Boletin 2012 Datos Industriales

Steel Edition
RHI Bulletin >1> 2012

The Journal of Refractory Innovations
CAS-OB Process
DELTEK Eco Gaskets and

Gas Purging Lance Design COMPAC ROX A93MAS-15 Insulation for Flow Control
Optimization Application in CAS-OB Bells Products
RHI Bulletin 1/2012

Steel Edition
Published by: RHI AG, Vienna, Austria

Chief Editor: Bernd Buchberger
Executive Editor: Alexander Maranitsch
Technical Writer: Clare McFarlane
Proofreaders: Bernd Buchberger, Clare McFarlane
Project Manager: Ulla Kuttner
Photography, Graphics
and Production: Christoph Brandner, Stefanie Puschenjak
Design and Typesetting: Universal Druckerei GmbH, Leoben, Austria
Printers: Universal Druckerei GmbH, Leoben, Austria
Contact: Ulla Kuttner

RHI AG, Technology Center
Magnesitstrasse 2
8700 Leoben, Austria
E-mail: [email protected]
Tel: +43 (0) 502 13-5300
Fax: +43 (0) 502 13-5237
www.rhi-ag.com

The products, processes, technologies, or tradenames in the
RHI Bulletin may be the subject of intellectual property rights
held by RHI AG or other companies.
2<
RHI worldwide
New Snorkel Production New Bag Filter Systems and Hardening Grate
Record at RHI’s Dalian Plant for the RHI Hochfilzen Plant
China >> In 2011, the RHI Dalian plant Austria >> In order to achieve future exhaust gas limits at RHI’s
(China) achieved a new plant record of Hochfilzen plant (Austria), the existing rotary kiln exhaust gas treat-
1717 prefabricated RH degasser snorkel ment facility (cyclone separation and gas washing system) will be
pieces. This figure was the result of high upgraded to a bag filter system. Concurrently, the hardening grate
domestic demand as well as increasing unit, which has now reached the end of its service life (built in 1958),
orders from customers worldwide. Prefab- will be replaced and modified for use with a downstream bag dust
ricated snorkels were delivered to the filter. The hardening grate required to harden briquettes will also be
USA, Japan, India, Brazil, and other coun- used to preheat raw magnesite following the conversion.
tries where RHI’s products are successful- The aims of the €8.6 million project are to reduce dust emission to
ly used in various RH degassers. < 10 mg/Nm³ (future BAT limit 20 mg/Nm³ obligatory as of 2013),
The future outlook is also promising; decrease air leakage by approximately 12000 m³/hour by decoupling
driven by an increased demand for high- the hardening grate from the Lepol kiln, reliably maintain low levels
quality steel the use of prefabricated snor- of SO2 in the exhaust gas, utilize waste heat from the kiln exhaust
kels from Dalian will increase accordingly. gas to preheat the raw magnesite, and increase the rotary kiln per-
Therefore RHI is already proactive in pro- formance by approximately 4.5%. Commissioning is scheduled for
viding the necessary capacity expansion. November 2012.
Sales budgets forecast a total turnover of
more than €12 million in 2012 for this
profitable business.
Opening Ceremony for the New Tundish Water
Model Took Place
RHI at AISTech 2012 Austria >> The tundish water modelling facility at the Technology
USA >> Together with approximately Center Leoben (Austria) was recently inaugurated and initial simula-
435 other exhibitors, RHI and INTER- tions have already been executed using scaled customer tundish
STOP took part in the Association for geometries.
Iron and Steel Technology (AISTech) Water modelling will assist in understanding flow phenomena in
2012 conference and exhibition, which the tundish as well as supplement CFD simulations. As an integral
ran from May 7–10 in Atlanta (USA). In part of RHI’s Tundish Technology Solutions, water modelling will
addition to presenting the INTERSTOP serve to optimize existing products and develop new technologies
Metering Nozzle Changer MNC at the dedicated to the increasing demand for clean steel production. The
trade fair, numerous lectures were also overall aim of the tundish water modelling facility is to realize tailor-
given by RHI personnel from both Aus- made solutions for RHI’s customers that meet the quality and safety
tria and the US. requirements.
AISTech is the largest steel trade fair
held in America with more than 6000 vis-
itors recorded at this year’s event.
MARVO Successfully Completes First Turn
around of 2012 at the MiRO Refinery in Karlsruhe
Tailor-Made Tundish Germany >> MARVO GmbH services were provided at the MiRO
ombustion engineering works in the petroleum coke area. Calcined
c
Solutions petroleum coke is processed in the coker on a rotating staged hearth,
Austria >> Once more it has been prov- by extreme heat treatment up to 1400 °C, to produce special con
en that thermochemical simulations are verted coke grades. These high quality calcinate grades are mainly
an extraordinary tool to predict wear required for the industrial production of electrodes used in carbon
phenomena of tundish wear linings. By baking furnaces.
considering the various aspects that lead The soaking pit cone section of this installation was lined with
to premature chemical wear, the linings COMPAC SOL M64COR-6, RESISTAL SK60C, and DIDURIT M60-6 pre-
and boundary conditions of several cus- cast components. The sidewalls and parts of the turntable were
tomers on all continents have been ana- reconstructed using COMPAC SOL M64COR-6 and DIDURIT M60-6
lysed and optimized. Tailor-made tundish precast components.
wear lining mixes adjusted to customer The installation also features rabbles that convey the final pet coke
conditions, including thermochemical from the turntable via the cone into the coke cooler. These rabbles
investigations, are another step forward were relined with DIPLASTIT 259 during the installation’s turnaround,
in RHI’s technology leadership. providing excellent abrasion resistance in this high wear area.
>
> 33
RHI worldwide
RHI’s Largest Fused Record Campaign Life of 1091 Heats in
Magnesia Plant Scheduled 40-Tonne EAF at AML
for Production in October India >> Adhunik Metaliks Ltd., (AML) achieved the highest cam-
paign life of 1091 heats from October 1, 2011, to November 20, 2011,
Norway >> As a major cornerstone of
in their 40-tonne EAF using RHI refractory bricks and monolithics.
the company’s backward integration
The previous average campaign life was 850 heats; however, it was
strategy, RHI purchased the SMA Miner-
extended beyond 100 heats by reengineering the slag conditions
al’s company in Porsgrunn (Norway) in
based on mutual interactions between AML and RHI as well as
2011. Concurrently, it was decided to
through using RHI’s ANKERJET NP12 T gunning mix. The brick
build a new state of the art MgO smelt-
brand installed was ANCARBON F6T10.
er at this location.
RHI have a supply management contract with AML for the EAF. At
By investing in buildings, smelting
the contract startup, an EAF lifetime of around 550 heats had been
furnaces, treatment facilities, and infra-
reached with other suppliers. Currently, the lining installation is
structure, it will be possible in future to
supervised by RHI and the EAF refractory maintenance, namely gun-
produce around 50000 tonnes of the
ning and fettling, is also performed under RHI supervision. On occa-
highest quality fused magnesia annual-
sion, the local RHI India team also provides refractory expertise to
ly, independent of the Chinese raw
improve the EAF performance.
materials market, chiefly for RHI’s own
AML is located near Rourkela in Eastern India and is part of
use.
Adhunik group who are also engaged in the mining and power
The total investment costs for the proj
sectors.
ect are approximately €72.5 million, of
which €9.8 million were spent in 2011.
Test operations will start in stages in
August 2012, with full production Lifetime Record of New EAF Burner Bricks in
scheduled for October 2012.
North America
USA >> RHI recently completed a trial using ANCARBON TB008 in
the high wear burner area of an EAF. The ANCARBON TB008 brick
CEMENTTECH China Hosts replaced the high wearing competitor brick (MgO-C) in this demand-
ing furnace area. The results were spectacular; the newly developed
More Than 400 Exhibitors brick achieved 440 heats with 254–304 mm remaining from the origi-
Including RHI nal brick length of 457 mm, as compared to the former competitor
brick that was normally replaced after around 250–300 heats in this
China >> For the 13th time, CEMENT- high wear zone without any residual thickness. The original trial tar-
TECH (China International Cement get for the high wear zone was 500 heats, so the ANCARBON TB008
Industry Exhibition) was a meeting point will exceed this significantly.
for experts and companies from the The additional good news from this trial is the customer has
Asian region. Held at the Beijing Exhibi- ordered approximately 20 tonnes of this new brick for ongoing
tion Center (China), from March 28–30, installations in each of the two EAF furnaces. The customer has also
2012, this international cement industry requested RHI submit a quote for 609 mm long ANCARBON TB008
trade fair was host to more than 400 bricks for the slag door area, since the customer considers the
exhibitors from mainland China, the ANCARBON TB008 to be perfectly suited for this application.
USA, and Europe and brought together This latest development was especially designed for the high wear
the most advanced international tech- rates in the EAF burner area. It is a further development of success-
nology and equipment. The event was fully implemented grades for the high wear areas in ladles and
visited by more than 10000 people and BOFs, which were introduced on the Brazilian market 3 years ago. To
included topics such as mine explora- outperform a local competitor, RHI developed highly oxidation and
tion, powder processing, cement manu- slag resistant grades based on special antioxidant addition and high
facturing, as well as concrete products quality raw materials. These results provided the basis for the subse-
and their construction. quent development of ANCARBON TB008. To withstand the high oxi-
The RHI stand focused on four major dative attack, special additives were used. On oxygen attack, these
topics: In-house high-grade sinter pro- compounds form liquid phases with MgO or other oxidic compo-
duction (HQM98), established standard nents of the brick and protect the carbon from oxidation by covering
brands (ANKRAL ZC, ANKRAL RC, and the pore surface with a thin film. For further trials, several grades
ANKRAL DC), high-grade refractories have been developed for EAF, BOF, and ladle applications. Already
based on HQM98 (ANKRAL R1, ANKRAL well established and tested grades with these special additives are
R2, and ANKRAL Z1), and new products ANCARBON F1T14B, ANCARBON F3T14B, and ANCARBON F6T14B
such as ANKRAL R8. for ladle and BOF applications.
4<
4 <
Fourth Tunnel Kiln and Addi Production Capacity Increase at RHI’s Trieben
tional Capacity Extensions Plant
at the Dalian Plant Austria >> To meet further organic growth in the nonferrous busi-
ness area, €2 million has been invested in the Trieben plant (Aus-
China >> Owing to strong growth in the
tria). The annual capacity limit of Trieben was approximately 53000
Asia-Pacific region, it is necessary to
tonnes of basic high-fired shaped products, depending on the prod-
increase the production capacity of basic
uct mix. However, by investing in a new press and a modern brick
high-fired bricks at the Dalian plant
milling machine, the production capacity has been increased to
(China) by an additional 35000 tonnes
63000 tonnes per annum. Following test operations in March 2012,
per year. To achieve this requirement,
the new facilities were officially commissioned on April 19, 2012.
€14.7 million has been invested in a
fourth tunnel kiln and additional facilities
for crushing, mixing, pressing, and fin-
ishing. The kiln was fired up at the
beginning of June and production using
RHI’s First Quarter Results for 2012
the new facilities will commence in mid- Austria >> RHI started 2012 with an improved revenues and earn-
July 2012. ings situation in the first quarter: Revenues increased 5.6% to
€436.9 million in the first quarter of 2012, comparable to the equiva-
lent period in 2011. The EBIT of the first quarter increased by 15.1%
to €33.6 million compared with the reference period of 2011 and the
RHI Participates at EBIT margin improved from 7.1% to 7.7%. The net profit even rose
by 74.9% up to €32 million.
ALUMINIUM BRAZIL While sales volume in the Steel Division fell slightly by 1.3% in
comparison with the first quarter of 2011, revenues were up 6.2% as
Brazil >> The nonferrous sector in Brazil
price increases were implemented.
is a very important market for RHI.
Steel EBIT amounted to €12.8 million in the first quarter, after
Therefore, at the recent ALUMINIUM
€6.1 million in the prior-year reference period.
BRAZIL, which ran from April 24–26, in
The sales volume in the Industrial Division dropped 5.9% in com-
Sao Paulo (Brazil), RHI not only had a
parison with the first quarter of 2011 because the cement business
stand at the exposition but also present-
was weaker. The recovery of the markets back to precrisis levels is
ed at the conference. The event, focusing
proceeding, but they still show a highly diverse picture depending
on a wide range of aluminium-associat-
on region and industry segment. Revenues of the Industrial Divi-
ed products and services, was held for
sion, at €143.0 million in the first quarter of 2012, fell slightly short
the first time in Brazil and immediately
of the €144.6 million revenues recorded in the first quarter of 2011.
received international praise.
EBIT amounted to €13.1 million in the first quarter, after €15.8 mil-
lion in the prior-year reference period.
Outlook: In a stable macroeconomic environment and with
unchanged foreign currency exchange rates, RHI expects similar
Second Magnesia Rotary revenue levels for the Steel Division in the second quarter and sig-
Kiln at RHI’s Eskisehir Plant nificantly higher revenues in the Industrial Division. Price increases
and the cost cutting programme initiated in 2012 in combination
Turkey >> China’s export policy, com- with a positive contribution to earnings of the higher level of back-
bined with a high demand for magnesia, ward integration leads RHI to expect a higher margin for the entire
is leading to price increases and the year 2012 than in the past financial year.
occasional shortage of high-quality mag-
nesia. To alleviate this scenario, RHI is
expanding its own production of sintered
magnesia in Turkey. A second rotary kiln Rotary Kiln Preheater Filter at Breitenau Will
at Magnesit Anonim Sirketi (MAS) in
Eskisehir (Turkey) will enable the addi-
Provide Enviromental Benefits
tional production of approximately 76000 Austria >> At RHI’s raw material and production plant in Breitenau
tonnes of sintered magnesia per annum (Austria) the existing electrostatic precipitator in rotary kiln 3 will be
and decrease the need to purchase this replaced with a bag dust filter. In addition, a raw magnesite preheat-
material at expensive prices. er will be installed prior to the filter in order to recover waste heat
The total investment costs for this new and enable the bag filter system to function.
rotary kiln facility are approximately €19 The total project costs are €3.5 million, of which €0.6 million
million, of which €6.14 million were were spent in 2011. The project aims are to reduce dust emission to
spent in 2011. An additional €4.75 mil- < 10 mg/Nm³ (future BAT limit 20 mg/Nm³ obligatory as of 2013), use
lion is estimated for raw magnesite sup- waste heat in order to increase energy efficiency, preheat the raw
ply. The test operation will start in magnesite and save approximately 2000000 Nm³ of natural gas per
August 2012, with full production annum (corresponding to 4000 tonnes of CO2), as well as decrease
planned for September 1, 2012. NOx emissions through primary measures.
>5
RHI worldwide
RHI Publishes First Sustaina EBT Taphole Lifetime Increased With
bility Report SYNCARBON TB028
Austria >> RHI has published its first sus- SYNCARBON TB028 is a new brand for EAFs, developed to with-
tainability report according to the report- stand the high wear rates in EBT tapholes. The carbon-bonded grade
ing standards of the Global Reporting Initi- is based on high-quality MgO and graphite in combination with spe-
ative (GRI), thereby taking a major step cial antioxidants. Whilst the addition of antioxidants is a well-known
towards systematically dealing with sus- practice to increase the oxidation behaviour of resin-bonded bricks,
tainability. The report titled “We write sus- it hadn’t previously been applied to such brick types due to the good
tainable (hi)stories” contains comprehen- intrinsic properties provided by carbon bonding. However, especially
sive data and facts on good corporate gov- for EBT taphole applications, the use of antioxidants provides advan-
ernance, product responsibility, environ- tages such as increased bonding strength and oxidation resistance.
ment and energy, employees, health and Further improvements to the brick properties were achieved by
safety, and social responsibility as well as impregnation to reduce pore volume and increase the carbon yield
targets for the coming years. An electronic after coking (during operation). This impregnation also improves the
version of this report is available on RHI’s carbon matrix of the entire brick. A new environmentally friendly
website www.rhi-ag.com at Group/Sustain- carbon binder was used for the carbon bonding and impregnation.
ability. The results of recent trials at three customers have confirmed the
RHI will publish a sustainability report- benefits of this brand. At Ferrostal Labedy Sp.z o.o.(Poland) the
ing according to GRI on an annual basis in standard lifetime of the cylindrical design EBT taphole was ~ 120
the future, in order to regularly report on heats, which was increased to ~ 170 heats after implementing a con-
trends, developments, and achievements. ical EBT taphole made from standard grades. However, a further life-
time increase to ~ 205 heats and a new EBT lifetime record was
achieved using SYNCARBON TB028. An EBT lifetime record was
also realized at Stahl Gerlafingen AG (Switzerland) where the num-
RHI Provides the Main Spon ber of heats with the conical EBT taphole was increased to ~ 200
with SYNCARBON TB028 from 130 with the standard conical EBT. In
sorship for MagMin 2012 in addition, a new EAF vessel lifetime record of 628 heats was achieved
Salzburg with SYNCARBON TB028 in the EAF slag zone at Elektrostahlwerke
Gröditz GmbH (Germany), where previously the average lifetime of
Austria >> The most important conference the EAF vessel had been approximately 500 heats.
for the magnesia industry, the Magnesia
Minerals Conference (MagMin), took place
from May 14–16, 2012, in Salzburg (Austria).
This annual global conference brings
together around 200 producers, dealers,
Nonferrous Metal Topics Presented at The
buyers, and other partners linked with the Minerals, Metals and Materials Society Con
magnesia industry in a setting focused on
speeches, panel discussions, field trips, and
ference
networking opportunities. USA >> The 141st TMS Annual Meeting and Exhibition took place at
This year Salzburg was chosen as the the Swan and Dolphin Hotel Resort in Orlando, Florida (USA). More
conference venue and with its long-estab- than 4000 of the world’s top materials science and engineering pro-
lished presence in the area, RHI was delight- fessionals participated in this event from March 11–15, 2012. RHI
ed to act as the principal sponsor. presented three technical topics during the conference: High-perfor-
On May 14, a field trip provided the mance brands for the nonferrous metals industry, slide gate systems
opportunity for delegates to visit RHI’s plant for copper tapping, and the chemical wear of basic brick linings in
in Hochfilzen, where alpine magnesite is the nonferrous industry.
mined and processed into high-quality sin- The main interest for RHI, in addition to the light metal processing
ter. More than 60 participants toured the of aluminium, centred on the event “International Smelting Technol-
mining and production facilities where ogy Symposium: Incorporating the 6th Advances in Sulfide Smelting
refractory mixes for the steel industry are Symposium”. Many of the participants are very well known to RHI
manufactured. as they are part of the customer base (e.g., Boliden, Umicore, Cam
Board Member Manfred Hödl officially pine, Metallo, Vale, Xstrata, Stillwater, Atlantic Copper, KCM, Mopa-
opened MagMin on May 15 with his wel- ni, and Eramet) or OEMs (Outotec, Xstrata Technologies, Mintek,
coming speech and outlined in his presenta- ANDRITZ Maerz, Kumera, Pyromet, Hatch, and SNC-Lavalin) RHI is
tion the strategic focus of RHI’s backward working with during daily business.
integration strategy, including the business RHI was also represented at the TMS 2012 Exhibition along with
rationale behind the two recent raw materi- approximately 100 different technical and analytical companies
als acquisitions in Ireland and Norway. working in the pyrometallurgical processing and mining industry.
6<
6 <
Editorial Contents
Sustainability has always been integral in RHI’s approach to 8 Comparison of Basic Oxygen Furnace
business, taking long-term responsibility for environmental, Bottom Gas Purging Options
economic, and social activities at a global level. However, in
recent months sustainability management has been restructured
at the company, with Management Board members strategically
16 New Oxycarbide Refractory Products
engaged in sustainable value creation. At a time when raw mate- Demonstrate Outstanding Properties—
rial availability and continually rising costs of raw materials, First Practical Results
energy, and climate control have such a significant impact, sus-
tainability at RHI is focusing on resource and energy efficiency as 20 Customer-Specific Analysis of
well as health, safety, and talent management, as exemplified in Steelmaking Slags to Provide Process
the first annual sustainability report published in April 2012. and Refractory Lining Lifetime
Improvements in Steel Treatment Ladles
In this edition of the Bulletin a number of papers describe RHI’s
and EAFs
direct commitment to sustainability including contributions to
resource efficiency in the context of European policy initiatives.
RHI’s proactive measures to address health and safety concerns 26 Gas Purging Lances: Improving
regarding certain ceramic mineral fibres used for high-tempera- Established Technology
ture insulation are also detailed in an article describing REACH
legislation. 34 Microscopic Examination of Premature
Wear Caused by Joint Opening and
Many of the additional articles highlight product developments Vertical Crack Formation in Magnesia-
and system improvements that can reduce specific refractory con-
Carbon Steel Treatment Ladle Linings
sumption as well as provide energy savings. For example a new
oxycarbide refractory material is introduced that demonstrates
excellent material properties including chemical and thermal 39 Thermomechanical Steel Ladle
shock resistance. The first trial results illustrate how the lifetime Simulation Including a Mohr-Coulomb
of CAS-OB bells can be doubled using this refractory, which is Plasticity Failure Model
also suitable for various steel treatment, hot metal, and foundry
applications. In a paper detailing customer-specific analyses of 44 Consequences of REACH on the Use of
steelmaking slags, various tools are discussed that enable the Ceramic Mineral Fibres
slag composition to be optimized, improving both lining lifetimes
and metallurgical processes. Further papers describe improve-
ments to gas purging lances, the development of a dynamic
50 Resource Efficiency—Global Context,
refractory wear test to improve quantitative evaluation of refrac- European Policy Initiatives, and RHI’s
tory dissolution, and a comprehensive overview of gas bottom Responses
purging in BOFs.
55 Full Integration of INTERSTOP Flow
Innovation was recognized by the European Commission as an Control Technology into RHI
essential precondition to improve resource efficiency and sustain-
able raw material supply. At RHI the “Power of Innovation” has
58 Dynamic Refractory Wear Test Method
been pivotal in the corporate strategy for many years and I hope
the Bulletin provides a forum in which the advances realized
for Magnesia-Carbon Products
through this approach, including those directly relating to
resource efficiency, can reach a wide audience.
In closing, I would like to thank all the authors involved in this

edition, many who regularly take time to write articles for the
Bulletin. I am also very grateful to the editorial team members, Subscription Service
whose continued commitment make this publication possible.
and Contributions
We encourage you, our customers and inter-
Yours sincerely ested readers, to relay your comments, feed-
back, and suggestions to improve the publica-
Bernd Buchberger tion quality using the contact details below.
Corporate Research and Development Furthermore, to receive the RHI Bulletin free of
RHI AG charge please e-mail or fax your details to the
Subscription Service using the form on the
back page.
E-mail: [email protected]
Phone: +43 (0) 502 13-5300
Fax: +43 (0) 502 13-5237
>7
RHI Bulletin > 1 > 2012, pp. 8–15
Thomas Kollmann, Christoph Jandl, Johannes Schenk, Herbert Mizelli, Wolfgang Höfer, Andreas Viertauer and Martin
Hiebler
Comparison of Basic Oxygen Furnace Bottom

Gas Purging Options
Introduction process technology (using only a top blowing oxygen lance)
to a process operating with a top blowing oxygen lance in
A higher level of product sophistication (e.g., clean steel,
combination with a bottom inert gas purging system (Fig-
interstitial-free, and ultra low carbon steel grades) and ure 1) [1–3].
unstable charging materials—dependent on the raw mate-
rial situation (e.g., availability and fluctuating prices)— Worldwide, different BOF philosophies (Figure 2) are in
require an economically optimized BOF process operation. operation using different bottom gas purging plug types,
In the early 1980s most of the steel plants, especially in arrangements, blowing practices, flow rate regulation sys-
Europe, made a decision to switch from the original LD tems, and patterns.
150 100
n No. of steel plants using

120 specific process 80
n Cumulative share
Cumulative share [%]

No. of steel plants
90 60
60 40
30 20
0 0
LD LD-BS LD-OB LD-OB KOBM OBM
Ar/N2 O2 /CO2 O2 /Cn Hm O2 /Cn Hm O2 /Cn Hm
Figure 1. Variety and application frequency of oxygen steelmaking processes worldwide [4]. Abbreviations include Linz-Donawitz (LD),
Linz-Donawitz bottom stirring (LD-BS), Linz-Donawitz oxygen bottom (Nippon Steel) (LD-OB), Klöckner oxygen bottom Maxhütte
(KOBM), and oxygen bottom Maxhütte (OBM).
Top only Soft Strong Combined Bottom only

Oxygen lance Oxygen lance Oxygen lance Oxygen lance
N2 N2 Hydrocarbon Hydrocarbon
Ar Ar Oxygen Oxygen
Top-blown Top lance plus Top lance plus Top lance plus Bottom-blown
(BOF) process permeable elements in bottom uncooled bottom tuyeres cooled bottom tuyeres (OBM or Q-BOP) process
Figure 2. Oxygen steelmaking processes [5]. Abbreviations include oxygen bottom Maxhütte (OBM), which is equivalent to Q-BOP.
8<
RHI Bulletin > 1 > 2012
Benefits of Bottom Gas Purging Argon and nitrogen are used as inert bottom purging gases.
Inert in this case means that no (i.e., argon) or hardly any
The internal motivation to install bottom gas purging sys-
(i.e., nitrogen) reaction with other dissolved elements in the
tems was nearly identical all over the world: The fundamen-
steel bath takes place even at the highest temperatures.
tal reasons were to improve metallurgical results and guar-
antee a highly effective and efficient oxygen steel produc-
tion at the lowest costs (Figure 3) [6–8]. Benefits Benefits in detail
High quality and economical steel >> Minimization of the tap-to-tap time
The common benefits of vessel bottom purging are listed in
production >> Reduction of the re-blow rate
Table I. By enhancing mass and heat transfer, the gas purg- numbers
ing system influences the equilibrium conditions in the steel >> Lower (Fet), [P] levels, and [Mn]
bath during the refining process enabling the system to oxidation loss
approach equilibrium at the end of blowing. As a result Realization of lower [C] x [O] levels/pCO >> Less deoxidation agents (e.g., Al)
decarburization and dephosphorization are considerably values are required
improved. Table II shows a detailed overview of the realized >> Minimization of the RH degassing
operation (cost saving)
metallurgical results with a bottom gas purging system
compared to the original LD process without bottom gas Improved steel bath homogenization/ >> Shorter and quicker reaction path-
purging [12,13]. kinetic and temperature distribution ways between the slag and steel
bath (better conditions for scrap/flux
additive melting, and higher scrap/
Influence of Gas Type and Purging Rate hot metal ratio)
>> Improved process control (higher
The indicator for an efficient gas purging performance is accuracy of the tapping temperature
the product of the dissolved carbon [C] and oxygen [O]. Due and element levels)
to the purging plug availability, inert gas supply, and plug >> Improved steel yield and flux addi-
regulation system (linked to the set flow rate patterns), tive levels (reduced slag volume and
[C] x [O] levels < 25 x 10-4 are realized without any problems slopping material)
(Figure 4) [14,15].
Table I. General benefits of gas bottom purging [9–11]
Parameter With bottom Without bottom

Cost savings gas purging gas purging
(Fet) in slag (wt.%) 18–20 > 20
[C] at end of blowing (ppm) 300–400 > 400
Bottom Optimization of [O] at end of blowing (ppm) 500–650 > 650
purging BOF process
[P] at end of blowing (ppm) 60–120 > 120
Aluminium consumption for 1.5–2 >2
deoxidation (kg/tonne)
Enhanced
productivity Re-blow rate (%) 10–18 > 18
Tap-to-tap time (min) 30–35 > 35
Figure 3. Advantages of BOF bottom purging. Table II. Metallurgical benefits of bottom gas purging.
1600 pCO
0.5 1.0 1.5 n Without bottom purging
1400 n With bottom purging
1200
1000
Oxygen [ppm]
800
600
400 [C] x [O]

37.5
200 25.0
12.5
0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14
Carbon [%]
Figure 4. Comparison of carbon and oxygen content at the end of blowing with and without bottom gas purging [15].
>9
The type of inert gas used depends on the critical nitrogen tonne x minute, respectively. This analysis indicates that
level for a particular steel grade because there is an increas- purging with argon during the initial refining phase pro-
ing level of nitrogen pick-up as the blowing progresses vides no metallurgical benefits and should be avoided since
when purging is performed with nitrogen. However, nitro- it is four times more expensive than nitrogen. However,
gen levels at tapping can be flexibly adjusted during the when aiming for the lowest nitrogen levels it is necessary to
refining process by shifting the point of switching from switch from nitrogen to argon at between 25–50% of the
nitrogen to argon and by controlling the specific nitrogen blowing time. A retarded switching point, especially at over
purging flow rate. 50% of refining stage, causes very high [N] levels at tapping
whilst higher nitrogen purging intensities also increase the
Figure 5 demonstrates that the gas type and nitrogen purg- final nitrogen content [16–18].
ing intensity used during the first 25% of the blow does not
influence the final [N] levels in the steel, since all the purg- Purging Plugs—Types and Installation
ing conditions examined resulted in a final value of 20 ppm. Arrangement
Furthermore, the influence of the nitrogen purging rate
The bottom gas purging system consists of different num-
even up to 50% through the blow has a minimal effect on
bers and types of plugs in a defined plug arrangement (i.e.,
the final steel [N] levels (22–25 ppm for nitrogen flow rates
elliptical, rectangular, and circular). Furthermore, the gas
between 0.02–0.1 Nm3/(tonne x minute), respectively). In
contrast, the nitrogen pick-up increases considerably in the purging system includes a level 1 and 2 automation and a
second half of the blow, with the final [N] values ranging purging plug valve regulation station. Level 1 includes digi-
between 32–48 ppm for nitrogen flow rates of 0.02–0.1 Nm3/ tal systems for measurement, control, and gas regulation.
Level 2 is the operating unit and regulates parameters, such
as flow rates and purging gas switching points, individually
for the different steel grades produced. In most instances,
60 the purging plug regulation systems are based on a total
Purging gas and rate flow rate regulation, which means the total set flow rate is
n N2 0.10 Nm3/(tonne x minute) distributed uniformly to the number of installed purging
n N2 0.06 Nm3/(tonne x minute) plugs. Moreover, each plug can be regulated separately
50
n N2 0.02 Nm3/(tonne x minute)
(typically only in manual mode) and the total set point flow
[N] after end of blow [ppm]
n Ar
rate in the individual process steps is kept constant, using
40 pressure regulation systems.
Commonly, there are three different bottom gas purging

30 plug types in operation: Multihole plugs (MHPs), single hole
plugs (SHPs), or tuyeres (Figure 6). The MHP is state of the
art and primarily chosen by steel plants using BOF bottom
20 gas purging. An overview of the different purging plug
characteristics is provided in Table III.
10 Purging Plug Features and Installation Require-

0 25 50 75 100
ments
Blowing time [%]
The MHP is characterized by a lower plug blocking poten-
Figure 5. Influence of gas type and nitrogen gas purging rate on tial, reduced infiltration affinity, and better purging plug
the final [N] levels in steel [16]. availability because MHPs reopen during the campaign.
(a) (b) (c)
Figure 6. Common purging plug types: (a) MHP, (b) SHP, and (c) tuyere.
10 <
Parameter MHP SHP Tuyere

Relative price Medium Low High
Bubble characteristics Well distributed small bubbles Ineffectively distributed large bubbles Ineffectively distributed large bubbles
Breakthrough safety High Low Low
Blocking Likely to reopen Likely to remain blocked Likely to reopen
Average flow rate range per plug (l/min) 200–1400 200–1200 2000–3500
Average total flow rate consumption 0.8–1.2 0.8–1.2 > 1.5
per heat (Nm3/t)
Average pipe diameter range (mm) 1–2 4–8 1.5–3.8
Number of pipes per plug 12, 24, 32 1 1
Open gas section per plug (mm )
2
9.4–100.5 12.6–50.3 100–120
Average wear rate (mm/heat) 0.40 0.42 0.40–0.45
Number of plugs per vessel 8–12 8–13 3–4
Additional information >> Less plug blocking potential >> Economically priced >> Defined drilling positions
>> Less infiltration affinity >> Increased plug blocking potential >> Complicated installation procedure
>> Reopening during a campaign during a campaign >> Installation during campaign startup
(purging availability increased) >> Installation during the relining period
>> Installation during relining procedure procedure >> N o purging availability at campaign
start (installation after 50–100 lining
heats)
>> P oor bath agitation caused by very
high flow rates (jetting)
Table III. Characteristics of MHPs, SHPs, and tuyeres.
To realize good bath kinetics, the aim is to achieve small circular tuyere arrangement (starting with four tuyeres) with
bubbles with a long dwell period in the liquid steel bath defined positions for the second drilling during the cam-
while jetting should be avoided. Steel plants that operate paign is depicted in Figure 7.
on the tuyere philosophy have the opportunity to drill and
set new tuyeres during a campaign. The function of the tuy- Commonly, 8–13 purging plugs (i.e., MHPs or SHPs) are
eres is nearly identically to a SHP and the tuyeres are installed per vessel, set with an average total purging inten-
installed at defined drilled bottom positions. These posi- sity of 1 Nm3/tonne per heat. The high gas flow rates
tions are preset by the gas connection points on the steel through the individual tuyeres result from the very low
shell bottom. Since the bottom lining moves during heat number of tuyeres, typically three or four per vessel, and
up, as a result of thermal expansion, the bottom purging guarantee purging through any slag layer present on the
system is activated (i.e., drilled) after 50–100 lining heats. vessel bottom, irrespective of its thickness. As a result jet-
However, installation takes several hours to complete per ting can occur, leading to poor bath kinetics and poorer
tuyere, with associated production loss and vessel cooling. metallurgical results in comparison to the outcome
Typically, the implemented tuyeres are set at very high flow achieved with SHPs or MHPs. The average [C] x [O] levels
rates using three to four tuyeres per vessel in combination obtained with various purging plug types are shown in
with an excessive slag splashing practice. An example of a Figure 8.
35.0
Positions for second drill 33.0
30.0
24.5
Average [C] x [O] level x 10-4
25.0
22.0
20.0
15.0
10.0
5.0
0.0
MHP SHP Tuyere
Purging plug types
Figure 7. Concept for tuyere installation. Figure 8. Average [C] x [O] levels achieved with different purging
plug types.
> 11
Typically, lower [C] x [O] levels are achieved with MHPs activation of the bottom gas purging system is limited by
when compared to steel plants operating with SHPs and wear. Most commonly premature wear is visible especially
tuyeres. As described, 8–13 purging plugs are installed in a in areas surrounding the plug. The bottom wear is influ-
circular, rectangular, or elliptical arrangement prior to the enced by the following parameters [19,20]:
vessel campaign start. The majority of the steel plants
adopt an elliptical bottom purging plug configuration (Fig- >> Bottom maintenance philosophy.
ure 9). >> Inert gas purity (primarily the {O2} level).
>> Tapping temperature.
Factors Influencing Purging Plug Availability >> Lining concept (quality and initial brick length).
During tapping and sampling (manually with a lance), direct

Maintenance Strategy
contact between the purging plug elements and the liquid
steel bath can be avoided. As a result the plug blocking To stabilize wear, slag splashing or coating are imple-
potential is minimized and a higher gas purging efficiency mented as bottom maintenance philosophies. However,
is achieved. On average, 50% of the total installed plugs are due to very thick or too sticky slag layers (related to the
required to realize the aimed metallurgical results using (MgO) level in the slag) in combination with very intense
SHPs or MHPs. However, it has to be considered that bottom maintenance or discontinuous production, the
Quality Quality
(a) (b)
Figure 9. (a) MHPs and (b) SHPs installed in an elliptical arrangement.
Inert gas distribution
Slag
Liquid steel
N2 /Ar N2 /Ar N2 /Ar N2 /Ar
Slag coating caused by
(a) slag splashing (b)
Figure 10. Inert purging gas distribution influenced by slag coating. (a) thick slag layer formed over the bottom and (b) thick slag layer
extending across the bottom and up the vessel walls.
12 <
bottom purging elements may become blocked and in the increase considerably, becoming close to the range
worst case they never reopen (bottom build up). If the slag detected when operating only with a top blowing lance.
layer formed is more than 50–100 mm, effective gas purg- Furthermore, the effect leads to unstable [C] x [O] levels
ing is not possible. As a result the inert gas diffuses during the vessel campaign. The influence of the slag
between the lining and the slag layer along the barrel to splashing rate on the obtained average [C] x [O] levels is
the vessel’s upper cone or mouth. The purging gas listed in Table IV. It is evident that an increase in the slag
streaming, dependent on the slag layer build up, is pic- splashing rate corresponds with a simultaneous rise of the
tured in Figure 10. This type of phenomenon has been average [C] x [O] levels.
seen and verified using natural gas, identifiable by a flame
(combustion reaction), which was detected coming out of
the areas described. Slag splashing rate (%) Average [C] x [O] level range (10-4)
10–15 20–26
Remedies to counteract this phenomenon include:
20 25–28
>> Immediately stopping slag coating until the plugs are 40 30–33
visibly open again.
Table IV. Influence of the slag splashing rate on the average
>> Bottom burning with an oxygen lance using hot metal
[C] x [O] levels.
or heating agents such as coke or FeSi to free the bot-
tom of the solidified slag layer.
For a more detailed understanding of this phenomenon,
Furthermore, the level of bottom gas purging availability is three different bottom maintenance strategies and their
limited by the slag layer that has formed (i.e, height and influence on the [C] x [O] levels were investigated including
consistency) and the slag splashing frequency. If the bot- the lower and upper [C] x [O] levels and their average
tom is completely covered with slag, the [C] x [O] levels course during a campaign (Figure 11).
Shutdown of the bottom gas

purging system between 1500
and 1700 lining heats
40 40
35 35
30 30
[C] x [O] level x 10-4
25 25
20 20
15 15
10 10
5 5
0 0
0 500 1000 1500 2000 2500 3000 3500 4000 0 500 1000 1500 2000 2500 3000 3500 4000
(a) Lining heats without bottom maintenance (b) Lining heats with 10–15% slag splashing rate
Shutdown of the bottom gas purging system

between 3000 and 4000 lining heats
40
35
30
25
20
15
10
0
0 500 1000 1500 2000 2500 3000 3500 4000
(c) Lining heats with > 60% slag splashing rate
Figure 11. Relationship between the lining maintenance strategy and the [C] x [O] levels. (a) without bottom maintenance, (b) 10–15%
slag splashing during the entire campaign, and (c) > 60% slag splashing when the bottom gas purging system was activated followed
by 100% slag splashing when bottom gas purging had shutdown.
> 13
The three cases were: 200 mm. The initial height of the implemented bottom gas
purging plugs is influenced by the BOF bottom design, ves-
>> Without bottom maintenance. sel capacity, and the installed purging plug type (production
>> With slag splashing (rate between 10–15%) during the length limitation of the brick press). Bottom bricks are man-
entire campaign. ufactured from MgO-C brands and contain 10 or 14 wt.%
>> With slag splashing (rate of > 60%) when the bottom gas <C> (residual carbon) with an initial length between 800–
purging system was activated and a 100% slag splashing 1200 mm. Two different philosophies for the bottom brick
rate to achieve the highest vessel lifetimes after the bot- lining design are in operation:
tom gas purging had shutdown.
>> Using the same quality material for the areas surround-
Without bottom maintenance it was observed that the [C] x ing the plug and the rest of the bottom.
[O] levels were in a range between 15–27 x 10-4 while the bot- >> Using a different quality material for the areas surround-
tom gas purging system was activated. After the bottom gas ing the plug and the rest of the bottom (higher <C> lev-
purging system had been shutdown due to bottom premature els in the surrounding plug areas).
wear, the values drifted to levels of 30–37 x 10-4. A slag splash-
ing rate between 10–15% resulted in higher vessel lifetimes The advantage of using lower <C> levels in the bricks sur-
and slightly increased average [C] x [O] levels and ranges, rounding the plugs is:
compared to gas purging with no bottom maintenance, as a
result of plug blocking and wear; however, the upper [C] x [O] >> An increase in the wettability that leads to better condi-
levels were not as high as those detected when no bottom gas tions for slag adherence (slag coating /splashing).
purging system was operational. For example, at advanced
vessel lifetimes, the [C] x [O] values tended to the upper limit Whilst the advantages of using higher <C> levels in the sur-
of more than 25 x 10-4. Using a slag splashing practice of 60% rounding bricks include:
corresponded to a very wide range of [C] x [O] levels between
20–37 x 10-4 from the initial stage of the campaign life to the >> Better thermal conductivity.
end of the bottom gas purging system activation. Further- >> More resistant to thermal stress.
more, from a metallurgical point of view, the process was very
unstable leading to potentially very high re-blow numbers and In addition, the wear rate of the plug and surrounding area
rising metallurgical treatment times and costs for secondary is about 0.1 mm/heat lower when the area surrounding the
metallurgy during the campaign period. Therefore, a consist- plug contains higher <C> levels than it is for bottom lining
ently reliable bottom gas purging efficiency (< 25 x 10-4) and designs where the same grade is used for the surrounding
plug availability was not achievable with this maintenance area and plug (Figure 12).
strategy. Finally, the bottom gas purging system was shut-
down after 3000 and 3500 lining heats due to premature bot- Outlook
tom wear. Afterwards an intensive slag splashing programme
In the future, a purging plug should provide very high inert
was carried out (rate of 100%) aiming for vessel lifetimes of
gas purging availability during the entire vessel lifetime and
more than 10000 heats per campaign. During this stage the
achieve average [C] x [O] levels between 20–25 x 10-4. The
[C] x [O] levels exceeded 30 x 10-4.
goal of steel plants to increase vessel lifetimes whilst lower-
ing maintenance practices and costs has demanded purging
Potential Plug Lifetime
plugs with reduced wear rates. Figure 13 demonstrates the
The critical plug thickness for closing is defined differently relationship between the calculated number of achievable
for each steel plant and ranges from nearly zero to about heats per campaign and the initial plug brick length for
0.6 8000
0.54 Wear rate [mm/heat]
Average wear of plug and surrounding area
7000 n 0.18
0.5 D 0.1 n 0.25
0.44 n 0.40
6000
0.4
Achievable heats
5000
[mm/heat]
0.3
4000
0.2 3000
2000
0.1
1000
0.0
Different material for plugs Same material for plugs 0
and surrounding bricks and surrounding bricks 500 700 900 1100 1300 1500
(higher <C>) Initial plug length [mm]
Figure 12. Comparision of plug and surrounding brick wear rate Figure 13. Influence of the initial plug length and plug wear rate
when the same or different material is used for the plug and sur- on the number of achievable heats.
rounding bricks.
14 <
three different wear rates. If the aim is 5000 heats per cam-
paign (critical residual brick thickness of 100 mm for plug
closing), the plug wear has to be 0.18 mm per heat with an
initial length of 1000 mm. Currently, the average wear rates
are in the range of 0.25–0.45 mm/heat. Therefore, RHI is
focused on developing a new generation of purging plugs
in the next few years that meet the requirements of steel
plant customers.
References
[1] Kreulitsch, H., Krieger, W., Antlinger, K. and Jungreithmeier, A. Der LD-Prozesse - ein ökologisch optimiertes Verfahren. Neue Hütte. 1992, 37,
313–321.
[2] Kohtani, T., Kudou K., Murakami, S., Okimori., M., Nakajima, M. and Aoki, H. On the Metallurgical and Blowing Characteristics of the LD-OB Pro-
cess. Iron and Steelmaker. 1982, 9, No. 12, 28–33.
[3] Wallner, F. and Fritz, E. Fifty Years of Oxygen-Converter Steelmaking. Metallurgical Plant and Technology International. 2002, 6, 38–43.
[4] Hüsken R., Fechner, R. and Cappel, J. Use of Hot Metal With High Phosphorus Content in Combined Blowing BOF Converters. Iron and Steel
Technology. 2011, 8, No. 11, 46–58.
[5] Fruehan, R. (Ed) The Making, Shaping and Treating of Steel: Volume 1 - Steelmaking and Refining. 11th edition; AIST Publications: Warrendale,
1998.
[6] Cappel, J. and Wünnenbeg, K. Cost-Saving Operation and Optimization on Metallurgical Reactions in BOF Practice. Iron and Steel Technology.
2008, 5, No. 11, 66–73.
[7] Cappel, J. and Wünnenberg, K. Kostengünstige Arbeitsweise und optimierte metallurgische Reaktionen beim Sauerstoffaufblasverfahren. Stahl
und Eisen. 1988, 128, No. 9, 55–66.
[8] Bruckhaus, R. and Lachmund, H. Stirring Strategy to Meet the Highest Metallurgical Requirements in the BOF Process. Iron and Steel Techno
logy. 2007, 4, No. 11, 44–50.
[9] Krieger, W., Hubner, F., Patuzzi, A. and Apfolterer, R. LD-Prozess mit Bodenspülung – Maßnahmen, Möglichkeiten, Ergebnisse. Stahl und Eisen.
1985, 105, No. 12, 673–678.
[10] Fiege, L., Schiel, V., Schröer, H., Weber, L. and Delhey, H-M. Einfluss des Bodenspülens auf die metallurgischen Ergebnisse in den LD-Stahlwerk-
en der Krupp Stahl AG. Stahl und Eisen.1983, 103, No. 4, 159–164.
[11] Krieger, W. and Poferl, G. Metallurgische und betriebliche Vorteile des LD-Prozesses mit Bodenspülung. Weiterbildungsunterlagen VOEST, Linz,
1982.
[12] Gudenau, H. Praktikum zur Metallurgie, RWTH Aachen, Germany, 2002.
[13] Chigwedu, C., Kempken, J. and Pluschkell, W. A New Approach for Dynamic Simulation of the BOF Process. Stahl and Eisen. 2006, 126, No. 12,
25–31.
[14] Schoeman, E., Wagner, A., Ebner, A. and Berger, M. Implementation of Basic Oxygen Furnace Bottom Purging at Mittal Steel Newcastle. RHI
Bulletin. 2006, No. 2, 7–11.
[15] Kollmann, T. Influence of Bottom Purging on the Metallurgical Results, Master’s Thesis, University of Leoben, Austria, 2010.
[16] Hiebler, H. and Krieger, W. Metallurgie des LD-Prozesses. BHM. 1992, 137, 256–262.
[17] Selines, R. Selection of Stirring and Shrouding Gases for Steelmaking Applications, Union Carbide Cooperation, New York, 1988.
http://www.praxair.com/praxair.nsf/0/FC4072B3D78AB3B5852573A8006EDB4A/$file/StirringandShroudingGases.pdf
[18] Genma, N., Soejima, T., Kobayashi, J., Matsumoto, H., Matsui, H. and Fujimoto, H. Application of CO as Bottom Stirring Gas in Combined Blown
Converter. Presented at 110th ISIJ Meeting, Niigata University, Japan, October 1985, Lecture No. S989.
[19] Messina, C. Slag Splashing in the BOF- Worldwide Status, Practise and Results. Iron and Steel Engineer. 1996, 73, 17–19.
[20] Mills, K., Su, Y., Fox, A., Li, Z., Thackray, H. and Tsai, H. A Review of Slag Splashing, ISIJ International, 2005, 45, No. 5, 619–633.
Authors
Thomas Kollmann, RHI AG, Steel Division, Mülheim-Kärlich, Germany.
Christoph Jandl, RHI AG, Steel Division, Vienna, Austria.
Johannes Schenk, Chair of Metallurgy, University of Leoben, Austria.
Herbert Mizelli, voestalpine Stahl GmbH, Linz, Austria.
Wolfgang Höfer, voestalpine Stahl GmbH, Linz, Austria.
Andreas Viertauer, Siemens VAI Metals Technologies GmbH, Linz, Austria.
Martin Hiebler, Siemens VAI Metals Technologies GmbH, Linz, Austria.
Corresponding author: Thomas Kollmann, [email protected]
> 15
RHI Bulletin > 1 > 2012, pp. 16–19
Jürgen Schütz, Alexander Maranitsch and Milos Blajs
New Oxycarbide Refractory Products

Demonstrate Outstanding Properties—First
Practical Results
Introduction creates a completely different pore structure. The matrix
structure is microporous with an average pore size approxi-
The initial idea behind the development of a new refractory
mately one-tenth that of traditional cement-bonded systems
material was to replace the traditional calcium aluminate
(Figure 1). This results in completely different material prop-
cement used as a binder in alumina-based refractory casta-
erties and facilitates water evaporation.
bles (e.g., low cement (LC) and ultra low cement (ULC)
mixes). Therefore, a new binding system was developed
Very complex reactions between the different carbon c arriers,
that avoids the disadvantages of the calcium aluminate
antioxidants, and binder generate a product with superior
cement. Refractory cement is not only an expensive raw
characteristics at high temperatures. These include:
material for bonding refractory products, it also has multi-
ple disadvantages during application including:
>> Excellent thermal shock resistance.
>> High chemical resistance against acidic as well as basic
>> Decrease in refractoriness (CaO forms low melting phas-
slag attack.
es with other oxidic raw materials used for refractories).
>> Hot erosion and corrosion resistance.
>> Time consuming curing, drying, and heating up proce-
dures.
>> Energy intensive drying and dehydration of the Ca-
Oxycarbide Product Properties
hydrate phases. Refractoriness Under Load
When compared to LC-bonded castables based on the same
Taking these facts into account, RHI developed a new type raw materials, the oxycarbide products demonstrate a 200–
of alumina-based refractory material for hot metal and steel 300 °C higher refractoriness under load (Figure 2). Outstand-
applications, comprising different carbon carriers, antioxi- ing hot modulus of rupture (HMOR) values (> 25 N/mm2 at
dants, a special liquid binder, and in certain cases silicon 1500 °C) have also been measured.
carbide.
The presences of carbon additives in the matrix in combina-
Philosophy of the New Oxycarbide Product tion with the microporous structure leads to a product with
Range more ductile characteristics, which is distinct from the very
brittle nature of traditional sintered ceramic materials. The
All oxycarbide products are completely cement-free con-
carbon present also eliminates the formation of glassy phases,
cretes that use a separate, special type of binder. Due to
whereas the micropores inhibit cracks from spreading.
the absence of Ca-hydrate phases there is no chemically
bonded water in the fluidized mix and cured product. There-
Thermal Shock Resistance
fore, a safe and rapid heating up is possible, including for
thick lined sections. The absence of CaO also guarantees a As shown in Figure 3, absolutely no cracks were visible after
much higher refractoriness. Furthermore, the special binder rapidly heating up (Figure 4) a wellblock with the new
1.4
n Bauxite LCC (T0.5 1464 °C)
1.2 n Oxycarbide bauxite (T0.5 > 1704 °C)
n Corundum LCC (T0.5 1681 °C)
1.0 n Oxycarbide corundum (T0.5 > 1750 °C)
Expansion [%]
0.8
0.6
0.4
0.2 Load: 0.1 N/mm2
0
0 300 600 900 1200 1500 1800
20 µm Temperature [°C]
Figure 1. Oxycarbide matrix prefired at 1500 °C. Figure 2. Comparison of the refractoriness under load of low
cement castables (LCC) with oxycarbide mixes based on the
same raw materials and prefired at 1500 °C.
16 <
oxycarbide bonding. In contrast, all standard cement- >> A much faster heating up rate is possible.
bonded blocks showed crack formation under the same test >> A lower overall temperature is necessary to dry out the
conditions, namely the blocks were heated up to 1700 °C in refractory castable.
5 hours from one side under oxidizing conditions. The
excellent thermal shock resistance makes the oxycarbide These two benefits are illustrated in the drying behaviour
products applicable for a diverse range of processes. curves shown in Figure 5, comparing cement and oxycar-
bide-bonded castables.
Chemical Resistance
In the case of LC and ULC mixes, the different Ca-hydrate
Another remarkable characteristic of the newly developed
phases created while curing the cement significantly affect
oxycarbide material is that it shows only a very thin decar-
the heating up process. A slow heating up rate, with hold-
burized zone of a few millimetres below the surface. Due to
ing times at several temperatures, is necessary to dehy-
the carbon content in the refractory products, the wettability
drate these phases. The total removal of the chemically
by steel, hot metal, and slag is strongly reduced. This prop-
bonded water happens at a temperature up to 600 °C. It
erty in combination with the microporous structure results
has to be taken into consideration that this temperature
in a much higher corrosion and infiltration resistance,
has to be reached throughout the entire refractory con-
including a reduced infiltration depth, compared to standard
crete installation to avoid any risk of damage during the
LC and ULC castables. Susceptibility to sulphur attack
heating up process. Depending on the application area
depends mainly on the cement-derived CaO content in tra-
and furnace geometry, this is difficult to realize and some-
ditional LC and ULC mixes; however, because there is no
times very long heating up schedules are necessary. In
cement in the oxycarbide products the sulphur resistance is
contrast, a temperature of ~ 150 °C is high enough to dry
excellent.
the new oxycarbide products. This remarkable advantage
results in a significant reduction of the heating up energy
Heating Up
and time as well as an associated reduction in CO2 emis-
In contrast to cement-bonded castables, there are two sions.
essential advantages when heating up and drying the oxy-
carbide products: Following the development and determination of the excel-
lent physical properties, the first practical tests were under-
taken with the oxycarbide products. The very aggressive
operation conditions of the CAS-OB process were chosen
for the initial service evaluation to provide significant prac-
tical test results.
The CAS-OB Process

The CAS-OB process (composition adjustment by sealed
argon bubbling-oxygen blowing) was developed by Nippon
Steel Corporation (Figures 6 and 7). During the process it is
possible to add all the necessary alloying elements into the
melt through a slag-free surface in the absence of atmos-
pheric air. This is achieved by immersing a bell into the
steel bath above an argon purging element. The bell also
enables oxygen to be lanced simultaneously with the addi-
tion of aluminium. In the resulting exothermic reaction,
Figure 3. Cross section of an oxycarbide wellblock heated up to Al2O3 is formed and considerable amounts of heat are gen-
1700 °C in 5 hours. erated; it is estimated that temperatures of around 2000 °C
1800 100
90
1500
80
Emitted water [%]
70
Temperature [°C]
1200
60
900 50
40
600
30
20
300 n Oxycarbide
10 n Cement (8 wt.%)
0 0
0 1 2 3 4 5 6 7 8 9 10 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0
Time [hours] Time [hours]
Figure 4. Heating up curve used to compare the thermal shock Figure 5. Comparison of the dehydration curves for cement-
resistance of oxycarbide-bonded and standard cement-bonded bonded versus oxycarbide-bonded materials.
wellblocks. The wellblocks were heated to 1700 °C from one side
in 5 hours under oxidizing conditions.
> 17
can be reached inside the bell. In summary, the basic func- heating up time can also be reduced dramatically. In addition
tions of the CAS-OB process are: to the described advanced physical properties, other very
important advantages of the newly designed products are
>> Homogenization and adjustment of the molten steel time, cost, and energy savings, as well as a reduction in CO2
composition and temperature. emissions.
>> No oxidation and loss of added alloying elements pro-
viding an exact and reproducible chemical composition After drying, both parts of the bell are assembled together
of the steel melt. and finished prior to application (Figure 8).
>> Effective method for attaining clean steel.
Trial Results—COMPAC ROX A93MAS-15 Instal-
Production of CAS-OB Bells Using the New lation in CAS-OB Bells at SSAB Tunnplåt
Oxycarbide Material
The Oxycarbide Bells in Operation
Typically, the CAS-OB bells consist of two parts: The so-
called “wine glass” or upper part is protected by refractory In 1992, SSAB Tunnplåt AB (Luleå, Sweden) took the decision
only on the inside whilst the “polo” or lower part is steel to build a new ladle treatment station. The CAS-OB process
reinforced refractory material. Since only the lower part is was chosen and the startup took place in August 1993. At
dipped into the steel bath during the CAS-OB process, this SSAB, the treatment time is up to 25 minutes per heat for
part is the most stressed by extremely high temperatures, a ladle capacity of 130 tonnes.
thermal shock, as well as chemical erosion and corrosion.
One major cost factor of the CAS-OB process is the refractory
Twelve hours after casting the lower bell section with material for the bell. This material is stressed by huge ther-
approximately 2.5 tonnes of the oxycarbide brand COMPAC mal cycles between each heat, which can limit the lifetime of
ROX A93MAS-15, it can be heated up and dried out. Since the bell (Figure 9). Periods of lower production and many
the material doesn’t contain any cement, it is not necessary stoppages and standstills can also have a negative influence
to have the prolonged curing time required for all cement- on bell performance because the bells cool down completely
bonded products. Furthermore, because there are no Ca- and are heated up very rapidly when they are dipped into the
hydrate phases in oxycarbide products the drying and hot steel again. This results in enormous thermal shock.
Upper part
Bell
Slag
Lower part
Ladle
Melt
Argon gas purging
element
Figure 6. Image of the CAS-OB process. Figure 8. CAS-OB bell in production.
Figure 7. CAS-OB process in operation. Figure 9. Magnesia-based competitor material after 17 heats in
operation.
18 <
Practical Results The bells can be operated at different heights, which means
that after the first segment is worn (~ 400 mm of the lower
Compared with standard competitor bells, the lifetime could
part), the bell is dipped deeper into the steel bath. Up to
be doubled using COMPAC ROX A93MAS-15 (Figures 10
three segments can be used in this manner. In comparison
and 11). In general at SSAB Tunnplåt there is no mainte-
to other steel producers who also use the CAS-OB technol-
nance, intermediate repair, or gunning of the CAS-OB bells.
ogy, the SSAB bells are relatively small and the treatment
time and ratio of Ca/Si treatment is long and intensive. In
addition, the chemical heating up is greater than at other
CAS-OB plants. Therefore, a direct comparison of the life-
time and performance of bells between different CAS-OB
plants is difficult. However, whilst the magnesia-based com-
petitor bells were destroyed by vertical cracks mainly
caused by thermal shock, the oxycarbide bells showed
absolutely no cracks until the end of operation and were
only slowly worn by hot corrosion and chemical dissolution
(Figure 12).
An additional significant advantage of the bells installed

with the oxycarbide material was a clean inner and outer
surface of the bell since the carbon and carbide content of
the oxycarbide product has an antiwetting effect (see Figure
Figure 10. COMPAC ROX A93MAS-15 bell after 52 heats. 12). As a result slag and oxides formed during the steel
treatment do not stick to the refractory surface in contrast
to the bells based on other raw materials. For this reason
no additional slag treatment with CaO-CaF2 or CaO-B2O3 is
necessary.
Conclusion
The superior properties including extremely good thermal
shock resistance, a microporous structure, the antiwetting
effect resulting from carbon and carbides, reduced brittle-
ness, and high hot strength caused by in situ carbide forma-
tion make the oxycarbide products highly suitable for differ-
ent steel treatment, hot metal, and foundry applications.
Currently, hot metal application field trials including blast

furnace runner systems (i.e., main runners, hot metal and
slag runners, tilters, skimmers, and spouts), torpedo cars
(i.e., mouth and impact areas), and hot metal ladles (i.e.,
Figure 11. COMPAC ROX A93MAS-15 bell after 64 heats.
bottom or full monolithic linings, spout areas, and well-
blocks) are planned or running.
In several steel plants the oxycarbide castables have been

installed for diverse applications including RH degasser
snorkels, CAS-OB bells, and steel ladles (i.e., full monolithic
lining or segments such as bottoms, sidewalls with and
without monolithic slag zones). Whilst there are no final
trial results at this stage, comparisons with traditional
installed linings are providing a very optimistic outlook for
these new products. In addition to the aforementioned tri-
als, prefabricated parts (e.g., wellblocks, and pocket blocks)
are in operation and showing very good results. On occa-
sions, the large and thick dimensions of refractory products
can cause problems during the heating up and for these
applications the oxycarbide bonding is proving to be an
ideal solution. Additional sectors where oxycarbide prod-
Figure 12. COMPAC ROX A93MAS-15 bell after 35 heats. Abso- ucts can be used include the foundry industry for long cam-
lutely no cracks and slag are visible. paign cupolas as well as transport ladles.
Authors
Jürgen Schütz, RHI AG, Steel Division, Mülheim-Kärlich, Germany.
Alexander Maranitsch, RHI AG, Steel Division, Vienna, Austria.
Milos Blajs, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Jürgen Schütz, [email protected]
> 19
RHI Bulletin > 1 > 2012, pp. 20–25
Marcus Kirschen, Simão Pedro de Oliveira, Elshad Shikhmetov and Matthias Höck
Customer-Specific Analysis of Steelmaking Slags

to Provide Process and Refractory Lining Lifetime
Improvements in Steel Treatment Ladles and
EAFs
Modern steelmaking processes require precise control of the slag operation in order to maxi-
mize mass and energy transformation efficiency and to minimize wear of the refractory lining
in the metallurgical unit. Regular slag sampling and analysis is state of the art in modern melt
shops, although slag optimization is sometimes neglected due to the additional effort and
costs. In this paper, the assessment of mass balances and chemical analyses of slags in
order to improve the metallurgical processes and lining lifetime in steel treatment ladles and
EAFs are presented.
Introduction Variance in the slag composition is common due to an input

of sand, concrete, and other contaminants with the scrap
The chemical composition of process slag has a significant
into the EAF, oxidation products from the metallurgical
impact on the customer’s steelmaking process and the life-
refining processes, and slag carryover from the primary
time of the refractory lining, comprising for example mag-
melting vessel to the ladle. Control and adjustment of the
nesia-, alumina-, or doloma-based materials. The physical
slag composition using slag analysis and detailed mass bal-
properties including viscosity, saturation status with respect
ance of slag formers are state of the art in modern melt
to periclase (MgO), dicalcium silicate (Ca2SiO4), and lime
shops. However, these prerequisites to minimize the impact
(CaO) determine the success of the slag operation during
various processes such as ladle treatment, slag foaming in of metallurgical treatments on the lining lifetime are some-
the EAF, and the impact on the lining lifetime. The viscosity times disregarded due to the additional effort and costs.
and chemical composition have to be in appropriate ranges
depending on the metallurgical constraints of the steel Mass Balance of Process Slags in the Steel
refinement. Significant corrosive wear of the ladle lining is Treatment Ladle
common if the MgO activity of the slag is too low and for
Mass balance of process slag is a valuable tool to determine
example the FeO and Al2O3 contents of the ladle slag are
the necessary input of slag modifiers in order to obtain opti-
too high (Figure 1). Minimizing corrosive wear of magnesia-
mum physical properties of the slag, such as CaO or MgO
based refractory linings requires a MgO-saturated slag;
saturation. All slag forming materials that are added to the
however, the MgO saturation point is particularly sensitive
ladle during tapping and steel treatment are taken into
to the FeO, Al2O3, and SiO2 concentrations.
account (Table I). Removal of highly oxidized slags after tap-
ping and the substitution by synthetic slag or mixes of lime,
dololime, and calcium aluminate are recommended for high-
quality steel treatment and clean steel production.
The MgO mass balance during the ladle transport and

refinement process is informative to determine the slag
potential to corrode the MgO-C lining (equation 1):
xlime MgO · mlime + xdolo MgO · mdolo + xCa aluminate MgO ·

mCa aluminate + xtaphole filling sand MgO · mtaphole filling sand + (1)
xcarryover MgO · mcarryover + xrefr MgO · mrefr = xslag MgO · mslag
Where x is the concentration of MgO in the slag former

(e.g., xlime MgO is the MgO content in lime) and m is the mass
of material added. In general, MgO addition is restricted to
the slag saturation limit because the corrosion potential of
the slag vanishes at the saturation point. Higher MgO val-
ues than the saturation limit come from either unnecessar-
Figure 1. Typical corrosion of a ladle lining by process slag with ily high MgO input from slag formers, unusually high losses
an unsuitable composition, namely an unusually high FeO con-
tent. from MgO-based repair or gunning mixes, or MgO losses
from the refractory lining due to erosion.
20 <
Slag formers CaO MgO SiO2 Fe2O3 Al2O3 Mass*

(wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (kg/tonnesteel)
Lime 89–95 1–4 1–2 0–10
Raw dolomite > 28–30 > 18–20 SiO2 + Fe2O3 + Al2O3 < 4–5 0–5
Dolomitic lime 56–60 37–40 SiO2 + Fe2O3 + Al2O3 < 2–4 0–5
Bauxite 2.5–7 2–7 74–82
Synthetic calcium aluminate slag A 0–40 27–35 < 8.0 < 5.0 0–10
Synthetic calcium aluminate slag B 0–40 20–26 1–5 < 1.0 0–10
Synthetic slag modifier C 10 0.5 33–37 0.5 5.0 0–5
Fluorspar < 1.5 < 18.3 < 0.02 CaF2 > 80 0–5
Synthetic lime-CaF2 mix 66 1 4 0.5 CaF2 > 34 0–5
Olivine taphole filling sand 0.5–3 40–50 39–45 6–9 0.5–3 0.5–1.5
EAF slag carryover 30–45 5–15 15–35 15–40 2–10 0–5
Al, FeSi oxidation products FeSi: 100 Al: 100 1–5
Table I. Composition ranges and input masses of slag formers added to the steel treatment ladle. * indicates the range of mass input
to produce CaO-SiO2-rich slags or CaO-Al2O3-rich slags for steel treatment.
If the initial composition of the process slag is MgO under- MgO Saturation of Slags
saturated, corrosion of the MgO-C lining by dissolution of
The MgO saturation concentration of a particular process
the MgO component occurs. Doloma linings require MgO
slag, xsaturated slag MgO, can be estimated from empirical mod-
and CaO saturation of the slag. The amount of MgO that
els, for example the Schürmann and Kolm model [1], the
will be corroded can be calculated from the slag mass bal-
Park and Lee model [2] (Figure 2), the Kwong model [3], and
ance and can be compensated by the appropriate addition
Pretorius ISD diagrams [4]. Both the Park and Lee, and Pre-
of MgO-containing material in order to decrease magnesia
torius and Carlisle models are based on the basicity ratio,
lining wear. The maximum amount of MgO corroded from
Bi, of the slag (Table II). Bi represents the ratio between the
the lining is estimated from equation 1 as the difference
between the MgO slag saturation level and the actual MgO
input into the slag (equation 2):
Basicity parameter Application
– xrefr MgO · mrefr loss max = [xlime MgO · mlime + B2 CaO/SiO2
xdolo MgO · mdolo + xCa aluminate MgO · mCa aluminate + B3 CaO/(SiO2+Al2O3) Oxidized slags, EAF, ladle
(2)
xtaphole filling sand MgO · mtaphole filling sand + B4 (CaO+MgO)/(SiO2+Al2O3) Oxidized slags, AOD
xcarryover MgO · mcarryover] – xsaturated slag MgO · mslag (CaO+MgO)/(SiO2+Al2O3+FeO+MnO) Reduced slags in ladle (FeO + MnO
considered)
B5
For example, a 1 wt.% MgO increase in the slag during ladle (CaO+MgO)/(SiO2+Al2O3+CaF2) Reduced slags in ladles (FeO + MnO
treatment indicates a MgO loss from the refractory lining of neglected), desulphurization slags
approximately 0.1 kg/tonnesteel, which is equivalent to a 10–20
Table II. Common basicity ratios from metallurgical guidelines
kg MgO loss per heat depending on the ladle volume. This
used in slag operations.
MgO loss corresponds with observed MgO lining wear rates
of 1–4 mm per heat and lining lifetimes of 50–150 heats.
16
The slag corrosion potential or presaturation level ΔMgO*, T = 1600 °C Mg wustite n Pretorius and Carlisle (1999)
14 + Ca2SiO4 n Schürmann and Kolm (1986)
can also be expressed as the difference between the MgO
n Park and Lee (1996)
content of the added slag formers and the slag saturation
12
level in wt.% (equations 3 and 4):
MgO saturation [%]
10
ΔMgO* = (xlime MgO · mlime + xdolo MgO · mdolo + MA spinel
xCa aluminate MgO · mCa aluminate + xtaphole filling sand MgO ·
8
(3)
mtaphole filling sand + xcarryover MgO · mcarryover)/mslag – 6 Mg wustite
xsaturated slag MgO
4
Or if slag analysis data is available: 2
ΔMgO* = xinitial analysed slag MgO – xsaturated slag MgO(4) 0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Basicity B3 = CaO/(SiO2+Al2O3)
The presaturation level is often more informative than the
analysed level of a slag sample taken during ladle treatment
as a certain amount of MgO from the lining may have dis- Figure 2. MgO saturation limits of CaO-SiO2-Al2O3 slags accord-
ing to the models of Schürmann and Kolm [1], Park and Lee [2],
solved at very low initial presaturation levels before the and Pretorius and Carlisle [4]. Abbreviations include magnesium
slag was sampled. aluminate (MA).
> 21
refractory oxides CaO and MgO, and the fluxing oxides Consistent with the ongoing increase in MgO presaturation
Al2O3, SiO2, FeO, MnO, and CaF2. Basicity ratios are also levels in the slag, the MgO loss from the ladle lining during
used as guidelines to estimate the adequate viscosity range ladle treatment decreased from 0.87 kg MgO/tonnesteel to
of the slag, for example B3 = 1.5 for ladle slags. MgO satura- 0.65 kg MgO/tonnesteel (see Figure 3). As a result, the rela-
tion can also be precisely calculated from phase equilibrium tive lining lifetime increased to 120% of the initial value due
calculations in multicomponent compositional space, for to a decreased wear rate during the optimization campaign
example using FactSage software [5,6]. One advantage of in 2010/2011 and the slag line repair interval was extended
the latter approach is that projections of the complex chem- (Figure 4).
ical composition to compositional subspaces such as B2 =
CaO/SiO2 or B3 = CaO/(SiO2+Al2O3) basicity concepts or pro- In addition to the optimization of slag operations and MgO
jections to compositional planes with particular MgO levels mass balance, other measures contributed to the increased
[8] are not necessary. Furthermore, the calculated MgO sat- ladle lining lifetime including improved control of ladle pre-
uration limit takes all the main slag components such as heating, maintenance personnel know-how, adjustments to
SiO2, Al2O3, FeO, and MnO into account. the brick formats, optimized ladle logistics, stirring, and
modification of the lining design to meet the specific
The corrosion potential of a process slag can then be mini- demands of the customer process.
mized by adding the appropriate amount of MgO to the slag
or—if MgO is already present in the slag—by achieving Efficient desulphurization requires stirring of the steel melt,
MgO saturation by adjusting the CaO level. a high steel temperature, and a high sulphide capacity of
Slag Analysis and MgO Optimization in a Steel

140
Treatment Ladle
An example of the assessment of ladle furnace slag presat- 120
uration figures is given in Figure 3 for Si-killed low alloyed
low carbon steels in a 100-tonne ladle. Following slag anal- 100
Relative units [%]
ysis, the presaturation levels were increased by appropriate

slag conditioning from ΔMgO* = -5 wt.% to -3.9 wt.%. Even 80
the small increase of 1.1 wt.% MgO had a measurable effect
on the lining wear rates because the chemical potential dif- 60
ference of MgO decreased between the slag and lining,
40
decreasing the driving force of the corrosion processes.
Other possible contributing factors to lining wear and MgO
20 n Total lining lifetime
input into the slag are due to erosion (e.g., stirring) and n Slag line repair interval
material loss following damage caused by thermal cycling
0
of infiltrated bricks or adverse stress patterns in the lining. Jan 10 Apr 10 Jul 10 Oct 10 Jan 11 Apr 11 Jul 11 Oct 11
Because these sources of MgO are always present in varia-
ble amounts, the slag optimization targets are in the range
Figure 4. Relative ladle lining lifetime and slag line repair interval
of ΔMgO* = -1 to -3, rather than MgO saturated slags, for a 100-tonne treatment ladle with Si-killed low alloyed low
namely ΔMgO* = 0 (Figure 3). carbon steels over the course of a slag optimization programme.
-1.0
Target
Slag presaturation level [ΔMgO*]
-2.0
-3.0
-4.0
-5.0
-6.0
MgO loss from lining to slag [kg/tonnesteel]
-7.0
-8.0 1.50
1.25
1.00
0.75
0.50
0.25
1–31 May 2010 1–13 June 2010 14–27 June 2010 28 June –19 July 2010
0.00
0 25 50 75 100 125 150 175 200
Heats
Figure 3. Analysed MgO presaturation levels (red line) of ladle furnace slags and the decreasing MgO loss from the lining (blue line)
during ladle treatment.
22 <
the slag, which is achieved at high slag basicity levels. Sul- indicating a high influence due to sand, concrete, and other
phide capacity slightly increases with the MgO content in contaminating additives in the scrap, as well as MgO input
CaO-SiO2 slags if SiO2 is not too low [7]; however, this effect from gunning and repair mixes and lining bricks, is shown in
vanishes at low SiO2. Therefore, an appropriate MgO con- Figure 5. A variance of the SiO2 mass input into the slag is
tent in the ladle slag near to saturation is also beneficial for not uncommon; however, due to unusually low amounts of
the desulphurization process with CaO-SiO2-rich slags, or at lime and dololime in the EAF the SiO2 composition scatter
least no negative impact on desulphurization has been was significant because under these circumstances the corro-
reported. sion potential of high SiO2 slags is intensified. A lack of slag
volume may also increase lining wear due to arc radiation.
Visualization of Slag Compositions Therefore, increasing the mass of both lime and dololime to
act as SiO2 buffering slag formers was recommended in this
Process Slags in Electric Arc Furnaces
case. Increasing the slag mass also improved slag foaming
Steelmakers often request the optimum slag composition and electric arc shielding.
suitable for steel melting and refinement during secondary
metallurgy that also maximizes the lining lifetime. The visu- Slag analysis from an 80-tonne EAF indicated MgO under-
alization of analysed slag compositions provides valuable saturation in all the EAF slag samples (Figure 6). In this case
information in order to characterize and evaluate particular replacement of lime by dololime was suggested in order to
process conditions. achieve 9 wt.% MgO saturation. In contrast, the analysis of
slags from a 60-tonne EAF, where mixtures of lime and
MgO saturation of the process slag in an EAF is not only ben- MgO-containing slag conditioners were used, revealed reg-
eficial for the MgO-based lining but is also a necessary pre- ularly saturated slags with good foaming properties and a
requisite for efficient slag foaming, as the presence of fine minimum lining corrosion potential.
solid MgO particles increases the slag viscosity to the appro-
priate level for foaming. The increased volume of the foam- The amount of proposed dololime addition in Figure 6 was
ing slag helps to decrease energy losses by arc radiation to calculated using MgO mass balance (see equation 1) and
the sidewalls, increase energy transfer from the arc to the the analogous CaO mass balance. The difference between
melt, and improve energy efficiency of the EAF process. the analysed slag composition and the target MgO-satu-
rated composition was used to determine the necessary
Figures for unusually high refractory lining wear may be due correction to the slag former input by mass balance.
to poor slag composition control, although the mean MgO
saturation level of the slag appears to be at the appropriate A further example shows slag samples from a 100-tonne
value. An example of slag compositions from a 60-tonne EAF EAF where there was high control of the slag composition
where the SiO2 and MgO contents scattered independently, so the MgO level was very close to saturation, although the
MgO SiO2 MgO SiO2

SiO2 SiO2
FeO FeO
CaO CaO
Under- Under-
saturated saturated
Saturated Saturated
MgO FeO MgO FeO

Analysis Target EAF Q1/2010 Analysis Target 80-tonne EAF, 2007
MgO 14.8 14.0 EAF Q2/2010 MgO 3.0 9.0 60-tonne EAF, 2011
Al2O3 5.2 5.0 ◆ EAF Q3/2010 Al2O3 8.7 8.0 n Saturation line, 35 wt.% CaO, 1500 °C
SiO2 18.5 17.0 n Saturation line, 35 wt.% CaO, 1600 °C SiO2 12.6 12.0 n Saturation line, 35 wt.% CaO, 1600 °C
CaO 36.3 35.0 n Saturation line, 35 wt.% CaO, 1650 °C CaO 34.8 33.0 n Saturation line, 35 wt.% CaO, 1700 °C
Cr2O3 1.1 1.5 l Recent slag analysis Cr2O3 1.9 1.5 l Recent slag analysis
MnO 5.6 6.6 l Target slag composition MnO 5.0 5.0 l Target slag composition
FeO 15.8 19.0 FeO 32.7 30
Lime (kg) 430 Correction to lime (kg) 600 Lime (kg) 2900 Correction to lime (kg) -906
Dololime (kg) 560 Additional dololime (kg) 1000 Dololime (kg) 0 Additional dololime (kg) 1236
Figure 5. Visualization of EAF slag analysis with respect to MgO Figure 6. Visualization of EAF slag analysis with respect to MgO
saturation levels indicating an initial poor control of the slag com- saturation indicating consistent MgO-undersaturated slags from
position due to a low input of lime/dololime into the 60-tonne EAF an 80-tonne EAF when 100% lime was used as a slag former ver-
and improved slag undersaturation over the course of the analy- sus slags from a 60-tonne EAF where efficient MgO slag condi-
sis. Saturation lines calculated with the thermochemical FactSage tioning had been implemented. Analysis and target values in
software [6]. Analysis and target values in wt.%. wt.%.
> 23
FeO content was poorly controlled with values between 20 line in Figure 8), or B5 > 1.6 for a CaO-SiO2-Al2O3 slag, and
wt.% and 60 wt.% (Figure 7). The slag analysis indicated B5 > 1.8 for a CaO-Al2O3-rich slag.
excellent control of the slag basicity, for example a bal-
anced CaO/SiO2 mass ratio of > 2. However, the large vari- The assessment of slag analysis data for Si-killed steels
ance in FeO, due to poor oxidation control by oxygen and from a 40-tonne ladle (Figure 9) showed remarkable
carbon injection, generated a high proportion of very oxi-
dized slags with a definite lining corrosion potential. In
addition, the slag viscosity dropped at high FeO levels and MgO SiO2
SiO2
the slag foaming index decreased. However, adjustment of
the slag former input improved the MgO saturation figures
in the fourth quarter of 2009 due to an increase in the CaO
Al2O3
level (see Figure 7).
CaO
→
Process Slags in Steel Treatment Ladles
w t.
2
iO
%A
%S
The ladle slag composition is adjusted to the type of steel
l 2O 3
w t.
killing strategy after tapping the BOF or EAF: Al-killed steels CaO/(SiO2+Al2O3) = 1.5
→
require a calcium aluminate slag with low SiO2 activity in
order to avoid reduction of SiO2 by Al added to the steel
melt. Mixed Al-Si-killed and Si-killed steels are usually cov-
M
ered by a calcium-silicate(-alumina) slag (Figure 8). The total fre gO
eC +
FeO and MnO concentration should be low, in the ideal case aO
< 2 wt.%, to avoid any mass exchange between the slag and
melt (e.g., oxidation of Si and Al by FeO and MnO). CaO Al2O3
80-tonne ladle n MgO saturation, 5 wt.% MgO, 1600 °C

For both slag types, high magnesia lining lifetimes require 150-tonne ladle n Ca2SiO4, Ca3SiO5 saturation, 5 wt.% MgO, 1600 °C
high MgO slag activity. In the case of high calcium silicate n CaO saturation, 5 wt.% MgO, 1600 °C
saturated slags, the Ca2SiO4 and/or Ca3SiO5 levels are close
to the MgO periclase saturation values of 7–12 % MgO. In Figure 8. Visualization of slag analysis from an 80-tonne ladle (Si-
the case of high calcium aluminate slags there is double killed steels) and a 150-tonne ladle (Al-killed steels) showing the
important saturation fields at 1600 °C and basicity values B3 =
saturation with lime and periclase at 7–12 wt.% MgO (Fig-
CaO/(Al2O3+SiO2) = 1.5. Saturation lines calculated with the ther-
ure 8). The common metallurgical rules for optimum slag mochemical FactSage software [6]. Stability fields at 5 wt.% MgO
composition reflect these saturation figures: B3 near 1.5 (red from [8]. Abbreviations include Ca2SiO4 (C2S) and Ca3SiO5 (C3S).
SiO2 SiO2
MgO MgO
SiO2 SiO2
FeO Al2O3
CaO CaO
→
2
iO
%S
wt.
wt.
Under-
Al 2
saturated
O3
→
CaO/SiO2 = 2
M
fre gO
eC +
Saturated aO
CaO FeO CaO Al2O3

Analysis Target EAF Q1/2009 EAF Q3/2009 Analysis Target Slag samples
MgO 4.6 8.0 EAF Q2/2009 ◆ EAF Q4/2009 MgO 9 11 n MgO saturation, 10 wt.% MgO, 1600 °C
Al2O3 2.5 3.0 n Saturation line, 5 wt.% MgO, 1550 °C Al2O3 30 26 n Ca2SiO4, Ca3SiO5 saturation,
SiO2 12.3 12.0 n Saturation line, 5 wt.% MgO, 1600 °C SiO2 15 14 10 wt.% MgO, 1600 °C
CaO 31.4 35 n Saturation line, 5 wt.% MgO, 1650 °C CaO 43 47 n CaO saturation, 10 wt.% MgO, 1600 °C
Cr2O3 0 0 l Recent slag analysis B3 0.956 1.542 l Recent slag analysis
MnO 4.6 4.0 l Target slag composition l Target slag composition
FeO 34.6 30.0
FeO + MnO < 1.5 wt.% Correction CaO (%) 4
Lime (kg) 4000 Correction to lime (kg) -455 MgO (%) 2
Sum 97 wt.%–102 wt.%
Dololime (kg) 0 Additional dololime (kg) 848 Al2O3 (%) -4
Figure 7. Visualization of slag analysis from a 100-tonne EAF Figure 9. Visualization of slag analysis from a 40-tonne ladle (Si-
with respect to MgO saturation indicating a significant propor- killed steels) indicating MgO saturated and undersaturated ladle
tion of high FeO-containing slags due to suboptimum control of slags during the steel treatment process. Stability fields at 10 wt.%
the oxygen versus carbon injection. Saturation lines calculated MgO from [8]. Abbreviations include Ca2SiO4 (C2S) and Ca3SiO5
with the thermochemical FactSage software [6]. Analysis and (C3S). Analysis and target values in wt.%.
target values in wt.%.
24 <
agreement between some heats and the calculated satura- Summary

tion lines. This indicated that undersaturated ladle slags
Assessment of slag mass balances and visualization of slag
with a variable initial Al2O3 content dissolved MgO from the
analyses provide essential information to optimize slag
lining until MgO saturation was reached. Other slags
compositions and improve both the lining lifetime and mul-
remained MgO undersaturated at the Ca2SiO4 saturation
tiple metallurgical processes. Using various tools, RHI is
line. Although most slag samples were MgO saturated, an
able to provide this customer-specific analysis, enabling tai-
increase of the initial MgO concentration in the slag formers
lored recommendations to be made regarding slag adjust-
could be effective in decreasing ladle lining wear.
ment using slag formers. In EAFs, where the slag composi-
tion can vary widely due to contaminants in the input mate-
Materials for MgO Slag Conditioning
rial, visualizing slag analysis enables, for example, the lin-
Metallurgically used lime and synthetic slag formers based ing lifetime as well as slag foaming to be improved. MgO
on calcium aluminates only contain a few wt.% MgO (see mass balance of process slags has also been effectively
Table I). As a result, the initial MgO content of the slag in an used to increase the lining lifetime in steel treatment ladles,
EAF or steel treatment ladle might be too low to prevent whilst the visualization of slag compositions enables slags
corrosion of the lining material. In this case the addition of to be precisely examined in relation to the type of killing
MgO is recommended and there are various MgO-contain- strategy adopted in the ladle.
ing mineral sources available on the market. RHI provides
high-quality dolomite and magnesia sinter in order to mod-
ify the slag composition, improve slag operation, and
increase the lining lifetime in EAFs as well as transport and
refinement ladles (Table III).
Material and origin Size LOI CaO MgO SiO2 Fe2O3 Al2O3
(mm) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%) (wt.%)
Raw dolomite, 1–6 47.7 30.9 21.2 0.13 0.1 0.1
Marone (Italy)
SLAGDOL, sintered 0–1 58.5 39.5 1.0 0.5 0.5
doloma, Marone (Italy)
PENTADOL 5-15, 4–13 58.5 39.5 1.0 0.5 0.5
sintered doloma,
Marone (Italy)
Magnesia brickets 20–50 39.3 8.3 45.2 0.7 3.4 0.3
HL15, Hochfilzen
(Austria)
KAUSTER RKM-S, 1–6 2.5 8.3 56.0 23.5 3.4 6.2
Radenthein (Austria)
Table III. Materials provided by RHI for MgO slag conditioning.

Abbreviations include loss on ignition (LOI).
References
[1] Schürmann, E. and Kolm, I. Mathematische Beschreibung der MgO-Sättigung in komplexen Stahlwerksschlacken beim Gleichgewicht mit
flüssigem Eisen. Steel Research. 1986, 57, 7–12.
[2] Park, J. and Lee, K. Reaction Equilibria Between Liquid Iron and CaO-Al2O3-MgOsat-SiO2-FetO-MnO-P2O5 Slag. Proceedings 79th Steelmaking
Conference, Iron and Steel Society, Pittsburgh, USA, March 24–27, 1996, pp. 165–171.
[3] Kwong, K., Bennett, J., Krabbe, R. and Thomas, H. Thermodynamic Calculations Predicting MgO Saturated EAF Slag for Use in EAF Steel
Production. The Minerals, Metals & Materials Society. Supplemental Proceedings. Materials Characterization, Computation and Modeling. 2009,
Vol. 2, 63–70.
[4] Pretorius, E.B. and Carlisle, R.C. Foamy Slag Fundamentals and Their Practical Application to EAF Steelmaking. Iron and Steelmaker. 1999, 26, No.
10, 79–88.
[5] Brüggmann, C. and Pötschke, J. Contribution to the Slagging of MgO in Secondary Metallurgical Slags. Presented at 53rd International Colloquium
on Refractories, Aachen, Germany, Sept., 8–9, 2010, pp. 145–149.
[6] Bale, C., Chartrand, P., Degterov, S., Eriksson, G., Hack, K., Ben Mahfoud, R., Melançon, J., Pelton, A. and Petersen, S. FactSage Thermochemical
Software and Databases. Calphad. 2002, 26, No. 2, 189–228.
[7] Taniguchi, Y., Sano, N. and Seetharaman, S. Sulphide Capacities of CaO–Al2O3–SiO2–MgO–MnO Slags in the Temperature Range 1673–1773 K.
ISIJ International. 2009, 49, No. 2, 156–163.
[8] Schlackenatlas, Slag Atlas; VDEh., Ed.; Verlag Stahleisen: Düsseldorf, 1981.
Authors
Marcus Kirschen, RHI AG, Steel Division, Vienna, Austria.
Simão Pedro de Oliveira, RHI Refratários Brasil, Belo Horizonte, Brazil.
Elshad Shikhmetov, RHI U.S., Ltd., USA.
Matthias Höck, RHI AG, Steel Division, Vienna, Austria.
Corresponding author: Marcus Kirschen, [email protected]
> 25
RHI Bulletin > 1 > 2012, pp. 26–33
Bernd Trummer, Bianca Heid, Manfred Kappel, Sarah Köhler, Alexander Maranitsch, Norbert Lebek and Volker Perl
Gas Purging Lances: Improving Established

Technology
Gas purging lances are a well-established tool in hot metal treatment and secondary steel
metallurgy. Lances consist of a steel pipe with steel reinforcement that is protected against
the heat by a refractory castable. The lifetime of a lance is controlled by different wear mech-
anisms that are presented in this paper. Computational fluid dynamics (CFD) was used to
investigate the influence of the reinforcement design on the temperature distribution in order
to allocate stress patterns. In addition, selected refractory castables were examined in the
laboratory regarding corrosion and thermal shock resistance. Lances comprising a selected
reinforcement design and various castables were produced and tested in field trials and the
results provide the basis for improving established lance technology.
Introduction lance under the very harsh service conditions. In particular,

the design of the reinforcement steel parts and the appro-
Inert gas purging is one of the key tools for hot metal desul-
priate choice of refractory has a significant influence on the
phurization and steel refining in secondary metallurgy. The
in-service behaviour.
purging gas is used to transport desulphurization agents
into the hot metal, mix and homogenize the liquid metal,
The following paper provides an overview of the refractory
as well as remove nonoxidic solid inclusions and dissolved
brands for lances available on the market and describes the
gases. Introduction of inert gas (mostly argon) into the liq-
main purging lance wear mechanisms. The influence of the
uid metal can be performed using purging lances as well as
steel design was evaluated by computer simulations of the
purging plugs. Hot metal applications almost exclusively
temperature distribution within the lance reinforcement.
use purging lances whereas in steel applications purging
The behaviour of selected refractory castables—alumina
lances are used mostly as emergency lances when bottom
castables as well as magnesia chromite castables—in
stirring via plugs doesn’t work properly or sufficiently. A
respect to thermal stress and interaction with slag was
detailed description of the use of plugs as well as lances is
tested in the laboratory. These laboratory findings were
given by Stolte [1]. As the older technology, purging lances
subsequently evaluated in field trials at customers, to asses
were expected to disappear with the upcoming purging
the accuracy of the assumptions.
plugs in the 1980s. However, even today gas purging lances
still have their position on the market.
Purging Lances: Market and Wear Mechanisms
Purging lances are a composite consisting of a central steel Purging lances are available from a large number of compa-
body necessary for gas transport and reinforcement, and a nies, each using its own refractory castables. Table I gives
refractory castable body protecting the steel parts against an overview of castable brands for purging lances used in
heat. The interaction of the steel parts with the refractory steel and hot metal applications, all based on alumina raw
material determines the performance and lifetime of the materials. Grades for steel treatment are in the high alumina
Brand 1 Brand 2 Brand 3 Brand 4 Brand 5 Brand 6

Application Steel treatment Steel treatment Steel treatment Steel treatment Hot metal treatment Hot metal treatment
Material composition Bauxite–mainly recycled Bauxite–mainly recycled Chrome corundum– Bauxite–mainly recycled Alumina–containing Alumina–containing
white fused alumina white fused alumina recycled white fused alumina raw materials raw materials
andalusite andalusite andalusite
fireclay
Binder Cement Cement Cement Cement Cement Cement
Steel fibres (wt.%) ~5 ~3 ~2.5 ~5.5 ~6 ~5

Chemical analysis (wt.%)
Al2O3 83.3 79.3 83.0 73.5 50 54
SiO2 11.2 12.9 3.4 22 45 39
Fe2O3 1.5 3.8 0.2 2.0 1
CaO 0.5 0.8 1.6 0.6 3 1
K2O 0.2 0.1 2.2 0.2
TiO2 2.8 2.4 0.9
Cr2O3 8.9
Table I. Overview of castable brands available on the market for purging lances used in steel and hot metal applications.
26 <
range (typically the Al2O3 content exceeds 80 wt.%) and are continuously and results in accelerated wear of the castable
mainly based on bauxite or recycled bauxite, often upgraded at the lance tip.
with minor amounts of corundum or andalusite to improve
the expansion behaviour. In addition, other high alumina Steel Reinforcement
recycled raw materials (e.g., chrome-corundum slag and
Purging lances consist of a central steel pipe that conducts
recycled corundum) are in use. Grades for hot metal treat-
the purging gas and additional reinforcing elements to affix
ment are mainly based on fireclay or alumina-enriched fire-
the refractory castable and provide mechanical stability. Fig-
clays. Most of the castables are hydraulically bonded low
ure 2 shows cross sections of three different reinforcement
cement castables (LCC); however, for hot metal applications
designs: A central pipe with hook anchors directly welded to
silica sol bonding is also common. All castables contain
the pipe (Figure 2a), a central pipe reinforced with V-shaped
steel fibres in a range from 3 wt.% to ~ 6 wt.%. Some brands
anchors also directly welded to the pipe (Figure 2b), and a
have the highest steel fibre content in the slag zone and at
central pipe encased by three angled steel plates running
the tip of the lance with lower levels in the rest of the lance.
lengthwise and V-shaped anchors welded to the obtuse
angled regions of the plates (Figure 2c). The thermal behav-
Purging lances are subject to extremely high operational
iour of these three designs under service conditions was
load, resulting in wear and consumption of the lance. The
main wear mechanisms are detailed in Figure 1. Different
modes of wear can be observed depending upon the posi-
tion along the lance. Chemical attack is the main wear fac-
tor in the slag zone, resulting in clogging or premature
wear. As the lance gets thinner, the pipework is attacked by
steel/hot metal, resulting in leakage or the lower part of the
lance breaking off. Whilst less stress affects the central
parts of the lance, thermal shock may result in the forma-
tion of vertical and longitudinal cracks with increasing crack
width after every cycle. These cracks allow liquid metal to
infiltrate the lance and cause damage to the pipework. The
highest stress levels are in the head/nozzle zone of the
lance, causing mostly discontinuous wear. Thermal shock
gives rise to the formation of cracks that widen with every
cycle and loosen the mechanical structure of the lance. (a)
Finally, the tip of the lance can break off and be lost. Infiltra-
tion of the cracks by molten metal may also create leaks in
the pipework. Additionally, the nozzles themselves can be
infiltrated by steel/hot metal that results in nozzle blockage.
Hot erosion caused by circulating steel/hot metal acts
Slag zone
>> Chemical attack by slag resulting in premature wear
or clogging
>> Leakage of pipework when attacked by steel/hot metal
(b)
Central part
>> Thermal shock during inserting and pulling out of the lance
>> Formation of vertical and longitudinal cracks
>> Increasing crack width with every cycle
>> Infiltration of steel/hot metal into open cracks
Head with nozzles

>> Thermal shock during inserting and pulling out of the lance
>> Formation of vertical and longitudinal cracks
>> Increasing crack width with every cycle (c)
>> Loss of tip/nose
>> Infiltration of steel/hot metal into open cracks and nozzles
>> Blockage of nozzles following steel/hot metal infiltration Figure 2. Cross sections of three purging lance steel reinforce-
ment designs: (a) central steel pipe with anchor hooks, (b) cen-
>> Hot erosion due to circulating steel/hot metal tral steel pipe with V-shaped anchors, and (c) central pipe
encased in three angled steel plates running lengthwise with
V-shaped anchors welded to the obtuse angled regions of the
Figure 1. Main wear mechanisms affecting purging lances. plates.
> 27
investigated using computational fluid dynamics (CFD). Fig- In general, the stress pattern is closely linked to the temper-
ure 3 shows the temperature distribution within these three ature distribution, with high temperature differences usually
designs for hot metal lance service conditions (Figures 3a–c) resulting in stress peaks in the immediate area. Such stress
as well as for steel treatment lance conditions (Figures 3d–f). peaks may generate cracks in the affected area when the
The simulation shows the development of three specific tem- mechanical strength of the castable is exceeded. Especially
perature distribution patterns that depend almost completely hook anchors seem to be very prone to causing cracks.
on the design of the steel reinforcement and are independent Increasing the anchor surface (e.g., V-shaped anchors) pro-
of the application temperature. Hook anchors (Figures 3a and vides a larger area for heat accumulation and an improved
3d) are subject to massive heat accumulation, and the tem- heat transfer to the refractory castable is possible. This
perature difference of about 80 °C between the anchor hooks results in smaller temperature peaks and therefore reduced
and steel pipe is extremely high. V-anchors directly welded stress levels.
to the steel pipe (Figures 3b and 3e) show a smaller tempera-
ture drop between the anchors and steel pipe; however, the Figure 4 illustrates that the wall thickness of the pipe has lit-
tips of the anchors are still significantly hotter than the steel tle influence on the temperature pattern within the steel
pipe. The steel pipe encased in angled steel plates running reinforcement. The heat transfer within the lance is primar-
lengthwise with V-shaped anchors welded to the plates (Fig- ily governed by the reinforcement design and its influence
ures 3c and 3f) shows the most homogenous temperature can be seen from the castable surface temperature pattern
pattern, with only very small temperature differences of the lance. Figure 5 shows the temperature distribution at
between the tips of the anchors and the steel pipe. The steel the refractory/liquid steel interface in the lower quarter of a
pipe temperature in this design is noticeably higher com- lance. Anchor hooks create a ring patterned temperature
pared to the other two designs. distribution (Figure 5a) with the anchor hooks located in the
1250 1275 1300 1325

(a) (b) (c)
1230 1340
Temperature [°C]
1450 1475 1500 1525 1550

(d) (e) (f) 1442 1563
Temperature [°C]
Figure 3. Temperature distribution within the three purging lance steel reinforcement designs detailed in Figure 2. (a–c) lance designs
under hot metal service conditions and (d–f) lance designs under steel treatment service conditions.
(a) (b)
1450 1475 1500 1525 1550

(c) (d)
1442 1563
Temperature [°C]
Figure 4. Temperature distribution within two of the purging lance steel reinforcement designs detailed in Figure 2 with a (a, c) thick
pipe wall and (b, d) thin pipe wall.
28 <
cooler rings. The resulting stress patterns favour the forma- distribution on the surface (Figure 5c). The stress pattern of
tion of cracks running circumferentially when the strength this design is almost neutral and doesn’t enhance crack for-
of the castable is exceeded. V-shaped anchors (Figure 5b) mation.
also create hot spots on the lance castable surface; how-
ever, the formation of rings with equal temperatures does
Refractory Castables
not occur. Whilst the resulting stress pattern also favours
the formation of horizontal cracks, these cracks will be Table II summarizes the chemical composition, physical
restricted locally and not propagate round the entire cir- properties, and cup slag tests, including the slag infiltra-
cumference. Lances reinforced with angled steel plates and tion behaviour, of several alumina-based castables and
V-shaped anchors show a very homogenous temperature two magnesia chromite castables for lances. The castables
(a) (b) (c)

1627 1627 1627 1628
1626.8 1627.8
Temperature [°C]
Figure 5. Castable surface temperature distribution in the lower quarter section of a steel treatment lance under service conditions for
the three reinforcement designs: (a) anchor hooks, (b) V-shaped anchors, and (c) three angled steel plates running lengthwise with
V-shaped anchors welded to the plates. The lance tip is the left-hand end of the lance section.
Mix Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Type 7 Type 8

Application Steel treatment Steel treatment Steel treatment Steel treatment Hot metal Hot metal treat- Steel and hot Steel and hot
treatment ment metal treatment metal treatment
Main raw material Sintered alumina Bauxite Recycled bauxite Sintered alumi- Alumina– Fireclay Standard MgCr Recycled MgCr
na, spinel, oxy- enriched fireclay
carbide
Typical chemical composition (wt.%)
Al2O3 96.0 82.5 83.5 93.4 61.0 52.5 7.5 6.2
SiO2 0.1 12.0 9.5 2.3 34.0 44.0 3.6 6
Fe2O3 0.1 1.2 1.0 0.1 0.9 0.8 15 13
Na2O 0.3 0.2 0.1
MgO 0.8 3.8 49.1 48
CaO 2.5 1.1 3.6 1.5 1 1.5
TiO2 2.6 1.8 2.2 1.6
Cr2O3 23 24
C 4.7
Typical physical properties
Bulk density (g/cm³) 3.12 2.84 2.66 2.92 2.52 2.39 3.10 3.00
Open porosity (vol.%) 11 15 22 12 12 14 16 18
Cold crushing strength (MPa) 110 56 43 40 104 91 80 80
Modulus of rupture (MPa) 12 8 4 4 11 9 12 12
Thermal shock resistance 30 cycles 30 cycles 30 cycles 30 cycles 30 cycles 30 cycles max. 7 cycles max. 6. cycles
(water quenching tests)
Cup slag tests
Test temperature (°C) 1600 1600 1600 1600 1300 1300 1600 1600
Wear area Low Medium High Little Little Medium Medium Medium
Infiltration depth (mm) 0 10 10 0 0 5 20–25 20–25
Infiltration area 0 20 20 0 0 5 > 90 > 90
(% of total area)
Cracks Few Few Few Some No No Many Many
Microstructure disintegration No Medium High Small No No High High
Table II. Chemical composition, physical properties, and cup slag test results for RHI purging lance refractory castables.
> 29
for steel application were in the high alumina range (e.g., very stable against basic slags, so no corrosion was
bauxite and sintered alumina) whereas lance castables for expected. However, the infiltrating slag extremely densified
hot metal application were mainly based on fireclay or the microstructure of the castable. This would significantly
alumina-enriched fireclays. The majority of these casta- deteriorate the thermomechanical properties of the castable
bles were hydraulically bonded LCC, and for hot metal and generate a brittle material that would crack in the case
applications silica sol bonding was also used. Magnesia of thermal shock.
chromite based castables were used for both steel treat-
ment and hot metal lances. Mechanical testing was per- When the alumina and magnesia chromite castable results
formed using standard testing equipment, and the ther- are directly compared, alumina shows far superior thermo-
mal shock resistance was examined using water quench- mechanical properties; however, alumina would theoreti-
ing tests. cally react more readily with the basic slag, making it sus-
ceptible to corrosion by slag attack.
LC alumina castables showed high mechanical strength
and good thermal shock resistance in the tests. However, The castable type also significantly influences heat trans-
whilst magnesia chromite castables are typically charac- port from the liquid steel/hot metal into the lance. Increas-
terized by a high resistance against basic slag attack, in ing the alumina content in the castable will result in higher
the tests they showed very poor thermal shock resistance thermal conductivities causing a higher heat up of the purg-
compared to the alumina materials. ing gas, as illustrated by the thermodynamically calculated
data in Figure 7 on page 32. Sintered alumina castables will
Interaction of the castables with a basic slag (Table III) was heat up the purging gas during its 4 m long passage
studied in detail using cup slag tests carried out at 1300 °C through the lance from 20 °C to almost 600 °C whereas a
for fireclay-based castables and 1600 °C for all the other maximum temperature of only 400 °C is reached with fire-
castables. The castable cups were filled with slag and then clay castables. High heat fluxes adversely affect the
heated for either 30 hours at 1600 °C or 60 hours at 1300 mechanical stability of the steel reinforcement and exces-
°C. The results of these cup slag tests can be seen in Figure
sive bending of the lance may occur. When solid desulphur-
6, showing the infiltration of the slag into the castable. Infil-
ization agents are transported through the lance this behav-
trated areas are delineated from the noninfiltrated areas
iour may also cause a temperature-related coagulation of
with a red line.
the particles and subsequent blockage of the lance.
Performance Benchmark in Customer Trials

CaO Al2O3 SiO2 CaF2
Lances were made from selected castables and tested in tri-
Test slag 60 20 15 5
als at several customers. Following the computer simulation
findings, the optimum steel reinforcement was selected
Table III. Chemical composition (wt.%) of the basic slag used in
the cup slag tests. consisting of a pipe reinforced with angled steel plates and
V-shaped anchors. The results of the customer trials are
summarized in Table IV. The numbers in the table give the
Slag infiltration into the alumina castables occurred at a
relative performance in percent compared to the standard
very low level and only minor portions of the castable were
lance used at the customer, which was bauxite for steel
infiltrated. The surface of the cup was sealed by a glassy
treatment and fireclay for hot metal applications.
layer, probably a reaction product of the slag and alumina
castable. This layer would slow down further chemical
Lances for Steel Treatment
attack of the slag on the lance surface and also prevent
steel infiltration into the lance. Ongoing corrosion and ther- The best performance was achieved with the type 2 bauxite
mal shocks would be the major wear factors under these castables, which performed in the same range or even out-
conditions. performed the standard lances. Lances based on corundum
(type 1) were 10–20% under the lifetime of the standard
An opposite behaviour was shown by magnesia chromite lances. Poor performance was seen with the type 7 and 8
castables, with the castable almost entirely infiltrated by magnesia chromite based lances, where the lifetime was
slag. Typically, magnesia chromite castables are chemically only 50% to about 70% of the standard lance.
Type 1 Type 2 Type 5 Type 6 Type 7 Type 8

Main raw material Sintered alumina Bauxite Alumina–enriched fireclay Fireclay Standard MgCr Recycled MgCr
Steel treatment lances
Customer A 90 100
Customer B 80 70 50
Customer C 80 110
Customer D 90
Hot metal lances
Customer E 50
Customer D 80 100
Table IV. Performance benchmark of selected RHI castables from customer trials (trials with castable types 3 and 4 in progress).
30 <
Lances for Hot Metal Treatment The marked performance differences observed in these tri-
The best performance was achieved with the type 6 fireclay- als are mainly related to the different wear behaviours of
based castables. Especially the use of silica sol as a bond- the castables. Figure 8 shows a typical picture of an
ing agent was very beneficial compared to standard LCCs. extremely cracked tip of a magnesia chromite castable
Poor results were also achieved with the type 7 and 8 mag- lance after use. A postmortem examination of the castables
nesia chromite castables, and their lifetime was only 50% clearly highlights the difference: Magnesia chromite casta-
up to 70% of the standard lance. bles (Figure 9a) show a large number of cracks
(a) (e)
(b) (f)
(c) (g)
(d) (h)
Figure 6. Cup slag tests with castable types 1–8 (a–h, respectively) from Table II. Samples a–d, g, and h were tested at 1600 °C for 30
hours and e–f at 1300 °C for 60 hours. (a) sintered alumina, (b) bauxite, (c) recycled bauxite material, (d) oxycarbide with sintered
alumina and spinel, (e) alumina-enriched fireclay, (f) fireclay, (g) standard MgCr, and (h) recycled MgCr.
> 31
deeply infiltrated with slag and steel resulting in complete off, and the cracks were infiltrated deeply with liquid metal.
disintegration of the castable structure. In contrast, the alu- This very different behaviour of alumina and magnesia chro-
mina castable (Figure 9b) shows no visible cracks and infil- mite castables can be attributed to the significantly higher
tration and the structure of the castable is fully intact. thermal expansion as well as thermal conductivity of magne-
sia chromite compared to alumina [2]. In addition, alumina-
Conclusion containing castables provide a range of compositions. From
a thermomechanical point of view, castables with low ther-
Computer simulations, experimental data, and customer tri-
mal expansion and low thermal conductivity (e.g., fireclay
als have provided an insight into the wear behaviour and
based castables) are preferable. However, whilst fireclay-
wear mechanisms of gas purging lances under laboratory
based castables are used very successfully for hot metal
and service conditions. Among various factors, chemical
lances, the refractoriness of fireclay castables is not sufficient
attack by the slag, infiltration and erosion by steel/hot metal,
thermomechanical properties, especially thermal shock
resistance, turned out to be crucial in determining the life-
time of the lance. Thermomechanical failures cause crack for-
mation followed by infiltration of the cracks and subsequent
damage of the piping or spalling of the refractory castable,
shortening the life of the lance considerably.
Approaches to improve the lance thermomechanical proper-

ties targeted two areas: Optimizing the steel reinforcement
and selecting the appropriate refractory castable. Computer
simulations revealed the larger the anchor surface and addi-
tional steel reinforcement, the more homogenous the tem-
perature distribution within the lance. Since temperature hot
spots in the metallic anchors always result in peaks within
the mechanical stress pattern, which can be the origin of
crack formation, especially thin, long anchors (e.g., hook
anchors) that create significant temperature hot spots in the Figure 8. Cracked tip of a lance after service.
castable should be avoided.
A huge influence on the thermomechanical properties can be

exerted by the composition of the refractory castable. In gen-
eral, alumina castables performed significantly better than
magnesia chromite castables. Under service conditions the
surface of alumina castables formed a glassy layer that pre-
vented the infiltration of slag or liquid metal into the casta-
ble. Alumina castables were less subject to crack formation
and if cracks formed they did not destroy the microstructure
of the alumina castable in its entirety. In contrast, with mag-
nesia chromite castables no protective layer formed on the
surface of the lance resulting in deep infiltration of the casta-
ble. Cracks formed in large numbers and completely disinte-
grated the microstructure of the refractory. Coarse grains
were loosened from the matrix, castable aggregates broke
(a)
700
600 n Castable type 1 (sintered alumina)

n Castable type 5 (alumina enriched fireclay)
500 n Castable type 6 (fireclay)
Temperature [°C]
400
300
200
100
0
0 1 2 3 4
Length [m] (b)
Figure 7. Influence of castable type on the purging gas temperature Figure 9. Different wear behaviour of magnesia chromite versus
within the lance determined from thermodynamic calculations. alumina castables: (a) postmortem sample of a magnesia chro-
mite castable and (b) postmortem sample of an alumina castable.
32 <
for steel applications. Therefore, castables based on bauxite

or even sintered alumina are used, which have much better
thermomechanical characteristics than magnesia chromite
but considerably lower thermomechanical properties than
fireclay. This fact, in combination with much higher tempera-
tures in steel applications, puts a limit on the lifetime of steel
treatment lances, which is far below the lifetime of hot metal
lances.
Previously, computer simulations of gas purging have mainly

focused on investigating and improving circulation patterns
in the ladle [3,4]. However, careful design studies of the steel
reinforcement in combination with computer simulations and
continuous development of the optimum refractory castable
will enable further improvements to purging lance perfor-
mance.
References
[1] Stolte, G. Secondary Metallurgy: Fundamentals, Processes, Applications; Verlag Stahleisen: Düsseldorf, 2002.
[2] Routschka, G. and Wuthnow, H. (Eds) Praxishandbuch Feuerfeste Werkstoffe. 5th Edition; Vulkan-Verlag: Essen, 2011.
[3] Krishnapisharody, K. and Irons, G. An Analysis of Circulation and Mixing Phenomena in Gas-Stirred Ladles. AISTech 2011 Proceedings (vol. I),
Indianapolis; USA, 2011; pp. 1367–1376.
[4] Alexis, J. and Björkvall, J. Mathematical Modeling of Stirring for an Optimized Ladle Furnace Process. AISTech 2011 Proceedings (vol. I),
Indianapolis; USA, 2011; pp. 1389–1399.
Authors
Bernd Trummer, RHI AG, Steel Division, Vienna, Austria.
Bianca Heid, RHI AG, Technology Center, Leoben, Austria.
Manfred Kappel, RHI AG, Technology Center, Leoben, Austria.
Sarah Köhler, RHI AG, Technology Center, Leoben, Austria.
Alexander Maranitsch, RHI AG, Steel Division, Vienna, Austria.
Norbert Lebek, RHI AG, Steel Division, Differdingen, Luxembourg.
Volker Perl, RHI AG, Steel Division, Duisburg, Germany.
Corresponding author: Bernd Trummer, [email protected]
> 33
RHI Bulletin > 1 > 2012, pp. 34–38
Christian Majcenovic, Johann Eder and Jens Rotsch
Microscopic Examination of Premature Wear

Caused by Joint Opening and Vertical Crack
Formation in Magnesia-Carbon Steel Treatment
Ladle Linings
In the steel industry there is ongoing process optimization concerning productivity and cost-
benefit ratio improvements. In this regard, the steel treatment ladle working conditions often
become very demanding. This happens in times of very high productivity when rapid ladle
turnover necessitates fast initial heating up as well as during low production periods with
unusually long idle times or operation stops that also require rapid ladle heating procedures.
A resulting lining wear phenomenon that can be observed in such cases is premature wear
due to joint opening and vertical crack formation. This paper provides a microscopic mineral-
ogical view of the wear that occurred at vertical cracks formed in a steel treatment ladle lined
with magnesia-carbon bricks. The detailed microscopic investigations not only provided infor-
mation regarding the refractory brick wear behaviour but were also a helpful tool to optimize
the brick microstructure during product development.
Introduction which may be caused by low Rc values. When thermal

shock and compressive failure occur at the same time the
Vertical crack formation and joint opening in steel treatment
joints are closed. However, when a distance in from the hot
ladles lined with magnesia-carbon bricks and the related
face there is thermal equilibration and expansion, the joints
premature wear with material loss in the crack and joint
open at the hot face. Additionally, tensile stresses occur at
area is highly undesirable. This wear phenomenon can be
observed with carbon-bonded magnesia-carbon brick lin- the hot face. In extreme cases of tensile stresses, vertical
ings independent of the carbon binder type or brick manu- crack formation can occur, typically bisecting the bricks in a
facturer, but always with the final consequence that the lin- vertical direction, which can also propagate when there is
ing must be replaced before the minimum residual thick- frequent thermal cycling. Often the cracks are not visible
ness is reached. after the first initialization but become macroscopically visi-
ble after several heats and accompanying wear. The typical
Fundamental investigations to verify the thermomechanical appearance is a ladle brick lining where the vertical joints
reasons for joint opening and vertical crack formation have are linked by vertical cracks (Figure 1).
been carried out for instance using finite element analysis
and are described in detail [1–3]. Joint opening is usually Under service conditions, not only is joint opening and
observed with magnesia-carbon linings that have been crack formation observed, but also a significant chemother-
exposed to severe thermomechanical stresses due to thermal and hot erosive load on the brick structure, resulting in
mal shock. This is mainly caused by a fast heating up pro- severe premature wear. Often a high erosive wear impact is
cedure combined with compressive failure of the brick hot indicated by the general ladle lining wear situation [2].
face in a circumferential direction, which entails irreversible
plastic deformation of the material at the hot face. Subse- To better understand this type of wear phenomenon, a
quent expansion of the brick material, a certain distance in microscopic mineralogical investigation was carried out on
from the hot face, results in final opening of the joints at the a postmortem magnesia-carbon brick sample from a steel
immediate hot face where the brick material has been irre- ladle. Thereby, the changes in microstructure in a particular
versibly damaged. vertical crack area could be clarified and used as the basis
for optimized product development as well as recommend-
Harmuth et al., [2] concluded that the irreversible strain ing countermeasures to the ladle operation conditions. The
caused by compressive failure at the hot face depends on: investigated postmortem brick sample originated from the
slag line and was a magnesia-carbon material with a 14
Rc = fc /α.E(1) wt.% carbon content. It showed typical bisection by a verti-
cal crack (Figure 2). Due to the premature joint wear and
Where Rc, which has formal similarity to the thermal stress vertical cracks the performance was only 40% of the stan
parameter R [4], is dependent on the compressive strength dard number of heats although the residual thickness was
(fc), Young’s modulus (E), and the coefficient of linear comparatively high. During the ladle cycle, shutdowns with
expansion (α). Therefore, a possible cause of vertical crack complete cooling of the ladle and subsequent partial tap-
formation in the brick is a high irreversible compression, ping into an insufficiently preheated ladle had occurred.
34 <
Investigation Procedures deeply worn and characterized by material loss and slag
coating. Additionally there was a vertical crack visible in the
The sample was investigated macroscopically and micro-
middle of the brick, which extended into the middle of the
scopically at the RHI Technology Center Leoben (Austria).
sample. This vertical crack was formed during operation and
The microscopic investigations were carried out on polished
also showed premature wear at the hot face surface. In the
sections by optical light microscopy using a Reichert
cut section, macroscopically the residual microstructure of
reflected light microscope and by scanning electron micros-
the brick appeared dense and compact up to the cold face.
copy (SEM) using a JEOL 6400 equipped with an energy
dispersive spectroscopy (EDS) analysis system to provide
Microscopic investigation of the zone containing the vertical
chemical microanalyses. To achieve the highest possible
crack revealed a slag coating at the hot face surface but pre-
quality and accuracy levels of the analysis results, investiga-
mature wear with increased decarburization and therefore
tions were carried out according to international standard
increased slag attack in the immediate vicinity of the vertical
procedures and calibrations were performed with interna-
crack (Figure 3). The slag coating was discontinuous; in the
tionally certified standards.
crack area at the surface it had broken open during cooling
down. In general there were different crack generations visi-
Results ble. One crack was mainly filled with steel and had reclosed
The investigated brick showed a residual thickness of about during further operation, whilst another crack had filled with
120 mm. The hot face surface was covered by a few mm slag (Figure 4). This indicated the possibility of ongoing, new
thick slag coating. One joint surface was exceptionally crack formation at every single thermal shock event.
Figure 2. Worn magnesia-carbon brick indicating premature

wear at the joints (1) and at a vertical crack (2). The vertical crack
had formed during operation.
(a)
1 2 mm
Figure 3. Reflected light image of the vertical crack area at the

hot face, with premature wear evident at the hot face. Decarbur-
(b)
ized zone showing slag attack (circle). Different crack generations
visible. One crack (green arrows) is mainly filled with steel
Figure 1. Prematurely worn steel ladle lining characterized by (bright) and has reclosed. Another crack (blue arrows) is filled
wear at joints and newly formed vertical cracks. (a) overview and with slag. For detail of the crack in rectangle 1 see Figure 4 and
(b) detail. for detail of rectangle 2 see Figure 6.
> 35
According to SEM-EDS microanalysis of the slag coating at

the hot face, the slag was of mayenitic type (i.e., 42.9 wt.%
Al2O3, 6.2 wt.% SiO2, 42.6 wt.% CaO, 0.7 wt.% MnO, 1.3
wt.% Fe2O3, and 6.3 wt.% MgO). At the hot face there was
4
slag coating and a decarburized microstructure. Decarburi-
zation was the cause of microstructure penetration by the 3 1
slag; therefore, increased corrosion of the magnesia com-
ponent, loss of bonding, and subsequent material loss by
hot erosion occurred. This type of wear is naturally
increased in an opened crack because of the increased reac-
tive surface. The corrosion mechanism of the magnesia at 2
the slag interface was characterized by formation of magne- 5
sium aluminate spinel and a MgO-containing slag phase
(Figure 5). Slag penetration into the microstructure was
only observed in the totally decarburized zone. The adjacent
partially decarburized zone showed no slag infiltration and
related corrosion. Here the residual carbon had obviously
stopped the infiltration and inhibited corrosion of magnesia Figure 6. SEM-BSE image of the hot face in the crack area (rectangle
embedded in the partly decarburized microstructure (Figure 2 in Figure 3) at the slag-brick interface showing slag attack after
6). Oxygen attack of the carbon-containing fines in the decarburization. Corroded MgO (1) with 4.9 wt.% Al2O3,
12CaO .7Al2O3 (2) from slag, residual graphite (3) in a partly decarbur-
matrix resulted in a partially decarburized microstructure ized area without any slag infiltration, magnesium aluminate spinel
with lower bonding strength (Figure 7). (4), and original, noncorroded magnesia fines (5) are indicated.
500 µm
100 µm
Figure 4. Reflected light image approximately 8 mm from the hot
(a)
face (rectangle 1 in Figure 3). Reclosed crack (1), partly filled with
steel (bright). Second crack, partly filled with calcium aluminate
slag (2). Minor decarburization and minor slag attack in this area.
5 3
1
100 µm
(b)
Figure 5. SEM-BSE image of the hot face, approximately 4 mm Figure 7. Reflected light image showing (a) partly decarburized
from the vertical crack at the slag-magnesia interface. Fused mag- microstructure approximately 0.4 mm from the hot face com-
nesia (1), magnesium aluminate spinel (2), MgO-containing pared to (b) the desired dense carbon bonding structure approx-
12CaO .7Al2O3 (3 and 4) slag phase, and an additional Ca2SiO4 (5) imately 20 mm from the hot face. Carbon components (brown-
slag phase are indicated. orange), magnesia (light grey), and open pore space (dark grey)
are visible.
36 <
There was a significant difference between the corrosion As a final consequence, joint opening as well as vertical
behaviour of magnesia with large single periclase (MgO) crack formation is highly dependent on ladle cycling, espe-
crystals and magnesia consisting of smaller single periclase cially the preheating conditions and refractory properties.
crystals. This was clearly depicted in the slag-infiltrated
crack without significant hot erosional impact where fused Magnesia-Carbon Brick Grade Development
magnesia with large MgO crystals showed significantly
From the refractory perspective, it is necessary to achieve a
higher resistance to slag attack (Figure 8) and grains with
permanent but controlled carbon-bonded magnesia-carbon
smaller crystals had lower resistance against decomposition
brick expansion during the ladle cycle. A well balanced
and a generally higher corrosion rate.
working lining stress state is essential to prevent joint open-
ing or discontinuous wear by spalling. This requires optimi-
Magnesia-Carbon Brick Wear Mechanisms zation of the microstructure regarding reversible thermal
The observed premature magnesia-carbon brick wear was a expansion and irreversible but controlled expansion. As indi-
combination of the following wear mechanisms: cated microscopically, the bonding structure with its carbon
components should demonstrate high resistance against
>> Initial thermomechanical wear with irreversible plastic oxygen attack. Additionally, the choice of the magnesia type
deformation of the hot face brick region resulting in ver- is important to minimize chemothermal attack on the micro-
tical crack formation and joint opening. structure. Fused magnesia with large single crystals shows
>> Chemothermal supply of oxygen that significantly the highest corrosion resistance. Furthermore, the brick’s hot
increased when the ladle was empty and preferentially compressive strength and “thermal stress resistance param-
occurred at the additional reactive surfaces in opened eter” [2] should be as high as possible.
joints and vertical cracks when the slag coating broke
open during thermal cycling. This led to decarburization Based on the aforementioned investigation results and con-
and partial decarburization of the microstructure. cluded requirements of the refractory material, the ANCAR-
>> The decarburized structure in the joints and vertical BON C brick series with controlled expansion and high
cracks became highly susceptible to deep reaching chemothermal resistance was recently developed [5].
chemothermal attack by slag and penetration by steel,
especially when the cold ladle went into operation. From the ladle operational point of view, measures can also
>> Slag attack was observed in decarburized microstructur- be performed to counterbalance the wear phenomenon dis-
al zones. cussed. A very important issue to minimize thermal shock is
>> Final increased material loss by hot erosion especially in ensuring a proper preheating procedure according to heating
thermomechanically and chemothermally affected joint up instructions and a high final preheat temperature before
and crack areas was observed. ladle filling. Severe and repeated thermal cycling due to dis-
>> Magnesia with a small periclase (MgO) crystal size continuous working practices such as shutdowns and tapping
showed lower corrosion resistance than fused magnesia into cold ladles should also be avoided. Furthermore, long idle
with large single crystals. times should be prevented to reduce oxidation and a high
>> The occurrence of different generations of vertical cracks number of heats per day generally reduce the degree of ther-
indicated that in the case of thermal cycling an ongoing mal cycling.
wear process like a chain reaction has to be expected.
In Service Performance of the Newly Devel-

oped ANCARBON C Brick Type
A customer was faced with severe premature wear in the
ladle. Characteristic for this plant was a suboptimal thermal
2
1 ladle situation, due to long waiting times, transport distances,
and chemical heating. The wear pattern showed cracks and
successive heavy spalling in the slag line, starting in the first
third of the ladle campaign (Figure 9). This led to intensifed
decarburization of the brick surface and subsequent infiltra-
tion and spalling of the infiltrated areas following sharp tem-
perature changes. The ladle lifetime could only be kept at the
usual level by increasing the gunning maintenance.
Several counteractive measures were examined and partially

tested. To lower the stresses in the slag line lining, an expan-
3 sion allowance for the bricks was considered to reduce the
thermomechanical stresses and a trial with special coated
bricks was performed. The coating was a substance that
burnt out during preheating to provide space for the thermal
500 µm expansion. In addition, the influence of the thermomechani-
cal behaviour of the brickwork was evaluated in trials using
Figure 8. Reflected light image of the vertical crack, 10 mm from different shapes. However, both measures did not provide
the hot face. Crack (1) filled with slag. Increased decomposition any improvement.
and corrosion by slag attack of magnesia with small periclase crys-
tals (2) compared to fused magnesia (3) with large MgO crystals.
Based on the described findings, further investigations led to
the conclusion that premature damage due to the formation
> 37
of cracks during preheating and the first heats was the Conclusion
main reason for the aforementioned wear pattern. The
Microscopic analyses provide comprehensive information
problem was subsequently solved by installing ANCAR-
regarding the wear behaviour of postmortem refractory
BON C bricks with controlled expansion and improving
materials. In the case of the ANCARBON C brick type, they
the ladle thermal situation by using lids and a better pre-
supported the fundamental physical, simulation, and pilot
heating practice in between heats. studies that resulted in this new product development.
Due to the special expansion behaviour of ANCARBON C, ANCARBON C was installed at a customer where severe
crack formation could be avoided and infiltration was premature ladle lining wear was occurring due to the
reduced. The brickwork temperature change during the demanding ladle cycle conditions. As a result of this brand
ladle cycle was less due to the accompanying measures change and improving ladle operating conditions, cracking
in the melt shop and spalling was completely eliminated. and spalling was avoided, the number of heats before the
Furthermore, the same lining lifetime was achieved with first gunning measure was significantly increased, and the
a massive reduction in the gunning requirement gunning requirement to achieve the same ladle lifetime was
(Figure 10). greatly decreased.
100
9
90
8
80
7
70
6
No. gunnings
60
Heats
50 5
40 4
30 3
20 2
10 1
0 0
Standard Controlled expansion
magnesia-carbon ANCARBON C type
n Total heats of one ladle campaign (standard magnesia-carbon)

n Total heats of one ladle campaign (ANCARBON C)
▲ Heats before first gunning measure
n No. gunning measures during one ladle campaign
Figure 9. Typical wear pattern of ladle bricks without controlled Figure 10. Ladle lifetime and gunning measures for a standard
expansion. magnesia-carbon lining design compared to ANCARBON C
bricks with controlled expansion.
References
[1] Gruber, D. and Harmuth, H. Durability of Brick Lined Steel Ladles from a Mechanical Point of View. Steel Research International. 2008, 79, No. 12,
913–917.
[2] Harmuth, H., Vollmann, S., Melcher, F., Gruber, D. and Majcenovic, C. Relevance of Numerical Simulation for Refractory Corrosion in Steel Indus-
try. Advances in Refractories V – The Michel Rigaud Symposium. Proceedings of the 49th Annual Conference of Metallurgists of CIM, Vancouver,
Canada, Oct., 3–6, 2010; pp. 453–463.
[3] Buchebner, G., Neuböck, R., Eder, J. and Studnicka, J. Thermomechanical Design of Magnesiacarbon Bricks for Steel Ladles. Presented at 51st
International Colloquium on Refractories, Aachen, Germany, Oct., 15–16, 2008; pp. 70–72.
[4] Kingery, W., Bowen, H. and Uhlmann, D. Introduction to Ceramics; John Wiley and Sons Inc: New York, London, Sydney, Toronto, 1976.
[5] Buchebner, G., Samm, V. and Rotsch, J. Latest Developments in Magnesia-Carbon Bricks. RHI Bulletin. 2011, No. 1, 23–28.
Authors
Christian Majcenovic, RHI AG, Technology Center, Leoben, Austria.
Johann Eder, RHI AG, Technology Center, Leoben, Austria.
Jens Rotsch, RHI AG, Steel Division, Vienna, Austria.
Corresponding authors: C
hristian Majcenovic, [email protected]
Jens Rotsch, [email protected]
38 <
RHI Bulletin > 1 > 2012, pp. 39–43
Shengli Jin, Dietmar Gruber, Harald Harmuth and Marc-Henri Fréchette
Thermomechanical Steel Ladle Simulation

Including a Mohr-Coulomb Plasticity Failure
Model
Due to the possible energy savings, insulating refractories are expected to play an ever more
important role in industrial furnaces and vessels. In order to quantitatively assess the influ-
ence of insulation on the thermomechanical behaviour of a steel ladle, Mohr-Coulomb plastic-
ity combined with a tension cut-off failure model was applied to the working lining during
simulations. The former describes the material failure under multiaxial compression condi-
tions and the latter states the material failure due to pure tension. The results show that both
failure mechanisms are found in the area close to the hot face of the working lining for mod-
els with and without insulation during the preheating and tapping processes. As a result of its
compliance, the insulation has a positive influence on decreasing compressive stresses act-
ing on the working lining and resulting irreversible strains in a circumferential direction. Fur-
thermore, tensile stresses in the steel shell are increased in a circumferential direction due to
the thermal influence of the insulation.
Introduction is essential for quantitative assessment of the thermome-

chanical behaviour of a steel ladle and provides more accu-
Modern steelmaking processes have altered the traditional
rate data regarding lining designs. In this paper the Mohr-
role of the steel ladle from a simple transport device to a
Coulomb plasticity failure model with tension cut-off was
more functional vessel in which secondary treatments are
applied to steel ladle working linings during the nonlinear
carried out, for instance RH degassing, alloying, and desul-
simulations. Irreversible plastic strains were investigated
phurization [1]. Usually the secondary treatments require
and failure mechanisms were identified for two typical ladle
the liquid steel to remain for a prolonged period in the
lining concepts with and without insulation.
ladle, which can lead to a temperature drop as high as
100 °C, mainly due to heat losses from the melt surface,
ladle walls, and bottom. Many efforts have focused on pre-
Material Behaviour
dicting and controlling liquid steel temperature fluctuations, Refractories are heterogeneous bulk ceramic materials that
from the refining process through to the ladle lining design, in many cases do not behave in a completely brittle man-
since the cast steel quality is strongly influenced by the liq- ner, and both elastic and inelastic behaviours may occur.
uid steel temperature [2–5]. The latter behaviour causes an irreversible displacement,
which brings about a so-called plastic strain after unload-
Decreasing heat losses through the vessel lining is one ing. The elastic limit is usually defined by a yield surface.
effective approach to control the temperature conditions Inelastic behaviour takes place at the yield surface, other-
and achieve energy savings and this can be realized with wise reversible elastic behaviour occurs. For the work
insulating refractories. By utilizing their low thermal con- described in this paper the Mohr-Coulomb yield surface
ductivities, the temperature drop of the liquid steel can be was applied:
reduced as well as the radiative and convective heat losses
from the steel shell. Systematic research on steel ladle lin-   c   tan (1)
ing concepts with and without insulation have been carried
out using linear elastic simulations from both thermal and Where c is the cohesion, φ the friction angle, σ the normal
thermomechanical points of view [6–8]. In these papers, the c cosτ the shear
stress, 2and 2c cos  on this plane. Theoretically,
stress
minimum efficient preheating time was investigated and the  tcohesion
the  ,  c friction angle can be calibrated from two
and
thermomechanical impact factors on the steel shell temper- simple  tan
  1ctests,
sin 1  sin a uniaxial tension test and a uni-
for example
ature and the lining stresses were classified. Highly accu- axial compression test [9]:
rate prediction models were generated, providing guidance
for the appropriate choice of ladle lining refractories. In  1  2c cos
1  
2c cos 
addition, the benefits and disadvantages of insulation dur-  t 3 cos  sin(, c 3 )  3 cos(   3 ) tan  q  p tan  c  0 (2)
ing an entire ladle cycle were qualitatively defined by com-
 1  sin  1  sin  
paring different lining concepts with and without insulation.
Where σt is the tension yield stress and σc the compression
11 r
Generally, linear elastic finite element (FE) simulation is less  stress.
yield arccos( )3  1  
time consuming than nonlinear elastic modelling, especially  3 sin(q  )  cos(   ) tan  q  p tan  c  0
 3 cos  3 3 3 
when simulating large models with complex geometries. By transforming equation 1 via Haigh-Westergaard stress
However, considering the nonlinear elastic lining behaviour coordinates, the general expression for the Mohr-Coulomb
1 1
p  1 I 1   r( 3 1   2   3 )
  3arccos( 3 ) > 39
3 q
1 1 2 3
q  33J 2  31   2   3   1 2   2 3   3 1
2 2 2
  cc   tan  (1)
  c   tan tan (1)
(1)
RHI Bulletin > 1 > 2012 2 2 2
q  3JJ2  ( 21   2 3 )( 212 23 )3( 21      )
 t  2c coscos 
22cc cos
 ,,  ccin terms cos 
22cc cos
2c cos of three stress invariants I1, J2, J3,
r 3 3 3  3 1 2
(2)
3 2 1 3 3 2 1
plasticity
t  1  sinmodel
 ,  c  11  sin  2 to the energy
equilibrated (2) consumption 2 due to the equiva-
 (2)
11  sin sin  Abaqus [9,10], can easily be
 sin
t lent plastic strain at the stress level of the cohesion:
as applied sin 
in the FE 1software J 3 3 ( 2 1   2   3 )  ( 2 2   1   3 )  ( 2 3   2   1 )
obtained: r 3 3 
 11  pl 2 1 2

 1  pl
 1 sin(    ))  11 cos(    )) tan   qq  pp tan  c 0  m   : d m (3) (8)
sin( cos( 
 sin(  33 )  33 cos(   33 ) tan  q  p tan
tan tan  cc  00 c (3)
 33 cos
cos 
(3) (3)
 3 cos  3 3 3 1
   : d
pl pl

Where pl
m em is the plastic
m strain due to multiaxial stresses.
1 rr 3 c2
The equivalent plastic strain from the tensile stress is
 11 arccos(     :  dt
pl pl pl

 r )3 (4)
  33 arccos( qq ))3
arccos( t (4)
(4) expressed
t as follows:
t
3 q
3 (4)
2 pl pl
 t    t :  t dt
pl
(9)
1 1 3
pp   11 II 1   11 ((  1   2   3 )) (5) (5) (5)
p   33 I 1   33 (  11   22   33 )
1 (5)
3 3 Where etpl is the plastic strain due to pure tension.
2 2 2
qq  33JJ 2   122   2 22   3 22   1 2   2 3   3 1 (6) Models and Boundary (6) Conditions
(6)
q  3J 22   11   22   33   11 22   22 33   33 11 (6) as a ladle metallurgical furnace
A typical steel ladle acting
during secondary treatment was chosen for the simulation.
JJ 3 3 (( 22 1   2   3 ))  (( 22 2   1   3 ))  (( 22 3   2   1 )) The ladle lining comprised MgO-C bricks in the working lin-
rr  33 33 J 33  33 ( 2 11   22   33 )  ( 2 22   11   33 )  ( 2 33   22   11 ) (7) ing, and fired MgO and(7)(7)chamotte bricks in the permanent
r 3 3 22  22 lining sidewall (Figure (7)
2). Simplifications were made to the
2 2
configuration and boundary conditions to avoid numerical
Whilst the Mohr-Coulomb criterion describes the behaviour problems and facilitate computation. Firstly, the steel ladle
1
  11  :: dd
pl
of different materials pl sufficiently well under compression, it
 pl pl was assumed to be symmetrical in a circumferential direc-
tion. A slice was cut in (8)
  cc  : d
m pl suitable form the
pl case of tensile loading. If the maxi- (8)

is less
m
m principal stress
m
m
an axial direction and the two verti-
(8)
mum
c σ1 reaches the tensile strength of the cal surfaces formed an angle of 2.76° in a radial direction.
material, the failure is associated with a tensile rather than Usually the slag zone experiences intense chemical and
a shear failure [11]. Therefore, a so-called tension cut-off thermomechanical impact; therefore, the simulation focused
22
  
equal pl to the Rankine
pl plcriterion was combined with the Mohr-
pl pl
2  ::  dt
pl
dt
on the linings in and above the slag zone including a steel
hoop (see Figure 2). In (9)
 
pl pl pl
t
Coulomb
t t
plasticity
t t
33  :  dt
model
t in the FE code Abaqus. Figure 1 (9)
addition, large integrated blocks
t t t (9)
provides 3
two representations of the Mohr-Coulomb com- replaced the bricks in the permanent lining. Model R is the
bined with the Rankine criterion. Figure 1a uses the invari- original lining concept without insulation whilst in model S
ants shown in equations 3, 5, and 6, where Rmc denotes the a 10-mm thick insulation layer (i.e., Pyrotek ISOMAG 70
coefficient of q in equation 3. Figure 1b is a representation XCO) replaced part of the chamotte brick thickness.
in the σ1/σ2 plane. When the tensile failure is predominant,
the yield surface will follow the Rankine criterion. For the simulation, the initial steel ladle temperature was
25 °C, equal to the ambient temperature. The preheating
In Abaqus, expressions for the equivalent plastic strain e- pl process lasted about 20 hours until the hot face of the work-
from the Mohr-Coulomb plasticity model and tensile cut-off ing lining reached 1100 °C. Liquid steel at 1660 °C was
are different. The former is expressed by equation 8, which instantaneously tapped from an electric arc furnace into the
is based on an energy balance. The energy consumed by preheated ladle up to its maximum capacity and held for
irreversible displacement caused by the principle stresses is several minutes before the secondary treatment. The heat
σ2
σ3 = 0
Rmcq
Mohr-Coulomb
Rankine
σt
σc σt σ1
Mohr-Coulomb c Mohr-Coulomb with

Mohr-Coulomb
tension cut-off
with tension
cut-off
Rankine
φ
σt p σc
(a) (b)
Figure 1. Mohr-Coulomb plasticity with tension cut-off. (a) including the invariants shown in equations 3, 5, and 6, where Rmc denotes
the coefficient of q in equation 3 and (b) representation in the σ1/σ2 plane.
40 <
transfer at the fluid/refractory interface was assumed to be restrained in an axial direction. The right vertical surface of
very high, whereas that between the linings was no more the model was restrained in the direction perpendicular to
than 60 W·m-2·K-1, which was inversely evaluated by verify- this surface because of the symmetry. Neighbouring bricks
ing the simulated temperatures of model R against in ser- were replaced by a fully constrained rigid body. Sliding and
vice data. Displacement of the entire model bottom was separation were allowed to happen at the interface between
the rigid body and the slice. The gap between the rigid
body and the slice was 0.2 mm, which was half the expan-
Model R Model S n Chamotte brick sion allowance for the lining interfaces. The steel shell was
n Fired MgO brick restrained in a circumferential direction and could freely
n MgO-C brick 1
expand in radial and axial directions.
n MgO-C brick 2
Insulation
n Steel Simulation Results
The cold round steel shell confined the refractory lining dis-
placement, caused by the thermal expansion, in a radial
direction. Consequently, the steel shell was subjected to
Steel hoop tensile stresses (Figure 3a), whilst the hot face of the work-
ing lining experienced compressive stresses in a circumfer-
ential direction (Figure 4a). In both models, the evident ten-
sile stresses were loaded on the steel shell after 4 hours
Insulation when the expansion allowance between linings had been
consumed. Between 4 and 8 hours into the preheating, the
(a) (b)
tensile stress increased dramatically; however, after 8 hours
the increase was much slower. In addition, model S showed
Figure 2. Three-dimensional representations of the steel ladle
and lining geometries used in the simulation. (a) model R with- slightly lower tensile stresses between 4 and 8 hours but
out insulation and (b) model S with insulation. distinctly higher tensile stresses after 8 hours in comparison
400 3.0
n Model R n Model R
350 n Model S n Model S
2.5
300
Tensile stress [MPa]
Displacement [mm]
2.0
250
200 1.5
150
1.0
100
0.5
50
0 0.0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
(a) (b)
Figure 3. Preheating and tapping period. (a) time-dependent tensile stresses in the steel shell in a circumferential direction and (b)
time-dependent displacement of the refractory lining in a radial direction.
50 6
n Model R n Model R
n Model S n Model S
5
40
Equivalent plastic strain [%]
Compressive stress [MPa]
4
30
3
20
2
10
1
0 0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
(a) (b)
Figure 4. Preheating and tapping period. (a) time-dependent compressive stresses on the lining in a circumferential direction and (b)
time-dependent equivalent plastic strains at the hot face of the working lining.
> 41
to model R. These results can be explained by the displace- rise to the counteracting sharply increased tensile stress in
ment of the entire refractory lining, which was determined the steel shell. The insulation in model S introduced an
from the difference between the total displacement of the additional expansion allowance, which contributed to the
steel shell and the steel shell displacement solely due to slightly lower tensile stress in model S compared to model
temperature (Figure 3b). Between 4 and 8 hours, the sharp R at the beginning of preheating. However, once the allow-
increase in the entire refractory lining displacement gave ance had been depleted, the insulating effect caused a
larger displacement of the refractory lining in model S than
model R (Figure 3b) due to the higher lining temperature.
Equivalent Equivalent After 8 hours the displacement increase of the entire refrac-
plastic strain plastic strain tory lining was slow due to the three dwell periods during
1.0 x10 -1
9.7 x10-2
8.0 x10-2 8.0 x10-2
the preheating. The high analogy between Figures 3a and
7.3 x10-2 7.3 x10-2 3b indicates that the lining displacement directly relates to
6.5 x10-2 6.5 x10-2
5.8 x10-2 5.8 x10-2
the tensile stress in the steel shell; namely, higher displace-
5.0 x10-2 5.1 x10-2 ment causes higher stresses.
4.3 x10-2 4.3 x10-2
3.5 x10-2 3.6 x10-2
2.8 x10-2 2.8 x10-2 Figure 4a shows that the compressive stress at the hot face
2.0 x10-2 2.1 x10-2
1.3 x10-2 1.4 x10-2
of the working lining increased sharply after 4 hours and
5.4 x10-3 6.4 x10-3 then slowly approached 43 MPa after 5 hours. Figure 4b
indicates that the plastic strain did not start until approxi-
mately 5 hours into the preheating and between 5 and 8
hours this plastic strain resulted in a moderate increase in
the compressive stress. Three dwell periods accounted for
(a) (b) the temporary decrease of the compressive stress and the
three plateau regions of the equivalent plastic strain.
Figure 5. Distribution of the equivalent plastic strain from multi- Although the maximum compressive stress in both models
axial stresses after tapping for (a) model R and (b) model S. was nearly the same, model S frequently showed lower
Equivalent plastic Equivalent plastic Equivalent plastic

strain from tensile strain from tensile strain from tensile
failure failure failure
2.696 x10-6 3.594 x10-4 5.355 x10-3
2.471 x10-6 3.295 x10-4 1.000 x10-3
2.247 x10-6 2.995 x10-4 9.167 x10-4
2.022 x10-6 2.696 x10-4 8.333 x10-4
1.797 x10-6 2.396 x10-4 7.500 x10-4
1.573 x10-6 2.097 x10-4 6.667 x10-4
1.348 x10-6 1.797 x10-4 5.833 x10-4 A
1.123 x10-6 1.498 x10-4 5.000 x10-4
8.987 x10-7 1.198 x10-4 4.167 x10-4
6.740 x10-7 8.985 x10-5 3.333 x10-4
4.493 x10-7 5.990 x10-5 2.500 x10-4
2.247 x10-7 2.995 x10-5 1.667 x10-4
2.220 x10-16 2.220 x10-16 8.333 x10-5
2.220 x10-16 B
(a) (b) (c)
Figure 6. Distribution of the equivalent plastic strain generated from tensile failure (a) 3.6 hours into preheating, (b) at the end of
preheating, and (c) at the end of tapping. The figures show the joint surface from the front.
0.10 0.6
n Model R n Model R
n Model S n Model S
0.5
0.08
0.4
0.06
Tapping 0.3 Tapping
0.04
Preheating 0.2
Preheating
0.02
0.1
0.00 0.0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
(a) (b)
Figure 7. Time-dependent equivalent plastic strains determined from tensile failure at (a) point A at joint surface and (b) point B at
protruding part.
42 <
values due to compliance of the insulation. Consequently, circumferential direction (Figure 6c and 7b). The reason for
the equivalent plastic strain in model S was less than that in these phenomena is the free expansion of the radial joint
model R at the corresponding time. surface and hot face of the working lining. Therefore,
spalling is predicted to happen at the radial joint surface at
The distribution of the equivalent plastic strain was also the beginning of the preheating period and the cracks that
compared for the working lining of both models after tap- form would propagate at an angle of approximately 45°
ping. The grey shaded areas indicate where the equivalent inclining to the hot face and at some distance from the edge
plastic strain was larger than 0.08 (Figure 5). In each model, during the tapping from the electric arc furnace to the ladle.
the lower left corner of each brick (hot face) showed higher
equivalent plastic strains, and the lower part of the entire Conclusions
working lining had a larger area with high equivalent plastic
strains. Compared to model R, model S showed a smaller Quantitative results regarding the irreversible plastic strains
volume with equivalent plastic strains exceeding 0.08. and failure mechanisms in ladle working linings, with and
without insulation, were obtained from simulations that
Tensile failure started at the protruding brick even as early incorporated the Mohr-Coulomb plasticity with tension cut-
as 3.6 hours into the preheating time for both models (Fig- off. The modelling indicates that shear and tensile failures
ure 6a) but only lasted about 20 minutes (Figure 7a). As are inevitable in the lining concepts examined. Furthermore,
shown in Figure 6b, nearly half of the radial joint surface the positive effects of an insulation layer, for example main-
underwent tensile failure, which began at some distance taining the liquid steel temperature by minimizing heat loss
from the working lining hot face. The tapping caused much from the steel shell and reducing the irreversible strain in
more intense tensile failure at the lower corners of the pro- the working lining, are more beneficial than the negative
truding bricks and the failure penetrated the entire brick in a effects on the thermomechanical behaviour of the ladle.
References
[1] Totten, G., Funatani, K. and Xie, L. (Eds). Handbook of Metallurgical Process Design; Marcel Dekker Inc: New York, 2004.
[2] Zimmer. A., Lima, À.N.C., Trommer, R.M., Bragança, S.R. and Bergmann, C.P. Heat Transfer in Steelmaking Ladle. Journal of Iron and Steel
Research, International. 2008, 15, No. 3, 11–14.
[3] Rahm, C., Kirschen, M. and Kronthaler, A. Energy Savings Through Appropriate Ladle Lining Concepts. RHI Bulletin. 2008, No. 1, 38–43.
[4] Min, Y. and Jiang, M. Exergy Analysis and Optimization of Ladle Furnace Refining Process. Journal of Iron and Steel Research, International.
2010, 17, No. 11, 24–28.
[5] Glaser, B., Görnerup, M. and Sichen, D. Thermal Modeling of the Ladle Preheating Process. Steel Research International. 2011, 82, No.12, 1425–
1434.
[6] Jin, S., Harmuth, H., Gruber, D., Auer, T. and Li, Y. Classification of Thermomechnical Impact Factors and Prediction Model for Ladle Preheating.
Journal of Wuhan University of Science and Technology. 2011, 34, No. 1, 28–31.
[7] Auer, T., Gruber, D., Harmuth, H., Jin, S. and Kronthaler, A. Thermo-Mechanical Simulation of a Steel Ladle Process Cycle With Special Emphasis
on the Preheating Process. Presented at 54th International Colloquium on Refractories, Aachen, Germany, Oct., 19–20, 2011.
[8] Jin, S., Harmuth, H., Gruber, D., Auer, T., Fréchette, M-H. and Li, Y. Thermo-Mechanical Modeling of a Complete Steel Ladle Process. Presented
at UNITECR2011, Kyoto, Japan, Oct., 30–Nov., 2, 2011.
[9] Chen, W.F. and Han, D.J. Plasticity for Structural Engineers; Springer-Verlag: New York, 1988.
[10] Abaqus 6.10 Analysis User’s Manual; Dassault Systèmes, Providence, USA, 2010.
[11] Gross, D. and Seelig, T. Fracture Mechanics: With an Introduction to Micromechanics; Springer-Verlag: Berlin, Heidlberg, 2006.
Acknowledgements
Financial support from the Austrian Federal Government and the Styrian Provincial Government (Österreichische Forschungsförderungsgesellschaft
and Steirische Wirtschaftsförderungsgesellschaft) within the K2 Competence Centre on “Integrated Research in Materials, Processing and Product
Engineering” (MCL Leoben) in the framework of the Austrian COMET Competence Centre Programme is gratefully acknowledged.
Authors
Shengli Jin, Chair of Ceramics, University of Leoben, Austria.
Dietmar Gruber, Chair of Ceramics, University of Leoben, Austria.
Harald Harmuth, Chair of Ceramics, University of Leoben, Austria.
Marc-Henri Fréchette, Pyrotek Inc., Drummondville, Quebec, Canada.
Corresponding author: Shengli Jin, [email protected]
> 43
RHI Bulletin > 1 > 2012, pp. 44–49
Matthias Höck, Robert Sorger, Christoph Eglsäer and Günter Deutsch
Consequences of REACH on the Use of Ceramic

Mineral Fibres
Introduction In addition to REACH, European authorities together
with national authorities introduced the regulation on
Refractory ceramic fibre (RCF), a high-temperature insu-
classification, labelling, and packaging of chemicals (CLP
lation wool, consists of several types of man-made vitre-
Regulation), which completely replaced the Dangerous
ous (silicate) fibres (i.e., aluminosilicate wool (ASW))
Substance Directive in 2010 and after a transitional
that are used for various high-temperature industrial
period will replace the Dangerous Preparations Directive
applications, including in the lining of metallurgical ves-
by 2015. The CLP Regulation is the European version of
sels [1]. Additional high-temperature insulations, devel-
the United Nations Globally Harmonized System of Clas-
oped after RCF, are alkaline earth silicate wool (AES) and
sification and Labelling of Chemicals (GHS), which aims
polycrystalline wool (PCW). The advantage of all these
to standardize labelling and the description of hazards
materials is they demonstrate high-temperature and
worldwide.
thermal shock resistance as well as low thermal conduc-
tivity. However, there are health concerns regarding RCF
It is a comprehensive approach to [3]:
since the fibres are small enough to penetrate deep into
the lungs. Although evidence linking these fibres to any
>> Describe the health, physical, and environmental haz-
human disease has been questioned, in the late 1990s
ards of chemicals.
RCFs of a particular size were classified as a carcino-
>> Create a classification process that uses available infor-
genic risk and any work involving these materials
mation on chemicals (e.g., REACH data) for comparison
became subject to more stringent control. Subsequently
with defined hazard criteria.
in 2007, the European Union (EU) Regulation termed
>> Define communication measures for hazard information,
Registration, Evaluation, Authorization and Restriction of
as well as protective measures in the form of labels and
Chemicals (REACH) came into force requiring many
Material Safety Data Sheets (MSDS).
chemical substances on the market in Europe to be reg-
istered, and has resulted in concerns about the future
availability of certain high-temperature insulation wools. REACH Registration of Fibre Materials and
Classification in the CLP Regime
REACH and the CLP Regulation In line with the REACH Regulation, manufacturers and
The REACH Regulation (EC) No. 1907/2006 requires importers of RCF/ASW, AES, and PCW submitted a joint
manufacturers and importers to gather specific data on registration dossier prior to the 2010 deadline, which
a chemical substance, including the potential health and enables these materials to currently be used on the
environmental hazards, and formally register this infor- European market. However, due to the material proper-
mation in the European Chemicals Agency (ECHA) data ties of RCF/ASW, the registration of these substances
base. Prior to this legislation there were a host of differ- will be subject to additional investigations. The reason
ent directives and regulations relating to chemical sub- for this is RCF/ASW were previously classified by the
stances; however, the system did not generate sufficient International Agency for Research on Cancer and the
information about the effects concerning the majority of EU’s Dangerous Substance Directive as a carcinogen cat-
existing chemicals (i.e., predating 1981) on human egory 2, which means substances that should be
health and the environment from the public authorities’ regarded as if they are carcinogenic to humans [4].
perspective [2]. In addition to the focus on improving Under the new CLP Regulation this is equivalent to a
health and environmental protection, an aim of REACH carcinogen category 1B “Known or presumed human
is to enhance global competitiveness of the EU chemical carcinogen; presumed to have carcinogenic potential for
industry by requiring research and innovation into safer humans, classification is largely based on animal evi-
substances. Furthermore, the regulation provides for the dence”. As a consequence, RCF/ASW have been listed
progressive substitution of dangerous chemicals, when under the EC index number 650-017-00-8 in Annex I of
suitable alternatives are identified. the Dangerous Substances Directive and under Annex VI
part 3 table 3.1 of the new CLP Regulation, which speci-
REACH is very broad ranging, covering all substances fies that the information depicted in Figure 1 is required
whether manufactured, imported, used as intermediates, on the label.
or placed on the market either on their own, in prepara-
tions, or in articles [2]. It is based on the idea that indus- In contrast, AES was exempted from classification as a
try is best positioned to assess the substances it manu- carcinogen after scientific assessment [4] since it fulfils
factures and markets, shifting the burden of proof from the exoneration criteria of the European Dangerous Sub-
public authorities to manufacturers, importers, and users. stances Directive based on biopersistence tests or long-
It also requires that the industry has certain knowledge of term animal studies. This means the fibres are suffi-
the substance properties and how to manage potential ciently soluble in vivo (i.e., biosoluble). Since AES fibres
risks, with the goal that information on hazards and risks containing > 18 wt.% alkali and earth alkali metal oxides
will be passed both down and up the supply chain. are not considered to be carcinogenic, they are
44 <
increasingly being used as a substitute for RCF/ASW and Fibre-Related Issues Concerning ASW/RCF
although they did not tolerate such high temperatures
The problematic status regarding the danger of RCF/
as RCF, the difference between the possible application
ASW has a long history and arises from the similarities
temperatures is decreasing, or in some cases no longer
between man-made fibres and asbestos. This has
exists, due to extensive AES product development. How-
resulted in high-temperature insulation fibres being
ever, for an international acting company like RHI it is
more extensively tested than any other material [4].
important to be aware that AES fibres have not been
Prior to REACH and CLP, RCF was classified as carcino-
evaluated by the International Agency for Research on
gen category 2. However, any doubts regarding the
Cancer or US authorities to date [5], especially since RCF
accuracy of this classification, since it has not been
and AES have been found to produce very similar expo-
confirmed as a human carcinogen [7], and new data
sure concerns during manufacture and use [6].
presented by the manufacturers did not influence the
subsequent process resulting in certain RCF/ASW and
PCW were not classified prior to the registration, and
zirconia aluminosilicate RCFs being included on the
self-classification led to the conclusion that this material
candidate list of substances of very high concern
type is not hazardous. However, the ECFIA, which repre-
(SVHC) and a potential authorization requirement. A
sents the European high-temperature insulation wool
main factor resulting in inclusion on the list is the fibre
industry, has highlighted that in Germany under TRGS
structure. The important properties in this context are
905 providing technical guidance on worker protection
fibre length, diameter, and bioavailability (i.e., degrada-
(supplementing or extending Annex 1 of the EU Danger-
tion rate in biological fluids) [8]. In humans, fibres with
ous Substance Directive) fibrous dusts emanating from
a diameter of > 3 μm are essentially nonrespirable,
the handling of PCW are classified as Category K3:
whereas the greatest pulmonary deposition occurs for
namely “cause concern for man owing to possible carci-
fibres with a diameter of ~ 1 μm and a length of ~ 8
nogenic effects but in respect of which the available
μm. Subsequent clearance of deposited fibres (i.e., bio-
information is not adequate for making a satisfactory
solubility) is also a function of the length to diameter
assessment” [4]. ECFIA also stated that although the
ratio. Fibres with a length smaller than the diameter of
TRGSs are not laws they are technical rules/recommen-
macrophages (i.e., 15 μm) are phagocytized and
dations that have a quasi-legal status.
removed, either by transport via the mucociliary sys-
tem or to local lymph nodes. [8,9]. However, dust parti-
cles with a length to diameter ratio exceeding 3:1, and
a length longer than 5 µm and a diameter smaller than
3 µm, so called “WHO-Fibres”, are considered health
critical. In the case of RCF/ASW, fibrous dusts can be
emitted that include fibres complying with the WHO
definition.
Possible REACH Authorization Requirement

for Fibre Materials
Authorization Procedure
Authorization from the ECHA will be required to use
and place SVHC on the market. The authorization
applies to substances that are carcinogenic, mutagenic,
or toxic for reproduction and/or persistent, bioaccumu-
(a) lative, and toxic [2]. As previously described, the RCF/
ASW fall into the category of potential carcinogenic
products.
Authorization is the means by which SVHC can be reg-

ulated centrally in a manner that ensures the risks
related to their actual uses are assessed, considered,
and the availability decided upon by the European
Community. The aim of authorization is to ensure that
substances are progressively replaced by suitable alter-
native substances, where economically and technically
viable. Applicants for authorization will have to include
plans in the authorization application to replace the use
of the SVHC with a safer alternative or, if no safer alter-
native exists the applicant must argue the socio-eco-
nomic benefit of the substance use [2]. As of January
2012, there are 73 SVHC nominated for a future author-
(b) ization requirement [10]. This list will be extended con-
tinuously. However, once a substance has been
included on the candidate list it will never be removed.
Figure 1. Labelling required for (a) carcinogen category 2 materi-
al under the EU Dangerous Substances Directive and (b) catego- As a result, inclusion in the authorization process is
ry 1B under the CLP Regulation [4]. only a question of time.
> 45
Fibre Materials and Authorization properties, and workers’ protection measures have to be
described.
On the initiative of a Member State’s Competent Author-
>> Required notification of articles containing the sub-
ities (CA) for the REACH process, two specific types of
stance in a concentration above 0.1 wt.% to the ECHA.
RCF/ASW products were included on the SVHC candi-
>> An authorization process may be necessary in the future
date list in January 2010. The candidate list was subse-
and in this case substitution will become obligatory for
quently extended on December 19, 2011, to include a
the industry.
further two RCF/ASW substances. The information nec-
essary for the inclusion in the SVHC candidate list was
It is the authorization requirement that will have the
stated in a Member State’s dossier. ECHA published an
most significant effect on fibre producers and users in
exact definition of the fibre substances included in the
the future, since the impact of such an authorization pro-
candidate list (Table I).
cess for the industry will be the search for feasible alter-
natives.
Industry experts and independent scientists have com-
plained that there are several important errors and
issues in the Member State’s dossier including: Planning for Substitution
One of the fundamental aims of authorization is the
>> Classification as a substance: Ceramic fibres do not fit replacement of the SVHC listed in Annex XIV of REACH
the official definition of a substance. According to the by suitable alternatives or techniques that are economi-
REACH legal definition substances are “a chemical ele- cally and technically feasible (Figure 2). If in the future a
ment and its compounds in the natural state or obtained substitution becomes necessary it will require [2]:
by any manufacturing process, including any impurities
and additives necessary to preserve its stability [11]”. >> An analysis of alternatives: It is a required element in
>> Stated worker exposure and risk management measures authorization applications and provides the basis to
(RMM) regarding the production and use of ASW/RFC: assess whether alternative substances or techniques are
Industry experts found a discrepancy between the RMM available.
stated in the Member State’s dossier and those currently >> Whether the transfer to alternatives would result in
used in the production and use of the fibres. reduced overall risks to human health and the environ-
ment.
Regardless of the concerns raised by manufactures and >> The technical and economic feasibility of alternatives.
downstream ceramic fibre user associations (e.g., PRE), and >> A substitution plan: Where the analysis of alternatives
the errors in the Member State dossier, inclusion in the shows that suitable alternatives are available.
SVHC candidate list has occurred and triggers a sequence
of obligations. These include: A granted authorization will be subject to a time-limited
review. The duration of this review period will be deter-
>> General aspects like providing a Material Safety Data mined on a case-by-case basis. The authorization is also
Sheet where the product composition, hazardous bound to a €50000 fee per substance, use, and applicant.
Substance name Description Inclusion date

Fibres fulfil the following conditions
Zirconia aluminosilicate RCF >> Oxides of aluminium, silicon and zirconium are the main components present (in the fibres) within 19/12/2011
variable concentration ranges
>> Fibres have a length weighted geometric mean diameter less two standard geometric errors of ≤ 6 µm
>> Alkaline oxide and alkali earth oixde (Na2O+K2O+CaO+MgO+BaO) content ≤ 18 wt.%
Aluminosilicate RCF >> Oxides of aluminium and silicon are the main components present (in the fibres) within variable 19/12/2011
concentration ranges
>> Alkaline oxide and alkali earth oxide (Na2O+K2O+CaO+MgO+BaO) content ≤ 18 wt.%
Aluminosilicate RCF >> Al2O3 and SiO2 are present within the following concentration ranges: 13/01/2010
Al2O3 43.5–47 wt.%, and SiO2 49.5–53.5 wt.%, or Al2O3 45.5–50.5 wt.%, and SiO2 48.5–54 wt.%
Zirconia aluminosilicate RCF >> Al2O3, SiO2 and ZrO2 are present within the following concentration ranges: 13/01/2010
Al2O3 35–36 wt.%, and SiO2 47.5–50 wt.%, and ZrO2 15–17 wt.%
Table I. Description of ceramic fibres on the SVHC candidate list [10]. All materials have the index number 650-017-00-8 in Annex VI,
part 3, table 3.1 of Regulation (EC) No. 1272/2008 of the European Parliament and of the Council of 16 December 2008 on classifica-
tion, labelling and packaging of substances and mixtures.
46 <
Timeline for a Possible Authorization The following examples describe where such substitu-
tions have already been successful or are being actively
As previously described, once a substance is listed as a
evaluated. Table II details some of the currently used
candidate SVHC it will never be removed. The require-
alternatives to RCF/ASW.
ment to apply for authorization depends on the prioriti-
zation process and the subsequent inclusion in Annex
XIV of the REACH regulation. This process is very much
RCF/ASW Substitution
dependent on the inherent substance properties and in AES (Organic fibres)
the case of RCF/ASW relates to the carcinogenic poten-
tial. According to current information, substances with PYROSTOP ROPE 1012 PYROSTOP ROPE SW 1200
persistent, bioaccumulative, and toxic properties will be PYROSTOP FORM 1260 PYROSTOP FORM SW 1250
prioritized. At the moment it is not possible to reliably
estimate how long RCF/ASW fibres can be used for with- PYROSTOP BOARD 1260 PYROSTOP BOARD SW 1260
out an authorization; however, RHI is committed to
replacing the RCF/ASW products in various applications. Table II. Examples of RCF/ASW alternatives currently used by RHI.
Identification as a
substance of
very high concern
Prioritization process
Member State
consultation
Substance is listed in
Annex XIV of REACH
No Is the risk during Yes

production/use adequately
controlled?
Assess suitable Assess suitable

alternatives alternatives
No Yes No Yes
R&D plans
Socio-economic analysis– Market Submission or submit
Substitution plan
benefits outweigh risk exit scientific report:
“No Safer Alternative Available”
Yes
Authorization
application
• Time limited
• For each use
• Review case-by-case
Figure 2. Flow chart of the REACH authorization and the substitution process.
> 47
Insulation and Gasket Applications in the Steel this application because they are quick and easy to install.
Industry Aiming for minimum thickness enables a slim design of the
entire lining and maximum tundish volume. Alternatives to
Steel Treatment Ladles ceramic fibre board insulation include insulating bricks or
Good insulation is an important part of the layered refrac- insulating gunning mixes. However, biosoluble fibre materi-
tory lining in a ladle (Figure 3) [12,13]. The steel shell has to als can also be considered for the tundish application
be protected from high temperatures coming from the liq- because typical temperatures in the installation area are
uid steel inside the ladle. If the insulation does not work below the application temperature of biosoluble fibre mate-
properly there is a risk of the steel shell being deformed, rials.
added maintenance for the shell, loss of containment, and
disruption of operations.
Apart from low thermal conductivity, the ability to maintain

dimensional stability over multiple campaigns in the ladle is
Steel shell
a key parameter for the correct choice of the insulating
material. Compression of the insulation layer can lead to
Permanent lining
shifting of the wear lining, joint opening, and vertical cracks
(Figure 4).
Insulation layer
RCF/ASW-carton material (e.g., PYROSTOP CARTO), with a
temperature resistance of up to 1250 °C, has been one of Working lining
the best choices for this application in the past. Up to now,
a biosoluble version of this type of material has not been Backfilling mix
able to achieve the same resistance against the combined
(a)
high thermal and compressive loads. The main alternative
is currently a vermiculite-based brick, but it has a higher
thermal conductivity. Therefore, research into alternative
insulating materials is currently underway. Steel shell
Shifted permanent lining

Ladle Covers
In the case of ladle covers and roofs for ladle preheaters,
easy to install KONTIBLOCK fibre bricks are a popular
choice. However, especially in the case of ladle covers Open joints
where there is minimal distance to the liquid steel and slag,
a layer of high-alumina mix is recommended to protect
such biosoluble materials from disintegration. Alternatively, Compressed insulation layer
lining with high-insulating castables can be performed.
Shifted working lining
Tundish (b) Shifted backfilling mix

In regard to the achievable sequence length and process
safety, tundish insulation is of major importance to prevent
heat loss. An appropriate insulation design results in slower
Figure 4. (a) cross section through a ladle depicting the multiple
heating up of the steel shell and efficient tundish use (Fig- lining layers and (b) effect of compressed insulation layer on the
ure 5). RCF/ASW boards are well established for wear lining.
1800 1600
MgO-C wear lining High-alumina Steel 1550 °C
Backfilling mix
Insulation
1 worn to 100 mm permanent lining n ANKERTUN 217

1600 1400
2 DIDURIT B83
1
PYROSTOP BOARD 1260
1400 2 1200
n ANKERTUN 217
1100 °C DIDURIT B83
1200 1000
Temperature [°C]
Temperature [°C]
3 PYROSTOP CARTO 125

3 1080 °C 910 °C
1000 960 °C 800
685 °C
600 °C
800 600
600 1 Steady state temperature [°C] 400

2 Maximum transient temperature [°C] 288 °C
400 3 Minimum transient temperature [°C] 200 290 °C 220 °C 218 °C
Wear lining Permanent lining Insulation Shell
1, 2, 3
200 0 Wear lining Permanent lining Insulation Shell
0 50 100 150 200 250 0 40 80 120 160 200 240 280
Thickness [mm] Material thickness [mm]
Figure 3. SIMU-THERM heat flow simulation of the temperatures Figure 5. Heat transfer curve of two tundish lining concepts.
across the various lining materials and steel shell for a ladle dur-
ing operation.
48 <
Isostatically Pressed Products formed shapes, RHI extended its product portfolio to include
fully biosoluble AES and thereby provides full flexibility for
Thermal insulations and preformed gaskets for isostatically
all customer-specific requirements. RHI’s DELTEK Eco Insula-
pressed products, used in the steel continuous casting pro-
tion and DELTEK Eco Gasket show excellent high-tempera-
cess, are characterized by their high-temperature stability,
ture performance characteristics, whilst also satisfying all
low thermal conductivity, and required flexibility. Tradition-
demands imposed by the European regulatory requirements.
ally, RCF/ASW has been the base material for blankets and
formed shapes utilized for:
Conclusion
>> Sealing between the steel ladle collector nozzle and ladle RHI has to carefully monitor the further developments
shroud (Figure 6a). regarding the impact of REACH legislation on the status and
>> Insulation of submerged nozzles (Figure 6b and c). availability of high-temperature insulation wools. In steel
>> Sealing between the tundish nozzle and submerged entry applications RHI is well prepared to provide environmentally
shroud. and user-friendly solutions, although for some applications
further development of fibre materials is necessary. Thereby
Anticipating potential restrictions in relation to the applica- a possible authorization process will neither lead to an inter-
tion and handling of RCF/ASW-containing insulations and ruption in supply nor to a decrease in lining quality.
Gasket
Insulation
(a) (b) (c)
Figure 6. (a) ladle collector nozzle and ladle shroud, (b) monotube, and (c) thin slab submerged entry nozzle.
References
[1] http://www.hse.gov.uk/foi/internalops/fod/oc/200-299/267_3v2.pdf
[2] “REACH in Brief”. Environment Directorate General, European Commission, Brussels, October 2007.
[3] A Guide to the Globally Harmonized System of Classification and Labelling of Chemicals (GHS). http://www.osha.gov/dsg/hazcom/ghs.html
[4] www.ecfia.eu/has_cal.htm
[5] Linnainmaa, M., Kangas, J., Mäkinen, M., Metsärinne, S., Tossavainen, A., Säntti, J., Veteli, M., Savolainen, H. and Kalliokoski, P. Exposure to
Refractory Ceramic Fibres in the Metal Industry, Ann. Occup. Hyg. 2007, 51, No. 6, 509–516.
[6] Class, P., Deghilage, P. and Brown, R. Dustiness of Different High-Temperature Insulation Wools and Refractory Ceramic Fibres. Ann Occup Hyg.
2001, 45, No. 5, 381–384.
[7] “Synthetic Mineral Fibres (SMF) and Occupational Health Issues Position Paper”. Prepared by Australian Institute of Occupational Hygienists
Exposure Standards Committee, Tullamarine, October 2011.
[8] “Recommendation From the Scientific Committee on Occupational Exposure Limits for Refractory Ceramic Fibers”. SCOEL/SUM/165, October
2010.
[9] Mast, R., Maxim, L., Utell, J. and Walker, A. Refractory Ceramic Fiber: Toxicology, Epidemiology, and Risk Analyses – A review. Inhal Toxicol.
2000, 12, 359–399.
[10] ECHA Website - Candidate List of Substances of Very High Concern for Authorisation. http://echa.europa.eu/chem_data/candidate_list_en.asp
[11] http://www.reachonline.eu/REACH/EN/REACH_EN/article3.html
[12] Maupin, M. Insulation of Steel Ladles. AISTech 2004 Proceedings (vol. I), Nashville; USA, 2004; pp. 1215–1220.
[13] Rahm, C., Kirschen, M. and Kronthaler, A. Energy Savings Through Appropriate Ladle Lining Concepts. RHI Bulletin. 2008, No. 1, 38–43.
Authors
Matthias Höck, RHI AG, Steel Division, Vienna, Austria.
Robert Sorger, RHI AG, Steel Division, Vienna, Austria.
Christoph Eglsäer, RHI AG, Steel Division, Vienna, Austria.
Günter Deutsch, RHI AG, Technology Center, Leoben, Austria.
Corresponding author: Günter Deutsch, [email protected]
> 49
RHI Bulletin > 1 > 2012, pp. 50–54
Thomas Drnek and Michaela Seelig
Resource Efficiency—Global Context, European

Policy Initiatives, and RHI’s Responses
Triggered by growing energy and raw material prices, a high dependency on certain critical
raw materials only available from a few countries, as well as rising demand for resources
globally, resource efficiency is gaining increasing importance. In the past couple of years, the
European Union has been focusing on this issue and has come up with a variety of policy
approaches and initiatives that deal with sustainable ways to use resources and render eco-
nomic growth more resource efficient. Industry plays a crucial role in this context, and RHI
contributes to resource efficiency by, inter alia, applying sustainable and innovative sourcing
methods, reducing the specific energy consumption for magnesia sintering, and continuous
process improvements.
Resource Efficiency in a Global Context

60
Resource efficiency is becoming ever more relevant in the
context of rising energy and raw material prices on a n Metals
50
global level and the high dependency on some critical n Fossil fuels
industrial raw materials with limited geographical availa- n Industrial and construction materials
40 n Biomass
bility. Global trends such as population increases and ris-
ing resource demands, especially in emerging economies
30
Tonnes per capita
such as China, India, and Brazil, are further giving impe-

tus to the development of ways to source and use
resources as sustainably and efficiently as possible. The 20
world population will grow from 6 billion in 2000 to 9–10
billion in 2050, which is equivalent to an increase of 50%. 10
Even if constant raw material consumption per capita is
assumed, by 2050 the total raw material consumption will 0
have increased 50% [1]. Furthermore, given the fact that 1980 2008 1980 2008 1980 2008 1980 2008 1980 2008 1980 2008 1980 2008
European Union (EU) legislation regarding energy and Africa Asia Average Europe Latin North Oceania
world America America
environmental issues is manifold with ambitious reduc-
total
tion targets for CO2, other industrial emissions and
energy efficiency issues as well as resource efficiency
have become ever more relevant during the last decades. Figure 1. Global resource extraction per capita by world region
for 1980 and 2008 [2].
Figure 1 illustrates global resource extraction (only eco-

nomically used extraction) per capita in 1980 and 2008 for 70
the major material categories (i.e., metals, fossil fuels,
GDP per capita [US$ thousand]; Metal production [million tonnes]
industrial and construction minerals, and biomass) [2]. In

World GDP [US$ trillion]; World population [billion people];
60
2008, the highest per-capita resource extraction was
observed for Oceania, an increase of 21% compared to 50
1980. North America ranked second in 2008 with 26.1
tonnes per capita, due to high extraction levels of indus- 40
trial minerals, fossil fuels, and biomass; however, the
30
amounts were lower than in 1980. The developing
regions of Africa and Asia were characterized by the low-
20
est per capita numbers in 2008, with 5 and 9 tonnes per
capita, respectively. The world average per capita extrac- 10
tion increased from 8.6 tonnes in 1980 to 10.1 tonnes in
2008, equivalent to 17%. Concurrently, Europe decreased 0
its resource extraction between 1998 and 2008. 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2008
Year
Figure 2 illustrates the development of world gross n World GDP n Copper
domestic product (GDP) between 1960 and 2008, which n World Population n Aluminium
n GDP per capita n Zinc
grew six times during this period, whereas GDP per cap-
ita only increased very slightly. In comparision, alumin-
Figure 2. World development of gross domestic product, popula-
ium production grew around four times, while copper
tion, gross domestic product per capita, and aluminium, copper,
and zinc increased at a lower level between 1960 and and zinc production levels from 1960 to 2008. The US$ figures
2008. are based on equivalent values in 2008 [3].
50 <
European Union Policy Initiatives innovative and research input from the private and pub-
lic sectors is required to rapidly and efficiently tackle
The policy responses by the EU to the challenges of
major societal challenges. In its recently published “Mak-
resource efficiency are manifold and can be found in a
ing raw materials available for Europe‘s future well-being
wide variety of policy initiatives. Already in its communi-
– proposal for a European innovation partnership on raw
cation “The raw materials initiative – meeting our critical
materials”, the European Commission recognized that
needs for growth and jobs in Europe” [4], published in
innovation is an essential precondition and key driver to
November 2008, and its follow-up communication in
improve efficient resource use and sustainable raw mate-
2011, titled “Tackling the challenges in commodity mar-
rial supply, as well as maintain and improve the competi-
kets and on raw materials” [5], the European Commis-
tiveness of the EU industry. Furthermore, the Commis-
sion highlighted that increasing resource efficiency, pro-
sion highlighted the importance of innovation along the
moting recycling, and thereby reducing the EU’s primary
entire raw materials value chain, and that a comprehen-
raw materials consumption, needed to be one of three
sive approach is required to address the various chal-
pillars to achieve raw material supply security in the long
lenges the EU will face in the future [10].
term.
To give impetus to the process, and speed up research

Under the Europe 2020 strategy for growth, which was
efforts and breakthrough technologies, the European
adopted by the European Council in 2010, resource effi-
Commission proposed concrete targets in this context
ciency is one of seven flagship initiatives [6] to generate
that should be achieved by 2020, such as [10]:
employment opportunities and growth in Europe as well
as boost competitiveness. In September 2011, the Euro-
pean Commission subsequently published its “Roadmap >> European standardized statistical instruments for the
to a resource efficient Europe” [7], presenting a strategic survey of resources and reserves (land and marine)
policy framework on how to achieve a more sustainable and a three-dimensional geological map.
way to use resources and make economic growth >> A dynamic modelling system linking trends in supply
resource efficient. and demand with economical exploitable reserves
and a full life cycle analysis including an assessment
The main objective of the Commission’s roadmap is to of the environmental, economic, and social impacts
achieve a competitive European economy by producing of various scenarios.
more with less resource use and input. By 2013, the Euro- >> Up to 10 innovative pilot actions (e.g., demonstration
pean Commission aims to establish resource efficiency plants) for exploration, extraction and processing, col-
targets together with stakeholders and conduct impact lection, and recycling.
assessments of the corresponding measures. Indicators >> Substitutes for at least three key applications of criti-
will be developed in order to measure progress in cal and scarce raw materials.
improving resource efficiency. The European Commission >> A network of research, education, and training cen-
aims to use a lead indicator, complementary macro indi- tres on sustainable mining and materials manage-
cators, and theme-specific indicators. To emphasize the ment (M³), whilst ensuring appropriate coordination
material resource aspects of resource efficiency, resource with the possible European Institute of Innovation
productivity calculated as GDP divided by domestic mate- and Technology (EIT) - Knowledge and Innovation
rial consumption (expressed in euro/tonne) was proposed Community (KIC) on sustainable exploration, extrac-
as the provisional lead indicator: This is the inverse of the tion, processing, and recycling.
term “intensity of use” (IU), which is widely used to refer >> Enhanced efficiency in material use and in prevention,
to the quantity of material used to produce goods and reuse, and recycling of valuable raw materials from
services. The complementary indicators will include waste streams, with a specific focus on materials hav-
water, land, materials, and carbon [7]. ing a potentially negative impact on the environment.
>> Identified opportunities and develop new ideas for
The way forward outlined in the roadmap is to provide innovative raw materials and products with market
incentives for companies and consumers to change con- potential.
sumption patterns and promote resource-friendly pro- >> A proactive strategy of the EU in multilateral organi-
duction and products. Life cycle approaches, increased zations and in bilateral relations, such as the US,
recycling, the phasing out of environmentally harmful Japan, Australia, in the different areas covered by the
subsidies, and shifting taxation away from labour to EIP.
boost employment and economic growth are also pivotal
in the roadmap [7]. These targets should also provide adequate follow-up
and will enable functioning of the EIP to be monitored,
In a draft report from March 2012, the European Parlia- including the results achieved and the work to be per-
ment’s Environment, Public Health, and Food Safety formed in the future.
Committee supported the Commission’s approaches to
decouple economic growth from resource use, however RHI’s Contribution to Increased Resource
stated that “it does not reflect the necessary sense of Efficiency
urgency” [8] and that it would like to see even clearer
Sustainable Magnesite Mining
and more concrete steps and targets regarding several
issues. Resource efficiency has to take place at all stages and in
all processes of the entire material flow, starting with the
The European Innovation Partnerships (EIPs) were intro- sustainable mining of raw materials. This means exploit-
duced in 2010 [9] and are launched when the combined ing the deposits and the existing infrastructure in an
> 51
optimal way, avoiding selective mining of high grade Sustainability and Resource Efficiency in Magnesia
material, and using as much of the material possible. Sintering
According to the principles of sustainable mining, RHI
After raw material extraction, RHI analyses material
uses sustainable and innovative methods to ensure long
flows from the mine to the final raw material product,
term raw material self-sufficiency whilst at the same time
minimizing environmental impact. This means precise which is the input material for refractory production,
deposit reconnaissance and adapting sourcing methods with the aim of optimizing material consumption and
as well as surveying the changing excavation areas in minimizing environmental impact.
the underground mines and if necessary adaptation and
new product development. A long term mine map con- Figure 3 illustrates the complexity of raw material refin-
tributes to sustainable planning of raw material sourcing. ing and details the multiple technical processes
Open cut Open cut

Dredging Open cut Open cut and underground Open cut Open cut
and under- and under-
ground ground
Mineral Uranium and Cu, Zn, Pb Cu/Au Au Ni Al Coal Fe Mg, Cr, Mn

sands Cu/U ores ores ores ores ores ores ores ores
Crush
Mill Mill Mill
Solids
from Heavies
Cu/U ore from
U leach Gravity Cu/Ag
Digestion Wash
Cu ores separation ores
Ni
Gravity Solvent Base metals Cyanide Precipita- Coal
concentra- Calcine
separation extraction concentration leach tion wash
tion
Cu concen-
trates
Product Calcine Au Calcination Dead
upgrade recovery burn
Ore Ore Ore

product product product
Refining SX/EW Smelter Smelter Smelter Smelter Electro-

Refining
refining refining refining refining refining fuse
Concentrate Concentrate Concentrate Concentrate Concentrate

products products products products products
Refined product
Figure 3. Flow diagram of the various processes required from raw material extraction to the refined raw material product [11].
52 <
required to generate a variety of different raw materials. Increasing Resource Efficiency at RHI’s
The wide variance in the processes makes it necessary Customers
to analyse each individual step separately. Therefore,
general approaches to process optimization are not RHI constantly aims for optimal product performance
successful, but rather multiple life cycle tools are neces- and reducing the energy consumption of customer pro-
cesses. The development of new materials during the
sary.
last decades has led to a considerable decrease in the
specific consumption of refractory products. Today, for
Innovative Filter Systems
example, steel is produced using significantly lower
With the aim of optimizing energy application, all areas amounts of refractory material compared to 1950
ranging from product development, production pro- (Figure 5).
cesses, and supply chain, to the use of refractory mate-
rials at RHI’s customers are constantly subject to pro- Use of Secondary Materials
cess improvements. In addition, the application of opti-
mal kilns, furnaces, and energy sources as well as opti- The use of secondary materials is an essential aspect of
mizing energy costs by achieving the lowest specific RHI’s raw material strategy as recycling refractory mate-
energy consumption per energy source is integral in rials compensates for rising energy prices, preserves
RHI’s strategy to increase energy efficiency. resources, and significantly helps to reduce the CO2
footprint and energy use, since the high temperature
If the example of firing magnesite to produce sinter raw material processing is eliminated.
magnesia (or dead burned magnesia) is examined,
RHI’s most commonly used input material for refractory
production, it is immediately evident that a burning pro- 120
cess combined with a sintering process, running at tem-
peratures in the range of 1600 °C up to 2000 °C, is n Government limit
100
energy intensive. In terms of costs and emissions, it is n Actual
in the interest of the process owner to have the lowest
Relative emissions [%]
energy consumption possible and hence increase 80

energy efficiency. An example of where RHI has simul-
taneously increased resource efficiency and reduced
60
emissions is the recently installed innovative filter sys-
tem at its plant in Breitenau (Austria).
40
Energy is the most important cost factor for magnesia
sintering. Therefore, one focus of a magnesia sinter
20
plant is reduction of the specific energy consumption
for the burning process. However, if this approach has
been used and all the physical and thermic potentials 0
are exhausted, it is easy to see that the possibility for Electrostatic precipitator Bag filter
further improvements in energy efficiency in such high

temperatures processes is limited. Nevertheless, in
Figure 4. Comparison of the old electrostatic precipitator and the
some cases further improvements are possible. In the new bag filter dedusting systems at RHI’s magnesia sintering
Breitenau plant, a preheater for raw magnesite was plant in Breitenau (Austria). The relative emissions achieved are
installed, which leads to a natural gas reduction of compared to the legislative limits.
350000 m3; corresponding to the annual natural gas
consumption of 180 households or 700 tonnes of CO2.
Through this process, energy efficiency is increased by 70
2%, which is an important step considering specific n Steel
energy consumption had previously been optimized. 60 n Glass
n Cement
Typically, industrial processes emit dust. This dust
kg refractories/tonne product
50
emission can be reduced by technical equipment. In the
past (i.e., 1970s) only cyclones and electrostatic precipi-
40
tators were available. With such equipment, the reduc-
tion of dust emissions was in the range of up to 90%.
30
However, new technology developments in the dust
separation process have resulted in bag filter systems
that can operate at higher temperatures. Such bag fil- 20
ters are manufactured from Teflon and can function at
temperatures up to 250 °C. With these filter systems, 10
the dust emissions can be reduced by 99%. The newly
installed bag dust filters in the Breitenau plant achieve 0
1950 1980 2000 2008
these emission reductions (Figure 4) and the dust,
which results from the burning of magnesite to become Year
sinter magnesia, is recycled and refed into the produc-
tion process, which helps decrease raw material use. Figure 5. Specific consumption of refractories from 1950–2008 [12].
> 53
Conclusion
To be globally competitive in the coming years, generate
employment opportunities, and boost economic growth,
Europe must secure raw material supply. Therefore, in
the last years the EU has focused on multiple policy initi-
atives to achieve this goal, including improving resource
efficiency. Through diverse and comprehensive strate-
gies that include sustainable mining, optimizing raw
material use, recycling, as well as increasing energy effi-
ciency in its own as well as customer plants, RHI is
actively engaged in improving resource efficiency and
the associated environmental and cost benefits it pro-
vides.
References
[1] Drnek, T. Lecture Mineral Economics: 2. Introduction, University of Leoben, Austria.
[2] SERI Global Material Flow Database. 2011 Version. www.materialflows.net
[3] Drnek, T. Lecture Mineral Economics: 2. Introduction, University of Leoben, Austria. (Aggregated data from World Bank and U.S Geological
Survey (USGS)).
[4] “The Raw Materials Initiative – Meeting Our Critical Needs for Growth and Jobs in Europe”. Communication From the Commission to the
European Parliament and the Council. COM(2008) 699. SEC (2008) 2741, Ed.; EU-Commission, Brussels, 2008.
[5 “Tackling the Challenges in Commodity Markets and on Raw Materials”. Communication From the Commission to the European Parliament, the
Council, the European Economic and Social Committee and the Committee of the Regions. COM(2011) 25 final, Ed.; EU-Commission, Brussels,
2011.
[6] “A Resource-Efficient Europe - Flagship Initiative Under the Europe 2020 Strategy”. Communication From the Commission to the European
Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions. COM(2011) 21, Ed.; EU-Commission,
Brussels, 2011.
[7] “Roadmap to a Resource Efficient Europe”. Communication From the Commission to the European Parliament, the Council, the European
Economic and Social Committee and the Committee of the Regions. COM(2011) 571 Final, Ed.; EU-Commission, Brussels, 2011.
[8] “Draft Report on a Resource-Efficient Europe”. Committee on the Environment, Public Health and Food Safety. (2011/2068(INI)), European
Parliament, Brussels, 2012.
[9] “Europe 2020 Flagship Initiative. Innovation Union”. Communication From the Commission to the European Parliament, the Council, the
European Economic and Social Committee and the Committee of the Regions. COM(2020) 546 Final, Ed.; EU-Commission, Brussels, 2010.
[10] “Making Raw Materials Available for Europe‘s Future Well-Being. Proposal for a European Innovation Partnership on Raw Materials.”
Communication From the Commission to the European Parliament, the Council, the European Economic and Social Committee and the
Committee of the Regions. COM(2012) 82 final, Ed.; EU-Commission, Brussels, 2012.
[11] Petrie, J. Life Cycle Approaches: Translating Life Cycle Thinking into Strategy and Action Plans for the Mining, Minerals and Metals Sector.
Presented at Euromines Internal Workshop on the Use of Life Cycle Assessments, Brussels, November 2011.
[12] Feytis, A. Between the Linings. Industrial Minerals. June 2010, 46–51.
Article reprinted by courtesy of Springer-Verlag Vienna.
Authors
Thomas Drnek, RHI AG, Raw Materials Division, Breitenau, Austria.
Michaela Seelig, RHI AG, Corporate Communications and Public Affairs, Vienna, Austria.
Corresponding author: Thomas Drnek, [email protected]
54 <
RHI Bulletin > 1 > 2012, pp. 55–57
Reinhard Ehrengruber, Walter Schaer and Arnold Haeni
Full Integration of INTERSTOP Flow Control

Technology into RHI
Introduction control technologies, including the four generations of ladle
gates (Figure 2), were in-house developments. The CS and
On January 18, 2012, RHI acquired 100% of Stopinc AG.
LC series, representing the fourth generation, is in opera-
Previously RHI had held a 50% stake in the Swiss company
tion at more than 170 steel plants worldwide. At present
with the other 50% owned by Steinegg AG (Dr. Tanner).
work is underway to design the fifth generation of INTER-
Stopinc, with its brand INTERSTOP, is among the market
STOP ladle gates.
leaders in flow control technology, providing innovative
systems from the converter through to the mould. With
this takeover, RHI will be in a position to meet future
Highlights of the Integration
requirements in the global flow control business. Stopinc, For many years, RHI and Stopinc have been approaching
based in Hünenberg (Switzerland), will be maintained as and developing the flow control market in very close coop-
an independent company and established as the new Flow eration and the major market activities will continue to be
Control Competence Center, providing increasing support performed by the same local teams and representatives.
to worldwide sales and service centres. Synergies in the areas of slide gate systems, controls and
automation, in combination with ceramic functional prod-
History Milestones ucts enable integrated products and services for flow con-
trol applications from converter to mould.
In 1966, Interstop AG was founded in Zürich by Didier
Werke AG (Germany) and Dr. Tanner (former CEO of Con-
The main focus of the new Flow Control Technology Com-
cast AG, Switzerland). Both parties each held 50% of the petence Center will include:
shares. The main task of the new company was to develop
and market a slide gate nozzle system to precisely control >> Full integration of the marketing and research and
the flow of molten metal from the ladle to the tundish. The development activities.
concept was based on a patent granted to David D. Lewis >> Stopinc AG, Hünenberg, will continue to cover world-
in 1885 (Figure 1) that had been revived and refined to fit wide demand concerning new projects, spare parts,
the current needs of the steel industry. The first slide gate engineering, and customer service with its offices in the
nozzle system on the market was the INTERSTOP BK tech- USA and China.
nology. Besides the mechanical slide gate system engi- >> RHI and INTERSTOP will continue to maintain the high-
neering by Interstop, Didier Werke developed and manu- est quality services and technologies.
factured the corresponding slide gate ceramics. >> With the full integration, research and development
activities will be further expanded to advance innova-
To avoid confusion between Interstop AG and Intershop tions by pooling resources.
AG, in 1971 a court decided that Interstop AG had to >> The centralized flow control know-how will enable opti-
change its company name to Stopinc Aktiengesellschaft; mal coordinated exchange between the specialists
however, INTERSTOP could be used as a brand name. regarding mechanics, electrics, hydraulics, automation,
refractories, and service, in order to provide exceptional
Due to tremendous progress regarding safety, economy, technical and economical solutions.
and handling of the system, the 2000th INTERSTOP slide >> Established supply and information chains with custom-
gate valve, mounted on a steel casting ladle, had already ers will be in no way affected.
been commissioned at Ovako Oy (Koverhar, Finland) by
1978.
Thanks to a worldwide network of sales representatives,

set up by both Didier Werke and Stopinc, the INTERSTOP
system became well established in the market. Over the
years Stopinc has grown from a slide gate supplier to the
steel industry’s flow control partner for ladle, tundish, and
converter applications with the related hydraulic and elec-
trical equipment. Next to product innovations, the highest
priority was the early expansion abroad to establish the
brand and technology in the global market. Therefore,
already in 1977 “Interstop do Brasil” was founded, fol-
lowed by a sales office in Singapore in 1980, Interstop
Corp. USA in Mokena in 1986, and last but not least Inter-
stop China in Shanghai in 2005.
Stopinc has always been dedicated to research and devel-

opment, and it is especially noteworthy that all flow Figure 1. Description of the first slide gate patent.
> 55
With the integration, the infrastructure is in place to remain Latest Trends in Flow Control Technology
in a leading position and actively participate in the develop-
The basic ideas behind the latest flow control developments,
ment of future trends in flow control.
to fulfil the increased demands of today’s steel industry, are
operator safety and process reliability. Therefore design solu-
tions have been pushed for a fully automated ladle prepara-
tion area. For example, tasks like oxygen lancing and chang-
ing ladle gate refractory parts will be automated once the
system is fully operational. Stopinc is now in the process of
a hot trial so its vision will successfully pass the milestone
for industrial scale-up (Figure 3).
The following advantages are offered by an automation system:
1966 First trial >> Cost saving due to higher process reliability. For example
the automation system ensures that the ladle gate is in
the closed position when the ladle leaves the preparation
area.
>> Cost reduction due to lower operating requirements.
A reasonable return on investment (ROI) is achieved by a
reduction of labour costs.
>> Increased operational safety is achieved due to constant
high-quality ladle preparation.
1970 BK Technology >> Potentially dangerous tasks for operators such as O2 lanc-
ing or handling heavy loads are performed by the auto-
mated system.
>> The tasks performed are recorded and can be statistically
evaluated for quality management issues.
1981 QC Technology
Figure 3. Layout of an automated ladle preparation system.
1989 LS Technology
2001 CS/LC Technology
Figure 2. INTERSTOP ladle gate generations. Figure 4. INTERSTOP Metering Nozzle Changer MNC-RSP.
56 <
The same aspects of operator safety and process reliability Summary

were the trigger for the newest ready to market INTERSTOP
Together with refractories from RHI, the INTERSTOP flow con-
products developed through the close cooperation between
trol technology is operational in more than 74 countries. A sub-
RHI’s and Stopinc’s Research and Development and Market- stantial amount of global steel production is running through
ing departments: one of the 7000 units in operation worldwide. With the full inte-
gration, technological focus, customer orientation, and local
>> New Metering Nozzle Changer MNC-RSP (Figure 4). presence of RHI and INTERSTOP, flow control technology will
>> New Mono Tube Changer MTC-ESP (Figure 5). be maintained on a high global level and further innovations
>> Ladle Shroud Crown Connection (Figure 6). will be supported and realized together with customers.
Figure 5. INTERSTOP Mono Tube Changer MTC-ESP. Figure 6. INTERSTOP Ladle Gate Type CS with the newly devel-
oped Ladle Shroud Crown Connection.
Authors
Reinhard Ehrengruber, Stopinc AG, Hünenberg, Switzerland.
Walter Schaer, Stopinc AG, Hünenberg, Switzerland.
Arnold Haeni, Stopinc AG, Hünenberg, Switzerland.
Corresponding author: Reinhard Ehrengruber, [email protected]
> 57
RHI Bulletin > 1 > 2012, pp. 58–62
Jürgen Goriupp, Andreas Rief and Johannes Schenk
Dynamic Refractory Wear Test Method for

Magnesia-Carbon Products
Introduction sometimes gives results that are not comparable to the in-
service behaviour. Nevertheless, for a general service suit
With the increasing demands on refractory quality in the iron
ability assessment this test can be used for MgO-C refracto-
and steel industry and a changing raw material situation,
ries; however, determination of small differences in MgO-C
continuous MgO-C product development is more and more
product performance, which is especially important for the
necessary. Prior to possible installation in a steelworks, the
development of new grades, is not possible.
products’ properties must be examined very carefully. Over
the years, many testing methods have been developed to
The use of a crucible or cup test can also raise problems
simulate different cases of refractory wear according to the
due to the aforementioned slag saturation. Especially the
product application field. Since all the tests have advantages
uncontrolled reaction of iron oxide rich slags (e.g., BOF
as well as disadvantages, the right choice of testing method
slag) with the carbon in the MgO-C sample (equation 1)
is crucial to obtain representative experimental results rele-
vant to the in-service conditions. According to Lee and negatively influences this test and the slag to brick ratio has
Zhang [1], the most commonly used refractory wear testing to be optimized.
methods can be described as follows:
FeOl + Cs → Fel + COg
(1)
Static testing methods (no motion of slag relative to refractory)
The above reaction causes a change in slag composition and
>> Sessile drop test. the formation of CO gas bubbles, which leads to uncontrolla-
>> Static finger test. ble slag buildup inside the crucible. Furthermore, the slag/
>> Cup or crucible test. brick ratio is more problematic than in the induction furnace.
>> Induction furnace test.
Dynamic testing methods (motion of slag relative to refractory)

Cover Gas offtake
>> Rotating finger test.
>> Rotary slag test.
Nearly all the static tests have the disadvantage that the slag Slag
quickly becomes saturated with corrosion products because
of the unfavourable ratio of slag to refractory sample. With Induction coil
static conditions it is also impossible to destroy any bound- Steel melt
ary layer between the slag and refractory sample. These neg-
ative issues can be eliminated by using a dynamic testing
method, but always with the disadvantage of increased
experimental effort [1]. Refractory samples
Standard Wear Test Methods for MgO-C Refrac- Figure 1. Experimental setup of an induction furnace test.
tories
One of the most commonly used refractory wear test meth-
ods for MgO-C products is still the induction furnace test
(Figures 1 and 2). In this test eight segments of various
refractory materials are arranged to form an octagonal cruci-
ble that is filled with molten steel and slag of a defined com-
position. However, for carbon-containing materials this test
has limitations and the results have to be interpreted care-
fully. Occasionally, contradictory results compared to field tri-
als as well as inconsistencies between repeated induction
furnace tests have been observed. Some of these misleading
results can be explained by the general test setup. High oxy-
gen levels, unintended modification of the steel bath or slag
chemistry, or a low slag/brick ratio are often problems experi-
enced with the induction furnace due to the practical design.
However, experiments with argon flushing of the furnace or
charging more slag have not solved these issues. Due to the
Figure 2. Tapping of an induction furnace test at the Technology
high oxidation rates during the induction furnace test and Center Leoben.
the significant changes in slag chemistry, the test method
58 <
To overcome these limitations, the decision was made to iden- experimental procedure and because it has to be removed by
tify an alternative wear test method, particularly focused on hand this increases the experimental effort. Furthermore, it is
examining MgO-C refractories. Therefore, a supplementary very difficult to remove the slag in its entirety out of the furnace,
dynamic testing method was evaluated and verified by experi- whereby an uncontrolled slag formation becomes possible.
mental research at the Chair of Metallurgy (University of Leo- Another very important shortcoming regarding slag formation
ben, Austria) by comparing the results of induction furnace cam- concerns the oxidation of certain steel bath components. This
paigns with this laboratory scale self-assembled test method. oxidation is caused by the intensive flow conditions and contact
between the steel bath and surrounding atmosphere during the
Testing Method Requirements for MgO-C Refrac- initial melting, as schematically represented in Figure 3.
tories
The influence of this possible oxidation can be seen in Figures
Following the pilot academic project, the main target has
4 and 5, detailing the SiO2 content and composition of multiple
been to establish a laboratory scale dynamic test method for
MgO-C products at the Technology Center Leoben (Austria), slag samples. In this case, a steel bath with a silicon content of
which eliminates the following main disadvantages of the ~0.2 wt.% was used during the test. The slag samples, taken
induction furnace test: when the slag was periodically replaced, showed the SiO2 con-
tent of the slag was significantly influenced and the Si was com-
>> Rapid (over-)saturation of the slag phase with MgO. pletely removed from the steel bath over the course of the trials
>> Possible interactions between the refractory sample due to the high O2 levels in the furnace atmosphere. It must be
carbon and the inductive field of the furnace. taken into account that such a phenomenon leads to an uncon-
>> Uncontrollable flow conditions. trolled slag formation, which makes it difficult to interpret the
>> Solid slag cover formation. results in terms of the in-service behaviour.
>> Challenging temperature measurements.
>> Uncontrollable furnace atmosphere. The formation of a solid slag cover over the molten slag surface
>> Challenging control of slag formation. is also detrimental because the conditions in the slag zone and
>> High experimental effort. during the experimental procedure become difficult to control.
A typical solid slag cover, which is caused by cooling of the slag
With the induction furnace, rapid saturation of the slag phase surface combined with too little bath movement to destroy it, is
makes it necessary to change the slag several times during the shown in Figure 6.
Atmosphere Slag 60
70 a- Gehlenite
Steel 2C
aO .
SiO2
bath 80 SiO 1
2
2
3
4 5
Induction
90 Lime 6
7
furnace 8 CaO
3CaO
CaO
0 10 20 30 40 50
Al2O3
n Analysed slag composition n Desired slag composition
Figure 3. Flow conditions in an induction furnace [2]. Figure 5. Slag compositions in the ternary system CaO-SiO2-
Al2O3, over the course of an induction furnace test. The slag sam-
ples 1–8 were taken during the slag changes detailed in Figure 4.
30
1 (sample number)
25
2
SiO2 in slag [wt. %]
20
3 SiO2
4
15 5
6 7 8
10
0
0 1 2 3 4 5 6 7 8
No. slag changes
Figure 4. SiO2 content of the slag over the course of an induction Figure 6. Solid slag cover on the surface of an induction furnace
furnace test. The test samples where taken when the slag was test.
replaced.
> 59
Dynamic Refractory Wear Test Setup The experimental setup is very similar to a standard disso-
lution test for CaO in different slags or other refractory tests
The rotating finger test setup was used as the basis for the
that are described in the literature [3–6]. The modified high-
dynamic laboratory scale wear test method. This setup is
temperature tube resistance furnace (Tamman type) enables
very suitable to counteract an unfavorable ratio of reactive
working temperatures of up to 1700 °C and the use of an
refractory sample surface to the slag bath volume, which is
inert gas atmosphere, which is required to protect the
necessary to prevent rapid saturation of the slag phase with
MgO-C samples from uncontrolled decarburization. How-
corrosion products. The rotating finger test (Figure 7) was
ever, because of the small furnace size, only one refractory
originally developed by refractory manufacturers for the
sample can be tested at a time. The rectangular shaped
glass industry, since it effectively simulates the convective
MgO-C sample is affixed to a steel rod and dipped into the
flow patterns present in a glass tank [1].
liquid slag bath once it has reached the required tempera-
ture. The sample is rotated on its axis using a laboratory
Commonly used finger tests operate at lower temperatures
stirring device. The most challenging issue is the appropri-
than those required to simulate the service conditions of
ate choice of crucible material. Since common ceramics are
MgO-C products in the steel industry, therefore an adaption
not resistant to slag attack, the use of these materials would
of the setup for higher temperatures was necessary. This
change the composition of the experimental slag or result
was realized by modifying a high-temperature resistance
in the crucible breaking. Additionally, many materials are
furnace as illustrated in Figures 8 and 9.
not resistant to molten metal and a crucible made of graph-
ite would react with an iron oxide bearing slag.
To solve this problem, a boron nitride crucible is currently

used that is resistant against all the aforementioned possible
types of attack in an inert gas atmosphere. Whilst dissolu-
tion of BN into the slag has been observed, it is considered
Rotating
samples negligible as during the test BN shouldn’t influence the cor-
rosion behaviour of the sample or the aggressiveness of the
Furnace Slag
slag as any oxidation of the BN to B2O3 can be excluded.
Crucible Steel melt However, for a realistic MgO-C test setup, a certain O2 level
will need to be incorporated to simulate the oxidative corro-
sion of the carbon. Therefore, completely new crucible
materials will be neccessary that prohibit oxidation of the
crucible and reaction with the slag at realistic field tempera-
Figure 7. Setup of a rotating finger test [1]. tures of above 1650 °C.
Slag sampling rod

Stirring device Refractory
sample
Rotating MgO-C sample High-temperature
Offgas analysis Steel rod resistance furnace
M
Slag sampling
rod
Ceramic
protection
Slag bath
Boron nitride
crucible
Slag bath
Crucible
Carbon
stamp
Thermocouple
Thermocouple
Figure 8. Modified high-temperature tube resistance furnace. Figure 9. Layout of the dynamic testing method .
60 <
Quantification of the Dynamic Test Method

MgO FeO Al2O3 CaO MnO SiO2
To quantify the dynamic test method, several experiments
with different MgO-C products and various slag composi-
CaO-Al2O3 slag 2.0 20.0 35.0 26.0 10.0 7.0
tions were carried out. The slag compositions (Table I) were
selected to provide a more aggessive environment than is CaO-rich slag 2.0 5.0 35.5 49.0 2.5 6.0
typically found during in service applications in the steel
SiO2-rich slag 2.0 10.0 10.0 33.0 5.0 40.0
industry. Furthermore, the low MgO values were chosen to
enhance the wear process and decrease the test duration.
Table I. Compositions of test slags.
Three different types of commonly used MgO-C products
were evaluated during the testing campaign described (i.e.,
pitch-bound, resin-bound, and antioxidant bearing). Approx-
imately 150 g of premelted slag powder was used in a
boron nitride crucible for each test (Figure 10). The rotating
speed of the refractory samples was 50 rpm with an overall
testing time of 2 hours at 1600 °C. Rotation of the MgO-
samples protected the slag bath from an uncontrolled
development of solid slag cover by homogenizing the tem-
perature and composition.
To analyse the refractory wear, a slag sample was taken out

of the liquid slag bath every 20 minutes using a steel rod.
The total MgO content dissolved in the slag samples was
Figure 11. Typical MgO-C sample at the end of a dynamic wear test.
measured by X-ray fluorescence (XRF) analysis and at the
end of the test the sample surfaces (Figure 11) were exam-
ined with a scanning electron microscope combined with an
SiO2
energy-dispersive X-ray analyser. The results of the XRF
slag analysis prior to the start of the test can be seen in Fig-
10 90
ure 12. In all cases, the slag was very homogeneous with Two liquids
little to no deviation from the desired composition. 20 80
Cristobalite
30 70
The MgO levels in the analysed slags samples are depicted Tridymite
40 60
in Figure 13. The results clearly demonstrate an increase in Pseudowollastonite
Mullite
the dissolved level of MgO in the slag samples as the test 50 50
Rankinite SiO2-rich slag
progressed. Furthermore, it is possible to distinguish not 60 40
only between the different slag types, but also between the a-
different MgO-C products (i.e., resin-bound (A), antioxidant 70 2C 30
aO Gehlenite
bearing (B) and pitch-bound (C)), which should facilitate S. iO Corundum
20
80
2 CaO-Al2O3 slag
product development. Lime
90 10
CaO-rich slag
An interesting result relates to the better performance of CaO Al2O3
0 10 20 30 40 50 60 70 80 90 100
the test products in the CaO-Al2O3 slag compared to their
behaviour in the CaO-rich slag. This can be explained by the
solubility of MgO in the tested slags, which were widely Figure 12. Analysed slag compositions in the CaO-SiO2-Al2O3
studied experimentally by Park and Lee [7]. system, prior to the test start.
10
n SiO2-rich slag B A
9
n CaO-rich slag C
Disolved MgO in slag [wt.%]
8 n CaO-Al2O3 slag
7
6
A C
5 B
A
4
C
3
B
2
1
0 20 40 60 80 100 120 140
Process time [minutes]
Figure 10. Furnace chamber at 1600 °C during a dynamic wear test. Figure 13. MgO content in the analysed slag samples. The
MgO-C refractories are resin-bound (A), antioxidant bearing (B)
and pitch-bound (C).
> 61
Park and Lee found that the MgO saturation is directly

dependent on the basicity (B3) of the slag used (Figure 14)
[7]. It was demonstrated that the solubility of MgO is higher
Slag
in CaO-rich slag, which directly leads to a greater wear of
the tested MgO-C samples.
Sample matrix
The decrease in the MgO solubility level for basicities lower
than 0.8 was caused by precipitation of magnesium alumi- Spinel
nate spinel [7]. This was evident experimentally in the cases
using CaO-Al2O3 slag. Figure 15 shows a scanning electron
micrograph of a MgO-C sample that was tested with this
slag type. It can be clearly seen that a spinel layer formed
on the sample surface and this layer protected the sample
against further corrosion. Figure 15. Scanning electron micrograph of a MgO-C sample
tested with the CaO-Al2O3 slag.
Conclusions
A dynamic refractory wear test, tailored to examining >> Possible comparison between different MgO-C products.
MgO-C refractories, is being established at the Technology >> Reduced experimental effort.
Center Leoben. The experimental setup will provide multi- >> No slag build-up.
ple advantages including: >> Controlled furnace atmosphere.
>> Controlled slag formation.
>> Simplified temperature measurement.
16
The initial experiments have indicated substantial improve-
14 ments over the induction furnace test, and the results pro-
12 vide highly quantitative evaluation of refractory dissolution.
One critical point is the current use of an inert gas atmos-
10
phere because it prohibits oxidation of the binder carbon,
MgO [wt.%]
CaO-rich slag
8 which is not representative of the in service conditions.
CaO-Al2O3 slag However, this adapted dynamic testing method is very suit-
6
able for studying the influence of different slag composi-
4 tions on the wear of the MgO in the matrix.
2
With the gas flow control in the new furnace and its well
0
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
defined and adjustable O2 partial pressure, studies of MgO
solvation to the slag depending on MgO quality and exclud-
B3=C/(A+S) ing the influence of decarburization can be performed. Fur-
thermore, the furnace construction enables all kinds of
Figure 14. MgO saturation in CaO-Al2O3-SiO2 slags as described dynamic material tests with molten metals or slags below
by Park and Lee [7]. Abbreviations include basicity (B3), CaO (C), 1700 °C under different atmospheric conditions (e.g., for
Al2O3 (A), and SiO2 (S). nonferrous, glass, and cement applications).
References
[1] Lee, W.E. and Zhang, S. Melt Corrosion of Oxide and Oxide-Carbon Refractories. International Materials Reviews. 1999, 44, No. 3, 77–103.
[2] Heinen, K.H. Elektrostahlerzeugung (4th Edition); Verlag Stahleisen: Düsseldorf, 1997.
[3] Bleck, W. and Senk, D. Annual Report aus dem Institut für Eisenhüttenkunde. RWTH Aachen, Band 46, 2001, 52.
[4] Jansson, S. A Study on Molten Steel/Slag/Refractory Reactions during Ladle Steel Refining. Licentiate Thesis, Royal Institute of Technology,
Stockholm, 2005.
[5] Chung, Y. and Schlesinger, M.E. Interaction of CaO-FeO-SiO2 Slags with Partially Stabilized Zirconia. J. Am. Ceram. Soc. 1994, 77, No. 3, 611–616.
[6] Cooper, A. and Nicholson, P. Influence of Glass Redox Conditions on the Corrosion of Fusion-Cast Chrome-Alumina Refractories. Ceramic Bulletin.
1980, 59, No. 7, 715–717.
[7] Park, J.M. and Lee, K.K. Reaction Equilibrium Between Liquid Iron and CaO-Al2O3-SiO2-FeO-MnO-P2O5- Slags. Proceedings 79th Steelmaking
Conference, Iron and Steel Society, Pittsburgh, USA, March 24–27, 1996, pp. 165–172.
Authors
Jürgen Goriupp, Chair of Metallurgy, University of Leoben, Austria.
Andreas Rief, RHI AG, Technology Center, Leoben, Austria.
Johannes Schenk, Chair of Metallurgy, University of Leoben, Austria.
Corresponding author: Andreas Rief, [email protected]
62 <
RHI Bulletin
To facilitate efficient distribution of the RHI Bulletin, amendments or e-mail the required alterations. In
we would be grateful if all recipients could check addition, if you would like to register and receive the
their delivery details are correct and select the pre- RHI Bulletin on a biannual basis, please complete
ferred format. Please use the form below to fax any the form below and return it to us by fax or e-mail.
New registrant Yes No
Preferred format Print E-mail Print & e-mail
Title
First name
Surname
Company
Department
Street / no.
Postbox
City (town)
Area code
Country
Phone
Fax
E-mail
Phone: +43 (0) 502 13-5300, Fax: +43 (0) 502 13-5237, E-mail: [email protected]

Boletin 2012 Datos Industriales

Uploaded by

Copyright:

Available Formats

Boletin 2012 Datos Industriales

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Boletin 2012 Datos Industriales

Uploaded by

Copyright:

Available Formats

Steel Edition

RHI Bulletin >1> 2012

DELTEK Eco Gaskets and

RHI Bulletin 1/2012

Published by: RHI AG, Vienna, Austria

Contact: Ulla Kuttner

In closing, I would like to thank all the authors involved in this

Comparison of Basic Oxygen Furnace Bottom

n No. of steel plants using

Cumulative share [%]

Ar/N2 O2 /CO2 O2 /Cn Hm O2 /Cn Hm O2 /Cn Hm

Top only Soft Strong Combined Bottom only

Parameter With bottom Without bottom

400 [C] x [O]

Commonly, there are three different bottom gas purging

10 Purging Plug Features and Installation Require-

(a) (b) (c)

Parameter MHP SHP Tuyere

Table III. Characteristics of MHPs, SHPs, and tuyeres.

During tapping and sampling (manually with a lance), direct

Figure 9. (a) MHPs and (b) SHPs installed in an elliptical arrangement.

Inert gas distribution

Shutdown of the bottom gas

[C] x [O] level x 10-4

Shutdown of the bottom gas purging system

Jürgen Schütz, Alexander Maranitsch and Milos Blajs

New Oxycarbide Refractory Products

0.2 Load: 0.1 N/mm2

The CAS-OB Process

Figure 6. Image of the CAS-OB process. Figure 8. CAS-OB bell in production.

An additional significant advantage of the bells installed

Currently, hot metal application field trials including blast

In several steel plants the oxycarbide castables have been

Customer-Specific Analysis of Steelmaking Slags

Introduction Variance in the slag composition is common due to an input

The MgO mass balance during the ladle transport and

xlime MgO · mlime + xdolo MgO · mdolo + xCa aluminate MgO ·

Where x is the concentration of MgO in the slag former

Slag formers CaO MgO SiO2 Fe2O3 Al2O3 Mass*

– xrefr MgO · mrefr loss max = [xlime MgO · mlime + B2 CaO/SiO2

Or if slag analysis data is available: 2

ΔMgO* = xinitial analysed slag MgO – xsaturated slag MgO(4) 0

Slag Analysis and MgO Optimization in a Steel

ysis, the presaturation levels were increased by appropriate

MgO SiO2 MgO SiO2

MgO FeO MgO FeO

80-tonne ladle n MgO saturation, 5 wt.% MgO, 1600 °C

CaO FeO CaO Al2O3

agreement between some heats and the calculated satura- Summary

Table III. Materials provided by RHI for MgO slag conditioning.

Gas Purging Lances: Improving Established

Introduction lance under the very harsh service conditions. In particular,

Brand 1 Brand 2 Brand 3 Brand 4 Brand 5 Brand 6

Steel fibres (wt.%) ~5 ~3 ~2.5 ~5.5 ~6 ~5

Head with nozzles

1250 1275 1300 1325

1450 1475 1500 1525 1550

1450 1475 1500 1525 1550

(a) (b) (c)

n No. of steel plants using

ΔMgO* = xinitial analysed slag MgO – xsaturated slag MgO(4) 0