1. Characterization of Inclusions in IF Steels from RH-OB Degasser to Mold
Tsai Hwan-Tang
蔡 煥 堂

2. Purpose of this study
Characterize the in-process steel cleanliness to develop countermeasures to improve nozzle clogging and steel surface quality.

3. Different mechanisms of nozzle clogging have been proposed.
Prior formation and transport
Inclusion formation by deoxidation or reoxidation
Transport of oxides to nozzle
Adherence of oxides to nozzle and to existing build-up
In-situ formation due to cooling

4. Steel grades studied
Grade C Mn P Ti Nb N A
ULC
Added B
ELC D

5. Steel and Slag Sampling Locations
RHOB
1,2,3,4,5,6,7,8
Minutes after Kill
Ladle
Start
Middle
End
Start
Middle
End x
Ladle x x
Well
Box
Well
Start
Middle
End
Mold
Mold x
First three heats of a sequence
Good- and bad-plugging casts
After Cast
Strand 1

6. Outline Indication of origin of plugging inclusion from
Cr2O3 pick-up in tundish slag
Variation of total oxygen content
Shape and distribution of inclusions
Electrochemical method
Remelt button
Shapes
Changes during processing
In Nozzle clogs
Trials with ladle sand with less reducible oxides
Keep initials and names together on a single line

7. Tundish slag picked up chrome oxide.
Pouring box
~5% Cr2O3
Above nozzle well
up to 9% Cr2O3

8. Total oxygen decreased from ladle at the RH-OB to the tundish pouring box.
Otot, Avg. Each Heat Last 2 RH Samples, ppm 40
Higher in Ladle 30 20 10
Higher in Tundish Pour Box 10 20 30 40
Otot, Avg. Each Heat in Tundish Pouring Box, ppm

9. In contrast, total oxygen increased from the tundish pouring box to the tundish well.
Otot, Avg. Each Heat in Tundish Pouring Box, ppm 40
Higher in Tundish Pour Box 30 20 10
Higher in Tundish Well 10 20 30 40
Otot, Avg. Each Heat in Tundish Well, ppm

10. Total oxygen in the tundish well also increased with increasing residence time in the tundish.
Heat Avg. Mean Residence Time, min 11 10 9 8 7 6 5 10 20 30 40
Otot, Avg. Each Heat in Tundish Well, ppm

11. The increase in total oxygen was much greater than the increase in nitrogen.
Otot, Heat Average in Well - Pouring Box, ppm 20
O:N for Air 10
-10
-20
-20
-10 10 20
Ntot, Heat Average in Well - Pouring Box , ppm

12. Total oxygen results suggested oxygen pickup in the tundish by reaction with tundish slag or ladle sand.
Total oxygen
Increased from the tundish pouring box to the tundish well.
Increased more with increasing residence time in the tundish.
Increase was greater than nitrogen increase.

13. Alumina inclusion shape - Electrochemical method

14. Alumina inclusion size - Electrochemical method

15. Alumina mass by inclusion size - Electrochemical method

16. Inclusion Classification
High Surface-Area
Faceted
Spherical
Agglomeration

17. Literature review – Nippon Steel
M. Akiyoshi et al. Nippon Steel Oita R&D (1991)

18. Literature review – Hoogovens (Corus)
Tiekink et al. Hoogovens Ijmuiden (1994)

19. The literature indicates that different alumina inclusions for from different conditions.
High Surface-Area
High super-saturation of O and/or Al
i.e. initial deoxidation or re-oxidation
Faceted
Formation or growth at lower super-saturation
i.e. later deoxidation or cooling
Spherical
'Ripening' of dendrites
Compaction of agglomerated small inclusions
Local chemical variations in steel
Agglomeration
Collection of inclusions by stirring or bubbling

20. SEM Analysis of Inclusions on Remelt Sample

21. High Surface Area Inclusions
Dendritic
O, Al
Starfish
O, Al
Gingerbread
O, Al

22. Faceted Inclusions Flat, Faceted Faceted, < 2 um Globular, Faceted
O, Al
O, Al
Globular, Faceted
> 5 um
O, Al

23. Spherical Inclusions Globular, Non-Faceted > 5 um Smooth Balls
O, Al, Mg
O, Al

24. Agglomeration Inclusions
Coral > 25 um
Fine Coral
O, Al
O, Al
O, Al
Lace Balloon

25. Inclusions - Steel grade & process location
There were no definite differences between grades in inclusion shape or size distribution.
But, there was a remarkable variation of shape and size distributions from ladle to mold.

26. The frequency of small, faceted inclusions (<2 um) was highest at the end of RH-OB treatment.
Pour
Well
Mold
Box

27. The frequency of high surface-area and coral inclusions were highest at RH-OB.
Pour
Well
Mold
Box

28. The frequency of dendritic inclusions at the end of RH-OB treatment increased with decreasing aO at kill.
RHOB aO deox1, ppm
300
250
200 5 10
Avg. No. of Inc's with Sec. Arms in Last 2 RH Samples

29. The frequency of dendritic inclusions increased with increasing oxygen activity in the tundish slag.
%MnO in Well Chamber 8 6 4 2 1 2 3
No. of Inc's with Secondary Arms in Well

30. The frequency of larger globular, faceted inclusions (> 5um) was highest in the tundish.RH
Pour
Well
Mold
Box

31. The frequency of globular, faceted inclusions >5 um decreased from pouring box to well in the tundish.
No. of Globular, Faceted Inc's >5 um in Pour Box
100
More in Tundish Pour Box 80 60 40 20
More in Tundish Well 20 40 60 80
100
No. of Globular, Faceted Inc's >5 um in Well

32. The frequency of globular, faceted inclusions >5 um decreased from the tundish to the mold.
No. of Globular, Faceted Inc's >5 um in Well
100
More in Tundish Pour Well 80 60 40 20
More in Mold 20 40 60 80
100
No. of Globular, Faceted Inc's >5 um in Mold

33. The number of globular faceted inclusions (>5um) increased as tundish superheat decreased.
Tundish Superheat, C 45 40 35 30 25 20 20 40 60 80
100
No. of Globular, Faceted Inc's >5 um in Pour Box

34. The size of globular faceted Inclusions increased during casting.
Percent 50 40
Ladle Start 30
Ladle End 20 10 2 3 4 5 6 7 8 9 10
>10
microns
Globular, Faceted Inclusions in the Pouring Box

35. Globular faceted inclusions > 5um
The results indicate that globular faceted inclusions grew in the ladle by cooling and were removed in the tundish.
Globular faceted inclusions > 5um
were not present in the ladle immediately after killing.
decreased from pour box to well to mold.
Increased during casting.
increased with decreasing superheat.
Globular faceted inclusions got bigger during casting.

36. Analysis of Well Nozzle Plugs
Grade A

37. Analysis of Well Nozzle Plugs – Loose powder

38. Analysis of Well Nozzle Plugs - Boundary between plugged material and steel

39. Analysis of Well Nozzle Plugs - Remelt sample from boundary region

40. The distributions of inclusion types were similar in the tundish well, well nozzle and mold.
No. of Inclusions per Six-Photo Strip
Thousands 10 1
0.1
0.01
0.001
Simplify this legend by putting the 820 in a note
How about brightening the colors?
0.0001
Coral
High Surface
Faceted <2 um
Sphere
Globular Faceted > 5
Well
Nozzle
Mold

41. Relationship of Inclusion Morphology to Clogging
The distribution of inclusion types is similar in the steel and the plugs.
Indicating that plugging comes from inclusions formed by deoxidation or reoxidation before the steel gets to the nozzle.

42. Overall, the results pointed to the reducible ladle sand as a cause of clogging.
Reduction of chrome oxide
Chrome oxide in slag
Chromium Pick-up in steel
Total oxygen
Increased from the pouring box to the well
Increased with longer time in the tundish
Lack of N Pick-up
Dendritic Inclusions
Increased with oxygen activity in the tundish slag

43. Nozzle Clogging Factor (NCF), derived from the slide-gate position, is used to quantify plugging.
Higher is better

44. Ladle sand chemistry Old New Cr2O3 33.1% 16.7% SiO2 29.0% 30.9% Fe2O3
18.7%
9.4%
Al2O3
11.0%
5.9%
MgO
7.2%
3.6%
CaO -
ZrO2
33.3% C

45. Nozzle Clogging Factor
Trial ladle sands with a lower percentage of reducible oxides resulted in less nozzle clogging.
Nozzle Clogging Factor
LCAK
ULC
(Fe, Cr, Si) O
94%
87%
(Cr, Si, Zr) O
96%
91%

46. Conclusions
Inclusion morphology in IF steels ranges from dendritic to globular depending on the degree of super-saturation of Al and O.
Inclusion morphology is similar between grades, but changes significantly from ladle to tundish.
Globular faceted inclusions are the most frequent in the tundish, nozzle clog, and mold.
At all locations, many inclusion forms coexist in the steel.

47. Conclusions All forms of alumina inclusions clog nozzles.
The presence of dendritic inclusions in the tundish indicates either insufficient rinsing or reoxidation.
Increase of total oxygen as tundish residence time increases and as the steel flows from pouring box to well indicated that the tundish design is less optimal and needs improvement.
Ladle sand is a significant factor in nozzle clogging.

48. Acknowledgements
Co-authors: Dr. Howard Pielet of R & D and Mr. Richard Gass of Operating Technology.
Members of the “Inclusion Characterization Team”.
The chemical analysis laboratories of Quality Department.