Intermetallic Fibre Materials for Hot Gas Filtration in Power Plants and
Combustion Plants

Simon Steigert, Zi Li, GKN Sinter Metals Filters GmbH (DE); Olaf Andersen, Günter Stephani,
FHG IFAM Dresden (DE); Theo Schrooten, INTENSIV-FILTER GmbH & Co KG (DE)

Introduction

In this BMBF sponsored project a high temperature resistant intermetallic filter for hot gas filtration
was successfully developed. As base material for the filter element, melt extracted Ni

3

Al short

fibres were used. With fibres a higher porosity can be reached. Up to 95 % porosity are possible and
after a sintering process the position of each fibre is fixed. With these fibres it is possible to produce
long tubular filter elements by pressing and sintering. Filter efficiency was increased by a thin
coating with Ni

3

Al powder.

Materials properties

The Ni

3

Al-alloy is an intermetallic material containing about 13 weight-% aluminium. The melting

point is about 1395 °C and the density is approximately 7.6 g/cm_ [1]. One advantage of this nickel
aluminide is its relatively high ductility. It is possible to bend fibres with 70 µm diameter more then
45°. This feature was enhanced through alloying with B [2]. Boron refines the grain structure and
increases the ductility.
However, the excellent resistance against high temperature oxidation is equally important. Like in
many other systems, the protection against oxidation is built through an oxide layer. The aluminium
in the alloy makes a thin layer of Al

2

O

3

on the metal surface.

Fig. 1: Grain structure of a Boron-alloyed Ni

3

Al fibre

Materials testing

Filters and other products with high porosity have a high surface area. Due to this, the use of such
products in oxidising or corrosive environment calls for a very good resistance to oxidation. The
oxide layer stops further oxidation and a slow oxidation rate increases the life time.

The oxidation behaviour can be characterised by the growth rate of the oxide layer. The oxidation
behaviour depends on the aluminium content of the Ni-Al-alloy. Higher aluminium content
improves the corrosion resistance because grain size becomes smaller and diffusion of aluminium to
the surface is faster and more aluminium is available to rebuild the protecting oxide surface.
Simultaneously, a small grain size on the fibre surface facilitates a strong adhesion of the
aluminium oxide layer to the matrix. After 1000 hours at 1100 °C in hot air the Ni

3

Al-alloys with

Ta + Zr or Hf + Zr had the lowest weight increasing of all tested materials.

Fibre processing

Although the used Ni

3

Al-alloy has a higher ductility than other intermetallics it is not possible to

make fibres by wire cutting, because it can’t be processed by wire drawing. For our project we
needed short fibres to manufacture a common filter geometry like a tube. With fibres, a higher
porosity can be reached and this improves the air permeability. With the melt extraction process it is
possible to make short fibres nearly from all fusible materials. A rotating wheel with a notched
surface is placed over a melt pool. The rotating extraction device is water cooled and thus generates
a high solidification rate. As a result, homogenous distribution of the alloying elements, small grain
sizes, reduced segregation and extended solubility, as well as the formation of metastable phases is
achieved.
The melt extracted fibres typically show a sickle or kidney shaped cross-section. The project partner
FhG IFAM Dresden has improved the crucible melt extraction process to produce fibres of a mean
equivalent diameter from 50 to 150 µm in batch sizes of one to several kilograms. The fibre length
can be set from 3 to 25 mm with a deviation of approximately ± 15 %.

Fig. 2: Schematic principle of the fibre melt extraction

melt bath

extraction wheel

solidified
filaments

Fig. 3: Fibre examples of different materials and sizes

Filter processing

A common geometry for industrial filters is the filter cartridge – a tube which is closed on one side.
In our process we used an isostatic pressing system to produce 1100 mm long seamless filter tubes.
The diameter for our samples was 60 mm and wall thickness was about 3 mm. A critical point was
the filling of the pressing form with the short Ni

3

Al fibres. It is important to separate the fibres by

the filling in the form. An inhomogeneously filled form causes a bad pore size distribution within
the filter and decreases the filtration performance. To avoid a welding step, we fitted a flange
directly to the filter before pressing and sintering - no more mechanical operations are needed after
sintering.
To improve the filtration behaviour for small particles, it is important to reduce the pore size of the
fibre filter. With a thin powder layer on the outer side of the filter elements the pore size is reduced
from 50 µm to 6 µm. Air permeability reached about 15 % after coating. Theoretically, this value
should be higher but the surface of the fibre filter is not so smooth as that of filters from pressed
powder. So some areas of the coating are thicker than normal.

Fig. 4: Structure of the coated filter 10 = 100 µm

Fig. 5: Finished filter element. 1.1 m long, ca. 5 kg

Filter testing

Besides some in-house tests like bubble-point or permeability, the filters were used in a semi-
industrial filtration device at temperatures up to 800 °C. In this device, the exhaust gas of an oil
burner was mixed with a test dust. This hot and contaminated air was cleaned by the filter elements.
In regular intervals the filter surface was cleaned with a short pressure impulse directed from inside
to the outside of the filter. The 6 bar overpressure pushed the filter cake away from the surface and
increased the permeability.
The mechanical strength was tested with the ring tensile test because of the tube geometry of the
filter. The results are nearly the same as with filters made of conventional stainless steel powder
with equal porosity.

Fig. 6: Chart of the semi-industrial test plant

Test results

Materials properties

After the filtration test run the filter element were tested for strength again. The applied ring tensile
test is a common method in order to measure the tensile strength of circular probes. The differences
to the results before and after 800 °C are not larger than the variation of the results at different
measurements. Other investigations showed that Ni

3

Al-alloys retain their mechanical properties up

to temperatures of 900 °C [1].

Our results of oxidation behaviour are showing that the Ni

3

Al-Alloys with microalloyed elements

have an excellent resistance to oxidation in hot air atmosphere. Figure 7 shows that there is only a
thin, well-adhering oxide layer on the fibre surface - a result of good self protection of the Al-oxide-
layer.

Fig. 7: Thin oxide layer after 1000 h at 1100 °C hot air

Fig. 8: Low corrosion on filter surface after test run at 800 °C

No mechanical deformation was found after the test run. The coated layer was not attacked by the
cleaning impulses or the temperature. The connection between the fibres and the mounting flange
was also strong enough although we used an CrNi-Alloy that has a higher coefficient of thermal
expansion.

The results of the weight gain test due to oxidation in hot air are shown in figure 9. Although the
FeCrAl-alloy have nearly the same good resistance it is fact that the Ni

3

Al-alloys have a

significantly better high temperature strength and hardness. An even better corrosion resistance for
both alloys can be achieved by microalloying with rare earth elements Hf, La and Zr.

White areas: Ni

3

Al base

Grey border: Oxide layer
Black border: Preparation holes

I------------I
100 µm

Big parts: cross section of fibres
Small parts: Coating powder

Fig. 9: Weight gain by oxidation in air at 1100 °C

Filtration

The filter elements reached a good efficiency at all tested temperatures. Test runs were made
between 300 °C and 800 °C. The run was taken for 8 weeks. The dust concentration was reduced
from 10 g/m_ before the filter to lower than 5 mg/m_ after the filter. The efficiency was better than
99.9 % at all times. The used test dust was PURAL with a particle distribution from 0.9 -120 µm
and d

50

= 30 µm. With 4 filter elements an air flow rate of more than 40 m_/h was given. The

pressure drop was nearly constant at 100 mbar due to regular pressure impulse cleaning.

Fig. 10: Results of the filter test run at 800 °C

0

5

10

15

20

25

24

100

200

300

400

500

600

800

1000

holding time (h)

spec. weight gain (mg/cm

2

)

Ni3Al + 0.5 Hf + 0.2 Zr

FeCrAl 23.15 + 0.15 La + 0.1 Zr

FeCrAl 23.15 + 0.2 Y

Ni3Al + 0.2 B

6,500

2,759

3,397

3,206

4,279

99,935

99,972

99,966

99,968

99,957

99,998

99,835

0,224

4,412

Directly after cleaning

impulse

16,464

99,956

0

15

30

45

60

75

90

105

1

2

3

4

5

6

7

8

Nr. of measurements

Dust content (dirty air) [g/m?];

Dust content (cleaned air) [mg/m?];

Efficiency [%]

Outlook

We see a great potential for porous materials based on Ni

3

Al. It is possible to design filters with

specialised geometry, porosity and filter grade. The intermetallic alloys have the potential for use at
temperatures above 800 °C in severe environments. High temperature filtration increases the
efficiency of many industrial processes. In some processes a costly cooling step for the waste
medium could be avoided by using filters that are resistant to high temperatures and thermal shock.

Hot gas filters are needed in power plants, waste combustion plants, steel mills or recycling plants.
Higher temperatures are mostly increase the efficiency of a process. Or expensive gas cooling step
before filtration is not more required.

Porosity, filter grade and geometry can be widely adjusted with these fibres and the kind of
production process.

Acknowledgements

The authors would like to thank Mr. Walter Häde and Mr. Peter Neumann for their vital support in
this project. This project was funded by the BMBF (German Federal Ministry of Education and
Research) under contract no. 03 N 2008.

References

[1]

V. K. Sikka, Commercialization of Nickel and Iron Aluminides, Oak Ridge Nat. Laboratory

[2]

A. I. Taub, S. C. Huang, K. M. Chang, Metall. Trans. 15A (1984) 399