JMEPEG (2013) 22:613–623 ASM International

DOI: 10.1007/s11665-012-0279-z 1059-9495/$19.00

Analysis of Formation and Interfacial WC Dissolution

Behavior of WC-Co/Invar Laser-TIG Welded Joints
P.Q. Xu, J.W. Ren, P.L. Zhang, H.Y. Gong, and S.L. Yang

(Submitted January 15, 2011; in revised form May 26, 2012; published online June 29, 2012)

During the valve fabrication, hard metal is welded to stainless steel or invar alloy for sealing purposes
because of its good heat resistance operating at 500 °C. However, WC (tungsten carbide) dissolution in weld
pool softens the hard metal and decreases mechanical properties near the hard metal/weld interface. In
order to analyze the WC dissolution in welded joint, joining of hard metal and invar alloy was carried out
using laser-tungsten inert gas hybrid welding method. Microstructures of the weld region, chemical com-
position were investigated using optical microscope, scanning electron microscopy, and EDAX, respectively.
Mechanical properties such as microhardness and four-point bend strength test were performed. Larger
and smaller WC dissolution and WC dissolution through transition layer based on thermo-dynamics were
discussed. The results thus indicate that WC dissolution led to cellular microstructure, columnar crystal,
and transition layer under the effect of laser beam and tungsten arc. WC dissolution was affected by metal
ions Fe+, Ni+, Co+ exchange in W-M-C system, and WC grain growth was driven by forces caused by laser
beam and tungsten arc in larger WC, smaller WC, and liquid Fe, Ni systems.

In practical applications, the joining method includes

Keywords hard metal, invar, laser, welding
brazing (soldering), diffusion, high energy beam, explosive
welding, and sinterbonding. The alloy used as filler materials
includes element of VIIIB group in element periodic table (iron,
cobalt, nickel), low melting point element (copper, phosphorus,
1. Introduction silver), and other nonmetal elements (carbon, boron). Brazing is
widely used while the service temperature is not very high.
Multiple layers of Cu and Ni alloys (Ref 12) or BAg8 brazing
Invar alloy has a very low coefficient of thermal expansion filler metals (Ref 13) are also used as insert metal. Past research
(CTE), approximately one-tenth as low as that of stainless steel, indicates that brittle phases, coarse WC, porosity, and crack are
which makes this alloy very attractive for cryogenic (liquefied easily formed and lowers the joint strength. Diffusion bonding
natural gas carrier, LNG), aerospace engineering, and telecommu- (or high-temperature soldering) is another method by selecting
nications applications to meet high-dimensional stability or expan- Ni-Cu-P (Ref 14) or Cu/Ni (Ref 15) as interlayer with better
sion characteristics matched with those of other materials, such as wettability at high temperature. Moreover, dynamic diffusion
glass, ceramics, or composites (Ref 1-4). However, its strength is bonding is reported using Ni-Cu as filler materials (Ref 16).
too low to meet the requirement of its applications (Ref 5, 6). Sinterbonding processing is used for high-temperature appli-
In order to overcome the problem, phase transformation- cations (Ref 17). During the above three processing (brazing,
induced hardening, liquid-nitrogen cooling, cold deformation diffusion, and sintering), enough time and energy can be
with aging treatment, and hard alloy-steel composite structure provided for interface reaction, including liquid flow, element
are widely used to improve the materials mechanical properties diffusion, and sinterbonding. From this point of view, it is
(Ref 7-11). Particularly, hard alloy-steel composite structure is beneficial for the metallurgical bonding of hard metal and steel.
used as pipe-valve component, as shown in Fig. 1. Valve is a On the other hand, it cannot prohibit the formation of brittle
special component and often serves at high temperature. phases: WC aggregation and WC coarsening. So, the issue is
Because the central part (‘‘2’’ in Fig. 1) is subjected to high how to prevent the brittle phase formation.
temperature and high pressure, wear and even corrosion, it is Laser technology has been used in powder metallurgy
required to select high-temperature alloy, for the central part. industry since 1990s (Ref 18). One of the salient features of
Accordingly, invar alloy is selected (‘‘1’’ in Fig. 1) because of this technology is to obtain narrow HAZ (Ref 19, 20). Their
its corrosion resistance and low CTE value. Stainless steels are researches undertaken with different laser sources lead to full
also used instead of invar alloy because of its lower cost. penetration and low cracking welded joints without filler
However, the cracks are easily formed during the welding of materials. Moreover, the increase of cooling rate assists to
hard metal and invar alloy. inhibit the carbon diffusion and brittle g phase formation. In our
previous paper (Ref 21, 22), we have chosen Ni-Fe and Ni-Fe-C
as filler alloys using tungsten inert gas (TIG) welding, which
P.Q. Xu, J.W. Ren, P.L. Zhang, H.Y. Gong and S.L. Yang, College inhibit the brittle g phases completely. But the presence of WC
of Materials Engineering, Shanghai University of Engineering Science, dissolution, aggregation, coarsening deteriorates the mechanical
Shanghai 201620, China. Contact e-mail: xupeiquan7810@yahoo.
properties. This phenomenon is also observed in the other study
com.cn.

Journal of Materials Engineering and Performance Volume 22(2) February 2013—613

Fig. 1 Application of hard metal-steel composite structures (1 steel or invar and 2 hard metal)

Table 1 Materials properties and mechanical properties of base materials

Thermal
Density, PoissonÕs Melting Specific conductivity, Modulus CTE (20 °C),
Materials kg/m3 ratio point, °C heat, J/(kg °C) W/(m °C) elasticity, GPa Strength, MPa ppm/°C

WC-30Co 12,250 0.22 2780 0.47 99.85 416 2000 (bend strength) 8.10
Invar 8120 0.29 1450 0.48 14.58 139 430 (yield strength) 6.14

(Ref 20). On this basis, laser-TIG hybrid welding method is used The microstructures of the samples are characterized by
to join hard metal and invar alloy (Ref 23). means of optical observation and SEM. A Hitachi S-3400N
The goal of this research is to analyze WC dissolution SEM and Electron Probe Microanalyzer EPMA-8705 equipped
during laser-TIG welding on the basis of weld formation with an EDAX analyzer are used for morphology observations
analysis. Consequently, WC dissolution through transition layer and composition analyses. The samples for optical and SEM
based on thermo-dynamics and WC dissolution in liquid W-M-C observations are etched in a solution of mixed acid (1% HF,
system are discussed. 1.5% HCl, 2.5% HNO3, 95% H2O) and reagent (20% KOH +
20% K3 [Fe (CN)6] + 60% H2O vol.%) at room temperature.
After polishing, HXD-1000-type microhardness tester is
used to investigate the microhardness distribution through hard
2. Materials and Welding Procedure metal/weld interface. The load is 200 g; and holding time is
15 s. Four-point bend strength test is also carried out using
Hard alloys used in this study have a chemical composition dynamic mechanical analysis system. The length of samples is
(wt.%) of 4.29 C, 30 Co, and 65.71 W. The other base material 48 mm curvature radius of the sample is 0.1-0.3 mm, and
is an invar alloy. Its chemical composition (wt.%) is 42 Ni, 0.6 chamfer angle in head face is 45 9 c (0.1-0.3 mm).
C, 3.5 Mn, 3 Nb, and balance Fe. Table 1 gives the material
properties and mechanical properties of hard metal and invar
alloy, respectively. The welded joints are prepared using CO2 3. Experimental Results
laser in combination with TIG welder. The laser-TIG hybrid
system consists of laser, CNC platform, TIG welder, and
3.1 Macrostructure and Weld Formation
shielding gas system. Before preparing the joint, all the base
materials are cut into circular shape (diameter: 48 mm; The macrostructure of sample 1 is illustrated in Fig. 2. From
thickness: 6 mm). Two samples with same dimensions are the images, it is a partial joint penetration (PJP) weld: lack of
used with different welding parameters, given in Table 2. Laser side-wall fusion and lack of fusion at the root of the weld. The
spot diameter is 1 mm. Helium gas is employed as the major weld cannot extend completely through the thickness of base
shielding gas; argon is used as the minor shielding gas. metal joined because of the lack of fusion. Their influencing

614—Volume 22(2) February 2013 Journal of Materials Engineering and Performance

Table 2 Laser-TIG hybrid welding parameters and sample number
Defocus Laser Gas flow Welding Welding Gas flow Welding
Sample amount, d0/mm power, P/kW rate, L1(L/min) current, I/A voltage, U1/ V rate, L2(L/min) velocity, v/(m/min)

1 8 10 He, 32 58 8.0 Ar, 10 1

2 8 10 He, 32 149 13.8 Ar, 10 1

Fig. 2 Macrostructure of welded joints (sample 1). (a) Face,

(b) back, and (c) cross section

factors contain insufficient heat input, residual stresses pro-

duced during weld solidification and cooling and the materials Fig. 3 Macrostructure of welded joints (sample 2). (a) Face,
properties difference between both of the base metals. Tunnel (b) back, and (c) cross section
defect near hard metal/weld interface and spatter slag on the top
surface are observed. At the root of welded joint, cold cracks
3.2 Microstructure and Morphology
occur near hard metal/weld interface and hot cracks happen
near invar/weld interface. Near hard metal/weld interface, In order to clarify the microstructure, a further analysis along
brittle carbide (Fe-W-C, WC, and WxC, will be discussed in the cross section is carried out as shown in Fig. 4. The
section 3.3), and constraint degree (caused by thermal stress, microstructures at top, middle, and lower surface regions are
restraint of jig plate) are the two key factors to lead to cold illustrated in Fig. 4(a), (b), and (c), respectively. At the region
crack initiation and propagation. Near invar/weld interface, near top, middle, and lower surfaces, the better metallurgical
because there is Ni-FeS (Ni-FeP) system, which leads to low welded joints are observed. At the region near top and lower
melting point eutectic liquid thin film formation. Consequently, surfaces, the WC dissolves into weld pool. It is also found that
under the effect of tensile stress, the hot crack occurs. WC emerges in a large range of welded seam, as shown in
Moreover, insufficient heat input and inadequate metal flow Fig. 4(d). The microstructure of welded joint includes three
make it impossible to cure the crack after the crack is formed. regions, clearly identified: hard metal/weld, invar/weld, and
The macrostructure of sample 2 is illustrated in Fig. 3. The weld, as shown in Fig. 5. The grain size near hard metal/weld
image shows that better welded joint is obtained without crack interface is larger than that of the near invar/weld interface
or gas porosity. Sufficient heat input and adequate metal flow at (Ref 22). Both of them manifest obvious directivity vertical to the
root weld are confirmed. fusion line. Figure 5(c) illustrates the interface between two

Journal of Materials Engineering and Performance Volume 22(2) February 2013—615

Fig. 4 Microstructures of cross section (sample 2). (a) Top, (b) middle, (c) lower, symbol ‘‘A’’ stands for the area in the black frame near hard
metal/weld interface at the weld root, and (d) an enlarged image of the area ‘‘A’’ in (c) (black arrows indicate the dissolving WC particles to the weld)

different directional solidifications in weld. The eutectic cellular 3.3 Composition

structure c-(Fe, Ni) (Fig. 5d) is observed at the center of welds.
The composition of hard metal/weld interface is carried out
Figure 6 shows SEM images of hard metal/weld interface.
using EDAX analyzer, as shown in Table 3. The spectrum S1
Figure 6(a) illustrates SEM image at low magnification. The
(Fig. 6b) is located in cellular precipitation. The primary
results indicate there are three different morphologies, as shown
element includes Fe, Ni, C, W, and small amount of Co which
in Fig. 6(b), (c), and (d), respectively. The observed columnar,
belongs to W-(Ni, Fe)-C system. The spectrum S2 (Fig. 6b) is
cellular, fir-tree, and dendrite morphology (Ref 22) are different
the matrix in weld which mainly consists of elements Ni and
from that of the observed in laser welds (Ref 20) and TIG welds
Fe. The ratio of Fe to Ni is 1.35. The solid solution includes
(Ref 21). SEM images of superfine crystals are shown in Fig. 7
c-(Ni, Fe), taenite iron nickel (Fe, Ni), Fe0.64Ni0.36, unnamed
formed at the weld near hard metal/weld interface. The
mineral (NR) iron nickel Fe0.93Ni0.056, awaruite FeNi3, kama-
morphology is shown in Fig. 7(b). The direction of superfine
cite iron nickel (Fe, Ni). Therefore, the matrix is nonequilib-
crystal is vertical to the fusion line and its width is <200 nm.
rium c-(Ni, Fe) solid solution and Fe0.64Ni0.36 iron nickel.
The length is 6 lm. In addition, two kinds of typical transition
Figure 6(c) shows fir-tree crystal (spectrum S3), the atomic ratio
zones are observed, as shown in Fig. 8. Figure 8(a) shows the
of W, Fe, Ni, Co, and C is 1:3.9:2.7:1:1.6, which is c-(Ni, Fe)
columnar crystal at weld near hard metal/weld interface. The
solution strengthened by WC phases. The composition of carbide
crystal growth direction is vertical to the hard metal/weld
aggregate (spectrum S4) and metallic binder (spectrum S5) is
interface. In Fig. 8(b), the width of transition layer is 25 lm.

616—Volume 22(2) February 2013 Journal of Materials Engineering and Performance

Fig. 5 Microstructure of laser-TIG hybrid welded joint. (a) Hard metal/weld interface, (b) invar/weld interface, (c) crystal orientation interface
in two directions in weld center. and (d) typical cellular structure

illustrated in Fig. 6(d). The carbide particles mainly consist of 3.4 Microhardness and Bend Strength
W and C; Ni and Fe elements can be omitted. So the carbide
After polishing, microhardness in sample 2 is measured
particles are WC. The matrix is NiFe(Co) solid solution. The
using HXD-1000 type microhardness tester. As shown in
spectrums S6, S7, and S8 are illustrated in Fig. 8(a). In
Fig. 9(a), d stands for the distance from the bottom of the
contrast, the average atomic percent of Fe increases from 14.80
sample to the measured points, the microhardness distribution
to 17.84 to 18.19; this value of Ni increases from 11.41 to 14.69
curve is shown in Fig. 9(b). The vertical axis is microhardness
to 15.06. However, the element Co atomic percent decreases
values (HV0.2); horizontal axis stands for distance from fusion
from 8.61 to 2.26 to 1.45. So the primary constituents of S6 are
line. From the curve in Fig. 9(b), microhardness of hard metal
NiFe(Co) solution; columnar crystal (spectrums S7) and the
is much larger than that of invar alloy (150 HV0.2). The
matrix (spectrums S8) are NiFe solution. Figure 8(b) illustrates
microhardness of weld is nearly 300 HV0.2. In detail, smooth
the spectrums S9, S10, S11, and S12. In transition layer,
curve of microhardness is achieved from weld to hard metal. It
nonequilibrium WC (spectrums S9) is observed. The average
indicates that the better metallurgical bonding has been
atomic percent is 81.57 W, 17.71 C, and 0.73 Fe, and the
obtained. The microhardness test results agree with the
corresponding atomic ratio of W and C is 4.6. So, according to
microstructure analysis.
chemical compounds WC, W2C, W2C0.85, CW3, W6C2.54,
Four-point bend strength test is carried out according to
WC1x, or WCx in current literatures, the nonequilibrium WC
JISR-1601-1995 standard, referenced GB/T14452-93 to esti-
(W23C5 means each WC has lost 18 carbon atoms) is
mate the mechanical properties of welded joints. The failure
characterized with carbon depletion. The gray phase (spectrum
crack initiates from HAZ near invar side (or HAZ near hard
S10) mainly includes Fe and C, so it might be iron carbide. The
metal/weld interface). It propagates through the invar, weld,
spectrum S11 (Fe and Ni) is solid solution. The primary
and hard metal; finally, complete failure occurs in HAZ near
constituents of needle-like structure (spectrum S12) are W, Fe,
hard metal. The bend strength of sample 1 (rbb = 923.54 MPa)
C, and Ni. According to the published literatures so far,
is much lower than that of the sample 2 (rbb = 1483.56 MPa)
chemical compound Fe3W3C, Fe6W6C, Fe2W2C, FeW3C,
because of the presence of cracks and lack of penetration in
Fe3W3C-Fe4W2C, and Fe3.57W9.43C3.54 are observed; therefore,
sample 1. In contrast with our previous research (Ref 22), the
the needle-like structure is nonequilibrium Fe-W-C (g) phase.

Journal of Materials Engineering and Performance Volume 22(2) February 2013—617

Fig. 6 SEM images of (a) hard metal/weld interface, (b) weld, (c) interface, and (d) hard metal

Fig. 7 SEM images of superfine crystals (a) formation position and (b) morphology

maximum value of bend strength for sample 2 is higher than The following three factors lead to the failure of welded
that of the TIG welding sample (rbb = 1350 MPa) using joint: (1) sulfur and phosphorus. If their composition is
Ni-Fe-C alloy as filler metal. SEM images of fracture are shown excessive, hot crack is easily formed near invar/weld interface
in Fig. 10. Figure 10(a) shows the fracture of invar metal; some where crack origin is located (sample 1); (2) WC dissolution,
particles (white arrows) and hole (black arrow) have been aggregation, coarsening, and broken occurs. These phenomena
observed. This might be the main reason for crack initiation. will deteriorate HAZ near hard metal/weld interface where cold
Concaves instead of porosity have been found in fracture of crack initiation and propagation occurs (sample 2). This is
weld (black arrow in Fig. 10b). Figure 10(c) shows the brittle determined by heat flow, forces (arc force, stirring effects of
fracture of HAZ near hard metal/weld interface. The high- welds, high pressure of molten materials and ionized shielding
magnification WC fracture morphology (Fig. 10d) indicates gas and surface tension), and key-hole effect; (3) constrain and
typical intercrystalline and transcrystalline fractures. overheating. This is determined by welding process. Large

618—Volume 22(2) February 2013 Journal of Materials Engineering and Performance

Fig. 8 Typical transition zone near hard metal/weld interface. (a) Columnar crystal and (b) transition layer

constrain and overheating will lead to crack initiation and Table 3 Composition analysis of typical phases
propagation: not only hot crack near invar/weld interface but in welded joints
also cold crack near hard metal/weld interface.
Spectrums Fe K WM CK Ni K Co K

S1
4. Discussions wt.% 33.81 34.80 3.11 22.30 5.97
at.% 39.45 12.33 16.87 24.75 6.60
From the above experimental results, we notice that WC S2
dissolution has led to nonequivalent brittle carbide (Table 3), wt.% 41.72 14.45 2.42 32.56 8.84
at.% 43.13 4.54 11.64 32.02 8.67
unconsolidated microstructures (Fig. 6d), and even crack
S3
failure (Fig. 10b). In general, WC dissolving in welds is wt.% 34.45 29.05 3.05 24.63 8.83
sensitive to heat input. The effect of heat input (tungsten arc or at.% 38.60 9.89 15.88 26.25 9.37
laser beam) on WC dissolution is significant. Under the effect S4
of tungsten arc, WC dissolution occurs at the top side near hard wt.% 1.41 91.05 6.29 1.26 …
metal/weld interface, as shown in Fig. 11(a). Nevertheless, at.% 2.37 46.49 49.14 2.01 …
according to our previous research (2 mm base metal) (Ref 24), S5
WC dissolution has been observed from the top to root across wt.% 41.46 12.13 1.41 28.83 16.18
the section. So, it is the heat input and not the position which at.% 43.90 3.90 6.94 29.04 16.23
determines WC dissolution. This is confirmed by WC disso- S6
wt.% 25.08 15.35 22.45 20.32 15.40
lution resulting from laser and tungsten arc during hybrid
at.% 14.80 2.75 61.59 11.41 8.61
welding (Fig. 11b to d). During hybrid welding, WC dissolu- S7
tion is observed near top surface, root surface, and the middle wt.% 30.81 13.45 22.91 26.69 4.13
section near hard metal/weld interface, as shown in Fig. 11(b) at.% 17.84 2.37 61.65 14.69 2.26
to (d). The influencing factors include stirring effects of welds, S8
high pressure of molten materials, and shielding gas of helium wt.% 32.92 9.10 24.34 28.65 2.78
and argon. at.% 18.19 1.53 62.53 15.06 1.45
WC dissolution changes carbon gradient near hard metal/ S9
weld interface, which lead to nonequivalent g phase, as wt.% 0.95 76.14 22.91 … …
indicated in Table 3. It has been confirmed that the formation at.% 0.73 81.57 17.71 … …
S10
and microstructure of g phase depend on the carbon level and
wt.% 20.90 33.20 25.83 14.03 6.05
temperature (Ref 25). Under the equilibrium condition, g phase at.% 12.28 5.93 70.58 7.84 3.37
is a ternary compound of tungsten, cobalt, and carbon that can S11
exist in two forms, either M6C or M12C (Ref 26). Moreover, the wt.% 26.33 22.99 22.07 20.60 8.01
study by Sarin (Ref 27) indicates that Co3W3C is formed at at.% 16.14 4.28 62.91 12.01 4.65
1425 C while Co6W6C formation occurred at 1000 C. So, S12
necessary thermal condition is different for M6C or M12C wt.% 21.99 44.36 14.12 13.45 6.08
formation. A large grain with a glassy, almost defect-free at.% 18.38 11.26 54.85 10.69 4.81
structure that tends to grow independent of the matrix
characterizes the morphology of Co3W3C and Co6W6C, both
of which have a cubic structure. In current study, under the gradient of carbon is large enough, carbon in c diffuses from
nonequilibrium condition, the g phase depends on the carbon W-Co-C system to weld, followed by diffusion of carbon from
level, temperature, nickel-iron ratio, and WC dissolution. When WC to c phase, thus carbon deficiency in W-Co-C system

Journal of Materials Engineering and Performance Volume 22(2) February 2013—619

 The volume of single WC is
VR ¼ 4pR3 =3 ðEq 1Þ

Vr ¼ 4pr3 =3: ðEq 2Þ

In the proposed system including larger WC, smaller WC
and liquid, the particle quantity per gram is obtained from
3 MR 3 Mr
nR ¼  ; nr ¼  3 ; ðEq 3Þ
4 pR3 q 4 pr q
where MR andMr are gram-molecular mass of larger and
smaller WC, respectively. For WC, MR = Mr = M, q is the
density of WC. In this system, when WC dissolution occurs,
free energy difference DEA between larger WC and smaller
WC is determined by
 
3 rM 1 1
DEA ¼ rDA ¼ r½nr Ar  nR AR  ¼  ; ðEq 4Þ
4 q r R
where AR and Ar are the area of each larger and smaller
WC particles, r is specific surface energy of WC and is
determined by
!
DH
r¼c 1=3 2=3
; ðEq 5Þ
N A Vm
where c is steric dimensionless coefficient. DH, NA, and Vm
are enthalpy difference, AvogadroÕs number, and molar vol-
ume, respectively. With the increase of WC dissolution in
liquid, there exists one dynamic equivalent between larger
WC and smaller WC. Their chemical potential can be
obtained from (Ref 28-30)
ur ¼ u0r ðT Þ þ R0 T LnCr ; uR ¼ u0R ðT Þ þ R0 T LnCR ; ðEq 6Þ
Fig. 9 Microhardness distribution. (a) Analysis position and (b) where u0r (T) and u0R(T) are standard chemical potential per
microhardness distribution curve unit concentration, so, u0r (T) = u0R(T). CR and Cr are the dis-
solvability of between larger WC and smaller WC in liquid.
R0 is the universal constant. Re-writing Eq 4 gives
appears. Subsequently, with the diffusion of Fe or Ni elements,  
W-M (Fe, Co, or Ni)-C system is formed. According to phase Cr
diagram of W-Ni-C, W-Co-C, and W-Fe-C system, range of c DEA ¼ ur  uR ¼ R0 T Ln ðEq 7Þ
CR
phase and WC in W-Ni-C is the largest, followed by that in
W-Co-C, and the range of c phase and WC in W-Fe-C is the Substituting Eq 6 and 7 into 4 leads to
smallest. Therefore, g phases are easily formed in W-Fe-C    
3 rM 1 1 Cr
system. At extreme deficiency level of carbon, the g phase   ¼ R0 T Ln : ðEq 8Þ
4 q r R CR
shows fine irregular shape while at minor carbon deficiency
levels and the g phase shows dendrite morphology. It is Therefore, if R > r, CR < Cr.
indicated that g phase shows large-bulk morphology under Dissolvability of smaller WC is larger than that of the larger
the effect of single tungsten arc during multilayers welding WC in the system. It means while larger WC attains its
(Ref 22). In contrast, when laser beam and tungsten arc are saturation degree in Co or c-(Ni, Fe) liquid, smaller WC has not
used, g phase indicates dendrite morphology consists of fine- attained its saturation degree. Therefore, smaller WC that has
stir structure. During g phase formation, carbon-deficient dissolved in the liquid begins to crystallize on the surface of
regions attract cobalt, which flow into the region until g phase larger WC, which leads to larger WC grain growth. The larger
is formed, and remains trapped until a carbon balance is WC grain growth and the dissolution of smaller WC in liquid
restored. And g phase chemically changes to cobalt and WC result in the increase of saturation degree of smaller WC in
after being decarburized. liquid. Therefore, the system among larger WC, smaller WC,
When WC dissolution happens in the molten pool (liquid and liquid cannot reach thermo-dynamics equilibrium until
Co, Fe, and Ni, see section 3.3), according to WC particles smaller WC disappears and larger WC grain growth. So, the
Fisher sub-sieve sizer (FSSS) in hard metal, they are classified above results indicate that the WC grain growth is driven by
into two types: larger WC ( ‡5 lm) and smaller WC (<5 lm). dissolution difference between larger WC and smaller WC
Assume that WC particles are spherical solid, while Co, Ni, and during nonequilibrium thermo-dynamics process.
Fe are liquid. Set, the radius of larger WC is R; the radius of Nonequilibrium Fe-W-C phase (Fig. 6c, 11) is another WC
smaller WC is r. dissolution, according to boundary layer (BL) theory proposed

620—Volume 22(2) February 2013 Journal of Materials Engineering and Performance

Fig. 10 Fractures of welded joint. (a) Invar alloy, (b) weld, (c) HAZ near hard metal/weld interface, and (d) morphology of WC brittle fracture

by Harrison and Wagner (Ref 31), dynamic factors play a key Interfacial phases can adjust atomic, electron concentration,
effect on WC dissolution. When metal ion Fe+ diffusion rate in and microstructure; control and contain the electrochemical
W-Fe-C system is less than that of the W-Co-C system, potential barrier. Chemical reaction is significant factor to
boundary layer indicates staggered phases. Oppositely, bound- maintain the interface equilibrium. During the WC dissolu-
ary layer consists of mixed aggregate structure (Ref 32-34). The tion, interfaceÕs destruction and reconstruction are affected by
interfacial phases in boundary layer are primarily W-Fe-C mechanical factors and thermo-dynamics factors. Failure of
(g phases). Within, diffusion rate of metal ions are affected by thermo-dynamics stability of hard metal/weld interface
not only intrinsic diffusion rate but also stirring effects of included dissolution, diffusion and chemical reaction. This is
welds, side blown, and even high energy arc and laser beam. affected by geometry of the interface and environmental
WC dissolution in transition layer (Fig. 8b) can be explained condition. Thermal stress can be provided by physical
using transition layer theory (TL theory) proposed by Adams and property difference (thermal expansion coefficient, Elastic
Pask (Ref 32). WC dissolution in transition layer consists of hard modular) of two adjacent phases or heterogeneous temperature
metal base material, boundary layer cluster, transition layer field.
cluster. In the inner of transition layer, metal ions Ni+ and Fe+
replacing Co+ combine with W and C through ionic bond and
covalent bond. Metal ions Ni+ and Fe+ combine with Ni+, Fe+ on 5. Conclusions
the other side through metallic bond. During WC dissolution,
transition layer might be singe molecular layer or multimolecular 1. Cold cracks and hot cracks easily emerge during the
layer. If there is no transition layer, the interface bonding belongs welding of hard metal and invar alloys. On the back side
to molecular bond. Therefore, the formations of transition layer of welded joint, cold cracks easily occur near hard metal/
provide energy for WC aggregation and broken behavior. weld interface and hot cracks happen near invar/weld
According to another model proposed by Walser and Bene interface. Brittle phases such as WC-Fe (Ni, Co), WxCy
(Ref 34), the interface is not only a bridge but also a barrier. are key factors to induce cold crack initiation.

Journal of Materials Engineering and Performance Volume 22(2) February 2013—621

Fig. 11 WC dissolution near hard metal/weld interface. (a) WC dissolution driven by tungsten arc at top side (4 mm, double-welded butt joint
(TIG)), (b) WC dissolution driven by laser beam at middle region, (c) aggregation of dissolved WC, and (d) microsize WC

2. During laser-TIG process, laser beam not only affects the References
bead geometry but also dominates the WC dissolution.
1. G.M. Kadzioka, W. Zarek, E. Talik, and E. Popiel, Mössbauer
However, on the top surface, tungsten arc has more influ- Investigations and Photoemission Studies of the Fe 3s Spin Splitting in
ence on the range of WC dissolution in the weld than Some Fe-Ni Alloys, Acta Phys. Pol., A, 2008, 114, p 1493–1500
laser beam. 2. V.M. Nadutov, S.G. Kosintsev, O. Svystunov, V.M. Garamus,
3. WC dissolution indicates WC grain growth, WC aggrega- R. Willumeit, H. Eckerlebe, T. Ericsson, and H. Annersten, Anti-Invar
tion during welding process. There is a dynamic system Properties and Magnetic Order in fcc Fe–Ni–C Alloy, J. Magn. Magn.
Mater., 2011, 323, p 2786–2791
among larger WC, smaller WC, and liquid, which leads 3. R.R. Mulyukov, K.Y. Mulyukov, and I.M. Safarov, Effect of Severe
to WC dissolution and WC grain growth. There also Plastic Deformation on the Properties of the Fe-36% Ni Invar Alloy,
exist transition layers and boundary layers to provide Phys. Met. Metallogr., 2006, 102, p 91–96
energy for WC aggregation and nonequilibrium W-Fe-C 4. Z.W. Yang, L.X. Zhang, Q. Xue, P. He, and J.C. Feng, Interfacial
(g) formation. Microstructure and Mechanical Property of SiO2-BN Ceramics and
Invar Joint Brazed with Ag-Cu-Ti Active Filler Metal, Mater. Sci. Eng.
A, 2012, 534, p 309–313
5. V.A. Shabashov, A.V. Litvinov, N.V. Kataeva, K.A. Lyashkov, S.I.
Novikov, and S.G. Titov, Formation of Boron Solid Solution in Fe-Ni
Acknowledgments Invar Upon Severe Plastic Deformation, Phys. Met. Metallogr., 2011,
112, p 245–255
This work was supported by the National Natural Science 6. P.Q. Xu, X.J. Zhao, G.X. Xu, L.N. Li, and Z.S. Yu, Microstructure
Foundation of China (Grants Number: 51105240). PQX thanks the Characterisation and CTE Study of Fe-42Ni-Nb Invar Alloys, Mater.
Foundation support from Shanghai Municipal Education Commis- Sci. Tech Lond, 2011, 27, p 655–660
sion and Dr. L. J. Li from Utah State University for the Visiting 7. T.K. Ha, K.D. Lee, J.H. Song, and H.T. Jeong, Effect of Aging
Treatment Conditions on the Microstructure and Strength of Fe-36Ni
Research Fellowship Award, which gives him more research time Based Invar Alloy, Key Engineering Materials, The Mechanical
and opportunities. And many thanks will be given to Dr. Bishal Behavior of Materials X Part 1: 10th International Conference on the
Silwal in Utah State University for his help during paper revision. Mechanical Behaviour of Materials, Vol. 345–346, 2007, p 109–112

622—Volume 22(2) February 2013 Journal of Materials Engineering and Performance

 8. O.A. Khomenko, Origin and Specific Features of Invar Anomalies of 20. C. Barbatti, J. Garcia, and G. Liedl, Joining of Cemented Carbides to
Physical Properties: Fe-Ni Alloys with an FCC Lattice, Phys. Met. Steel by Laser Beam Welding, Materialwiss. Werkstofftech., 2007, 38,
Metallogr., 2007, 104, p 146–156 p 907–914
9. M.K. Kim, J. Namkung, and Y.S. Ahn, The Effect of Si and 21. P.Q. Xu and X.J. Zhao, Analysis of Microstructure and Properties of
Microstructure Evolution on the Thermal Expansion Properties of Cemented Carbide and Invar Alloy Weldment, J. Mater. Eng. Perform.,
Fe-42Ni-Si Alloy Strips, J. Mater. Sci., 2008, 43, p 3112–3117 2010, 19, p 294–300
10. A.I. Uvarov, V.A. Sandovskii, V.A. Kazantsev, E.I. Anufrieva, N.F. 22. P.Q. Xu, X.J. Zhao, D.X. Yang, and S. Yao, Study on Filler Metal
VilÕDanova, and Y.I. Filippov, Effect of Heat and Thermomechani- (Ni-Fe-C) During GTAW of WC-30Co to 45 Carbon Steel, J. Mater.
cal Treatments on the Structure and Physical and Mechanical Sci., 2005, 40, p 6559–6564
Properties of the N30K10T3 Invar, Phys. Met. Metallogr., 2008, 105, 23. P.Q. Xu, Dissimilar Welding of WC-Co Cemented Carbide to
p 45–55 Ni42Fe50.9C0.6 Mn3.5Nb3 Invar Alloy by Laser-Tungsten Inert Gas
11. V.M. Nadutov, Y.O. Svystunov, S.G. Kosintsev, and V.A. Tatarenko, Hybrid Welding, Mater. Des., 2011, 32, p 229–237
Mössbauer Analysis and Magnetic Properties of Invar Fe-Ni-C and 24. P.Q. Xu, D. Ma, and C.W. Ma. Analysis of WC Dissolution
Fe-Ni-Mn-C Alloys, Hyperfine Interact., 2006, 168, p 929–935 Phenomenon Happened in TIG Welded Joint of Cemented Carbide
12. W. Lee, B. Kwon, and S. Jung, Effect of Cr3C2 on the Microstructure and Invar alloy. 2012 Int. Conf. Front. Mech. Eng. Mater. Eng.,
and Mechanical Properties of the Brazed Joints WC-Co and Carbon MEME2012, Hong Kong, July 27–29, 2012
Steel, Int. J. Refract. Met. H, 2006, 24, p 215–221 25. G.S. Upadhyaya, Cemented Tungsten Carbides: Production, Proper-
13. O. Kenichi and K. Mitsuhiko, Elasto-Plastic-Creep Characteristics of ties, and Testing, Noyes Publications, Park Ridge, NJ, 1998
BAg8 Brazing Filler Metal Estimated by Tension-Strain Maintenance 26. O. Eso, Z.G.Z. Fang, and A. Griffo, Kinetics of Cobalt Gradient
Test, Q. J. Jpn. Weld. Soc., 2009, 27, p 96–103 Formation During the Liquid Phase Sintering of Functionally Graded
14. M.I. Barrena, J.M. Salazar, and L. Matesanz, Interfacial Microstructure WC-Co, Int. J. Refract. Met. H, 2007, 25, p 286–292
and Mechanical Strength of WC-Co/90MnCrV8 Cold Work Tool Steel 27. V.K. Sarin, Morphology of Eta Phase in Cemented WC-Co Alloys,
Diffusion Bonded Joint with Cu/Ni Electroplated Interlayer, Mater. Proc. 5th Int. Powder. Metall. Conf., Chicago, IL, 1976
Des., 2010, 31, p 3389–3394 28. Z.Y. Xu, Thermal Dynamics of Materials, Higher Education Press,
15. L. Zhu, L.M. Luo, J. Luo, J. Li, and Y.C. Wu, Effect of Electroless Beijing, 2009, p 121–126
Plating Ni-Cu-P Layer on the Wettability Between Cemented Car- 29. K. Mannesson, J. Jeppsson, A. Borgenstam, and J. Ågren, Carbide Grain
bides and Soldering Tins, Int. J. Refract. Met. H, 2012, 31, Growth in Cemented Carbides, Acta Mater., 2011, 59, p 1912–1923
p 192–195 30. J. Weidow, S. Johansson, H.O. Andrén, and G. Wahnström, Transition
16. F. Andreatta, L. Matesanz, A.H. Akita, L. Paussa, L. Fedrizzi, C.S. Metal Solubilities in WC in Cemented Carbide Materials, J. Am.
Fugivara, J.M. Salazarb, and A.V. Benedettic, SAE 1045 steel/WC-Co/ Ceram. Soc., 2011, 94, p 605–610
Ni-Cu-Ni/SAE 1045 Steel Joints Prepared by Dynamic Diffusion 31. J.D. Harrison and C. Wagner, The Attack of Solid Alloys by Liquid
Bonding: Microelectrochemical Studies in 0.6 M NaCl solution, Metals and Salt Melts, Acta Metall., 1959, 7, p 722–735
Electrochim. Acta, 2009, 55, p 551–559 32. R.B. Adams and J.A. Pask, Fundamentals of Glass-to-Metal Bonding:
17. J. Rodelas, G. Hilmas, and R.S. Mishra, Sinterbonding Cobalt- VII, Wettability of Iron by Molten Sodium Silicate Containing Iron
Cemented Tungsten Carbide to Tungsten Heavy Alloys, Int. J. Refract. Oxide, J. Am. Ceram. Soc., 1961, 44, p 430–433
Met. H, 2009, 27, p 835–841 33. L.G. Hagan and S.F. Ravitz, Fundamentals of Glass-to-Metal Bonding:
18. N.L. Tian and Y.Q. Yang, Study of Laser Molten Welding of Cemented VI, Reaction Between Metallic Iron and Molten Sodium Disilicate,
Carbides and Steel, Proc. SPIE, 1996, 2888, p 185–193 J. Am. Ceram. Soc., 1961, 44, p 428–429
19. A. Costa, R.M. Miranda, and L. Quintino, Materials Behavior in Laser 34. R.M. Walser and R.W. Bene, First Phase Nucleation in Silicon-
Welding of Hardmetals to Steel, Mater. Manuf. Process., 2006, 21, p 459– Transition-Metal Planar Interfaces, Appl. Phys. Lett., 1976, 28, p 624–
465 625

Journal of Materials Engineering and Performance Volume 22(2) February 2013—623