Development of New Active Filler Metals
in a Ag-Cu-Hf System
Hafnium-containing active filler metals show promise for brazing
nonoxide ceramics to themselves and to steel
BY E. LUGSCHEIDER AND W, TILLMANN
ABSTRACT. Due to the increasing applications of nonoxide ceramics in technical
structures, new active filler metals have to
be developed for joining ceramics either
to other ceramics or to metals. In commercially dominant active braze alloys, titanium as the reactive agent is added to
the base filler metal, particularly to the
Ag-Cu eutectic. Apart from titanium and
other reactive elements, hafnium additives are also known for promoting the
wetting of ceramics by conventional
braze metals.
This investigation deals with hafniumadded Ag-Cu active filler alloys and the
examination of their brazing properties
for joining SiC and Si3iN4 ceramics to
themselves and to steel. Therefore, active
filler metals in a system of Ag-Cu-(ln)-Hf,
with hafnium contents ranging from 2 to
5 wt-% and an eutectic Ag-Cu composition, have been fabricated by arc melting.
Differential thermoanalysis showed that
the alloys have a melting behavior similar
to titanium-containing Ag-Cu alloys. Wettability tests conducted on SiC and Si3N4
proved that they exhibit good wetting
properties at elevated temperatures
(>1000°C = 1273 K). Metallographic observations showed that a typical reaction
layer on the ceramic surface was formed,
which was enriched in hafnium. This indicates that the desired reaction between
hafnium and Si3N4 or SiC has taken place.
Correlating thermodynamic calculations
of possible reactions between hafnium
and the ceramics suggests that formation
of hafnium nitrides or carbides will occur.
Silicide reactions are not very probable
due to low brazing temperatures.
Brazing test joints at different brazing
conditions showed that hafnium-containing filler metals are suitable for joining
E. LUGSCHEIDER and W. TILLMANN are with
Aachen University of Technology, Aachen,
West Germany.
Paper presented at the 21st A WS International
Brazing and Soldering Conference, held April
23-26, 1990, in Anaheim, Calif.
416-s | NOVEMBER 1990
SiC/Si3N4 ceramics to themselves, as well
as for joining them to steel. The quality of
the joint strongly depends on the quality
of the vacuum because of the high reactivity of hafnium, even to traces of atmospheric gases.
Regarding the mechanical properties of
joints brazed with haf nium-conl aining filler
metals, four-point bending test specimens
have been fabricated. The results gained
in the tests indicated that a brazing temperature above 1000°C (127.3 K) is essential for a strong joint w i n bending
strengths above 149 MPa (21.6 ksi) in the
case of Si3lM4. SiC ceramics display weaker
mechanical properties due to their thicker
reaction zone.
Introduction
Structures that require exceptional
properties, especially heat or corrosion
resistance, and hardness or wear characteristics, offer a high potential lor the application of ceramic materials. In order to
exploit this potential sufficiency, strong
and reliable joining technologies have to
be provided. Apart from joining ceramics
to themselves, the question of joining ceramics to metals is of particula' interest.
Active brazing is a new techrology that
offers very promising opportunities to
braze ceramic materials either to themselves or to metallic base materials. In
contrast to techniques that are employed
KEY W O R D S
Hf/AgCu Filler Metal
Active Filler Metals
SiC/Steel Brazing
Si3iN4/SteeI Brazing
SiC/Si3N4 Brazing
Reaction Layers
SiC/Si 3 N 4 Wettability
High-Temp. Wetting
Vacuum Brazing
Joint Bend Strength:;
when ceramics are brazed using nonreactive filler metals, active brazing is very easy
to carry out.
Among other elements of the IVA and
VA groups of the periodic table of the elements, hafnium is well known for its high
reactivity. Thus, an application of hafnium
as a reactive agent within filler metals for
ceramic brazing should be possible.
Thermodynamic calculations can give a
first evaluation of the reactivity of hafnium-ceramic reactions. The joints were
brazed under vacuum conditions, and the
braze joint was characterized and analyzed by electron- and light-microscopic
methods. Bending tests were used to
evaluate the mechanical strength of the
joints.
Silicon nitride and silicon carbide are the
ceramic base materials, and 5X18Cr10Ni
(AISI 304) and 5X29Ni18Co are the metallic base materials investigated. The filler
metals were fabricated in the Ag-Cu-(ln)Hf system. Thermodynamic analyses of
the reactions between hafnium and silicon
nitride and silicon carbide, respectively,
were conducted.
The atomic structure of ceramic materials is determined by strong covalent or
ionic interatomic bonds. Due to their different atomic structure, metals are not
able to wet ceramics directly without developing a new interface. However, interaction between ceramics with liquid
metals is only possible as long as there is
a partial or complete dissociation of the
interatomic bonding in the solid ceramic
body (Ref. 9). Hence, a major requirement
for active or direct brazing of ceramics is
that the active agent within the filler metal
is able to change the chemistry and the
energy of the ceramic-metal interface. By
doing this, a wetting and bonding reaction
can occur. Apart from chemical reactions
at the interface liquid metal-ceramic, there
is a strong dependence of the degree of
wettability on the brazing temperature.
After a chemical dissociation reaction
between the active filler metal has taken
place, a layer of reaction products is
formed that has a small free energy of
formation accompanied by a higher electrical conductivity (Ref. 10). In general,
these metal-like compounds that are
formed by chemcial reactions are wetted
better than covalent or ionic bonded
compounds. Carbides and nitrides of transition metals belong to that class of metallike wettable compounds (Ref. 9). By adding highly reactive transition metals to a
conventional filler metal, a wetting and
bonding reaction between the ceramic
and the liquid metal can be achieved. Ideally, a brazing alloy should have a low
content but high activity of the reactive
agent because otherwise an excessive reaction can occur accompanied by the
formation of a thick fragile reaction layer
at the ceramic surface (Ref. 10).
Whereas the brazing temperature at
which a sufficient degree of wetting takes
place can only be determined empirically,
thermodynamics offer some important
indications on the probability of possible
reactions between active metals and ceramic base materials.
As mentioned above, the wetting and
brazing process of active filler metals on
ceramics strongly depends on chemical
reactions between the reactive agent of
the braze and the ceramic itself. In order
to achieve a better understanding of this
process, possible reactions have to be
compiled. By applying the Gibbs equation
for the Gibbs free energy, C = H — TS, a
thermodynamic analysis of possible reactions can be achieved, presupposing a
thermodynamic equilibrium. Due to the
lack of data concerning the thermodynamic activity of the reaction system, a
complete analysis could not be carried
out. This means that the influence of the
Hf concentration could not yet be taken
into consideration.
Depending on the ceramic to be brazed,
t w o basic reactions systems occur apart
from various possible silicide reactions:
Hf/Ti/Zr + SiC
TiC/ZrC/HfC + Si
<= =>
Hf/Ti/Zr + '/4Si3N4 < = = >
HflM/TiN/ZrN + %Si
(1)
(2)
Using a computer program to calculate
the values of the Gibbs free energy of reactions (1) and (2) lead to the results presented in Fig. 1. The values have been calculated on the basis of data published in
the literature (Refs. 1-3, 11). In order to
enhance the comparability, the data are
plotted per mole active metal.
Comparing these curves with each
other leads to the conclusion that the formation of HfC and HfN is very probable.
The G values for Zr- and Ti-Si3iN4/SiC reactions lie above those calculated for the
corresponding hafnium reaction. Thus, for
all reaction systems, HfC and HfN are the
most stable reaction products, and this indicates that hafnium additions to conven-
Gibbs free energy IkJ/mol]
100
TiC
120
140
-160
180
-200
0
200
400
Curves plotted per mole of active metal
i
i
i
1
1200
1400
600
800
1000
1600
Temperature [K]
Chemical Reactions:
SiC * x -> xc • Si
1/4Si3N4 • X -> XN * 3/4Si
X • Hf/Zr/Ti
Fig. 1 - Diagram of Gibbs free energy for different reactions between active components and SiC
and SijNj.
tional filler metals should promote a wetting and bonding reaction.
Thermodynamic data for hafnium silicides Hf5Si3, HfSi and HfSi2 were not completely available; therefore, thermodynamic evaluation of these reaction products could not be carried out. But by
comparing the standard enthalpy of formation as published (Ref. 1), it can be
concluded that their formation is not very
probable. Furthermore, other investigations (Refs. 4, 5) showed that the formation of hafnium silicides could first be observed at temperatures above 1200°C
(2192°F), where the amount of silicides is
much lower than that of hafnium carbide.
For the corresponding titanium and zirconium reactions, the necessary thermodynamic data are available. But similar to
hafnium reactions, the G values of the
corresponding carbides and nitrides are
much lower than those of the silicides.
As already mentioned, the calculation
of the Gibbs free energy does not take
into consideration the necessary activation energies. From the thermodynamic
point of view, hafnium-containing filler
metals are supposed to exhibit a good
wetting and bonding behavior, as corresponding experiments could prove.
Experimental Procedure
In order to examine the technical benefit of hafnium as a reactive agent in active
brazes, filler metals in the systems Cu-Hf
(10, 15 wt-% Hf); Ag-Hf (3, 4 wt-% Hf);
Ag-Cu-Hf (in the range 1-5 wt-% Hf,
Ag-Cu: eutectic); and Ag-Cu-ln-Hf (70,19,
5, 3 wt-%) have been fabricated in an arc
furnace under shielding gas. This method
guarantees that hafnium does not react
with other elements due to the short
melting time. By using other melting techniques, such as induction heating or melting in a vacuum furnace, sufficient results
could not be achieved. Hafnium either reacted with the crucible materials or with
atmospheric gases. After the alloy was
fabricated, it was rolled into foil.
Chemical analysis of selected alloys by
atomic absorption spectroscopy proved
that the desired composition could be
maintained after melting within a limit of
± 0 . 2 wt-%. X-ray diffraction analysis of
the foils showed that part of the hafnium
also reacted with copper under formation
of copper-hafnium phases. The following
phases have been detected by x-ray diffraction analysis of an Ag-Cu-4Hf and AgCu-5Hf alloy: Ag, Cu, Hf; H f 0 2 , C u 2 0 ,
(Cu 4 0 3 ); Cu8Hf3 (nonequilibrium phase
according to ASTM data sheets); Cu7Hf2.
The melting behavior of the brazes was
characterized by differential thermoanalysis. As expected, the melting range of the
Ag-Cu-Hf alloys corresponded in general
with Ag-Cu eutectic, namely 837°-927°C
(1010-1100 K). The melting ranges of the
t w o component systems are in accordance with the Cu-Hf phase diagram or
with the melting data of silver. Indium
lowered the melting range of the Ag-CuHf systems, which was expected from the
corresponding Ti-brazing alloy (Ref. 7). In
general, hafnium has no significant influence on the melting range of the braze.
Brazing Process and Metallurgical Reactions
As expected from thermodynamic calculations, hafnium reacts with the ceramics and thus enables the filler metal to wet
the Si3N4 as well as SiC ceramics. Referring
to the structure of the brazing zone and
the strengths measured in the four-point
WELDING RESEARCH SUPPLEMENT 1417-s
(7.5 X 10 7 torr) at temperatures between
950°-1050°C (1223-1323 K). Prestudies
showed that these elevated brazing temperatures are necessary for a good and
reliable bonding reaction. At these temperatures, the matrix of the filler metal is
molten and hafnium diffuses to the ceramic surface in order to react with the
ceramic. Besides, a vacuum of this quality
is necessary in order to achieve good
bonding results. Otherwise, part of the
hafnium would have reacted with atmospheric gases.
Fig. 2 — Cross-section of an active brazed
Si3N4/Si3N4joint. Filler metal: 90Cul0Hf(240X).
bending tests, alloys in the Ag-Cu-Hf system seem to be the most promising. The
wetting behavior of the Cu-Hf or Ag-CuIn-Hf alloys was sufficient; whereas, the
Ag-Hf, in contrast to Ag-Ti, does not
exhibit a very good wetting behavior on
both ceramics.
The brazing process itself was conducted under a vacuum of 10~ 6 mbar
In Fig. 2, the brazing of a Si3N4-Si3N4
joint with a Cu-Hf alloy is presented. A
very thin reaction zone can be detected
that mainly consists of hafnium. This reaction zone had an average thickness of only
1 nm as measurements at higher magnification showed. Dark phases within the
braze are pure Cu-phases that are embedded in a Cu-Hf matrix. Despite the high
reactivity of hafnium with silicon nitride, a
certain amount of hafnium is still found in
the center of the brazed joint
Figure 3 demonstrates an active SiC
joint brazed with a Ag-Cu-Hf alloy. The
element distribution presented helps to
identify the different phases within the
joint. The reaction zone mainly consists of
hafnium and in contrast to Fig. 2, it is much
wider. Apart from hafnium, a small amount
of copper could also be detected. The
thickness of the reaction has an important
influence on the mechanical properties of
the joint. O n the one hand, it is brittle as
measurements of the microhardness
proved: reaction zone —680; matrix —
120 [HV003]. On the other hand, it has a
different thermal expansion behavior. The
following information sums up thermal
expansion coefficients for various transition metals —carbides and nitrides (Ref.
12):
SiC Si3N4 TiN TiC
4.1 3.2 9.3 7.7
(20-500)
(20-1100)
ZrN ZrC HfN HfC
7.2 6.7 6.9 6.7
AISI304 X5NiCo2818
5.7
18.5
(20-600) [" 10-<VK]
[°C]
The thermal expansion behavior of the
reaction layer plays an important role in
the intrinsic stress that is built up at the interface between ceramic and braze. The
thicker the reaction zone and the larger
the misfit in thermal expansion behavior
between ceramic and reaction layer, the
higher is the intrinsic stress at the interface.
Referring to thermodynamic calculations, a vehement reaction between Hf
and SiC forming HfC was expected, which
could be verified in the experiments, as
Fig. 3 shows. In contrast to Fig. 2, the reaction zones of the SiC-joints exhibited a
much thicker reaction layer. Measurements showed that they were 5jum in
thickness.
Similar to the Hf-SiC-reaction system,
the reactions between Hf and Si3N4 are
supposed to be very strong, too —Fig. 1.
While HfC was the most probable reaction product in the Hf-SiC reaction, HfN is
the most expected reaction product for
the Hf-Si3N4 reaction. According to what
is published in the literature (Ref. 6), the
properties of HfN concerning wetting behavior are very interesting. After a first reaction layer has been formed on the surface of the ceramic, any further reaction
between the ceramic and the active metal
is controlled by diffusion mechanisms. But
hafnium nitride is supposed to build a diffusion barrier on silicon nitride. This means
as soon as a dense layer of HfN is formed
on the surface, it can act as a diffusion
barrier to protect the braze from an
•-••ite
>..
excessive formation of brittle reaction
products. On the one hand, the HfN
Hf
coating makes a wetting and bonding reaction possible, and on the other hand,
2352
1 0 . 9 U 6 GFE
the thickness of the fragile and brittle reFig. 3 — Cross-section and element distribution of SiC/SiC joint for Ag, Cu and Hf. Filler metal: action layer is rather low in comparison to
72Ag24.5Cu3.5Hf (600X).
418-s I NOVEMBER 1990
t^MJ fM,J^.Jmm.
trM^j^,^MM^lrul,JMMML ^ " •»»)? J.|(.» j / f f W * ! ^ , *' . " . ' . J ^ f ' , ? ^
'"i"4'g'JyiB*
% . 4 - Cross-section of an active
metal:73Ag25Cu2Hf (500X).
brazed
hafnium-silicon-carbide reactions. C o n s e q u e n t l y , t h e mechanical p r o p e r t i e s o f
these joints are m u c h b e t t e r than those o f
the SiC joints.
Figure 4 gives a g o o d e x a m p l e for this
reaction m e c h a n i s m . T h e surface of t h e
Si3N 4 ceramic is c o v e r e d w i t h a thin a n d
dense layer o f H f N (thickness: 2jum). Elect r o n m i c r o s c o p i c analysis p r o v e d that t h e
w h o l e brazing z o n e is f r e e of silicon,
w h i c h is additional e v i d e n c e that n o Hf-silicides are f o r m e d at this fairly l o w brazing
t e m p e r a t u r e . The dark c o l o r e d phases are
Ag-rich A g - C u - H f phases w i t h i n t h e A g Cu-eutectic matrix c o m p o s i t i o n .
In o r d e r t o exploit the technical p o t e n tial o f active brazing c o m p l e t e l y , filler
metals f o r c e r a m i c - t o - m e t a l joining must
b e available as w e l l . From the metallurgical p o i n t of v i e w , active filler metals in t h e
system A g - C u - H f have v e r y g o o d w e t t i n g
p r o p e r t i e s o n metallic base materials, t o o .
A d e q u a t e tests h a v e b e e n carried o u t o n
austenitic steel 5X18Cr10Ni (AISI 304) and
5 X 2 8 N H 8 C o , w h o s e t h e r m a l expansion
b e h a v i o r is m o r e a d a p t e d t o that o f silicon
nitride t h a n that o f austenitic steel. T h e
joints c o u l d easily b e f a b r i c a t e d , especially
w h e n Si3N 4 w a s used as ceramic base m a terial. Joining Si 3 N 4 t o austenitic steel leads
t o a b r a z e that is similarly f o r m e d as the
Si3N 4 -Si3N 4 joints, as can b e seen in Fig. 5.
Again, a thin and dense reaction z o n e w a s
f o r m e d , but the a m o u n t o f Hf-containing
phases increased d u e t o a r e d u c e d area of
ceramic material being in c o n t a c t w i t h the
active filler m e t a l . F u r t h e r m o r e , t h e c o p p e r d e n d r i t e s can b e d e t e c t e d .
Joining silicon nitride t o t h e N i - C o alloy
caused the f o r m a t i o n of Ni-, C o - and Fee n r i c h e d phases —Fig. 6. T h e influence o f
these phases o n t h e mechanical p r o p e r ties h a v e n o t yet b e e n investigated. T h e
f l o w b e h a v i o r of t h e A g - C u - H f alloy o n
this base metal w a s v e r y g o o d .
Mechanical Properties
Besides metallurgical investigations, m e chanical tests are essential in o r d e r t o
Si3N4/Si3N4
joint.
Filler
b ^ ^ H
'
Fig. 5-Cross-section of an active brazed Si3N4 (bottom)/AISI
joint. Filler metal: 69Ag26Cu5Hf (500X).
evaluate the quality of a b r a z e d joint. C e ramic materials are generally tested a p plying t h e f o u r - p o i n t b e n d i n g test, a n d
t h e r e f o r e , the b r a z e d ceramic samples
(dimensions: 25 X 4.5 X 3.5
mm/0.79
X 0.18 X 0.14 in.) in this research have
also b e e n tested e m p l o y i n g this m e t h o d .
In general, f i v e specimens w e r e b r a z e d
f o r testing w i t h i n each g r o u p of p a r a m e ter variation. H o w e v e r , in s o m e cases, n o t
all o f t h e m c o u l d be e x p o s e d t o the f o u r p o i n t b e n d i n g test d u e t o brazing faults.
Different
system h a v e
t h e influence
t h e brazing
304 (top)
,t\r
alloys in t h e Ag-Cu-(ln)-Hf
b e e n investigated t o w e i g h
of Hf c o n t e n t s . For o n e alloy,
t e m p e r a t u r e w a s varied as
20 um,
well.
Evidently a Hf c o n t e n t higher than 3.5
w t - % is necessary f o r Si3N 4 t o reach high
b e n d i n g strength values —Fig. 7. Filler
metal A g - C u - 1 . 7 5 H f is an e x c e p t i o n b e cause the foil that was used h a d a thick-
Fig. 6 — Cross-section of an active brazed Si3N4
5X29NH8Co joint: 1-eutectic structure;
2Cu crystals; 3 — diffusion band of Fe and Ni with
traces of Hf and Ag; 4 — reaction zone consisting of Hf and diffused Fe and Ni. Filler metal:
69Ag26Cu5Hf (600X).
B e n d i n g S t r e n g t h [MPa]
5 9 - 2 2 8 MPa
121-165 MPa
109-148 MPa
.130-167 MPa
92-203 MPa
23.2
Iksil
3 3 - 9 4 MPa
LL
Ag-Cu-Hf1.75 Ag-Cu-Hf2
5.8
Ag-Cu-H(3.5 Ag-Cu-Hf4
Ag-Cu-Hf5 AgCu20ln5Hf3
Filler M e t a l
• i
HPSN B-Temp.: 1323 K
TZZ SSiC B-Temp.: 1323 K
KH
HPSN B-Temp. : 1223 K
HH
SSiC B-Temp.: 1223 K
the spread is presented on top of the bars
Pressure: 1E-6mbar (7.5E-7 torr)
F o i l - T h i c k n e s s : 100 um ; * : 2 0 0 y m
Fig. 7 - Four-point bending strength for Si3N4/Si3N4 joints brazed with various active filler metals.
WELDING RESEARCH SUPPLEMENT 1419-s
Weibull-Plot
Silicon Nitride - AgCuHf4 - AISI 304
Cumulative
lnln[1/(1-F)]
Failure P r o b a b i l i t y
1
0,5
x
-
[%]
93
X
80
0
63
* /
-
-0,5
-1
45
' *
*
'
30
20
- 1,5
Weibu ll-Modulus m =3.23
13
-2
-2,5
'55
'
A*
67
81
i
100
122
A v e r a g e S t r e n g t h : 149 M P a (21.61 ksi)
S t a n d a r d D e v i a t i o n : 3 8 . 3 MPa ( 5 . 5 5 k s i )
B r a z i n g C o n d . : T = 1 3 2 3 K; p = 7 . 5 E - 6 t o r r
8
14 8
181
2 21
Bending Strength [MPa]
5 0 MPa • 6.9
ksi
Fig. 8 - Four-point bending strength of Si3N4-AgCu4Hf-AISI 304 joints employing Weibull analysis.
Weibull modulus of bulk material m = 15-25; average bending strength of bulk materialcr 600-800
MPa (87-116 ksi).
ness of 200 / i m in contrast t o the 100-/xmthick foils that h a v e b e e n used in the o t h e r
experiments. Naturally, the thicker foil
c o n t a i n e d d o u b l e the a m o u n t of h a f n i u m ,
a n d t h e r e f o r e , it is a d e q u a t e f o r filler metal
Ag-Cu-3.5Hf.
The Ag-Cu-ln-Hf braze d i d n o t reach
t h e high values o f t h e A g - C u - H f alloys o n
SJ3N4 d u e t o m a n y large brittle phases that
c o u l d be d e t e c t e d metallurgically. Interestingly, this alloy p r o d u c e d fairly high
strength values o n SiC. T h e reason for this
b e h a v i o r has n o t b e e n f o u n d yet.
F u r t h e r m o r e , Fig. 7 s h o w s that the
influence of the brazing t e m p e r a t u r e o n
t h e strength o f the joints is of m i n o r
i m p o r t a n c e , especially f o r Si 3 N 4 .
C o m p a r e d t o silicon nitride, the strength
of the silicon carbide sample is m u c h
l o w e r . A p p a r e n t l y the f o r m a t i o n of t h e
b r o a d reaction z o n e a n d t h e intense reaction cause a p r e d a m a g e of the joint that
c o n s e q u e n t l y decreased the mechanical
strength. Thus, a f u r t h e r m o d i f i c a t i o n of
the alloy c o m p o s i t i o n f o r joining siliconcarbide has t o be c o n d u c t e d . T h e
a c h i e v e d values lie a b o u t 2 0 t o 3 0 %
b e l o w the values measured f o r c o r r e s p o n d i n g Ag-Cu-Ti filler metals (Ref. 7). In
c o m p a r i s o n t o t h e b r a z e d joints, the b e n d ing strengths o f the bulk material are:
4 0 0 - 6 0 0 MPa ( 5 8 - 8 7 ksi) in the case o f SiC
a n d 6 0 0 - 8 0 0 MPa ( 8 7 - 1 1 6 ksi) in the case
of Si3N 4 .
Joining ceramic t o metals is a v e r y difficult task because of the d i f f e r e n t t h e r m a l
expansion behavior. W h i l e c o o l i n g d o w n
such a joint, t h e r m a l stresses are i n d u c e d
in the ceramic that can w e a k e n the joint
420-s | N O V E M B E R
1990
u n d e r certain circumstances. In o r d e r t o
assess the strength o f these joints, A g - C u Hf filler metals w e r e used t o join AISI 304
a n d silicon nitride. In general, the results
have b e e n exceptionally high. In accordance t o SiC-SiC joints, the b e n d i n g
strengths of t h e SiC-AlSI 304 joints have
b e e n similarly l o w .
Even
brazing
silicon
n tride
to
5 X 2 9 N i 1 8 C o w i t h a A g 2 8 C u 2 T i (Ref. 8)
c o u l d not reach the high vali.es of the
combination
Si 3 N 4 -AgCu4Hf-AISI
304,
w h i c h is a v e r y critical c o m b i n a t i o n regarding t h e r m a l expansion beha vior. Probably, the mechanical p r o p e r t i e s of the
braze, as w e l l as the reaction z o n e , m a t c h
better w i t h the ceramic base material in
t h e case of Hf-containing filler metals but
the effects h a v e not y e t b e e n investigated
in full detail.
Figure 8 s h o w s the W e i b u l l plot o f the
measured strengths a c h i e v e d after joining
SJ3N4 t o steel w i t h a A g C u 4 H f f Her m e t a l .
T h e average strength of 149 MPa (21.61
ksi) w a s a b o u t 2 0 % higher than the
strengths that have b e e n a c h i e v e d w i t h
Ti-containing filler metals f o r t h e less critical c o m b i n a t i o n m e n t i o n e d a b o v e . T h e
fracture course was f o u n d in t h e reaction
z o n e , a n d only those samples that r e a c h e d
b e n d i n g strengths o v e r 180 M P a (26.1 ksi)
f r a c t u r e d partly w i t h i n the ceramic base
material.
Conclusions
Brazing experiments s u p p o r t e d b y therm o d y n a m i c calculations s h o w that filler
metals in the A g - C u - H f system are a p p r o -
priate f o r joining silicon nitride or silicon
carbide either t o themselves or t o metal.
Especially o n silicon nitride, Hf-containing
filler metals p e r f o r m w i t h excellent w e t ting a n d b o n d i n g b e h a v i o r b y d e v e l o p i n g
a thin a n d dense H f N reaction z o n e that
acts as a diffusion barrier, as m e t a l l o graphic analysis p r o v e d . Thus, the ceramic
is p r o t e c t e d f r o m a t o o v e h e m e n t react i o n that leads to brittle f o r m a t i o n s . O n
the c o n t r a r y , a thick a n d brittle reaction
layer is f o r m e d w h e n joining SiC w i t h A g Cu-Hf active brazes.
M e c h a n i c a l tests m a t c h v e r y w e l l w i t h
these metallurgical observations. T h e
b e n d i n g strengths o f Si3N 4 joints b r a z e d
b y A g - C u - H f alloys a c h i e v e d exceptionally
high values. Particularily, t h e Si 3 N 4 -AISI 304
joints, w i t h average b e n d i n g strengths o f
149 M P a (21.61 ksi), are o f special t e c h n i cal interest.
The investigations exhibited impressively that the A g - C u - H f system is a p r o m ising alternative t o Ti-containing active
filler metals. Further research should t r y t o
scrutinize t h e r e a c t i o n mechanisms as w e l l
as d e t e r m i n e the o p t i m i z e d alloy c o m p o sition of the Ag-Cu-(ln)-Hf filler metals in
m o r e detail. Besides, those filler metals
should be c o m p a r e d t o active brazes o n
the basis o f o t h e r IVA a n d V A g r o u p elements in o r d e r t o get a b e t t e r evaluation
o f possible active filler metals.
A ckno wledgment
T h e w o r k r e f e r r e d t o in this p a p e r w a s
s p o n s o r e d b y the D e u t s c h e Forschungsgemeinschaft, B o n n , G e r m a n y .
References
1. Spencer, P. H„ et al. 1981. Hafnium,
physiochemical properties of its compounds
and alloys. International Atomic Energy Agency,
Wien.
2. Kubaschewski, O., and Alcock, C. B.
1979. Metallurgical Thermochemistry. Pergamon Press, Auflage.
3. Naka, M., and Okamoto, O. 1985. Wetting of silicon nitride ceramics by coppertitanium or copper-zirconium alloys. Trans.
JWR114(1):29-34.
4. Cottselig, B., Gyramati, E., Naoumidis, A.,
and Nickel, H. 1989. Development of methods
for joining nonoxide silicon ceramics. Proceedings of the International Conference on joining
Ceramics, Glass and Metal. Bad Nauheim (FRG),
pp. 191-199.
5. Gottselig, B., Gyramati, E., Naoumidis, A.,
and Nickel, H. 1989. Beitrag zur Verbindungstechnik von SiC-Keramik uber metaliische Zwischenschichten. Berichte der KFA Julich. Nr.
2288.
6. Schuster, I. C., etal. 1988. Joining of silicon
nitride ceramics to metals: the phase diagram
base. Materials Science and Engineering, A
105/106 (1988), pp. 201-206.
7. Lugscheider, E., and Boretius, M. 1989.
Active brazing of silicon carbide and silicon nitride to steel using a thermal-stress-reducing
metallic interlayer. Proceedings of the International Conference on joining Ceramics, Glass
and Metal. Bad Nauheim (FRG) pp. 25-33.
8. Wielage, B„ and Ashoff, D. 1989. Brazing
of ceramic and metal-ceramic compounds.
Proceedings of the International Conference on
joining Ceramics, Glass and Metal. Bad Nauheim (FRG) pp. 385-399.
9. Cadenhead, D. A., and Danielli, J. F. 1981.
Progress in Surface and Membrane Science.
Academic Press, New York, London, Toronto,
Sydney, San Francisco, pp. 354-470.
10. Nicholas, M. G. 1986. Bonding ceramicmetal interfaces and joints. Ceramic Micro-
structure '86: Role of Interfaces. Plenum Press,
New York, p. 349.
11. Barin, I. 1989. Thermochemical data of
pure substances. VCH-Verlag, Weinheim.
12. Appen, A. A., and Petzold, A. 1984. Hitzebestandige
Korrosions-, Warmeund
verschlei^-schutzschichten. YEB Deutscher Verlag Fur Crundstoffindustrie, Leipzig.
I-
Z
ui
s
o.
o
x
o
cr
<
UJ
tn
Ui
tr
WRC Bulletin 343
May 1989
} -
z
ui
Destructive Examination of PVRC Weld Specimens 202, 203 and 251J
2
a.
o
-J
ui
This Bulletin contains three reports:
>
UJ
( 1 ) Destructive Examination of PVRC Specimen 202 Weld Flaws by JPVRC
By Y. Saiga
o
o
tr
( 2 ) Destructive Examination of PVRC Nozzle Weld Specimen 203 Weld Flaws by JPVRC
By Y. Saiga
Ul
<
tn
( 3 ) Destructive Examination of PVRC Specimen 251J Weld Flaws
By S. Yukawa
s
The sectioning and examination of Specimens 202 and 203 were sponsored by the Nondestructive
Examination C o m m i t t e e of the Japan Pressure Vessel Research Council. The destructive examination of
Specimen 251J was p e r f o r m e d at the General Electric Company in Schenectady, N.Y., under the
sponsorship of the S u b c o m m i t t e e on Nondestructive Examination of Pressure Components of t h e
Pressure Vessel Research C o m m i t t e e of the Welding Research Council. The price of WRC Bulletin 343 is
$24.00 per copy, plus $5.00 for U.S., or $8.00 for overseas, postage and handling. Orders should be sent
with payment to the Welding Research Council, Room 1 3 0 1 , 345 E. 4 7 t h St., New York, NY 10017.
a.
O
-J
ui
>
Ul
Q
O
tr
<
Ul
tn
Ul
tr
^~~
i-
z
s
WRC Bulletin 356
August 1990
Q.
>
o
ui
o
This Bulletin contains three reports involving welding research. The titles describe the contents of the
reports.
( 1 ) Finite Element Modeling of a Single-Pass Weld
o
tr
<
UJ
By C. K. Leung, R. J. Pick and D. H. B. Mok
tn
( 2 ) Finite Element Analysis of Multipass Welds
UJ
z
2
a.
By C. K. Leung and R. J. Pick
o
_l
( 3 ) Thermal and Mechanical Simulations of Resistance Spot Welding
ui
>
By S. D. Sheppard
Publication of the papers in this Bulletin was sponsored by the Welding Research Council. The price of
WRC Bulletin 356 is $35.00 per copy, plus $5.00 for U.S. and $10.00 for overseas postage and handling.
Orders should be sent with payment to the Welding Research Council, 345 E. 4 7 t h St., Room 1 3 0 1 , New
York, NY 10017.
X
o
tr
<
Ul
to
UJ
tr
WELDING RESEARCH SUPPLEMENT 1421-s