Journal of Materials Processing Technology 75 (1998) 259 – 265

Ultrasonic welding of ceramics/metals using inserts

Shin-ichi Matsuoka 1
Department of Mechanical Systems Engineering, Toyama Prefectural Uni6ersity, Kosugi-machi, Toyama 939 -03, Japan

Received 22 January 1997

Abstract

This paper gives a description of an experimental study of the ultrasonic welding of ceramics and metals using inserts.
Ultrasonic welding has made it possible to weld various ceramics such as Al2O3, SiC, Si3N4, AlN, to metals at room temperature,
quickly and easily compared to other welding methods. For example, the ultrasonic welding of Si3N4/Al/Cu can be accomplished
under the conditions of: amplitude of ultrasonic horn top, 23 mm; welding pressure, 20 MPa; and required duration, 1 s. The
ultrasonic-welding technique can be carried out in many different ways, such as the direct welding of ceramics/metals, welding
with a metal coating on the ceramics surface by vacuum deposition and welding with a metal insert. The insert material and the
vaporized film used for auxiliary purposes play the role of binders to facilitate welding, but little difference in welding strength
was found between the welds with the metal insert and the vaporized film. Since ultrasonic vibration cleans the contact surface,
there is no need to carry out surface treatment prior to welding. © 1998 Elsevier Science S.A.

Keywords: Welding; Metal insert; Vaporized film

1. Introduction scopic observations of such joints and studies of the

structure of the interface between the materials and
When metal and ceramics are welded, the welding reports on the effects, properties, etc. when using this
characteristics are influenced to a great extent by the method.
boundary properties. The results of previous attempts
to perform ultrasonic welding of various kinds of ce-
ramics and metals have confirmed that there are opti- 2. Welding methods and conditions
mum welding conditions (ultrasonic amplitude, welding
pressure, required duration, etc.) for each ceramic/ Fig. 1 shows ultrasonic-welding equipment together
metal combination and that when welding is done with an outline of a joint. This welding method includes
under conditions that conform closely to these opti-
mum conditions, it is possible to greatly improve the
welding strength and other properties of the joint [1–3].
With this method, it is possible to easily perform weld-
ing in all environments, either in the atmosphere or in
a vacuum for example and it appears that it will be
possible to apply the method to ceramics of all kinds.
However, there are numerous materials and combina-
tions of materials that have to be joined, and it has not
been confirmed that all of these can indeed be joined,
so naturally it is important to develop welding technol-
ogy and adhesive agents to join these materials.
This paper discusses ultrasonic welding using a
binder as one way to promote joining, describes micro-
1
Fax: +81 766 568029. Fig. 1. Basic principle of ultrasonic welding.

0924-0136/98/$19.00 © 1998 Elsevier Science S.A. All rights reserved.

PII S 0 9 2 4 - 0 1 3 6 ( 9 7 ) 0 0 3 7 2 - 5
260 S. Matsuoka / Journal of Materials Processing Technology 75 (1998) 259–265

2. Welding methods and conditions

Fig. 1 shows ultrasonic-welding equipment together

with an outline of a joint. This welding method includes
one version in which the ceramics and metal material
are joined directly and another method using a binder,
and this paper investigates the latter method. One
binder-based technique involves first vacuum depositing
metal onto the surface of the ceramic, as on the model
shown in Fig. 2, putting the metal on this surface and
applying pressure to the metal and this surface, then
applying ultrasonic vibration (Fig. 2(a)), whilst another
technique involves the joining of low-melting-point
metal as an insert material (Fig. 2(b)).
In this work, for both of the cases shown in Fig. 2,
the output of the welding equipment is 600 W, the
amplitude of ultrasonic vibration a = 23–35 mm, the
welding pressure Pc = 5–40 MPa, and the required
duration T= 0.1–5 s. The ceramic used for the welding
is an atmospheric sintered body containing oxides
Al2O3, ZrO2, nitrides Si3N4, AlN and carbide SiC, their
surface roughness having an average of 1 mm.
Al(Al100) and Cu(C1020) 0.5 mm thick, 0.2 mm thick

Fig. 2. Model of an ultrasonically-welded interface between metal

sheet and ceramic. (a) Effect of coating metal on a ceramic surface;
(b) effect of insertion.

Fig. 4. Scanning electrons micrographs showing the internal structure

of ultrasonically-welded products, where the coating is: (a) copper on
Fig. 3. Welding conditions and bonding region. SiO2; and (b) nickel on SiC; by vacuum deposition.
 S. Matsuoka / Journal of Materials Processing Technology 75 (1998) 259–265 261

Fig. 5. Scanning electron microscope photographs showing the internal structure, and line analysis by electron probe X-ray microanalyser, for
ultrasonically-welded products: (a) Si3N4/Cu/Al; and (b) AlN/In/Ni.

Ni sheet metal, 50 mm thick rolled-copper foil and cantly with the combination of ceramics and metal,
indium sheet, were used as inserts. Visual and structural since the conditions also change according to the com-
observations were made: (i) tension tests, to compare bination. An example is shown in Fig. 3, which shows
the welding strength; and (ii) by electron probe X-ray the welding strength in cases where various combina-
microanalyzer (EPMA) structural analysis and scan- tions of materials were welded at an amplitude of 23
ning electron microscope (SEM) observation of the mm, the lines under the combined materials in the figure
joint; to evaluate the characteristics of the 2× 2 mm2 indicating insert materials formed by first vacuum de-
contact area of the bonded material. positing a metal onto the ceramics surface. Properly-
joined material was formed with a welding pressure of
10–40 MPa and a duration of 0.2–2 s for all combina-
3. Welding conditions and the bonding region tions of materials. When the welding pressure is in-
creased overall, the required duration is reduced and
The welding strength and characteristics vary signifi- inversely when the welding pressure is low, the required

Fig. 6. Surface analysis by electron probe X-ray microanalyser for ultrasonically-welded products: (a) Si3N4/Cu/Al and (b) AlN/In/Cu.
262 S. Matsuoka / Journal of Materials Processing Technology 75 (1998) 259–265

Fig. 6. (Continued)

duration must be increased by a few minutes. Within quired duration fall below the levels in the figure, the
the scope of these tests, little difference was observed ultrasonic vibration stops or it can not be used effec-
between cases where a deposited membrane was used tively, so that it is impossible to form a good joint.
() and cases where insert materials were used (, ).
The testing also revealed that the joining process is
completed in a shorter period of time using Al or In as 4. SEM joint observation
the binder material than when Cu is used for this
purpose. This is assumed to be caused by the wettabil- Fig. 4 shows an example of SEM joint observation
ity, deformation performance, and fusion point of the of: (a) SiO2/Cu/Al; and (b) SiC/Ni/Al. In these cases,
materials. Also, when the welding pressure or the re- the Cu and Ni were vacuum deposited on the SiO2 and
 S. Matsuoka / Journal of Materials Processing Technology 75 (1998) 259–265 263

SiC surfaces, respectively, then aluminium foil was 5. Properties and structure of the joint boundary
welded ultrasonically to the surfaces. The results
were good-quality joint boundaries, without any It has already been clearly shown that direct join-
trace of separation or voids of the insert mater- ing forms a reaction layer [1,2]. The boundary struc-
ial, indicating that this insert material is a useful bind- ture was thus investigated, particularly the joint
ing agent that effectively promotes joining. The same material including the insert material. Fig. 5 shows
held true for other joining materials: SiO2/Ni/Al; the results of element distribution (line analysis by
Al2O3/Cu/Al; SiC/Ti/Al; Si3N4/Al/Cu; and others. Fig. EPMA) at the joint boundaries of Si3N4/Cu/Al and
5 shows an example of SEM observations of joints AlN/In/Ni, with excellent boundaries being formed
formed by: (a) Si3N4/Cu/Al; and (b) AlN/In/Ni. In on the joint in both cases. A change of width of the
these cases, Cu and In were used as insert material. The transition layer of 1–2 mm was found, from which it
welding conditions in both cases were an ultrasonic was concluded that a very slight reaction layer or
vibration amplitude of 23 mm, a duration of 0.5 s, and diffusion layer was formed.
a welding pressure of 20 MPa. Judging from the SEM That no noticeable difference was observed in
observations, the boundaries of all of the joints were the joint immediately after joining and after it had
good, these combinations of materials being able to be been annealed for 2 h at a temperature of 550°C
joined using ultrasonic welding. These observations suggests that even if the dispersion mechanism
played a role, it was within an extremely narrow
demonstrate that the vacuum-deposited membranes
range.
and insert materials used to promote joining act as
Fig. 6 shows the results of EPMA-based surface
effective binders and are extremely useful in this pro-
analysis of the chemical elements of the joint materi-
cess.
als: (a) Si3N4/Cu/Al; and (b) AlN/In/Cu. Fig. 6(a)
shows that the Cu adhesioned on the Si3N4 side. Al-
though the welding strength was  35 MPa (max) as
shown in joint combination (D) in Fig. 7, which is a
little less than the shear strength of both base
metals, separation occurred along the joint boundary.
Also, the EPMA image in Fig. 6(b) shows that be-
cause the indium fused and scattered during joining,
it became part of the joint structure and remained
dispersed, its thickness reduced to  1/10, showing
that it functioned as an effective insert material. Sim-
ilar tendencies were observed in specimens where a
vacuum-deposited membrane was used.

6. Strength of the joint material and cleft surfaces

Microscopic observations with an SEM revealed

good joints, but the welding strength is still a serious
problem hampering the practical application of the
method. Fig. 7 shows an example of the welding
strength obtained from the shear tensile testing of
various joint materials. Almost no elongation was
found in any of the joint material, confirming that,
overall, either fracture or shearing occurred at the
joint boundary, or breakage occurred in the base-
material part of the ductile material. This finding
agrees with an earlier report [1,2] of the results in
the case of joint material formed by direct joining. It
also demonstrated that the use of an insert material
or a vacuum-deposited membrane promoted joining,
Fig. 7. Welding strength sw for various ultrasonically-welded prod- but there was no clear difference in welding strength.
ucts (required duration: 2–4 s; welding pressure: 10–30 MPa); (metal- In any case, it is assumed that a connection reaction
lic thick film by vacuum evaporation). (reaction between atoms) occurs, in addition to
264 S. Matsuoka / Journal of Materials Processing Technology 75 (1998) 259–265

Fig. 8. Fracture surfaces of ultrasonically-welded products of SiC/Al/Cu observed by scanning electron microscope.

mechanical bonding, regardless of the materials. When 7. Estimation of the temperature of the joint boundary
two clean surfaces make contact, plastic deformation
and elastic deformation bring them close together, A study was performed to determine how much
causing the atoms to attract to one another. The heat is generated at the contact boundary by the re-
boundary temperature-rise due to vibration should action described in the previous section. Although it
bring about the formation of linkage between atoms, is difficult to directly measure the temperature gener-
or a reaction layer. The analytical results in Figs. 5 ated between the ceramics and the metal, as one
and 6 support this view. These results show that it is yardstick in this study, the temperature of the
impossible to deny that the quality, composition, etc.
boundary obtained by converting the electromotive
of the joint materials, and the ultrasonic conditions,
force generated when joining: (a) copper/constantan;
caused the differences.
and (b) Cu/Al was recorded as shown in Fig. 9. The
Fig. 8 is an example of a photograph obtained
results show that in both cases, when the welding
during a SEM observation of a fracture in SiC/Al/
Cu joint material after tensile testing, showing the pressure is 50 MPa and the required duration is 0.7
adhesion of Al and Cu deposited on to the surface s, the boundary temperature rose to 500°C, then
of the SiC. It also confirms that there was an adhe- later stabilized. Little difference was observed when
sion of Al or Si deposited on to the Cu surface, the welding pressure Pc ranged from 30 to 50 MPa.
suggesting that its welding strength is high. In this This means that it is not possible to simply make a
way, at the contact boundary, powerful ultrasonic vi- conversion, but it is hypothesized that the tempera-
bration caused adhesion, with the result that some ture also rises this much or even more at a ceramics/
degree of welding strength was obtained, the same metal boundary. In other words, if frictional heating
being true for direct joints [1]. is generated at the contact surface, increasing the
 S. Matsuoka / Journal of Materials Processing Technology 75 (1998) 259–265 265

temperature, this will without a doubt cause a binding

reaction if only to a small degree, which suggests that
there is good reason to count on the bonding reaction
at the contact surface, described in Section 5 and
Section 6.

8. Conclusions

This study has examined the effects of a binder in

the ultrasonic welding of ceramics and metal. The
results have revealed that by using an activated
metal or low-melting-point metal as an insert mate-
rial or by metallizing (vacuum-deposition) the con-
tact surface, it is possible to obtain good bonding
effects. It also indicated that these encourage joining
even with Cu, which has been difficult to join in the
past. The ultrasonic-welding method can be used to
perform joining easily and quickly at room tempera-
ture, and is fully capable of joining dissimilar mate-
rials.

Acknowledgements

The author wishes to thank the AMADA Foun-

dation for Metal Work Technology for their sup-
port.

References

[1] S. Matsuoka, Trans. J. Soc. Mech. Eng. 55 (517) (1986) 2481.

Fig. 9. Measured temperature of the welding interface in ultrasonic [2] S. Matsuoka, J. Mater. Process. Tech. 47 (1994) 185.
welding: (a) Cu – C, Pc = 50 MPa; and (b) Cu–Al, Pc = 50 MPa. [3] Japanese Patent No. 1607227, 1991.