Content uploaded by Pavel Kejzlar
Author content
All content in this area was uploaded by Pavel Kejzlar on Mar 22, 2016
Content may be subject to copyright.
Kovove Mater. 54 2016 1–7 1
The influence of ternary alloying element in iron aluminides
on coefficient of thermal expansion
M. Švec1*, P. Kejzlar2
1Technical University of Liberec, Faculty of Engineering, Department of Material Science,
Studentská 1402/2, 46117 Liberec, Czech Republic
2Institute for Nanomaterials, Advanced Technologies and Innovation, Technical University of Liberec,
Studentská 1402/2, 461 17 Liberec, Czech Republic
Received 17 July 2015, received in revised form 13 October 2015, accepted 5 November 2015
Abs tract
The iron aluminides appear as suitable materials with many advantages for use in high-
temperature applications. However, a sharp drop in their strength above 600◦
C and limited
ductility at room temperature are the major obstacles to their mass expansion. There are
several possibilities how to improve these negative properties. One possibility is alloying of
binary alloys by another element.
The main aim of present work is to evaluate the effect of different alloying elements on the
coefficient of thermal expansion with respect to phase composition and structure. Generally,
the high mechanical temperature properties are affected by alloy structure, and the same could
be supposed for the coefficient of thermal expansion (CTE). In this paper, there were studied
Fe-26Al-2X(X= Cr, Si, Ta, Zr) alloys. Structure and phase composition were evaluated by
SEM, EDX, and EBSD. The CTE’s were examined in a temperature range 460–1200◦
Cusing
horizontal dilatometer.
Precipitates of ternary phases can beneficially affect the CTE’s values. The lowest and
most stable CTE was observed for iron aluminide with Ta-addition.
K e y w o r d s: iron aluminides, Fe3Al, structure, coefficient of thermal expansion (CTE)
1. Introduction
The iron aluminides seem to be very perspective
materials for structural applications. They are char-
acterized by a wide range of benefits, for example by
good strength up to 600 ◦
C, excellent corrosion resis-
tance, lower density than stainless steel and low price
of raw materials. However, some negative properties,
for example, a sharp drop in strength above 600◦
C,
are the major obstacles to their mass expansion [1–
4]. Industrial applications require better mechanical
properties at high temperature than can be achieved
by binary Fe-Al alloy. This is the reason for alloying
of iron aluminides by different alloying elements. Nat-
urally, it can be expected that coefficient of thermal
expansion is also affected by the addition of these el-
ements.
Four basic possibilities were described how to
strengthen iron aluminides [5–7].
*Corresponding author: e-mail address: martin.svec@tul.cz
First – by solid solution hardening. This option
is possible in systems with sufficient solubility of al-
loying element in the matrix. The alloying element
is dissolved into the matrix, and thus it strengthens.
Suitable alloying element for this type of reinforce-
ment is, for example, chromium or silicon. Chromium
is dissolved in the Fe-Al matrix up to 50 at.%, and no
other secondary phases can form [6].
Second – by alloy ordering. This option is based
on D03structure stabilization at high temperature.
Through the addition of suitable alloying element (in
this case, for example, titanium), it is achieved that
the transition temperature TC(transition from D03to
B2 lattice) is shifted to higher values than the binary
diagram Fe-Al indicates. Thereby the existence of D03
structure is extended [6].
Third – by incoherent precipitates. Solid solubility
of the most of alloying elements in binary Fe-Al sys-
tem is limited. The higher addition of alloying element
2M. Švec, P. Kejzlar / Kovove Mater. 54 2016 1–7
than is its solubility limit in the alloy leads to the for-
mation of incoherent precipitates. The precipitates are
frequently formed by intermetallic phases, carbides or
borides, and they also contribute to hardening of iron
aluminides [7]. An example of an element that forms
incoherent precipitates can be zirconium.
The last one and the most reinforcing effect is hard-
ening by coherent precipitates. Coherent precipitates
are formed due to miscibility gap. It is a region of the
phase diagram in which two phases with essentially
the same structure are not miscible or soluble. Mis-
cibility gap allows the formation of coherent precip-
itates with strong reinforcement effect. The coherent
precipitates can be observed for example in the sys-
tems Fe-Al-V, Fe-Al-Ni, Fe-Al-Ti or Fe-Al-Ta [6–8].
Iron aluminides appear as ideal materials for high-
temperature applications (and possible substitute for
P91, P92 type stainless steels) due to their excellent
resistance to high-temperature oxidation and corro-
sion. Until now, almost all publications dealing with
Fe-Al alloys are closely related to the study and de-
scription of the structure, high-temperature mecha-
nical properties (strength, creep) and corrosion be-
haviour. The measurement of coefficient of thermal
expansion is important because CTE together with
the high-temperature strength limits the use of these
materials [9]. The knowledge of their behaviour at high
temperatures is essential for the use of iron aluminides
in high-temperature applications.
The dimensional changes are measured by a me-
chanical digital indicator; temperature is measured by
a thermocouple. The true or medium coefficient of lin-
ear thermal expansion (CTE) can be determined from
the recorded data [9–11]. The medium coefficient αmed
is determined from the Eq. (1):
αmed =lT−l0
T−T0
·1
l0
,(1)
where lTis the length of the sample at the given tem-
perature, l0is the length of the sample at the room
temperature (20◦
C), Tis given applied temperature
and T0is room temperature (20◦
C)
The intention to use iron aluminides in high-tech
applications as e.g. steam turbine blade makes the
knowledge of CTE indispensable.
2. Materials and experimental methods
The investigated alloys were Fe3Al-type iron alu-
minides with 26 at.% of Al and addition of 2 at.% of
different elements (chromium, silicon, tantalum, and
zirconium). The samples were prepared by induction
vacuum melting.
For the study of their microstructure, they were
oxide-polished by suspension OP-S. The structure
Fig. 1. A scheme of the apparatus used for measurement
of dilatation cycle.
Ta b l e 1. The applied temperature cycle for dilatation
measurement
Step Conditions
1 25–250◦
C→heating rate 7 ◦
Cmin
−1
2 250–1200◦
C→heating rate 4 ◦
Cmin
−1
3 1200◦
C→delay at temperature 15 min
4 1200–700◦
C→cooling rate 4 ◦
Cmin
−1
5 700–25◦
C→cooled at a rate less than 4 ◦
Cmin
−1
was observed by the light optical microscope Nikon
Epiphot 200 (with the use of differential interference
contrast – DIC). The microstructural details and the
phase composition were studied by the scanning elec-
tron microscope (SEM) Zeiss ULTRA Plus equipped
with Oxford detector for energy-dispersive analysis
(EDX). Presented phases were confirmed using EBSD
analysis (Oxford NodlysNano) on SEM Zeiss ULTRA
Plus.
Thermal expansion was measured using horizon-
tal dilatometer. A scheme of the measuring appara-
tus is in Fig. 1. During the measurement, the sample
was placed into a holder of dilatometer and the whole
system was inserted into the furnace. The holder and
transmissive rod were made from sapphire. Applied
temperature cycle for dilatation measurement is given
in Table 1. The heating cycle lasts about 4.5 h. Re-
peatedly reached deviation was from 5 to 7 % in the
temperature range 25–600◦
C, 2–4 % in range 600–
1000◦
C and 0.5–1 % in range 1000–1200◦
C. The first
cycle (without sample) serves for calibration of the
system. An example of recorded dilatation curves is
given in Fig. 2. Insignificant hysteresis between heat-
ing and cooling curves (Fig. 2) suggests sufficient time
for phase transformation. The CTE values were calcu-
lated from heating curves according to Eq. (1). Slopes
M. Švec, P. Kejzlar / Kovove Mater. 54 2016 1–7 3
Fig. 2. An example of expansion curves: green curve (cir-
cle) – the total extension of the system (sample + sapphire
adapter rod); pink curve (square) – the absolute extension
of sample; blue curve (star) – the relative extension of the
system; red curve (triangle) – the relative extension of the
sample.
Fig. 3. The structure of Fe-Al-Ta alloy with the particles of
Laves phase on grain boundaries and with the precipitates
of Heusler phase inside the grains (initial state).
of the CTE curves were calculated using a linear re-
gression in the temperature ranges corresponding to
the occurrence of investigated phases.
3. Results and discussion
3.1. Microstructure and phase identification
Almost all high-temperature properties are af-
fected by alloy microstructure. Therefore, the mi-
crostructure knowledge is important for the explana-
tion of alloy behaviour and properties at high tempe-
ratures.
There occurs the combination of two hardening
mechanisms in the Fe-Al-Ta system – strengthening
Fig. 4. The structure of Fe-Al-Ta alloy after dilatation cy-
cle. Only Laves phase is presented in the structure – on
grain boundaries and also inside the grains, no precipitates
of Heusler phase were observed.
by incoherent precipitates and by coherent precipi-
tates. The structure of material Fe-Al-Ta is shown
in Fig. 3. The coarser particles on the grain bound-
aries were identified by EDX and EBSD as incoher-
ent C14 Laves phase with chemical composition Fe-
10.1Al-24.5Ta (at.%). Inside the grains, there were
observed very fine and densely distributed particles
of coherent Heusler phase with a thickness about 10
nm, and, therefore, they are very suitable for blocking
of dislocations movement, and they can improve the
creep resistance [12, 13]. The precipitates of Heusler
phase are metastable in the system Fe-Al-Ta. In equi-
librium, the Laves phase is formed. However, the nu-
cleation of Laves phase is a kinetic problem, so the
metastable Heusler phase is formed first. The nucle-
ation problem of the Laves phase is visible in Fig. 3.
The Laves phase particles start to nucleate at the most
appropriate energy location – on grain boundaries. In
Fig. 3 there is also seen a denuded zone without pre-
cipitates, no precipitates of Heusler or Laves phase
were observed in this area. It is caused by the forma-
tion of stable Laves phase precipitates on grain bound-
aries. Tantalum from the matrix goes into the Laves
phase particles. Therefore, the stripe along the line of
precipitates is deprived about tantalum. The amount
of tantalum around Laves phase particles falls below
the solubility limit in Fe-Al-Ta system, the remain-
ing tantalum can be dissolved into the matrix and
the denuded stripe is occurred. In [13] Heusler phase
was observed after annealing at 700 ◦
C. However, only
Laves phase particles were observed after annealing
at 800◦
C. In agreement with TTT diagram [13] and
phase diagrams [14], no precipitates of Heusler phase
were observed in the structure after dilatation cycle
(see a cut in Fig. 4), only Laves phase particles were
observed. Laves phase particles were distributed on
4M. Švec, P. Kejzlar / Kovove Mater. 54 2016 1–7
Fig. 5. The as-cast structure of Fe-Al-Zr alloy (initial
state). λ1Laves phase creates a fine lamellar eutectic with
Fe3Al phase.
grain boundaries and also inside the grains, what is
obvious in Fig. 4, and they can improve the mecha-
nical properties at temperatures above 800◦
C, which
is limiting temperature for the existence of Heusler
phase [12, 13].
The solid solubility of Zr in Fe-Al matrix does not
exceed 0.1 at.% in a whole temperature range, which
is why even a minor addition of Zr leads to precip-
itation of λ1and/or τ1phase according to Al-Fe-Zr
ternary diagram [15]. In the Fe-Al-Zr alloy in an as-
cast state, there was observed a fine lamellar eutectic
composed of λ1Laves phase and Fe3Al in Fe3Al ma-
trix (see Fig. 5). A total volume fraction of the eu-
tectic is approximately 20 %. Small bright particles
appearing in the structure were identified as ZrC and
Zr2CS; their amount is <0.5 %. The presence of ZrC
and Zr2CS could be explained due to a high affinity
of Zr to carbon and sulphur, which were presented
as impurities in a raw material. The structure after
dilatation cycle is in Fig. 6. The fine lamellar eutec-
tic was replaced by coarser particles of Laves phase.
In the detailed cut, there is visible, that ZrC particles
act as a nucleus for precipitation of coarse Laves phase
particles.
In the Fe-Al-Cr system there exists a wide area of
solid chromium solubility in the iron aluminides ma-
trix (see phase diagram in [16]). Therefore, no precip-
itates were observed in the structure, what is obvious
from Fig. 7. Even after the dilatation cycle no changes
in the structure were observed. The same hardening
mechanism applies for Fe-Al-Si system. It is evident
from phase diagram [17] that solid solubility of sili-
con is higher than 2 at.% (the investigated material
has this composition). Therefore, again there were no
precipitates in the structure – see Fig. 8. Also after
the dilatation cycle no changes in the structure were
observed. According to the Hägg’s ratio the atoms of
Fig. 6. The structure of Fe-Al-Zr alloy after dilatation cy-
cle. In the Fe3Al matrix, there are coarser particles of Laves
phase (sometimes with ZrC core).
Fig. 7. The structure of Fe-Al-Cr alloy in initial state – no
particles were observed.
Fig. 8. The structure of Fe-Al-Si alloy in initial state – no
particles were observed.
M. Švec, P. Kejzlar / Kovove Mater. 54 2016 1–7 5
Ta b l e 2. The values of the slopes of CTE curves and phase transformation temperatures
Structure
System D03D03+
precipitates*
B2 B2 +
precipitates*
αFe αFe +
precipitates*
Fe- A l -Ta – SoC: 0.0111
Tr: up to 580◦
C
*Heusler + Laves
–
SoC: 0.0076
Tr: 580–760◦
C
*Heusler + Laves –SoC: –0.0019
Tr: above 920◦
C
*Laves phase
SoC: 0.0016
Tr: 760–920◦
C
*Laves phase
Fe-Al-Zr – SoC: 0.0197
Tr: up to 560 ◦
C
* Laves phase
– SoC: 0.0059
Tr: 560–780◦
C
*Laves phase
– SoC: –0.0029
Tr: above 780◦
C
*Laves phase
Fe-Al-Cr SoC: 0.0166
Tr: up to 600◦
C
– SoC: 0.0042
Tr: 600–960◦
C
– SoC: –0.0019
Tr: above 960◦
C
–
Fe-Al-Si SoC: 0.0093
Tr: up to 680◦
C
– SoC: 0.0055
Tr: 680–1080◦
C
– SoC: –0.0075
Tr: above 1080◦
C
–
Note: SoC is the slope of the CTE curve, and Tris temperature range.
Fig. 9. The chart of CTE curves for the investigated alloys.
alloying element have to have mutual size ratio lower
than 0.59 (for interstitial elements) or in the range of
0.9–1.1 (for substitutional elements) to be solid solu-
ble. In the case of Fe-Al-Cr alloy, there is the atom
size ratio for Cr/Fe = 1.06 and Cr/Al = 1.41. For
Fe-Al-Si alloy the atom size ratio is Si/Fe = 0.71 and
Si/Al = 0.94. With respect to the Hägg’s ratio, it can
be supposed that in the system Fe-Al-Cr there will be
replaced iron atoms by chromium while in the case of
Fe-Al-Si alloy the silicon will replace aluminium.
3.2. Coefficient of thermal expansion
Figure 9 shows the chart of CTE curves for the
investigated alloys. In Table 2 there are summarized
the slopes of CTE curves in different phases for all
investigated alloys.
For the binary Fe26Al alloy phase, transformation
temperatures were calculated from the phase diagram
[1]. The binary alloy undergoes two phase transforma-
tions (D03↔B2 ↔αFe) at approximately 550◦
Cand
790◦
C.
From the chart (Fig. 9) is seen that the CTE val-
ues for Fe-Al-Ta are the lowest in a whole range of
measured temperatures. The CTE values are also rel-
atively stable from medium to high temperatures (the
values fluctuate between 14.5 ×10−6and 18 ×10−6
K−1). The first change in the slope of CTE curve is
seen at 580◦
C when the structure is changed from D03
to B2. The slope change is relatively minimal due to
the existence of coherent Heusler phase and incoherent
Laves phase in the structure. Heusler phase disappears
at approximately 760◦
C – the second slope change in
the chart. In the temperature range 760–920◦
Cthere
exists B2 matrix with Laves phase precipitates. The fi-
nal slope change occurs at 920◦
C when the ordered B2
transforms to disordered αFe. The Laves phase parti-
cles remain in the structure.
The initial structure of Fe-Al-Zr is composed of
amixtureofD0
3+ Laves phase. The presence of
ZrC precipitates can be neglected, because their vol-
ume fraction is too low to affect the CTE. The CTE
curve corresponding to D03+ Laves phase structure
rises the most steeply of all investigated alloys. The
first change in the CTE curve slope is evident at
560◦
CwhereD0
3lattice transforms to B2. CTE fur-
ther grows linearly up to 780◦
C (transformation from
ordered B2 to disordered αFe). Above 780◦
C, the CTE
values decrease with the temperature.
6M. Švec, P. Kejzlar / Kovove Mater. 54 2016 1–7
The CTE curve of Fe-Al-Cr alloy grows in the tem-
perature interval from 460 to 960◦
C. The maximum of
CTE value is reached at 960◦
C. The first decrease in
the slope of the CTE line is evident at a temperature
of 600◦
C where occurs a phase transformation from
D03to B2. A decrease of CTE values with increasing
temperature starts at 960◦
C. The decrease is related
to a transformation from the ordered B2 structure to
disordered αFe.
In the temperature range from 460 to 1080◦
C, the
CTE of Fe-Al-Si alloy grows almost linearly with the
temperature. A small decrease in the slope of the CTE
curve could be observed around 680◦
C approximately
when the phase transformation from D03to B2 occurs.
The CTE values decrease above 1080◦
C, what could
be explained by a transformation from B2 to the disor-
dered αFe lattice. The Fe-Al-Si alloy shows the high-
est CTE values in the whole temperature range for
all investigated samples. In the Fe-Al-Si, there applies
the same reinforcing mechanism as in the Fe-Al-Cr
system. However, the coefficient of thermal expansion
is significantly higher for Fe-Al-Si. This is probably
caused by the fact that in the Fe-Al-Si system silicon
replaces aluminium, however in the Fe-Al-Cr system
chromium substitutes iron. With respect to mecha-
nical properties, it is more beneficial to replace iron
than aluminium in the case of Fe-Al-Xternary sys-
tem.
3.3. Summary
Besides the temperature, CTE values are strongly
affected by the structural lattice. The steepest increase
in CTE curve shows the D03lattice. The milder in-
crease of CTE with the temperature was observed in
the case of the B2 lattice. On the other hand, the
CTEs of αFe decrease with the growing temperature.
Stable and low values of CTE are required for the
use at high temperatures. Due to this fact, B2 and
αFe lattices are more beneficial than D03. For keep-
ing the lowest possible CTE values, it is desired that
D03↔B2 and B2 ↔αFe transformations took place
at lowest temperatures. It is necessary to keep in mind
that Fe-Al-type alloys are suitable for machine parts
up to 800◦
C with respect to their high-temperature
mechanical properties.
It is obvious (see Table 2) that transformation tem-
peratures are affected by the alloying elements. While
the addition of Cr, Ta and Zr does not affect the
D03↔B2 phase transformation temperature signifi-
cantly, the addition of Si shifted the transition tempe-
rature from 550 to 680◦
C. Except Zr addition sim-
ilar phase transformation temperature shift due to
the addition of ternary element was observed in the
B2 ↔αFe transformation.
The effect of ternary element addition on the slope
of CTE curves is also obvious from Table 2. The
most stable CTE values were observed in the case
of Ta- and Cr-alloyed samples. The addition of Zr
caused the steepest grow of CTE at lower tempera-
tures, above 560◦
C the SoC (slope of the CTE curve)
values are comparable to the other alloys. Si-alloyed
sample shows the steepest decrease of CTE in the αFe
area.
With respect to CTE, the addition of Ta is the
most beneficial for the required application tempera-
tures for Fe-Al alloys (approximately up to 900◦
C).
Results obtained by Cr- and Zr-doped alloys were
similartoeachother(Crseemstobebetterattempe-
ratures below 860◦
C, above 860◦
C Zr should be pre-
ferred).
The addition of Si led to an increase of CTE in the
whole tested temperature range.
Other studies in Fe-26Al-2Xsystems performed by
DTA show well comparable results of transition tem-
perature values to the results obtained from CTE mea-
surements. E.g., in [18] there were measured D03↔B2
transition temperatures (TC) by DTA. According to
[18], TCD03↔B2 for Fe-26Al-2Ta alloy (in at.%)
is approximately 580◦
C, for Fe-26Al-2Zr (in at.%) is
550◦
C and for Fe-26Al-2Cr (in at.%) is 610 ◦
C(for
comparison with present CTE data see Table 2).
4. Conclusions
Obtained CTE values are essential for construc-
tion engineers in designing structural parts for high-
temperature application. Also, the data obtained from
CTE curves can be used for explanation of high-
temperature deformation behaviour (relation between
high-temperature strength and phase transformation)
of Fe-Al-Xalloys.
– The lowest CTE was measured for material Fe-
Al-Ta. Low and relatively stable CTE is very suitable
for the materials used in structural applications. The
most stable CTE values were observed in the case of
Ta- and Cr-alloyed samples. It is obvious that tanta-
lum assign to similar positive influence on CTE values
stability as chromium (that is considered as a strategic
element), the replacement of chromium by tantalum
should be beneficial. In addition, it could be reached
saving of alloying element (usual addition of tantalum
in iron aluminides is <2 at.%, while chromium 5–10
at.%.
– Precipitates formed due to ternary addition can
beneficially affect the values of CTE similarly as high-
temperature mechanical strength.
– The present study can be further extended by:
– the effect of other ternary alloying elements;
– the effect of different Al concentration with fixed
ternary element concentration;
– the effect of different ternary element concentra-
tion with constant Al content.
M. Švec, P. Kejzlar / Kovove Mater. 54 2016 1–7 7
Acknowledgements
The results of this project LO1201 were obtained with
co-funding from the Ministry of Education, Youth and
Sports as part of targeted support from the programme
“Národní program udržitelnosti I”.
References
[1] Mc Kamey, C. G.: Physical Metallurgy and Process-
ing of Intermetallic Compounds. Eds.: Stoloff, N. S.,
Sikka, V. K. New York, Springer 1996.
doi:10.1007/978-1-4613-1215-4 9
[2] Stoloff, N. S.: Mater. Sci. Eng. A, 258, 1998, p. 1.
doi:10.1016/S0921-5093(98)00909-5
[3] Stoloff, N. S., Liu, C. T.: Microstructure and Proper-
ties of Materials. Iron Aluminides. Volume 2. Singa-
pore, Word Scientific Publishing Co. Ptc. Ltd. 2000.
[4] Deevi, S. C., Sikka, V. K.: Intermetallics,4, 1996, p.
357. doi:10.1016/0966-9795(95)00056-9
[5] Morris, D. G.: Intermetallics, 6, 1998, p. 753.
doi:10.1016/S0966-9795(98)00028-4
[6] Palm, M.: Intermetallics,13, 2005, p. 1286.
doi:10.1016/j.intermet.2004.10.015
[7] Palm, M., Schneider, A., Stein, F., Sauthoff, G.:
Mater. Res. Soc. Symp. Proc., 842, 2005.
doi:10.1557/PROC-842-S1.7
[8] Risanti, D. D., Sauthoff, G.: Intermetallics,13, 2005,
p. 1313. doi:10.1016/j.intermet.2004.12.029
[9] Švec, M., Hanus, P., Vodičková, V.: Manufacturing
Technology, 13, 2013, p. 399.
[10] Porter, W. D., Maziasz, P. J.: Scripta Metall. Mater.,
29, 1993, p. 1043. doi:10.1016/0956-716X(93)90175-R
[11] Kanagaraj, S., Pattanayak, S.: Cryogenics, 43, 2003,
p. 399. doi:10.1016/S0011-2275(03)00096-1
[12] Hanus, P., Bartsch, E., Palm, M., Krein, R., Bauer-
Partenheimer, K., Janschek, P.: Intermetallics, 18,
2010, p. 1379. doi:10.1016/j.intermet.2009.12.035
[13] Risanti, D. D., Sauthoff, G.: Intermetallics, 19, 2011,
p. 1727. doi:10.1016/j.intermet.2011.07.008
[14] Raghavan, V.: J. Phase Equilib. Diff., 34, 2013, p. 328.
doi:10.1007/s11669-013-0239-9
[15] Raghavan, V.: J. Phase Equilib. Diff., 31, 2010, p. 459.
doi:10.1007/s11669-010-9746-0
[16] Raghavan, V.: J. Phase Equilib., 24, 2003, p. 257.
doi:10.1361/105497103770330587
[17] Ghosh, G.: Aluminium-Iron-Silicon. In: Landolt-B¨orn-
stein Database.New Series IV/11 A2.
[18] Anthony, L., Fultz, B.: Acta Metall. Mater., 43, 1995,
p. 3885. doi:10.1016/0956-7151(95)90171-X
