The brazing of Si3N4 with Ni-Cr-Si alloys

Citation for published version (APA):

Ceccone, G., Nicholas, M. G., Peteves, S. D., Kodentsov, A., Kivilahti, J., & Loo, van, F. J. J. (1995). The
brazing of Si3N4 with Ni-Cr-Si alloys. Journal of the European Ceramic Society, 15, 563-572.
https://doi.org/10.1016/0955-2219(95)00015-M

DOI:
10.1016/0955-2219(95)00015-M

Document status and date:

Published: 01/01/1995

Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be
important differences between the submitted version and the official published version of record. People
interested in the research are advised to contact the author for the final version of the publication, or visit the
DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page
numbers.
Link to publication

General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please
follow below link for the End User Agreement:
www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
providing details and we will investigate your claim.

Download date: 09. ago. 2023

 Journal of the European Ceramic Society 15 (1995) 563-512
0 1995 Elsevier Science Limited
Printed in Great Britain. All rights reserved
0955-2219/95/$9.50
0955-2219(95)00015-1

The Brazing of S&N4 with Ni-Cr-Si Alloys

G. Ceccone,a M. G. Nicholas,” S. D. Peteveqa A. A. Kodentsoyb J. K. Kivilahtib
& F. J. J. van Loob
“Institute for Advanced Materials, CEC, Petten, The Netherlands
‘Solid State Chemistry and Materials Science, Eindhoven University of Technology, Eindhoven, The Netherlands

(Received 29 July 1994; revised version received 30 October 1994; accepted 10 January 1995)

Abstract joining metal workpieces do not wet technologi-

cally important ceramics and hence a sub-class of
The use of commercially available and laboratory brazes, active metal brazes, have been developed
made Ni-Si-Cr brazes for wetting and bonding for the joining of ceramics. The distinctive feature
S&N, has been explored. The microstructure and of these active metal brazes is that they react to
chemistry of brazed joints has been characterized by change the surface chemistry and wettability of
techniques such as optical microscopy, scanning the ceramics.
electron microscopy, electron probe micro-analysis Commercially available active metal brazes are
and X-ray d@raction. Si,N, was wetted well by all generally based on the Ag-Cu-(In, Pd)--Ti and
the braze alloys at 11.50-1200°C but the joint qual- Cu-Al-Si-Ti systems and can be used for applica-
ity obtained was poor because of the formation of tions that do not require service temperatures
brittle phases and solid solutions. higher than 400°C. 4-7 Si,N, can be joined with
these brazes and important applications have been
found in the automobile industries. However,
1 Intruduction many of the proposed future uses of Si,N, involve
service temperatures in excess of 700°C. The selec-
The increasing interest in the application of struc- tion of a high temperature active metal braze
tural ceramics such as S&N4 and SIC is due to involves two tasks: (a) choice of a solvent and,
their high-temperature strength and creep resis- (b), identification of a reactive component. The
tance, excellent wear, corrosion and thermal shock first task requires consideration of metallurgical
resistance. However, reliable joining technologies and technological factors; the solvent must be
are needed if ceramic inserts are to be used to sufficiently refractory and chemically inert with
exploit these properties or complex components respect to the workpieces and the environment so
are to be made from simple ceramic bodies. that the joined structures can withstand service
Brazing is an attractive technique for joining conditions such as high temperature oxidizing at-
ceramic-ceramic and, particularly, ceramic-metal mospheres. The development of brazes for high
systems. However to profit from the high tempera- temperature metals such as super alloys can pro-
ture mechanical properties of ceramics, high tem- vide some guidance for solvent selection but the
perature brazes have to be employed and several second task is often more difficult to accomplish.
studieslm3 have suggested Ni-based alloys as suit- Identifying a suitable reactive component requires
able brazes for joining Si,N, and Sic. information about what type of products will be
This paper describes an evaluation of the poten- formed by the specific workpiece/braze combina-
tial usefulness of some Ni-Cr-Si brazes for joining tions during the joining process. Equilibrium
Si3N4, employing a material science approach thermodynamics can provide insight about the
based where possible on phase stability and ther- chemistry of the products, but not about their
modynamic factors. formation kinetics or properties such as wettabil-
ity or strength. Thus experimentation is almost
always necessary to identify the optimum brazing
2 Background materials and process conditions.
Several studies of the active metal brazing of
Brazing is a versatile and cost-effective technique engineering ceramics aimed to increase the service
for joining both similar and dissimilar material temperatures have been reported. In particular,
systems. However, many of the brazes used for Au-Ni-Ti and Pd-Ni-Ti alloys have been used to
563
 564 G. Ceccone, M. G. Nicholas, S. D. Peteves, A. A. Kodentsov, J. K. Kivilahti, F. J. J. van Loo

join S&N, and SIC ceramics8 The results obtained Si

indicate that a major problem is the enhanced /A\ Cr-Ni-Si
I \
chemical reaction between the ceramic and the \ 1027 OC
/I \
braze. 0.8,; if.2

The present work focused on the utility of / \

\
1’ \
Ni-Cr-Si brazes. These are based on the Ni-21%Si O.S,k
eutectic* which melts at 1143°C and were chosen /
for several reasons: their commercial availability, +s
,’
If
CrSi +
the expectation that the presence of Si will reduce
the dissolution of the Si3N4,’ the successful use of
Ni-Cr alloys to diffusion bond Si3N4,10,1’their
potentially good high temperature mechanical
properties, the ability of Cr to form bonding lay-
ers of CrN by reacting with S&N,,” and the good Cr 0.8 0.6 0.4 0.2 Ni
-x0
oxidation resistance of Ni-Cr alloys. However, a
problem to be faced when using Si as a tempera-
si
ture depressant for Ni brazes is the brittleness of Cr-Ni-Si
IA\
the joints caused by the formation of silicides if / \
\ 1127 OC
/I \
the Si content exceeds about 10%.
0.6,; if.2
Thermodynamic modelling can be a useful design \
I/ \
tool in predicting the behaviour of the ceramic-
braze systems and predictions for the SiCYNi-Cr-Si $4 +
I $1
braze systems are in good agreement with the ex-
perimental results.‘,* Unfortunately Si,N,/Ni-Cr-Si CrlSi + I
‘1
,___-__, \\
d 0.6
systems are more complicated because the reac-
tion processes are strongly influenced by the N2
local partial pressure. 3.10~11
Furthermore, to under-
stand the behaviour of Si,N,/Ni-Cr-Si systems
requires a thorough knowledge of the Ni-Cr-Si-N 0.6 0.6 0.4 0.2
Cr Ni
phase diagram. This is lacking but some data,
cited below, are available for the Ni-Cr-Si and
Ni-Cr-N systems. Furthermore, a few (Ni-Cr-Si)/
N, nitridation experiments were conducted as a
precursor to this brazing study, as discussed later.
Some experimental data for Ni and Cr rich
Ni-Cr-Si alloys have been published’2-14 and
Ansara et aLI3 have assessed the data for Si lean
alloys. The Ansara isotherms at 1027 and 1127°C
are reproduced in Fig. 1. An obvious deficiency of
this evaluation is that even the isotherm for v
LV V \
1127°C does not contain a liquid field, yet the 0.4 0.2 Ni
commercial BNi-5, Ni-19Cr-18Si braze has a
solidus of 1078°C. Hence, it became clear early on -XCr
during this study that a more sound knowledge of Fig. 1. Calculated partial isothermal section for the Ni-Cr-Si
this system was necessary if braze microstructures system at 1127°C and 1027°C (after Ref. 13). The insert dis-
were to be interpreted, so a parallel pro- plays the compositions of the braze alloys used in this study:
BNi-5 is shown as a solid circle and the laboratory made
gramme 15,16was initiated to derive a full Ni-Cr-Si alloys as hollow circles.
diagram and reference to its results will be made
in the Discussion.
The stability diagram of nitride phases in vari- decreasing temperature down to 1000°C a ternary
ous Ni-Cr alloys at 950, 1000 and 1125°C as a phase, r, is stable in the Ni-Cr-N system.
function of N, fugacity (partial pressure) is given Nitridation of Ni alloyed with at least 28%Cr pro-
in Fig. 2. More details concerning the thermody- duced a r phase, Cr,,Ni,N,” or Cr,,Ni,N,,” at
namic calculations and experimental verification 1000-l 125°C which thermodynamic analyses
of this diagram can be found elsewhere.” With demonstrated to be in equilibrium with the FCC
Ni solid solution and Cr,N at modest nitrogen
*Unless otherwise specified, composition at% are given. activities. l9
 Brazing of S&N, with Ni-Cr-St’ alloys 565

1000 ‘C

FCC + ON
4:1\,,,,,
i .l-
y3 .01- * + Cr,N -

.001- FCC .001- FCC

lE-4 I I I I lE-4 I I I I
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0

U(Cr) = N,/ (N, + N,) U(Cr) = N, / (N, + N,)

(4 (b)
950 ‘C

FCC + CrN

0 0.2 0.4 0.6 0.8 1.0

UfCr) = N,/ (N, + N,)

w
Fig. 2. Calculated stability diagrams of nitride phases in various Ni-Cr alloys as a function of nitrogen fugacity: (a) at 1125”C,
(b) at lOOOT, and (c) at 950°C.

3 Experimental were ultrasonically cleaned in methanol, rinsed in

water and dried using an air blast immediately
The ceramic used in this work was a commercial before being assembled as a sandwich and placed in
hot-pressed S&N, containing 1.8% A&O, and 51% the brazing furnace. The brazing experiments were
Y,O, as sintering aids. It was obtained from Cera- performed under vacuum of 6 X lo4 Pa (1 Pa = 1
dyne Inc., Costa Mesa, CA, USA, as a plate from
which were cut samples measuring 7 X 15 X 3 Table 1. Composition of the alloys used for wetting and
mm. The faces to be bonded were polished, using brazing experiments
a 1 pm diamond paste, to a final average rough-
Alloy Liquidus (“C) Composition (at%)
ness (&) of about 0.3 ,um. The braze alloys used Designation
were 90 pm thick foils of a commercial BNi-5, Ni Cr Si
obtained from Wesgo Inc., Belmont, CA, USA, and
BNi-5 1135 Bal. 18.9 18.4
500 pm thick foils of laboratory produced alloys. Wesgo
Their compositions are indicated in Fig. 1 and N-21 1150 Bal. 6 22
summarised in Table 1. Table 1 also includes their Lab. Made
N-4 1350 Bal. 5 10
liquidus temperatures, which were determined by Lab. Made
differential thermal analysis and ranged from N-5 1345 Bal. 10 10
1135°C for BNi-5 to 1400°C for one of the labora- Lab. Made
tory alloys. N-35 1400 Bal. 10 5
Lab. Made
The ceramic pieces and the braze alloy foils
 566 G. Ceccone, A4. G. Nicholas, S. D. Peteves, A. A. Kodentsov,J. K. Kivilahti,F. J. J. van Loo

x lO-5 bar) using temperatures of 1150 or 12OO”C,

times of 0.08-l h and heating and cooling rates of 0

lO”C/min. In ‘accord with recommended practice

for the BNi-5 alloy*‘, each brazing cycle included
a 1 h sub-solidus dwell at 1050°C during the cool-
ing stage in an attempt to homogenize the seam
microstructure. A slight uniaxial force of about
0.5 MPa was applied during brazing to maintain
alignment of the workpieces.
Before commencing the main series of brazing 20
-‘..._,.
“...,
experiments, the wetting behaviour of the brazes x .. . . . .a
0’t I -
in vacuum at 1200°C was assessed. The alloy foils 0
00
6 10 crcod I

20 26
I
30

were placed on polished and cleaned S&N, sub- (at%)

strates and heated at 10Wmin to 1200°C in a Fig. 3. Wetting behaviour of Ni-Cr-Si brazing alloys on
vacuum chamber evacuated to a pressure of less S&N, substrates at 1200°C (filled circles). The composition of
than 5 X lOa Pa. The samples were kept for 30 the alloys is given in Table 1. Also included are literature
data for BNi-5 (open squares) Ref. 9, for Ni-17P-Cr brazes
min at temperature and then cooled to room tem- (open triangles) Ref. 21 and for BNi-2, a Ni-13EHSi-7Cr-3Fe
perature at 1O”C/min. braze at 1150°C (open circles) Ref. 22.
After being used in the wetting or brazing
experiments, the samples were cross-sectioned,
polished with diamond paste and finished with a 120”,22but our observation of excellent wetting by
0.3 pm alumina slurry. The contact angles of wet- BNC5 confirms a literature report.’ Using the crite-
ting samples and the interfacial microstructures of rion of a maximum contact angle of 30” for effec-
brazed samples was characterized by optical tive wetting of a ceramic suggested by Moorhead
microscopy and scanning electron microscopy. and Keating,23 it can be seen that all the Ni-Cr-Si
Electron probe microanalysis, (EPMA), and X-ray brazes used in this study are adequate but some of
diffraction, (XRD), techniques were used to char- the other compositions are unsuitable.
acterise the micro-chemistry of braze seams. The
Vickers hardness of phases within the seams were 4.2 BNi-5 microstructure
derived using a 50 g load applied by Zeiss MHT4 Comparison of the quoted Ni-19Cr-18Si compo-
equipment. sition of BNi-5 with the phase stability diagrams
Nitridation of the Ni-Cr-Si alloys under total for both 1027 and 1127°C Fig. 1, reveals that it
gas pressures up to 100 bars, took place in a sin- lies within the solid-solid Ni,Si,-FCC Ni two
tering furnace (KCE, Germany). Temperature phase field, yet very close to the boundary of the
control was carried out with an optical pyrometer solid-solid-solid Ni,Si,-FCC Ni-Cr,Ni,Si, three
within +5”C accuracy. All experiments were car- phase field. It was surprising, therefore, that the
ried out in pure nitrogen (0, 5 10 ppm; H,O I 10 microstructure of the as received BNi-5 braze,
ppm). After nitridation under these conditions, shown in Fig. 4, contained three phases. These
formation of oxides was not detected at the sur- were: (1) a matrix labelled M with a hardness of
face of the samples. The surface of the alloys were HV463, (2) a substantial amount of a hard irregu-
examined after nitriding by XRD analysis with a lar phase labelled A, and (3) a small amount of a
cylindrical texture camera using nickel-filtered hard angular phase labelled B.
Cu-Ka radiation.

4 Results and Interpretation

4.1 Wetting
The contact angles derived from the 1200°C wet-
ting studies are presented in Fig. 3. The commer-
cial BNi-5 braze wetted well with an angle of lo”,
while the laboratory produced braze alloys wetted
even better with contact angles of virtually zero.
Also included in the figure are some contact
angles reported for other Ni braze alloys.9,21,22It
can be seen that not all Ni brazes wet, the contact Fig. 4. Scanning electron micrographs of as-received BNi-5
angle for BNi-2 (Ni-13B-8Si-7Cr-3Fe) being microstructures.
 Brazing of Si,N, with Ni-Cr-Si alloys 567

Table 2. Phases identified in BNi-5 foils and braze seams. (Compositions are given in at%). The stoichiometry of the Q-phase was
found to be NilSCr2,SisNs

Condition Phase M Phase A Phase B

As -received foil Ni S.S. Ni,Si, Q-phase

Ni-24Cr-15Si Ni-26Si-1 OCr Ni-45Cr-15Si-3.5N
Melted foil, Ni S.S. Ni,Si, Q-phase
115O”C, 15 min Ni-24Cr-12Si Ni-27Si-7Cr Ni-SlCr-15Si-3.5N
Braze seam, Ni S.S. Ni,Si, Q-phase
115O”C, 15 min Ni-24Cr-12Si Ni-28Si-2.8Cr Cr-29.6Ni-
14,9Si-7.3N
Braze seam, N&-S& Ni,Si Q-phase
12OO”C, 15 min Ni-27Si-8.4Cr Ni-32Si-3.5Cr Cr-29Ni-
14.2Si12.2N
Braze seam, Ni$i, Ni,Si Q-phase
12OO”C,30 min Ni-27Si-8.2Cr Ni-32Si4.30 Cr-28.2Ni-13.5Si-139N

The compositions of the phases in this and After brazing at 1200°C (Fig. 5) it was found
other cross-sections are summarised in Table 2. that the Vickers hardness of the braze matrix
Comparison of these compositions with the phase had been increased to about HV 1100
Ansara diagrams shown in Fig. 1 identifies the (Table 3) and that the composition of the braze
Ni-24Cr-15Si Matrix, M as a Cr and Si saturated seam had been changed (Table 2).
FCC Ni and the hard irregular Ni-26StlOCr The matrix became depleted in Cr but Fig. 6
phase A as a Cr enriched Ni,Si,. However, assign- shows that extending the brazing cycle caused the
ing an identity to the Ni-5 lCr-15Si-3.5N compo-
sition of the angular phase B is difficult. The
angular phase contains nitrogen and hence cannot
be identified validly as any of the ternary com-
pounds shown in the Ansara diagrams. Even if the
N content is ignored, the normalised composition
is not in accord with any of the phases shown in
Fig. 1 but lies within the two phase field between
Cr3Ni,Siz and the Cr,Ni,Si, or u, phase. Therefore,
it was decided somewhat arbitrarily for the mo-
ment, (it will be justified later on), to regard the
angular phase (B) as a quaternary compound, Q.
The microstructure of the BNi-5 foil was little
changed by subjecting it to a simulated brazing
cycle by melting in an A120, crucible and holding
at 1150°C for 15 min, Table 2. The microstructure
contained three phases; 53% of the Ni-based solid
solution matrix phase, 40% of Ni,Si, and 7% of
the quaternary Q-phase. However, major changes
occurred when the alloy was subjected to braze
cycles in contact with S&N,.

4.3 Si,N,/BNi-5 interactions

The good wetting behaviour of BNi-5 ensured that
brazed sandwiches were formed successfully at both
1150 and 1200°C. The wetting was so good that
the braze also spread over the sides of the ceramic
pieces, but the structures of the joints were poor.
Optical microscopy revealed that there were voids
within the brazed seam and that both the seam and
the ceramic-metal interfaces were cracked, presum-
ably due to stresses caused by the mismatch between (W
the 9.5 X lOA ‘Cl thermal expansion coefficient Fig. 5. Microstructure of BNi-5 joints (a) brazed at 1200°C
of the braze,’ and the 3.3 X 10” “C-r of S&N,. for 15 min, (b) brazed 1200°C for 30 min.
 568 G.Ceccone, M. G. Nicholas, S. D. Peteves, A. A. Kodentsov, J. K. Kivilahti, F. J. J. van Loo

Table 3. Hardness of the matrix phase in Ni-C-Si alloys

before and after use in various experiments

Alloy or System Experiment Hardness

(Hv 5Ogr)

BNi-5 None 463.5f 11.4

Si,NdBNl-5 Brazingat 1200°C 1090.3f 40.7 6000-
Si,N,/N-21 Brazingat 1200°C 1109.0f 34.0
S&NJ/N-35 Wettingat 1200°C 1061.5i 67.8
S&N,@-4 Wettingat 1200°C 1102.3+ 31.6

Si content of the braze seam to increase progres-

sively as more Si rich silicides were formed. With
increasing the brazing time, the Ni,Si, phase iden-
tified in Fig. 4 as A was converted to Ni,Si, Fig. 7,
and became very abundant. Finally, the EPMA
studies showed that quite large amounts of a com-
plex Ni-Cr-Si-N phase was formed, the N con-
40 42 4.4 46 48 SO
tent of which increased with the brazing
2 theta
temperature and time. By comparison with the mi-
crostructure of the as received filler metal this Fig. 7. X-ray diffraction spectrum of a fracture surface of a
Si,N,/BNi-5 joint brazed at 1200°C for 30 min.
phase was identified as the quaternary Q-phase.
The large amount of Si found in the braze seam
clearly demonstrates that S&N, reacts with the 500 pm thick foil of Ni-22Cr, Fig. 8. In this case,
liquid filler metal during the brazing process. The &Ni,Si was not found within the braze seam after
conclusion is reinforced by the fact that decreasing bonding at 1200°C for 30 min. However, there
the brazing time to 5 min, still caused the conver- was an extensive formation of the Q;phase within
sion of the matrix to a brittle silicide. Rapid reac- a matrix of N&based solid solution and Ni,Si,. In
tion of the ceramic with the molten braze is places, the Q-phase was formed as spherical parti-
reported also by Hadian and his co-workers,’ who cles, that suggested it had grown in the molten
used the BNi-5 alloy for brazing S&N, to MO at matrix. Another reaction product worth noting
1200°C. They measured a smaller increase in the was a layer of ‘CrN’ at the braze-ceramic inter-
Si content of the braze than it was observed in our face that was observed exclusively in the central
study, but this difference may be explained by the region of the joint.
higher partial pressure of nitrogen in their experi- Although the brazing temperature suggested for
ments (-8 Pa) compared to ours (-8 X 10m6Pa), the commercial BNi-5 alloy ranges between 1150
which leads to a lower chemical potential of sili- and 1216°C,24 the present results demonstrate that
con in S&N4 and, therefore, to lower driving force a temperature of 1200°C is too high if it is to be
for diffusion of silicon. used to join Si3N4. Hence, further brazing experi-
In some final experiments, the configuration of ments at lower temperatures were performed. The
the joints was changed by the introduction of a

Ni22Cr

J
12

Fig. 6. Concentration of Si in the braze seam of Si,NdBNiJ

joints plotted as a function of the brazing time at 1200°C.
Each point is the average of 5 measurements at different Fig. 8. Microstructure of the Si,N4/BNi-WNi-22Cr interface
5 X 5 pm2 locations. (12OO”C, 15 min).
 Brazing of Si,N, with Ni-Cr-Si alloys

Fig. 9. Scanning electron micrographs of a Si,N@Ni-5 joint

brazed at 1150°C for 15 min.

microstructure of a cross-section of a Si,N,/BNi-5

joint brazed at 1150°C for 15 min, shown in Fig.
9, is not significantly different to that produced at
12OO”C, Fig. 5. Once more, three phases are pre-
sent in the brazing seam. Quantitative electron
probe micro-analysis of the braze seam, sum-
marised in Table 2, showed that the Ni solid solu-
tion matrix was still present in the seam even
though large amounts of Ni,Si, had been formed;
the amount of Cr dissolved in the Ni,Si, was
decreased from 8 to about 3%, and the angular
Q-phase was even more deficient in nitrogen.
Fig. 10. Cross-sections of Si,N&21 joints brazed at, (a)
1200°C for 15 min, (b) 1150°C for 15 min.
4.4 Interactions of S&N, with other Ni-Cr-Si
brazes
The cross-section of a S&NJ+21 alloy (Ni-6Cr-22Si) this reaction layer because of spatial resolution
joint bonded at 1200°C for 15 min, shown in Fig. limitation of the electron probe micro-analysis
10(a), has features similar to those observed in technique, but from the optical appearance and
S&N,/BNi-5 joints. Once again, the results indicate morphology as well as from qualitative EDX ele-
a brazing temperature of 1200°C is too high mental maps we may conclude that this layer is
because of excessive ingress of Si. This conclusion the quaternary Q-phase.
is supported by results obtained using the other EPMA analysis of the braze seam revealed that
Ni-Cr-Si laboratory alloys, N-4 (Ni-SCr-lOSi), the matrix was Ni& and that the angular inclu-
N-5 (Ni-lOCr-IOSi) and N-35 (Ni-lOCr-5Si). sions shown in Fig. 10 had similar composition
Melting occurred even though the liquidus tem- with the phases identified as Q in the other joint
peratures of the alloys are 1350-14OO”C, Table 1. microstructures. Finally, cracks and voids were
The melting was associated with a matrix hard- present in the seam, so the microstructure was
ness increase, as shown in Table 3, and the forma- unacceptable as in the case of the Si,NJBNi-5 joints.
tion of the brittle quatemary, Q-phase.
A significantly different microstructure was 4.5 Nitridation of Ni-Cr-Si alloys
produced when the laboratory alloy N-21 In general, as in the case of the Ni-Cr-N system,
(Ni-6Cr-22Si) was used to braze at 1150°C for 15 see Fig. 2, the activity of N, largely determines the
min, as illustrated in Figure 10(b). In this case a stability of various phases in the Ni-Cr-Si-N sys-
reaction layer with an average thickness 2.4 + 0.5 tem. Nitridation of Ni-15Cr-1OSi alloy at 1125°C
pm was formed at the ceramic/braze interface. A under 1 bar N, pressure or higher, produced a
quantitative composition cannot be quoted for surface layer and internal precipitates of the
 570 G. Ceccone, M. G. Nicholas, S. D. Peteves, A. A. Kodentsov, J. K. Kivilahti, l? J. J. van Loo

interaction of the nitrogen with the liquid

Ni-Cr-Si filler metal is as follows:
S&N4 + L @ Q + L’
where L’ is the remnant Cr-depleted liquid filler
metal. The second eutectic ‘isotherm in the binary
Ni-Si system (L’ @Ni,Si, + SNi,Si) lies at
1215°C.25 Chromium that is still present in the liq-
uid may depress the temperature of this eutectic.
However, the significant ingress of silicon from
S&N, into the molten braze leads to the forma-
tion, inside of liquid, of precipitates of SNi2Si (or
its high temperature modification t$Ni,Si) having
a higher liquidus temperature.26 Then, these precip-
itates grow through the liquid. Such a mechanism
Fig. 11. The formation of the quatemary phase, Q,
(Ni,,Cr2,Si8N6) after nitriding Ni-15Cr-1 OSi alloy at 1125°C
seems to be confirmed by the resultant morphology
under 1 bar of Nz (C = Metal-gas interface). of the reaction zone after brazing at 1200°C for
15-30 minutes (Fig. 5(a), (b)).
At 1150°C the interactions between the ceramic
quaternary phase Q, Fig. 11. Careful EPMA and the braze are less severe. Still the liquid filler
analysis of equilibrated samples pointed out to a metal is enriched with Si and depleted from Cr.
Ni,,Cr,,Si,N, stoichiometry for the Q-phase. Besides the formation of the quaternary Q-phase,
Additional nitriding experiments of different excess of Si within the molten filler metal result in
Ni-Cr-Si alloys (N&(20-35%Cr)-(2-5% Si)) have the precipitation of Ni,Si, at the brazing tempera-
shown that the Q-phase exists as an equilibrium ture rather than upon cooling.
phase in the quaternary Ni-Cr-Si-N system over In order to decide whether any other Ni-Cr-Si
the temperature range of lOOO-1250°C. braze might be suitable, it is necessary to under-
stand the reaction paths followed by the
ceramic/braze systems, especially the formation of
5 Discussion the nitride phases inside the braze seam. For in-
stance the formation of a continuous layer of CrN
It is clear that S&N4 interacts strongly with at the interface is desirable”Ti5 because this could
Ni-Cr-Si brazes. Thus all the brazes wetted well impede the ingress of Si, whereas the formation of
at 12OO”C, and the behaviour of the laboratory- the brittle Q-phase in the seam deteriorates the
made brazes (N-4, N-5 and N-35) is evidence of joint quality. Quantitative understanding of the
a melting temperature depression caused by the reaction paths cannot be achieved because of the
ingress of Si from the ceramic. Vigorous work- lack of phase equilibrium data for the
piece/braze interactions are often associated with Cr-Ni-Si-N system but some qualitative guidance
good wetting, and all the brazes used in this study can be obtained from considering the recent
wetted very well. The least wetting braze was the Ni-Cr-Si phase diagram study in parallel with
commercial BNi-5 alloy which also has the highest this brazing work.‘5v16
Cr content. However, the correlation between Cr The ineffectiveness of the Si contents of the
content and wetting behaviour is unclear because, brazes in inhibiting further detrimental ingress of
for instance, the Ni-Cr-P brazes with the highest Si from the ceramic is further illuminated by these
Cr contents also formed the lowest contact angles studies. Figure 12 displays two experimentally de-
with Si,N,.2’ termined isothermal sections for 1050 and 1125°C
While braze/ceramic interactions promote wet- that should be relevant to phase structures of
ting, the resultant microstructures are excessively molten and solidified brazes. Comparison of this
brittle and none of the Ni-Cr-Si brazes used in new information with the assessments for Si lean
this study can be considered satisfactory joining alloys shown in Fig. 1 reveals both similarities and
materials for S&N,, not even the commercially differences which impinge on our observations.
developed BNi-5. There is agreement about the phase structure of
Upon brazing at 1200°C the liquid filler metal the laboratory made alloys: N-4, (Ni-SCr-lOSi),
reacts with the ceramic leading to an ingress of Si N-5, (Ni-lOCr-1OSi) and N-35, (Ni-lOCr-SSi),
and nitrogen into the braze. Since the melting lie within the Ni solid solution and N-21,
point of the Q-phase is higher than 125O”C, as (Ni-6Cr-22Si), lies within the two phase field be-
determined by the nitridation experiments, the tween Ni,Si, and the FCC or r_Ni solid solution.
 Brazing of S&N4 withNi-Cr-Si alloys 571

Si dissociating ceramic. Ingress of Si will cause the

Ni solid solution braze matrix to be replaced pro-
1125 OC gressively by silicides and we have observed the
formation of Ni,Si, and Ni,Si.
The same considerations apply to the interac-
tion of solid Ni-Cr alloys during the diffusion
bonding of Si,N,, but the approach to equilibrium
will be much slower. The diffusion coefficient of Si

-*yNi*si in Ni-22Cr at 1200°C is 4.9 X 1W14m*s-i, proba-

bly 10 000 times slower than in the molten braze
because the self diffusion coefficient for molten Ni
kz4i&i,
i Si
lies in the range 10~8-10~9m2s~1.27
A substantial displacement Cr is caused by the
conversion of the Ni-24Cr-12Si matrix of the
braze to Ni,Si, and Ni,Si containing only 8% of
Cr. The ingress of Si from the ceramic that causes
Cr 0.8 0.6 0.4 0.2 Ni
this displacement of Cr is accompanied also by a
--x0 release of nitrogen and our experimental results
Si show the displacement to result in a formation of
a Cr rich nitride, Ni,,Cr2,SisN6, which can be in

A 1050°C equilibrium with the Ni-solid solution and Ni,Si,

or Ni,Si. The presence of Ni,,Cr2,Si,N6, therefore
provides proof of one logical step in the progress
of a reaction path in the uncharted Ni-Cr-Si-N
system created when a Ni-Cr-Si alloy is used to
braze S&N,.
CrSi>/,, _\, _\ J\
The 1050°C isotherm in Fig. 12 shows that one
or more of the Cr-35Ni-25Si (T), Cr-30Ni-1OSi
(a) or Ni-30Cr-20Si (71’)phases will be in equilib-
rium with TNi braze matrices and that the T and
n’ phases will be in equilibrium with Ni,Si, matrices.
Experimental measurements of the Cr enriched
angular Q-phase in solidified BNi-5 seams with
both YNi and silicide matrices yielded composi-
Cr 0.8 0.6 0.4 0.2 Ni tions of approximately Ni42Cr-16Si-12N, inter-
--XC, mediate in metal stoichiometry between that of
Fig. 12. Experimentally determined isothermal sections for the T and u phases The a-phase is not in equilib-
the Ni-Cr-Si system at 1125°C and 105O”C, after Ref. 16. rium with the Ni,Si, present in all the solidified
braze seams, but there are tie lines between the T-
phase and both YNi and Ni,Si,. Thus the designa-
tion of the angular Q-phase in BNi-5 braze seams,
Similarly, there is agreement that any a-phase pre- and possibly the reaction product layer formed by
sent in the as-received BNi-5 will have been a non- N-21, as a N modified T phase is a logical, albeit
equilibrium or a quaternary phase. tentative, outcome of the Ni-Cr-Si phase diagram
In contrast, the isothermal sections due to studies.
Ansara et al., I3 shown in Fig. 1 predict that the It is clear that the prospects of improving the
commercial BNi-5, Ni-18Cr-19Si, lies on the edge joint characteristics by varying fabrication condi-
of a solid-solid two phase field between FCC Ni tions such as time and temperature are bleak.
and Ni,Si,. This is not confirmed by the new Diffusion barriers will need to be applied to the
isotherms shown in Fig. ‘12. In these, the BNi-5 ceramic surfaces to prevent Si and nitrogen ingress
composition lies close to the edge of the L, liquid when using Ni-Cr-Si braze alloys.
field at 1125°C and presumably will lie on the
actual edge of an expanded L2 liquid field at the
braze liquidus of 1135°C. Even more informa- 6 Conclusions
tively, the new diagram for 1050°C casts light on
the microstructure of the solidified braze seams Ni-Cr-Si braze alloys wet and adhere to Si,,N,
produced by the equilibrium of the brazes and the and in some cases spreading is observed. Despite
572 G. Ceccone, M. G. Nicholas, S. D. Peteves, A. A. Kodentsov, J. K. Kivilahti, F. J. J. van Loo

the large range of brazing conditions explored, the 5. Iwamoto, N., Kamay, M., Fujii, K. & Sato, N., ibid,
joint quality obtained was very poor. Unbonded p&l-51.
6. Mixuhara, H., Brazing alloy of copper, silver, titanium,
areas, cracks and voids were observed in almost aluminium, US Patent 5,087,416, February 1992.
all the samples. 7. Santella, M. L., Advan. Ceram. Mat., 3(5) (1988) 457-62.
Contact between the molten Ni-Cr-Si brazes 8. Lugsheider, E., Boretius, M. & Tillmann, W., DKG 68
No1/2 (1991) 14-22.
and S&N, caused ingress of Si and nitrogen from 9. Hadian, A. M., McDermid, J. R. & Drew, A. L., Ceram.
the ceramic. In the case of the commercial BNi-5 Trans., 19 (1991) 221-7.
alloy, the relatively ductile YNi matrix of the 10. Peteves, S. D., Moulaert, M. & Nicholas, M. G., Metall.
Trans. A, 23A (1992) 1773-81.
braze was converted to brittle Ni,Si, or Ni$i.
11. Moretto, P., Moulaert, M., Glaude, P., Frampton, P.,
Other changes caused by the ingress of Si were the Ceccone, G. & Peteves, S. D. In Proc. 2nd European Collo-
formation of a Ni-Cr-Si-N compound as inclu- quium on Design Ceramic Interfaces, Petten (NL), (1992),
sions or a reaction product layer. ed. S. D. Peteves, Commission of the European Communi-
ties, Luxembourg, EUR 15306 EN, 1993 p. 519-538.
The difficulties encountered when using the 12. Lugscheider, E., Knotek, 0. & Klohn, K., Thermochim.
Ni-Cr-Si alloys are related to fundamental mate- Acta, 29(2) (1979) 323.
rials science effects rather than inappropriate 13. Ansara, I., Chart, T. G., Chevalier, P. Y., Hack, K.,
McHugh, G., Rand, M. H. & Spencer, P. J., Report No
choice of brazing conditions. None of the EUR 9657/1/EN, (1985) Commission of the European
Ni-Cr-Si brazes can offer direct resistance to the Communities, Luxembourg, and references therein.
embrittling ingress of Si and nitrogen from the 14. Gupta, K. P., Rajan, N. S. & Beck, P. A., Trans. TMS-
AZME, 218 (1969) 617.
ceramic.
15. Kodentsov, A. A., Kivilahti, J. K., van Loo, F. J. J. &
Peteves, S. D., Ceram. Trans., 35 (1993) 13 516.
16. Kodentsov, A., Kivilahti, J. K., van Loo, F. J. J. &
Peteves, S. D., to be published.
Acknowledgements 17. Kodentsov, A. A., Gulpen, J. M., Kivilahti, J. V. & van
Loo, J. F., to be published.
18. Ono, N., Kajihara, M. & Kikuchi, M., Metall. Trans.
Part of this work was carried out within the Com- 23A (1992) 1389.
mission Research and Development Programme. 19. Frisk, K., Ph. D. Thesis, Royal Institute of Technology,
It was also partially supported by the Innovative Stockholm, Sweden, (1990).
20. Lugscheider, E., Knotek, 0 & Klohn, K., Weld. J., Weld.
Programme for Technical Ceramics from the
Res. Suppl., 57 (1978) 319s.
Dutch Finance Ministry (Grant # 91.B230). 21. Nicholas, M. G. & Peteves, S. D., In Proc. of&h Znterna-
tional Symposium on Ceramic Materials and Components
for Engines, eds. R. Carlsson, T. Johansson & L.
Kahlman, Elsevier Applied Science, London, 1992, p. 332.
22. Holmstrbm, M., Ljungberg, L., Warren, A. & Warren, R.,
References In Proc. of 4th International Symposium on Ceramic
Materials and Components for Engines, eds R. Carlsson,
McDermid, J. R. & Drew, R. A. L. J. Am. Cer. Sot, T. Johansson & L. Kahlman, Elsevier Applied Science,
74(8) (1991) 185560. London, 1992, p. 364-71.
McDermid, J. R., Pugh, M. D. & Drew, R. A. L. Metall. 23. Moorhead, A. J. & Keating, H., Report No ORNL-6262,
Trans., 20A (1989) 1803-10. Oak Ridge National Laboratory, Oak Ridge, TN, March
Mehan, R. L., Jackson, M. R., McConnell, M. D. & 1986.
Lewis, N., J. Mat. Sci., 18(2) (1983) 508-14. 24. Metal Handbook, 9th ed, (ASM), Ohio, 6 1983, 1014-21.
Foley, A. G. & Andrews, D. J., In Proc. 2nd European 25. Massalski, T. B., Binary Alloy Phase Diagrams, (ASM),
Colloquium on Design Ceramic Interfaces, Petten (NL), Ohio, 1986 p. 842 & 1756.
1992, ed. S. D. Peteves, Commission of the European 26. Nash, P. & Nash, A., Bull. All. Ph. Diugr., 8(l) 1987 613.
Communities, Luxembourg, EUR 15306 EN, 1993, p. 27. Iida, T. & Guthrie, R. I. L., The Physical Properties of
41 l-27. Liquid Metals, Clarendon Press, Oxford, 1988, p, 217.