Journal of Materials Processing Technology 239 (2017) 240–250

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

journal homepage: www.elsevier.com/locate/jmatprotec

Reduction of ductility-dip cracking susceptibility by

ultrasonic-assisted GTAW
Chen Hua a , Hao Lu a,b,∗ , Chun Yu a,∗∗ , Jun-Mei Chen a , Xiao Wei a , Ji-Jin Xu a
a
School of Materials Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240, PR China
b
Shanghai Key Laboratory of Materials Laser Processing and Modification, School of Materials Science and Engineering, Shanghai Jiao Tong University,
Shanghai 200240, PR China

a r t i c l e i n f o a b s t r a c t

Article history: The ability to reduce ductility-dip cracking (DDC) susceptibility in nickel filler metals is one of the keys
Received 8 May 2016 to their reliable applications in nuclear industry. Here, it is demonstrated that improvement in DDC
Received in revised form 10 August 2016 susceptibility can be attained for nickel filler metal 52M (FM–52M) through the superimposition of ultra-
Accepted 16 August 2016
sonic field during gas tungsten arc welding (GTAW). A concept of Detrimental Grain Boundary Length
Available online 18 August 2016
(DGBL) has been proposed to elucidate DDC features in two kinds of samples: without ultrasound and
with 20 kHz ultrasonic-assisted GTAW. Electron backscattered diffraction, thermo-mechanical simula-
Keywords:
tion tests, scanning electron microscopy and optical microscopy were utilised to quantitatively analyse
Ultrasonic-assisted GTAW
Filler metal 52M
the microstructure and the cracking features in FM–52M. The results reveal that the external ultrasonic
Ductility-dip cracking field can refine the grain of FM–52M welds, and can further decrease the DGBL from 2890 ␮m to 1700 ␮m.
Detrimental grain boundary length This decrement of ∼41.2% in DGBL arises from the violent stirring induced by ultrasonic cavitation. The
suppressed DGBL, under ultrasonic circumstance, shortens the potential path for crack propagation, and
improves the DDC resistance in FM–52M.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction with coarse grains are easy to suffer from DDC. Ramirez et al. (2006)
evaluated the effect of GB morphology on DDC tendency in nickel-
Alleviating ductility-dip cracking (DDC) by a suitable welding base alloys. Our previous research (Chen et al., 2014) revealed that
process is an effective method to improve the performance of the GBs within a particular range of misorientation, 30–45◦ , can
nickel-base components in nuclear industry. The DDC is a type of easily induce DDC in FM–52M. These particular GBs are denoted
hot cracking that occurs in the temperature range between 50% as detrimental GBs in the present paper, since they are harmful
and 80% melting point of an alloy (Noecker and DuPont, 2009). to DDC prevention. Obviously, reducing the detrimental GBs can
It has been identified as a serious weldability issue associated benefit DDC prevention with the help of an appropriate welding
with nuclear materials, such as nickel filler metal 52M (FM–52M) process.
reported by Mo et al. (2015). The DDC in FM–52M may reduce the DDC propagation tendency relies heavily upon the connection
reliability of nuclear equipment and even lead to catastrophic acci- and gathering of these detrimental GBs. However, such connection
dents if it is not in control. Therefore, numerous attempts have and gathering were scarcely assessed in a quantitative way in previ-
been made to explore the cause of DDC. For instance, Nissley and ous studies. Moreover, few investigators have proposed a welding
Lippold (2008) developed a strain-to-fracture (STF) test based on procedure which can control the detrimental GBs effectively.
thermo-mechanical simulation. Motivated by these shortcomings, ultrasonic-assisted arc weld-
Using STF tests, many investigators have suggested the possible ing is considered as a promising method to tailor detrimental GBs
mechanisms of DDC, one of which deals with the grain boundary in this study. This is because an ultrasonic field can induce a
(GB) character. Lippold and Nissley (2008) reported that materials unique physical circumstance for weld solidification. For example,
Watanabe et al. (2010) introduced ultrasound into welding of fer-
ritic stainless steel, Dai (2003) into aluminium alloy and Morisada
et al. (2013) into titanium alloy. Evident changes of GB morphol-
∗ Corresponding author at: School of Materials Science and Engineering, Shanghai
ogy were demonstrated in these alloys after the ultrasonic field
Jiao Tong University, Shanghai 200240, PR China.
∗∗ Corresponding author. was employed. Therefore, it is promising to tailor GB morphology
E-mail addresses: [email protected] (H. Lu), [email protected] (C. Yu).

http://dx.doi.org/10.1016/j.jmatprotec.2016.08.018
0924-0136/© 2016 Elsevier B.V. All rights reserved.
 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250 241

Table 1
Chemical composition of FM–52M.

Element Ni C Mn Cr Fe S P Ti Al Si Cu Nb

Wt Pct 59.54 0.02 0.8 30.06 8.22 0.015 0.020 0.224 0.11 0.09 0.02 0.84

Fig. 1. Experimental system of the U-GTAW that generates ultrasonic field by using the welding arc as an ultrasound emitter.

and detrimental GBs in FM–52M welds by introducing ultrasonic- Table 2

Welding parameters of the conventional welding without ultrasound, and of the
assisted arc welding.
U-GTAW.
The present paper studies the DDC characteristics in FM–52M,
under conditions of (1) conventional welding without ultrasound, Parameter
and (2) ultrasonic-assisted gas tungsten arc welding (U-GTAW). DC welding current 80/A
Two kinds of FM–52M bead-on-plate weldments were prepared Arc voltage 10.5/V
by U-GTAW and conventional welding without ultrasound, respec- Welding speed 0.5/mm s−1
Filler wire feed rate 7/mm s−1
tively. A number of characterisation techniques, including electron
Shielding gas Ar (99.995% in purity)
backscattered diffraction (EBSD), scanning electron microscopy Flowrate 10/L min−1
(SEM), optical microscopy (OM) and STF test, were adopted to Ultrasonic current amplitude in conventional welding 0/A
characterise microstructures and DDC features of the FM–52M Ultrasonic current frequency in conventional welding 0/kHz
Ultrasonic current amplitude in U-GTAW 10/A
weldments. A quantitative DDC criterion, based on the length of
Ultrasonic current frequency in U-GTAW 20/kHz
continuous detrimental GBs of ␥-matrix in FM–52M, is proposed
correspondingly to interpret the cracking. This allows for possible
manufacturing strategies for improving DDC resistance in FM–52M
via U-GTAW.

The ultrasonic generating system outputs high-frequency and

high-amplitude current in sine waveform. Such a function is
2. Materials and methods realised by the software LabVIEW in PC, the data acquisition card,
and the ultrasonic current power amplifier. Small sine control sig-
2.1. U-GTAW setup nal of frequency 20 kHz is generated by the data acquisition card.
This sine wave signal is then transformed to about 10 A current by
The U-GTAW system was designed to generate an ultrasonic the ultrasonic power amplifier. The DC GTAW power supply gener-
field during welding. The key body of the U-GTAW system includes ates direct current to drive the welding arc, while the sine current of
three parts, i.e., (1) the ultrasonic generating system, (2) the DC high-frequency, output by the ultrasonic generating system, is cou-
GTAW power supply, and (3) the ultrasonic wave monitors. A pled to the welding arc through an insulator inside the ultrasonic
detailed illustration of the experimental system is shown in Fig. 1. power amplifier.
242 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250

Fig. 2. The waveform of the current passing through the welding arc: (a) without ultrasound (0 kHz) and with (b) 20 kHz U-GTAW. The schematic in (b) illustrates that an
ultrasonic sine current was superimposed on the direct current, while the schematic in (a) demonstrates that there was no ultrasonic current through the arc.

The modulated sine current of 20 kHz makes the arc vibrate in ucts Company (Newton, North Carolina, USA; http://specialmetals.
ultrasonic frequency. Then an ultrasonic field is generated by the com/smw). The composition of the FM–52M is given in Table 1.
arc itself—a natural loudspeaker. The arc is just above the weld pool The U-GTAW system was utilised to manufacture bead-
and the ultrasonic field is superimposed upon the weld pool. on-plate welds on 304 stainless steel plates with FM–52M
The resultant ultrasonic field was measured by the ultrasonic wire of 0.9 mm in diameter. The interpass temperature was
wave monitors in two ways. The ultrasonic current waveform was controlled to be 50 ◦ C. Two groups of FM–52M overlays
acquired by Hall current sensor, and the ultrasound emission of with dimension 130 mm × 50 mm × 13 mm were prepared
the arc was detected by ultrasonic wave sensor. In addition, the under the conditions (1) without ultrasound (0 kHz) and
high-speed camera was utilised to monitor the arc, and it is demon- (2) with 20 kHz ultrasound (20 kHz U-GTAW), shown in
strated that the arc was vibrating in the frequency of the sine Fig. 2.
ultrasonic current. The ultrasound signal of the 20 kHz U-GTAW, collected by the
ultrasonic wave sensor shown in Fig. 3, reveals that the ultrasonic
field does exist physically in U-GTAW and leads to the weld pool
2.2. Bead-on-plate welding oscillation.
Detailed experimental parameters are summarised in Table 2.
The material under investigation was nickel-base filler wire,
FM–52M, which is produced by the Special Metals Welding Prod-
 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250 243

Fig. 3. Ultrasound detection of 20 kHz U-GTAW. (a) Sensor located near the welding arc, corresponding waveform shown in (b); (c) sensor attached to welding workpiece,
corresponding waveform shown in (d). CH1 is the control signal from the data acquisition card. CH2 is the ultrasound signal from the ultrasonic wave sensor.

Fig. 4. (a) Simulation curve for STF test, and (b) DDC measurement in an isothermal region of each STF specimen. Each crack length is denoted as li and the local strain of
each crack is denoted as εi .
244 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250

Fig. 5. EBSD characterisation of FM–52M overlays welded by conventional GTAW and U-GTAW of frequency 20 kHz. (a) 0 kHz (conventional welding without ultrasound),
longitudinal section; (b) 0 kHz (conventional welding without ultrasound), cross section; (c) 20 kHz, longitudinal section; and (d) 20 kHz, cross section.

2.3. Sample preparation and characterisation zones by transverse laser-marked lines. The area of each zone is
1 × 10 mm.
EBSD samples and dog-bone STF specimens, of dimension
70 mm × 15 mm × 2.5 mm, were taken from the cross section and
longitudinal section beneath the top of the FM–52M overlay, to 2.3.1. EBSD characterisation
avoid any dilution caused by the 304 stainless base metal. Both EBSD data sets were collected from a field emission scanning
samples were cut by electrodischarge machining. The surface for electron microscope NOVA 230 equipped with AZtec HKL Max
observation was ground to 2000 grit SiC paper and fine polished backscattered electron detector. The operating voltage was 20 kV.
using 4 and 2.5 ␮m diamond paste. The data were acquired from the Oxford Instruments AZtec HKL
The EBSD samples were finally electropolished in a solution of system, using a scan area of 2500 ␮m × 2500 ␮m. Post-analysis
20% perchloric acid and 80% alcohol, using DC 12 V at room tem- was conducted by the Oxford Instruments Tango software pack-
perature. The surface of STF specimen was finally divided into 20 age, where noise filtering was applied and the wild spikes were
removed. Both the equivalent diameter and the aspect ratio of the
 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250 245

Fig. 6. Diameter of grains in longitudinal section of FM–52M overlay.

austenitic grains in FM–52M overlay were measured by Image-Pro Table 3

Data for calculating ultrasonic pressure of U-GTAW and cavitation threshold in liquid
Plus 6.0 software package.
FM–52M.

Description Value/unit
2.3.2. Strain-to-fracture (STF) test
STF test was conducted at DSI Gleeble 3500 thermo-mechanical U Voltage output by ultrasonic current 43/V
power amplifier
tester, and the simulation curve is illustrated in Fig. 4a. During the
I Ultrasonic current measured by Hall 10/A
test, the heating chamber was in vacuum to prevent STF specimens current sensor
from oxidation. The specimen experienced an accurately controlled C Capacitance of the insulator 10 × 10−6 /F
heating cycle using thermocouples welded on the midpoint of the  Density of liquid FM–52M metal 8.2 × 103 /kg m−3
sample. Following the STF test, OM and SEM were employed to v Ultrasonic wave velocity in liquid 4.03 × 103 /m s−1*
FM–52M metal
record the morphology of each crack that falls in an isothermal
D Width of weld pool of FM–52M 6 × 10−3 / m
zone, where the thermocouples were fixed, shown in Fig. 4b. The overlays
isothermal zone is shown to be a 4 mm region symmetrical about P0 Atmospheric pressure 1.01 × 105 /Pa
the thermocouple spot, as described by Chen et al. (2014). Since Pv Vapour pressure of liquid FM–52M 297.792/Pa*
some cracks may be confused with GBs under OM observation, SEM metal
 Surface tension of liquid FM–52M 1.3/N m−1*
was employed to distinguish these unclear cracks. metal
Image-Pro Plus 6.0 software package was utilised to measure R0 Radius of a cavity in liquid FM–52M ∼10−6 /m**
the length of each crack and its local strain, and the weighting crack metal
length, wi , was calculated by Eq. (1). Data marked * are deduced from Iida and Guthrie (1988), and ** from Han (2015).

li max(ε1 , ε2 , ..., εn )
wi = (1)
εi
shown in Fig. 5a. Similarly, large columnar grains dominated the
Where li and εi are illustrated in Fig. 4b. cross section of the weld, seen in Fig. 5b
The equivalent maximal DDC length, lmax , was determined as Following ultrasound of frequency 20 kHz in U-GTAW, the size
the average of five largest weighting crack lengths in each STF of equiaxed grains in longitudinal section was decreased signif-
specimen. icantly, shown in Fig. 5c; some columnar grains in cross section
were fragmented and evolving into a more globular morphology,
seen in Fig. 5d.
3. Results

3.1. Grain morphology in the weld 3.1.2. Quantitative comparison

Figs. 6 and 7 quantitatively demonstrate the diameter and the
3.1.1. General characteristics aspect ratio of the austenitic grains in the prepared FM–52M over-
Inverse pole figure maps of two kinds of FM–52M overlays, lays. Compared with conventional GTAW without ultrasound, the
welded by U-GTAW of frequency 20 kHz and by conventional GTAW grain diameter in longitudinal section was decreased by ∼50.8%
without ultrasound respectively, are shown in Fig. 5. Both cross when the ultrasonic field of frequency 20 kHz in U-GTAW was
sections and longitudinal sections were characterised to reveal applied. The ultrasonic field played a significant role in refining
detailed features of the microstructure. longitudinal microstructure of FM–52M overlays.
Without ultrasonic field, the microstructure of longitudinal sec- As for cross section, similar effects of U-GTAW on grain mor-
tion consisted of coarse equiaxed austenitic grains predominantly, phologies can be seen in Fig. 7. The grain diameter was decreased
246 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250

Fig. 7. (a) Diameter, and (b) aspect ratio of grains in cross section of FM–52M overlays.

by ∼34.4% (Fig. 7a), and the aspect ratio of the grains was decreased mal grain boundaries by using OM only, because the greyscale of
by ∼19.3% (Fig. 7b). It indicates that the coupled ultrasound in U- these DDC get close to that of grain boundaries under OM observa-
GTAW has a distinguished effect on fracturing coarse columnar tion. Therefore, any doubtful dark curves observed in OM images
grains. were re-examined by SEM, which has higher resolution, to avoid
confusion, as demonstrated in Fig. 8c and d.
3.2. Alleviation of ductility-dip cracking Compared with conventional welding without ultra-
sound, the ultrasonic field in U-GTAW yielded a ∼40.7%
The equivalent maximal DDC length of STF specimens is shown decrement in maximal DDC length when the ten-
in Fig. 8, under both U-GTAW of frequency 20 kHz and conven- sile stroke of Gleeble tester is 0.25 mm, as shown in
tional welding without ultrasound. Typical OM images of DDC are Fig. 8a.
also shown in Fig. 8. Generally, the dark and wide curves repre- Similarly, the equivalent maximal DDC length was decreased by
sent DDC, while much shallower curves represent normal grain ∼37.1% using U-GTAW of frequency 20 kHz, when the tensile stroke
boundaries, as illustrated by the white circles in Fig. 8a and b. of Gleeble tester is 0.2 mm, as shown in Fig. 8b.
However, sometimes it is difficult to distinguish DDC from nor-
 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250 247

Fig. 8. Equivalent maximal DDC length of STF specimens under conventional welding without ultrasound and U-GTAW: (a) tensile stroke of 0.25 mm; (b) tensile stroke of
0.2 mm. SEM was utilised to differentiate DDC and GBs: (c) typical DDC morphology in FM–52M; (d) typical austenitic grain boundary in FM–52M.

4. Discussion tion will occur in weld pool. The method compares the ultrasonic
pressure with one critical value called the cavitation threshold.
4.1. On grain refinement of FM–52M overlays Provided that the ultrasonic pressure is higher than the cavitation
threshold, the ultrasonic cavitation will take place.
The presence of high-frequency sine wave in Fig. 2 and ultra- As reported by Li et al. (2008) and Liu et al. (2012), ultrasonic
sound signal in Fig. 3 indicates that an ultrasonic field, excited by pressure Pk is quantitatively dependent on both ultrasonic param-
the arc, was imposed on the molten pool of FM–52M weld, which is eters and liquid metal features, as shown in Eq. (2).
similar to the results reported by Da Cunha and Bohorquez (2015) 
and He et al. (2014). 2Wu v
Pk = (2)
Irradiated with ultrasound, the weld pool of FM–52M expe- S
rienced periodic expansive force and compressive force. If the
ultrasound energy is high enough, a cavity in the weld pool will be Where
created. Then the steady growth and collapse of cavities, referred to Wu is the ultrasonic power,
as cavitation, is likely to occur. However, it is difficult to observe the  is the density of liquid metal where ultrasonic wave propa-
cavitation phenomenon directly since the weld pool is not a trans- gates,
parent liquid and the temperature of the weld pool is higher than v is the ultrasonic wave velocity,
1300 ◦ C. Therefore, an indirect method reported by Kuijpers et al. S is the area of ultrasonic field projecting on liquid metal.
(2002) is adopted here to determine whether the ultrasonic cavita- Considering the features of U-GTAW in this study, the ultrasonic
power Wu can be determined by Eq. (3), and the area S by Eq. (4).
248 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250

Fig. 9. Maximal detrimental GB length of (a) longitudinal, and (b) cross sections in FM–52M weld.

Then Eq. (2) can be rearranged as Eq. (5), yielding the expression of and also break down coarse grains. These two factors account for
arc induced ultrasonic pressure, Pua . the grain refinement (Figs. 6 and 7) in FM–52M under U-GTAW of
20 kHz.
I
Wu = Re[(U − )I] (3)
j2fC
1 4.2. On DDC reduction
S= D2 (4)
4
 STF test results shown in Fig. 8 reveal that the equivalent max-
I
8Re[(U − j2fC
)I]v imal DDC length in FM–52M overlays was decreased by using
Pua = (5) U-GTAW. Before the effect of ultrasound on DDC reduction could be
D 2
discussed, however, it should be noted that DDC is an intergranular
Where each parameter and its value are described in Table 3, and j cracking reported by Collins et al. (2004). That is, the trail of each
is the imaginary unit. crack is extending along austenitic grain boundaries in FM–52M.
Thus, the ultrasonic pressure of U-GTAW can be calculated as Therefore, it can be deduced that maximal DDC length is related
Pua ≈ 2.6 × 107 Pa. to one eigenvalue of GB. Such an eigenvalue should be determined
Using the cavitation theory reported by Kuijpers et al. (2002), first.
the cavitation threshold for liquid FM–52M can be expressed by The intergranular cracking phenomenon suggests that a GB pos-
Eq. (6). sessing high energy is prone to cracking. According to the extended
 Read-Shockley model reported by Wolf (1989), GB energy gets its
4 2 maximum value when the misorientation of the GB is 45◦ , as shown
PB = P0 − Pv + 2
(6)
3 3R03 (P0 + − Pv ) in Eq. (7).
R0

Where all parameters and their values are shown in Table 3. E() = E0 sin(2){1 − r ln[sin(2)]} (7)
Thus, the cavitation threshold in liquid FM–52M metal can be
determined as PB ≈ 106 Pa. Obviously, Pua » PB , where Pua is the Where
ultrasonic pressure in the U-GTAW. As a result, the ultrasonic field E is the GB energy,
in this study is capable to induce cavitation in FM–52M weld pool.  is the misorientation of a GB,
During cavitation, the alternative growth and collapse of cavi- E0 and r are constants.
ties provides a unique physical circumstance for grain refinement. Thereby, DDC is most likely to occur along GBs of misorienta-
Firstly, cavity collapse will produce localised high pressure in tion 45◦ , theoretically. This theoretical deduction was verified by
very short lifetime. These transient hot spots have pressure over Chen et al. (2014) and Chen et al. (2013)’s investigations on nickel-
108 –109 Pa, reported by Bang and Suslick (2010) and Zhai et al. base alloy. It was further revealed that, during STF tests, the GBs,
(2015), and will increase the melting point of FM–52M. Therefore, with misorientation ranging from 30 to 45◦ , are easy to suffer from
a melt under cooling will occur near these spots, and will promote DDC. Those GBs with such misorientation are detrimental to DDC
homogenous nucleation in FM–52M welds during U-GTAW. Sec- prevention.
ondly, the violent nature of cavitation will enhance the flow of Clearly, the connection of these detrimental GBs makes different
the liquid metal (Shu et al., 2012). The weld pool will be stirred DDC trails join together easily and causes longer cracks. Therefore,
dramatically, increasing the wettability between impurity particles the maximal length of connected detrimental GBs of misorientation
and liquid metal. This acts to enhance heterogeneous nucleation, 30–45◦ , referred to as detrimental GB length, was determined to
 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250 249

be the GB eigenvalue that influences the equivalent maximal DDC

length.
The detrimental GB length evaluated by EBSD is shown in Fig. 9.
The related GB maps are reconstructed from Fig. 5. The GB maps
indicate that the morphology of detrimental GBs was much more
dispersive after U-GTAW being applied. A quantitative analysis
demonstrates that detrimental GB length was decreased by ∼41.2%
in longitudinal section of FM–52M overlays, and was decreased
by ∼60.1% in cross section with the help of U-GTAW. The reduc-
tion of detrimental GB length arises from the grain refinement in
FM–52M weld, and the grain refinement is induced by the ultra-
sonic cavitation described in 4.1. Due to the cavitation effect, the
FM–52M molten pool was stirred by ultrasonic wave, and the weld
microstructure becomes more homogeneous. In other words, the
dispersion degree, among the GBs misorientations in the weld, is
decreased when U-GTAW was applied. This proposal can be val-
idated by analysing standard deviation Spgb , as shown in Fig. 10.
Spgb is related to GBs features in FM–52M longitudinal section and
cross section, and is defined by Eq. (8) that quantitatively assesses Fig. 10. Standard deviation of GB misorientation distribution.
the dispersion of GB misorientation.

 2
 1 
n
1
n needs extra energy to transfer from one connected detrimental GB
Spgb =  (pi − pj ) (8) to another (see Fig. 11b), thus leading to a suppressed maximal DDC
n−1 n length in FM–52M overlays. Under conventional GTAW without
i=1 j=1
ultrasound, however, the detrimental GB length is easy to combine
Where with one another. In this case, cracks will propagate through longer
n is the number of GB misorientation data acquired by EBSD and connected detrimental GBs (see Fig. 11a) more easily, yielding
system, a longer maximal DDC length.
pi and pj are the occurrence probability of each GB misorienta-
tion.
5. Conclusions
It is evident that the standard deviation under U-GTAW of fre-
quency 20 kHz is smaller than that under conventional welding
Ultrasonic field has been introduced in the welding process of
without ultrasound. Thus, it has been proven that the ultrasound
nickel-base FM–52M to alleviate ductility-dip cracking. The cause
in U-GTAW homogenised the distribution of GB misorientations in
of cracking in FM–52M has been studied in a quantitative way
FM–52M.
based on grain boundary features. The following conclusions can
Thereby, the effect of U-GTAW on DDC reduction can be clari-
be drawn from this work.
fied by means of detrimental GB length. Due to the grain refinement
in the weld, detrimental GBs of misorientation 30–45◦ will not
gather together at one localised region and will not connect with 1. A concept of detrimental grain boundary length has been pro-
each other seriously either. Therefore, the detrimental GB length posed to interpret quantitatively the prevention mechanism of
in FM–52M overlays was decreased under U-GTAW. This decre- ductility-dip cracking in FM–52M. It was revealed that a longer
ment obstructed the connection and extension of DDC. The DDC detrimental grain boundary length aggravates the extension of

Fig. 11. Schematic of the influence of (a) conventional welding without ultrasound, and the influence of (b) ultrasonic-assisted GTAW on the reduction of DDC susceptibility.
250 C. Hua et al. / Journal of Materials Processing Technology 239 (2017) 240–250

cracks, thus making the weldment susceptible to ductility-dip Da Cunha, T.V., Bohorquez, C.E.N., 2015. Ultrasound in arc welding: a review.
cracking. Ultrasonics 56, 201–209.
Dai, W.L., 2003. Effects of high-intensity ultrasonic-wave emission on the
2. Ultrasonic field, of frequency 20 kHz, was successfully superim- weldability of aluminum alloy 7075-T6. Mater. Lett. 57, 2447–2454.
posed on the gas tungsten bead-on-plate welding of FM–52M by Han, Q.Y., 2015. Ultrasonic processing of materials. Metall. Mater. Trans. B: Proc.
modulating high frequency current upon the welding arc. The Metall. Mater. Proc. Sci. 46, 1603–1614.
He, L.B., Yang, P., Li, L.M., Wu, M.S., 2014. The ultrasonic characteristics of high
consequent ultrasonic effect has induced grain refinement and frequency modulated arc and its application in material processing.
grain fragmentation in FM–52M welds significantly. Ultrasonics 54, 2178–2183.
3. The violent stirring of weld pool, caused by ultrasonic cavitation, Iida, T., Guthrie, R.I.L., 1988. The Physical Properties of Liquid Metals. Oxford
University Press, New York.
has made the FM–52M weld more homogenous. Detrimental
Kuijpers, M.W.A., van Eck, D., Kemmere, M.F., Keurentjes, J.T.F., 2002.
grain boundaries were then distributed as a dispersive mor- Cavitation-induced reactions in high-pressure carbon dioxide. Science 298,
phology but not gathered together, which prevents different 1969–1971.
Li, J.W., Momono, T., Tayu, Y., Fu, Y., 2008. Application of ultrasonic treating to
detrimental grain boundaries from connecting with each other.
degassing of metal ingots. Mater. Lett. 62, 4152–4154.
Thus, the detrimental grain boundary length was decreased Lippold, J.C., Nissley, N.E., 2008. Ductility-Dip Cracking in High Chromium, Ni-Base
by ∼41.2% and the susceptibility of ductility-dip cracking in Filler Metals. Springer-Verlag Berlin, Berlin.
FM–52M was lowered. Liu, Z.W., Han, Q.Y., Li, J.G., Huang, W.D., 2012. Effect of ultrasonic vibration on
microstructural evolution of the reinforcements and degassing of in situ
TiB2p/Al-12Si-4Cu composites. J. Mater. Process. Technol. 212, 365–371.
Acknowledgement Mo, W.L., Hu, X.B., Lu, S.P., Li, D.Z., Li, Y.Y., 2015. Effects of boron on the
microstructure, ductility-dip-cracking, and tensile properties for NiCrFe-7
weld metal. J. Mater. Sci. Technol. 31, 1258–1267.
This work was supported by the National Natural Science Foun- Morisada, Y., Fujii, H., Inagaki, F., Kamai, M., 2013. Development of high frequency
dation of China (Grant Nos. 51575347 and 51405297). tungsten inert gas welding method. Mater. Des. 44, 12–16.
Nissley, N.E., Lippold, J.C., 2008. Ductility-dip cracking susceptibility of
nickel-based weld metals part 1: strain-to-fracture testing. Weld. J. 87,
Appendix A. Supplementary data 257S–264S.
Noecker, F.F., DuPont, J.N., 2009. Metallurgical investigation into ductility dip
Supplementary data associated with this article can be found, in cracking in Ni-based alloys: part I. Weld. J. 88, 7S–20S.
Ramirez, A.J., Sowards, J.W., Lippold, J.C., 2006. Improving the ductility-dip
the online version, at http://dx.doi.org/10.1016/j.jmatprotec.2016. cracking resistance of Ni-base alloys. J. Mater. Process. Technol. 179, 212–218.
08.018. Shu, D., Sun, B.D., Mi, J.W., Grant, P.S., 2012. A high-speed imaging and modeling
study of dendrite fragmentation caused by ultrasonic cavitation. Metall. Mater.
Trans. A: Phys. Metall. Mater. Sci. 43A, 3755–3766.
References Watanabe, T., Shiroki, M., Yanagisawa, A., Sasaki, T., 2010. Improvement of
mechanical properties of ferritic stainless steel weld metal by ultrasonic
Bang, J.H., Suslick, K.S., 2010. Applications of ultrasound to the synthesis of vibration. J. Mater. Process. Technol. 210, 1646–1651.
nanostructured materials. Adv. Mater. 22, 1039–1059. Wolf, D., 1989. A read-shockley model for high-angle grain boundaries. Scr. Mater.
Chen, J.Q., Lu, H., Yu, C., Chen, J.M., Zhang, M.L., 2013. Ductility dip cracking 23, 1713–1718.
mechanism of Ni-Cr-Fe alloy based on grain boundary energy. Sci. Technol. Zhai, W., Hong, Z.Y., Wen, X.L., Geng, D.L., Wei, B., 2015. Microstructural
Weld. Join. 18, 346–353. characteristics and mechanical properties of peritectic Cu-Sn alloy solidified
Chen, J.Q., Lu, H., Cui, W., Chen, J.M., Huang, Y.F., 2014. Effect of grain boundary within ultrasonic field. Mater. Des. 72, 43–50.
behaviour on ductility dip cracking mechanism. Mater. Sci. Technol. 30,
1189–1196.
Collins, M.G., Ramirez, A.J., Lippold, J.C., 2004. An investigation of ductility-dip
cracking in nickel-based weld metals—part III. Weld. J. 83, 39S–49S.