Anais do 44º Congresso Brasileiro de Cerâmica
23801
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
CHARACTERIZATION AND TESTING OF REFRACTORIES FOR GLASS TANK
MELTERS
M. Velez, M. Karakus, M. R. Reidmeyer, W. D. Headrick, R. E. Moore, University
of Missouri-Rolla, Ceramic Engineering Dept., Rolla, MO 65409-0330, USA;
mvelez@umr.edu
ABSTRACT
Current and future research goals in our laboratories include the study of
the attack of glass-contact refractories and of crown refractories in glass tank
melters, under either air-gas or oxyfuel conditions. There is an emphasis on
evaluation and characterization of critical parameters of commercial refractories
such as microstructure, porosity and mechanical properties. A second focus is
the evaluation of the performance of refractory alternatives to traditional crown
and superstructure refractories including their physical aspects, crown design,
and joint quality.
Key Words: Cathodoluminescence, Corrosion, Glass, Refractories
INTRODUCTION
Several approaches have been taken to minimize corrosion in glass tank
melters: improved construction techniques, new crown designs, and use of new
materials (for instance alumina-based compositions in oxyfuel-fired furnaces
where silica brick has failed, as well as AZS, and spinel based compositions).
Other approaches are based on understanding the corrosion mechanisms and
the modeling of the combustion space, linked to the melting of the glass and
operation of the furnaces. The degradation of refractories used in glass tank
Anais do 44º Congresso Brasileiro de Cerâmica
23802
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
melters can normally be assessed after a campaign when the furnace is partially
or totally disassembled
(1).
Corrosion tests to predict degradation usually employ
small specimens exposed to accelerated working conditions that might not be
simulative of actual industrial conditions.
The current study covers three
sections: post mortem refractory characterization, initial testing using current
techniques, and representative test development.
Crown Refractories. Silica is the preferred material for crown construction,
in terms of cost and of defect potential, among the possible choices for oxyfuel
combustion
(2).
The typical chemical and mineral composition of a Type A
(3)
silica brick is: 96% SiO2, 0.2-0.4% Al2O3, 2.5-3% CaO, 0.02-0.06 Na2O+K2O, 0.20.8 Fe2O3+TiO2+MgO; 45% cristobalite, 50% tridymite, <1% residual quartz and
<3% glassy Ca-silicate phase
(4).
During furnace operation, the chemical and
mineral composition, and therefore the properties of the brick change.
The
original SiO2 transforms to either cristobalite or tridymite, depending on local
temperatures and presence of alkalis.
Corrosion studies of different
superstructure materials have been performed
(i.e., 5-10).
A qualitative corrosion
model has been proposed for the corrosion of silica under oxyfuel conditions
(5):
(1) deposition of alkali on silica surface, (2) penetration/diffusion of alkali into the
silica, (3) reaction between alkali and silica and formation of low-melting glassy
phases, and (4) dripping of the glassy phases from the crown. Under oxy-fuel
conditions stable Na-silicate glass phases are formed in the higher temperature,
near-surface zones of the silica bricks. Thermodynamic calculations indicate a
higher driving force for NaOH(g) to react with silica refractories under oxy-fuel
conditions (11).
Glass-Contact Refractories. The chemical reactions on the surface of
refractories take place between molten glass, fluxing agents, and/or volatile
components.
Erosion can follow; washing away refractory grains after the
original bond has dissolved.
Corrosion studies are based on operating and
reaction temperatures, reaction rates, and the formation of any product coatings
on the refractory surface. The corrosion tests have been reviewed having in
mind that they must show good reproducibility
(12).
Different tests are compared
Anais do 44º Congresso Brasileiro de Cerâmica
23803
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
regarding reproducibility, trying to establish the most useful test for each
particular case (dynamic finger test, rotating cylinder face method, small rotating
furnace test, crucible test, static finger, and static plate corrosion test).
EXPERIMENTAL TECHNIQUES
UMR Dynamic Corrosion Apparatus. A test apparatus was constructed at
the University of Missouri at Rolla (UMR) for the dynamic corrosion of refractory
materials in molten glass. This dynamic test is an alternative to the ASTM C621
static corrosion test. The equipment is designed to provide test temperatures
from 1300 to 1600 °C, simultaneous multiple samples, separate glass sources,
and variable velocity. A 0 to 90 rpm DC gear motor with a variable speed control
provides precise velocity adjustments. A reticulated foam baffle between the
elements and the test chamber provides uniform heat and lengthens the life of
the heating elements. Standard testing parameters used for screening materials
is rotation at 1400 °C for 48 h. Rotation speed is usually 10 rpm for a 1.2 cm
(1/2”) diameter specimen that results in a surface velocity of 40 cm/min (15.7
in/min).
To accurately report the corrosion, volumetric refractory loss is
determined using image analysis (Fig. 1)
Figure 1. Refractory Loss Evaluation
Anais do 44º Congresso Brasileiro de Cerâmica
23804
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
Oxy-Fuel Simulation Furnace. The Refractories Satellite of the Center for
Glass Research (CGR) has completed construction of an oxy-fuel simulator
furnace at the University of Missouri-Rolla. The furnace has been designed to
reproduce the environment of a working commercial glass furnace. The facility
has a BOC Gases liquid oxygen tank supply (3,000 gallon vertical storage) that
can support running the furnace continuously for up to 2-4 weeks. The main
objective is to study the corrosion behavior of refractories under simulated
industrial conditions. The furnace can function as a pilot-scale furnace that can
melt over 100 kg/day, depending on the experiments planned
(13,14).
The work
comprises also analysis of the gases in the combustion chamber obtained by
sampling the gases with a special probe. A commercial stainless steel watercooled probe produces gas sampling and the temperature of the gases at the
location of sampling. Gases are condensed for chemical analysis of NaOH and
H2O.
Hydrocarbon chemistry is frequently monitored using a UMR adapted
chromatograph system. The sampling is conducted when the combustion space
chemistry reaches stabilization (OFS in steady state). The experimental results
will be fitted to volatilization and thermodynamic models for equilibrium species.
Optical Microscopy Assisted with Cathodoluminescence. Besides the usual
evaluation techniques such as chemical analysis, optical and scanning electron
microscopy (SEM) with microprobe analysis, and X-ray diffraction (XRD),
cathodoluminescence microscopy (CLM) is featured in this work. In general, the
applications of CLM to refractories include the study of phase identification,
distribution, and quantification; follow-up of refractory corrosion by slags and
glasses; sintering effects upon thermal treatment; and glass defects (stones) and
crystallization
(15).
CLM can be used in conjunction with reflected-light (RL) and
transmitted-light (TrL) methods to rapidly identify phases, such as stones in
glasses (16). The technique uses an electron gun that interacts with the specimen
surface. As a result, original refractory minerals and alteration phases produce
CL characteristic colors. Refractories usually leave a “characteristic fingerprint”
Anais do 44º Congresso Brasileiro de Cerâmica
23805
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
or pattern related to their crystal habit in the stones they produce. Table I lists
some main crystalline phases found in glass plant refractories.
Table I. Refractory Phases identified according to CLM
Phase
CL property and habit
Baddeleyite
Dendritic
Corundum
Red CL
Zircon
Sand shape grains, light blue to violet CL
Beta-alumina
Lath-like and bladed crystals, green CL
Alpha-alumina
Red to violet CL
Chromite
Non-CL phase
Na-Ca-aluminosilicate
Non-CL phase
glass
Magnesium lime silicate
Needle-like crystals, blue CL
Ca-silicate phase
Light blue
Cristobalite
Dark blue CL
(“fish-scale” in reflected light, Fig. 2)
Tridymite
Intense dark blue
Figure 2. Reflected light microscopy of silica crown brick.
Anais do 44º Congresso Brasileiro de Cerâmica
23806
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
Measurement of Thermomechanical Properties. The most widely used
parameter to characterize the fracture strength of Refractories, especially at high
temperatures, is the Modulus of Rupture (MOR) in three-point bending. This test
has been standardized for many segments of the ceramic industry and it is
relatively easy to perform
(17).
One example of application has been the study of
the mechanical behavior as a function of temperature of magnesia based
refractories, by evaluating MOR and the static Young’s modulus in three-point
bending and fracture strength by diametrical compression of discs.
The measurement of Modulus of Elasticity and of creep behavior has been
conducted at UMR and at the Oak Ridge National Laboratory
(4).
UMR can
currently measure creep up to 1700 °C at loads corresponding to a maximum of
35.51 KPa (50 psi). Atmospheres, typically inert gases, or gases not corrosive to
oxide specimens can be introduced.
The creep facility at ORNL has been used in several projects involving
creep of glass plant Refractories. It requires a relatively small specimen and can
measure creep parameters and steady state creep to a temperature of 1450 °C.
The large load frame permits stresses in excess of 1.42 MPa (2000 psi). Strain
measurement sensitivity is extremely high, 1x10-6 cm/cm so that the ORNL
equipment can be used for MOE as well.
EXAMPLES OF CHARACTERIZATION
Characterization of Inclusions. Polished-thin sections are systematically
studied by reflected light/cathodoluminescence (RL/CL) microscopy and the
composition of phases in inclusions is determined by SEM-EDS if needed.
Figure 3 shows the microstructure of an inclusion taken with SEM. The inclusion
consists of needle-like  -alumina crystals in a matrix of leucite.
 -alumina
crystals (corundum) exhibit characteristic red (and also violet red) CL color.
Leucite is identified by an intense blue CL color and a dentritic structure. Using
CL, it can also be noted that the inclusion contains another red CL phase. This
Anais do 44º Congresso Brasileiro de Cerâmica
23807
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
phase is identified to be gahnite displaying euhedral crystals and exhibiting dark
and intense red CL (different from CL color of alumina). The source of this
inclusion is probably high alumina refractories.
Figure 3. SEM-BSE image of inclusion showing needle-like  -alumina
(corundum), gahnite, ZnAl2O4, (white crystals) and leucite.
Characterization of Refractories. This work usually involves either the
characterization of brand new refractories and/or the post mortem analysis of
refractories.
refractories
Common examples of analysis are corrosion study of silica
(i.e., 18),
alumina, aluminosilicate and AZS refractories
(i.e., 19)
for glass
tank melters. For instance, two brands of AZS refractories were characterized,
which do not differ much in surface porosity, however with different pore size and
shape. The refractories were characterized for stone potential in TV glasses and
Anais do 44º Congresso Brasileiro de Cerâmica
23808
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
microstructure details. Such analysis allows to indicate, almost always, whether
the origin of inclusions is located at the raw materials or at the refractories. One
refractory (K) had mostly larger needle-shape pores size, around 30-230 µm.
The second refractory (P) had mostly irregular small pore size, <5µm, and about
one third larger pores size 10-25µm. Both refractories had almost identical
mineralogical characteristics. Table II shows the quantified mineralogy results
under the optical microscope (digital analysis) for both refractories.
Table II. Estimated composition of two AZS refractory samples.
Polished
Depth (cm)
Zirconia (%)
Alumina (%)
Section
KS-1
0
22.12
42.26
Glass (%)
35.52
K-2
3.5
25.51
42.16
32.33
PS-1
0
23.63
44.73
31.63
P-2
3.5
20.21
52.78
27.00
The SEM-EDS analysis corroborates the pore morphology described
above, and shows the phase distribution and texture of the specimens (glassy
phase, alumina needles, and pore size and distribution). A summary of the
glassy phase composition is shown in Table III.
Table III. EDS Analysis of Glass Phase in Refractory Surface
Element
KS-1 Glass Phase (wt%)
PS-1 Glass Phase (wt%)
Na2O
2.42
4.94
Al2O3
19.96
24.23
SiO2
76.41
69.41
P2O5
1.21
0.30
ZrO2
0.00
0.54
CaO
0.00
0.17
TiO2
0.00
0.42
Total
100.00
100.00
Anais do 44º Congresso Brasileiro de Cerâmica
23809
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
CONCLUSIONS
Refractories performance affects the economics of industrial glass
manufacture in two ways: by entering into the production cost and by affecting
the quality of the glass through the introduction of defects. Refractories research
at UMR has focused on thermochemistry, phase equilibria, corrosion, creep,
thermal shock, mechanical properties and the constitution and texture of unused
and used refractories.
The current work provides a foundation for developing corrosion test
techniques and ultimately establishing simulative tests for highest wear areas.
The study is currently divided into three sections: post mortem refractory
characterization, initial testing using current techniques, and representative test
development.
Simulation of the conditions present in a glass melter is essential for
conducting crown corrosion experiments. Temperature, gas phase chemistry,
and flow must be reproduced if a small-scale simulation is expected to reflect
industrial conditions.
REFERENCES
1. M. Velez, J Smith, R. E. Moore, “Refractory Degradation in Glass Tank
Melters. A Survey of Testing methods,” Ceramica (Brazil) 43[283-284] 180184 (1997).
2. C. J. Windle, V. K. Bentley, “Rebonded Magnesia-Alumina Spinel Products
for Oxy-Fuel and Alkali Saturated Atmospheres,” pp. 219-225, in Proc.
UNITECR’99.
3. ASTM C-416, Standard Classification of Silica Refractory Brick, Vol. 15.01
4. A. Wereszczak, M. Karakus, K. C. Liu, B. A. Pint, R. E. Moore, T. P. Kirkland,
“Compressive Creep Performance and High Temperature Dimensional
Stability of Conventional Silica Refractories,” Technical Report ORNL/TM13757, March 1999.
5. A. J. Faber, O. S. Verheijen, “Refractory Corrosion under Oxy-Fuel Firing
Conditions,” Ceram. Eng. Sci. Proc., 18[1], 109-119 (1997).
6. L.H. Kotacska, T. J. Cooper, “Testing of Superstructure Refractories in a GasOxy Atmosphere Against High-Alkali Glasses,” Ceram. Eng. Sci. Proc., 18[1],
136-145 (1997).
Anais do 44º Congresso Brasileiro de Cerâmica
23811
31 de maio a 4 de junho de 2000 - São Pedro – S.P.
7. A. Gupta, S. M. Winder, "Ongoing Investigation of Oxy-fuel Firing Impact
on Corrosion of Nonglass Contact Refractories," Ceram. Eng. Sci. Proc.,
17[2], 112-120 (1996).
8. G. Duvierre, A. Zanoli, Y. Boussant-Roux, M. Nelson, “Selection of
Optimum Refractories for the Superstructure of Oxy-Fuel Glass Melting
Furnaces,” Ceram. Eng. Sci. Proc., 18[1], 146-164 (1997).
9. D. Shamp, "In situ Testing of Superstructure Refractories," Amer. Ceram.
Soc. Bull., 76[3] 64-67 (1997).
10. C. A. Paskocimas, E. R. Leite, E. Longo, W. Kobayashi, M. Zorrozua, J. A.
Varela, “Determination of Corrosion Factors in Glass Furnaces,” pp. 89-98
in Proc. 58th Conference on Glass Problems, The American Ceramic
Society, 1997.
11. K. E. Spear, M. D. Allendorf, “Mechanisms of Silica Refractory Corrosion in
Glass Melting Furnaces: Equilibrium Reactions,” to be published in Proc.
of the 192nd ECS Joint Meeting, 2000.
12. M. Dunkl, “Corrosion Tests: A Very Important Investigation Method for the
Selection of Refractories for Glass Tanks,” Glastechn. Ber., 67[12] 325334 (1994).
13. http://www.umr.edu/~lcarroll/cgra.html
14. C. Carmody, M. Velez, L. Carroll, W. D. Headrick, R. E. Moore,
“Refractory Corrosion in Oxy-Fuel Systems,” Ceram. Ind., 149[13] 61-63
(1999).
15. M. Karakus, R. E. Moore, “Post-Mortem Study of Glass Melting Furnace
Refractories,” pp. 179-191, in Corrosion of materials by Molten Glass, The
American Ceramic Society, Westerville, Ohio, 1996.
16. M. Velez, M. Karakus, R. E. Moore, “Identifying Stones Using
Cathodoluminescence Microscopy,” Ceram. Industry, pp. GMT 7-10, Sept.
1998.
17. C. Baudin, W. D. Headrick, R. E. Moore, “Young’s Modulus and Fracture
Stress Testing of Magnesia-Graphite Refractories at High Temperature,”
pp. 413-415, in Proc. UNITECR’99.
18. A. Wereszczak, H. Wang, M. Karakus, W. Curtis, V. Aume, D. Verdow,
”Postmortem Analyses of Salvaged Conventional Silica Bricks from Glass
Production Furnaces,” to be published in Glass Technology, 2000.
19. M. Karakus, R. E. Moore, “Seeking Solutions to Glass Defects from
Refractory Corrosion, the Glass Researcher, 7[2] 13-14; 17 (1998).