Erhöhung der Beständigkeit von Hochtemperaturwärme-dämmmaterialien in Industrieöfen mit H2-haltiger Schutzgasatmosphäre (EBEST)

Förderinstitution:AiF/ GVT
Bearbeiter:Dipl.-Ing. Jürgen Rank
E-Mail:juergen [dot] rankatiwtt [dot] tu-freiberg [dot] de
Projektlaufzeit:01.02.2005, 2 Jahre


The most important protective gases in industry furnaces are H2, H2/N2-mixtues and gases produced by sub stoichiometric combustion of natural gas. These hydrogen containing atmospheres are used in:

  • heat treatment of steel (e.g. bright annealing)
  • production of metallic or ceramic sinter materials
  • chemical industry

The number of innovative applications in the powder metallurgy and ceramic industry using hydrogenous atmospheres was increasing throughout the last years. In order to minimize construction and energy costs more and cheaper mullite and other SiO2 containing materials are used for insulation purposes instead of expensive corundum materials.

The limiting factor of this development is the knowledge of the behaviour of silica containing materials in hydrogenous atmospheres. The goal of the research project is the quantification of the destructive influences on the insulation materials by hydrogenous atmospheres in industrial furnaces. The experiments were made with focus on corrosion mechanism, corrosion rate depending on the experimental setup (temperature, time, atmosphere), material characteristics (chemical composition, porosity, structure) and determination of relevant material properties before and after the exposure to hydrogen. This results in a durability prediction of the refractories.

Corrosion mechanism

The reducing atmosphere avoids the oxidation of the processed goods, at the other hand the insulation material of the system Al2O3/SiO2 are being reduced due to reaction 1 [2, 3, 4].

EBEST Formel    (reaction 1)

The corrosion rate depends on different parameters like:

  • Chemical composition of the insulation
  • Porosity and pore structure
  • Composition, pressure and velocity of the atmosphere
  • Temperature
  • Time
  • Moisture
  • Material impurities

The SiO equilibrium partial pressures are calculated for different parameters in order to get the direction for critical operating conditions. Figure 1 shows the SiO partial pressures depending on temperature, moisture and pressure.

EBEST Diagramm 1    EBEST Digramm 2
Figure 1: SiO equilibrium partial pressures depending on moisture and pressure (calculated with Factsage 5.4)

The moisture of reducing combustion atmospheres is between 0.1 and 1 %. Under these conditions the SiO equilibrium partial pressure is reduced and the reaction is slows down. A pressure reduction in the furnace leads to a more intensive corrosion.

To investigate the influence of the chemical composition different types of insulation (light weight insulation brick – FL, fireclay brick – S, and insulation wool -F) with varying SiO2 content have been tested in hydrogen atmospheres. Table 1 shows the properties of the tested materials.
Table 1: Materials

 FL 1FL 2S 1F 1F 2F 3
classification temperature in °C184015401660140016001500

chemical composition in %

density in kg·m-3150010002600128130100
compressive strength in MPa124≥120   
thermal conductivity in W·m-1·K-1 at      
600 °C
800 °C
1000 °C
1200 °C
1400 °C

The stone S 1 and FL 2 have almost the same chemical composition but different densitiy and porosity. FL 1 is the reference material made of alumina. The dependence on chemical, time and temperature is shown in figure 2. The materials have been tempered in 100 % H2 for 192 h.

EBEST Diagramm 3    EBEST Digramm 4
EBEST Diagramm 5    EBEST Digramm 6












Figure 2: Relative mass loss (in 100 % H2)

Comparing the absolute mass loss, the massive S 1 is showing higher mass losses than the FL 2, although this material has a higher porosity and consequently a larger surface area. In the pores of the material SiO partial pressure seems to be build up, that the overall corrosion rate is decreasing.

EBEST Diagramm 7






Figure 3: Absolute mass loss FL 2 and S 1

The mass losses fit to the chemical composition change of the material according to reaction 1. The corrosion is not homogeneous. After 192 h at 1500 °C (100% H2) there was only corundum found at the edge of the sillimanite stones, while in the middle of the specimen the original composition was found by x-ray diffraction. Table 2 shows the found chemical compositions (energy dispersive x-ray) and Figure 4 shows the found structures (scanning electron microscopy).
Table 2: SiO2-content S1, 192 h at 1500°C, 100 % H2 (EDX-Messung)

 surface20 mm from the surface35 mm from the surface
Al2O398,40 %60,79 %64,72 %
SiO20,43 %36,90 %33,91 %


EBEST Bild 1  EBESR Bild 2  EBEST Bild 3



Figure 4: S 1 a) edge b) 20 mm and c) 35 mm from the edge (192 h at 1500 °C, 100 % H2)

The progressive corrosion is shown on the F 1 material in figure 5.

EBEST Bild 4  EBEST Bild 5  EBEST Bild 6




Figure 5: F1 new, after 96 und 192 h, 1400 °C, 100 % H2)

In sub stoichiometric combustion atmospheres no mass change is recognized up to 1050 °C [7]. Experiments in formier gas showed that the corrosion rates are smaller than in pure H2 atmosphere.

EBEST Diagramm 8







Figure 7: Corrosion rate at 1450 °C FL 2

The materials did not show obvious change in heat conductivity and hot bending strength under the tested conditions [5]. The changes are in the range of measuring accuracy. Cold compressive strength of the silica containing materials (S 1 and FL 2) is reduced depending on the time and temperature exposed to H2 (figure 6). With the prediction of SiO2 loss of the material a durability expectancy can be calculated (figure 6).

EBEST Diagramm 9    EBEST Diagramm 10







EBEST Diagramm 12

EBEST Diagramm 11








Figure 6: Compressive strength of the stones and durability expectancy of S 1

The flexibility and resilence of the wools are reduced by corrosion and recrystallization [6]. In flowing atmospheres wools are partly eroded.

Under certain conditions the experiments allow to estimate the damage on Al2O3/SiO2 materials by hydrogeneous atmospheres. This is expressed by a "damaged layer" of the insulation material. It is not possible to estimate the behaviour of the entire insulation in the lifetime of furnaces. Two scenarios have to be considered:

  • If there is no abrasive stress (mechanical/atmospheric) the corundum matrix remains after the SiO2 is corroded.
  • If there is a abrasive stress there will be a material removal of the insulation (erosion, flake off). At high temperature this could be less because the corundum starts sintering.

In the worst case the damage of the whole insulation must be considered, but up to now it is not predictable.
The corrosion rates of all tested wools the neglectable up to 1100 °C in 100 % H2. Nevertheless the wools loose their resilence even at this temperature. They should be used only as back insulation.

The corrosion rates of ceramics in hydrogeneous atmospheres are determined by mass loss. This mass loss depends on the SiO2 content of the material, temperature and exposure time to H2 atmospheres. For all specimen there is an influence of geometry and size. The mass loss increases with the surface/volume ratio of the specimen. The porosity has a weaker influence on the corrosion rate, although the light weight materials and the stones loose less mass than the wools. The corrosion rates are increasing strongly at higher temperatures. Also the results are influenced by the flow conditions in the furnace.
The cold compressive strength of the materials is reduced by H2 corrosion. Other properties like hot bending strength and thermal conductivity did not change under the tested conditions.
With the reduction of SiO2 according to reaction 1 the SiO partial pressure of the atmosphere increases. At temperatures below 1000 – 1050 °C the equilibrium shifts to the reactants and the SiO and the H2O vapor form H2 and SiO2. Therefore it must be ensured that the SiO2 is not influencing product quality or furnace operation. Vitrifications have to be considered in the flue gas tract of the furnace.


[1] Z. Hrabě u.a: Einfluss der Zusammensetzung der Ofenatmosphäre auf die Eigenschaftsänderungen von Wärmedämmerzeugnissen. Stavivo 64 (1986) 4, S.158-160
[2] M. S. Crowley: Hydrogen-silica reactions in refractories. Am. Ceram. Soc. Bull. 46 (1967), S.679-682
[3] M. S. Crowley: Hydrogen-silica reactions in refractories – Part II. Am. Ceram. Soc. Bull. 49 (1969), S.527-530
[4] S. T. Tso, J. A. Pask: Reaction of Silicate Glasses and Mullite with Hydrogen Gas. J. Am. Ceram. Soc. 65 (1982) 8, S.383-387
[5] Wulf, R.; Groß, U.; Barth, G.: Wärmeleitfähigkeit keramischer Fasermatten – vergleichende Messungen mit unterschiedlichen Methoden. Keramische Zeitschrift 9/10 (2004) S. 554-561
[6] T. Bolender: Untersuchung zur Temperaturstabilität von "biolöslichen" Keramikfasern im Vergleich zu Keramikfasern des Systems Al2O3/SiO2. Gaswärme International 50 (2001) 9, S.405-410
[7] P. Dietrichs, G:S: Dhupia, W. Kröhnert: Das Verhalten keramischer Hochtemperaturfasern und Faserwerkstoffe in CO-Atmosphäre. Beitr. Elektronenmikroskop. Direktabb. Oberfl.16 (1983), S.233-244