JOURNAL OF MATERIALS SCIENCE 29 (1994) 4889-4896

The morphology and microtexture of M703

carbides in Fe-Cr-C and Fe-Cr-C-Si alloys of
near eutectic composition
G. L. F. P O W E L L , R. A. C A R L S O N
CSIRO Division of Manufacturing Technology, P.O. Box 4, Woodville, South Australia 5011
V. R A N D L E
Department of Materials Engineering, University College of Swansea, Singleton Park,
Swansea SA2 BPP, UK

The microtexture of M7C 3 carbides in undercooled 40 g samples of hyper- and hypo-eutectic

Fe-Cr-C alloys was determined by electron back scatter diffraction. In the hyper-eutectic alloy
the carbides were monocrystalline, while those in the hypo-eutectic alloy were polycrystalline.
While in the former the preferred growth direction of the M7C 3 carbides was [ 0 0 0 1 ] , in the
hypo-eutectic alloy there was a relatively weak texture near [1 0 1 1 ] . There was no evidence
for the presence of growth twins in either the M7C3 carbide rods or in the branching mech-
anism in the joint between the carbide rods of the hypo-eutectic sample. The morphologies of
the M7C 3 carbides resulting from undercooling were used to explain the microstructure of
hardfacing Fe-Cr-C weld deposits applied by the manual metal arc process. The effect of
silicon additions on the morphology of M7C 3 carbides in Fe-Cr-C-Si alloys is explained in
terms of the effect of silicon on undercooling.

1. I n t r o d u c t i o n bides is obvious, whereas if it is perpendicular to the

High chromium white irons based on the Fe-Cr-C rod axis the primary carbides appear to be equiaxed.
eutectic alloy are used extensively to combat abrasive The fine eutectic microstructure shown surrounding
wear in the minerals processing and related industries the equilateral carbide array in Fig. la decreases in
[1]. The white irons are utilized both as castings [2] quantity from the first to the last layer as the quantity
and hardfacing weld deposits [3]. While castings are of primary carbides increases.
produced with a eutectic or slightly hypo-eutectic Undercooling is prevalent in surfacing processes.
microstructure (primary austenite), hardfacing de- Recently, one of the authors demonstrated that a solid
posits are usually hyper-eutectic (primary carbide). state microstructural change which occurred in the
This is achieved by a multi-layer weld deposit to undercooling of alloys could be used to explain the
overcome dilution of the weld layer by the material variation in microstructure of plasma sprayed coat-
from the substrate. Hardfacing deposits are subject to ings [7]. This paper reports on the morphology and
tension cracks which often run vertically from the microtexture of the MvC3 carbide in undercooled
surface of the deposit to the substrate [4]. The fracture Fe-Cr-C alloys and the data applied to explain the
surface reveals fractured (Cr, Fe)TC 3 (M7C3) carbides development of microstructure in Fe-Cr-C hard-
indicating that the fracture path is through the carbide facing weld deposits, particularly in the first and sec-
or through the interface between the carbide and the ond layers of a multi-layer deposit applied by the
matrix I-5]. manual metal arc process. Finally, the role of silicon in
Typically, the first layer of a manual metal arc changing the morphology of MTCa barbides as the
deposit exhibits a substantially eutectic microstruc- result of undercooling is presented and the results
ture with a few austenite dendrites at the interface with compared with those reported by Atamert and Bhade-
the steel substrate. Occasionally additional M 7 C 3 car- shia [8, 9] for the effect of silicon in Fe-Cr-C-Si alloys
bides arranged in equilateral triangular patterns are for consideration as hardfacing alloys.
observed, as shown in Fig. la [6]. A similar pattern of
carbides is sometimes found at the interface between
adjacent layers (Fig, lb) [6]. Second and successive 2. Experimental procedure
layers are hyper-eutectic with the appearance of the The hardfacing welds were deposited in four layers
primary carbides dependent upon the geometrical using a manual metal arc technique with commercial
relationship between the plane of section and the axis 4 mm diameter electrodes of the chromium carbide
of the M 7 C 3 rods. If this plane is parallel to the rod austenitic iron type. This procedure was used to
axis the rod morphology of the primary MTC 3 car- minimize dilution of the top layers from the carbon
0022-2461 9 1994 Chapman & Hall 4889
Figure 1 (a) The microstructureof a perpendicularsectionthrough the first layer of a multi-layerFe-Cr~C alloyhardfacingweld deposit. The
substrate carbon manganesesteel is at the bottom. Adjacent to the substrate are dendrites of austenite. Note the apparentlyequiaxedcarbides
arranged in an equilateral triangular pattern surrounded by a fine unresolvableeutectic of MTC 3 and austenite. (b) The microstructureof
M7C3 carbides in section perpendicularto the substrate at the junction of the first and second layers. Note the arrangement of carbides in
equilateral triangular arrays (arrowed). In these arrays the carbide rods were vertical to the plane of section. In the middle bottom of the
micrograph the carbide rods are parallel to the plane of polish.

manganese steel substrate. A current of 150 A was 3. Results and discussion

used with the d.c. electrode positive. After each weld 3.1. Microtexture of M7C3 carbides in
pass the deposit was allowed to cool tO room temper- undercooled hypo- and
ature before the next layer was deposited. Using this hyper-eutectic alloys
procedure the microstructure of the first layer was Fig. 2 shows the boundary between two "grains" in a
usually eutectic with succeeding layers hyper- sample of F e - C r - C alloy of hyper-eutectic composi-
eutectic. tion 20 wt % Cr and 4 wt % C undercooled > 100 K
Forty gram melts of F e - C r - C alloys were under- below the MTC 3 liquidus temperature. The boundary
cooled > 100 K with respect to both the liquidus and runs from the upper left corner to the bottom right
eutectic temperatures. Final melting was carried out in corner in the micrograph. The term grain is used
a radio frequency furnace using a boron-free soda- following the terminology in Elliott [133 to indicate
lime silicate glass slag. The details of the sequence of growth of a faceted-non-faceted eutectic from a single
melting and alloying prior to the final melting have nucleation centre. Both Laue photographs [103 and
been reported previously [10]. Both hypo-eutectic EBSD [11] showed that the interconnected carbide
(Fe-20 wt % Cr-3 wt % C) and hyper-eutectic (Fe- within each grain had the same orientation, i.e., it was
20 wt % Cr-4 wt % C) alloys were undercoaled. Indi- monocrystalline. In the bottom left-hand corner, the
vidual MvC 3 carbides were examined by electron baek carbide rods were parallel to the plane of polish while
scatter diffraction (EBSD) to determine the relative those in the top right-hand grains were approximately
orientation of adjacent carbides [11]. perpendicular to the plane of section. In the grains the
Samples (50 g) of F e - C r - C - S i alloys of hyper- [ 0 0 0 1] direction was parallel to the rod axis indicat-
eutectic composition with silicon contents up to 7% ing that the rods grew in the [0 0 0 1] direction perpen-
were melted in a radio frequency induction furnace as dicular to the basal plane. The rods were connected as
described previously [12] and cooled at 0.5 K s- 1. The shown by the arrows in Fig. 2. In the upper right-hand
compositions were similar to those described by grain the MvC3 carbide rods formed patterns showing
Atamert and Bhadeshia [8, 9]. For optical metallo- three-fold rotational symmetry. Two common arrays
graphy, the specimens were electrolytically etched in were an equilateral triangle or three carbide rods
aqueous oxalic acid. Prior to scanning electron micro- separated by included angles of approximately 120 ~.
scopy (SEM) for examination of carbide morpho- In this sample the thermal data [103 showed that the
logy, the samples were deeply etched to remove the carbide grew slowly with slow release of latent heat.
matrix. The mierostructure of a hypo-eutectic 20 wt % Cr

4890
Figure 2 Micrograph of hyper-eutecticsample, 20 wt % Cr and 4 wt % C, undercooled > 100 K below the M7C 3 liquidus. The arrows show
joining between the carbide rods (white). In the top right-hand grain the plane of section is transverseto the rod axis. M7C3rods aligned in
equilateral arrays are marked l, while 2 shows carbides with an included angle of approximately 120~

and 3 wt % C sample undercooled > 100 K below rods. The reason for this texture wilt be investigated
both the austenite liquidus and the eutectic temper- by analytical electron microscopy of thin foils of the
atures is shown in Fig. 3. There v~as a boundary joints of the carbides. It may represent branching by
concave towards the left between the eutectic grain on nucleation on active growth sites on the sides of the
the left where the plane of section was approximately carbide rods as occurs in the growth of silicon carbide
parallel to the M~C3 eutectic carbides, and a grain on rods in the vapour-liquid-solid process [14].
the right where the plane of section was transverse to
the carbide rods. Note that the diameter of the carbide
rods was much smaller than in the hyper-eutectic 3.2. Application of microstructural data
sample of Fig. 2. Again, the carbide rods were joined from undercooled specimens
as shown by the arrows. These fine carbide rods were 3.2. 1. Carbide morphology in hardfacing
consistent with a much higher growth rate as evid- weld deposits
enced by substantial recalescence [10]. The fine eutectic of Fig. 1 is illustrated in Fig. 6, where
The carbide rods within a eutectic grain were poly- the joints between adjacent rods are shown by arrows.
crystalline as s h o w n by Laue photographs [10] and An indication of the growth process of the fine eutectic
EBSD [11]. Fig. 4 shows a typical area of an under- of Fig. 6 is shown by Figs 7 and 8. In Fig. 7, the SEM
cooled hypo-eutectic alloy similar to that in Fig. 3 but micrograph shows the distribution of eutectic MTC 3
at a higher magnification. The microtexture of the carbides after the matrix of a conventionally polished
eutectic carbides of Fig. 4 is shown in the inverse pole specimen was removed by acid etching. Note that the
figure of Fig. 5. Unexpectedly, there was no texture carbides appeared to have grown in a colony or
near [ 0 0 0 1 ] but a weak texture close to [ 1 0 ] 1]. No cellular mode from the bottom left to top right with an
twins were observed within the individual carbide elongated cell tip similar to that reported by Matsu-

4891
 1120

0001 /~ 1010
lo~1
Figure 5 Inverse pole figure for the normal to the specimen surface
of the sample in Fig. 4 (i.e. the direction of the carbide rod axes if the
specimen was polished perpendicular to the rods). There was a
texture, near [101 1] predominantly and not near to [0001].

Figure 3 Microstructure of hypo-eutectic sample, 20 wt % Cr and

3 wt % C, undercooled > 100 K below the eutectic temperature.
The arrows show joining of the euteetic MvC3 carbide rods (white).

Figure 6 Micrograph of the fine eutectic of M7C 3 carbide plus

austenite similar to that shown in Fig. 1. Joints in the eutectic
carbides are arrowed.

b a r a a n d c o - w o r k e r s [-15, 16] for u n i d i r e c t i o n a l l y

solidified F e - C r - C alloys. I n a g r e e m e n t with this
suggested m o d e of growth, Fig. 8 shows t h a t the
c a r b i d e s (arrowed) at the p e r i p h e r y of the tip were
c o a r s e r t h a n those n e a r the centre. N o t e the hole in the
centre of the c a r b i d e r o d s which resulted from dissolu-
tion of the m a t r i x core d u r i n g deep etching. T h e fact
t h a t the c a r b i d e s r e t a i n e d the spatial d i s t r i b u t i o n of a n
e l o n g a t e d cell-tip c o l o n y g r o w t h after deep etching
p r o v i d e d further evidence t h a t the eutectic c a r b i d e s
were j o i n e d together.
Figure 4 Typical area of sample in undercooled hypo-eutectic alloy It is n o w possible to suggest h o w the m i c r o s t r u c t u r e
similar to that of Fig. 3 subjected to EBSD. develops in F e - C r - C h a r d f a c i n g weld d e p o s i t s laid

4892
 welded microstructure and those of deliberately
undercooled specimens. In addition to the growth of
the carbide arrays from an undercooled weld deposit,
it is probable that the fine eutectic of Figs 6-8 also
grew from an undercooled weld deposit. Thus, the
carbide in both cases would be branched with the
same textures as those for the undercooled samples
shown in Figs 2 and 3. It follows that if a crack was
initiated as the result of stresses produced by dimen-
sional changes during cooling the continuity of the
carbide would facilitate crack propagation.

3.2.2. Effect of silicon additions on

morphology of M 7C3 carbides in
Fe-Cr-C-Si alloys
Fig. 9a-c shows the top surfaces of solidified 50 g
ingots of nominal 0 wt % Si (S1), 3.6 wt % Si ($2) and
6.9 wt % Si ($3), respectively, prepared from pure
metals. The nominal and actual compositions are
shown in Table I. The suffix (N) indicates the nominal
composition while (P) shows the analysis of an alloy
prepared with pure materials. $2(C) and $3(C) are the
analyses for alloys of nominal compositions $2 and
Figure 7 SEM micrograph of fine eutectic M7C 3 carbides after deep $3, respectively, using commercial purity starting
etching to remove the matrix. Cellular growth from bottom left to
top right is apparent with elongated cell tips. materials. The nominal compositions are those invest-
igated by Atamert and Bhadeshia 1-8, 9] to simultate
the effect of silicon in F e - C r - C - S i hardfacing alloys.
Inspection of Fig. 9 revealed that growth from
nucleation centres on the top surface occurred in all
three alloys with fewer nucleation centres for the
3.6 wt % Si alloy ($2) and the 6.9 wt % Si alloy ($3).
Similarly, in Fig. 10 it can be seen that growth pro-
ceeded from nucleation centres on the top surfaces
when the ingots were prepared from commercially
pure starting materials. It appeared that silicon had a
large effect on reducing the number of nucleation
centres but that the purity of the starting materials
had only a minimal effect.
The microstructures of all five ingots shown in Figs
9 and 10 were similar. The microstructures of two
specimens with similar compositions but made from
starting materials of different purities are shown in
Fig. 11. The nominal silicon content for the specimens
of Fig. 11 was 6.9 wt % Si. In both micrographs a
boundary between a grain on the right, where the
plane of polish was transverse to the carbide rods, and
a grain on the left, where the plane of polish was
parallel to the carbide rods, could be observed. Note

TABLE I Chemical compositions of Fe-Cr-C Si alloys

Figure 8 SEM micrograph of eutectic carbides near end of cell tip
showing coarser carbides (arrows) around the periphery of the tip. No. C Cr Si Mn AI S P
Note holes in carbides from dissolution of matrix core by deep
etching. SI(N) 3.90 33.6 . . . . .
S2(N~ 4.48 34.4 3.6 - -
S2(C) 4.11 34.0 3.57 0.22 0.01 0.01 0.01
S3(N) 3.60 31.2 6.90 - -
down with the manual metal arc process. According to S3(P) 3.50 29.9 7.03 < 0.05 < 0.01 < 0.01 < 0.01
Svensson et al. [17], the cooling rate of the weld $3(C) 3.21 31.6 7.10 0.29 0.06 0.01 0.01
deposit in this process is ~ 2 5 K s -1. At such a fast
(N), nominal; (P), alloy prepared using pure materials; (C), alloys
cooling rate, undercooling before nucleation and prepared using commercialmaterials;C and S by combustion, other,
growth occurs, as evidenced by the similarity of the elements by optical emission spectroscopy.

4893
 Figure 9 Macrographs of the top surfaces of alloys S1 (a), $2 (b) and
$3 (c). Starting with pure materials the silicon content varied from
zero (S1) to 3.6wt% ($2) to 6.9wt% ($3). Note growth from
nucleation centres on top surfaces.

that in the transverse plane the carbide rods tended to

be arranged with three-fold rotational symmetry. This
is especially apparent in Fig. l l a where the carbide
rods were arranged in equilateral triangles (arrowed).
In these ingots the appearance of the M7C3 carbides
in the plane of polish transverse to the axis of the
carbide rods was similar to that shown in Fig. 2 for an
undercooled hyper-eutectic alloy and indicated that
undercooling occurred prior to solidification. Since
silicon was more effective in eliminating nucleation
centres (Fig. 10) than the purity of the starting mater-
ials (Fig. 9a) silicon appeared to be effective in produ-
cing undercooling, in agreement with a previous

Figure 10 Macrographs of the top surfaces of alloys $2 (a) and $3 (b) made from commerciallypure materials.The silicon content varied from
3.57 (a) to 7.1 wt% (b).

4894
Figure 11 Micrographs of S3(P) in (a) and S3(C) in (b). There are two grains in each micrograph. In (a) an equilateral array of carbides is
indicated by the arrow.

report [18]. The appearance of the carbides in the 4. The morphology of the MTC3 carbides in com-
transverse plane of polish was changed from the ex- positions varying from hypo- to hyper-eutectic is
pected microstructure of a hyper-eutectic alloy, The continuous.
normal microstructure of a hyper-eutectic alloy 5. The effect of silicon on hyper-eutectic Fe-Cr-C-Si
consists of large randomly nucleated and oriented alloys with silicon up to 7 wt % was to produce
primary M7C 3 rods in a eutectic matrix of fine inter- undercooling and therefore a change in the morpho-
connected M7C3 rods and austenite. Thus, the car- logy of the M7C 3 carbides from that in the normal
bides on the right-hand side of Fig. 1la and b could be microstructure of a hyper-eutectic alloy to that char-
interpreted as a more equiaxed morphology if not acteristic of an undercooled alloy. This was not a
examined in a SEM after deep etching to remove the change to a more equiaxed morphology and the car-
matrix. This is the likely explanation for the conclu- bides were still continuous.
sion by Atamert and Bhadeshia [8, 9] that the addi-
tion of silicon produced a change in morphology of
the MTCa carbides towards a more equiaxed shape, as Acknowledgements
these authors relied solely on optical metallography. The authors thank Mr Guenter Herfurth for optical
metallography, Mr Peter Lloyd for SEM and Mr
Trevor Kenyon for macrophotography.
4. Conclusions
1. In an undercooled hyper-eutectic Fe-Cr-C alloy
slow growth from the melt produced MvC 3 carbide
References
1. D.E. DIESBURG and F. BORIK, Symposium "Materials for
rods that were coarse, monocrystalline and joined
the Mining Industry" ed. R. G. Barr, 1974, Climax Molyb-
together with a [0 0 0 1] growth texture. denum, Vail, USA, p. 15.
2. In an undercooled hypo-eutectic Fe-Cr-C alloy, 2. ASTM A532-87, Standard Specification for Abrasion-
rapid growth from the melt resulted in a eutectic Resistant Cast Irons.
M7C3 carbide morphology consisting of fine rods 3. "Metals handbook", 9th Edn, Vol. 6 (Welding, Brazing, and
Soldering, ASM OH, 1983) p. 779.
joined together with a weak [1 0 i 1] texture. There
4. ibid. p.776.
were no twins in the individual rods and the branching 5. "Fractographic atlas of steel weldments" (Japan Welding
was inconsistent with a simple twinning mechanism. Society, 1982) p. 234,
3. The microstructure of the first and second layers of 6. G . L . F . POWELL, Australian Welding Research 6 (1979) 16.
a multi-layered Fe-Cr-C hardfacing weld deposit 7. G . L . F . POWELL, J. Mater. Sei. Lett. 10 (1991) 745.
8. S. ATAMERT and H. K. D. H. BHADESHIA, "Proc. of heat
applied by the manual metal arc process could be
treatment '87" (Institute of Metals, London, 1988) p. 39.
explained in terms of undercooling of hypo- to hyper- 9. S. ATAMERT and H. K. D. H. BHADESHIA,"Recent trends
eutectic compositions. in welding science and technology TWR '89", edited by S. A.

4895
 David and J. M. Vitek (ASM International, Materials Park, 15. Y. MATSUBARA, K. OGI and K. MATSUDA, A F S Trans.
USA, 1990) p. 273. 1989 (1981) 183.
10. G . L . F . P O W E L L , Mater. Trans. J I M 31 (1990) 110. 16. K. OGI, Y. M A T S U B A R A and K. MATSUDA, ibid., 197.
11. V. R A N D L E and G. L. F. P O W E L L , J. Mater. Sci. Lett. 12 17. L . E . SVENSSON, B. G R E T O F T , B. U L A N D E R and H. K.
(1993) 779. D. H. BHADESHIA, J. Mater. Sci. 21 (1986) 1015.
12. G . L . F . P O W E L L and L. S. HEARD, Trans. Jpn Inst. Metals 18. G~ LAIRD and G. L F. P O W E L L , Metall. Trans. 24A (1993)
22 (1981) 543. 981.
13. R. E L L I O T T , "Eutectic solidification processing" (Butter-
worths, London, 1983) p. 85.
14. J . V . M I L E W S K I , F. D. GAC, J. J. P E T R O V I C and S. R. Received 17 December 1992
S K R A G G S , J. Mater. Sci. 20 (1985) 1160. and accepted 9 February 1994

4896