group. A l l these phases show v i r t u a l l y complete mutual s o l u b i l i t y . 4) The feldspar group includes phases of the general form: MO•A1 20 3•2Si0 2 where M represents Mn and Ca. These phases also show mutual s o l u b i l i t y and take into solution c e r t a i n quantities of FeO or MgO replacing MnO or CaO. 5) Cordierites. This i s a group which encompasses compounds of the following general chemistry: 2MO«2Al 20 3« 5Si0 2. M there represents Fe, Mn and Mg. (91) (92) Kie s s l i n g and Lange i n agreement with Pickering have c l a s s i f i e d the most common compounds i n the CaO-Al 20 3~ S i 0 2 (C-A-S) system, i n the following manner: 1) C«A«S 2 , anorthites. 2) C 2A«S, gehlenite. 3) C 2 « A 2 « S 5 , Ca-Corderite and 4) C 3«A«S 3, g l o s s u l a r i t e . Kiessling and Lange have established that the C 2«A«S, C 3«A«S 3 and the C2* A2* S5 t v P e s a r e n o t common inc l u s i o n phases. ( 9 1 9 2 ) It i s generally agreed ' that to avoid misleadin chemical analysis by EPMA, due to the extensive mutual solu-92 b i l i t i e s and the wide v a r i a t i o n of compositions around the stoichiometric values, i t i s required to know not only the Ca, Al and S i but also the amount of Mn, Mg, Fe and T i . (91) A summary of work performed by Kiessling and Lange i n the C-A-S system i s graphically shown i n Figure (12). The L^ and l i n e s represent the maximum and the minimum MeO:SiC>2 r a t i o s in the phases of the inclusions. Although the binary oxide MeO p r i n c i p a l l y represents CaO, i t can frequently contain various amounts of FeO, MnO and MgO. It has also been pointed out that the central area between L^ and L 2 largely corresponds to the low-melting parts of the above systems. The open c i r c l e s represent chemical composition of indigeneous inclusions determined by K i e s s l i n g and Lange. The major area, on the Al 202-Si02 side are chemical analysis of extracted samples from a Ca-Si A -A- A • a. (207) deoxidized ingots (91 92) It i s suggested ' that to trace the o r i g i n of the deoxidation products a knowledge of the furnace and ladle slag and refractory composition as well as the deoxid-ation practice i s required. Salter and P i c k e r i n g ' 4 ^ have reported that i n c l u s i o n phases i n the range of the C2*A'S-C2*M*S2 types are commonly found i n Ca-Si de-(208) oxidized melts. Other studies on deoxidation of s t e e l with complex deoxidizers (CaSiAl and MgSiAl) have reported i n c l u s i o n compositions as follows: A^O^/ 5.0 - 82.7 %, 93 CaO, 6.6 - 37%; FeO, 1.4 - 6.0%; and S i 0 2 , 2.4 - 64.4%. Lindon and B i l l i n g t o n ^ 1 1 5 ^ have also found that the alumina content i n the C-S-A products increases to approach pure alumina as the A pet. 0: pet. A l added decreases to less than the stoichiometric r a t i o . The chemistry of the deoxidation products indicate that the degree of u t i l i z a t i o n (181 of calcium i s maximum only for a short period of time ' 202 20 5) •' . Hence, i f there i s not s u f f i c i e n t residual c a l -cium i n the melt although CaO i s present i n the deoxidation product, i t w i l l mainly contribute to reduce the a c t i v i t y of s i l i c a and thus to achieve a lower oxygen content. (94) Holappa's re s u l t s also exhibit similar trends. Experiments in t h i s research show that the aluminum i n solution controls the Al 20 3:Ca0 r a t i o and also the S i 0 2 content i n the deoxidation products. If the aluminum i n solution increases from 0.05 to 0.4% the Al 20 3:CaO r a t i o also increases whereas the S i 0 2 gradually decreases i n the endogenous p r e c i p i t a t i o n products. As depicted i n the Ca0-Si0 2~Al 20 3 ternary, t h i s be-(91) havior agrees with Kiessling's and Lange's proposal, i . e . i n c l u s i o n chemistry w i l l follow a tendency to f a l l within the area of low melting point enclosed by and 94 2.4 Inclusions i n ESR-ingots In spite of the improvement i n mechanical properties mainly due to i n c l u s i o n size and quantity obtained by the ESR-process/ l i t t l e work has been published with respect to the i n c l u s i o n chemistry. One of the major drawbacks found in approaching the i n c l u s i o n chemistry either i n s i t u by metallographic or microprobe (EPMA) techniques or by ex-t r a c t i o n methods i s the small i n c l u s i o n s i z e . While en-dogeneous (primary deoxidation products) inclusions i n con-ventional steelmaking practices are 15-40 ym i n diameter, products manufactured by the ESR-technology ; they range between 2 to 10 ym i n diameter. The second major d i f f i c u l t y of t h e i r study i s the complexity of the reaction scheme. The p r e c i p i t a t i o n of inclusions i n the ESR-process has been studied by B e l l ( 2 1 8 ) and M i t c h e l l ( 1 5 ' 2 1 9 ) under d i f -ferent slag and deoxidation practices. B e l l ' s findings on the Fe-O-Al system and CaF 2~30% A^O^ slag, indicate that the f i n a l i n c l u s i o n composition cannot be explained by the c l a s s i c a l nucleation theory applied to the l i q u i d pool, i . e . i n s u f f i c i e n t supersaturation for nucleation. It was observed, however, that t h i s requirement was f u l f i l l e d e xclusively i n l a t t e r stages of s o l i d i f i c a t i o n . (15) M i t c h e l l i n an attempt to influence the p r e c i p i t -ation s i t e i n laboratory ESR-ingots has induced an instant-aneous supersaturation by adding f e r r o s i l i c o n into the 95 metal pool through the slag. Under these conditions, i t was anticipated that the i n c l u s i o n size d i s t r i b u t i o n i n the ingot would show the e f f e c t of growth time and possibly f l o t a t i o n . His r e s u l t s i n agreement with B e l l ' s , however, produced no detectable change i n either oxygen content or i n c l u s i o n d i s t r i b u t i o n . Other studies on F e - C u ( 1 5 ) and F e - N i ( 2 1 8 ) (ESR)-alloys which exhibit well developed dendrites, have shown either inclusions aligned along the i n t e r d e n d r i t i c region or dendrites folded around inclusions. These observations have enabled them to suggest that inclusions i n the f i r s t case were formed at early stages and i n the second case - they might be formed at the beginning of s o l i d i f i c a t i o n . Hence, i t was concluded that deoxidation i n the ESR process occurs by chemical e q u i l i b r a t i o n of oxygen and deoxidant with the slag and since almost a l l inclusions are nucleated and grown i n i n t e r d e n d r i t i c regions during s o l i d i f i c a t i o n , the i n c l u s i o n f l o t a t i o n mechanism was discarded. (15 i M i t c h e l l ' has reported that only very few inclusions from the electrode penetrate to the l i q u i d pool bulk and those which do so are s u b s t a n t i a l l y altered i n composition. It has been widely recognized that due to the surface area available for thermochemical and electrochemical re-actions and temperature differences i n ESR-furnaces; two 96 d i f f e r e n t behaviors i n terms of inclusion chemistry are also observed. Work on the Fe-O-Al and on the Ni-O-Al (218 220) systems performed at U.B.C. ' i n agreement with (38) Miska's and Wahlster's work i n laboratory ESR-furnaces has shown that the Al-0 behavior does not obey the s t o i c h i o -(218) metric r a t i o expected from equilibrium conditions. B e l l who has remelted electrodes i n CaF 2-25% A^O^ under an argon atmosphere has found that i f an (anodic) electrode with a t o t a l oxygen content of 30 ppm i s remelted, a d r a s t i c i n -crement of oxygen i s observed at the droplets 1600 ppm) and t h i s i s reduced down to about 70 ppm af t e r they have passed through the cathodic ingot surface. Under the above conditions hercynite or a mixture enriched i n hercynite and alumina as inclusions would be expected. Miska et a l . ' s (38) work on aluminum alloyed steels remelted i n a labor-atory ESR-furnace under alumina saturated CaF 2~slags has also shown that the oxygen content far exceeded the e q u i l i -2 3 brium, [Al] [0] , product. B e l l has also performed (ESR) laboratory scale experi-ments i n which aluminum deoxidation was carried out. Elec-trodes containing 700 and 30 ppm of oxygen were refined through a CaF 2~25 wt% A l 2 ° 3 s l a 9 ~ s ' deoxidized at various lev e l s (1.15, 10.3 and 44.0 grams). The introduction of deoxidizer i n the melt was carried out by attaching aluminum 97 wires to the electrodes. Results of these series of ex-periments showed an 'apparent equilibrium temperature' of 2000 to 2100°C / rather than 1700°C. Thus, whereas the metal was expected to lose aluminum and oxygen as alumina to the slag sometimes the opposite was observed. I t was suggested that t h i s difference i s due to the i n e f f i c i e n c y of the aluminum deoxidation and that an excess of aluminum i s required to lower the oxygen content. Since calculations have shown that losses of aluminum to atmospheric oxidation would be as important as the slag deoxidation reaction then i t was proposed that the reactions: 2 A 1 ( 1 ) + 3(FeO) t F e ( i ) + A l 2 0 3 ( s ) (12-iv) and 2 A1 ( 1 ) + | ( O ) t A l 2 0 3 ( s ) ( l l ' - i i ) take place. This l a s t reaction was the most probable, since the condensation s i t e would be the cold mold wall where the aluminum i s not refluxed but removed from the system. Miska et a l . ' 8 ^ i n agreement with Burel's findings'"'"' 2 2 0 ^ have found . that the amount of alumina or hercynite i n (ESR) laboratory scale ingots was strongly increased as compared to the i n i t i a l quantity traced i n the electrode. (55) Kajioka et a l . have reported that better cleanliness i s achieved i n larger rather than i n smaller ESR-furnaces. Sev-e r a l explanations have been proposed to account for the above 98 facts; i t i s , however, accepted that electrochemical rather than chemical reactions control the ingot and hence (63) the i n c l u s i o n chemistry. Boucher's work on equivalent slag compositions to the previous works have suggested that i n i n d u s t r i a l ESR-ingots where there i s a larger surface area exposed to the slag, chemical instead of electrochemical reactions govern the ingot and thus the in c l u s i o n chemistry. Boucher's findings, although with some scatter show that there i s a li n e a r r e l a t i o n s h i p between the "FeO" content i n the slag and the aluminum i n the ingot. This behavior was an in d i c a t i o n that thermo-dynamic equilibrium was p r a c t i c a l l y achieved. While B e l l ' s findings suggest an "apparent equilibrium temperature" i n the 1900° to 2000°C range for alumina i n small ESR ingots, Boucher's i n d u s t r i a l scale ESR-experiments suggest that an almost true thermodynamic e q u i l i b r i a i s reached at 1700°C. He also claims that almost pure alumina inclusions were i d e n t i f i e d when the protective (Ar) atmosphere was t i g h t l y maintained throughout the r e f i n i n g period. Retelsdorf and Winterhager^ 9^ i n the i r aim to produce alumina free high carbon ferrochrome and metal chrome by ESR. have found that r e l a t i v e l y large diameter ingots, about 200 mm, contained either a-A^O^ 'C^O^ or a-alumina (corundum) and the aluminum and oxygen contents always cor-responded to the Al-0 equilibrium. Their "unsuccessful" findings were claimed to be due to the high oxidation p o t e n t i a l of the slag (CaF2-CaO-Al2C>3 with 30% A l 2 0 3 ) and to the i n e f f i c i e n c y of the protective atmosphere. (15) M i t c h e l l who has remelted electrodes containing complex (calcium alumina s i l i c a t e ) i n c l u s i o n phases has reported that ingots remelted through CaF 2~20 wt.% A1 20 3 and high aluminum deoxidation (0.5% Al) are prone to con-t a i n calcium bearing inclusions. It has also been reported i n t h i s study that although less than one percent of these inclusions contained s i l i c a , where i t was found i t was as high as 50%. (74) Holzgruber's work on the e f f e c t of s i l i c a of the CaF 2~CaO-Al 20 3 slag system, has found that at low s i l i c a l e v e l s (4.2%) i n the slag, calcium aluminates containing 84% Al 2C» 3, 12% CaO and about 1% s i l i c a with a peripheral s u l f i d e phase—probably (CaMn)s—are commonly found. Another peculiar in c l u s i o n composition (14% S i 0 2 , 32% A l 2 0 ^ and 56% CaO) reported by Holzgruber which also contained a peripheral s u l f i d e phase was obtained under the above experimental conditions, the s i l i c a i n the slag, however was about 12.0 wt.%. Holzgruber's findings indicate that about 12-15% s i l i c a i n the slag y i e l d s the highest Ca 100 content i n inclusions. These re s u l t s were l a t e r confirmed by A l l i b e r t et (53 54) a l . ' . I t should be mentioned that Holzgruber's experi-ments were performed under variable CaOrSiC^ r a t i o s and a l -though s i l i c o n was used as deoxidizer the amount was not s p e c i f i e d . (53 54) A l l i b e r t et a l . ' s studies ' on the acid-basic reactions i n the CaF2-Al202-CaO-SiC>2 slag system i n agree-ment with Holzgruber's findings have reported that at low s i l i c a content i n the slag, calcium aluminates associated with a s u l f i d e phase are c o i n c i d e n t a l l y precipitated. Holzgruber's and A l l i b e r t ' s et a l . ' s findings i n terms of slag and i n c l u s i o n chemical composition are shown i n (53 54) Figures (12) and (13). A l l i b e r t et a l . ' have c l a s s i -f i e d the i n c l u s i o n composition i n oxides plus s u l f i d e s and s u l f i d e s . The f i r s t types are located i n the ternary where a r e l a t i v e l y high a c t i v i t y of s i l i c a i s found, Zone (1). The s u l f i d e zone (2) i s located at moderate s i l i c a a c t i v i t i e s . Holzgruber*s r e s u l t s , c i r c l e s enclosing a s o l i d square, also r e f l e c t similar trends. (74) It has been reported that i f th,j s i l i c a content i n the ESR-slag i s higher than 10%, the percent of s i l i c a i n inclusions w i l l be higher than the percent i n the slag. Sim-ultaneously to t h i s s i l i c a increment the size of inclusions w i l l be larger and hence the t o t a l oxygen w i l l also be i n -creased. (92) Holzgruber i n agreement with Pickering and Kiessling (91) and Lange has also found that higher s i l i c a content 101 in the slag, the alumina s i l i c a t e phase i n inclusions contains either CaO or MnO, but not both compounds together. Holzgruber has also noted that i f r e f i n i n g i s carr i e d out i n alumina free CaO-CaF2 slags small alumina free manganese s i l i c a t e s which contain a maximum of 10 wt % CaO w i l l be the most common type of inclusions found. (83) Rehak et a l . who have studied the e f f e c t of elec-trode and slag chemistry on inclusions i n ESR-ingots have remelted a s t e e l (CSN 19.426 used for cold r o l l i n g rods) produced v i a e l e c t r i c arc furnace under either CaSi or A l deoxidation practices. The chemical composition of i n -clusions i n electrodes Al-deoxidized i s : 92.25 wt% A^O^, 2.23 wt% CaO and 5.43 wt % MgO. In the CaSi deoxidized electrodes i t was as follows: 58.48 wt.% A l 2 0 3 , 19.46 wt. % CaO, 5.36 wt. % S i 0 2 and 16.7% MgO. Three major slag types were selected to refi n e these electrodes; namely basic, neutral and a c i d i c . While the series of aluminum deoxidized ingots showed almost pure A l 2 0 3 i n inclusions and only traces of Ca (less than 1%), s i l i c a was detected exclusively i n one case where the s i l i c a con-tent i n the slag was as high as 25%. The CaSi deoxidized series of electrodes showed, after r e f i n i n g a higher Ca-content (2.41 to 6.91%). The e f f e c t of the s i l i c a content i n the slag was strongly r e f l e c t e d i n the i n c l u s i o n chemistry. The slag and inclusion chemistry are also shown i n Figures (12 and 13). A slag con-taining 50% CaF 2 30% CaO and 20% S i 0 2 and a pure CaF 2 slag, both considered as neutral and the highly a c i d i c slags pro-duced S i 0 2 enriched inclusions (53.0 to 71.36% S i 0 2 ) . Their alumina content ranged from 27.0 to 39.4%. The CaO content varied from 2.4 to 5.0%. The MgO ranged from approximately 1.5 to 3.5%. The CaSi series of electrodes remelted through basic slags showed a s l i g h t l y higher CaO-contents (3.2 to 6.9 5%) than i n the previous ESR ingots. Traces of s i l i c a i n i n c l u -sions of about 0.3 to 0.57% were f o u n d s The MgO contents, on the other hand varied from 5.1 to 10.0% and the alumina content which was the major component, was 86-87%. Rehak et a l . have concluded that oxide inclusions are sequentially formed as their thermochemical a f f i n i t y for oxygen i s dictated. They have suggested that the a-A^O^ (corundum) w i l l form f i r s t and as the a c t i v i t y of aluminum in the melt decreases other elements with lower a f f i n i t y for oxygen, such as calcium-aluminum-silicate inclusions, w i l l subsequently be formed. Thus, the parameters which influence the f i n a l i n -clusi o n chemical composition are the slag chemical comp-(49 51 53 54 74) o s i t i o n and the deoxidation of the elec-t r o d e ( 8 3 ) . Several studies on mechanical properties of ESR-ingots obtained under d i f f e r e n t practices have been recently pub-(133) li s h e d . Work carr i e d out by Boldy et a l . has dealt with the e f f e c t of s u l f i d e inclusions on the "overheating" phenomenon. Overheated materials, as a r e s u l t of p r e c i p i -t a t i o n of manganese s u l f i d e s onto high temperature (1100°-1400°C) aus t e n i t i c grain boundaries show a reduction i n toughness. These researchers describe that a faceted ap-pearance of the fracture surface i s c h a r a c t e r i s t i c of an "overheated" material. It has been found that although r e f i n i n g processes such as ESR and VAR ingots are capable of reducing the sulfur content i n ingots down to very low levels and hence promoting the p r e c i p i t a t i o n of a fine dispersion of s u l -104 fid e s , i t enhances the intergranular (austenitic) p r e c i p i t -ation which causes overheating. Several counteracting measures have been proposed to overcome such a problem, i . e . changes i n ingot chemistry or deoxidation practices or cooling rates other than a i r cooling (^100°C/min). Another communication (221) which has evaluated and compared mechanical properties between VAR and calcium-treated ESR-ingots under several slags and deoxidation rates, has shown that impact strength i s strongly affected by these variables. This work shows that at high calcium deoxidation rates (0.14% Ca) the aluminum and oxygen and p a r t i c u l a r l y the s i l i c o n content are sharply increased during remelting. The ESR ingots exhibited a gradual increment i n the inc l u s i o n size as the Ca-deoxidation l e v e l was increased (0.032%, 0.047% and 0.14% Ca). As a r e s u l t of the above parameters the lowest toughness was found at the highest calcium deoxidation l e v e l s . In these studies neither the inclu s i o n chemical composition nor a self - c o n s i s t e n t explanation as to why the mechanical prop-e r t i e s exhibited such a behavior, have been considered. Although the Ca/CaF 2 solution has been used i n d u s t r i -(86) a l l y i n ESR for removing phosphorous from st a i n l e s s steels (85) and sulfur from rotor steels and as a r e s u l t improved mechanical properties have been reported, the inclu s i o n chemical composition has not been investigated. 105 (222) Viswanatan and Beck have reconfirmed findings from R a t l i f f (223) and Brown i n terms of determining the influence of A l i n the mechanical properties of a rotor (Cr-Mo-V) s t e e l . Their r e s u l t s c l e a r l y show that the presence of aluminum (more than 230 ppm) i n s o l i d solution without forming n i -t r i d e s , markedly reduces the rupture d u c t i l i t y and hence leads to premature f a i l u r e s . CHAPTER III 106 NATURE OF THE PROBLEM A comprehensive work on semi-industrial or f u l l - s c a l e ESR experiments, which would account for a l l of the steps-at the electrode, electrode-slag, s l a g - l i q u i d pool and i n the process of s o l i d i f i c a t i o n during r e f i n i n g and the ef-fects of slags and deoxidizers on the f i n a l ingot and hence the i n c l u s i o n chemistry, has yet not been performed. 3.1 Inclusions i n the Electrode (2-5) Several studies have attempted to elucidate the nature of the transformations of inclusions from electrodes to ingots during r e f i n i n g . Mathematical models^ 9^ which have analysed the thermal history of electrodes have demon-strated that c r i t i c a l thermal gradients are developed at the electrode t i p . Regarding the mechanism by which inclusions i n the electrode are removed, controversial and inconsistent models -,(1,5) ., . (1,12,13,15,16) have been proposed ' . While some researchers ' have reported that inclusions are gradually dissolved as a consequence of the thermal gradients at the electrode t i p , (17 19 20) others •' have reported the opposite. Other research-( ? 3 5 i ers ' have suggested that the elimination of inclusions from electrodes i s by mechanical action. On the other hand, 107 (15) studies performed on inclusions at the l i q u i d f i l m have demonstrated that inclusions do not chemically show any simi-l a r i t y with inclusions i n areas where electrodes experience (22) low thermal gradients. Other studies i n f u l l scale (ESR) electrodes, i n agreement with t h i s proposal, have also sug-gested that there i s a c r i t i c a l length above the l i q u i d f i l m where ce r t a i n volumetric changes of inclusions take place. (19 20) Russian investigators ' have also pointed out that the c h a r a c t e r i s t i c liquidus-solidus length of a l l o y s also plays an important role i n the removal of inclusions from the electrode. As seen i n t h i s summarized review, there i s a vast quantity of q u a l i t a t i v e information but only a l i m i t e d amount of quantitative information a v a i l a b l e . Since the chemical nature of inclusions i s related to t h e i r transformation as a r e s u l t of the steep thermal gradi-ents at the electrode t i p i t has only been approached on a q u a l i t a t i v e basis (in terms of the chemical composition of the electrode) and quantitative analysis (in terms of i n -clusion size d i s t r i b u t i o n s ) , i t was, however, very clear that a deeper study which could r a t i o n a l i z e the reported research, would be very valuable. 108 3.2 The Chemical Influence of the ESR Components on the  Composition of Inclusions T T • (37,55,58,75-77,79) . . - , Various studies have been performed to optimize the chemical homogeneity of ingots manufactured by the ESR technology. It i s widely accepted that i f re-melting i s conducted under atmospheric conditions, an en-hanced and continuous accumulation of iron oxide i n the slag occurs. It i s also known that the oxidative state of the slag with respect to the l i q u i d pool i s related to the slag ^ (38,48,82) system ' ' On an a p r i o r i basis, i t i s understood that i f deoxi-dation i s not c a r r i e d out during remelting then a s a c r i f i -c i a l oxidation of reactive a l l o y i n g elements takes place. Several workers'^'^'^' have pointed out that the chemical composition of the slag strongly enhances or suppresses cer-t a i n reactions during r e f i n i n g . The net production of iron (28 29 58 71) oxides and the production of calcium or alum-inum at the e l e c t r o - a c t i v e interfaces dictates the oxidative state of the molten pool and hence the f i n a l ESR-ingot and i n -clusion chemical composition. It has been proposed that p o l a r i z a t i o n due to current passing through the slag-^skin/mould wall interface, which generates small arc contacts, increases the r e c t i f i c a t i o n i n the A.C. ESR process and hence i t enhances the net as-symetry of reactions (1-5). This p o l a r i z a t i o n increases the 109 2+ Fe content of the slag bulk and thus, accentuates a l l oxid-ation rates i n the system. This proposal has been used to ex-p l a i n the chemical composition of inclusions in the Fe-Al-0 . (219,220) system • ' . The b a s i c i t y index of the slag (as a measure of i t s chemical potential) and i t s e f f e c t on the chemical composition (49-51) of ingots and to a limited extent on the chemical na-ture of i n c l u s i o n s ' 8 ' 7 4 ^ , has been strongly supported i n the German l i t e r a t u r e . These studies, however, have not con-c l u s i v e l y determined the r o l e of the deoxidizer, the chemistry of the slag and/or deoxidizer, and/or the chemistry of the electrode or the combined e f f e c t of these parameters on the incl u s i o n chemical composition of ESR ingots. Several deoxidation practices have been suggested i n / ^  g 74 8 3) the l i t e r a t u r e ^  ' ' . Among the uneconomic deoxidation techniques, to overcome losses of reactive elements (Al, T i , S i , Mn, e t c . ) , i t has been proposed to: eliminate the hard scale from r o l l e d electrodes, deliberately increase the r a t i o of these species in the electrode, use protective paint-ings based on Mg or A l , to deposit oxidative elements on the electrode surface, etc. The most frequent deoxidation tech-nique developed excl u s i v e l y on an empirical basis has been the external addition of either aluminum or s i l i c o n as wire, p e l l e t s , f e r r o a l l o y s , etc. into the slag to achieve a desired ingot chemistry. The conventional widespread deoxi-110 dation l e v e l i s about 0.2 wt. % A l or S i . I t has been pro-(74) posed that e f f i c i e n t deoxidation i s achieved when a de-oxidizer i s introduced into the slag which does not contain i t s oxide. On the other hand, other researchers have pro-(39 48) posed ' that i f an element i s prone to oxidation ( i . e . reactive elements as T i , S i , Zr, A l etc.) during r e f i n i n g , then additions of i t s respective oxide into the slag pre-vents i t s losses. The pot e n t i a l harm of inappropriate de-oxidation w i l l manifest i t s e l f i n an uneven d i s t r i b u t i o n (55 75 79) of c r i t i c a l a l l o y i n g elements i n ingots ' ' and re-(221) s u i t i n deleterious mechanical properties These studies, however, have not approached the re-action mechanisms and hence ingot and inc l u s i o n chemical composition has remained unexplored. In summary, although the need to introduce a deoxidant into the slag, has been i d e n t i f i e d , the net e f f e c t of i t has not been elucidated. Thus, to a large extent the explanation for the ingot chem-i s t r y and hence the composition of inclusions has remained obscure. In addition, since the reaction mechanisms which control the chemistry of ingots have not been completely understood, the exploration of other a l t e r n a t i v e means of deoxidation—with th e i r p o t e n t i a l advantages and disadvant-ages—has never been properly investigated. I l l 3.3 The P r e c i p i t a t i o n of Inclusions from Liquid Pool to  Ingot , . ,. (1,33,36,48,58,61,64) .. . , Several studies ' ' • ' ' ' ' which have ap-proached the reaction scheme i n the ESR process have c l e a r l y demonstrated the existence of oxidation-reduction reactions at the e l e c t r o a c t i v e (electrode t i p - s l a g and s l a g - l i q u i d pool) interfaces which largely contribute to control the (218—220) ingot and the i n c l u s i o n chemistry. Research car-r i e d out at U.B.C. has c l e a r l y revealed the electrochemical nature of i n c l u s i o n p r e c i p i t a t i o n i n the Fe-O-Al system. Chemical analysis (in s i t u by metallographic and electro-microscopic techniques and by extracting inclusions and an-alyzing them by X-ray techniques) have conclusively shown that a state of thermochemical equilibrium i s only q u a l i t a t -i v e l y obeyed. From thermochemical conditions i n slags where alumina was expected as the only type of i n c l u s i o n , a mix-ture of iron oxide and hercynite, hercynite and alumina were found instead of pure alumina. The majority of studies on the chemical composition of inclusions have been performed on samples from the l i q u i d (15) f i l m (at the electrode t i p ) , droplets i n process of form-ation which have been i n contact with the s l a g ' ^ ' ^ " ^ (also at the electrode tip) and from samples of already s o l i d i f i e d ingots. 112 I t should be pointed out that these studies have been performed i n ingots r e f i n e d i n laboratory ESR-furnaces and as previously c i t e d , the surface area a v a i l a b l e for reactions i s smaller than i n i n d u s t r i a l s i z e furnaces. In a d d i t i o n p h y s i c a l l i m i t a t i o n s i n the small furnaces ( i . e . electrode mould wa l l spacing) pose another drawback for e x t r a c t i n g samples from e i t h e r l i q u i d pool or slag. These f a c t o r s have not allowed researchers to elucidate the o r i g i n of i n c l u s i o n s . Thus, the need to investigate the thermochemical or e l e c t r o -chemical influence of the s l a g - l i q u i d pool i n t e r f a c e on the chemical composition of i n c l u s i o n s and therefore to unambigu-ously i d e n t i f y t h e i r o r i g i n i s e s s e n t i a l . 113 3.4 D i s t r i b u t i o n of Inclusions during S o l i d i f i c a t i o n It i s generally accepted that inc l u s i o n size d i s t r i -butions i n ESR-ingots are markedly smaller compared to ingots produced under conventional and most of the secondary (refining) steelmaking processes. While i n conventional processes l o c a l i z e d concen-t r a t i o n of inclusions have been widely reported due to the i r c h a r a c t e r i s t i c c r y s t a l l i z a t i o n mode, i n semi-industrial or f u l l scale ESR-ingots t h i s phenomenon has almost never been reported. Other in t e r e s t i n g features commonly observed i n ESR-ingots obtained under conventional deoxidation practices, (224) are t h e i r i n c l u s i o n size d i s t r i b u t i o n s (about 2 to 12ym) (225) and t h e i r r a d i a l l o c a l s o l i d i f i c a t i o n times , expressed as a regular decreasing v a r i a t i o n of th e i r primary and second-ary dendrite arm spacings along t h e i r r a d i a l directions from (65) the centreline towards the mould wall . Several detailed (218) studies on inclusions i n ESR ingots where r e f i n i n g of a ferrous Ni enriched a l l o y was performed through a CaF^^CaO slag, have found round inclusions aligned along primary dendrite arms. These observations led to the b e l i e f that p r e c i p i t a t i o n of inclusions takes place homogeneously from the i n t e r d e n d r i t i c l i q u i d . Complementary studies also per-(15 219) formed at U.B.C. ' have also corroborated t h i s proposal. In t h i s study i t i s reported that inclusions were located i n in t e r d e n d r i t i c spaces and very r a r e l y were dendrite arms seen 114 folded around inclusions. Since attempts^ 1 5^ to generate a s i l i c o n supersaturation i n laboratory ESR melts did not produce any detectable change i n the i n c l u s i o n size or i n the t o t a l oxygen analysis, i t was proposed that nucleation and growth of inclusions takes place almost exclusively as a r e s u l t of r e j e c t i o n of solutes during s o l i d i f i c a t i o n . It was therefore concluded that under a con-ventional degree of deoxidation the inclu s i o n f l o t a t i o n mech-anism was not applicable, in disagreement with other research-(1,12,35) ers ' ' As previously stated since there are physical l i m i t -ations and d i f f e r e n t k i n e t i c s at the electroactive interfaces i n small ESR furnaces researchers have not been able to ap-propriately monitor a l l of the reactions i n the various com-ponents and stages of r e f i n i n g , a more complete investigation i s required. 115 3.5 Establishment of the Proposal and Objectives Sought  Through th i s Research A clear necessity to understand and thereby to control the sequence of events to which inclusions and the l i q u i d metal are subjected i n the various stages of r e f i n i n g i s re-quired, as seen from the previous review. Thus, i n order to cover a l l the e x i s t i n g gaps and to extend our present under-standing regarding the nature of inclusions i n t h i s f i e l d a series of four questions was addressed: 1) How are electrode inclusions removed? 2) Is the i n c l u s i o n composition controlled by the chemistry of electrodes, slags or deoxidizers? 3) Are inclusions i n the l i q u i d pool the same as in the ingot? and 4) Is the i n c l u s i o n size d i s t r i b u t i o n r e l a t e d to r a d i a l distances from the centreline to the mold wall of ESR ingots. These questions were stated i n such a way that a l l of the phenomena involved i n the process of r e f i n i n g , i n terms of inclusions were addressed. Once the mechanisms which govern the o r i g i n of inclusions (reactions) were determined then the deoxidant, deoxidation technique and the slag chemical composition would be selected and a comparison between re-su l t s as a function of electrode and slag compositions as well as deoxidants could be c a r r i e d out. 116 CHAPTER IV EXPERIMENTAL WORK AND TECHNIQUES 4.1 Experimental Procedure Ingots were refined through several slag systems using laboratory and semi-industrial scale ESR-furnaces des-cribed e l s e w h e r e ^ 2 2 ^ ' 2 2 7 ^ . Electrodes of several chemical compositions and diameters were refined using several slag systems (commercial grades) and deoxidants. Tables (VII) and (VIII) summarize t h i s information. Electrodes 31.75 and 44.75 mm i n diameter were melted in the laboratory size ESR-furnace at melting rates ranging from 1.2 to 1.5 Kg min - 1. 1020, 4340 and rotor (Ni-Cr-Mo) steels which were 76.2, 88.9 and 114.3 mm i n diameter were remelted i n the semi-industrial (200 mm) ESR-furnace at melting rates of approximately 1 Kg min . Both furnaces use l i n e frequency A.C. power. Refining was ca r r i e d out with and without a protective atmosphere. In the f i r s t case the system enclosed an argon gas sh i e l d and deoxidant additions could be made at monitored rates, Figure (14). The deoxidi-zers were granular aluminum 99.99% purity, calcium s i l i c i d e a l l o y s and aluminum 65 wt.% S i , i n the size range of 8-32 mesh- S p e c i f i c compositions are given i n Table (IX). Several deoxidation practices were followed, namely i) Constant addition i n small ESR-ingots, 117 i i ) intermittent additions and i i i ) continuously increasing. The l a s t two practices were performed i n ingots 200 mm i n diameter. Experiments to determine the inc l u s i o n removal mechanisms i n electrode t i p s were carried out i n 1020 mild s t e e l produced v i a acid e l e c t r i c furnace, 89 mm diameter 4340 Ca-Si-Al treated, and 114 mm diameter rotor (Ni-Cr-Mo) ste e l Ca-Si-Al treated. Once r e f i n i n g experiments were com-pleted electrodes were rapidly withdrawn from slags to achieve as f a s t a cooling rate as possible. Electrode t i p s were sectioned and metallographic (optical) and electro-micro-scopic (SEM and EPMA) analysis were performed. Qualitative dispersive-X-ray-spectrum analysis by the SEM and back-scattered and electron composition maps by the EPMA as well as quantitative analysis (by EPMA) on inclusions were car r i e d out. Samples from l i q u i d pools and slags were taken as deoxidation and r e f i n i n g was taking place. Liquid metal samples were extracted by suction from l i q u i d pools and slag samples by means of a small copper c h i l l . Chemical analysis of inclusions i n samples d i s c r e t e l y extracted from l i q u i d pools as well as from s o l i d i f i e d ingots were ca r r i e d out by EPMA. Oxygen analysis by vacuum fusion also on both types of samples were performed. The chemical composition of slag samples were analyzed 118 by standard spectrophotometric techniques. The chemical composition of ingots was ca r r i e d out by spectrographic analysis along the v e r t i c a l axis of ingots. F i n a l l y , complementary experiments to determine where and when inclusions were nucleated and grown were per-formed. S i l i c a tubes which contained either rare earth metals (mischmetal) or Zirconium wires were introduced through the slag while helium was gently blown into the l i q u i d pool to extract l i q u i d metal by the suction technique. After-wards samples were polished and q u a l i t a t i v e and quantit-ative analysis of inclusions were performed. EPMA and SEM studies on inclusions to determine t h e i r chemical composition and the d i s t r i b u t i o n of the i r phases (composition maps) were ca r r i e d out. 4.2 Analysis of Inclusions A t r i p l e spectrometer Jeolco JXA-3A electron micro-probe and an Etec "Autoscan" scanning electron microscope with a dispersive X-ray analyser were used to determine the chemical nature of inclusions. Since the diameter of the electron beam which excites the sample, the specimen current density and the accelerating voltage determine the steadiness and the magnitude of the signals (X-rays, secondary and backscattered electrons, etc.) 119 emitted they were continuously c a l i b r a t e d to give a beam approximately 1.0 ym i n diameter when the ex c i t a t i o n v o l t -age was 25 kV and the specimen current density was about 0.08 yA . Hence, the e l e c t r i c a l - o p t i c a l conditions i n these instruments were cal i b r a t e d i n such a manner that the maxi-mum possible current was obtained i n the smallest electron probe. Aluminum, calcium, s i l i c o n , s u l f u r , manganese, chromium, titanium, and magnesium were determined at 25 kV. Oxygen and fl u o r i n e were determined at 10 kV. The i d e n t i f i c a t i o n of inclus i o n phases was also obtained by the v i s i b l e l i g h t prod-uced by the electron beam and photon radiati o n (cathodo-luminescence). Since compound standards are known to produce more ac-(228229) curate and reproducible analysis ' , compounds l i k e CaC0 3, A^O^, S l 0 2 ' M ( ? 0 ' Z n S a n d pure Mn were used as standards to determine the chemical composition of the in c l u s i o n phases. Raw data (specimen background, accelerating voltage, take-off angle, specimen counts, compound standard information, background from standards, X-ray counting time, etc) were translated into chemical composition, taking into account cor-r e c t i o n for atomic number e f f e c t s , absorption and secondary fluoresecence, by Colby's MAGIC IV program' 3 0* with U.B.C.'s 120 AMDAHL computer. Relative accuracy of ± 7-10% was obtained. These values are i n f a i r agreement with reports avai l a b l e i n the l i t e r a t u r e ' 0 6 ' 2 2 8 ' 2 2 9 ! (±5-7%). I t i s also worthwhile to mention that, as reported i n the l i t e r a t u r e ' 4 0 ' 2 1 6 ' 2 1 7 ^ the calcium to aluminum r a t i o i n inclusions i n the Ca-deoxidized ingots gave a closer indi c a t i o n of the deoxidation sequence. These r a t i o s (Ca:Al) i n agreement with Faulring's et a l . • s ( 2 0 9 ' 3 1 6 ) and Salter's and P i c k e r i n g • s ' 4 0 } findings were found to approximate the corresponding stoichiometric r a t i o s of those phases given by the pseudo-binary CaO-A^O^ diagram. Spectrographic and oxygen analysis of samples from i n -gots and l i q u i d pool and spectrophotometric and c r y s t a l l o -graphic analysis of inclusions (by Debye-Scherrer and d i f -fractometer techniques) extracted from ingots corroborated (207) these findings. Based on a reported work on calcium aluminate inclusions a minimum of twenty and a maximum of 50 single assays were performed to obtain a representative analysis of a sample, i . e . this represents one point on the graphs. The inclusions were grouped according to siz e , chemi-c a l composition, fluorescence, shape and representative quan-t i t i e s . Inclusions smaller than 3 ym i n diameter were chemi-c a l l y analyzed; they were, however only q u a l i t a t i v e l y con-sidered due to the cumbersome in t e r a c t i o n e f f e c t s of the i n -clu s i o n chemical composition and the metal matrix. Aluminum 121 deoxidized ingots contained either FeO-A^C^, pure alumina, or calcium hexaluminate as inclusions phases which were very small single or clustered. Since i n c l u s i o n diameters were less than 6-8 pm inclusions smaller than 3.0 ym were also analysed on a q u a l i t a t i v e basis. Analyses were carr i e d out by scanning diagonally across the sample and the i n -clusions so that re s u l t s s t a t i s t i c a l l y represented the i n -clusion chemistry of the sample. Typical inclusions of a given sample were micro-photographed, o p t i c a l , scanning and backscattered (EPMA) analyses and composition maps were also obtained. 4.3 Total Oxygen Analyses The t o t a l oxygen content of samples extracted from l i q u i d pools and ingots was determined by the standard i n e r t gas vacuum fusion technique using a Leco 537 induction f u r -nace (507-800) and a Leco oxygen analyzer (509-600). Samples weighing 1.0 - 1.5 grams were cut, ground, and washed, u l t r a -s o n i c a l l y cleaned and rinsed with anhydrous 1,1,1, three-chloroethane. Since the accuracy and r e p r o d u c i b i l i t y of t h i s analysis was strongly influenced by the weight of the sample and i t s preparation, a meticulous procedure was followed to obtain a minimum of three assays with a maximum of 2-5% deviation amongst them. If the deviation i n analyses were greater than t h i s , a new set of three analyses was per-122 formed. To obtain an appropriate c a l i b r a t i o n of the analy-t i c a l equipment standards of known oxygen content were analyzed. Analysis on samples containing an upper and lower l i m i t of oxygen as well as a blank test to check the v a r i -ation i n gas (helium) flow rate were continuously performed to maintain them constantly throughout any series of analyses. Samples extracted from l i q u i d pools were c a r e f u l l y selected because p o r o s i t i e s and slag entrapment were occasionally de-tected . 4.4 Inclusion Extraction Method It i s an accepted f a c t that chemical extraction meth-ods of non-metallic inclusions followed by analysis pre-sent several advantages over chemical analysis performed i n situ.(by microprobe or scanning E.M.). The major ad-vantages are the following: 1. A n a l y t i c a l r e s u l t s are far more representative because a much larger sample i s taken for extraction than for i n s i t u methods. 2. Phases can be s p e c i f i c a l l y i d e n t i f i e d after ex-t r a c t i o n . 3. The t o t a l oxygen content and the t o t a l amount of most phases i n s t e e l can be estimated by chemical analy-s i s of the residue. 123 There are, however, several disadvantages as well. 1. Some phases cannot be quantit a t i v e l y extracted. This f a c t i s due to either i n c l u s i o n size (< 5ym i n d i a -meter) or to the chemical nature of inclusions and re-agents (too agressive). 2. Phases containing common elements and amorphous or/and isomorphic structures may int e r f e r e with c e r t a i n analy-s i s . 3. Since i n c l u s i o n size i n ESR-materials i s r e l a t i v e l y small (S 10ym i n diameter) analysis by X-rays i s rather d i f -f i c u l t . Several chemical methods of in c l u s i o n extraction were performed. Among them bromine i n methanol, iodine i n meth-anol, bromine-ester-methanol and iodine-methyl acetate-methanol. From a l l the above methods only the l a s t two were suitable to thi s purpose. The iodine-methanol-methyl acetate however, was found to be the most convenient because of i t s accuracy and r e p r o d u c i b i l i t y . 4.4.1 Apparatus and Experimental Procedure The apparatus used was integrated i n four units. The f i r s t unit (I) was used to pour iodine c r y s t a l s and the methyl acetate-methanol mixture under a protective atmo-sphere consisting of) anhydrous argon. This part of the sys-tem was also used as a di s s o l u t i o n chamber. An u l t r a -sonic cleaner or a magnetic s t i r r e r with con t r o l l a b l e temp-erature served to speed up the iodine d i s s o l u t i o n . The second unit (II) consisted of a m i l l i p o r e f i l t e r which contained a whatman paper No. 50 or a t e f l o n m i l l i -pore f i l t e r (0.5 ym) . The reaction chamber was the third (III) unit. This was a glass container with four outlets which served as a) breathing system b) reagent-level regulator, c) vacuum c o n t r o l l e r and d) argon-flux valve. This cham-ber was inside the ultrasonic agitator which was used to speed up the d i s s o l u t i o n of the metal sample. The fourth unit (IV) was also equipped with another m i l l i p o r e f i l t e r s imilar to the one used i n unit no. 2 but t h i s had a 0.5 ym diameter porous size f i l t e r (made of teflon), Figure (15) . Since humidity i s one of the major concerns i n any halogen-methanol-methyl acetate technique a c a r e f u l pro-cedure was followed. Reagents used such as methanol and methyl acetate were 99.99% i n purity and iodine c r y s t a l s were previously dried i n a dessicator which contained s i l i c a gel and d r i e r i t e . This extraction method was based on Rooney's and Staple (231) ton's , i t was, however, improved i n terms of avoiding cert a i n complexity In i t s design and minimizing the r i s k of contaminating the apparatus by moisture and other vapours ca r r i e d away during evacuation and heating. Combustion tube containing s p i r a l s of copper and nickel as well as s u l f u r i c acid and contaminated iron turnings were also avoided. 125 Experimental Procedure 20-30 grams of an ESR-steel sample free of oxide, cleaned with 1,1,1-trichlorethane and dried with hot a i r were introduced into the reaction chamber. Argon was completely dehydrated by passing i t through 3 U-tubes con-taning d r i e r i t e , s i l i c a - g e l - i n d i c a t i n g and activated carbon i n a U-tuhe immersed i n l i q u i d nitrogen. The argon had two functions: to sweep out the a i r and humidity i n the four units and to pump the l i q u i d s from one unit to another. For th i s l a s t purpose a vacuum was also used. The whole drying operation was executed i n two hours. Once the system was free of moisture and a i r d i s s o l u t i o n of the predried iodine c r y s t a l s was c a r r i e d out. The next step in t h i s technique was to pump the iodine solution either by vacuum or by argon to the f i r s t f i l t e r i n g unit. Then the solutio n was pumped to the reaction chamber. When the s t e e l sample was completely dissolved a f t e r 10-24 hours, the liquor containing the extracted inclusions was sent to the l a s t f i l t e r i n g unit. At t h i s l a s t unit, r i n s i n g of the f i l t e r e d inclusions with 99.99% methyl alcohol was performed several times u n t i l the f i l t e r e d alcohol was completely c o l o r l e s s , i . e . iodine free. This l a s t operation was performed under an argon atmosphere. Extracted inclusions were dried and weighed before and after s u l f i d e inclusions were eliminated. Samples from sev-126 e r a l ESR ingots were subjected under t h i s extraction tech-nique for various purposes, namely 1) to corroborate the EPMA inc l u s i o n chemistry by quantitative (spectrophotometry) and q u a l i t a t i v e crystallographic (X-rays) analysis 2) to v e r i f y the v a l i d i t y of the t o t a l oxygen analysis and 3) to determine how inclusions were present i n the f i n a l product. 4.5 Crystallographic X-ray Analysis of Extracted Inclusions The extracted inclusions were c r y s t a l l o g r a p h i c a l l y an-alysed by a P h i l i p s high angle diffractometer and a Debye-Scherrer camera 114.83 mm i n diameter. The machine setting was 40 kV and 15 uA. The scanning rate used i n the d i f f r a c t o -meter was 1° 26/min. The camera i s designed such that 2 mm measured on the f i l m corresponds to 1° 6. The distance along the f i l m between the zero point and the reference end i s 180mm. In both X-ray techniques the iron ka^ r a d i a t i o n (X = 1.9 36) was used. Although the amount of extracted inclusions ranged from 10 to 30 milligrams, p r i o r to t h e i r X-ray analysis, micro-photographs and q u a l i t a t i v e SEM analysis were performed on a portion of the sample. The crystallographic X-ray analysis of inclusions c l e a r l y showed a sequence of A^O^ up to CaO*Al 20 3 and 12 CaO*7Al2C>3 (as given by the CaO-Al 20 3 pseudo binary phase 127 diagram) as the Ca-Si deoxidation l e v e l was increased. It i s worthwhile to mention that i n the calcium aluminates en-riched i n calcium (CaO•2Al 20 3~CaO•A1 20 3 and CaO•Al20^-12CaO• 7Al 20 3) very wide d i f f r a c t i o n peaks and bands were observed. This was a clear i n d i c a t i o n of t h e i r lower degree of cry-s t a l l i n i t y . The d i f f r a c t i o n bands corresponding to the 12CaO«7Al 20 3 were very weakly traced. Generally, the c r y s t a l patterns showed from 6 to 16 p r i n c i p a l d i f f r a c t i o n bands with various i n t e n s i t i e s . Mix-tures of at least two i n c l u s i o n c r y s t a l structures and some n i t r i d e s and carbides were observed. 4.6 Atomic Absorption Analysis (Spectrophotometry) A Perkin Elmer 306 spectrophotometer with a controller model HGA-220 was used to analyze slags and extracted i n -clusions from metal matrices. The lithium metaborate fusion (232 233) procedure available in the l i t e r a t u r e ' was followed. The concentration of the elements of i n t e r e s t were determined by using appropriately matched standards and blanks with the routine procedure given i n the general information of the Perkin-Elmer manual. Generally speaking the accuracy of the solution analysis was Ca ± 0.2%, A l ± 0.06%, S i ± 0.04%, . F + 0.2%, Mg ± 0.002% and Fe ± 0.001 wt %. Stoichiometric balances were ca r r i e d out on the bases that Ca, A l , S i , Fe, Mg, and F were present as CaO, A l 2 0 3 , S i 0 2 , FeO, MgO and CaF 2 res-pectively. Under t h i s assumption the stoichiometric calcu-l a t i o n s gave 10 0% with a minimum accuracy of + 1.0%. 128 4.7 Metallographic Analysis Ingots were sectioned l o n g i t u d i n a l l y i n such a manner that a s l i c e 2.5 cm i n thickness across the diameter was ob-tained. These plates were surface ground and etched with a 50% HCl-H^O solution at 70-80°C for approximately one hour. Thus l i q u i d pool marks made with m e t a l l i c tungsten powder to id e n t i f y deoxidation l e v e l s and ingot structure were re-vealed. Longitudinal discrete sampling (matching the pro-gressive sampling of l i q u i d pools and slags) to perform spectrographic, i n c l u s i o n chemical and t o t a l oxygen analysis were thus obtained. Samples obtained r a d i a l l y were used to determine i n -clu s i o n s i z e d i s t r i b u t i o n s and dendrite arm spacings. The o p t i c a l analysis was performed using a Zeiss ultraphot i n d i f f e r e n t i a l interference mode. Samples from electrode t i p s , l i q u i d pools and r a d i a l and longitudinal specimens from ingots were ground on emery paper and diamond polished. Polishing was ca r r i e d out to determine incl u s i o n sizes and t h e i r l o c ation with respect to dendrites i n samples ex-tracted from l i q u i d pool and ingots using: diamond compound pastes (Metadi II) down to 0.25 ym diamond p a r t i c l e s i z e , tex-met, microcloth and nylon polishing (Buehler) cloth ( o i l and water r e s i s t a n t and alcohol and an o i l based lubricant, (Geomet thinner Micrometallurgical MM218). By following the above sequence removal of inclusions from metal matrices i n any siz e range was completely avoided. To determine the loc a t i o n of inclusions and dendrite arm spacings standard Oberhoffer's reagent was used to l i g h t l y etch these specimens. Four and f i v e faces of each specimen were polished, etched and microphotographs were taken at a magnification such that representative number of inclusions were i n each photograph. The magnification used was very dependent on the l e v e l of deoxidation and the chemistry of the deoxidant. 130 CHAPTER V RESULTS AND DISCUSSION 5.1 Mechanism by Which Electrode Inclusions are Eliminated In the l i t e r a t u r e review on inclusions p a r t i c u l a r emphasis has been given to the p r e c i p i t a t i o n sequence. It has been established that inclusions according to the de-oxidation technique exhibit sequential changes i n t h e i r chemical composition as s o l i d i f i c a t i o n takes place. These tr a n s i t i o n s may take place at subsolidus temperatures; p a r t i -c u l a r l y where s u l f i d e phases p r e c i p i t a t e . Thus, to f u l l y understand the mechanism by which i n -clusions are removed, electrode t i p s from 1020 M.S., 4340 and rotor (Ni-Cr-Mo) steels were studied. 1020 M.S. ingots were o r i g i n a l l y produced by acid e l e c t r i c furnace p r a c t i c e . The 4340 and the rotor s t e e l were calcium aluminum treated i n the l a d l e . Subsequently, they were hot r o l l e d into electrodes 76.2, 88.9 and 114.3 mm i n diameters res-pe c t i v e l y . 5.1.1 Behavior of Oxisulfide Inclusions i n 1020 M.S.  Electrode Tips The most complete picture of the i n c l u s i o n d i s s o l u t i o n and the transformation of the metal matrix i s given by the 1020 M.S. electrodes. Typical inclusions i n electrodes i n t h i s series of experiments, as received, are shown i n Figures (16) and (17). Their q u a l i t a t i v e (SEM) chemical 131 composition are also presented (spectrum X-ray a n a l y s i s ) . Two well-defined phases are i d e n t i f i e d . The darker phase i s enriched i n Si and the l i g h t phase i s enriched i n mang-anese s u l f i d e . Iron was also found i n both phases. Metallographic studies of electrode t i p s which were sub-jected to c r i t i c a l (ESR) thermal gradients have c l e a r l y re-vealed the existence of several heat affected areas, Figures (18) to (21). Findings from t h i s research i n q u a l i t a t i v e (8) (13 agreement with previous t h e o r e t i c a l and experimental work ' 19 20 22) ' ' . show that au s t e n i t i c grain growth and changes i n the morphology of inclusions occurs between 0.5 - 0.8 cm above the l i q u i d f i l m . Sulfide inclusions are f i r s t spherodized and subsequently dissolved i n t h i s region as well. The chem-i c a l composition of inclusions determine the c r i t i c a l length at which the above changes take place. The d r i v i n g forces to produce these changes i n order of importance are: the heat produced by the high r e s i s t i v i t y of the slag and hence the au s t e n i t i c grain growth of the e l e c -trode t i p . The deformation of inclusions and metal matrix which produce sharp concentration gradients and the non-equilibrium nature of the incl u s i o n p r e c i p i t a t i o n — a s indicated i n Figure (5) by the single dotted l i n e — c a n also provide a d r i v i n g force for t h i s sequence ' 6 4 ~ 1 6 6 \ The Mn depletion around inclusions i n the metal matrix should also be considered as another d r i v i n g f o r c e ' 6 4 166,199)^ T h e c n e m i c a - ] _ nature of inclusions according to t h e i r appearance and thermal h i s -tory (location) can be c l a s s i f i e d as: a) Deformed double phase o x i - s u l f i d e s and deformed s u l f i d e s , Figures (16,17). These inclusions were commonly found i n areas where i n c i p i e n t grain growth was observed, b) Comparatively large globular s u l f i d e s and o x i s u l f i d e s which were located i n r e l a t i v e l y grown au s t e n i t i c grains, Figure (22). c) Spherical single oxide phas dark i n appearance, Figure (23) and d) Relatively small re-p r e c i p i t a t e d complex (Ca, A l and Si) oxides. These two kinds of inclusions were located i n p a r t i a l l y and completely l i -quid areas. The l a s t type was p r e f e r e n t i a l l y located i n the l i q u i d f i l m and i n droplets, Figures (24,25). As shown i n Figures (18) to (20), areas at which physical and chemical changes take place either i n inclusions or i n the metal ma-t r i x are very well defined. The presence of the d r i v i n g forces previously described compensate the r e l a t i v e l y short periods of time at which the volume of electrode i s therm-a l l y affected. This promotes quasi-equilibrium thermo-chemical conditions approaching the predicted changes given i n Figures (4) and (5). Based on the experimental q u a l i t a t -ive (SEM) and quantitative (EPMA) information about how the electrode t i p and inclusions are p h y s i c a l l y and chemically tran formed as a p r i n c i p a l consequence of the thermal gradients and the chemical composition of i n c l u s i o n s , a mechanism which d e s c r i b e s the removal of o x y s u l f i d e i n c l u s i o n s i s proposed. Once a g i v e n r e g i o n i n the e l e c t r o d e t i p i s a f f e c t e d by the heat coming from the l i q u i d s l a g the i n c l u s i o n s and metal m a t r i x s t a r t to experience c e r t a i n changes. The spher-o i d i z a t i o n of s u l f i d e s and o x i s u l f i d e s takes p l a c e almost s i m u l t a n e o u s l y to the growth of the a u s t e n i t i c g r a i n s of the metal m a t r i x . E x p e r i m e n t a l l y estimated temperatures along a x i a l p o s i t i o n s i n the e l e c t r o d e and the t h e o r e t i c a l c a l c u l a t i o n s (8) i n d i c a t e t h a t t h i s t r a n s i t i o n (a + p e a r l i t e -»• y) s t a r t s to occur a t about 0.5 to 0.8 cm above the l i q u i d f i l m where the e l e c t r o d e i s a t a temperature of about 8 5 0 - 9 5 0 ° C , F i g u r e s (2, 21). T h i s f i n d i n g approximately corresponds to the alpha i r o n p l u s p e a r l i t e to gamma i r o n t r a n s f o r m a t i o n . In the zone where r e l a t i v e l y l a r g e g r a i n s were observed (about 1000°C to 1350°C) s u l f i d e s e n r i c h e d i n manganese are almost t o t a l l y d i s -s o l v e d , a f a c t which i s i n agreement w i t h Turkdogan's and co-workers' f i n d i n g s ^ ^ 4 166)^ F i g u r e s (4) and (5). These r e -s u l t s should a l s o be dependent on the Mn:S r a t i o as i n d i c a t e d i n the literature'°"*' 1"'" 3* . O x i s u l f i d e i n c l u s i o n s c o n t a i n i n g s i l i c o n , a c c o r d i n g t o Van V l a c k e t a l . ' 3 0 * , S i l v e r m a n ' ^ 9 * * r, (129,163,168) ,. ., , . . and H i l t y and C r a f t s • are d i v i d e d i n two c a t e g o r i e s , namely those i n which the S i : 0 r a t i o i n wt. % i s e i t h e r s m a l l e r or g r e a t e r than u n i t y . Low S i phases, S i : 0 r a t i o s l e s s than u n i t y q u a l i t a t i v e l y obey Turkdogan e t a l . ' s ' * * 4 * e q u i l i b r i u m s t a b i l i t y phase diagram. In the temperature range of 900° to 1225°C the equilibrium phases in the Fe-Mn-S-0 system are ruled by the univariant h i n Figure (5). This univariant i s constituted by the gamma iron, "0"as [Fe(Mn)0], "MnS" and %^ as l i q u i d o x i s u l f i d e . Under normal ESR-operating conditions, however, i t i s not expected that the composition of inclusions s t r i c t -l y follows t h i s univariant (h). Instead the behavior given by the "triple-dashed"-lines i n t h i s diagram i s expected, Figure (5). Simultaneous with the above changes solute ac-cumulated i n grain boundaries, growth of some inclusions and sulf u r depletion from s u l f i d e inclusions are observed. These events f u l l y coincide with the formation of l i q u i d o x i s u l f i d e , i^, i n Figures (4,5). It i s important to note that the s p e c i f i c temperature at which %^ forms, i s s t r i c t l y a function of the o r i g i n a l amount of Mn present (Mn:S ratio) and the degree which the iron i s saturated with Mn(Fe)0 and M n ( F e ) S ' 6 4 ~ 1 6 6 ^ . I t also has to be emphasized that with commercial steels i n these series which have standard S content, (0.02-0.05 wt % ) , there i s s u f f i c i e n t Mn and S i present that no l i q u i d o x i s u l f i d e (^) i s present at temp-eratures lower than 1150°C. Above th i s temperature the l i q -(131 132) uid o x i s u l f i d e i s expected *~' to show up as material penetrating the "original" a u s t e n i t i c grains i a fa c t which was indeed observed. In areas closer to the f u l l y transformed (austenitic) grains the Mn content i n inclusions, due mainly to the d i s s o l -ution of sulfur i n the metal matrix i s increased, Figure (4). Although the Mn content was not s i g n i f i c a n t l y greater the inc l u s i o n composition q u a l i t a t i v e l y obeyed the univariant h. From approximately 1150° to 1250°C the composition of low Si o x i s u l f i d e inclusions changes and £^ i s expected to flux the s o l i d s u l f i d e as Silverman'*' 9* and Van Vlack et a l . ' 3 0 * have indicated. Mn w i l l continue increasing at slower rates than i n previous transformations as the temperature i s increased. Upon quenching a sample taken from areas closer to the f u l l y ytransformed region, duplex ( r e l a t i v e l y grown) o x i -s u l f i d e s low i n S i should be precipitated; a f a c t which i s c l e a r l y 'seen i n Figure (22 a-b). Their major constituents were i d e n t i f i e d as a Mn r i c h s u l f i d e , Mn(Fe)S, and a Mn r i c h oxide, Mn(Fe)0. At temperatures higher than 1250°C the manganese s u l f i d e enriched phase p r a c t i c a l l y disappears; a fact which was observed i n the q u a l i t a t i v e (SEM X-ray spectrum analysis) and semiquantitative (EPMA) analysis, F i g -ures (23 and 25). At temperatures about 1370° to 1420°C r e l a t -i v e l y large a u s t e n i t i c grains are observed. The region which i s exposed to thi s temperature range experiences the gamma to delta iron transformation. This t r a n s i t i o n i s i d e n t i f i e d i n Figure (5) as the in t e r s e c t i o n of the "triple-dashed" l i n e crossing the univariant g_ i n which delta and gamma iron, 0 136 and £^ are i n equilibrium. The major changes i n inc l u s i o n composition w i l l s t a r t just a f t e r univariant f i s reached, f i s a univariant equilibrium which i s constituted by delta iron, o, ^ and &2, where &2 i s the l i q u i d metal. After f_ i s reached the only s o l i d compounds may be the Fe(Mn)0 and a very small amount of ir o n oxide. It i s important to r e c a l l that t h i s event takes place only i f Si content i s low • 4-u i x-o- (130,169) i n these l a t t e r stages ' Under t h i s condition, between f_ and e univariants where delta i r o n , o and £,2 are i n equilibrium the remaining S from inclusions goes i n solution i n delta iron. Furthermore, the remaining inclusions w i l l be exclusively constituted by Mn(Fe)0. This l a s t step was p a r t i c u l a r l y clear i n the range at which the l i q u i d f r a c t i o n was about 0.5. In areas where the l i q u i d f r a c t i o n was greater than 0.5 inclusions were completely dissolved and reprecipitated. Since t h i s region was i n contact with the slag some reprecipitated small i n -clusions (very few) Figures (24, 25), show the slag char-acter, i . e . , some A l and Ca. The second category of o x i s u l f i d e inclusions which Si:0 r a t i o i s larger than unity exhibit an almost equivalent pat-tern as the previous behavior ( i . e . Si:0 < 1), except i n the l a s t stages. Instead of p r e c i p i t a t i n g Mn(Fe)0 and FeO, S i 0 2 and MnO or 2Mn0 w i l l p r e c i p i t a t e i n the f u l l y austenitized zone and when l i q u i d fractions are smaller than 0.5 These transformations as proposed by Silverman'** 9* and Van Vlack et a l . ^ 1 3 ^ * can be represented by quaternary diagrams, Figures (6) and (7). As shown i n Figures (6) and (7) as the temperature increases Mn ( and thus 'MnO') and the proportion of o r t h o s i l i c a t e increases. As a r e s u l t the liquidus for the "MnS" i s decreased, thus over a s i g -n i f i c a n t range of compositions the amount of l i q u i d o x i -s u l f i d e i s increased. Silverman'** 9* has pointed out that i f the proportion of s i l i c a t e i s increased, as observed i n th i s case then at temperatures above 1300°C what i s l e f t as a s o l i d (from the previous Fe-O-S-Mn-system, FeO and Mn(Fe)O) w i l l become almost completely l i q u i d . These transformations as proposed by Van Vlack et a l . ' 3 0 * occur from C to B i n Figure (8a) and C to B i n Figure (8b). During t h i s temperature increase "FeS", "MnS" and "FeO" are d i s -solved, Figure (25). Subsequent chemical reactions of the i n c l u -sion components which take place i n the f u l l y austenitized met-a l and the p a r t i a l l y l i q u i d zones are shown i n macrophotographs (18a-b) and (21) and photographs (23a) and 23b). These reactions are represented by: 2MnO + S i 0 2 Z 2MnO«Si0 2 «- Mn 2Si0 4 or 2MnO + S i 0 2 X MnO«Si0 2 + MnO t Mn 0 + MnSi0 3 138 If a sample with t h i s composition and thermal history i s r a p i d l y cooled then . single phase inclusions enriched i n manganese and s i l i c o n must be found, a fact which i s l a t e r corroborated with the inclusion chemical analysis. Their f i n a l composition, according to Van Vlack et al . ' s work ' 3 °V i s dictated by the Si:0 r a t i o and the amount of manganese present i n the electrode. Thus, i f the S i content exceeds the amount of oxygen, as seen i n analysis from samples of _3 thi s series (0.25 wt% Si and 90 ppm = 9.0 x 10 wt!) then a v i t r e o u s ( s i l i c e o u s ) type of in c l u s i o n i s formed i n place of the monophasic tephroite (MnO'SiC^) or rhodonite (2MnO'SiO^) as the temperature i s increased, Figures (23a,b). Some electrodes which were very rapidly withdrawn from the slag showed small s i l i c e o u s and (iron) s u l f i d e s as re-pre c i p i t a t e d inclusions exclusively in the l i q u i d f i l m . Their q u a l i t a t i v e (SEM) and semiquantitative (EPMA) analysis, (15) as reported, i n the l i t e r a t u r e i d e n t i f i e d them as non-stoichiometric f a y a l i t e ^FeOSiC^) compounds. To confirm the previously described mechanism a seq-uence of t y p i c a l analysis i s shown. The electrode chemical (spectrographic) analysis i n wt. % as received, i s as follows: C Mn S Si P 0.19 0.71 0.026 0.025 0.01 The average t o t a l oxygen analysis as received, was 90 ppm., i . e . at centreline 87 ppm., midradius 91 ppm. and edges 93 ppm. The average i n c l u s i o n chemistry i n at.% as re-ceived, was S S i Mn Fe 35.0 27.0 37.0 balance The average chemical composition of inclusions located bet-ween the r e c r y s t a l l i z e d area and l i q u i d f r a c t i o n s smaller than 0.5,were (in at.%) as follows Si Mn S Fe 51.0 46.0 2.0 balance This composition immediately suggest the formation of teph-r o i t e ( 1 3 0 > . Due to the number and size of inclusions found i n the neighbourhood of the l i q u i d f i l m chemical analysis from inclusions located i n l i q u i d f r a c t i o n s greater than 0.6 were performed only on a q u a l i t a t i v e basis. Calcium and aluminum i n these inclusions as shown i n Figures (24) and (25) were traced. This p a r t i c u l a r electrode was remelted through a (50 wt% CaF 2, 30 wt% Al 2C> 3 and 20 wt% CaO) slag which was Ca-Si deoxidized. The chemical analysis of inclusions from l i q u i d pool (a) and from the (ESR) ingot (b) i n at.%. 140 are as follows: A l Ca S Mn Fe a) 49.0 30.0 20.43 0.3 balance b) 46.0 30. 5 22.80 balance Thus, i n summary the change in inc l u s i o n chemical com-po s i t i o n i n terms of sulfides and low Si-oxysulfides and the morphological changes of inclusions take place i n austenitic temperature ranges i n the metal matrix. In intermediate stages the d i s s o l u t i o n of s u l f i d e s i n the metal matrix also occurs. These series of events represent approximately 10-30% of the t o t a l "transformation-dissolution" mechanism and i t takes place at about 0.4 - 0.8 cm above the l i q u i d f i l m . The next 20-40% of the transformation of inc l u s i o n phases occur either between the g_-f univariants, Figure (5), i n the low S i con-tent phases or i n the C-B sequence of transformations, Figure (8a), i n the high S i content phases. The next sequence of transformations takes place between f-e and B-A for low and high S i content phases respectively. I t occurs between the f u l l y austenitized region and the point where the l i q u i d f r a c -t i o n does not exceed 0.5. It represents another 20-40%. The remaining "transformation-dissolution" takes place i n the neighbourhood of the l i q u i d f i l m where inclusion d i s s o l u t i o n i s the major mechanism and in c l u s i o n slag reactions (form-ation of lower melting point i n c l u s i o n phases with slag char-acter) acts as a very limited mechanism (^1.0-3.0%). 141 5.1.2 Removal of Oxide and Sulfide Inclusions i n 4340 and  Rotor Steels Since the aim of th i s research i s to gain an extensive understanding of the inclusion-removal mechanism; electrodes with d i f f e r e n t matrix-inclusion compositions were also anal-yzed. 4 340-electrodes with alumina type of inclusions and tool s t e e l electrodes with calcium-aluminum-silicates were also included i n t h i s research program. Macrophotographs (19) and (20) show the sequence of trans-formations i n 4340 and rotor steels respectively. One of the f i r s t differences to notice between these two types of steels i s that i n the l a t t e r type a u s t e n i t i z a t i o n did not occur due to t h e i r chemical composition (electrode) and to the chemical composition of inclusions. Another conse-quence of t h i s i s that solute penetration (between austeni-t i c grains) at subsolidus temperatures i n the rotor s t e e l did not take place. On the other hand i n the p a r t i a l l y molten electrode much more"segregated material," as a re-su l t of the i n c l u s i o n d i s s o l u t i o n and the in t e r a c t i o n of the electrode with the slag, was observed at l i q u i d f r a c t i o n s larger than 0.5. Once the d i f f e r e n t areas were i d e n t i f i e d ( i n c i p i e n t l y heat affected area, semi or austenitized region, l i q u i d f i l m i n or out of the droplet) i n each type of electrode, meticulous chemical analysis of inclusions (by EPMA) was c a r r i e d out at every 250-300 ym. Results obtained from 4340 and rotor s t e e l electrodes are shown respectively i n Figures (26) to (28). Among the most important conclusions from t h i s study are the following: a) since the au s t e n i t i c transformation i n the rotor s t e e l was almost absent the inclusion transform-at i o n - d i s s o l u t i o n started to take place p r i n c i p a l l y i n the p a r t i a l l y l i q u i d zone. It takes place about 0.3-0.5 cm above the solidus. b) A l and S i n inclusions i n both types of electrodes (4340 and rotor steel) were gradually dissolved as the electrode experienced higher thermal gradients. These changes, shown i n Figures (26) to (28) s t a r t to take place i n a very discrete manner just above the solidus isotherm for the rotor s t e e l and i n e a r l i e r (lower temperature gradi-ents) subsolidus temperatures i n the 4340 ingots. c) In graph (27), i t can be observed that due to a strong grain growth and thus intergranular segregation, the A l , Mn and S compositions were s h i f t e d i n 2500-4500 ym range. These re-u . . ... . (1,17,18,129,131-133, suits i n agreement with previous research ' ' ' ' ' 163 169) a n ( j w ^ t n p r e v i o u s findings i n the 1020-series show that s u l f i d e s are the most thermally affected phases. Oxide inclusions i n deformed 4340 electrodes also experi-enced gradual morphological changes. d) The largest changes i n the i n c l u s i o n chemical composition i n both electrodes i n a manner equivalent to the 1020 M.S. electrodes oc-curred i n the p a r t i a l l y l i q u i d region. The presence of strong intergranular and i n t e r d e n d r i t i c (segregated) material which contained inclusion-formers c l e a r l y indicate that i n -clusions were completely l i q u i d i n this transient zone. Figure (19a) from 4340-electrodes and (20) from the rotor s t e e l electrodes strongly corroborate these findings, e) Chemical analysis i n 4340 electrodes, Figures (26,27), suggest that Si and Mn i n inclusions, p a r t i c u l a r l y i n the l i q u i d f i l m follow the same behavior as i n 1020 M.S., i . e . inclusions get richer i n S i and Mn j u s t before they are t o t a l l y dissolved. 144 5.1.2.1 Removal of Oxides and Sulfides i n 4340-electrodes The sequence of changes i n the chemical composition of inclusions i n these electrodes i s summarized as follows. The spheroidization and a subsequent d i s s o l u t i o n of s u l -fides just as i n previous observations i n the 1 0 2 0 M.S. electrodes i s the f i r s t step i n the removal mechanism. It takes place i n subsolidus temperature ranges. Figures (19), (26) and (27) indicate that these changes occur at about 2000-3000 ym above where the solidus of the a l l o y was metallographi-c a l l y i d e n t i f i e d . The most d r a s t i c changes, however, take place in areas where complete a u s t e n i t i z a t i o n was observed. In f u l l y austenitized areas and i n the region where the melting started the A l and S i n inclusions de-crease while the S i , Mn and the Ca correspondingly increase. This behavior i s accentuated as the l i q u i d f r a c t i o n i n -(9192) creases. A mixture of feldspars ' (MO "A^O^ «2Si02) and garnets (3M0«Al20.j *3Si02) ° r e v e n c o r d i e r i t e s (2M0* 2 A I 2 O . J • 5Si02) instead of the o r i g i n a l aluminates should p r e c i p i t a t e as the l i q u i d f r a c t i o n approaches unity. The actual stoichiometry of t h i s compound i s i r r e l e v a n t since M i n a l l of these compounds cart be either Fe, Mn or Ca and these series of compounds show v i r t u a l l y complete mutual (91 92) s o l u b i l i t y ' . The most important finding, however, i s that oxides i n t h i s matrix are suddenly transformed not so much in subsolidus temperature ranges as occurs i n 145 the 1020 M.S. as i n the semi-liquid stage. The presence of i n c l u s i o n r e l i c s and " l i q u i d " enriched i n i n c l u s i o n constituents i s shown i n Figures (19c) and (2 0a) as i n t e r -d e n d r i t i c segregates. In the l i q u i d film, about 50ym from the edge of the electrode t i p a similar e f f e c t as that seen i n the 1020 M.S. electrodes was observed. Since t h i s volume of l i q u i d electrode was subjected to a d i r e c t i n t e r a c t i o n with the slag a s h i f t i n the chemical composition of inclusions was detected. This indicates that the electrode i n c l u s i o n composition was completely transformed. The chemical analysis of inclusions i n the l i q u i d f i l m (a) i n the droplet or i t are (in at.%) as follows: A l Ca S Si Mn a) 19.0 2.78 _ 78.0 0.22 b) 7.0 3.90 1.5 70 .20 17.40 a) 8.27 9.30 _ 56.0 26.46 b) 6.30 11.00 - 56.0 27.60 4340 (I) 4340 (II) F i n a l l y , the chemical composition of inclusions i n either (a') l i q u i d (pool) stage or .(b')the ingot correspond to: 434 a') Al„0_ and traces of s i l i c a + MnS II 0 (I) 2 3 b') A1 20 3 + MnS I I 4340 (II) a') Ca«Al 20 3 b') 12CaO-7Al 20 3 146 It i s important to note that 4340 (I) was Al-deoxidized and 4340 (II) was CaSi treated. Ingot 4340 (I) was s l i g h t l y deoxidized 0.02 kg ton-"'') and ingot 4340 (II) was heavily CaSi deoxidized (10 Kg t o n - 1 ) both ingots were refined through a 50wt % CaF 2, 30 wt% Al 2C> 3 and 20 wt% CaO. If the above chemical analyses are compared i t can c l e a r l y be seen that inc l u s i o n r e l i c s from the bulk of the electrode, i f any, are dissolved and reprecipitated i n c l u s -ions which show the slag character. Inclusions i n the l i q u i d f i l m are small and complex i n composition. Therefore, i n -clusions located in t h i s narrow f i l m do not represent i n any way what happened i n the transformation-dissolution of previous stages. 5.1.2.2 Calcium-Aluminum S i l i c a t e s i n a Rotor (Ni,Cr,Mo) Steel The i n c l u s i o n chemical composition was meticulously determined at discrete locations i n electrode t i p s . These analysis were carr i e d out at every 250-300 ym st a r t i n g from the l i q u i d f i l m . For the sake of c l a r i t y as i n previous studies (1020 and 4340), they were arranged according to the s p e c i f i c area at which they belonged i . e . , l i q u i d f i l m (in and out of droplet), p a r t i a l l y molten area at several l i q u i d f r a c t i o n s and at subsolidus temperatures. 147 A summary of the results obtained by EPMA and t h e i r c l a s s i f i c a t i o n according to the thermal history at which these volumes were subjected i s shown i n Figure (28) * Since these electrodes experienced almost no grain growth during a u s t e n i t i z a t i o n , the composition of inclusions was exclu-s i v e l y changed i n regions close to the solidus temperature of the a l l o y . It i s important to notice that the size of the mushy zone i n the rotor s t e e l (> 2800 pm) i s larger than i n 1020 M.S. or the 4340 s t e e l s . The sl a g - i n c l u s i o n d i s s o l u t i o n (in only very few remaining inclusions) took place i n a manner equivalent to a l l of the other electrodes, i . e . l i q u i d f i l m . T y pical analysis of inclusions from these areas i n at. %, are as follows: P a r t i a l l y L iquid (f^ ^ 1) A l Ca S i S Mn ingot (1) 25.18 61.15 8.65 5.0 ingot (2) 25.88 61.36 9.40 3.5 Liquid Film out of droplet 38.48 20.20 34. 85 6.50 i n droplet 61.16 21.63 16.38 0. 82 After ESR 62.35 2.27 3.16 0.16 1.16 148 These ingots were refined through the 50% CaF 2, 30% A^O^ and 20% CaO slag. Their deoxidation was ca r r i e d out with the CaSi a l l o y at approximately 0.1-0.2 Kg ton 1 . 5.1.3 F i n a l Remarks About the Removal Mechanism Inclusions i n electrode t i p s during ESR are removed by di s s o l u t i o n i n the metal matrix in well defined steps ac-cording to the electrode i n c l u s i o n composition and location i n the electrode t i p . The thermal gradient to which the electrode and hence inclusions are subjected i s the main dri v i n g force for the i r removal. This mechanism, however, does not correspond to that previously described i n the l i t e r a t u r e review. Gradual d i s s o l u t i o n along the heat affected zones or s t r i c t d i s s o l u t i o n of inclusions at the l i q u i d film, and the "washing o f f " or the mechanical re-moval of inclusions are not operative mechanisms. In-stead a mechanism based on substantial experimental and the o r e t i c a l evidence which suggests the quasi-thermo-chemical equilibrium of inclusions and the i r location i s indicated. The series of actual chemical transformations and the di s s o l u t i o n of inclusions in electrode t i p s are s t r i c t l y confined to a distance no greater than 0.5 - 0.8 cm above the l i q u i d f i l m . Sulfides i n the Fe-Mn-S-0 system and o x i -s u l f i d e s i n the Fe-O-S-Mn-Si system are p a r t i a l l y dissolved and p a r t i a l l y transformed at subsolidus temperatures.. 149 Reactions between i n c l u s i o n components and the metal matrix leads to the solution of ce r t a i n inclusion components such as "FeS", "FeO", "MnS", etc., which i n i t i a t e solute penetra-t i o n i n au s t e n i t i c grain boundaries. This f a c t implies that these inclusions are a mixture of phases, namely "MnS", "FeO" and a small amount of "FeS". The chemical reactions to which inclusions are subjected and the i r d i s s o l u t i o n i n the metal matrix are q u a l i t a t i v e l y predicted by using diagrams available (164 —16 6) i n the l i t e r a t u r e . Sulfides belonging to the Fe-Mn-S-0 system and o x i s u l f i d e s low i n s i l i c o n which belong to the general Fe-O-S-Si-Mn system i n electrode t i p s follow a quasi-thermochemical equilibrium dictated by the Figure (5). This equilibrium phase diagram obtained by Turkdogan and co-workers i s n o t completely obeyed and instead a be-havior given by the " t r i p l e dashed l i n e " i s followed. Inclusions belonging to the Fe-Si-Mn-O-S system which have a s i l i c o n content such that the Si:0 r a t i o i s greater than 0.5, follow the FeO-MnS-MnO-Si02 quaternary system dev-eloped by Silverman'** 9^ as indicated by the arrow i n F i g -ure (9). The behavior of these inclusions i s also complemented by the binary MnO-Si02 as part of the Si02~MnO-FeS-MnS quat-ernary diagram developed by Van Vlack et a l . , Figures (8a) and (8b). This proposal corresponds largely to that a l -(168) ready suggested by H i l t y and Crafts , i . e . the pseudo 150 ternary behavior of the metal-oxide-sulfide phases. Thermo-dynamically more stable phases such as Mn(Fe)0, MnO, Si02^ MnO«Si02A 2Mn0«Si02/ calcium aluminates, calcium s i l i c o n a l -uminates, etc., are transformed and dissolved i n the neigh-bourhood of the l i q u i d f i l m between the f u l l y austenitized area and the f u l l y l i q u i d metal. Electrode i n c l u s i o n r e l i c s ( i f any) i n the l i q u i d f i l m react, up to a limited extent with the slag producing complex in c l u s i o n phases. F i n a l l y , inclusions i n ESR ingots which show the e l e c t -rode i n c l u s i o n chemical composition are found only under un-stable ESR conditions, i . e . at the s t a r t i n g and during the "hot topping" stages. 151 5.2 The Chemical Influence of the Electrode, Slag and Deoxidizer on the Chemical Composition of  Inclusions 5.2.1 Description of Experimental Findings 5.2.1.1 Preliminary Studies on the E f f e c t of the Slag and  the Deoxidation 4340-electrodes 31.75 and 44.75 mm i n diameter and with d i f f e r e n t i n c l u s i o n chemical compositions were refined to 75 mm i n diameter ingots at melting rates of about 1.3 Kg m i n - 1 through d i f f e r e n t slag systems and under a protective (argon) atmosphere. The components of the slag systems (CaF 2, CaO, Al-^O^ and Si0 2) were previously dried at 650°C and the "cold s t a r t i n g procedure" was followed. Electrodes were refined through slags which had several S i 0 2 contents. Two d i f -ferent slag systems were chosen to be deoxidized with a CaSi a l l o y and A l , Table (X). The aluminum was i n the form of p e l l e t s (99.9%). The CaSi a l l o y contained 62.5 152 wt. % S i , Table (IX). The deoxidation rates were constant (^  2.3 Kg/ton) and they were performed when steady remelting conditions were observed. The purpose of these experiments were: i) to determine the chemical composition of inclusions as an exclusive ef-fe c t of the slag composition (and electrode). i i ) to i n -vestigate what changes i n i n c l u s i o n compositions could be achieved by using the same slag system and deoxidizers and to compare r e s u l t s from (i) against ( i i ) and ( i i i ) to det-ermine the most appropriate slag systems to be used i n the 200 mm i n diameter ESR-furnace. Experimental r e s u l t s are shown i n Table (X) where the slag chemical composition i n wt. % # the incl u s i o n chemical composition i n at. %, the chemical composition of inclusions i n electrodes and the major in c l u s i o n phases are given. It i s important to note that although chemical analysis of i n -clusions was performed on a large number (30-40) a wide scatter (± 5.0 - 7.0%) i n th e i r analysis was found. The scatter i s p r i n c i p a l l y attributed to the i n c l u s i o n size 153 d i s t r i b u t i o n s (less than 5ym i n diameter) and to the un-steady r e f i n i n g conditions due to the use of inappropriate slag systems. As a consequence of t h i s lack of s t a b i l i t y occasionally an uneven surface of the ingot was observed and l i q u i d enriched i n deoxidizers and oxygen pr e c i p i t a t e d alumina type of inclusions i n a confined volume, Figure (29). The f i r s t set of experiments performed without a de-oxidizer showed s i l i c o n i n inclusions where SiC^ i n slags was higher than 10 wt %. Calcium-aluminum s i l i c a t e s i n inclusions were found above t h i s l e v e l . The presence of more than two i n c l u s i o n phases was commonly observed i n the same sample. Table (X). The second series of experiments i n which deoxidation was c a r r i e d out, s p e c i f i c a l l y ingots (7) and (9), showed v i r t u a l l y the same behavior as ingot (1). The chemical composition of inclusions i n ingot (1), used as a reference showed almost exclusively calcium aluminates. Ingots (7) and (9) also showed calcium aluminates, the Ca:Al r a t i o , however i s larger than i n previous cases. The s i l i c o n con-tent of inclusions i n ingots (1) and (7) were equivalent, whereas i n ingot (9). i t was higher. If the analysis of inclusions from electrodes used i n ingots (1) to (7) are compared against (9) and (10) then 154 i t can be inferred that the increased S i content comes from the electrode, Table (X). The maximum calcium content, as calcium aluminates, was found i n ingots remelted through r e l a t i v e l y high CaO (15-22 wt %) and r e l a t i v e l y low S i 0 2 (less than 10 wt %) slags. The i n c l u s i o n chemical analysis performed i n ingots (2) and (10) shows that by r e f i n i n g electrodes through the (55/ 15/15/15) slag system, t h e i r composition with and without Ca-Si deoxidation remains unaltered. On the other hand, for electrodes refined under the same slag system and de-oxidized with aluminum; lower s i l i c o n and r e l a t i v e l y higher calcium and hence higher aluminum i n incl u s i o n phases i s found. Results i n t h i s i n v e s t i g a t i o n c l e a r l y show that the i n -t r i n s i c slag e f f e c t i n the chemical composition of inclusions follows a very well defined pattern. Slag systems i n which the S i 0 2 content i s lower than 10 wt % as i n ingots (1), (4), (5), (7), and (9) yielded i n c l u s i o n compositions which pre-dominantly l i e i n the CaO-Al 20 3 pseudo binary phase diagram on the A l 2 0 3 r i c h side, i . e . alumina types and low CaO-aluminates. To corroborate these findings other series of ex-periments i n the 200 mm ESR-furnace were performed. Ex-perimental d e t a i l s and a summary of findings are given i n 155 Tables (VIII) and (XI). The main point to be considered i n t h i s set of experiments i s the low l e v e l of deoxidation (0.02 kg ton 1 ) to which the s l a g - l i q u i d pool was sub-jected and the electrode surface preparation. As seen i n Table (XI) consistent q u a l i t a t i v e r e s u l t s are found i n terms of the chemical composition of inclusions i n both ESR-furnaces. P a r t i c u l a r emphasis should be given to r e s u l t s ob-tained from ingot (11). A rotor s t e e l electrode with chemical composition i s given i n Table (VII) and with an average i n c l u s i o n chemical composition (in at%) as follows: A l Ca S i S Mn Fe 28.5 24.1 42.0 2.2 2.5 balance was remelted through a 49 wt % CaF 2, 16 wt % CaO, 17 wt % A^O-j, 12 wt % S i 0 2 and 6 wt % MgO slag. The average chemical composition obtained from 40 inclusions i n the (ESR) ingot, i n at.% i s as follows: A l Ca S i S Mn Mg + Fe 44.70 13.60 24.5 7.4 9.2 balance Since the at.%Mn:at.%S r a t i o i s approximately one then i t can be assumed that a MnS phase was pr e c i p i t a t e d . The s u l f i d e phase was commonly found surrounding the oxide phase. 156 X-ray composition p r o f i l e s and maps as well as dispersive X-ray spectrum analysis confirmed these findings. The oxide phase was not spherical (as calcium aluminates-calcium s u l f i d e inclusions), instead angular oxides were ob-served. If the above values are normalized then the re-maining elements can be further analyzed; thus the o v e r a l l composition i s A l Ca S i ^54 ^16 ^30 From these computed values the r a t i o s for A l , Ca and S i res-pectively are approximately 4:1:2. Thus, the f e a s i b l e phase present i n these type of inclusions could be: CaO«2Al 20 3'2Si0 2 This compound, as i n d i r e c t l y stated i n the l i t e r a t u r e review* 9 2 ^ has not been reported as an i n c l u s i o n phase; i t can be instead a feldspar of the type M0«A1 20 3«2Si0 2 i n which MO can be either FeO, MgO, MnO or CaO. Based on i n d i v i d u a l analysis of inclusions; alumina r i c h phases and complex CaO-Al 20 3 s i l i c a t e s were indeed the major phases present. Hence, the only f e a s i b l e compounds i n t h i s ingot are the Al 20 3«CaO•2Si0 2 (anorthite) phase i n conjunction with an Al_0_ (corundum) enriched phase. 157 5.2.1.2 Intermittent CaSi Additions and the Reaction Scheme Since no difference was found between deoxidizers (Al and the Ca-Si alloy) i n small ESR-furnace, i n terms of the chemical composition of inclusions ( i . e . p r e c i p i t a t i o n of calcium aluminates and up to a given extent calcium sulphides), an extension of these experiments i n the semi-i n d u s t r i a l size ESR-furnace was ca r r i e d out. These experi-ments are l i s t e d i n Table (VIII). 50 grams of FeO and the Ca-Si all o y were alt e r n a t e l y added at two discrete time i n t e r v a l s during r e f i n i n g under two d i f f e r e n t slag systems; namely 50/30/20 and 70/30/0. These figures represent the CaF 2, A^O^, and the CaO compositions i n wt %. 1020 M.S. electrodes whose chemical composition and i n c l u s i o n chem-i s t r y have already been described, were refined at about 1 kg min 1 , with and without (RIII-W and RII-W respectively) an argon atmosphere shielding. Total oxygen content, slag chemical analysis from samples taken during r e f i n i n g , ingot chemical analysis and incl u s i o n chemical composition as well as size d i s t r i b u t i o n s from ingots were determined. These re s u l t s are shown i n Figures (30) to (36). Ca, F, A l , S i , Mn and Fe and Mn, C, P, S, S i , Mo, Cr, Al and N i were analyzed i n the slags and ingots respectively. Elements with s i g n i f i c a n t changes i n t h e i r composition are plotted. Figures (32) and (33a-b) from RII-W and RIII-W, show the 158 e f f e c t of the Ca-Si a l l o y (intermittent) additions. These graphs show that the deoxidizer produces a sharp decrement in a l l of the slag constituents. Correspondingly, the chem-i c a l composition of ingots show a sharp increment i n Mn, A l and S i , Figures (34,35). Other important points to be considered from these graphs are that s u l f u r i s decreased when the Ca a l l o y was added and iron oxide additions in the slag did not produce as sharp changes i n the ingot composi-tion as did additions of the deoxidizer. The s i l i c o n which comes from the electrode gradually increased i n the slag as r e f i n i n g takes place, Figure (32) and (33b). The response time of the system was also observed. RII-W showed a response to the deoxidizer i n about 100-150 seconds and RIII-W within 200-250 seconds. The e f f e c t of the deoxidizer i n the slag and ingot decrease i n a r e l a t i v e l y slow manner. The sudden changes i n FeO content (expressed as wt % of Fe) i n the slag as well as the abrupt changes i n the t o t a l oxygen content and the i n c l u s i o n chemical composition as a d i r e c t e f f e c t of these intermittent(Ca-Si) additions are the most important responses i n the refined ingot, Figures (30) to (33). I t i s important to notice that there was not enough reaction time to show the•individual Ca or FeO effects i n terms of oxygen i n (RIII-W), Figure (30). The difference between RII-W and RIII-W are the s t a r t i n g slag compositions and the presence of an argon shie l d i n g atmosphere i n the l a t t e r , Table (VIII). Thus a 159 lower oxygen content (15-20 ppm) i n RIII-W was expected and observed, Figures (30) and (31). The main conclusion from the above r e s u l t s i s that the ESR-process i n terms of slags and deoxidizers i s a very com-plex reactor. The series of reactions taking place i n the slag do not occur independently of the ones taking place i n the l i q u i d pool and i n the ingot, i . e . the i n c l u s i o n com-po s i t i o n i s not c o n t r o l l e d by one single factor. 5.2.1.3 Refining of 1020 M.S., 200 mm Diameter Ingots  Deoxidized Continuously with Aluminum To gain a better understanding of the above sequence of reactions, complementary and more detailed experiments by which reactions i n the l i q u i d slag l i q u i d pool and ingots could be monitored were planned. The next set of experiments included the r e f i n i n g of electrodes through equivalent slag compositions. The purpose of these experiments was to discriminate the i n t r i n s i c chemical e f f e c t of the electrode i t s e l f on the slag ( i . e . without any deoxidizer).and to i d e n t i f y separately the ef-f e c t of deoxidizers on the slag during the three stages of the r e f i n i n g process. The i n t r i n s i c electrode-slag-in-clusion chemistry of a refined (1020) ingot, without deoxid-ation and remelted through an i n i t i a l 50 wt% CaF^, 30 wt % A^O^ / and 20 wt % CaO, was used as the basis of comparison for other ingots, CaSi and Al-deoxidized. 160 The ingot i d e n t i f i e d as RI-Il was refined under an argon atmosphere and i n the absence of a deoxidant. The "FeO" i n the slag and the t o t a l oxygen content remained approximately constant i n the i n i t i a l stages, Figure (37). Inclusions i n samples extracted from the l i q u i d pool and from the ingot were e s s e n t i a l l y i d e n t i f i e d as spherical single p a r t i c l e s or as small clusters of alumina type (FeQ'A^O^ and a-A^O-j) and less frequently as f a y a l i t e type (2FeO«Si02) • The alumina type was usually associated with manganese s u l -fides, (Fe,Mn)S and MnS I I . The gradual increase of S i 0 2 i n the slag was considerable. I t ranged from 0.75 wt.% at the bottom and up to about 2.0 wt.% at the top of the ingot. The s i l i c o n content in the ingot ranged from 0.095 up to 0.125 wt.% from s t a r t to f i n i s h . The ingot chemical an-a l y s i s i s shown i n Figure (38). Experiments which are equivalent to the small (Ca-Si and Al) deoxidized (4340) ingots were also c a r r i e d out. 1020 M.S. electrodes were refined through equivalent slags and samples from l i q u i d slags and pools were extracted while continu-ously increasing additions of deoxidizers to the slag were made. Table (VIII) summarizes the d e t a i l s of t h i s set of experiments. Refined ingots referred to as R I I - I l and RII-I2 were aluminum treated by using two d i f f e r e n t deoxid-ation sequences. Deoxidation rates were 3.63, 6.1, 6.8 and 7.6 kg t o n - 1 and 1.21, 2.42, 3.64, 4.85, 6.06 and 12.12 kg ton - 1 161 for R I I - I l and RII-I2 respectively. RII-I2 was CaSi (50 grams) deoxidized i n the i n i t i a l r e f i n i n g stage. .. The iron oxide content, given as wt- % iron i n Figures (39) and (40), slowly and continuously decreased from about 0.6 wt.% down to about 0.4 wt-% as the aluminum addition rates were increased. While high deoxidation rates produced r e l a t i v e l y steady t o t a l oxygen content (RII-Il) and hence equilibrium behavior i n i n c l u s i o n compositions, the low de-oxidation rates produced an o s c i l l a t i n g behavior i n both parameters, (RII-I2), Figures (41, 42) and (43,44). Findings i n these experiments, although not as drama-t i c as the (Ca-Si) intermittently deoxidized ingots, show that the aluminum as a deoxidizer also produces simult-aneous exchange reactions between two l i q u i d s of the general type (11) and reactions of the type .(12). The elements i n -volved i n these reactions, i n a manner equivalent to the i n -termittent Ca-Si additions are the S i , Mn and the A l by i t s e l f , Figures (37-40) and (43-46). The most sensitive parts of the system to the aluminum deoxidation were the t o t a l oxygen content and the chemical composition of inclusions represented by the at.%Ca:at% A l i n Figures (43) and (44). Figure (40) from RII-I2 shows that the aluminum was able to slowly overcome the s i l i c o n e f f e c t from the electrode. 162 Hence, from the calcium to aluminum, r a t i o of inclusions, the chemical composition of ingots and slag, i t i s inferred that the major reactions which govern the chemical composition of inclusions d e f i n i t e l y involve the CaO and A^O^ from the slag. On the other hand the s i l i c o n - from electrodes, although i t i s transported i n the ingot , Figures (45,46), does not play a role i n the deoxidation scheme. The sequence of in c l u s i o n formation was as follows: 1) i n the bottom part of these ingots, where r e f i n i n g was unstable some inclusions (approximately 5%) contained s i l i c o n . This was almost invariably located i n the i n -clusion core and associated with manganese and calcium. This i s considered as a clear i n d i c a t i o n that some inclusions, at low r e f i n i n g e f f i c i e n c i e s , come from the electrode and i n subsequent stages they are transported d i r e c t l y into the ingot which i s also s o l i d i f y i n g under unsteady state condi-tions. The bulk of these inclusions are, however, mainly represented by small single or clustered type (alumina galaxies) of inclusions, Figures (47) to (48). 2) As the degree of deoxidation i s increased, globular and faceted single inclusions were observed (FeO-A^O^ and a-A^O^ res-p e c t i v e l y ) , Figure (49). These types were usually associated with a s u l f i d e phase (MnS II) and 3) At the highest deoxidation rates, 163 a mixture of spherical and faceted alumina with hexagonal aluminates were observed, Figure (50). This l a t t e r type had a peripheral double, (Mn,Ca)S, s u l f i d e . These findings are shown i n Graphs (51) and (52) where the Ca:Al r a t i o against the S:Mn i n at.% are plotted. These figures were obtained from inclusions i n R I I - I l and RII-I2 respectively. R I I - I l which was a heavily deoxidized ingot showed very highly segregated material. These segregates which under the electron beam produced a red fluorescence occurred i n the t h i r d deoxidation l e v e l , Figure (53) and t h e i r t y p i c a l composition i n at %, was as follows: A l Ca S i Mn 26.23 45.67 17.19 Balance Composition maps shown i n Figures (53a-d) are t y p i c a l of these segregates. Their area ranged from 10ym2 up to 65-70ym2. This finding also confirms the "multi-exchange-reacting" nature of the deoxidation i n the ESR-process, e.g. exchange reactions of the type (12 a-b). These types of segregates were commonly seen i n samples extracted from l i q u i d pools. In these types of samples inclusions containing s i l i c o n and occasionally f a y a l i t e type of inclusions were also found. Based on mass balances the "FeO" content of the slag during r e f i n i n g changed from 0.45 wt % down to 0.27 wt % 164 i n RII-Il and from 0.6 wt % to 0.26 wt % in RII-I2. The lowest "FeO" l e v e l i n slags was also accompanied by a change i n the A^O^tCaO r a t i o ; whereas the C a F ^ content was only s l i g h t l y changed (from 0.5 to 1.0 wt % ) . Thus, the minimum l e v e l of deoxidation reached without s h i f t i n g the slag composition i s about 0.28 - 0.30 wt % FeO for 1.3 and 1.5 CaOiA^O^ r a t i o s i n slags of RI I - I l and RII-I2 respectively, Figures (54, 55). 5.2.1.4 1020 M.S. Ingots Deoxidized Continuously with a  CaSi Alloy Since the e f f i c i e n c y of the Ca-Si all o y i n the small 4340 ingots and i n the large diameter (200 mm) 1020 M.S. intermittently deoxidized was observed to be higher than * i n the aluminum deoxidized ingots, a series of experiments were conducted using equivalent slag systems and degrees of deoxidation as well as melting rates (1 kg min *) with the CaSi a l l o y . Experimental techniques used to determine the t o t a l oxygen content from l i q u i d pool and ingot, the slag chemical analysis, the i n c l u s i o n composition and t h e i r size d i s t r i b u t i o n s and the chemical analysis of ingots were the same as those used to study the 1020 M.S. ingots re-fined through the CaF 2-Al 20 3-CaO system. The Ca-Si a l l o y was * added to the slag i n R I I I - I l and RIII-I2 i n about equivalent * Notice that 1 gram A l i s approximately equivalent to 3.12 grams of the CaSi a l l o y ; Table (IX). 165 aluminum rates as i n RII-Il and RII-I2 namely: 5.11, 11.23, 16.83, 22.24 and p a r t i a l l y 28.0 kg t o n - 1 and 5.11, 11.23, 16.83, 22.44, 28.05, and 56.1 kg t o n - 1 , Table (VIII). The p r i n c i p a l differences between RII I - I l and RIII-I2 are .their slag chemical composition and their deoxidation rates. R I I I - I l and RIII-I2 were refined through 60/36/4 and 50/30/20; (CaF 2, A1 20 3 and CaO) respectively. The three lowest (CaSi) rates were added i n shorter periods of time at the bottom of the ingot i d e n t i f i e d as RIII-I2 and the fourth l e v e l of additions (22.4 4 kg ton ^),was longer than i n R I I I - I l . Due to these differences the t o t a l oxygen analysis shown i n Figures (56) and (57) and hence the i n -clusi o n chemical compositions, Figures (58) and (59) from RII I - I l and RIII-I2 respectively, were the most c r i t i c a l l y affected parameters. Their slag and ingot compositions followed equivalent behavior, Figures (60) and (61) and Figure (62) and (63). The S i and Ca contents i n slags gradually increased whereas the "FeO"and the A l 2 0 3 given as Fe and A l in wt % gradually decreased as the (Ca-Si) deoxidation rates were increased. An important point to note i s that the iron oxide i n slags from both ingots was reduced to about 0.2 wt% as shown i n Figures (64) and (65) without producing strong changes i n the composition of the slags. These graphs consistently show 166 that the aluminum and s i l i c o n gradually increased as the l e v e l of deoxidation i s increased. The i n c l u s i o n chemical composition p r i n c i p a l l y given as the at.% Ca: at% A l r a t i o i n Figures (58) and (59) for R I I I - I l and RIII-I2 respectively, followed an equivalent pattern. This r a t i o , however, was larger in RIII-I2 due to heavier deoxidation i n i t s fourth l e v e l . Another import-ant finding i n regard to the i n c l u s i o n composition was the proportional changes of sulfur (as a CaS) with the Ca:Al r a t i o s and hence with the deoxidation rates, Figures (66) to (68). Inclusions at the bottom of ingots where low rates of deoxidizer were used were mainly faceted and round alumina types, (a-A^C^ and FeO*Al 20 3). At intermediate de-grees of deoxidation faceted alumina and hexagonal aluminates, (a-A^O^ and CaO6Al 20 3) as c l u s t e r s , together with s u l f i d e s (MnS III and (Mn,Ca)S) were found. In the high deoxidation (rate) ranges spherical aluminates enriched i n calcium, (CaO«2Al 20 3, CaO«Al 20 3 and 12CaO•7A1203) associated with peripheral CaS were i d e n t i f i e d . Extreme deoxidation con-d i t i o n s , as i n RIII-I2 and p a r t i a l l y i n R I I I - I l , produced a CaS phase with i n c i p i e n t amounts of aluminum (usually located i n t h e i r core) and the largest segregated material enriched i n deoxidizers (Al, S i , Ca and sometimes Mn). These r e s u l t s indeed corroborate the q u a l i t a t i v e findings previously described for RII-W and RIII-W as well as for the small 4340-ingots which were Ca-Si deoxidized. 167 The above findings (RII-W, RIII W, RIII-Il and RIII-I2) c l e a r l y indicate that the CaSi as a deoxidizer plays by f a r a more important r o l e i n the chemical composi-ti o n of ingot and inclusions than the chemical composition of the slag. The low l e v e l of SiC>2 and the gradual decrement of A^O^ i n the slag, the amounts of Si and Al which are continu-ously increasing i n the ingot and the Ca:Al r a t i o i n i n -clusions immediately suggest that the Si i s u t i l i z e d i n the reacting (ESR) system as a c a r r i e r . I t i s also observed that simultaneous exchange reactions between the two l i q u i d s (slag and metal pool) d e f i n i t e l y contribute to the deoxid-ation mechanism, Figures (60-68). Further evidence that t h i s mechanism, reactions of the type (11) and (12 a-b) , rules the chemistry of the melt, i s seen i n the Si content i n inclusions at high deoxidation rates i n RIII-I2, Figure (68). Excessive (CaSi) deoxidation performed i n t h i s ingot induced the formation of calcium aluminates with peripheral CaS and some s i l i c o n as well as the formation of segregates enriched i n deoxidizers, Figure (69). 168 5.2.1.5 Corroboration and Extension of Previous Findings  to a 4340 Steel CaSi (continuously) Deoxidized Once the l i q u i d pool-slag deoxidation mechanism was unmistakenly i d e n t i f i e d through the previous work, a fu r -ther set of experiments was ca r r i e d out to reconfirm and extend these r e s u l t s to more complex deoxidizers and a l l o y systems such as 4340 and a rotor (Ni-Cr-Mo) s t e e l . Re-melting of these electrodes was performed using the equi-valent experimental conditions as i n the 4 340 electrodes remelted by the small ESR-furnace and the 1020 M.S. e l e c t -rodes remelted by the semi-industrial ESR-furnace. The melting rates were kept i n the 1 kg min 1 range. The l a s t ingot included i n the f u l l y monitored set of experiments (through the three r e f i n i n g stages) was a 4 340-electrode which was refined through a 50 wt. % CaF 2/ 30 wt. % A1 20 3 and 20 wt. % CaO slag and CaSi deoxidized. The degree of deoxidation was continuously increased from 4.17 to 41.67 kg (CaSi) t o n - 1 for almost equivalent periods of time as i n R I I I - I l . The l a s t l e v e l of deoxidation was also suddenly decreased from 41.67 down to 20.83 kg ton Chemical composition of the slag and ingot followed the same pattern as the 1020-ingots, deoxidized with the CaSi a l l o y , Figures (70) and (71). The major difference found i n t h i s ingot was i t s iron oxide and i t s oxygen content, Figure (72). The ir o n oxide given i n Figure (73) changed 169 from 0.4 down to 0.15 as the rate of deoxidation was i n -creased. The average t o t a l oxygen content was about 10-20 ppm. Whereas i n the 1020 ingots either A l or CaSi deoxidized the l e v e l ranged from 30 and up to 80 ppm. This substantial difference i s e s s e n t i a l l y attributed to the t i g h t Ar-atmosphere enclosure of the furnace, the de-gree of (Ca-Si) deoxidation of these ingots (50-70 ppm of oxygen) and the chemical composition of the electrode. The chemical composition of inclusions given i n Figures (74 a,b), followed the general trend observed in the equivalent 10 20 M.S. ingots. The chemical composition of inclusions i n similar manner to previous 1020 M.S. ingots deoxidized with the CaSi a l l o y exhibited the gradu-a l t r a n s i t i o n from aluminate to calcium aluminate phases. Their proportional increment i n sulfur content (as CaS) up to 25 at. % was also observed, as the deoxidation rate was increased, Figure (75). A very important fact to address i n these experiments i s the r a t i o of the S i i n the CaSi-alloy even though i t i s constituted by approximately 62.0 wt. % S i , i t has not played a role in the deoxidation scheme previously presented, i . e . involving the CaO and A^O^ of the slag and the Ca and A l i n inclusions and deoxidizers . It i s also important to emphasize that the S i from either the electrode (in 170 the 1020 electrodes) or the CaSi all o y i s v i r t u a l l y trans-ported conjointly to the A l and Ca into the ingot and i t does not appear i n inclusions. At t h i s point i n t h i s description, i t becomes e s s e n t i a l to r e c a l l that although the Al as a deoxidizer seems to be operating within the same frame of reactions as the CaSi (slag-deoxidizer and l i q u i d pool), i t due to k i n e t i c fac-tors, does not deoxidize the ESR-melt as e f f i c i e n t l y as the CaSi a l l o y . In the l i g h t of these findings a new set of three experiments was planned. The major purpose was to corroborate the conclusion that the s i l i c o n i n the de-oxidizer works exclusively as a c a r r i e r , to reconfirm the v a l i d i t y of the proposed mechanism (reaction scheme) and to compare the degrees of deoxidation reached through three d i f f e r e n t deoxidizers, Table (VIII). Three rotor (Cr-Mo-V) ste e l electrodes were refined in the 200 mm diameter ESR-furnace using equivalent melting rates, slag system and deoxidation rates, Table (XII). The (Si-based) deoxidizers were CaSi, SiAlCaBa ("hypercal") and an A l S i a l l o y . Their chemical compositions are given i n Tables (XII a-c). Despite the f a c t that the chemical composition of the deoxidizers i s quite d i f f e r e n t calcium aluminate phases with peripheral calcium sulphide are precipitated. The Ca:Al r a t i o vs. the sulfur content are plotted i n Figure (76). 171 Again i t i s shown that exchange reactions of the type (12 a-b) play the most important role i n the deoxidation (reaction) scheme and since the S i appears i n the ingot but i t does not i n inclusions i t s role i s p r i n c i p a l l y as a c a r r i e r of Ca and A l into the l i q u i d pool and ingot. These r e s u l t s support the three previous proposals. The chemical composition of inclusions indicate that s i l i c o n free, calcium aluminates with peripheral calcium s u l f i d e (i.e. others than C a O 6 A l 2 0 3 and (Mn,Ca)S) were the pre-c i p i t a t i n g phases. The s i l i c o n content i n the slag was considerably increased and the "FeO"content was also held i n the ranges previously described, Table (XII). Furthermore, segregates which were found i n the ex-cessively and abruptly A l and CaSi deoxidized ingots were the same as observed i n (size, fluorescence under the e l e c t -ron beam and t h e i r chemical composition) the A l S i deoxidized ingot. The d i s t r i b u t i o n of elements in a segregate enriched i n strong oxide formers, namely aluminum, s i l i c o n and p a r t i -c u l a r l y calcium, i s shown i n Figures (77a-d). Their t y p i c a l composition was (in at.%) as follows: Ca A l Si Mn + Fe 40.6 41.5 19.7 balance This finding i s considered as another evidence that the pro-posed mechanism i s indeed operating. 172 5.3 Discussion of Results i n Terms of Electrode and Slag Composition, Related to the Second Question 5.3.1 The E f f e c t of the Electrode on the Inclusion Comp-o s i t i o n of ESR-ingots The experimental findings previously described have been used to address the second question stated i n Chapter III, i . e . i s the i n c l u s i o n composition controlled by the chemical composition of electrodes, slag or deoxidizers? This question as envisaged i n the l i t e r a t u r e review (in terms of the complexity of the reaction scheme and the i n -cl u s i o n chemical composition of ingots i n the ESR-process) and as shown in the previously described r e s u l t s cannot be answered unless the reactions between the three d i f -ferent stages of r e f i n i n g and i t s components (electrode, slag, s l a g - l i q u i d f i l m , deoxidizer, l i q u i d pool-slag and ingot) are c a r e f u l l y monitored. Through the f i r s t part of t h i s research i t has been concluded that inclusions from electrodes under stable r e f i n i n g conditions are completely dissolved i n the e l e c t -rode t i p . Thus, further discussion assumes t h i s f a c t . The elucidation of the role played by the electrode and the slag on the chemical composition of inclusions i n the ESR-ingot have been shown through several experiments: 1) The small 4340 ingot .(6), Table (X), refined through 173 an alumina free slag and i n the absence of gaseous oxid-ative and deoxidant (external) sources, i . e . under argon and without deoxidant. 2) A large diameter (200 mm) 1020 M.S. electrode refined through a CaF 2-CaO-Al 20 3 slag under the above conditions, (RI-Il), Table (VIII) and 3) A series of 1020 M.S. electrodes refined i n the 200 mm i n diameter ESR-furnace which were intermittently deoxidized with CaSi and others continuously deoxidized with A l and a CaSi a l l o y . 1. The ingot (6) which was remelted through the 31 wt.% CaF 2, 46 wt. % CaO, and 23.0 wt. % S i 0 2 has con-c l u s i v e l y shown that the electrode composition indeed plays a dominant role i n the f i n a l i n c l u s i o n chemical composition of the refined ingot. The alumina content i n inclusions represented as Al in Table (VIII) with respect to s i l i c a , given as s i l i c o n , shows that the aluminum which came ex-c l u s i v e l y from the electrode (in solution) has co n t r o l l e d the i n c l u s i o n chemical composition. This finding i n agree-(83) ment with Rehak 1 s et al.'s i n Al and CaSi treated electrodes c l e a r l y demonstrates that the chemical composition of e l e c t -rodes s p e c i f i c a l l y due to the presence of deoxidizers i n solution, under a given slag system can play an important r o l e in the f i n a l i n c l u s i o n composition of refined ingots, i . e . self-deoxidation. 174 2. Results found from a 1020-ingot remelted under a protective atmosphere without the influence of a deoxidizer are shown i n Figures (37,38). These graphs from RI-Il c l e a r l y i l l u s t r a t e that as the remelting of the electrode i n the 200 mm mould takes place an accumulation of s i l i c a in the slag and a gradual change i n the alumina and calcium oxide occurs. The SiC^ i n the slag acts exclusively as a diluent and only contributes to a gradual s h i f t i n the CaOiA^O^ II II (38 r a t i o s and to control the i n t r i n s i c "FeO"content i n the slag It does not, however, influence the chemical composition of i n c l u s i o n s r A deeper discussion related to t h i s matter w i l l be pursued i n a subsequent section. It i s pertinent to c l a r i f y that the change i n the CaOrA^O^ r a t i o was not i n -fluenced by the formation of v o l a t i l e f l u o r i d e s (AlF^, S i F 2 or HFl) since the f l u o r i n e analysis was changed only in the measure of the experimental error, (± 0.2%). The continuous increment of S i 0 2 and "FeO"in the slag produces a s h i f t in the CaOtA^O^ r a t i o i n the s l a g — g i v e n as the t o t a l A l and Ca content i n Figure (37)--and an A l depletion i n the ingot, Figure (38). It i s worthwhile to point out that i n spite of the (gradual) increase i n Mn and Si (0.65 to 0.72 wt. % and 0.09 to 0.13 wt. % respectively) i n the ingot, only aluminates (A^O^ and A^O^'FeO) and s u l f i d e s ((Fe,Mn)S and MnS II) were pre c i p i t a t e d . 3. A t y p i c a l a l t e r a t i o n of the otherwise continu-ously increasing content of s i l i c a i n the slag and hence Si i n the ingot i s shown i n Figure (32). This ingot i d e n t i -f i e d as RIII-W was remelted under argon and under an equi-valent slag to R I - I l . The abrupt changes i n the S i and Mn contents i n the slag and ingot and'the Ca:Al r a t i o i n i n -clusions are a r e s u l t from the intermittent additions of the (CaSi) deoxidizer. Since the s i l i c o n comes from the e l e c t -rode and the deoxidizer a more appropriate analysis should be performed on the Mn since i t comes exclusively from the electrode. The Mn content changes from about 0.65 down to 0.45 wt. % i n the slag, Figure (32). The behavior of the S i from the electrode becomes more important when the 1020 M.S. electrodes are A l deoxidized. The ingot i d e n t i f i e d as RII - I l c l e a r l y shows that as the degree of deoxidation i s increased the l e v e l of S i and Mn in the slag and i n g o t — F i g u r e s (39) and (45)—are held to a constant l e v e l . This suggests that the chemical composition of the electrode no longer plays a role i n the r e f i n i n g pro-cess. This ingot was refined, as shown i n Table (VIII), through a Si and Mn free slag under an argon atmosphere. Thus, the only source of S i (0.25 wt. %) and Mn (0.6 - 0.7 wt. %) either in solution or as inclusions i s the electrode. The above r e s u l t s c l e a r l y indicate that i f an electrode 176 contains oxide forming elements i n solution or as inclusions, which are not present as oxides in the slag and are stronger deoxidizers than those present i n the slag, a cooperative deoxidation takes place. The mechanism by which these re-actions take place i s within the general scheme given by reactions (11) and (12 a-b). The most important conclusion from these experiments i s that these reactions predominate sol e l y i n the absence of or under i n e f f i c i e n t deoxidation. A conventional Al-deoxidation (0.05 - 0.2 wt. %) i s able to overcome the e f f e c t of the Si from the electrode i f the r e f i n i n g slag i s low (< 10 wt, %) i n SiO,,. 5.3.2 Elucidation of the Ef f e c t of Slag and Deoxidizers  (Preliminary Studies) The elucidation of the e f f e c t of slags and deoxidizers was approached through the small 4340 ingots. Ingots (1) to (6) which were refined through d i f f e r e n t Si0 2-containing slags, yielded calcium aluminate and calcium-aluminum s i l i -cate phases. The former type of inclusions did not exceed 3.75 at. % Ca and they were i d e n t i f i e d only where the S i 0 2 ~ content in the slag was less than 10 wt. %. Above t h i s per-cent the l a t t e r type of phases were i d e n t i f i e d . Table (X) summarizes the main features of these experiments. 177 I f the above analysis i s extended to both slags and i n c l u s i o n compositions and they are plotted i n tern-(14 53 54) ary diagrams as suggested by A l l i b e r t et a l . ' ' (207) for ESR ingots and slags and Bruch and K i e s s l i n g and (91) Lange for CaSi deoxidized ingots i n conventional s t e e l -making processes then the chemical compositions of i n -clusions and t h e i r probable o r i g i n i s e a s i l y followed. Figures (12) and (13) . Several important points should be considered i n these diagrams: i) the slag chemical composi-ti o n i s plotted under the assumption that the CaF 2 acts as an i n e r t d i l u e n t ( 3 3 ' 3 8 ' 4 5 ' 5 3 ' 8 8 ' 8 9 } . i i ) regarding the t e r -nary diagram which describes the i n c l u s i o n composition, (91) Kiessling's and Lange*s studies as an extension of (207) Bruch's have considered a replacement of either Mn, Mg or Fe by Ca as oxides i n the CaO corner. The replacement (34) of Mn by Ca as oxides has also been reported by Holzgruber in ESR ingots. And the two divergent l i n e s which emerge from the A^O-j corner to the CaO-Si0 2 binary are l i n e s which rep-resent the maximum and minimum MeO:Si0 2 r a t i o s , L^ and L 2 respectively, i n inclusions. i i i ) The t r i p l e l i n e which also emerges from the A l 2 0 3 ~ c o r n e r are Holappa et a l . ' s ' 1 0 ^ findings from Ca-treated (conventional) ingots and iv) data obtained by other researchers i n c o n v e n t i o n a l ' 4 0 ' 1 4 4 ' 15115783) ' ' and i n ESR-research i s also included i n t h i s diagram. 178 Slags containing less than 10 wt. % S i 0 2 produce i n -clusions l y i n g along the A^O^-CaO binary and s p e c i f i c a l l y located closer to the A l 0_-corner. These types were 2 3 J c usually associated with s u l f i d e s (MnS II, MnS III or (Ca, Mn)S). The type and composition of these s u l f i d e s (34) depended on the Ca:Al r a t i o in the oxide phase. Holzgruber (14 53 54) and A l l i b e r t et a l . ' ' who have studied the reaction (12 v i i i ) , have also reported these i n c l u s i o n phases, Figures (12,13). The chemical composition of inclusions which were ob-tained from slags containing more than 10 wt. % S i 0 2 were located along the SiC^-A^O^ axis. Ingots (2), (3), (8), (10) and (11) were located i n an area confined i n the 20 to 85 wt. %. Al 202 range along the SiC^-A^O^ axis and about 30 wt. % CaO. These r e s u l t s are in agreement with Bruch's and Kiessling's and Lange's^^' and q u a l i t a t i v e l y with A l l i b e r t ' s et a l . ' s ( 5 3 ) , Rehak's ( 8 3 ) and Holzgruber ' s ( 3 4 ) . Findings from the l a t t e r two researchers were obtained from ESR experiments where experimental conditions were not d i r -e c t l y concerned with the slag chemical composition. Thus, i f CaF 2 i s s t r i c t l y considered as an i n e r t diluent (234 ) and Rein's and Chipman's data are used to substantiate the thermodynamic behavior of s i l i c a i n ESR-slags then i f a_.__ £ 0.01, a,,. S 0.5 and a„ . 5 0.1 , aluminates S i 0 2 AlO^ 5 CaO w i l l p r e c i p i t a t e . On the other hand i f the S i 0 2 content of slags exceeds 10 wt. % then aluminum-silicate inclusions with some calcium (up to about 30 wt. % CaO) w i l l p r e c i p i t a t e . The a c t i v i t y r e l a t i o n s are as follows: a„.^ ^0.01, S i O „ ' a C a 0 < 0.10 and a A 1 Q ^ 0.5. Holappa's r e s u l t s v ' 1. 5 from Ca-treated ingots suggest that as the a c t i v i t y of Al in the melt i s reduced and the a c t i v i t i e s of Si and Ca are i n -creased, a gradual change i n composition of the inclu s i o n phase as indicated by the 'triple-dotted' l i n e i n Figure (12) should be observed. This behavior i n ESR ingots (1) to (6) was not completely followed and instead a mixture of calcium- s i l i c o n -aluminate and aluminate phases were found, Table (X). The best deoxidation measured as the CaO:Al 20 3 r a t i o i n i n c l u s i o n s * 2 1 * 3 ^ was found i n the CaF 2-CaO-Al 20 3 system, s p e c i f i c a l l y where the CaF 2 and CaO contents were 50 wt. % and 20 wt. % respectively i n the slag. Polish researcher's work ^ ^ ' o n slags belonging to t h i s system have reported a maximum of 71.0% reduction of non-metallic inclusions in r e l a t i o n to the st a r t i n g s t e e l electrode i n exactly the same slag composition (50/30/20) at which the largest Ca:Al r a t i o i n inclusions was found i n t h i s research, Table (X). Thus, through the previous description i t has been found that the slag d e f i n i t e l y plays a ro l e i n the r e f i n i n g process. What has not been answered yet i s how large the 180 slag e f f e c t i s i n comparison with the deoxidizer. This question can be answered in a t r i v i a l manner by analyzing r e s u l t s i n Table (X). By comparing the i n c l u s i o n chemical composition of ingot (1) against (7) and (9), one can ob-serve that the Ca-content in inclusions i n the l a t t e r two ingots i s double that i n the former. Thus, on an a p r i o r i basis, one could e s t a b l i s h that since deoxidation rates and slag were equivalent (2.3 kg ton ^, 50/30/20) and the Ca:Al r a t i o s i n inclusions i s twofold i n the A l and the CaSi deoxidized ingots (7) and (9) respectively then the de-oxidizer overcomes the slag e f f e c t at deoxidation rates larger than 1.15 kg ton And i f these findings are ex-trapolated, i t could be estimated that since the Ca i s almost insoluble in molten iron then a r e l a t i v e l y higher deoxidation rate could generate aluminates riche r in calcium and hence much lower t o t a l oxygen contents. These premises, however, do not rest on any mechanism and consequently they do not explain why the A l and the CaSi deoxidation produce an almost equivalent trend i n r e s u l t s i n terms of i n c l u s i o n compositions. I t i s important to c l a r i f y that although the amounts of deoxidizers (2.3 kg ton 1 ) were equivalent the CaSi a l l o y contains 62.5 wt.% S i . The e f f e c t of deoxidizers on the i n c l u s i o n chemical composition, i n the small 4 340 ESR ingots refine d through the 55 wt.% CaF 2, 15 wt.% A l 2 0 3 , 15 wt. % CaO and 15 wt. % S i 0 0 slag i s shown in Table (X). 181 I f i n g o t s (8) and (10) which were A l and CaSi de-o x i d i z e d are compared a g a i n s t i n g o t (2), one can see t h a t the CaSi d e o x i d i z e d i n g o t d i d show the i n f l u e n c e of the s i l i c a . The S i C ^ e f f e c t from the s l a g i n i n g o t (8), how-ever, was almost completely suppressed by the d e o x i d i z e r , ( A l ) . To c o r r o b o r a t e these f i n d i n g s a 1020-electrode 76.2 mm i n diameter was remelted under e q u i v a l e n t c o n d i t i o n s . T h i s p a r t i c u l a r i n g o t (11), however, was A l - d e o x i d i z e d at a c o n s t a n t r a t e (0.02 Kg/ton ). The e f f e c t of the d e o x i d i z e r , Table (X) and i n c l u s i o n composition p r e v i -o u s l y c i t e d , was not enough to counterbalance the S i C ^ e f f e c t o f the s l a g . The i n c l u s i o n phases were " a n o r t h i t e " and alumina. I t i s worthwhile to p o i n t out t h a t the s i l i c o n content i n the i n c l u s i o n phases of i n g o t s (2) and (8) are one order of magnitude l a r g e r than the aluminum-deoxidized i n g o t s . And although t h e r e i s a d i f f e r e n c e i n the chemical compo-s i t i o n o f e l e c t r o d e s (1) to(8) and (9) and (10) i n terms of t h e i r s i l i c o n c o n t e n t (SiC^ i n i n c l u s i o n s ) , the main e f -f e c t i s a t t r i b u t e d to the l a r g e q u a n t i t i e s i n the (CaSi) deoxidant and the chemical composition of the s l a g , i . e . , l a r g e r than 10 wt. %. Thus, the chemical composition of the s l a g a t which an a p p r o p r i a t e d e o x i d a t i o n i n terms of the CaSi treatment, should be c a r r i e d out i s a t s i l i c a c o n t ents s m a l l e r than 10 wt. %. 182 Hence, the most important conclusion from previous findings i s that the deoxidation phenomenon i s a net re-s u l t of cooperative reactions between deoxidizers and slags. Thus, to obtain an e f f i c i e n t deoxidation an appropriate se l e c t i o n of these parameters i s e s s e n t i a l . A further proof of these statements w i l l be approached in the next section. It i s also worthwhile to c l a r i f y that although there i s a common trend of res u l t s i n both ESR-furnaces the v a l i d i t y of previous findings i s q u a l i t a t i v e . The lower sur-face area available for reactions, the higher thermal gradi-ents and the unsteadiness of the s o l i d i f i c a t i o n conditions i n the small ESR-furnace are factors that must be con-sidered. 5.3.3 Preliminary Discussion on the Deoxidation Mechanism In the l i g h t of previous discussion of findings, the evaluation of the o r i g i n a l q u e s t i o n — I s the inclu s i o n comp-o s i t i o n c o n t r o l l e d by the chemical composition of electrodes, slag or deoxidizers?—becomes i r r e l e v a n t and instead other questions emerge, i . e . what i s the mechanism by which the deoxidation (in i n d u s t r i a l slags and deoxidants) takes place? What i s the ro l e played by the slag? and What are the condi-tions which an appropriate deoxidation i s carr i e d out under? To s a t i s f a c t o r i l y answer th i s set of questions a summary 183 of experimental r e s u l t s correlated to the general theory i s presented i n advance. The following section emphasizes the r e s u l t s obtained with the 20 0 mm ESR ingots continu-ously or constantly deoxidized. Findings from the small (4340) ESR ingots, Table (X) suggest that the chemical composition of ingots and i n -clusions (Ca:Al ratios) i s determined solely by the CaO: A^O^ r a t i o in the slag when the Si02 content i s lower than 10 wt. % and deoxidation i s absent. The r e s u l t s obtained through the 200 mm ESR-furnace have confirmed these f i n d -ings and they have also contributed to the formulation of a more self - c o n s i s t e n t deoxidation model. The inherent reaction scheme (1-5) at the electroactive • ^ * • t v „r.T, (27,28,30, 33,34) interfaces i n the ESR-process • ' ' / t h e atmos-n • (1,34,38) , . . phere-slag-liquid pool in t e r a c t i o n (in terms of the oxygen transport). and the amount of scale (as a re-s u l t of the thermal history of the electrode) introduced i n the slag are the main factors which determine the o x i -dative state of the slag and hence the oxygen po t e n t i a l i n the l i q u i d pool. Thus, the deoxidant, the sequence and degree of de-oxidation as well as the slag system are parameters which should be adequately selected to optimize deoxidation without s a c r i f i c i n g the chemical i n t e g r i t y of ESR-ingots. The appropriate control of the '^ed1 i n the slag by the de-184 oxidation w i l l influence the l e v e l of oxygen i n the molten pool and consequently the p r e c i p i t a t i o n of inclusions. The importance of the electrochemical reactions on the "FeO" l e v e l of a melt and t h e i r influence on the chemical composition of inclusions i s seen through ingot (14) in Table (XI). This rotor (Cr-V-Mo) s t e e l electrode was sur-face ground, thus scaling formed during mechanical working was removed from i t s surface (about 1 mm). This electrode was also coated with an Al-Mg spinel painting to prevent i t s oxidation during r e f i n i n g . The chemical analysis of i n -clusions i n t h i s (ESR)-ingot show that despite the pre-vious surface preparation, the argon atmosphere enclosure and a s l i g h t Al-deoxidation (0.02 kg ton ^ ) , round, single and c l u s t e r s of aluminates (mostly FeO-A^O^ and A^O^) and iron oxides and s u l f i d e s (FeO, FeS) and (Mn,Fe)S) were de-tected instead of the a-A^O^ expected from the s l a g - l i q u i d (218) metal i n thermodynamic equilibrium , i . e . according (234) to Kuo Chu Kun ' s phase diagram th i s slag i s A^O^-saturated. On the other hand, despite large deoxidation rates (10 kg t o n - 1 ) there i s a minimum "FeO" l e v e l i n the slag which can be achieved for a given deoxidation practice. The e f f e c t of the atmosphere as well as the a r t i f i c i a l l y introduced "FeO" i n the slag and hence in the l i q u i d pool are c l e a r l y shown in Figures (32,33) and (34,35) which belong to the ingots i d e n t i f i e d as RII W and RIII W. 185 The i n i t i a l stages of r e f i n i n g are controlled by the slag composition which i n i t s e l f allows a given amount of " F e O ( 1 , 3 8 ' 8 2 ) . If the "FeO" content i n the slag i s per-mitted to r i s e above this l e v e l by any of the described mechanisms the r e s u l t i s an increased rate of oxidation of the reactive species from the electrode or the slag, hence moving the system towards unacceptable r e f i n i n g conditions, i n terms of slag or ingot composition. The reaction (12-iv) 2[A1] + 3 (FeO) t ( A l ^ ) + 3Fe (12-iv) which has contributed to produce the gradual depletion of A l i n the ingot; i d e n t i f i e d as R I - I l , Figures (37,38), at expense of the slag (change i n the CaOiA^O^ r a t i o i n the slag) has also been influenced by the presence of the gradual introduction of "FeO" (mostly as scale) into the slag as the r e f i n i n g took place. The CaOtA^O^ r a t i o i s controlled by 2[A1] + 3(CaO) J (A1 20 3) + 3[Ca] (20-C) Although the extent to which t h i s reaction occurred was very limited, i t s r e s u l t s were detected, Figures (37) and (38). These reactions can take place only when there i s not enough deoxidizer to suppress the continuously increasing amount of "FeO" in the slag. Under these conditions the major pre-c i p i t a t i o n reactions are governed by the oxygen po t e n t i a l as follows: 186 Fe + 2[A1] + 4[0] t (FeO«Al„0 ) 2 3 inclusion (22) or 2[A1] + 3[0] t (Al o0_) 2 3 inclusion (11-i) These p r e c i p i t a t i o n reactions are controlled by reaction (12-iv) i n the A l deoxidized ingot. Once the oxygen pot-e n t i a l has been decreased as a r e s u l t of the low but f i n i t e "FeO" content i n the slag then the reaction (20) plays a very important r o l e . Figures (45) and (46) which represent the behavior of the Al-deoxidized ingots show that the Al as a deoxid-i z e r i s v i r t u a l l y introduced into the ingots (RII-Il and RII-I2). I t can also be seen that the"FeO"content i n the slag i s decreased to about 0.3 wt. %. 7 Figures (54,55). The i n c l u s i o n composition, however,•changes gradually from FeO«Al 20 3 to a - A l 2 0 3 and at. very high deoxidation rates the CaO«6Al 20 3 i n c l u s i o n phase i s p r e c i p i t a t e d , Figures (47-50). The t o t a l oxygen analysis also r e f l e c t s the deoxidation trend, Figures (41,42). From a consideration of these r e s u l t s i t i s i n f e r r e d that to a limited extent the "deoxidation Phenomenon" i s also attributed to the reaction (20). This reaction induces the p r e c i p i t a t i o n of the CaO-eA^O^ phase and also (up to the same degree) the p a r t i a l substitution of Mn by Ca i n the s u l f i d e phases,. Figures (50,52). As a consequence of these reactions inclusions are precipitated according (140,144,147,148,216) to: 187 m CaO + n A l 2 0 3 J [m(CaO) «n(A1 20 3)] i n c l u s i o n (19) or i n a more s p e c i f i c formulation: CaO + 6(A1 20 3) t CaO-6Al 20 3 (19-a) Since these reaction products are i n equilibrium with oxygen and sulfur then a s u l f i d e phase can be pre c i p i t a t e d (147, 148,197,210,215),. according ' ' ' ' to: ic ic CaO + [S] t CaS + [0] (21) For low CaO content phases such as the CaO«6Al 20 3, a su l f i d e phase such as the double s u l f i d e (Mn, Ca)S should be expected to (heterogeneously) p r e c i p i t a t e on oxides. Although very rare when the s u l f i d e phase did not contain Ca, i t was faceted and contained only Mn and S. Hence, i t was c l e a r l y i d e n t i f i e d as MnS II I . The 1020-M.S. ingots intermittently deoxidized with the CaSi a l l o y have c l e a r l y reconfirmed that the deoxidation reactions are not exclusive of each other. Instead a cooperative process between the deoxidizer, slag and l i q u i d metal takes place. At a given discrete addition of CaSi into the slag i t s "FeO" content decreases, Figures (32) and (33). Simultaneously to t h i s change an increment of A l i n the ingot and a decrement of A l 2 0 ^ in. the slag i s also observed. As a r e s u l t of the above coincidental reactions a net change i n t o t a l oxygen content which i s a consequence of the inc l u s i o n quantities and compositions i s expected. A reduction from 75 down to 30 ppm was ob-it represents a peripheral phase on inclusions 188 served. Another inter e s t i n g finding i s that the reduction of t o t a l oxygen content was very dependent on the melting conditions. RII-W was melted under a i r whilst RIII-W was under an argon blanket. The in c l u s i o n phases i d e n t i f i e d i n t h i s experiment (RII-W and RIII-W) were e s s e n t i a l l y a-A^O^ and FeO-A^O^ as clusters of small spherical pre-c i p i t a t e s . The samples c r i t i c a l l y affected by the additions of d e o x i d i z e r — F i g u r e (36)—showed Ca:Al r a t i o s (at. %) which c l o s e l y correspond to the formation of the CaO«6Al 20 3 phase. This phase was metallographically i d e n t i f i e d be-cause of i t s faceted hexagonal appearance and i t s peripheral envelope of ( C a , M n ) S ( 1 5 ' 9 4 ' 1 4 5 ' 2 0 3 ' 2 1 0 ) . The a - A l 2 0 3 and the FeOA^O^ were observed closer to the FeO additions and they were associated with (Mn,Fe)S or MnS II. The series of findings from the continuously CaSi de-oxidized ingots remelted through the CaF 2-Al 20.j-CaO slag i n the 200 mm diameter moulds can be condensed as follows. The slag chemical analysis have shown—Figures (60) and (61)-^that as the degree of deoxidation i s increased the CaO content of the slag i s increased while the A l i s de-creased. Although the f l u o r i n e analysis showed a s l i g h t decrement suggesting the formation of v o l a t i l e fluorides by reactions (8 to 10), the major changes, however, were 189 due to the reaction between the deoxidizer and the slag. The "FeO'1 Content of the slag was gradually reduced as the deoxidation rates were increased. The i n i t i a l deoxidation changes are adequately described by the following reactions: [Ca] + (FeO) t Fe(1) + (CaO) (12-v) 2[A1] + 3(FeO) t 3Fe(l) + ( A l 2 0 3 ) (12-iv) At t h i s degree of deoxidation of the slag alumina type of inclusions and MnS II were v i r t u a l l y the only phases pre-c i p i t a t e d . As the l e v e l of deoxidation i n the slag was increased Ca-aluminates with increasing CaO-content were observed, Figures (58,59) and (66 ,67). The in c l u s i o n chem-i c a l analysis revealed the following p r e c i p i t a t i o n se-quence: a - A l 2 0 3 , CaO«6Al 20 3, CaO«2Al 20 3, CaO«Al 20 3 and traces of 12Ca0«7Al 20 3 at extreme degrees of deoxidation (^0.2 wt "FeO"). Beyond th i s deoxidation l e v e l the form-ation of segregates enriched i n Ca, A l , Si (and Mn) and the formation of (Al, Ca)S were seen. These su l f i d e s showed (EPMA analysis) that the A l ( A l 2 0 3 ) although i n very low amounts was p r e f e r e n t i a l l y located i n the incl u s i o n core. Coincidental to the p r e c i p i t a t i o n of the oxide phases a s u l f i d e phase, which was enriched i n calcium proportional to the amount of CaO i n the Ca-aluminate phase was ob-served. These analysis lead to the conclusion that the mechanism by which the deoxidation-precipitation occurs i s 190 as follows: once the l e v e l of the "FeO" i n the slag has reached i t s minimum the transport of aluminum and calcium into the l i q u i d pool by: 3[Ca] + (A1 20 3) t 2[A1] + (CaO) (20-c) takes place. Thus, the lev e l s of deoxidation are propor-t i o n a l to the amount of A l and Si in the ingot and also to the amount of CaO i n the i n c l u s i o n phases ( i . e . Ca:Al r a t i o s ) . Hence the p r e c i p i t a t i o n sequence, i n terms of the degree of deoxidation i s as follows: i n the absence of or at very low deoxidation rates i n a conventional ( i n -d u s t r i a l ) CaF 2-Al 20 3-CaO slag, the p r e c i p i t a t i o n i s gov-erned by: 2[A1] + 4[0] + Fe = (FeO«Al 20 3) in c l u s i o n (24) 2[A1] + 3[0] = (A1 20 3) i n c l u s i o n (11-ii) The former oxide (Fe0*Al 20 3) i s usually associated wi th (Mn,Fe)S or MnS II. With the l a t t e r oxide (a-Al 20 3) MnS II or MnS III are usually observed. The degree of (Al) de-oxidation dictates the formation of a s p e c i f i c s u l f i d e . At intermediate deoxidation lev e l s the reaction (20-c) st a r t s to operate according to reaction (19) or (19a). 191 This p a r t i c u l a r oxide phase (CaO«6Al 20 3) was commonly observed with either MnS III or (Ca,Mn)S. Hence suggesting that i n addition to reaction (20-c) the following reaction [Ca] + MnS t CaS + [Mn] (25) was also taking place. At r e l a t i v e l y high deoxidation levels, where the re-action (20-c) e n t i r e l y controls the transport of deoxidizers i n the melt and when very low "FeO" content i n the slag i s reached the p r e c i p i t a t i n g phases are CaO«2Al 20 3 with either (CaMn)S or CaS and CaO«Al 20 3 and 12CaO«7Al 30 3. These l a t t e r two oxide phases were surrounded by an envelope of CaS. This mixture of phases (oxide and s u l -fides) suggests that the calcium oxide from the calcium aluminates strongly interacts with the sulf u r and oxygen i n solution according to: (CaO)* + [S] = (CaS)* + [0] (21) This reaction generated a CaS enriched phase wherever the minimum oxygen content was reached (^20-30 ppm). F i n a l l y , at extremely high deoxidation rates where the CaO:Al 20 3 r a t i o in the slag i s d r a s t i c a l l y and suddenly shifted, the formation of small segregates enriched i n Ca, A l and Si can occur. Under these deoxidation conditions 192 the formation of (Al,Ca)S and a peripheral envelope of Si-phase around the CaS (which surrounds the calcium aluminate) were also observed. Similar r e s u l t s were obtained with 4 340 and rotor (Cr-Mo-V) s t e e l s . This allows the formulation of a comprehensive mechanism discussed i n the next section. It i s worthwhile to mention that while there are some , . (151,153, 157,214,221,235,236) . . . ^  . . advantages • ' • ' • ' ' ' to p r e c i p i t a t i n g aluminates enriched on calcium oxide (because of their round shape and t h e i r CaS envelope) instead of the alumina galaxies and the manganese su l f i d e s the coincidental transport of aluminum into the refined ingot constitutes a p o t e n t i a l prob-lem. This behavior can be represented as i n the calcium i n -(210) , j e c t i o n processes by: X[Ca] + Y ( A l 2 0 3 ) i n c l u s . o n i x C a O ( Y - f ) A l ^ + § X(A1) (20-b) This reaction as Holappa* 2 1 0^ has pointed out i s equival-ent to reaction (20-c). Thus, i f an excess of deoxidizer (CaSi) i s used A l i s introduced i n the melt and hence the p o t e n t i a l • (221-223) to reduce the mechanical properties i s enhanced 5.3.4 Comprehensive Discussion on the Deoxidation Mechanism When Al or Ca i s added to the ESR-slag as a deoxidant, i t becomes part of a reaction scheme represented by: 193 [Ca] + (FeO) % Fe(l) + (CaO) (12-v) 2[A1] + 3(FeO) + Fe(l) + A l 2 0 3 (12-iv) 3[Ca] + (A1 20 3) % 2[Al] + 3(CaO) (20-C) At high l e v e l s of "FeO", reactions (12-v) and (12-iv) w i l l pre-dominate leading to a simple deoxidation scheme i n which the deoxidant addition appears as the appropriate slag oxide component. At low lev e l s of "FeO" reaction (20-c) w i l l take over (12-v) and (12-iv) and i t w i l l be observed that an ex-(46 47 52 change reaction (similar to those already reported ' ' ' 54) for S i / A l and Ti/Al) i n which the a l l o y Ca:Al r a t i o i n the ingot w i l l be determined by the CaO:Al20 3 r a t i o i n the slag low i n s i l i c a , (less than 10 wt. % S i 0 2 ) , not by the rate, form or composition of the deoxidant. In determining the point at which reactions (12-v) and (12-iv) w i l l give way to (20-c) i n the sense of producing ingot composition, the l e v e l of slag FeO-activity i s evidently of prime import-ance. It i s more important than the i n t r i n s i c slag and the electrode chemical composition. At low FeO-activity i n the slag i f large quantities of Ca are added as a deoxidizer, the r e s u l t i n a conventional ESR-slag composition w i l l be a corresponding increase i n the a l l o y A l l e v e l , through reaction (20-c). 194 In order to i l l u s t r a t e the role of reactions (12-v) to (12-iv) the re s u l t s obtained from the 4 340 ingot were selected as the prototype of the general behavior to approach t h i s discussion. The deoxidation sequence i n terms of the chemical composition of the ingot and the slag are shown i n Figures (70) and (71). Applying a mass balance to t h i s ingot and the slag, i t i s apparent that the s i l i c o n addition (from the deoxidizer), appears almost quantitat-i v e l y i n the ingot with very l i t t l e of SiC^ to the slag. I t i s also observed that the concentration of A^O^ i n the slag decreases while the CaO content (represented as a part of the t o t a l Ca content) increases. U t i l i z i n g t h i s data i n conjunction with the increased A l assay of the ingot leads to an excellent closure of a mass balance drawn on the system using equation (20-c). In consequence, i f the following stoichiometric r e l a t i o n s h i p * 2 1 0 ^ which re-lates the reduction of alumina by Ca i n the chemical comp-o s i t i o n of the inclusions, X[Ca] + Y(Al o0,). , t XCaO-(Y - *-) A l o 0 , + \ X[A1] z J i n c l u s i o n J z 6 s (20-b) i s u t i l i z e d to perform the balance, equivalent r e s u l t s and a very close prediction of the inclu s i o n chemical composition i s obtained. These re s u l t s are taken as a strong evidence that the calcium component of the CaSi a l l o y addition has re-duced A^O^ from slag, following (20C) producing A l i n the 195 ingot and leading to a CaO increase i n the slag. This re-duction i s stoichiometric, as would be e x p e c t e d ' 4 7 ' 1 4 8 , 210 216) ' from an equilibrium analysis of (20) . During the process of r e f i n i n g , as expected ' ' 2 7 ' 3 8 ' ^ 7 ' 8 2 ^ the "FeO" l e v e l i n the slag remained at low but f i n i t e l e v e l of 0.1 - 0.2 wt. %, Figure (73). These r e s u l t s lead to the conclusion that at an 'FeO" a c t i v i t y corresponding to approximately 0.1 - 0.2 wt. % in t h i s slag (50/30/20), the reaction of A l 2 0 3 to give an increase i n the ingot w i l l take place i f calcium i s used as a deoxidant. When either A l or aluminum s i l i c o n was used as deoxidant at equivalent rates, i t was observed that the slag "FeO"concentration was again held at low l e v e l s — Figures (39, 40) and (45,46) and Tables XII and XIII (a-c) — a n d that both A l and Si were qu a n t i t a t i v e l y transferred to ... . . . , . ,(147, 148,210,216) the ingot. This r e s u l t i s to be expected ' ' from reaction (20-C) as only a very s l i g h t degree of alum-inum reduction of Ca from CaO would arise from reaction (20-C), producing composition changes not detectable within the accuracy of t h i s mass balance. The behavior of the i n -clusion compositions i s a more sensitive guide than the mass balance i n r e l a t i o n to the Ca/Al exchange reaction i n the slag. The Ca:Al r a t i o i n the oxide phase and the p a r t i a l substitution of Mn by Ca i n the s u l f i d e phase i n i n -clusions i n Al-deoxidized i n g o t s — F i g u r e s (43,44) and (51, 52)--as well as the Ca:Al r a t i o i n the oxide (inclusion) 196 phases and th e i r proportional amounts of CaS i n the CaSi and i n the A l - S i based deoxidized i n g o t s — F i g u r e s (58,59), (66,67), (71,75) and Figure (76)—indeed monitor the magni-tude and d i r e c t i o n of the equilibrium dictated by reactions (20-b,c). Figure (74) which shows the behavior of the 4340 ingot shows that the Ca:Al r a t i o i n the oxide i n -clusions r i s e s r a p i d l y to a constant value which i s approxi-mately equal to a composition CaD'A^O^. These r e s u l t s may be compared to those reported for calcium i n j e c t i o n (144 148 154 157) processes ' ' ' and for the basic e l e c t r i c arc steelmaking p r o c e s s ' 4 0 , 2 0 6 ^ p a r t i c u l a r l y those shown i n (178) Figure (78) where an equivalently low-sulfur s t e e l shows the same behavior. Faulring and Ramalingam* 2 1^ have studied the r e l a t i o n s beween CaO and A^O^ i n generating oxide inclusions i n t h i s system. Their conclusions are represented by the p r e c i p i t a t i o n (equilibrium) phase diagram shown i n Figure (11). When the Ca a c t i v i t y i s estimated i n the ESR slag/metal system following reaction (20-c), (234 ) (19 8) using the data of Rein et a l . and Sponseller and F l i n n and assuming that CaF 2 acts as an i n e r t diluent, an a c t i -— 8 v i t y of h = 6.7 x 10 at 0.1 wt.% A l i s obtained. This (_a conclusion would indicate from Faulring's data, Figure (11), that one should observe an i n c l u s i o n composition close to CaO-A^O^/ which i s indeed the case. It i s c l e a r , there-fore, that in spite of an excessively high addition of 197 calcium ('vlO kg/ton) , the in c l u s i o n composition i s controlled e n t i r e l y by reaction (20-c) and that i n c l u s i o n calcium contents w i l l not r i s e above those permitted by reaction (20-c), despite the calcium addition. It i s i n -teresting to note also that at th i s l e v e l of calcium ad-d i t i o n , the s u l f i d e inclusion surrounding the calcium a l -uminates were exclusively composed of CaS not MnS. This finding can also be compared to those reported i n (Ca,CaO) i n j e c t i o n processes in Al deoxidized melts of r e l a t i v e l y low oxygen a c t i v i t y , where i n c i p i e n t amounts of Ca induce the p r e c i p i t a t i o n of either MnS m ' ^ " ' ^ ^ o r 4 - n e pre-c i p i t a t i o n of aluminates (low i n CaO content) with a p e r i -l s , , u i /„ „ ^ ( 1 4 0 , 144, 146,153, 159, 160, pheral double s u l f i d e , (Mn,Ca)S ' 19 7 212) ' '. This t r a n s i t i o n i s obviously dictated by: MnS + [Ca] t CaS + [Mn] (25) and i t i s expected to occur when the Ca i n the melt i s ap-proximately 5-10 ppm. At higher lev e l s of Ca, the s u l -. N • . ,(140, 144, 146,147, 153,157,211,216) . f i d e phase i s expected • ' • ' ' ' • to be ruled by: * * CaO + [S] t CaS + [0] (21) and hence p r e c i p i t a t i n g a peripheral s u l f i d e , namely (Mn,Ca)S or pure CaS. In the case of deoxidation with the A l S i a l l o y , the observed changes in slag composition were almost equi-198 valent to the CaSi, the only change being a s l i g h t increase i n both SiC^ and Al^O^ at low "FeO" levels i n the slag, Table (XII). The observed i n c l u s i o n composition was that of Ca-aluminates containing approximately 20-30% CaO and about 5-7% S as CaS i n t h e i r periphery, Figure (76). These findings indicate that since the equilibrium p o s i t i o n of reaction (20-c) re s u l t s i n reduction of CaO by A l , the t o t a l mass of Ca ca r r i e d into the metal by t h i s means with the AISi a l l o y i s s i m i l a r to that i n the CaSi and the "hypercal" case, Figure (76). The equilibrium r a t i o s of Ca:Al reported by Faulring and Ramalingam^ 2 l^ r hence are attained by the CaSi, "hypercal" and the AISi a l l o y . The chemical r e s u l t of deoxidation by AISi and the "hypercal" a l l o y i n respect of ingot composition i s therefore v i r t u a l l y i d e n t i c a l with that obtained by conventional CaSi deoxidation at about the same rate. I t i s i n t e r e s t -ing to note that aluminum deoxidation follows a pattern f a m i l i a r from e l e c t r i c furnace prac t i c e . In the case of aluminum deoxidation at the present l e v e l s (10 kg/ton), a range of A l content i s achieved i n the ingot as shown i n Figure (71). The corresponding l i q u i d and s o l i d ingot oxygen analysis are shown i n Figure (72 a-b) where i t can be seen that the minimum oxygen content i s at approximately 0.1 wt% A l . This minimum i s at the composition expected, (175 177) Figure (10) ' , from a consideration of 2 [Al] + 3 [0] t (Alo0-,) . i . (11-ii) 1 J 1 J ' 2 3 i n c l u s i o n using the compiled information by Gustaffson and Melberg 199 (175) A l Figure (10), for e Q at an equilibrium temperature of ap-proximately 1600°C. It can therefore, be assumed that the deoxidation and p r e c i p i t a t i o n of inclusions are c l o s e l y equivalent to those established i n the e l e c t r i c f u r n a c e ' 4 0 , 2 0 6 ^ . 5.3.5 F i n a l Remarks It i s important to emphasize that an inadequate de-oxidation of a ESR-melt with A l or Ca-bearing deoxidants becomes very important where ever the aluminate-CaS i s not properly adjusted. The A l deoxidation or the lack of CaSi deoxidation seen as generators of alumina galaxies or as (133) inducers of "burning" or the excess of A l , A l S i or CaSi a l l o y s , because of the excess of A l i n ingots introduced through reaction (15 a-b) constitute a p o t e n t i a l source of s u l -fides and n i t r i d e s and hence to a degradation of mechanical (221-223) properties of the refined ingots. Hence, the f i n a l remarks, which at the same time answer the l a t t e r set of questions to be drawn are: i) ESR-deoxidation with Ca w i l l follow the reaction (20-c) at low oxygen po t e n t i a l s , i . e . slag "FeO" contents below 0.2 wt.%; i i ) At intermediate oxygen potentials (slag "FeO" content ranging between 0.4-0.6 wt,%), the deoxidation process using either A l or Ca i s equivalent and serves only to control the slag composition. This l a t t e r factor w i l l therefore de-termine the choice of deoxidant; i i i ) Deoxidation with A l w i l l also follow the reaction (20-c) but w i l l not reach the predicted ingot Ca:Al r a t i o due to unfavorable k i n e t i c factors, unless Si i s added as a c a r r i e r for Ca; iv) The maximum SiC^ content i n the slag for e f f i c i e n t deoxidation through reaction (12-v), (12-iv), and (20-c); i s less than 10 wt. %; v) The chemistry of the electrode does not play a rol e i n the deoxidation (reaction) scheme unless inappropriate (low) deoxidation rates are used, i . e . s a c r i f i c i a l e l e c t -rode deoxidation which leads to change the ESR slag or ingot composition; vi) Excessive and/or abrupt deoxidation can lead to deleterious mechanical properties, i . e . i n introduction of high l e v e l s of A l or deoxidizers (as small segregates) i n the ESR ingot; and v i i ) The shape of inclusions depends upon the type and degree of deoxidation: 1) i n A l deoxidized ingots spherical single or cl u s t e r s of alumina phases ( a - A l ^ ^ and FeO'A^O^) associated with manganese s u l f i d e s are found at r e l a t i v e l y low deoxidation l e v e l s and faceted aluminates and low calcium alum-inates (CaO-eA^O-j) as single or clust e r s at r e l a t i v e l y high deoxidation rates. 2) i n CaSi, "hypercal" or AISi deoxidized 201 ingots, faceted ( a - A l 2 0 3 and CaO6Al 20 3) and single spherical calcium aluminates (CaO*2Al 20 3, CaO«Al 20 3 and 12CaO«7Al 20 3) with peripheral s u l f i d e s are found at r e l a t i v e l y low and high deoxidation rates respectively. 202 5.4 Findings and Discussion Related to the Third Question 5.4.1 Description of Experimental Results 5.4.1.1 The Inclusion Mean Diameter Tables (XIV a-b) show the standard information drawn from the in c l u s i o n (EPMA) analysis. As previously described, a minimum of twenty and a maximum of 40 single analyses were performed i n each sample. Solid ingot samples 2.5 x 2.5 cm2 i n area and l i q u i d pool samples approximately 75% of this area were systematically analyzed. Analyses were performed i n longitudinal and transversal directions i n both types of samples. The mean inc l u s i o n diameter was s t a t i s t i c a l l y obtained. Figures (79 a,b) show an example of each type of sample. The l i n e a r i t y shown i n these graphs c l e a r l y indicates that the size of inclusions i s represented by a normal d i s t r i b u t i o n . The c o r r e l a t i o n c o e f f i c i e n t s ranged from 0.97 to 0.99 which are excellent for the p a r t i c l e size found. Based on thi s information plots of mean in c l u s i o n diameter, given as the 50% of the cumulative frequency, against ingot height (or deoxidation levels) were obtained. To correlate t h i s inform-ation to the deoxidation behavior, these findings are also plotted along with the t o t a l oxygen analysis from both types of samples. 203 5.4.1.2 Findings from Individual Experiments The Figure ( 36 ) which corresponds to ingot l a b e l l e d as RII-W i n which the CaSi was d i s c r e t e l y added, shows that the i n c l u s i o n average size depends strongly on the deoxidation practice. This p l o t c l e a r l y shows that coincidental to the largest Ca:Al r a t i o s i n inclusions the smallest mean i n -clusi o n s i z e i s found. The behavior corresponding to RII-Il i s shown i n Figure (41). This shows that the mean inc l u s i o n size varies with the t o t a l oxygen content. This graph also shows that at low deoxidation rates (3.60 and 6.1 kg ton ^ ) , the i n -clusion size of samples from the l i q u i d pool are smaller than the ones from the ingot. At moderate and high deoxidation l e v e l s , however, the opposite behavior i s observed. The d i f -ference i n average size i s approximately 1 pm. V a r i -ations i n the Ca:Al r a t i o s i n i n c l u s i o n s — F i g u r e ( 4 3 ) — a l s o r e f l e c t t h i s behavior. Findings from RII-I2 are shown i n Figure (42). Since t h i s ingot was deoxidized by lower Al-addition rates than R I I - I l the mean in c l u s i o n diameter i n samples from ingot and l i q u i d pool were almost equivalent. The average Ca:Al r a t i o s of inclusions i n t h i s ingot also changed i n an equivalent manner, Figure (44). The ingot i d e n t i f i e d as R I I I - I l , which was CaSi de-oxidized, i n some respect behaved as RII-I2. Variations in the Ca:Al r a t i o s i n inclusions and oxygen contents are also r e f l e c t e d in the mean size of inclusions. It i s important to note that the CaSi addition rates were almost equivalent from one to the other and gradually increased during r e f i n i n g , Figures (56 a-b) and (58). RIII-I2 used an equivalent deoxidation procedure to R I I I - I l . RIII-I2, however, was refined under d i f f e r e n t de-oxidation regimes. The three lowest CaSi additions were added i n shorter periods of time at the bottom of RIII-I2, while the fourth l e v e l of additions (22.4 4 kg ton ^ ) was much longer than i n R I I I - I l . In t h i s experiment, i t i s unmistakenly shown that inclusions from l i q u i d pool, at moderate and higher CaSi additions, are larger i n size than the ones from the ingot. The Ca:Al r a t i o s , the t o t a l oxygen content as well as the mean i n c l u s i o n diameters behaved i n exactly the same way as R I I - I l , Figures 57(a-b) and (59). F i n a l l y , Ingot 4340 which was CaSi deoxidized shows the same pattern, i n terms of t o t a l oxygen analysis, Ca:Al r a t i o s and p a r t i c l e sizes, as RII-I2 and R I I I - I l , Figures (72 a,b) and (74a,b). 205 5.4.1.3 Complementary Studies To gain a better understanding about i n c l u s i o n form-ation and growth and to elucidate whether the f l o t a t i o n mechanism operates under the ESR-conditions one more set of experiments was performed. Samples were sucked from the l i q u i d pool into s i l i c a tubes which contained either a mishmetal or Zr wire. Refining was carried out under argon and at a constant deoxidation rate (10 kg/ton of either AISi or CaSi). Three major areas, i n terms of the composition of inclusions were i d e n t i f i e d i n these samples. Regions where inclusions were mainly constituted by either rare earth or zirconium enriched phases. A region where the Zr or the rare earth phases were mixed with complex C a - A l - s i l i c a t e s and a region where a mixture of pure Ca-aluminates and very few inclusions with peripheral rare earth or zirconium s u l f i d e phases were i d e n t i f i e d . This l a t t e r type of i n -clusions i s shown i n Figures (80,81) and (82,83) where the composition maps for A l , Ca,Ce, La, S and Zr are given. It i s very important to emphasize that t h i s type accounted for as much as 7-10% of the t o t a l amount of inclusions analyzed (40-50) i n each sample. Metallographic analysis on specimens obtained from l i q u i d pools (1020 and 4340)show that a considerable amount 206 of inclusions was found close to the wall of the s i l i c a tube. Most of them, however, were located in i n t e r d e n d r i t i c regions and only approximately 5-7% were trapped by primary dendrites, Figures (85) and (86). 207 5.4.1.4 Summary of Experimental Findings In order to approach the answer (discussion ) to t h i s question a summary of findings, in terms of 1) i n c l u s i o n mean siz e , 2) i n c l u s i o n composition and 3) t o t a l oxygen analysis, i s presented. 1) - Inclusion size d i s t r i b u t i o n s obey the normal d i s t r i -bution (the c o r r e l a t i o n c o e f f i c i e n t ranged from 0.96 - 0.99) . The mean i n c l u s i o n diameter was approximately 6-8 ym. The average incl u s i o n size increases approximately 1 ym i n diameter i n both types of samples (ingot and l i q u i d pool) as the deoxidation rate i s i n -creased. In ingot heads (CaSi deoxidized) some inclusions were seen as large as 30 ym i n samples extracted from l i q u i d pools and t h i s size was very r a r e l y seen i n the ingot. The i n c l u s i o n density, number of inclusions per area, i s gradually increased from the center to the mould wall. This fa c t was accentuated i n CaSi deoxidized ingots. 2) - Calcium aluminates were almost always observed to contain a core enriched i n alumina ( i . e . A l by EPMA). 208 The Ca:Al r a t i o s i n inclusions i s almost always smaller i n samples extracted from l i q u i d pools than from ingots The peripheral S content as CaS i s proportional to the Ca:Al r a t i o s in the calcium enriched aluminate phases. At r e l a t i v e l y large deoxidation rates (^10-12 kg/ton) the formation of segregates enriched i n deoxidizers (Al, Ca and Si) was observed. It i s important to r e c a l l that these segregates were commonly seen i n samples extracted from the l i q u i d pool and very r a r e l y i n ingots. They also contained some Mn. 3) - The average difference i n t o t a l oxygen content between samples extracted from l i q u i d pool and ingot i s approxi-mately 20 ppm. 5.4.2 P r e c i p i t a t i o n of Inclusions i n the Fe-Al-Ca-O-S(Mn) System By using Faulring's and Ramalingham's d a t a ^ 2 1 ^ f given i n Tables (IV) to (VI), to construct the Fe-Al-Ca-0 p r e c i p i -t a t i o n diagram, conjointly with observations of other i n v e s t i -gators, previously described i n sections 2.3.5, 2.3.7.2 and 5.3.4, the p r e d i c t i o n of the p r e c i p i t a t i o n of the Ca-aluminate/ Ca-sulfide phases can be pursued a l i t t l e further. If the Fe-Al-Ca-0 system i s now approached by superimposing the (CaS) + [0] t CaO + [S] (21) equilibrium to the Fe-Al-Ca-0 system, as the Fe-Al-Ca-O-S system, and care i s taken to consider the phase r u l e , a diagram representing the p r e c i p i t a t i o n of Ca-aluminates and t h e i r corresponding s u l f i d e phases was constructed. The procedure used was equivalent to that used by Wilson et a l . ( 2 3 9^ and Faulring et a l . . The a c t i v i t y co-e f f i c i e n t s for CaS i n equilibrium with l i q u i d [CaO-Al 20 (liquid) - CaS(liquid)] and [CaO-CaS]solid were estimated N (240 241) from Sharma's and Richardsons 1s investigations ' , Table (XV). This diagram based on Henrian a c t i v i t i e s i s isothermal (1550° C) and was developed for l e v e l s of A l ( h A l = 0.001, 0.01 and 0.1) i n the range of major in t e r e s t . The data generated from these e q u i l i b r i a i s summarized i n Table (XV). The major reactions are given i n Appendix (I). Although the Fe-Al-Ca-O-S system was exclusively de-veloped as a four f o l d component e q u i l i b r i a , the MnS i n equilibrium with A^O^ and the double [(Ca,Mn)S] s u l f i d e i s • 4- • • n -A -,(140,146,210) , , . i n t r i n s i c a l l y considered ••' . The gradual s u b s t i t -ution of Mn by Ca i n the MnS phase i s governed by the f o l -lowing equilibrium: (MnS)*+ [Ca] t (CaS)*+ [Mn ] (25) The presence of Ca i n solution i n the melt generates several t r a n s i t i o n s . Incipient amounts of Ca originate the 210 (91 92 140 144) t r a n s i t i o n of MnS II to MnS Relatively higher Ca contents i n the melt induce the double s u l f i d e s , e.g. (Ca,Mn)S. If the amount of Ca i n t r o -duced i n the melt i s increased a m i s c i b i l i t y gap between the CaS and MnS appears' 4* 5^. This t r a n s i t i o n occurs somewhere between the formation of the CaO ' e A^O^ and the CaO^A^O^. The s u l f i d e phase, as reported by several researchers i s heterogeneously p r e c i p i t a t e d on the oxide phases (previously described) and i t i s very dependent on the a c t i v i t y of the A l . In previous discussions i t was acknowledged that as the Ca(Al) deoxidation l e v e l i s increased, the amount of Ca as CaO i n the Ca-aluminate phase i s increased by reaction (20) and hence the amount of CaS phase increases by reactions (25) and/or (26). The fact that the i n j e c t i o n of CaO slags do not produce pure CaS i s also considered. As Saxena and cowork-(147 148 214) ers •' ' have reported, t h i s t r a n s i t i o n w i l l take place (144 153 a f t e r Ca-aluminates have formed. As several researchers ' ' 154,156 .185,197.211) . _ . . . . . . _ _ • • • ' work suggest, the p r e c i p i t a t i o n of pure CaS i s expected to occur once the Ca-content i n the melt (or the oxygen l e v e l i s approximately 10-40 ppm) i s such that a comp-o s i t i o n i n the double s u l f i d e , (Mn,Ca)S i s greater than 4 3.0 to 50.0 wt. % (140,146)^ This t r a n s i t i o n i s reached once the (202) CaO-A^O^ stoichiometric phase i s formed . Ototani's and Kataura's r e s u l t s * 2 1 5 ^ confirm Kiesslings and Westman 1s* 1 4^^ (140) (159) Salter and P i c k e r i n g ' s v / and Church's ' findings. They report a "pure" CaS phase after the " A l 2 0 3 " content i n the Ca-aluminates i s reduced by Ca to approximately 40.0%. It i s also r e p o r t e d ' 1 ^ that once t h i s percent of CaO (40.0%) i s reached a sharp increment in the CaS i s noted. A schematic representation of the above description i s shown i n Figure (87). This Figure (87) i s necessarily i s o -thermal (1823K) and at a f i x e d A l a c t i v i t y , h f t l = 0.1. To represent the s t a b i l i t y of these composite phases over the ranges of i n t e r e s t (h g = h A l = 0.001, 0.01 and 0.1) i n a com-prehensive manner the h ^ t h ^ r a t i o s are plotted against either h A ^ or hg i n logarithmic scales i n Figures (88) and (89). These Figures (88,89) summarize the behavior of the A^O^/ MnS I I - I I I , the Ca-aluminates (C•6A,C•2A)/(Mn,Ca)S and the CaO-Al 20 3(liq) and CaO/CaS e q u i l i b r i a . 5.4.3 Discussion of Results 5.4.3.1 Nucleation, Growth and F l o t a t i o n o f I n c l u s i o n s The difference in the mean diameter (1-3 ym) of inclusions from l i q u i d pool and from ingots indicates that the f l o t a t i o n mechanism i n the ESR-conditions operate. The conditions under which th i s behavior was displayed were: a) when the difference i n t o t a l oxygen content between the l i q u i d pool and ingots i s greater than 20 ppm and b) Where smooth, gradual and equivalent changes i n the Ca:Al r a t i o s of i n -clusions and i n the t o t a l oxygen content i n samples from l i q u i d pool and ingot were observed. Case (a) was speci-212 f i c a l l y observed i n the Al-deoxidized ingots, i . e . R I I - I l and p a r t i a l l y i n RII-I2. These two ingots exhibit t h i s d i f -ference in oxygen content when the highest deoxidation rate with A l was applied. I t i s important to note that because of the changes in oxygen analysis i n the l i q u i d pool and ingot (RII-I2), a difference greater than 20 ppm i s observed i n t o t a l oxygen. This behavior i s also r e f l e c t e d i n the Ca:Al r a t i o s i n inclusions. Case (b) i s found i n RIII-12 and p a r t i a l l y i n RII-Il where the deoxidation sequence, given by the behavior of the t o t a l oxygen content and the Ca: A l r a t i o s i n inclusions, Figures (57,59) and (43,44), shows smooth and p a r a l l e l changes. The s o l i d i f i c a t i o n conditions i n samples from ingots (center part) are less d r a s t i c than in samples extracted from the l i q u i d pool. The secondary dendrite arm spacing (DAS 1 1 ) i n ingots where the samples were obtained was about 250-300 ym whilst i n samples from the l i q u i d pool 30-50 ym. A larger D A S 1 1 , as acknowledged i n the l i t e r a t u r e ( " ' 1 0 1 ' 1 0 3 - 1 0 6 > provides more advantageous conditions for the nucleation and growth of inclusions. In spite o f . t h i s inclusions i n samples extracted from the l i q u i d pool, under the described conditions, showed a larger mean diameter. It i s also important to note that the difference i n i n c l u s i o n mean size was 1 ym, above where the faceted (a-A^O^) a l -213 umina was observed. At lower concentrations of aluminum round iron aluminates (FeO'A^O^) were i d e n t i f i e d whereas at higher deoxidation rates the aluminates with some calcium and some angular alumina were i d e n t i f i e d . These findings (99 183 strongly agree with Turpin's and E l l i o t ' s and others ' ' 184) (99) observations. These researchers who have studied the nucleation phenomenon under sub-cooled conditions, have suggested that the angular alumina phase was nucleated i n the melt at equilibrium temperature and i t simply grew at sub^-liquidus temperatures. Coincidentally to t h i s ob-(99) servation , i t was also reported that i n very early stages of sub-cooling a scum was formed on the surface of t h e i r melts. Thus, in d i c a t i n g that at the beginning of sub-cooling some inclusions simply floated to the surface. To determine the extent at which the f l o t a t i o n mechanism i s allowed i n the ESR-liquid pool a more elaborate (SEM and EPMA) study, through the extracted samples containing either RE or Zr, was c a r r i e d out. The peripheral RE and Zr as o x i -s u l f i d e s enclosing the Ca-aluminates have c l e a r l y revealed that the l a t t e r phases were already present i n the l i q u i d pool. These experiments also suggest that inclusions are i n a l i q u i d - s o l i d s t a t e ( 9 2 ' 9 4 ' l 2 l ) , Figures (80) to (84). These r e s u l t s which are also i n agreement with the metallography observations, show that the f l o t a t i o n of inclusions can occur 214 i n as much as 7-10% out of the t o t a l i n c l u s i o n content i n the ingot. These r e s u l t s , however, do not rep-resent the amount of inclusions removed from the s o l i d i f y i n g ingot. I f an analysis of Figures (79a) and (79b) i s made one can see that i n order to account for the di f f e r e n c e i n size (1.0 - 1.5 ym i n diameter) a displacement i n the cumul-ative frequency from 50 to 70% produces the expected d i f f e r -ence. This indicates that an elimination of i n c l u s i o n s of approximately 20% i s c a r r i e d out by f l o t a t i o n . Although these experiments do not c l e a r l y r e v e a l the nature of saturation, i t i s believed that i t i s reached by the three mechanisms' 0 0^ namely by cooling, additions of de-oxi d i z e r s and during s o l i d i f i c a t i o n . The highest degrees of deoxidation which r e s u l t from the introduction of large amounts of A l in the melt eith e r from the deoxidizer or through rea c t i o n (20a-c) and the high c r y s t a l l i n e character of the alumina phase lead to the b e l i e f that t h i s phase i s uniformly nucleated i n early stages of undercooling at the beginning of s o l i d i f i c a t i o n . The presence of segregates enriched i n deoxidizers i n some samples from l i q u i d pool and ingot also suggest that " l o c a l supersaturation" (by additions) can be achieved. It i s i n f e r r e d that i f growth of inclusions r e s u l t s from a mechanism other than d i f f u s i o n and p r e c i p i t a t i o n , s p e c i -f i c a l l y growth due to c o l l i s i o n coalescence between i n -clusions of d i f f e r e n t sizes and there i s s u b s t a n t i a l con-215 vective mixing i n the ingot pool then t h i s phenomena should be r e f l e c t e d i n inclusions i n the ESR-ingot. The s i z e , shape and arrangement of inclusions extracted by the iodine-methyl acetate-methanol method from A l deoxidized ingots i s shown i n Figures (91) and (92). These (SEM) photographs reveal that the growth by c o l l i s i o n and coalescence of A^O^ and CaC"6Al2C>2 i n the l i q u i d pool indeed has taken place. 4. A u i u (38, 121,242,243) As suggested by several researchers ' , since there i s not a complete assimilation of inclusions by the slag some inclusions are c a r r i e d back into the s o l i d i f y i n g ingot. The type of inclusions, c l u s t e r s of aluminates, i d e n t i f i e d i n t h i s research very much resemble those reported i n conventional mechanically, thermally, or electromagneti-, , . . , (38,120,183, 189,193-195) . . . . c a l l y s t i r r e d melts ' ' ' ' . Thus, although the growth of inclusions i n ESR-ingots can be accounted for by the simultaneous d i f f u s i o n - p r e c i p i t a t i o n mechanism, the difference i n size found in l i q u i d pool and ingot cannot be explained by a mechanism other than the f l o t a t i o n . It i s also r e a l i s t i c to suggest that the i n c l u s i o n size d i s t r i b u t i o n seen by the s o l i d i f y i n g interface i s not s t r i c t l y controlled simply by buoyance considerations but instead by o v e r a l l (244 ) bath hydrodynamics as suggested by Engh and Lmskog (245) and Linder . From t h i s discussion, i t i s evident that the i n c l u s i o n f l o t a t i o n mechanism cannot be approached by 216 st r a i g h t Stokes 1 Law unless the appropriate corrections to this equation (13) are considered, Table (I). The second important conclusion from these r e s u l t s i s that at deoxidation rates which produce an A l content of 0.1 - 0.15 wt. % i n (ESR) ingots, inclusions are exclusively nucleated and grown i n i n t e r d e n d r i t i c spaces during s o l i d i -f i c a t i o n . These findings i n agreement with A l deoxidized 4 . • 4 . - n 4 . i , • (90, 172,183-187) ingots i n conventional steelmakmg practice indicate that the nucleation phenomenon at these deoxidation leve l s i s c o n t r o l l e d by the formation of the FeO«Al 20 3 i n -clu s i o n phase, e.g. l o c a l supersaturation during s o l i d i -f i c a t i o n . Recent s t u d i e s * 2 4 6 ^ i n u n i d i r e c t i o n a l l y s o l i d i f i e d ingots deoxidized with 0.1 and 2.0 % A l which c l o s e l y re-semble the Al(ESR) deoxidized ingots, have shown that the FeO«Al 20 3 phase i s p r e c i p i t a t e d at a s o l i d f r a c t i o n of 0.65 i n the low A l (0.1%) content and at 0.89 i n the other. These r e s u l t s suggest that the FeO«Al 20 3 i s the i n c l u s i o n f i r s t phase pr e c i p i t a t e d and therefore as suggested by other re-s e a r c h e r s ( 1 8 3 , 1 8 4 , 1 8 6 _ 1 8 8 ) t h i s phase i s transformed to a l -umina as the s o l i d i f i c a t i o n proceeds. 217 5.4.3.2 Comparison Between Theoretical and Experimental  Results The t o t a l oxygen content of samples from the l i q u i d pool and ingot under an appropriate deoxidation sequence, Figures (41) and (57), have c l e a r l y shown that there i s a difference between the samples of approximately 20 ppm. This (63) behavior i s expected from an equilibrium s i t u a t i o n i n Figure (10). The average Ca:Al r a t i o s in inclusions from both types of samples ( l i q u i d pool and ingot) also indicate that under a gradual deoxidation sequence with either an A l S i , CaSi or hypercal a l l o y the calcium which remains i n solution p r e c i p i t a t e s on calcium aluminates during s o l i d i -f i c a t i o n . This, acts to rai s e the Ca:Al r a t i o s by allowing the reaction (CaO)*+ [S] X (CaS)*+ [0] (18) , , , ,. ^. (147,148) ^ . to take place i n the forward d i r e c t i o n ' . It i s imp-ortant to emphasize that despite the 20 ppm of oxygen i n solu t i o n i n the l i q u i d pool, the equilibrium i s achieved i n the d i r e c t i o n indicated by the reaction (18). As the temperature decreases the calcium, oxygen and sulfur.pre-c i p i t a t e as peripheral oxide and s u l f i d e , i . e . CaO and CaS. The t r a n s i t i o n of these phases i s presented i n Figures (92 * a-c) where the CaS/CaO interphase i s c l e a r l y revealed. This behavior i s observed when the l e v e l of "FeO" i n the slag i s such that the calcium aluminates are either CaO«2Al nO_ 218 or 12CaO«7Al 0_. At higher levels of "FeO" where the Al o0_ 2 3 3 2 3 i s transformed to CaO-GA^O^, the reaction which controls the p r e c i p i t a t i o n of s u l f i d e i s : MnS + [Ca] t CaS + [Mn] (19) This reaction enables the p r e c i p i t a t i o n of double s u l f i d e s , i . e . (Ca,Mn)S, Figures (51) and (52). The most important fa c t to point out i s that the sec-ondary p r e c i p i t a t i o n which i s heterogeneous i n nature i s s t r i c t l y c o ntrolled by the Ca:Al r a t i o i n the i n c l u s i o n phases. The e f f e c t of the temperature on the p r e c i p i t a t i o n sequence, p a r t i c u l a r l y i n the Ca:Al r a t i o s where aluminates enriched i n CaO are i n equilibrium i s equivalent to an increment i n the aluminum content i n the melt i n the Ca-Al-0 system, Figure (11) This behavior i s also expected from the Ca0-Al 20.j pseudo binary equilibrium diagram. This indicates that as the CaO content increases in the aluminates their s t a b i l i t y i n terms of temperature decreases and a more stable compound i s formed. The completion of the s u l f i d e p r e c i p i t a t i o n reactions (18) and (19) i s e x p e c t e d ' 4 ^ to occur at approximately 1000°C where the m i s c i b i l i t y gap i n the MnS-CaS binary diagram d i s -appears. I t i s also important to mention that during s o l i d i -f i c a t i o n some iron or Cr can be i n solution with the . (92, 140,159) s u l f i d e phase ' ' Another point to be considered i n t h i s analysis i s that where the aluminate phase (A^O^ or CaO'GA^O^) i s stable the s u l f i d e phase i n equilibrium with i t i s only the MnS i n 219 any of i t s shapes, i . e . MnS I, II or III which are also dependent on the chemistry of the melt. The o v e r a l l be-havior of the aluminate-sulfide t r a n s i t i o n i s condensed i n F i g -ures (88,89) which are an extension of results shown in Figure (87). Figure (87) also indicates that the p r e c i p i t a t i o n of "pure" CaS cannot occur unless lower oxygen potentials, than those required to p r e c i p i t a t e CaOiA^O^ and 12CaO • 7AI2O.J are reached. If these findings are compared against studies on Ca-i n j e c t i o n processes then i t can be seen that the simultane-ous deoxidation-desulfurization mechanism i s also r e f l e c t e d • 4.1. ^  4.- J 4. (140,147,148,153,197, 211,214) in the deoxidation products ' ' ' ' ' . (14 7 This diagram i n agreement with Saxena et al.'s work ' 148 214) ' c l e a r l y reveals that Ca either as a CaSi or as CaO does not d i r e c t l y contribute to the CaS p r e c i p i t a t e i n inclusions unless the alumina i s f i r s t transformed into Ca-aluminates. While calcium aluminates with peripheral Zr or RE oxides s u l f i d e s , i n the samples extracted from the l i q u i d pool by the s i l i c a tube containing either Zr or mischmetal, were not commonly found, a l l of the inclusions which con-tained these elements (Zr or Ce and La) homogeneously d i s -tributed were p r e f e r e n t i a l l y composed of phases enriched i n calcium. Inclusions generally show that among the elements 220 traced by the X-ray spectrum (SEM) analysis (Al, Zr, Ca and Si as oxide-sulfide and almost pure CaS) calcium i s one of the main constituents of these phases. This finding i s taken as one more evidence that calcium, due to i t s low solu-(198 199) b i l i t y i n iron ' . i s gradually rejected and hence i t contributes to increase the Ca:Al r a t i o i n inclusions, by reaction (18) as the s o l i d i f i c a t i o n progresses. In previous discussion of r e s u l t s (section 5.4.1.4, 5.4.2 and 5.4.3.1), i t was established that the p r e c i p i -t a t i o n of inclusions i s ruled by the reaction scheme (12-iv), (12-v), and (20-C). Reaction (20-C) and (19) dictate the chemistry of the oxide phase, i . e . the gradual t r a n s i t i o n of A1 20 3 to 12CaO«7Al 20 3. The reactions (21) and (25), which occur simultaneously to the above ones determine the s u l f i d e t r a n s i t i o n , i . e . MnS II MnS III (Ca,Mn)S -»-CaS. Figures (87) and (88) reveal t h i s behavior. Results given i n Tables (XVI) and (XVII) to a certain extent des-cribe these r e s u l t s . The equilibrium, isothermal calculations shown i n Table (XVI) were performed by assuming constant ( f i r s t order) i n t e r a c t i o n c o e f f i c i e n t s for Al-O, Ca-0 and Ca-S. The values used were - 5 . 2 5 ( 1 7 5 ) , - 6 2 . 0 ( 1 7 5 ) and -40 respectively. The l a s t value for the Ca-S was assumed on the basis that since the free energy of the CaO i s approximately 1 1/3 larger than that for CaS, Table (XV), then the r a t i o of t h e i r f i r s t order i n t e r a c t i o n parameters could be equivalent. Hence, i f S-analyses from the 1020 M.S. electrode are taken as the s t a r t i n g point i n early (low) deoxidation stages and they are compared against those given i n Table (XVI) for the Al 20 3-CaO«6Al 20 3-CaS equilibrium, i t i s noted that i n terms of s u l f u r , very good agreement i s observed. If these values are compared i n terms of Ca- and S- contents and t h e i r respective oxide phases against those reported (144) by H i l t y and Popp given i n Figure (78), good agreement i s found. On the other hand, predicted values for oxygen are overestimated, Figure (72). Since, Gustafsson and Melberg' 7^? suggest the use of variable f i r s t order i n t e r a c t i o n c o e f f i c i e n t s then a second set of calculations was performed. Under these conditions, r e s u l t s shown i n Table (XVII) were computed. The f i r s t order i n t e r a c t i o n c o e f f i c i e n t s w e r e ; - 5 3 5 " ^ , -400, -300, -350 and -200 for the Ca-O, -62.0 ( 1 7 5 ) for Al-0 and -110 ( 2 1 1 ) for Ca-S, for the Al 20 3-CaO•6Al 20 3~CaS, CaO•6Al 20 3~Ca0•2Al 20 3 CaS, CaO«2Al 20 3-CaO«Al 20 3-CaS, CaO'Al^-CaO + A l ^ - C a S and Ca0^ sy - CaO + Al 20 3~CaS e q u i l i b r i a , Figure (87) and Table (XV). While the Ca, A l and 0 are predictable by f o l -lowing t h i s approach the sulfur i s not. The sulfur content i s underestimated (10 6 - 10 ^ wt.%). Regarding the independence of reaction (20-C) on the 222 oxygen pote n t i a l t h i s i s c l e a r l y revealed through these ca l c u l a t i o n s . Table (XVII). The oxygen content i n the melt ranges between 10-40 ppm regardless of the amount of A l and Ca i n solution. I t i s also important to note that the predictions stated i n section 5.3.4 i n terms of the expected in c l u s i o n compos-i t i o n which were based on Faulrings et a l . ' s ^ 2 l 6 ^ findings are also i n agreement with the r e s u l t s presented i n Figure (88) . F i n a l l y , two important facts are worthwhile to mention: F i r s t , the Ca-0 int e r a c t i o n parameters are very important i n the Al-Ca-O-S p r e c i p i t a t i o n and second a confident pred i c t i o n of the p r e c i p i t a t i o n sequence cannot be f u l l y r e l i a b l e unless the i n t e r a c t i o n ( f i r s t and second order) c o e f f i c i e n t s are appropriately determined. 223 CHAPTER VI THE RADIAL DISTRIBUTION OF INCLUSIONS IN CaSi AND Al DE-OXIDIZED INGOTS 6.1 Experimental Details and Techniques The l a s t part of t h i s i n v e s t i g a t i o n was undertaken to elucidate how inclusions are d i s t r i b u t e d , i n terms of s t a t i s -t i c a l l y determined sizes, i n ingots. For t h i s purpose a series of samples from several 1020 M. S. and one 4340 ESR ingots deoxidized with A l and the CaSi a l l o y , were r a d i a l l y s l i c e d at known deoxidation l e v e l s and samples were polished and s l i g h t l y etched on four and sometimes f i v e of t h e i r faces. Prior to performing measurements, microprobe analysis and X-ray spectrum analysis were car-r i e d out to i d e n t i f y the major i n c l u s i o n phases. Measurements of mean diameter of inclusions or second-ary dendrite arm spacing were performed almost invariably at every half centimeter on each face. Since inclusions i n A l deoxidized ingots were very small and complex i n shape (alumina galaxies associated with manganese s u l f i d e s ) , several approaches to determine the mean size of inclusions were followed, i . e . normal d i s t r i b u t i o n s and "the f i v e largest inclusions technique." 224 6.2 Experimental Findings Results of t h i s research are shown in Figures (9 3) to (95) and (96) to (99) for secondary dendrite arm spacing and i n c l u s i o n mean sizes respectively. The A l deoxidized ingots, as previously described showed alumina galaxies and manganese s u l f i d e . Only traces of calcium were found. The morphology of inclusions i n the 1020 M.S. A l -deoxidized ingots varied from globular single and double phase (spherical aluminates and manganese sulfide) at the ingot core, to elongated with double phase (aluminates and manganese sulfide) at midradius and almost exclusively small alumina galaxies associated with manganese s u l f i d e at the mould wa l l . The 1020 M.S. CaSi deoxidized ingots showed calcium a l -uminates and single and double calcium s u l f i d e s , i . e . , CaS and (Ca,Mn)S. The deoxidation l e v e l was such that calcium aluminates of the type CaO«6Al 20 3 and CaO•2Al 20 3 were found as the p r i n c i p a l i n c l u s i o n phases. The s u l f i d e phase en-closed the oxide phase. The 4340 CaSi deoxidized ingot showed v i r t u a l l y the same types of inclu s i o n phases as the 1020 M.S. ingots deoxidized with the CaSi a l l o y . I t i s im-portant to mention that the amount of the peripheral CaS phase was proportional to the Ca:Al r a t i o i n the oxide phase and i t was not necessarily dependent on the size of the i n -clusions. F i n a l l y , the i n c l u s i o n density (number of i n -elusions/per unit area) i n a l l of the ingots was increased considerably with the radius. 6.3 Discussion of Results In the discussion of the previous r e s u l t s , i t has been emphasized that some f l o t a t i o n of inclusions (about 10-20%) w i l l take place. This was determined to originate during s o l i d i f i c a t i o n i n early stages of cooling. This statement, however, i s true only when the required supersaturation r a t i o for the p r e c i p i t a t i o n of aluminates i s achieved, i . e . above 0.1 to 0.15 wt. % A l i n the ingot . The previous discussion has also established that i n -clusions i n samples from l i q u i d pool were larger i n d i a -meter (under an appropriate deoxidation sequence) than i n ingots although the s o l i d i f i c a t i o n conditions (given by t h e i r corresponding secondary DAS) were more d r a s t i c i n the former samples. Consequently, i t was concluded that the i n c l u s i o n growth was almost e n t i r e l y controlled by the d i f -f u s i o n - p r e c i p i t a t i o n of solutes on prenucleated phases. On the other hand, r e s u l t s obtained from the r a d i a l i n -cl u s i o n size d i s t r i b u t i o n s show that- the p r e c i p i t a t i o n i n ingots i s e s s e n t i a l l y c a r r i e d out during s o l i d i f i c a t i o n . Hence, i t i s very dependent on the l o c a l thermodynamic ,... (99,100,101,104,10 5,10 6,121,143,184,186,189) conditions ' ' ' ' ' ' • ' ' • Since the i n c l u s i o n mean size ranges r a d i a l l y from 6.0 -5.0 ym i n the ingot centreline to 3.0 - 2.0 ym i n the mould wall and the secondary DAS equivalently varies from 20 0-250 to 60-100 ym then the p r e c i p i t a t i o n of inclusions, p a r t i -c u l a r l y i n ingots deoxidized with the CaSi a l l o y i s almost e n t i r e l y controlled by the r e j e c t i o n of s o l u t e s ' 0 1 , 1 0 2 - 1 0 6 ' 121 243) ' , e.g. by the l o c a l thermodynamic conditions of the i n t e r d e n d r i t i c micropools formed as the s o l i d i f i c a t i o n pro-ceeds . Thus, although the l o c a l s o l i d i f i c a t i o n time given b y ' 4 7 ^ : DAS X 1 * = a t£ = b(GR)" n = 707.946 ( G R ) " 0 ' 3 8 2 5 (26) (65 225) i s very short, about 5-30 seconds ' i n locations close (1-1.5 cm) to the ESR-mould wall; t h i s i s s u f f i c i e n t to grow inclusions on prenucleated phases^" ' 1 0 0 ' 1 0 6 ' 1 1 5 ^ , by the d i f f u s i o n - p r e c i p i t a t i o n mechanism up to even larger * DAS 1 1 = secondary dendrite arm spacing, i n ym a and b = constants n = exponent which ranges from 1/2 to 1/3 t^ = l o c a l s o l i d i f i c a t i o n time, i n seconds GR = product of qrowth rate by the thermal gradient, i n °C/min. s i z e s ' 2 1 ^ . Thus, the c o n t r o l l i n g steps are either the nucleation of inclusions where the supersaturation r a t i o . - . , - , . ( 9 9 , 1 0 0 , 1 0 1 ) i s not reached i n early stages of cooling or the r e j e c t i o n of solutes (oxygen and deoxidizers) which are o ^ o.u -i • ^ • • ( 1 0 2 - 1 0 6 , gradually b u i l d i n g up as the s o l i d i f i c a t i o n progresses 1 2 1 , 1 8 5 , 1 2 1 , 2 2 7 , 2 4 3 ) I t i s also important to emphasize that the growth of i n -clusions during cooling i s very dependent on the density J • * ^ • • -i • i, 4- ( 9 1 , 1 0 1 , 1 2 1 ) T + T and size of the o r i g i n a l phases present . I f 5 7 these parameters are 1 0 - 1 0 inclusions/cm 2 and 1 - 1 0 um i n radius, as expected i n ESR-melts, only a very s l i g h t growth should be e x p e c t e d ( 1 0 1 ' 1 2 1 , 2 4 3 ) , Table (XVIII). It can also be elucidated that since "oxygen segregation" does not take place along r a d i a l d i r e c t i o n s i n i n d u s t r i a l size i n g o t s ' 4 8 2 5 ° ) f i t i s expected to observe a gradual change i n the i n c l u s i o n density, a f a c t which agrees with the observations i n t h i s research. Hence the i n c l u s i o n r a d i a l size d i s t r i b u t i o n should be inversely proportional to the i n c l u s i o n density. 228 CHAPTER VII * 7.0 Conclusions 7.1 Inclusions from the electrode are p h y s i c a l l y and chemi-c a l l y transformed i n the electrode t i p by the thermal gradients. Inclusions are chemically altered by the presence of the l i q u i d slag at the l i q u i d f i l m and they are almost e n t i r e l y removed (by d i s s o l u t i o n -reactions) when the droplet i s completely formed. Thus, electrode inclusions only play a r o l e i n the f i n a l ESR ingot insofar as t h e i r solution product enters into the slag-metal reactions experienced during processing. 7.2 Inclusions i n ESR-ingots are more strongly influenced by the deoxidation practice than by the electrode composition and/or the slag system used i n low SiC^ slags. I t i s important to comment, however, that under low deoxidation rates the i n t r i n s i c chemistry of the slag predominates, i . e . at high slag oxygen po t e n t i a l s . 7.3 As a consequence of the above conclusion, i t has been confirmed that the p r e c i p i t a t i o n of complex Al-Ca-s i l i c a t e inclusions i s predictable i n high s i l i c a slags where t h e i r o r i g i n i s s t r i c t l y given by the slag chemistry, i . e . i f wt. % SiO„ > 10.0. For ease of reference, the nomenclature of reactions i n previous texts are rewritten i n t h i s chapter. 229 7.4 The most appropriate slag system i n which to perform an e f f i c i e n t deoxidation i s the CaF2~CaO-Al203 system at 50, 30 and 20 wt. % respectively, i . e . the highest Ca:Al r a t i o i n inclusions i n the absence of deoxidation. 7.5 The inc l u s i o n chemistry expected, from the most common slag system (CaF 2-CaO-Al 20 3) used i n the ESR-process, i s controlled by the following e q u i l i b r i a : 2[A1] + 3(FeO) t (CaO) + Fe (7.5.i) [Ca] + (FeO) t ( A ^ O ^ + 3Fe (7.5.ii) (A1 20 3) + 3[Ca] t 3(CaO) + 2[Al] ( 7 . 5 . i i i ) This reaction scheme depends on the type and degree of deoxidation. Hence the p r e c i p i t a t i o n of i n -clusions i s dictated by: mCaO + n ( A l 2 0 3 ) J mCaO-nA^O.^ (7.5.iv) or more appropriately by: X[Ca] + Y(Alo0-.) ? 'XCaO'(Y - *-) Al o0_ + \ X[A1] *• J i n c l u s i o n J A J J (7.5.v) 7.6 From the above e q u i l i b r i a , i t can be seen that i f an excess of deoxidizer i s added to the slag (in the forward d i r e c t i o n of reactions 7 . 5 . i i i or 7.5.v) 230 undesirable composition w i l l r e s u l t i n the ingot, i . e . high Al contents, i n which case there exists a pote n t i a l problem of p r e c i p i t a t i n g Al n i t r i d e s or s u l f i d e s . 7.7 Deoxidation of slags during r e f i n i n g by using A l S i , CaSi, CaAlSi and CaSiBaAl alloys i s more e f f i c i e n t than by using Al p e l l e t s alone. The s i l i c o n i n a l l of the above alloys under the appropriate slag system (7.3), acts exclusively as a c a r r i e r of the deoxidant into the metal l i q u i d pool. 7.8 A l and A l S i a l l o y are e f f i c i e n t deoxidizers, however, they do not control the shape of the deoxidation prod-ucts, i . e . A l generates alumina galaxies and MnS I I . Although the A l S i a l l o y does produce spherical Ca-aluminate inclusions i t does not so strongly induce the p r e c i p i t a t i o n of CaS at the periphery of the oxides. The CaSi, CaAlSi and CaSiBaAl alloys are very strong i n c l u s i o n shape c o n t r o l l e r s . At low CaSi or high Al or A l S i deoxidation rates MnS II-III or (Ca,Mn)S are formed. 7.9 Inclusion p r e c i p i t a t i o n can be explained by a detailed Henrian-precipitation diagram i n which superimposing the reaction: 231 (CaO)* + [S] t (CaS)* + [0] (7.5.vi) on the previously reported (Ca-Al-0) diagram and using i n t e r a c t i o n c o e f f i c i e n t s from the l i t e r a t u r e a r e a l i s t i c p r e d i c t i o n can be made. This diagram i n t r i n s i c a l l y includes the tr a n s i t i o n s of s u l f i d e s , ruled by: 7.10 F l o t a t i o n of inclusions to some extent (10-20%) occurs i n moderate or highly deoxidized melts. At low deoxidation rates, as expected from 7 . 5 . i i and 7.5.v uniform nucleation of inclusions i n dendriti c spaces during s o l i d i f i c a t i o n takes place. 7.11 If electrode inclusions, slag system, deoxidant and deoxidation rates are known, by using a s i m p l i f i e d ternary - Si02, (Ca,M)0 and A^O^-diagram, the f i n a l i n c l u s i o n composition can be estimated. 7.12 Inclusion s i z e , as expected from (7.10) i s a function of the l o c a l s o l i d i f i c a t i o n time and hence on the l o c a l (interdendritic) thermochemical conditions. * * [Ca] + (MnS) [Mn] + (CaS) 232 SUGGESTIONS FOR FUTURE WORK Based on research c a r r i e d out i n the past, i n terms of inclusions i n conventional steelmaking and i n ESR and the type of reactions studied i n t h i s research, several immediate research proposals are suggested: 1. To determine with a higher degree of accuracy the t r a n s i t i o n point between reactions which involve the oxygen pot e n t i a l i n the s l a g - l i q u i d metal such as: 2[A1] + 3 (FeO) t ( A l ^ ) + Fe and [Ca] + (FeO) 1 (CaO) + Fe and the exchange reactions (between two l i q u i d s ) , such as: 3[Ca] + (A1 20 3) X 3(CaO) + 2[A1] 2. With exactly the same purpose as above, to de-termine the series of tr a n s i t i o n s i n the s u l f i d e i n c l u s i o n phases: MnS II -y MnS III (Ca,Mn)S CaS by designing s p e c i f i c experiments where various oxygen potent-i a l s (several amounts of A l and Ca) i n the melt should be involved. 233 3. As an extension of the r e s u l t s found through t h i s research, i t i s suggested to perform experiments with ex-ac t l y the same techniques and purposes as i n t h i s work i. e . to determine the possible deoxidation and slag best combination i n terms of the following reaction schemes [Si] + 2 (FeO) t (Si0 2) + Fe [Ca] + FeO t (CaO) + Fe and [Si] + 2(CaO) Z s i 0 2 + t C a l and [Mn] + (FeO) % MnO + Fe and [Si] + 2(FeO) J S i 0 2 + Fe or [Ca] + (FeO) J CaO + Fe and either [Si] + 2(MnO) t s i 0 2 + 2 [ M n J or [Ca] + (MnO) t 2(CaO) + [Mn] These reactions can be compared to those already reported for A l 2 0 2 / A l / S i / S i 0 2 and for T i 0 2 / T i . It i s worthwhile to study these reactions since the excess of A l i n ingots can cause deleterious mechanical properties. 234 4. I t becomes obvious that the tr a n s i t i o n s s p e c i f i e d i n point (1) w i l l be also necessary for any other set of e q u i l i b r i a i n proposal (3). 5. I t also becomes apparent that once these conditions are f u l l y c o ntrolled the best slag/deoxidation and hence the most appropriate mechanical properties, as a r e s u l t of the ingot and in c l u s i o n chemistries can be selected. Thus, t h e i r evaluation i n terms of mechanical and chemical resistance as a function of these parameters should be evaluated. 6. It i s also very i n t e r e s t i n g to note that since the Ca and the Al a c t i v i t i e s are c o n t r o l l i n g parameters, i n 8a A l and eo which have been reported with a great deal of scatter should be once and for a l l appropriately determined. 7. Experimental and t h e o r e t i c a l work along the same l i n e s as those proposed i n points (1) and (2) could be per-formed; i n t h i s case, i t i s suggested to use Ce as (RE) deoxidizer i n the presence of a CaO-CaF2 (A^O^) or a RE-enriched slag. The reaction scheme might be as follows: [RE] + (FeO) t RE-O + Fe and/or Ca + (FeO) t (CaO) + Fe LIST OF REFERENCES 235 1. Kay, D.A.R.: Proc. of the F i r s t International Sym-posium on ESR and Casting Technology. Pittsburgh, Pa, 1967, Part II. 2. Myzetsky, V.L., et a l . : Proc. Int. Symp. on Elec t r o -metallurgy, Kiev , 1972, 119. 3. Klyueb, M.M. and Mironov, M. Yu: Stal in English, June, 1967, 6, 480-483. 4. Kuslitsky, A.B., et a l . : i b i d . , 85. 5. Volkov, S.E.: i b i d . , 12. 6. M i t c h e l l , A., Joshi S. and Cameron J . : Met. Trans. 2, February 1971, 561-567. 7. E l l i o t , J.F., and Maulvault, M.A.: Elect . Furn. Conf. P r o c , AIME, 1970, 2%_, 13. 8. Mendrykowski, J. et a l . , Met. Trans., ASM-AIME, 1972, 3, 1761. 9. Tacke, K.H., and Schwerdtfeger, K.: Arch. Eisenhuttenwes, 1981, 52^ , 4, 137-142. 10. M i t c h e l l , A., and Joshi J. ; Met. Trans. 1973, 4_, 631. 11. Fredriksson, H. and Jarleborg, 0.: J. Metals, 1971, 2_3, 9, 32-40. 12. Kay, D.A.R. and Pomfret, R.J.: J.I.S.I., 1971, 209 962-964. 13. Chan, J.C.F., et a l . : Met. Trans. B., March 1976, IB, 135-141. 14. Wadier, J.F., et a l . : Tool Alloy Steels, A p r i l 1978, 127-136. 15. M i t c h e l l , A.: Ironmaking and Steelmaking (Quarterly) 1974, 3, 172-179. 16. Hajra, J.P. and Ratnam, U.: Trans of the Indian Inst, of Metals. 31_, 1, Feb. 1978, 20-23. 236 17. Medovar, B.I. e t a l . : Proc. of the 5th S o v i e t - J a p -anese Symp. on P h y s i c a l - C h e m i c a l B a s i s of M e t a l -l u r g i c a l P r o c e s s e s , 1975. 18. M i t c h e l l , A.: Proc. o f the 1st I n t . Symp. on ESR, P a r t I I I , M e l l o n I n s t . , P i t t s . , Pa., August 1967. 19. Roshchin, E. e t a l . : S t e e l i n the USSR, Dec. 1978, 679-681. 20. Roshchin, E. e t a l . : i b i d . , Feb. 1980, 80-81. 21. B u r e l , B. and M i t c h e l l , A.: Met. Trans. 1970, 3^ , 2253. 22. Paton, B.E. e t a l . : Proc. 5th I n t . Symp. on ESR and Other S p e c i a l M e l t i n g T e c h n o l o g i e s , P i t t s b u r g h , Pa., 1974, P a r t I, 433-448. 23. Zhengbang, L i , Wenhiu, Z. and Yi d a L i ; I r o n and S t e e l V o l . 15, No. 1, January 1980, 20-26. 24. J i e , Fu: Act a M e t a l l u r g i c a S i n i c a , V o l . 15, No. 4, Dec. 1979, 526-539. 25. Wadier, J.F., e t a l . : T o o l A l l o y S t e e l s , A p r i l 1978, 127-136. 26. M i t c h e l l , A. and Beynon B.: Met. Trans. 2L, 1971, 3333-3345. 27. M i t c h e l l , A.: "The E l e c t r o s l a g Remeltinc Process," B u l l e t i n 669, U.S. Bureau of Mines, 1976, 15-19. 28. Schwerdtfeger, K.: Proc. 5th I n t l . Conf. Vac. Met., Munich, 1976, 133-141. 29. Beynon, B., Ph.D. t h e s i s , D e p t . of Met., U.B.C., 1973. 30. M i t c h e l l , A.: " E l e c t r o s l a g and Vacuum Arc Remelting P r o c e s s e s , " t o be p u b l i s h e d i n E l e c t r i c Furnace S t e e l -making, AIME p u b l i c a t i o n . 31. Peover, M.; J . I n s t . Met. 1972, 100, 97-106. 32. Maximovitch, B.I.; Avtom. Svarka, 19 61, 8, 86-88. 237 33. Hawkins, R.J. et a l . : Proc. Conf. i n Electroslag Refining, S h e f f i e l d , Jan. 1973, 21-34. 34. Holzgruber, W.: Proc. of the 1st I n t l . Symp. on ESR Mellon Inst. P i t t s , Pa., August 1967, Part I I. 35. Klyguev, M.M. et a l . : S t a l in English, Feb. 1969, 168-171. 36. M i t c h e l l , A.: Proc. 1st I n t l . Symp. on ESR, Mellon Inst., P i t t s , Pa., Aug. 1967, Part I I . 37. Buzek, Z. and Hlineny, A.: Prob. Special E l e c t r o -metallurgy. Naukova Dumak, Kiev 1972, Part I. (Sbornik Vedeck. Prac. Vys. Skol b.v.s. Vol. 11, 3, 1965, 483-487) . 38. Miska, H. and Wahlster, M.: Archiv. fur Eissen, 44, (1), 1973, 19-25. 39. Tobias, J.B., and Bhat, G.K.: Proc. of the 1st I n t l . Symp. on ESR., Mellon Inst., P i t t s . , Pa., August, 19 76. 40. Schwerdtfeger, K. and Klein, K.: Proc. of the 4th I n t l . Symp. on ESR, Tokyo, Iron and Steel Institute of Japan, June, 1973, 81-90. 41. M i t c h e l l , A.: Can. Met. Quart., 1981, Vol. 20, 101-112. 42. P o v o l o t s k i i , D.Y., et a l . : Izv. Vyssh. Ucheban. Zared Chem. Metall. 1977, 2, 40-42. 43. Zhmoidin, G.I.: Izv. Abad. Nauk. CCCP Metally, 1971, 6, 46-52. 44. Choudhury, A.: Stahl und Eisen, 1980, 100, No. 17, 1012-1018. 45. M i t c h e l l , A.: Trans. Farad. S o c , 1967, 6_3, 1408-1418. 46. Pateisky, G.: Prob. Special Electrometallurgy, Naukova Dumka, Kiev, 19 72, Part I I. 238 47. Pateisky, G. et a l . : Journal of Va. S c i . Technol. 1972, 9, 1318-1326. 48. Kay, D.A.R.: Prob. Spec. Electrometallurgy, Naukova Dumka, Kiev, 1972. Part I I . 49. Kopoywuy, B.: Problemy Spetsial'nai Electrometallurgy, Vol. 13, 1980, 12-23. 50. Korousiv, B. Berg-und Htlttenmannische Monatshefte 122, Jahrgang, July 1977, Heft 7, 287-291. 51. Korousid, B. and Holzgruber, W.: i b i d . , 123, Jahrgang, January 1978, Heft 1, 17-22. 52. Krucinski, M.: Hutnik, Rok 1974, Nr. 11, 539-548. 53. A l l i b e r t , M. et a l . : Iron Making and Steelmaking, 1978, No. 5, 211-216. 54. Wadier, J.F. et a l . : Mem. S c i . Metall., Feb. 1978, 61-78. 55. Kajioka, H., et a l . : Proc. of the 4th Int. Symp. on ESR, Tokyo, Iron and Steel Institute of Japan, June 1973, 102-114. 56. Fraser, M.E.: Ph.D. thesis, Dept. of Met. , U.B.C., 1970. 57. Wei, Chi-ho and M i t c h e l l , A.: Proc. I n t l . Conf. on Process Modelling, AIME, Pittsburgh, 1982, to be published. 58. Knights, C.F., and Perkins R.: Proc. of a Conference on ESR, Sh e f f i e l d , The Iron and Steel Inst., Jan. 1973, 35-40. 59. Kusamichi, H. et a l . : Proc. of the 1st I n t l . Symp. on ESR, Part III, Mellon Inst., Pittsburgh, August, 1967. 60. Krichevec, M.I. et a l . : Prob. Special Electrometallurgy, Naukova Dumka, Kiev 1972, Part I, 61-72. (Neue Hdtt. Sept.-Oct. 1971, 16, (9-10), 614-618). 61. Patchett, B.M. and Milner, D.R.: Welding Research Sup-plement, (Paper presented at the AWS 53rd Annual Meeting i n Detroit, A p r i l , 1972), 491-505(s). 239 62. M i t c h e l l , A.: Ironmaking and Steelmaking (quarterly) 1974, 3, 172-179. 63. Boucher, A.: Prob. Spec. Electrometal. Naukova Dumka, Kiev 1972, Part I I , 47-62. 64. Cooper, CK. et a l . : Trans. AIME, Elec t . Furn. Steel Conf. Proceedings, 1970, 26_, 8. 65. Ballantyne, A.S.: Ph.D. thesis, Dept. of Met., U.B.C., 1977. 66. M i t c h e l l , A. et a l . : Proc. of a Conference i n ESR. Sh e f f i e l d , Iron and Steel Inst., 1973, 3-15. 67. Ballantyne, A.S. and M i t c h e l l , A.: Ironmaking and Steelmaking, 1977, 4_, 222-239. 68. Wahlster, M.: Proc. of the 5th I n t l . Conf. on ESR and Other Special Melting Technologies, Mellon Inst., Pittsburgh, Pa., Part I, October 1974, 40-61. 69. Retelsdorf, J.H., and Winterhager, H.: Prob. Spec. Electrometallurgy, Naukova Dumka, Kiev, 1972, Part II, 122-135. 70. Pho, P. and Eckstein, F. H.-J.: Neue Hutte., Heft 9. Sept. 1980, 342-244. 71. M i t c h e l l , A.: Prob. Spec. Electrometall., Naukova Dumka, Kiev, 1972, Part II, 95-109. 72. Holzgruber, W. and Peterson, K.: Prob. Spec. Electo-metall., 1970, Naukova Dumka, Kiev, 4_, 90. 73. Hozgruber., W. and Peterson, K. : i b i d . , 1973, 6, 190. 74. Holzgruber, W. : Proc. of the 1st I n t l . Symp. on ESR and Casting Technology, Mellon Inst. Pittsburgh, Pa., Part I I , 1967. 75. Rohde, L.E., Choudhury, A.U., Wahlster: A r c h . E i s e n . 42, 1971, 165-174. 76. Holzgruber, W. and Plockinger, E.: Stahl und Eisen, 88, 1968, Nr. 12, 13 Juni. 240 77. Kusamichi, H. et a l . : Proc. of the 1st. I n t l . Symp. on ESR and Casting Technology, Mellon Inst., Aug. 10, 1967, Part II. 78. Choudhury, A., Jauch, R. und VOlklingen, F.T.: Stahl und Eisen, 100, 1980, Nr. 17, 25 August. 79. Jager, H. and Kdhnelt, G. : Berg und Htlttenmannische, Monatsheft, Vol. 190, Nr. 9, 1975, 423-429. 80. Kamardin, V.A. et a l . : Russian Metallurgy (Metally) 4, 1974, 59-63. 81. Opravil, 0.: Proc. Conf. "Metal-Slag-Gas-Reactions," Electrochemical Soc. Toronto, 1975, 923-931. 82. Kay, D.A.R., M i t c h e l l , A., and Ram, M.: J. Iron and Steel I n s t i t u t e , Feb. 1970, 141-146. 83. Rehak, B. et a l . : Proc. of the 5th I n t l . Conf. on Vacuum Metallurgy and ESR Processes, Munich, Oct. 19 76, 147-152. 84. Iodkovskiy, S.A. and Panin, V.V.: Izvest, Akad. Nauk. SSSR, Metally, Marz-Apr., (2), 1968, 115-121. 85. Nakamura, Y. et a l . : Trans. ISIJ. , Vol. 16, 1976, 623-627. 86. Marakhorskii, I.S. et a l . : F i z . Khim. Osn. Proisvod. S t a l , 1968, 1, 277-279. 87. Tokumitsu, N., Nakamura, Y., et a l . : ' Proc. 6th I n t l . Vac. Met. Conf.) on Special Melting, San Diego, Ca. A p r i l 19 79. 88. Somerville, J. and Kay, D.A.R.: Metall, Trans., 1971, 2, 1727. 89. Chai, C.S. and Eagar, T.W.: Wilding Research Supplement of the AWS, 1982, 229-S. (Paper presented at the 62nd Annual Meeting i n Cleveland, Ohio, A p r i l 1981.) 90. Sims, C.E.: The 1959 Howe Memorial lecture: "The Non-Metallic Constituents of Steel," Trans, of the Met. Soc. of AIME, 215, June 1959, 367-393. 241 91. K i e s s l i n g , R. and Lange, N. : "Non-Metallic Inclusions i n Steel,", 2nd Edition. Published by the Metals Society-London, 1978. 92. Pickering, F.B.: "Inclusions" , the I n s t i t u t i o n of Metallurgists, Monograph No. 3, 1979. 93. "Sulfide Inclusions i n Steel," Proc. of an I n t l . Symp. Nov. 1974. Port Chester, N.Y. Ed. J. Barbadillo and E. Snape, ASM, N.6. 94. "Swedish Symposium on Non-Metallic Inclusions i n Steel", arranged by Uddeholm Research Found. Swedish Inst, for Met. Research, Apr. 1981. 95. Van Vlack, L.H.: International Metals Reviews, Sept. 1977, 187-228. 96. Scaninject., I n t l . Conf. on Injection Metallurgy, Lule.a Sweden. June 9-10, 1977. 97. Scaninject I I . , Second. I n t l . Conf. on Injection Metal-lurgy, Lulea, Sweden, June 12-13, 1980. 98. Holappa, L.E.K.: "Ladle Injection Metallurgy," I n t l . Metals Reviews, 1982, Vol. 27, No. 2, 53-76. 99. Turpin, M.L. and E l l i o t , J.F.: JISI, 204, 1966, 217-225. 100. Forward, G. and E l l i o t , J.F.: Journal of Metals, 54, 1967. 101. Turkdogan, E.T.: JISI, 204, 1966, 914-919. 102. Sigworth, G.K. and E l l i o t , J.F.: Met. Trans 4_, Jan. 1973, 105-113. 103. Malm, S.: Scand. Journal of Metallurgy 5_, 1976, 134-137. 104. Malm, S.: Scand. Journal of Metallurgy, 5, 1976, 248-257. 105. Yarwood, J.C., Fleming, M.C. and E l l i o t , J.F.: Met. Trans., Vol. 2, Sept. 1971, 2573-2581. 106. E l l i o t , J.F. and J.K. Wright.: Can. Metall. Quart. 1972, Vol. 11, 574-584. 107. Fredriksson, H. and Hammar, 0.: Met. Trans. B., Vol. 11B, Sept. 1980, 383-408. 242 108. Volmer, N. and Weber, A.: Z. Physik Chem. 119, 1926, 277. 109. Becker, R., Doring W.: Ann. Physik., 2£, 1935, 719. 110. Popel, S.I.: Izv. VUZ Chem. Met 4, 1962, 5-13. 111. Chipman, J . : Trans. AIME, 224, 1962, 1288. 112. McLean, A. and Ward, R.G.: J. of Metals, 2/7/ 1965, 526. 113. Kim, CM. and McLean, A.: Proc. Conf. "Metal-Slag-Gas Reactions," Electrochemical S o c , Toronto, 19 75, 284-299. 114. Von Bodgandy, L. et a l . : Archiv. Eissenhut., 32, Nr. 7. 1961, 451-460. 115. Lindon, P.H. and B i l l i n g t o n , J . C : JISI, 207, 1969, 340-347. 116. Shewmon, P.G.: Diffusion i n Solids, McGraw-Hill Co., New York. 1963, 19. 117. Lindborg, U. and T o r s e l l , K.: Trans-TMS, AIME, 1968, 94-102. 118. Wert, C. and Zener, C. : J. Appl. Phys. 21_ , 1950, 5. 119. Miyashita, Y. : Trans. ISIJ, 1_, 1967, 1-8. 120. T o r s e l l , K. and Olette, M.: Rev. de Met., Dec. 1969, 814-822. 121. Iyengar, R.K.: D. of Ph., Department of Metallurgy, C.M.U., Pittsburgh, Pa., 1970. 122. Mathew, M.P. et a l . : Arch. Eisenhtlt. , 4_6, Juni 1975, Nr. 6 ( i b i d . 45_, Sept. 1974 , Nr. 9, 569-573.) 123. Jacobs, J.E.: Open Heart Steel Conf., AIME, Proc. 40, 1957, 315-318. 124. Hartman, F. : Stahl und Eisen, 65_, Nr. 3-4, 1945, 29-36. 125. Kawawa, T. et a l . : Trans. ISIJ, 8, 1968, 203-219. 243 126. Fuchs, N.A.: The Mechanics of Aerosols. Pergamon Press. N.Y., 1964. 127. Korber, F. : Stahl und Eisen, 19 37, Vol. 57, 19 37, 1349-1355. 128. Urban, S.F. and Chipman,J.: Trans, ASM. Vol. 27, 1935, 645-671. 129. Crafts, W. and H i l t y , D.C.: E l e c t r i c Furnace Steel Conf. P r o c , AIME, Vol. 11, 1953, 121-145. 130. Van Vlack, L.H., et a l . : Trans of the Met. Soc. of AIME, Vol. 221, A p r i l 1961, 220-228. 131. Weinberg, F.: Met. Trans B,, Vol. 10B, June 1979, 219-227. 132. Weinberg, F.: i b i d . , Dec. 1979, 513-522. 133. Boldy, N.D., et a l . : MIT, technical report, Jan. 1981. DAAG-46-78-C-32; AMMRC-TR^-81-4. 134. Sims, C E . and Dahle, F.B.: Trans. Am. Foundrymen1s Ass. 46, 1938, 65-132. 135. Dahl, W. et a l . : Stahl und Eisen, 1966, 86_, 796. 136. Fredriksson, H. and H i l l e r t , M.: Scand. J. of Metal-lurgy 2, 1973, 125-145. 137. Mohla, P.P. and Beech, J . : Journal of the Iron and Steel I n s t i t u t e , Feb. 1969, 177-180. 138. Takada, H. et a l . : Trans. ISIJ, Vol, 18, 1978, 564-573. 139. Ito, Y. et a l . : i b i d . , Vol. 21, 1981, 477-573. 140. Salter, W.J.M. and Pickering, F.B.: JISI, July 1969, 992-1002. 141. Wilson, G.W. and McLean A.: "Desulfurization of Iron and Steel and Sulfide Shape con t r o l , " ISS of AIME, pub. 1980, 1-40. 142. Baker, T.J. and Charles, J.A.: JISI, Sept. 1972, 702-706. 244 143. Steinmetz, E. et a l . : Arch. Eisenhtltt. 4_7, Feb. 1976, 71-76 ( i b i d . 47, Sept. 1976, 521-524). 144. H i l t y , D.C. and Popp, V.T.: AIME Elec t . Furn. Steel Conf. Proc. 1969, Vol. 7, 1952, 56-66. 145. Tahtinen, K. et a l . : Scaninject Conf. June 1980, paper 24. Lulea, Sweden, (ref. 97). 146. K i e s s l i n g , R. and Westman, C.: JISI, July 1970, 669-670. 147. Saxena, S.K. and coworkers: Scand. J. of Metals, Vol. 5, 1976, 105-112. ( i b i d . , Vol. 4, 1975, 42-48). 148. Saxena, S.K. et a l . : Proc. of a Conf. i n Steelmaking, Chicago, 1978, Vol. 61, 561-573. 149. Wilson, A.D.: Metal Progress, A p r i l 1982, 41-46. 150. Kobayashi, S. et a l . : Trans. ISIJ., Vol. 11, 1971, 260-269. 151. Uemura, T. et a l . : AIME, Proc. Open Heart Comm., 1976 Vol. 59, 457-478. 152. FOrster, E. et a l . : Stahl und Eisen, 94, 1974, Nr. 11 23 Mai, 474-485. 153. Nashiwa, H. et a l . : Special Rep. ISIJ., 1977, (6th Japan-USSR-Joint Symp. Phys. Chem. Metall. Processes, 81-94. 154. Takenouchi, T. and Suzuki, K.: Trans. ISIJ. Vol. 18, 1978, 344-351. 155. Oelschlagel, D. et a l . : Iron and Steel International, Dec. 1981, 332-330. 156. Folmo, G. et a l . : Scand. J. of Metall., 9, 1980, 99-104. 157. Haida, O. et a l . : Tsetsu-to-Hagane, 1980, 6_6, 3, 48-56. 158. Haida, 0. et a l . : Proc. of the 8th, Japan-USSR Joint Symp. on Phys. Chem. of Me t a l l u r g i c a l Processes., June 19 81, Tokyo, 56-6 7. 245 159. Church, C P . et a l . : Journal of Metals., Jan. 1966, 62-78. 160. Eklund, G. : Jerkont Annali, Vol. 154 , 1970, 321-325. 161. Rosenqvist, T., and Dunicz, B.L.: Trans AIME, 1952, Vol. 194, 604-608. 162. Turkdogan, E.T.: Trans. TMS-AIME, 1961, Vol. 221, 546-563. 163. H i l t y , D.C and Crafts, W.: J. of Metals, Dec. 1952, 1307-1312. 164. Turkdogan, E.T. et a l . : Met. Trans., Vol. 2, June 1971, 1561-1570. 165. Kor, G.J.W. and Turkdogan, E.K.: i b i d . , Vol. 2, June 1971, 1571-1578. 166. Kor, G.J.W. and Turkdogan, E.K.: i b i d . , Vol. 3, May 1972, 1269-1278. 167. H i l t y , D.C. and Craft, W.: Journal of Metals, Sept. 1954, 959-967. 168. H i l t y , D.C. and F a r r e l l , J.W.: I. and S.M., May 1975, 17-27. 169. Silverman, E.N.: Trans, of the Met. Soc. of AIME, Vol. 221, June 1961, 512-517. 170. H i l t y , D.C. and Crafts, W.: Trans. AIME, 188, 1950, 414 . 17.1. D'Entremont, J.C. et a l . : Trans. AIME, 1963, 227, 14 (ib i d . , 1967, 239^ , 123) . 172. Wentrap, H. et a l . : Arch. Eisen, 13_, 1939, 15. 173. Gokcen, N.A. et a l . : J. of Metals, 1953, 173. 174. McLean, H. and B e l l , H.B.: JISI, 203, 123. 175. Gustafsson, S. and Melberg, P-O.: Scand. J. of Metal-lurgy, Vol. 9, 1980, N. 3, 111-116. 176. Janke, D. and Fisher, W.A. : Arch. Eisen 4_7, 1976, 195-197. 177. Fruehan, R.I.: Met. Trans. 1, 1970, 3403-3410. 246 178. Rohde, L.W. et a l . : Afd. Eisenhiitt. , £2, 1971, p. 1966. 179. Ototani, T. et a l . : Trans. ISIJ. 16, 1976, 275-282. 180. Miyashita, Y. et a l . : Tetsu-to-Hagane', 5_7, 1971, 1969-1975. 181. Kobayashi, S. et a l : Trans. ISIJ., 11, 1971, 260-269. 182. Kusakawa, R. et a l : Japan S c i . Assn. 19th Committee, 1976, paper 21. 183. Watanabe, S. et a l . : Trans ISIJ., Vol. 19_, 1979, 683-688. 184. Steinmetz, E. et a l . : Arch. Eisen. 4_7, 1976, Nr. 4, A p r i l , ( i b i d . , Arch. Eisen 4_8, 1977, Nr. 11, November 569-574. 185. H i l t y , D.C. and F a r r e l , J.W.: 13th Annual Conf. of Metallurgy, Toronto. Aug. 27, 19 74. 186. Hammar, 0.: "Progress Through Research Sandvik." Steel Research Centre Lab. Sandviken-Sweden. 187. Waudby, P.E. and Salter, W.J.M.: JISI, July 1971, 518-522. 188. Morgan, E.L. et a l : JISI, 1968, 206, 987. 189. Cremer, P. and Driole, J . : Met. Trans. B., 13B., March 1982, 45-51. 190. Plockinger, E.: Stahl Eisen, 1956, 76, 810 ( i b i d . , 1957, 77/ 701, i b i d . , 1960, 8_0, 656). 191. Straube, H. et a l . : Arch. Eisenhtlt. July 1967, 509 ( i b i d . , 1967/ 38, 607). 192. Sloman, H.A. and Evans, E.L.: JISI, 165, 1950, 81. 193. Braun, T.B., E l l i o t , J.F. and Flemings, M.C.: Metall.. Trans. B.,1979, Vol. 10B., 171-184. 194. Okohira, K. et a l . : Trans. ISIJ., Vol. 14, 1974, 102-109. 195. Ooi, H. et a l . : Trans. ISIJ., Vol. 15, 1975, 371-379. 247 19 6. Waudby, P.E. and Wilson, F.G.: Proc. I n t l . Symp. Chemical M e t a l l u r g y o f Ir o n and S t e e l , S h e f f i e l d , J u l y 1971, 195-197. 197. G a t e l l i e r , C. e t a l . : ABM. 3_3 , No. 234 , Maio, 1977, 275-283. 198. S p o n s e l l e r , D.L. and F l i n n , R . A . : Trans. AIME, 230, 1964, 876. 199. Shurmann, E.B. e t a l . : A r c h . E i s e n h t l t t . 45_, Nr. 7, J u l i 1974, 433-436. i b i d . , 46, Nr. 10, Oktober 1975, 619-622. i b i d . , 46, Nr. 8, August 1975, 473-476. i b i d . , 4_5, Nr. 6, J u n i 1974 , 336-371. i b i d . , 46, Nr. 12, Dez. 1975, 767-771. 200. Kepka, M. : Neue Htltte, V o l . 21(11), Nov. 1976, 645-652, 201. Lindon, P.H. and B i l l i n g t o n , J.C : J I S I , March 1969, 340-347. 202. Jaeger, H. and Holzgruber, W. : Proc. I n t l . Symp. Chemical M e t a l l u r g y o f I r o n and S t e e l , S h e f f i e l d , J u l y 1971, 195-197. 203. B o r i s , M.J. : AIME, E l e c . Furn. S t e e l Conf. P r o c , 1970, 28, 89. 204. Dunn, E . J . : i b i d . , 1971, 29, 117. 205. Koch, W.: Stahl und E i s e n , 8 1 , 1961, 1592. 206. P i c k e r i n g , F.B.: P r o d u c t i o n and A p p l i c a t i o n o f C l e a n S t e e l , 1972, London, 75-95. 207. Bruch, J..: Mikrochim. A c t a , S u p l . 5., J a n . 1974, S p r i n g e r V e r l a g , 137-146. (Bruch, J . e t a l . : A r c h i v . E i s e n h t l t t , H e l f t 11, Nov. 1965, 799-807) . 208. Lin d o n , P.H. and B i l l i n g t o n , J . C : J I S I , March 1969, 340-347. 248 209. Faulring, G.M. et a l . : Iron and Steel Maker, 1980, 14-20. 210. Holappa, L.E.K. (ref. 94), 19-34. 211. Sanbongi, K.: Trans. ISIJ., Vol. 19, 1979, 1-10. (Emi, T., Sanbongi, K., et a l . : Tetsu-to-Hagane, 1978, 64 (10), 1538-1547) . (Kobayashi, S., Sanbongi, K., et a l . : Trans. ISIJ. 1961, 198, 260-269.) 212. R i k a t a l l i o , P.: Paper 13, Ref. 97, Scaninject I. 213. Usui, T.: Paper 12, Ref. 98, Scaninject I I . 214. Saxena, S.K.: Ironmaking and Steelmaking, 1982, Vol. 9, No. 2, 50-57. 215. Ototani, T. and Kataura, Y.: Trans. ISIJ, Vol. 12, 1972, 334-342. 216. Faulring, G.M. and Ramalingam, S.: Met. Trans. B. 11B, March 1980, 125-130. 217. Faulring, G.M. and Ramalingam, S.: Met. Trans. A. 10A, Nov. 1979, 1781-1787. 218. B e l l , M.M.: M.A. Sc. thesis, Department of Metallurgy, U.B.C., 1971. 219. M i t c h e l l , A. and B e l l , M.M. : Can. Met. Quart. 2, 1972, 363-369. 220. Burel, B.: M.A. Sc. thesis, Department of Metallurgy, U.B.C., 1969. 221. Vaccari, J.A.: Design Engineer, May 1980, 57-60. 222. Viswanathan, R. and Beck, G.: Met. Trans. A., 6A, Nov. 1975, 1997-2003. 223. R a t l i f f , J.L. and Brown, R.M.: Trans. ASM. 1967, V. 60, 176. 224. Kelly, T.N. et a l . , Proc. of the 3rd I n t l . Conf. on ESR and other Spec. Melting, Techs., Mellon Inst. P i t t s -burgh, Pa., June 1971* 125-140. 249 225. Ballantyne, A.S. and M i t c h e l l , A.: Proc. of an I n t l . Conf. on S o l i d i f i c a t i o n . S h e f f i e l d Metall, and E g. Assn. and U. of S h e f f i e l d . July 1977, 363-370. 226. Etienne, M.: Ph.D. th e s i s , Dept. of Metallurgy, U.B.C., 1971. 227. S i d l a , G. and M i t c h e l l , A.: "The Design, Construction and Operation of an ESR Furnace I n s t a l l a t i o n , " Special Report to DREP/DSS, June 1980. 228. Ridal, K.A. et a l . : JISI, Oct. 1965, 995-1003. 229. Malm, S.: Scand. J. of Metallurgy, 4_, 1975, 231-237. 230. Colby, J.W.: Company report, B e l l Telephone Labs. Inc., Allentown, Penn. 1971. 231. Rooney, T.E. and Stapleton, A.G.: JISI, Vol. 131. 1935, 249-254. 232. Yule, J.W. and Swanson, G.A.: At. Absorption News-l e t t e r , 8, 1969, 30. 233. Ingamels, CO.: Anal. Chim. Acta' 5_2, 1970, 323. 234. Rein, R.H. and Chipman, J . : Trans, of the Met. Soc. of AIME, Vol. 233, Feb. 19 65, 418. 235. K l i s i e w i c s , Z.. et a l . : Hutnik 1980, 47 (2), 56-60. 236 . Kuo C-K. and Yen T-S. : Acta Chim. Sinica, 30_, 1964, 381. 237. Nakai, Y. et a l . : Trans. ISIJ., Vol. 19, 1979, 401-410. 238. Bercik, P., Hutnicke, L i s t y , : Vol. 31, No. 12, 1976, 867-872. 239. Wilson, W.G., Kay, D.A.R. and Va hed A.; J. of Metals, 1974, 26, 14-23. 240. Sharma, R.A. and Richardson, F.D.: JISI, Aug. 1961, 386-390. 241. Sharma, R.A. and Richardson, F.D.: Trans, of the Met. Soc. of AIME, Aug. 1965, Vol. 233, 1586-1592. 250 242. Iyengar, R.K. and Philbrook, W.O.: Met. Trans. Vol. 3 July 1972, 1823-1830. 243. Lindskog, L. and Sanberg, H.: Scand. J. Metallurgy, 1973, 2, 71-78. 244. Engh, T.A. and Lindskog, N.: i b i d . 1975, 4, 49-58. 245. Linder, S.: i b i d . , 1974, 5, 137-150. 246. Aritomi, N. and Gungi, K.: Trans of the Japan Inst, of Metals, Vol. 22, No. 1, 1981, 43-56. 247. Suzuki, A. et a l . : Nippon Kinzoku Gakkai Shuho, 32, 1968. (see Brutcher Transl. 7804 (1969). 248. Choudhury A. et a l . : Stahl u. Eisen 96, 1976, Nr. 20, 946-951. 249. Mochizuki, T. et a l . : 4th I n t l . Symp. on ESR, June 1973 ISIJ, Tokyo, Japan, 260. 250. Kadose, M. et a l . : i b i d . 246. 251 Figure(1) -Schematic i l l u s t r a t i o n of an ESR system. 252 Axial length (cm) Figure (2) - Predicted and measured temperature p r o f i l e s for a 1018 M.S. electrode, 25 mm i n diameter. 253 Figure (3) - Manganese content of the metal for univariant equilibrium gamma iro n + "MnO" + "MnS" + l i q u i d (1) for the Fe-Mn-S-0 system and univar iant equi l ibr ium gamma iron + "MnS" + l i q u i d su l f i de for Fe-Mn-S system. 800 1000 1200 1400 1600 Temperature (°C) Figure (4) Univariant e q u i l i b r i a i n Fe-Mn-S-0 system the presence of gamma iron and Mn(Fe) 0 (164) phases 255 >i •P > •H 4-> O — < CD CO CD (0 CO .£ CD ft c C i—I (0 O S co |Fe-Mn-0 ToFe-5-0'S> 1100 1300 1500 1700 Temperature (°C) —•>• actual conditions (cooling) equilibrium conditions (cooling) s t a r t i n g composition ~ — heating conditions Figure (5) f 16 6> ) - Univariant e q u i l i b r i a involving s o l i d metal and Mn(Fe)0 i n Fe-Mn-S-0 system bonded with ternary Fe-Mn-0 and Fe-S-0 terminal-phase f i e l d s ; (e) 6, 'O', 1 2; (p) f S , l 1 ,1 2; (f) 6, 'o', 2 ; (n) 6, 'o\ 1 1, 1 2; (g) 6, y, 'o'; 1 ^ (h) Y# 'o'; 'MnS', 1 1 . 256 MiS .MO-&0, (a) MiS MnO F i g u r e ( 7 ) * 1 6 6 ) - E q u i l i b r i u m phases i n three planes o f the FeO-MnO-MnS-Si0 2 system, a) MnS-FeO-2MnSiOj b) MnS-2FeO'Si0 2-2MnO«Si0 2 and c) MnS-FeO-MnO, 257 (95) Figure (8) a) MnO-Si02 binary phase diagr (T-Mn_SiO. and R-MnSiO-.) b) Schematic i l l u s t r a t i o n of l i q u i d compositions versus Mn/Si/O r a t i o s . (b) i s a section of the Fe-Mn-Si-S-0 system. ABC are s i m p l i f i e d l i q u i d compositions. A'B'C* are l i q u i d compositions saturated with s o l i d s u l f i d e at 1315°C. 258 gure (9) - Schematic i l l u s t r a t i o n of changes i n inclusion composition (enriched i n Si and Mn) i n a 1020 MS electrode. 260 F i g u r e (11) - Isothermal. Fe-Al-Ca-0 p r e c i p i t a t i o n (Henrian a c t i v i t i e s ) d i a g r a m ' 1 * ^ . to 263 Figure (14) - Schematic i l l u s t r a t i o n of the used i n this investigation. ESR arrangement I II & 01 IV Figure (15) - Schematic illustration of the "inclusion extractor". 265 Figure (16) - Inclusions from 1020-steel used as electrode l i g h t microscope. 4 30 X. Spectrum X-ray analysis are aiven i n Figure 17 (a-b) . X-ray energy (KeV) F i g u r e (17) Deformed i n c l u s i o n i n a 1020 M.S. (a) Spectrum X-ray analyses of dark phase. (b) Spectrum X-ray analyses of l i g h t phase. 267 F i g u r e (18) - M a c r o s t r u c t u r e of a 1020-electrode t i p . (a) unetched s u r f a c e (125X). (b) m a c r o s t r u c t u r e where l i q u i d f i l m , s e m i l i q u i d r e g i o n and y - g r a i n growth, areas are shown. 268 (c) (d) Figure (19) - Macrostructures from a 4340-electrode t i p . (a), (c), and (d) show the l i q u i d f i l m , p a r t i -a l l y l i q u i d region and the f u l l y and p a r t i -a l l y austenitized zones- (b), (c) and (e) 30X. (a) shows a droplet i n process of forming (b) 6.6 X. 269 270 F i g u r e (21) Schematic i l l u s t r a t i o n o f a 1020 e l e c t r o d e t i p s u b j e c t e d to ESR thermal g r a d i e n t s . 271 Figure (22) - Multiphase ( r e l a t i v e l y grown) inclusions i n a 1020 electrode. 400 X. 272 Figure (23) - Single phase inclusions i n p a r t i a l l y and f u l l y molten regions i n a 10 20 electrode t i p . (a) 45 X and (b) 400 X. F i g u r e (24) - C o m p l e x ( C a , A l , S i , Mn) i n c l u s i o n s f o u n d i n t h e l i q u i d f i l m a n d d r o p l e t s o f 1020 e l e c t r o d e s ( a ) , ( c ) , (d) - 1 .8 x 1 0 2 X a n d (b) 3 . 6 x 1 0 2 X . 274 X-ray energy (KeV) F i g u r e (25) - (a) T y p i c a l complex (Ca, A l , S i , Mn) i n c l u s i o n i n the l i q u i d f i l m and d r o p l e t of 1020 e l e c t r o d e s (2.4 x 10 3X) (b) Spectrum X-ray a n a l y s i s . 275 Composition i n at. % Figure (26) - Changes i n inclu s i o n chemical composition i n a 4340(1) electrode t i p subjected to (ESR) thermal gradients. 276 Figure (27) - Changes i n in c l u s i o n chemical composition in a 4340 electrode t i p with strong r e c r y s t a l l i z a t i o n . 277 E •H fr, •a cr •rl £ O 4-1 CD U C fO 4 J w •H Q 12000 8000 4000 20 40 • A! ACa o & • Mn spherodization of sulfide inclusions •f 5500iim Incipient heat affected zone solidus " -|- 285Cym Liquid-solid region — i $- 350Mm 60 80 liqiidus Composition ( at. % x) Figure (28) - Behavior of oxide i n c l u s i o n s i n an electrode t i p of a rotor s t e e l subjected to ESR-thermal g r a d i -ents. F i g u r e (29) - A l - s i l i c a t e i n c l u s i o n s i n a 4340 ESR i n g o t , 75 mm i n diameter. Deep etched sample by i o d i n e methyl a c e t a t e methanol (Ingot 3) (a) 1000 X and (b) 2000 X. 279 RIII-W 20 40 60 80 100 T o t a l Oxygen Content (ppm) Figure (30) - Influence of CaSi and FeO i n t e r m i t t e n t additions on the oxygen content i n a 1020 M.S. 280 RII-W 25 20 g 3 15 4-> 2 io o U l c Ca, 2ndaddJ FeO, 2 nd add. 'Ca', 1st add. FeO, 1st add. 40 60 80 100 T o t a l Oxygen Content (ppm) F i g u r e (31) - Changes i n t o t a l oxygen content r e s u l t i n g i n CaSi and FeO i n t e r m i t t e n t a d d i t i o n s d u r i n g r e f i n i n g of a 1020 M s t e e l . 281 Figure (32) - Changes i n slag chemical composition i n a 1020 (RIII-W) s t e e l r e s u l t i n g from CaSi and FeO intermittent additions. 282 RII-W 01 1 1 1 U I | | L 0 0.2 0.4 2 3 4 5 (wt.% Mn) (wt.% Fe X 10 ) Figure (33a) - Changes i n slag chemical composition as a r e s u l t of intermittent additions of CaSi and FeO i n slag during r e f i n i n g . 283 RII-W Slag Chemical Composition (wt.%) Figure (33 b) - Changes i n S i , A l and Ca as a r e s u l t of d i s -crete additions of CaSi and FeO i n the slag during r e f i n i n g . 284 RIII-W 1 i—I—I I—I—i—i—I—r i— i— i— i—I I—i—i—i—I I i i i i i | I • . . . I 0 6 0.75 0 B 5 0.1 02 0 0 1 0 .03 0 0 5 0 0 7 0 2 0 0 2 4 0 .28 Ingot Chemical Composition (wt.%) F i g u r e (34) - Changes i n i n g o t chemical composition as a r e s u l t of CaSi and FeO a d d i t i o n s i n s l a g . 285 Figure (35) - E f f e c t of CaSi and FeO additions in the slag on the chemical composition of a 1020 MS ESR-ingot. 286 R I I I - W +J x: Cn •H (1) BC JJ O Cn C 251 20 15 10 5 h ~i i 1 1 r o inclusion size x 10>m power off 2nd Ca-addition 2nd FeO addition 20 40 60 80 100 120 140 160 a t . % Ca 1 n 4 a t . % A l X 1 0 F i g u r e (36) a,b - Changes i n i n c l u s i o n composition (a) and mean s i z e (b) i n a 1020 MS i n g o t as a r e s u l t o f i n t e r m i t t e n t o x i d a t i o n and d e o x i d a t i o n o f the s l a g . 287 Figure (37) - Chemical analysis of slag samples in RI-H 288 0.1 0.2 0.1 0.2 0.6 0.7 0.8 Ingot Composition (wt. %) Figure (38) - Ingot chemical analysis in R l - I l . (Ingot used as a reference). 289 R I I - I l Slag Composition (wt. %) Figure (39) - Slag chemical analysis (wt.%) i n R I I - I l . 290 4-> tn •H cu K 4-1 o Cn c H 0.2 0.6 1.0 28 30 Slag Composition (wt. % X) Figure (40) - Slag chemical analysis (wt.%) i n RII-I2. 291 R I I - I l 30h 25h 20 15 10 i—r i—I—r \ V o i l i I i o 1 — • Solid O Liquid o I t o 4th add. V 3rd add. A 1 P 2nd add.J J L st add. i  I 2 3 4 5 6 7 8 9 60 80 I00 I n c l u s i o n Mean Diameter (ym) T o t a l Oxygen Content (ppm) Fi g u r e (41) - I n c l u s i o n mean diameter and t o t a l oxygen content i n R I I - I l . 292 30 25 -i 20 4-» x: tn •H 0) K 4-> o tn CJ 15 10 r or T 1 r () V o Liquid \ pool •Solid / ingot I \ / j T r 5 th odd oi: ~o—-~ 4th add 3rd odd 2Qd.Qq!d Ca addition * \ 1 1 *A L ° Liquid pool • Solid ingot 1st add. * Calculated total oxygen content from t extracted inclusions 2 4 6 8 30 40 50 60 70 80 90 100 Inclusion Mean Diameter (ym) Total Oxygen Content (ppm) Figure (4 2) - Total oxygen content and inclusion mean d i a -meter i n RII-I2. 293 T 1 1 1 1 1 1 1 1 r ,* 20 40 , 6 0 , 8 9 liguidpod, 10 20 30 40 50 60 70 80 ingot ,at.% Ca, 1 Q 2 'at.% A1 J Figure (4 3) - Inclusion chemical composition (at.%) as a function of continuously increasing deoxid-ation rates (ingot height) i n ingot RII - I l . 294 30 25 20 e £ 15 4-1 Cn i 10 4-1 o in H 5 r -— 3 » 0 o-=^-~ / o o ' / Ca addition / ^ ^ o u • solid ingot g j_ ^ / ° liquid pool «/ . , , 50 100 Figure (44) Inclusion chemical composition as a function of the ingot height (or continuously increasing deoxidation rates) i n RII-I2. 295 R I I - I l Ingot Chemical Composition (wt.%) Figure (45) - Ingot chemical composition against ingot height (or deoxidation r a t e ) . 296 Ingot Composition (wt.%) Fi g u r e (46) - Ingot chemical composition vs. ingot h e i g h t (or d eoxidation rate) i n RII-I2. 2 9 7 298 (c) (d) Figure (4 8) - "Alumina galaxies* associated with MnS II i n an A l deoxidized ingot, ( R I I - I l ) . 2000 X. (a) BE photograph, (b) A l (c) Mn and (d) S maps. 299 X - r a y E n e r g y (KeV) F i g u r e (49) - a - A l ^ O ^ ( c o r u n d u m ) i n c l u s i o n s i n A l d e o x i d i z e d i n g o t s . (a) u n e t c h e d s u r f a c e , ^ 1000 X; (b) a n d (c ) a r e d e e p ( i o d i n e m e t h y l a c e t a t e ) e t c h e d s a m p l e s , * 3000 X; (d) s p e c t r u m a n a l -y s i s o f (a) a n d (b) r e s p e c t i v e l y ; •v 2000 c o u n t s . 300 (c) (d) Figure (50) - Calcium-aluminate i n c l u s i o n from a heavily A l -deoxidized ingot. (a) backscattered electron photograph taken at 4000 X. Reverse p o l a r i t y . (b) , (c), and (d) are A l , Ca and S maps. RIII-I2-S8-SLD. 301 R I I - I l o o X to u < 0\° 0\° • - p +J f0 8 6 h 1 -•• —"1 1 1 ' / / / / — try Ratio / / / / S - Stoichiome / / / •/ / / 4 /« / • » • c s / / • / r /*! ' i i i i i 0.4 0.6 0.8 1.0 1.2 1.4 1.6 .at.  L a t . Mn" F i g u r e (51) - Composition dependence of s u l f i d e phases on the Ca-aluminate i n c l u s i o n phases i n R I I - I l . Figure (52) - Composition dependence of sul f i d e phases the Ca-aluminate inclusion phases in RII (a) (b) (c) (d) 'igure ( 5 3 ) - Segregated material i n an Al-deoxidized ingot, (a), (b) , (c) and (d) are A l , Ca, Si and Mn maps. -v 1 0 0 0 X. 3 0 4 Figure (54) - Dependence of "FeO" content on the ( — ^ 3 , CaO r a t i o i n slag i n a continuously Al-deoxidized mgot, (RII-H). y ueoxiaizea 305 u 1.3 1.2 1.1 A 1 2 ° 3 A l 0 Figure (55) - Dependence of the "FeO" on the ( —-) r a t i o , CsO i n slag i n a continuously Al-deoxidized incrot. ( R I I - I 2 ) . y ' 306 R I I I - I l Figure (56) - Total oxygen content and i n c l u s i o n mean diameter in a CaSi-deoxidized ingot (RIII-Il) . 307 RIII-I2 T — i — i — i — I | 1 ' 1 1 1 r 0 4 8 3 0 4 0 6 0 8 0 Inclusion Total Oxygen Content (ppm) Mean Diameter (ym) Figure (57) - Inclusion mean diameter and t o t a l oxygen content in RIII-I2. 308 025 0.5 0.75 r a.t • % Ca i L a t . % A1 J F i g u r e (58) - I n c l u s i o n chemical composition (at. % ) , as a f u n c t i o n of d e o x i d a t i o n r a t e s i n R I I I - I l . 309 Figure (59) - Inclusion chemical composition as a function of deoxidation rates i n RIII-I2. 310 R I I I - I l F i g u r e (60) - Changes i n the s l a g composition i n a c o n t i n u -o u s l y Ca-Si d e o x i d i z e d i n g o t ( R I I I - I l ) . RIII-I2 1 r T 1 1 r 30 25 ^ 6 th addition - • \ J " 20 -P si tn •H CD -P O tn C H 15 10 \ c / 3 r d add.. / \ 2nd add. • Fe J < / ^ \ o Si \ 1st add. J I I ' I L 0.2 0.6 1.0 1.4 r t t A Al \ II 13 15 Slag Composition (wt. %) Figure (61) - Changes i n slag composition i n a continu ously (CaSi) deoxidized ingot, (RIII-I2) 312 Figure (62) - Changes in Al and Si i n ingot (RIII-Il) as CaSi deoxidation i s increased. 313 RIII-I2 o Si T r ~ o — i r I r 1 7 6th addition 5 th addition 4th add. 0.2 0.6 0.7 0.8 3rd add. / _i _o 1 2nd add. 6 i 1st i i add. i ) 1 2 3 4 Wt.% X (Al,Mn,Si) i n ingot Figure (63) - Changes i n chemical composition i n a continu-ously CaSi deoxidized ingot, (RIII-I2). 314 R I I I - I l Figure (64) - Dependence of "FeO" contents i n the slag on the deoxidation rates i n R I I I - I l . 315 R H I - 1 2 0 . 8 I , 2 . 5 3 .0 3 . 5 Wt. % Co Wt. % Al ] . slag Figure (65) - Dependence of "FeO" contents i n the slag the deoxidation rate i n RIII-I2. 316 RII I - I l Figure (66) - Sulfur content in inclusions as a function of the Ca:Al r a t i o i n the Ca-aluminate phases (RIII- I l ) . 317 R1I-I2 (Ca-Si deoxidized ingot) ~i r- 1 r~w ( A t % S ). , . i n c l u s i o n Figure (67) - Inclusion composition in samples from pool and ingot as deoxidation rate i s l i q u i d increased. 0 I n g o t H e i g h t (cm) 3 o rt H C 3 P-0 n 3 0) I o a o i- g c t> g o H - 01 01 ~~ 01 cr — 01 3 m Ul » c H (-• M rr M I o M rt\ & (t O X H " a 01 01 3 a ro O ' O cn * o OD O a a. a. ro a. o a. a. ro o • 1 — OJ o a. a. zr \ a Q. _1_ 2 • V ro cn o o. o - I — cn Q a. a. A t . % Ca 1 0 0 i n s a m p i e s from l i q u i d p o o l A t . % A l ro cn oo > r o cn ro o >>. *\ \ 1 a R r ^ 0 — \ * — s \ \ \ \ \ x - \ 0) X \ -v. 1 1 J — 1 8 T £ 319 F i g u r e (69) - Segregate e n r i c h e d i n (a) A l , (c) S i and (d) Mn. * 5000 X. (b) Ca, 320 R-4340(1) 35 30 25 ~ 20 15 10 F L 5. 5 h T — r I.. ... ... M o Fe Mn J L 7^ t / / 4J I f A S i -I I 1 I u 0 01 02 03 0.7 OS 11 13 15 17 % 16 38 AO Slag Composition (wt. %) Figure (70) - Slag chemical analysis of a 4340 ingot continuously deoxidized with a CaSi a l l o y ([R-4340 (1)]. 321 Figure (71) - Ingot chemical composition of a 4340-ingot continuously deoxidized with a CaSi a l l o y . [R-4340 (1)]. 322 Figure (72) a - Variation i n mean inclusion s i z e . b - Oxygen analysis i n a 4340-ingot continuously deoxidized with a CaSi a l l o y , [R-4340(1)]. 323 0.1 25 26 2.7 2.8 29 3.0 3.1 (wt. % A l } S l a ^ Figure (73) - Changes i n slag composition as a r e s u l t of continuously increasing CaSi deoxidation rates, R-4340 (1). 324 Figure (74) - Inclusion chemical composition, i n (a) ingot and (b) l i q u i d pool, as a r e s u l t of continu-ously increasing CaSi deoxidation rates in a 4340 ingot R-4340(1). 325 R-4340 (I) l i q u i d ingot pool 140 120 100 |At.%Caxl00 At.% Al 80 60 40 20 1-700 600 h500 400 ,o o' / / / / / / / / / / / / / 300 / / / / / / 200 / p .00// o LIQUID POOL • SOLID 10 (At. % S ) 20 Figure (75) - Inclusion composition i n terms of the Ca:Al r a t i o and sulfur content (as CaS) in R-4340(1). 326 10 c o CO r—I u c •H CO U -P cd 0.8 0.6 0.4 0.2 • (§) Ca-Si , solid and liquid * ® Hypercal, (§) A l - S i , '» »» »« 2.0 4.0 6.0 80 10.0 12.0 Sulfur Composition i n Inclusions (at. S) Figure (76) - Inclusion chemical composition (oxide and s u l f i d e phases) i n a rotor steel deoxidized with three deoxidizers; R-RS-I, R-RS-II and R-RS-III. 327 (c ) (d) F i g u r e (77) - S e g r e g a t e e n r i c h e d i n A l (40 a t . %) , C a (41 a t . %) , S i (17 a t . % ) t a n d Mn ( b a l a n c e ) i n t h e r o t o r s t e e l d e o x i d i z e d w i t h A l - 6 5 w t . % S i . (a) A l , (b) C a , ( c ) S i a n d (d) M n , 250 X. 328 4 0 1 — i — i — i — i — i — i — i — i — r — r Calcium conlent of steel ppm Figure (78) - Inclusion"precipitation sequence"in a s t e e l containing two levels of s u l f u r . 329 Figure (79) S t a t i s t i c a l determination of the mean inclusion diameter (ym). (a) sample from an ingot, (b) sample from l i q u i d pool. 330 Ca A l Ce Figure (80) - A l , Ca and Ce d i s t r i b u t i o n i n an i n c l u s i o n of a sample extracted from the l i q u i d pool by a quartz tube containing a RE-wire. (a) BE photograph, 4000 X (b) A l , Ca and Ce d i s t r i b u t i o n across the i n c l u s i o n . 331 Figure (81) - (a) - BE photograph = 6COO X, and A l , Ca, Ce and La d i s t r i b u t i o n s i n a l i q u i d pool CaSi deoxidized. La and Ce come from a RE-wire p r e v i -ously located i n the s i l i c a t e tube. 332 Figure (82) - (a) - BE photograph, 4000 X and A l Ca and Zr d i s t r i b u t i o n s i n an i n c l u s i o n of a sample extracted from a l i q u i d pool deoxidized with "hypercal". The Zr was previously located i n the s i l i c a tube. 333 Figure (83) - Composition p r o f i l e s and maps of an i n c l u s i o n i n a sample extracted from a (ESR) l i q u i d pool deoxidized with "hypercal". (b), (c), (d) and (e) are A l , Ca, S, and Zr. 334 (e) (f) Figure (84) - BE photograph (a) * 4000 X and composition maps from an i n c l u s i o n extracted from a l i q u i d pool deoxidized with "hypercal"; (b) Ca, (c) A l , (d) S, (e) Si and (f) Zr. 335 F i g u r e (85) - I n c l u s i o n d i s t r i b u t i o n i n a d e n d r i t i c s t r u c t u r e of 1020-steel samples taken from l i q u i d p o o l d u r i n g r e f i n i n g . ( a), (b) and (c) show commonly found i n c l u s i o n s i n an A l - d e o x i d i z e d i n g o t . (a) 50 X, (b) and (c) 170 X. 3 3 6 337 Figure (87) - Isothermal (1823°K) p r e c i p i t a t i o n (Fe-Al-Ca-O-S) diagram at 0.1 a c t i v i t y of aluminum. 338 Figure (88, - E f f e c t of the a c t i v i t y of A l (h = 0.001 Symbols: A - A1 20 3 S - MnS (11,111) C-6A - CaO-6Al 20 3 C-2A I ?a"2A? 0 ^ " ' ( C a ' M n ) S ° r C a S "2 3 C-A - CaO.Al 20 3 C - CaO A ( l ) - A 1 2 ° 3 ( 1 ) 339 Figure (89) - E f f e c t of the a c t i v i t y of S (h = 0.1, 0.01 and 0.001) on the " p r e c i p i t a t i o n sequence" of calcium aluminates. Symbols are defined i n Figure (88). 340 (b) Figure (90) - Inclusions extracted from a Ca-Si-deoxidized ingot by the Iodine-Methyl Acetate-methyl alcohol method. (RIII-Il-Sl-SLD). Photo-graphs were taken at: (a) 3000; (a 1) and (b) 8000 X. 341 (a) (b) (c) Figure (91) - Inclusions extracted from a Ca-Si-deoxidized ingot by the Iodine-methyl acetate-methyl alcohol"method. (RIII-I1-S3-SLD). (a) and (b) 4000 X and (c) ^ 1000 X. (calcium aluminates) 3 4 2 F i g u r e (92a) - C a l c i u m a l u m i n a t e / c a l c i u m s u l f i d e i n t e r f a c e s o f i n c l u s i o n s i n C a S i d e o x i d i z e d i n g o t s , (a) and (b) are SEM and EPMA photographs 1.2 x 10 3 and 6.0 x 10 3 X. (b) a l s o i n c l u d e s Ca and S a n a l y s i s . (c) are t y p i c a l comp-o s i t i o n s o f core and p e r i p h e r y of i n c l u s i o n s , r e s p e c t i v e l y . 343 Figure (92b) : Spherical calcium aluminate (core) v/ith ( p e r i -pheral) s u l f i d e phases i n CaSi deoxidized ingots (a) and (b) EPMA photographs. (c), (d) and (e) are A l , Ca and S maps, respe c t i v e l y . 3 4 4 RII-I2 Figure ( 9 3 ) - Secondary dendrite arm spacing i n a round 1 0 2 0 MS — ESR ingot. 345 R I I I - I l i n g o t r a d i u s (cm) m ° U l d W a l 1 F i g u r e (94) - Secondary d e n d r i t e arm spacing i n a round (200 i n diameter) ESR i n g o t . 1020 MS. 346 R-4340 (1) wall Ingot Radius (cm)' (95) - Secondary dendrite arm spacing in a round (200mm in diameter) ESR-ingot (4340)'. 347 RII-I2 8 9 mold wall Ingot radius (cm) Figure (96) - Radial size d i s t r i b u t i o n of inclusions in an Al deoxidized ingot. 348 R I I I - I l u cn -u 0) e (0 •H Q C O •H tfi i H C J c 3 o o R o o o o o o o o o o o 8 § 8 A 3 measurements same points • average large dia. inclusions measurements o average value from one picture 1 i i i i ' I o o 4 5 6 Ingot Radius (cm) -I I 8 9 mould wall Figure (97) - Radial inclusion size d i s t r i b u t i o n i n a low Ca-deoxidized ingot (1020 MS). 349 R I I I - I l F i g u r e (98) - R a d i a l i n c l u s i o n s i z e d i s t r i b u t i o n i n a 200 mm ESR i n g o t . ('5-biggest i n c l u s i o n s technique') 350 R 4340 (1) -> r CD +J CD E nJ • H Q C O -H CO H U C CD tn rd S-i CD > 5.0 4.0 3.0 2.0 1.0 8 mold wall Ingot Radius (cm) F i g u r e (99) - R a d i a l i n c l u s i o n s i z e d i s t r i b u t i o n i n a 4340 i n g o t (200 mm i n diameter) d e o x i d i z e d with Ca-65% S i A l l o y . 351 TABLE I M o d i f i c a t i o n t o S t o k e s ' Law f o r D e v i a t i o n from I d e a l i t y I d e a l c o n d i t i o n U = 2 no 2 9 — 9 r = U s t C o r r e c t i o n due t o c r u c i b l e w a l l u = u s t (1 + bJL) d i s t a n c e o f p a r t i c l e c e n t e r from w a l l b s 0. 5 t o 2 P r e s e n c e o f o t h e r p a r t i c l e s U = U . / (1 + K* /,) s t 3 * = c o n c e n t r a t i o n o f p a r t i c l e s by volume K = 1.3 t o 1.9 I n e r t i a l e f f e c t U = U s t / ( 1 + I Re) Re = R e y n o l d s number o f p a r t i c l e L i q u i d p a r t i c l e P r e s e n c e o f s u r f a c e a c t i v e a g e n t s on l i q u i d p a r t i c l e U = 0 ^ 3 s t 2-j + 3p' v i s c o s i t y o f l i q u i d p a r t i c l e v + u' + Y n s t 2u + 3 p 1 + 3y. = r e t a r d a t i o n c o e f f i c i e n t S l i p a t the p a r t i c l e - f l u i d i n t e r f a c e . s t Br + 2 L; c o e f f . o f s l i d i n g f r i c t i o n D • Ust< l i-H' E = 6 ' (exp'a'S (Wk - K ) / k T ) - l ) = s l i p f a c t o r N o n - s p h e r i c a l p a r t i c l e s S l i g h t l y deformed i n c l u s i o n 9 v X r g ^ = e q u i v a l e n t r a d i u s X = dynamic shape f a c t o r R = , 7 X R J eq sed r , = s e d i m e n t a t i o n r a d i u s s e a x = f ( x B ) X S = c o e f . o f s p h e r i c i t y v = 2 Ap_ g r ; 1  3 u g 3 (3 + | Re + j i - We) 2 r p U Re = R e y n o l d s number 2 r p , U J We = Weber number * o r = i n c l u s i o n r a d i u s = d e n s i t y o f l i q u i d m e t a l o = s u r f a c e t e n s i o n TABLE II-A Data for Invariant E q u i l i b r i a i n Fe-S-0 System -RT i n p0 2 -RT i n p g Invariant E q u i l i b r i a pO^i atm P S 2 / atm k c a l k c a l 2 Iron, wustite I. 560°C magnetite, 4.8 X 10~ 2 7 5.5 X l O - 1 " 100.3 50.5 p y r r h o t i t e , gas Iron, wustite I I . 915°C p y r r h o t i t e , 3.2 X 10~ 1 7 2.2 X l o " 8 89.6 41.6 l i q u i d (1), gas Wustite, magnetite, I I I . 942°C p y r r h o t i t e , 1.1 X l o " 1 " * 5.4 X 10~ 6* 77.6 29.3 l i q u i d (1), gas Composition of at 915°C: N^ e = 0.50, N Q = 0.19, N g = 0.31. Composition of l j at 942°C: N F e = 0.49, N Q = 0.19, N g = 0.32. a Fe = 0.09. LO 354 TABLE I I - B E s t i m a t e d Data f o r I n v a r i a n t E q u i l i b r i a i n Fe-Mn-0, Fe-Mn-S and Mn-S-0 T e r n a r y Systems 1527°C Fe-Mn-0 t e r n a r y system 6 - i r o n : L i q u i d i r o n : S o l i d o x i d e : L i q u i d o x i d e : Gas : 5 80 ppm Mn 800 ppm Mn a =0.65 FeO a F e O - ° - 7 3 52 ppm 0 1130 ppm 0 = 0.35 aMnO o MnO 0.27 1.2 X 10 ~ 9 atm Fe-Mn-S t e r n a r y system Gamma i r o n : S o l i d (Fe) s u l f i d e : 980°C S o l i d (Mn) s u l f i d e : L i q u i d s u l f i d e : Gas : 15 ppm Mn a F e S £ 1 83 ppm S aMnS ° ' 3 7 2 1 wt. p e t . Mn p„ = 1.0 X 10~ 7 atm = 30 wt. p e t . S B 1230°C 1225°C Mn-S-0 t e r n a r y system L i q u i d manganese: S o l i d o x i d e : S o l i d s u l f i d e : L i q u i d o x y s u l f i d e : Gas: S o l i d manganese: L i q u i d manganese: S o l i d o x i d e : S o l i d s u l f i d e : Gas: T r a c e s o f S and 0 *MnO S ° - 9 8 >MnS ~' ° ' 9 8 E 30% Mn, s 35%S, s 351 0 p„ = 7.6 X 1 0 _ 2 0 a t m "2 atm T r a c e s o f S and O T r a c e s o f S and 0 s 0.98 p_ = 1.5 X 10" b2 MnO MnS = 0.98 w 2 atm = 5.8 X 1 0 _ 2 0 a t m , p = 1.2 X 10 _ i 2 TABLE III Estimated Data for Invariant E q u i l i b r i a i n  Fe-Mn-S-0 Quaternary System y- i r o n = 10 ppm. Mn > 1 ppm. 0 S o l i d (Mn) s u l f i d e : a M = 0.4 MnS =900°C S o l i d (Fe) s u l f i d e : aFeS s 1 Fe(Mn)0 oxide: a ^ g 0.5 a F g 0 = 0.5 L i q u i d (1) o x y s u l f i d e : s 26 pet. FeO, 54 pet. FeS, 15 pet. MnS and 5 pet. MnO by weight Gas: p0 2 s 3.8 X 10 _ 1 8atm, pS 2 = 1.4 X 10" 8 atm S o l i d Fe/Mn: "MnS" I I . =1225°C "MnO" L i q u i d (1): L i q u i d (2): 90 pet. Mn MnS MnO s 1 = 1 s 0.1 pet. FeO, 0.3 pet. FeS, 65.2 pet. Mns and 34.4 pet. MnO by weight (pet. O/pct. S) for 1 2 ( m e t a l l i c ) < (pet. O/pct. S) for l j (oxysulfide) Ul 356 TABLE IV C a l c u l a t e d and P u b l i s h e d F r e e E n e r g y D a t a  f o r F e-O-Ca-Al System a t 1823 K ( 1 5 5 0 ° c / 2 1 6 ' E q u a t i o n J/kg 3 Atom Ref. Sigworth"'' Sigworth''' E l l i o t t 2 JANAF 3 JANAF 3 Ca(g) = C a ( l wt. %) 50 574 .68 (-39 481 . 52 + 49.4T) A l ( l ) = A l (1 wt. %) -114 137 . 1 (-63 220 .68 - 27.93T) 1/2 0,(g) = 0 ( 1 wt. %) -122 498 .9 ( -117 230 .4 -• 2.89T) 2 A l + 372 0 2 = A 1 2 ° 3 -1 085 230 . 1 C a + 1/2 0 2 = CaO -437 996, .22 2 A l + 3 0 = A l 2°3 -489 480. .46 Ca + 0 = CaO -366 081. .65 CaO + A l 2 ° 3 = C a O - A l 2 0 3 -45 845 . .46 CaO + 2 M 2 0 3 = Ca0'2 A 1 2 0 3 -50 869 . ,62 CaO + 6 A 1 2 0 3 = CaO-6 A 1 2 0 3 -60 708. ,60 Ca + 2 A l + 4 0 = C a O - A l 2 0 3 -901 407 . ,57 Ca + 4 A l + 7 0 = CaO-2 A 1 2 0 3 -1 395 912. 20 Ca + 12 A l + : 19 0 = CaO-6 A 1 2 0 3 -3 363 673. 03 Ca + S = CaS -257 659 . 0 K i r e e v ^ T a y l o r 5 1. G.K. S i g w o r t h and J . F . E l l i o t t : Met• S c i . , 1974, v o l . 8, pp. 298-310. 2. J . F . E l l i o t t and M. G l e i s e r : T h e r m o c h e m i s t r y f o r S t e e l m a k i n g , A d d i s o n - W e s l e y Pub. Co., 1960. 3. JANAF T h e r m o c h e m i c a l T a b l e s , 2nd e d . , N a t i o n a l S t a n d a r d R e f . E d i t i o n , 1971. 4. V.A. K i r e e v : Sb. T r . Mosk. I n z h - S t r o i t , I n s t . , 1971, v o l . 69, pp 3-18. 5. J . T a y l o r : P r o c . B r i t . Ceram. S o c , 1967, v o l . 8, pp. 115-23. TABLE V (216) Equilibrium Constants for Deoxidation Reactions Compound A c t i v i t y Product K (1823 K) CaO 1 / ( h C a X h 0 ) 3 - 0 5 X l o l ° CaO-Al 20 3 l / ( h C a X h A l X X h 0 } 6 , 4 6 X 1 q 2 5 CaO-2 A1 20 3 1 / ( h C a X h A l X h 0 ) 9.46 X 10 3 9 CaO-6 A1 20 3 1 / ( h C a X h A l X ^ > ) 2 ' 1 3 X 1 q 9 6 A1 20 3 1 / ( h A l X h 0 ) 1 ' 0 4 X 1 q 1 4 TABLE VI E q u a t i o n s o f L i n e s Between t h e I n d i c a t e d Phases A l 2 0 3 - C a O - 6 A l 2 0 3 h C a X h Q = 6.01 X 1 0 ~ 1 3 h A l 3 X h O = 2.13 X 1 0 " 5 h C a / h l ( 3 = 2.82 X l O " 8 C a O - 6 A l 2 0 3 - C a O - 2 A l 2 0 3 h C a X h Q = 1.59 X 1 0 " 1 2 h A l 3 X h O = 2 , 0 1 X 1 0 " 5 h C a / h 2 { 3 = 7.90 X 1 0 " 8 C a O - 2 A l 2 0 3 - O a O - A l ^ h C a X h Q = 2.24 X 1 0 " 1 2 • , 2 / 3 Y h a A l X h 0 h2 { 3 X h „ = 1.90 X 10 5 h „ / h 2 { 3 = 1.18 X 1 0 " 7 C a O - A l 2 0 3 ( s o l i d ) ^ 0 3 0 ( 4 2 p e t . ) + A 1 2 0 3 (58 p e t . ) ( l i q u i d ) h C a X h Q = 2.27 X 1 0 1 2 h A l 3 X h 0 = 1.89 X 1 0 " 5 h C a / h 2 { 3 = 1.20 X l O " 7 CaO ( s o l i d ) -CaO (57.4 p e t . ) + A 1 2 0 3 (42.6 p e t . ) ( l i q u i d ) h C a X h Q = 3.29 X 1 0 - 1 1 h A l 3 X h 0 = 5 - 7 7 X 1 0 ~ 6 h c a / h 2 { 3 = 5.70 X l O " 6 TABLE VII Chemical Analysis of Electrodes Used i n t h i s Research: Electrode Chemical Composition in wt.% 1020 M Steel 4340 - Steels Rotor Steel (1) (1) (2) (1) (2) c 0 . 19 0 . 415 0.422 0.213 0 . 202 p 0 . 099 0 . 014 0.014 0 . 007 0.007 s 0.026 0.016 0.015 0.005 0 .006 Mn 0 . 709 0.697 0.696 0.673 0 . 673 Cu - 0.094 0.091 0 .049 0.048 Ni - 1.882 0 . 184 0 .357 0.0362 Cr - 0 . 873 0. 867 1. 151 1.1150 Si 0 . 25 0.357 0.353 0 . 246 0 . 244 V - 0. 005 0 .005 0 . 246 0 . 244 Mo - 0. 189 0.188 0.940* 0.940* A l - 0 . 029 0.029 0.006 0 .006 Sn 0.005 0 .005 0.005** 0.005* where (*) stands for more than 0.940 wt.% and (**) indicates less than 0.005 wt.%. TABLE V I I I - L i s t of E x p e r i m e n t s I n i t i a l S l a g Run Type o f E l e c t r o d e C o m p o s i t i o n Type o f No. o f A d d i t i o n s Type o f D e o x i d a t i o n r a t e s No. E l e c t r o d e Diameter (mm) C a F 2 / A l 2 O 3 / C a O / S i O 2 / M g 0 D e o x i d i z e r Rates A d d i t i o n (deox.) (kg t o n - 1 ) 1 4340(I) 31.75 50 30 20 - - N i l - - -2 •' " 55 15 15 15 - - - - -3 •• 40 20 20 20 - - - - -4 40 22 23 5 - - - - -5 » 55 15 22 8 - - - -6 31 0 46 23 - - - - _ i 7 •• 50 30 20 - - A l a l o n g r e m e l t i n g c o n s t a n t 2.3 kg t o n 8 •• " 55 15 15 15 - A l " 9 4340(II) 44 .75 50 30 20 - - C a - S i 10 4340 ( I I ) " 55 15 15 15 - C a - S i 11 r o t o r s t e e l 114.3 49 16 17 12 6 A l " c o n s t a n t ^ 0.0 2 kg ton 12 1020 M.S. 76. 2 70 0 30 - - A l " " 13 r o t o r s t e e l 114 ,3 70 15 15 - - A l " 14 " 70 0 30 - - A l II II " 15 " •• 50 20 30 - - A l " RII-W 1020 M.S. 76 . 2 50 30 20 - - C a - S i 2 i n t e r m i t t e n t 50 grams (each) RIII-W 1020 M.S. 76.2 70 30 - - - C a - S i 2 50 grams (each) R I - I l 1020 M.S. 76.2 50 30 20 - - N i l - - -R I I - I l 50 30 20 - - A l 4 c o n t i n u o u s l y i n c r e a s i n g 3.63, 6.1 and 3 7.6 kg t o n RII-I2 » 50 30 20 - - A l 5 1.21, 2.42, 3.6' 4.85,6.06 and 12.1/ R I I I - I l " •• 60 36 4 - - C a - S i 4 . 5 " 5.61,11.23,16.83,2; and p a r t i a l l y 28.0J R I I I - I 2 '• 50 30 20 - - C a - S i 6 5.61,11.23,16.83,2; 28.05 and 56.10 . R-4340 4340 88.9 50 30 20 - - C a - S i 6 4.17,8.35,12.5,16. 20.85,41.7, and 20 R-RSI r o t o r s t e e l 114 . 3 50 30 20 - - C a - S i a l o n g r e m e l t i n g c o n s t a n t 36.0 kg t o n - 1 R-RSII 50 30 20 - - A l - S i c o n s t a n t 33.0 R-RSI 11 50 30 20 - - Hyperca1 c o n s t a n t 36.0 oo o 361 TABLE IX Chemical Composition of D e o x i d i z e r s C a l c i u m - S i l i c o n Hypercal A l - S i Calcium 29 . . 50 10 , . 50 -S i l i c o n 6 2 . . 50 39 , . 00 65 Iron 4 , . 50 18 , .00 -Barium 0. . 50 10 , . 30 -Aluminum 1. . 20 20 , . 00 35 Manganese 0 , . 2 5 0 . . 30 Carbon 0 . . 55 0 , . 50 Chromium 0 . . 10 0 , . 0 3 Copper 0 , . 01 0 . . 03 N i t r o g e n 0 . . 0 3 0 , . 0 5 N i c k e l 0 . . 01 0 , . 02 Oxygen 0 . . 50 0 , .70 Phosphorous 0 . . 01 0 , . 02 S u l f u r o , . 0 5 5 0 , . 12 Titanium 0 , . 08 0 , . 06 Bulk D e n s i t y 110 l b . / c u . f t . 95 l b . / c u . f t . TABLE X I n c l u s i o n C h e m i c a l C o m p o s i t i o n as a F u n c t i o n o f S l a g and D e o x i d i z e r i n 4340 ( s m a l l d i a m e t e r ) ESR I n g o t s a ) I n c l u s i o n C h e m i c a l C o m p o s i t i o n as a F u n c t i o n o f S l a g System No. Nominal S l a g System (wt.%) Atom P e r c e n t Types o f I n c l u s i o n s C a F 2 A 1 2 0 3 CaO S i 0 2 A l Ca S i 1 50 30 20 - 96.1 3.733 0.170 C a l c i u m a l u m i n a t e s , Manganese s u l f i d e s 2 55 15 15 15 75.86 1.8470 22.28 A l u m i n o - S i l i c a t e s , Manganese s u l f i d e s , and " F a y a l i t e . " * 3- 40 20 20 20 85.33 2.470 13.160 C a l c i u m , Aluminum-s i l i c a t e s , Manganese s u l f i d e s and " F a y a l i t e " * 4 40 22 33 5 99.20 0.19 0.60 "Alumina", Manganese s u l f i d e s 5 55 15 22 8 96.27 3.462 0.265 "Alumina" and Ca-a l u m i n a t e s , " F a y a l i t e " 6 31 0 46 23 84.88 4.38 10.74 C a l c i u m - a l u m i n o s i l i c a t e s . Manganese s u l f i d e s and " f a y a l i t e " . * b) S l a g - D e o x i d a n t E f f e c t on F i n a l I n c l u s i o n C h e m i s t r y wt. % (X) at. % (X) + No. CaF 2 A1 20 3 CaO S i 0 2 Deox. A l Ca S i 7 50 30 20 - A l 92.42 7.11 0.170 C a l c i u m a l u m i n a t e s -C a l c i u m s u l f i d e s 8 55 15 15 15 A l 89.20 7.85 2.940 A l u m i n a t e s , C a l c i u m Aluminum s i l i c a t e s and " F a y a l i t e " . * 9 50 30 20 - C a - S i 91.00 7.20 1-80 C a l c i u m a l u m i n a t e s , (*) Manganese s u l f i d e s 10 55 15 15 15 C a - S i 76.217 2.143 21.64 Aluminum c a l c i u m s i l i c a t e s , (*) f a y a l i t e * and Manganese s u l f i d e s c) I n c l u s i o n C h e m i s t r y o f E l e c t r o d e s A l Ca S i Electrodes f o r 1 to 8 (31.75 im) 89.00 10.00 balance C a l c i u m A l u m i n a t e s -c a l c i u m s u l f i d e s Electrodes!*) for 9 and 10 (44.75 nm) 81.08 5.237 13.685 Aluminum C a l c i u m s i l i c a t e s and Manganese s u l f i d e s Remarks: 1. Melting rates 1.2 - 1.5 Kg min. - 1, 2. deoxidation rate"^2.3 Kg t o n - 1 3. Fay a l i t e * was not always observed to follow the theoretical stoichiometry. TABLE XI C h e m i c a l E f f e c t o f S l a g and E l e c t r o d e S u r f a c e P r e p a r a t i o n on I n c l u s i o n C o m p o s i t i o n  ( E x t e n s i o n o f R e s u l t s Found i n 4340 S m a l l and L a r g e E S R - i n g o t s , and 1020 S t e e l s t o a Cr-V-Mo R o t o r S t e e l * I n g o t E l e c t r o d e Type Nominal S l a g C o m p o s i t i o n I n c l u s i o n Type and Shape C a F 2 CaO A 1 2 ° 3 s i 0 2 M g 0 * 11 r o t o r s t e e l 49 16 17 12 6 A l - C a s i l i c a t e s and a l u m i n a t e s . Semiround t y p e s ; t h e y were o c c a s i o n a l l y seen w i t h p e r i p h e r a l MnS. Mg t r a c e s were a l s o de-t e c t e d . 12 r o t o r s t e e l 70 15 15 - - A l u m i n a t e s o f t h e t y p e F e 0 - A l 2 0 3 and A 1 2 0 3 . Round, e l o n g a t e d ( F e O - A l 2 0 3 ) and c l u s t e r s and a n g u l a r A 1 2 0 3 . * 13 r o t o r s t e e l 70 30 - - A l u m i n a t e s , s i n g l e r o u n d and c l u s t e r s . I r o n o x i d e s and some i r o n s u l f i d e s were a l s o o b s e r v e d . * 14 r o t o r s t e e l 50 20 30 - - A l u m i n a t e s , s i n g l e r o u n d and c l u s t e r s and a minor amount o f r o u n d C a - a l u m i n a t e s and MnS I I . 15 1020 70 - 30 - - A l u m i n a t e s g e n e r a l l y as c l u s t e r s and MnS I I . Remarks 1. These E S R - i n g o t s were s l i g h t l y d e o x i d i z e d w i t h A l a t a c o n s t a n t r a t e , 0.02 Kg t o n ^. 2. E l e c t r o d e s 11, 12, 13 and 14 were s u r f a c e g r o u n d and c o a t e d w i t h an Al-Mg s p i n e l p a i n t i n g t o p r e v e n t s c a l e ("FeO") o x i d e f o r m a t i o n d u r i n g r e m e l t i n g . 3. E l e c t r o d e and E S R - i n g o t c o m p o s i t i o n a r e g i v e n i n T a b l e s (VII) and ( V I I I ) . 4. R e f i n i n g c o n d i t i o n s were e q u i v a l e n t f o r a l l e x p e r i m e n t , A r - a t m o s p h e r e and m e l t i n g r a t e s a b o ut 1 Kg min TABLE XII Slag Chemical A n a l y s i s from (Ni-Cr-Mo) Rotor ESR-ingots * Deoxidized with CaSi, A l S i and Ca-Al-Ba-Si A l l o y s CaF 2 CaO A 1 2 0 3 S i 0 2 "FeO" Nominal ( i n i t i a l ) s l a g composition (wt.%) 50 20 30 C a S i * * 47.33 20.15 32.2 0.175 0.14 A l S i 43.32 18.27 38.11 0.141 . 0.147 H y p e r c a l * (Ca, A l , Ba, S i a l l o y ) 43.51 18.19 37.88 0.25 0.165 * * Remarks: Deoxidation r a t e used i n t h i s experiment was s l i g h t l y lower than i n the other two experiments, i . e . ^ 33 grams/min and ^ 36 grams/min r e s p e c t i v e l y . Remelting c o n d i t i o n s were approximately constant f o r the above runs, i . e . , m e l t i n g r a t e s were ^ Kg/min. Experiments were c a r r i e d out under a p r o t e c t i v e atmosphere (argon). cn TABLE XIII - A Chemical Analysis (wt.%) of a Cr-Mo-Mn-Ni-V-Steel Deoxidized with Al-65.0 wt. % Si C P s Mn Cu Ni Cr Si V Mo Al Sn 1 0 . , 261 0 . 008 0 . 003 0 . 707 0. ,050 0 , . 353 1. 176 0 . ,594 0 . 239 +0 , .94 + 0 .250 -0 .005 2 0 . 252 0 . 007 0. ,003 0. , 700 0. .051 0, . 369 1. 165 0. , 536 0 . 246 + 0, .94 + 0, .250 -0 .005 3 0 . 255 0. .007 0. .003 0 . 695 0 . 050 0 . 350 1. , 160 0. , 532 0 . 242 + 0 .94 + 0 .250 -0.005 4 0. . 253 0. .007 0. .003 0. .695 0. ,050 0 . 348 1. 160 0. . 536 0. 242 +0 .94 + 0 .250 -0.005 5 0. . 251 ' 0 . 007 0. .003 .0 , . 689 0. .051 0 . 362 1. . 147 0 . 536 0 . 245 + 0 .94 + 0 .250 -0 .005 6 0 .249 0. .007 0. .003 0, .691 0. .049 0 . 355 1. . 155 0 , .534 0 . 241 + 0 .94 + 0 .250 -0.005 7 0 . 255 0, .007 0, .003 0 .693 0, .051 0 . 354 1. . 155 1 0 , . 538 0 . 243 + 0 .94 + 0 .250 -0.005 8 0 . 250 0 .007 0 .003 0 .695 0, .051 0 . 359 1. . 160 0 . 538 0 . 243 + 0 .94 + 0 .250 -.0.05 9 0 . 260 0 .008 0 .003 0 .701 0, .052 0 . 361 1, . 161 0 . 539 0. 245 + 0 .94 + 0 .250 -0 .005 10 0 .254 0 .007 0 .003 0 . 694 0 .052 0 .361 1 . 154 • 0 . 543 0 . 246 + 0 .94 + 0 .250 -0 .005 TABLE X I I I - B C h e m i c a l A n a l y s i s (wt.%) o f a Cr-Mo-Mn-Hi-V S t e e l D e o x i d i z e d w i t h a C a - S i A l l o y p i e No. C P c Mn Cu N i C r S i V Mo A l Sn 1 0 . 262 0 .007 0. .002 0, .691 0 .049 0 .337 1.103 0 .627 0 . 247 +0 .94 0. . 157 -0 .005 2 0 .271 0. .007 0 .002 0 .696 0 .048 0 . 334 1. 116 0 .674 0 . 236 + 0, .94 0. . 150 -0 .005 3 0 .242 0 .007 0 .002 0, . 673 0 .046 0 . 321 1.091 0 .646 0 . 235 + 0. .94 0. . 147 -0 .005 4 0 . 258 0, .007 0, .002 0, .688 0 .045 0 . 324 1.118 0 .650 0 .235 +0, .94 0 . 117 -0 .005 5 0 .266 0. .007 0. .002 0. .691 0 .04 8 0 . 343 1.115 0 .648 0 .244 + 0. .94 0 . 126 -0 .005 6 0 . 226 0, .007 0. .002 0. .690 0 .047 0 . 340 1. 119 0 .654 0 . 242 +0. .94 0 . 126 -0 .005 7 0 . 273 0. .007 0. .002 0 . 687 0 .047 0 .341 1. 116 0 .652 0 .652 + 0. .94 0 . 125 -0 .005 8 0 .264 0. .007 0. .002 0. . 680 0 .046 0 . 337 1. 1031 0 . 645 0 . 240 + 0. .94 0 . , 121 -0 .005 9 0 . 273 0. ,007 0 . 002 0. . 685 0 .048 0 .351 1. 114 0 .660 0 . 246 + 0 . 94 0. , 121 -0 .005 10 0 . 276 0. .007 0. .002 0. , 687 0 .047 0 .332 1. 118 0 . 652 0 . 241 + 0. .94 0 . , 119 -0 .005 11 0 . 228 0. .007 0. .002 0. , 700 0 .049 0 . 355 1. 125 0 . 665 0 . 245 + 0. .94 0 . , 122 -0 .005 (+) - more t h a n a c e r t a i n c a l i b r a t i o n (-) - l e s s t h a n a c e r t a i n c a l i b r a t i o n cn TABLE X I I I - C C h e m i c a l A n a l y s i s o f a C r - M o - M n - N i - V - S t e e l  D e o x i d i z e d w i t h a C a - S i - A l - B a A l l o y %C 9 p % s %Mn %Cu ? N i % C r S i V %Mo % A l Sn 1 0 . 253 0 008 0. 004 0.706 0.056 0 366 1. 126 0 815 0 229 +0 .94 + 0 .250 -0 005 2 0. 255 0 007 0. 003 0.702 0.053 0 366 1. 146 0 772 0 241 + 0 94 + 0 250 -0 005 3 0. 253 0 008 0. 004 0. 704 0.051 0 34 7 1. 160 0 757 0 234 + 0 94 + 0 250 -0 005 4 0. 267 0 008 0. 004 0.718 0.052 0 356 1. 174 0 774 0 236 + 0 25 + 0 250 -0 005 5 0.263 0 008 0. 003 0.714 0. 056 0 382 1. 164 0 793 0. 248 + 0 94 + 0 250 -0 005 U l CTi TABLE XIV-A Example o f i n c l u s i o n s i z e d i s t r i b u t i o n i n R l l l - I l - L Q D - S a m p l e 1. Sample I n c l u s i o n Shape F l o r e s c e n c e D i a m e t e r No. No. under t h e i n pm e l e c t r o n beam 1 r o u n d b l u e 4 . 5 2 e l o n g a t e d b l u e 5.0 3 s e m i r o u n d b l u e 6.0 4 e l o n g a t e d b l u e 7.0 5 r o u n d b l u e 4.5 6 S-shape b l u e 6.0 7 r o u n d b l u e - g r e e n 6.0 8 a n g u l a r b l u e 4 . 5 9 d u p l e x - r o u n d 2 - b l u e 5.0 10 r o u n d b l u e 4.0 11 t r i a n g l e b l u e 5.5 12 d u p l e x b l u e 8.5 13 e l o n g a t e d b l u e 5.5 14 e l o n g a t e d b l u e 5.5 15 t r i a n g l e b l u e 5.5 16 a n g u l a r b l u e 4.5 17 e l o n g a t e d b l u e 6.5 18 r o u n d l i g h t - b l u e 8.0 19 d u p l e x 2 - b l u e 8.0 20 i r r e g u l a r b l u e 6 . 0 21 r o u n d b l u e 8.0 22 r o u n d b l u e 6.0 369 TABLE XIV-B Example o f I n c l u s i o n S i z e D i s t r i b u t i o n i n R I I I - I l - S L D - S l Sample No. L u s i o n No. Shape F l u o r e s c e n c e D i a m e t e r i n ym 1 s e m i - r o u n d b l u e 4.0 2 e l o n g a t e d b l u e 6.0 3 r o u n d v i o l e t 5.5 4 e l o n g a t e d b l u e 4 . 5 5 r o u n d b l u e 4 . 5 6 i r r e g u l a r b l u e 5.5 7 e l o n g a t e d b l u e 5.5 8 r o u n d b l u e ^ 8.0 9 i r r e g u l a r b l u e 6.0 10 e l o n g a t e d b l u e 5.0 11 r o u n d b l u e 6.0 12 i n a c l u s t e r -r o u n d b l u e - g r e e n 5.5 13 i n a c l u s t e r -r o u n d b l u e 6.5 14 e l o n g a t e d b l u e 5.0 15 r o u n d b l u e 5.0 16 h a l f - m o o n shape b l u e 7.5 17 e l o n g a t e d -i r r e g u l a r b l u e 5 .0 18 e l o n g a t e d b l u e 5.0 19 e l o n g a t e d b l u e 4 .0 20 e l o n g a t e d b l u e 5 .0 TABLE XV Data o f P l o t F i g u r e s (87 - 89) E q u i l i b r i a A l 2 0 3 / C a O - 6 A l 2 0 3 / C a S 6 A l 2 0 3 - C a O / C a O - 2 A l 2 0 3 / C a S C a O - 2 A l 2 O 3 / C a 0 - A l 2 0 3 / C a S + C a O - A l 2 0 3 / ( C a O ) * + ( A l ^ W C a S + + C a 0 s o l i d / ( C a 0 ) t + ( A l 2 0 3 ) + / C a s V h S 1.4155 x 10' 1.995 x 10" Ca S -3 0.758 x 10 4 x 1 0 ~ 3 ^ 1.35 x 10" 2.3155 x 10 1.83 x 1 0 ~ 9 3.360 x 1 0 - ! 6.87 x 1 0 ~ 9 ^ 1.35 x 10 -10 -8 A l S 1.4228 x 10 7.85 x 1 0 ~ 4 1.2457 x 10 2.5 x 1 0 " 3 -4 -3 ^ 3.84 x 10 -3 h A l h O -5 2-014 x 10 ! - 5 6 7 x 1 0 " 5 1.2156 x l 0 - 5 1-0 x l o " 5  % 9 -61 x 1 0 ~ 6 * t S l a g c o m p o s i t i o n s a c c o r d i n g t o the C a O - A l ^ p s e u d o - b i n a r y d i a g r a m a t 1827° K (1550°C) 0.7, a, + 3 C a S = ° - 0 3 " , * A l 2 o 3  + + a C a S = °- 9*8 -1.0, a A l a CaO 0.1, a 2 U 3 0.0625 CaO (238,239) 0.8 - 0.9 CaS ~ f r o m Sharma and R i c h a r d s o n @ x = 0 568 C a 0 ' Y C a S ( s o l i d ) They s u g g e s t v /< Y C a S ( s o l i d ) / J = Y C a S ( l i 65 i q u i d ) and X = 1 x 10 3 - 2 x 1 0 ~ 3 . T h i s i s a l s o an a v e r a g e v a l u e s i n c e @ 1650°C X CaS 6.3 x 10 and a v = l^ f i f l r C a S ( 1 6 5 0 ° C ) ' i b 6 B 371 T A B L E XVI E q u i l i b r i u m ( i n v a r i a n t ) and e A 1 =-5.25, e^a = -62, and «? = - 4 0 %Ca %0 %A1 %S (1) 1 x 10" 3 2.65 x 10~ 2 2.35 x 10" 3 7.4 x 10" (2) 2.5 x 10~ 3 2.6 x 10~ 2 5 x 10~ 2 5.35 x 10 (3) .5 x 10~ 3 2.35 x 10~ 2 1 x 10~ 2 3.95 x 10 (4) 8 x 10" 3 2.27 x 10" 2 1.2 x 10~ 2 2.75 x 10 (5) 1 x 10" 2 2.15 x 10" 2 1.89 x 10" 2 1.56 x 10' 2 2 -2 2 TABLE XVII Computed Compositions by Using Data i n Table XV Interaction Parameter Composition (ppm) e£ a Ca A l 0 Equilibrium (invariant) (i) -535 10 15 20 II (2) -400 25 38 32 (3) -300 65 97 34 ti (4) -250 80 120 36 ii (5) -200 100 150 42 The in t e r a c t i o n parameter for the C a - 0 was assumed variable and the A l - 0 and Ca-S were: e A l = -62 and -110 TABLE X V I I I E f f e c t o f I n i t i a l Number o f I n c l u s i o n s on Growth D u r i n g C o o l i n g o f L i q u i d M e t a l Number o f I n c l u s i o n s I n i t i a l F i n a l Growth Time I n i t i a l l y R a d i u s R a d i u s (Lowe r L i m i 1 0 3 / c c 1 ym 40 72 um 279. 5 s e e s » 2 40 72 268. 0 5 40 75 238 . 9 9 40 87 208 . 3 10 40 92 201. 7 1 0 4 / c c 1 18 90 57. 37 2 18 91 52. 79 5 19 02 42. 40 " 9 19 56 32. 87 " 10 19 79 30. 98 1 0 5 / c c 1 8 78 11. 23 2 8 81 9 . 58 5 9 28 6 . 38 9 11 2 4 . 15 10 11 88 3. 79 1 0 6 / c c 1 4 09 2. 02 2 4 23 1. 5 " 5 5 77 0. 76 9 9 27 0 . 44 10 10 22 0. 39 1 0 7 / c c 1 1 98 0. 31 It 2 2 45 0. 19 " 9 9 03 0. 04 10 10 02 0. 0397 374 APPENDIX Thermodynamic r e l a t i o n s h i p s developed to generate the sur-faces of s t a b i l i t y f o r the Fe-Ca-Al-O-S system, u s i n g data from the l i t e r a t u r e ( 1 4 7 ' 1 4 8 ' 2 1 6 ' 2 3 2 ' 2 3 8 ' 2 3 9 > . E q u i l i b r i a ( I ) : Al 20 3/CaO/6Al 20 3-CaO/CaS 1(1): 6 A 1 2 ° 3 + 2 C a 0 + [ S ] = 6Al 20 3-CaO + CaS + [0] A G ° = A G o ^ + A G O A S = B A G - ^ - 2 A G ° a Q = R T l n K I ( 1 ) l f a 6 A l 2 0 3 - C a O ~ a"CaS" ~ a A l 2 0 3 ~ aCaO ~ 1 3 h 0 7.082 X 10 = -RT l n [^] h S -1 h O l n K I ( l ) = " 1 - 9 5 5 K i ( l ) = i ^ i s s x 10 = — • S h Q = 1.4155 X 1 0 _ 1 h g A-I (1) 1(2): 6 A 1 2 ° 3 + 2 f C a ^ + + [S] = 6Al 20 3-CaO + CaS -RT InK = -RT l n [ - 5 - ^ ] = -1.6781 X 10 5 h C a h O h S 1 ?n l n K I ( 2 ) = 4 - 6 3 2 7 x 1 0 t h u s K i ( 2 ) = 1 - 3 1 6 8 x 1 0 h C a h O h S = 7 ' 5 9 4 2 x i O " 2 1 / b Y s u b s t i t u t i n g A - I ( l ) h C a h s = 2.316 X 1 0 " 1 0 A-I(2) 375 1 ( 3 ) : 12 [Al] + 18[0] + CaO + [S] = 6 A l 2 0 3 « C a O + CaS -7.8183 X 10 5 = -RT l n K l ( 3 ) -> l n K I ( 3 ) = 2.158 X 10 2 9 3 K I ( 3 ) = 5.46657 X 10 and by s u b s t i t u t i n g A-I(2) h A l h O = 2 ' 0 1 4 X 1 0 5 A-I(3) E q u i l i b r i a ( I I ) : 6Al 20 3•CaO/CaO/2Al 20 3«CaO/CaS 11(1): |(6A1 20 3-CaO) + y(CaO) + [S] = 2 A l 2 0 3 « C a O + CaS + [0] AG° = AG° + A G ° A S - | ( A G ° A 0 ) - | ( A G ) = 2 D - R T k n K I I ( 1 ) ± f aCaS " a C A 0 " aCaO " a C A c " 1 2 O - R T l n K I I ^ 1 j = 1.4257 X 10 4 + l n K ^ ^ j = - 3.9359 K I I ( D = 1 ' 9 5 3 X 1 0 ~ 3 h 0 -2 thus y- = 1.953 X 10 A - I I ( I ) h s 11(2): j ( 6 A l 2 0 3 - C a O ) + |[Ca] + |[0] + [S] = 2A1 20 3 «CaO + CaS - R T l n K I l ( 2 ) = 1. 3146 X 10 5 -»• l n K I l ( 2 ) 3.629 X 10 1 5 2 K I I ( 2 ) = 4 - 3 1 1 2 x 1 q 1 5 * h c a h O h S = 2 - 3 1 9 5 x i O " 1 6 by s u b s t i t u t i n g A - I I ( l ) h C a h S £ 1 , 6 0 8 X 1 0 ~ 9 A - I K 2 ) 11(3): 4[A1] + 6[0] + [Ca] + [S] = 2Al2<D3-CaO + CaS - R T l n K ( l I ( 3 ) = - 3.11824 X 10 5 -+ l n K I I ( 3 ) = 8 ' 6 0 8 X l o 1 •* K n ( 3 ) = 2 - 4 3 2 4 X 1 0 3 7 by s u b s t i t u t i n g A-II(1) h A l h O = 1- 7 0 4 X 1 0 ~ 5 A - I K 3 ) E q u i l b r i a ( I I I ) : 2Al 20 3-CaO/CaO/Al 20 3•CaO/CaS 1 3 I I I ( l ) : 2(2A1 20 3-CaO) + j CaO + [S] = A l ^ - C a O + CaS + [O] 377 AG° = A G ° A + A G C a S = | ( A G ° A 2 ) - § ( A G ° a 0 ) = - R T l n K I I I ( 1 ) l f aAl 20 3.CaO " aCaS ~~ aCA 2 ~ aCaO " 1 = R T l n K I I l ( 1 ) = 1 . 6 7 6 9 7 X 1 0 4 + L N K I I I ( 1 ) = - 4 . 6 3 h Q = 9 . 7 5 8 X 1 0 " 3 h g A - I I I ( l ) 111(2): y(2Al 20 3-CaO) + |(CaO) + [Ca] + [S] = A l ^ - C a O + CaS - R T l n K I I l ( 2 ) = 7.0675 X 10 4 l n K I I l ( 2 ) = 1.95111 X 10 1 -»• K I I I ( 2 ) = 2.97546 X 10 8 by substituting A - I I I ( l ) hCa hS = 3 , 3 6 X 1 0 ~ 9 A-IIK2) 111(3): 2[A1] + 3[0] + [Ca] + [ s ] + CaO = A l ^ - C a O + CaS 5 - R T l n K I I I ( 3 ) = -1.9366 X 10 ^ l n K I I l ( 3 ) = 5.346 X 10 1 K I I I ( 3 ) = 1-6568 X 10 2 3 378 by s u b s t i t u t i n g A - I I I (2) 2 h A l h O = i - 2 1 5 6 x !0~ 5 A - I I I (3) E q u i l i b r i a ( I V ) : A1 20 3-CaO/CaO(42 .0 wt.%) + A12C>3 58.0 wt. % ( l i q u i d ) / C a S I V ( 1 ) : 2CaO + A l 2 0 3 + [S] = C a O - A l ^ + CaS + [0] - R T l n K I V ( 1 ) = 1.06873 x 10 4 -> l n K I V ( l ) = " 2 - 9 5 8 6 + K I V ( 1 ) = "5.18915 X 10~ 2 u 2 0 a h ~ = CaO a A 1 0 X 5.18915 X l O - 2  S aCaS 2 3 l f aCaO - 0 . 0 6 2 5 ( 2 1 6 ' 2 3 2 ) , a 5 - 0 . 7 ( 2 3 2 ) and a . C a S = 0 . 0 3 5 5 ( 2 3 8 ' 2 3 3 > then h Q = 4.0 X 10~ 3 h g A-IV (1) I V ( 2 ) : 2[Ca] + 2[A1] + 4 [0] + [S] = C a O - A l ^ + CaS -RTlnK = -2.8111 X 10 5 -> K I V ( 2 ) = 5 ' 0 5 3 3 4 x 1 q 3 3 by s u b s t i t u t i n g a = 0.0355 ( 2 3 8 ' 2 3 9 ^ and A-IV (1) (-3 O h A l h O = 1 X 1 0 5 A _ I V ( 2) I V ( 3 ) : CaO + 2[Al] + 3[0] + [Ca] + [S] = C a O - A l ^ + CaS A Gt° = A G C A + A G C a S " A GCaO = - R T l n K I V ( 3 ) i f aCaO = A l 2 ° 3 = 1 KTT7,,> = 1.65287 X 1 0 2 3 = C a S [ IV(3) " — - - i 2 .3. T~-aCaO h A l h O h C a h S by s u b s t i t u t i n g a , a and A-IV (2) L 3 b (_,civj h C a h S = 6 , 8 7 X 1 0 _ 9 A - I V ( 3 ) E q u i l i b r i a (V): CaO ( s )/CaO(57.4 wt.%) + A l ^ (42.6 wt.%) ( l i q u i d ) / C a S ( s ) 380 ± f aCaS B 0.988 - 1 . 0 ( 2 3 8 ' 2 3 9 ) , a A l O , , = 0 . 1 ( 2 3 2 ) and an _ s 0.8 - 0 . 9 ( 2 1 6 ' 2 3 2 ) CaO h C a h s s 1.35 X 10 8 A-V (1) h A l h O = 9.6 X 10 6 A-V (2) hCa hO S 3.29 X 10 1 1 A-V (3) h Q = 2.35 X 10" 3 h s A-V (4) 

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}]}"
                            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.831.1-0078752/manifest

Comment

Related Items