The microstructural evolution of infrared brazed Fe
3
Al by
BNi-2 braze alloy
Y.L. Lee
a, R.K. Shiue
b, S.K. Wu
a,*
aDepartment of Materials Science and Engineering, National Taiwan University, Taipei 106, Taiwan bDepartment of Materials Science and Engineering, National Dong Hwa University, Hualien 974, Taiwan
Received 7 March 2002; accepted 23 May 2002
Abstract
The present work reports a novel approach in joining Fe3Al by infrared vacuum brazing using BNi-2 (7.0 Cr, 3.1 B, 4.5 Si, 3.0 Fe, 0.06C, and Ni balance) as the brazing filler alloy. Based on the related alloy phase diagrams and solidification theory, the species and morphology of phases in the brazed joint are extensively examined. Many transient phases, including: (Ni,Fe)3Al, (Ni,Fe)3(Si,Al), (Fe,Ni,Cr)3B and BCr, were observed after infrared brazing. With increasing the homogenization time of the brazement at 1000C, the stoichiometry of both (Ni,Fe)3Al and (Ni,Fe)3(Si,Al) phase has changed into (Ni,Fe)2Al and (Ni,Fe)2(Si,Al), respectively. Meanwhile, the amounts of the chromium boride and (Fe,Ni,Cr)3B phases are greatly decreased. Finally, a single-phase joint can be obtained if sufficient time is applied in homogenization treatment of the brazement.
#2003 Elsevier Science Ltd. All rights reserved.
Keywords:A: Iron aluminide; based on Fe3Al; B: Phase identification; D: Microstructure
1. Introduction
Iron aluminides including Fe3Al and FeAl have been
extensively studied because of their low cost, low den-sity, fairly good corrosion and oxidation resistance [1–3]. Considerable efforts have been focused on both materials processing issues and alloy design of the iron aluminides with modified mechanical and metallurgical properties for structural applications [1,4–20]. The development of a joining process is very important in the application of iron aluminides, but the study of welding and/or brazing iron aluminides is very limited in literatures[2,21–23]. Joining of iron aluminides plays an important role in practical application of such alloys. Welding of iron aluminides is difficult due to its inherent low temperature ductility and poor weldability [23]. Cold cracks can be initiated from the weld even for low energy input processes such as laser welding. Water vapor presence in the atmosphere has shown severe embrittlement of most iron–aluminide alloys during welding [23]. On the other hand, vacuum brazing pro-cessed under a high vacuum condition up to 5106
mbar provides an alternative bonding technology to join iron aluminides.
Infrared vacuum brazing makes use of infrared energy generated by heating a tungsten filament in a quartz tube as the heating source [24]. The infrared rays can transmit quartz tube and be focused on the specimen. The specimen is locally heated by infrared rays, and the rest of the furnace is not heated during brazing process. Therefore, infrared brazing is a highly potential process with the characteristics of rapid thermal cycle. It has been successfully applied in brazing of TiAl, NiAl and Ni3Al intermetallics[25–27].
Nickel base braze alloys both have good corrosion resistance and creep strength, so it was selected as filler metal. However, some nickel base braze alloys contain-ing phosphorus are not suitable due to the formation of brittle nickel phosphides. BNi-2 has high creep strength and low liquidus temperature among all Ni base braze alloys. Therefore, it was chosen as a brazing filler metal. The present work reports a new approach in joining Fe3Al by infrared brazing using the BNi-2 as brazing
filler alloy. The microstructural evolution of the as-brazed joint and the subsequently homogenized braze-ment will be extensively studied. With the aid of fast infrared heating rate, the early-stage of reaction kinetics in brazed Fe3Al is also discussed.
0966-9795/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0966-9795(02)00252-2
www.elsevier.com/locate/intermet
* Corresponding author. Tel.: 2363-7846; fax: +886-2-2363-4562.
2. Experimental procedures
The Fe3Al with the nominal composition Fe-28Al (in
at.%) was prepared by vacuum arc remelting of high purity ( > 99.99 wt.%) Fe and Al pellets. The final weight loss of the master alloy was below 0.1 wt.%. A nickel base filler alloy, Nicrobraz1
LM alloy tape made by Wall Colmonoy Co., was used as brazing filler alloy, and its chemical composition in weight percent was 7.0 Cr, 3.1 B, 4.5 Si, 3.0 Fe, 0.06C, and Ni balance. Based
on the AWS specification of Ni base braze alloys, the chemical composition of Nicrobraz1
LM braze is con-sistent with BNi-2 braze alloy [28]. The solidus and liquidus temperatures of this alloy are 970 and 1000C,
respectively. The thickness of Nicrobraz1
LM tape was 100 mm throughout the experiment.
Compared with traditional furnace brazing, infrared brazing is featured with rapid thermal cycle. Unlike furnace brazing, higher infrared brazing temperature will have less damage to the base metal. Additionally,
Fig. 1. SEM observations of (a) Nicrobraz1LM alloy tape prior to infrared brazing, (b) Fe
high brazing temperature can greatly speed up the microstructural evolution of the brazed joint. Therefore, a temperature 150C higher than the liquidus of braze
alloy was chosen as brazing temperature. Infrared brazing was performed in a vacuum of 5105mbar at
1150 C for 2, 24 and 240 s in the study. In order to
study the phase stability in the brazed joint, some infrared brazed specimens were subsequently annealed in a vacuum furnace at 1000C for 24, 48, 72 and 144 h,
respectively. All joined surfaces were firstly polished by SiC papers up to grit 600, and subsequently ultra-sonically cleaned by acetone prior to infrared brazing. The area of the LM foil was approximately the same as that of base metal. To enhance the absorption of infra-red rays, a graphite fixture was used during brazing. Specimens were sandwiched between two graphite plates. A thermal couple was inserted into the upper graphite plate, and contacted with the brazed specimen. The thickness of Fe3Al substrate is 1 mm, and the total
thickness of Fe3Al/BNi-2/Fe3Al joint is about 2 mm.
Because the thickness of brazed joint is not very thick, it is reasonable to assume that the temperature gradient in the specimen is negligible. An ULVAC SINKO-RIKO RHL-P610C infrared furnace with the heating rate of 900C/min was used throughout the experiment. There
is a time delay between the actual specimen temperature and programmer temperature, so time compensation is necessary in the experiment. The brazing time specified in the test is the actual specimen holding time in the experiment[24].
The brazed sample was cut by a low speed diamond saw, and followed by a standard metallographic pro-cedure. The cross section of the brazed specimens was examined using a Philips1
XL30 scanning electron
microscope (SEM) with an accelerating voltage of 15 kV. Quantitative chemical analysis was performed by using a JEOL JXA-8800M electron probe micro-analyzer (EPMA) equipped with a wavelength dis-persive spectrometer (WDS). Its spot size is 1 mm, and its operation voltage is 15 kV.
3. Results and discussion
Fig. 1(a)displays the secondary electron image (SEI) of the Nicrobraz1
LM alloy tape prior to brazing. It shows the morphology of brazing powders. Fig. 1(b) shows the backscattered electron image (BEI) of SEM observations for Fe3Al/BNi-2/Fe3Al specimen brazing
at 1150 C for 2 s. The SEI displays topographic
con-trast of the cross section in the joint. The BEI does not provide topographic contrast but primarily shows the element distribution in the brazed joint [29]. The mor-phology of braze alloy after infrared brazing is very different from that of the original powder. It demon-strates the melting of braze alloy powders during infra-red brazing. According to Fig. 1(b), the element distribution in the joint is not quite uniform, and many phases can be observed in the joint. Fig. 2 shows the BEI of the infrared brazed specimen at 1150C for 24 s.
It is noted that huge difference is found in comparison betweenFigs. 1(b) and 2. According to chemical analy-sis results, at least two major phases can be identified in Fig. 2, one is the matrix and the other is the gray streak-like boride phase.
Fig. 3 shows the SEM BEI observation and EPMA
chemical analysis results of the Fe3Al/BNi-2/Fe3Al
specimen brazed at 1150 C for 240 s. Very limited
oxygen content is detected in the EPMA analysis, so it implies that there is little oxidation during infrared brazing. Avoiding oxidation during brazing is a pre-requisite to obtaining a good bonding. At least five phases can be observed in the analysis. First, the base metal as marked by 1 in Fig. 3 is primarily Fe3Al
sub-strate solid solution with minor Ni due to the high brazing temperature. A continuous layer close to the interface of Fe3Al/BNi-2 is shown inFig. 3as marked
by 2 in the SEM BEI photograph. The phase is mainly comprised of Al, Fe and Ni. Based on the stoichiometric ratio of Ni, Fe and Al in this phase, a (Ni,Fe)3Al phase
is suggested in the analysis. Both Ni3Al and Fe3Al
pha-ses can be found in related binary alloy phase diagrams, and Fe can be completely dissolved into Ni at elevated temperature[30]. Therefore, it is reasonable to deduce the formation of (Ni,Fe)3Al phase after infrared
braz-ing. There are three different phases in the central part of the joint. The small black spots marked by 5 in the figure are identified as chromium boride, BCr. It is alloyed with other minor elements, e.g. Fe, Ni, Si and C. Similarly, the gray matrix in the middle part of the joint marked by 3 in the figure can be identified as (Fe,Ni)3
(Si,Al) according to the stoichiometric ratio of Fe, Ni, Si
and Al. Finally, the dark gray streak-like phase marked by 4 inFig. 3contains B, Cr, Fe and Ni. It is basically a type of boride, and its stoichiometric ratio is close to (Fe,Ni,Cr)3B phase.
To unveil the transient evolution of the brazed microstructure, a multi-component equilibrium phase diagram is preferred in the study. However, there is no such phase diagram currently available. Therefore, it is essential to consult related ternary alloy phase diagrams [31]. The chemical composition of the LM braze in atomic percent is 6.6 Cr, 14.1 B, 7.9 Si, 2.6 Fe and Ni balance, and the base metal consists of Fe and Al.Fig. 4 displays the liquidus projection of B–Cr–Ni ternary
alloy phase diagram in atomic percent [31]. Many important invariant reactions are included in this figure. It is well known that silicon acts as a melting point depressant in BNi-2 braze alloy. According to EPMA chemical analysis as shown inFig. 3, BCr exists in the braze alloy after infrared brazing. The melting point of BCr is 2100 C, and the eutectic reaction of a molten
braze can explain its formation. On the other hand, select B–Ni–Si ternary phase diagram cannot explain the formation of BCr after infrared brazing. Conse-quently, the use of B–Cr–Ni ternary alloy phase dia-gram is important in elucidating the solidification of the braze alloy. Although Si is not included in the B–Cr–Ni
ternary alloy phase diagram, the B–Cr–Ni phase dia-gram can still provide a first approximation of the phase evolution during infrared brazing.
According to Fig. 4, there are three ternary eutectic reactions, E1, E2and E3, in the diagram. The chemical
composition of the molten BNi-2 braze is located in the primary field of BNi3 and closes to that of eutectic
liquid at E2inFig. 4. The ternary eutectic temperature
at E2 in B–Cr–Ni phase diagram is 1050 C, which is
lower than the brazing temperature (1150 C) in the
experiment. Three phases are expected to form after the
eutectic reaction, including BCr, BNi3 and Ni phase.
According to the EPMA chemical analysis shown in Fig. 3, BCr is readily observed after brazing. It is also noted that Fe in the base metal can be dissolved into the molten braze. Consequently, BNi3is alloyed with Fe. A
new phase (Fe,Cr,Ni)3B is formed after infrared
braz-ing. The dissolution of Fe and Al elements from Fe3Al
substrate into the molten braze may result in isothermal solidification of the BNi-2 braze alloy during brazing. The Ni phase may react with Fe and Al dissoluted from Fe3Al matrix. Therefore, no pure Ni phase is observed
Fig. 5. The SEM BEI observation and EPMA chemical analysis results of Fe3Al/BNi-2/Fe3Al specimens brazed at 1150C for 240 s and followed
in the joint. The formation of both (Ni,Fe)3Al close
to the interface and (Ni,Fe)3(Si,Al) in the middle of
joint has been observed in this study. According to current experimental results, the kinetics of the soli-dification path is consistent with the related alloy phase diagrams.
In order to verify the phase stability in the brazed joint, some infrared brazed specimens were sub-sequently homogenized in a vacuum furnace at 1000C
for 24, 48, 72 and 144 h, respectively.Fig. 5shows SEM BEI observation and EPMA chemical analysis results of Fe3Al/BNi-2/Fe3Al specimens brazed at 1150C for 240
s and followed by annealing at 1000C for 24 h.
Com-pared with Fig. 3, the width of brazed joint is greatly decreased due to high temperature homogenization of the joint. The chemical composition of base metal is close to Fe3Al as marked by 1 inFig. 5. The initial
as-brazed (Ni,Fe)3Al phase has changed its stoichiometry
into (Ni,Fe)2Al phase as marked by 2 in Fig. 5.
Simi-larly, the initially as-brazed (Ni,Fe)3(Al,Si) phase
chan-ges its chemical composition close to (Ni,Fe)2(Al,Si) as
marked by 3 in Fig. 5. Two borides, BCr marked by 5 and (Fe,Cr,Ni)3B marked by 4, are still observed in the
figure. It is reasonable that the amount of BCr in the
joint is greatly decreased due to the homogenization of the specimen. Meanwhile, both the morphology and the amount of (Fe,Cr,Ni)3B phase are also changed. The
width of the joint is greatly decreased. With increasing homogenization time of the infrared brazed specimen, the width of the joint is consistently decreased as shown inFig. 6(a)–(c). If the brazed specimen is homogenized at 1000 C for 120 h, most phases near interface are
disappeared. Most of BCr and (Fe,Cr,Ni)3B phases are
dissolved into the Fe3Al matrix. Consequently, a
single-phase joint can be formed if sufficient time is applied in the homogenization treatment.
The interdiffusion between the braze alloy and base material resulting in isothermal solidification may play a crucial role in determining the final microstructure of the joint. The melting point depressants, e.g. boron and silicon, in BNi-2 braze alloy can fast diffuse into Fe3Al
substrate, and Fe3Al will also dissolute into the molten
braze. Both metallurgical phenomena can increase the melting point of the molten braze. Therefore, it is reported that isothermal solidification takes place dur-ing traditional furnace brazdur-ing usdur-ing Ni–B–Si filler metal as a result of boron removal from the filler metal [32–34]. A simple x ¼ ðDtÞ12estimation can indicate
whe-Fig. 6. The SEM BEI observations of Fe3Al/BNi-2/Fe3Al specimens brazed at 1150C for 240 s and subsequently followed by 1000C
ther isothermal solidification is possible for these experimental conditions. The diffusion distance (x) is equal to the square root of diffusion coefficient (D) times the time period of diffusion (t). The diffusion coefficient of boron in Fe3Al at 1150 C is
approxi-mately 4.11011 cm2/s [35]. The estimated diffusion
distance of B into Fe3Al at 1150 C for 240 s is only
about 1 mm. Consequently, the isothermal solidification due to the removal of B from braze alloy is very limited in infrared brazing due to its fast thermal cycle. How-ever, the dissolution rate of Fe3Al into the molten braze
at elevated temperatures needs further study in order to estimate the possibility of isothermal solidification dur-ing infrared brazdur-ing.
The subsequent homogenization treatment of braze-ment results in interdiffusion of elebraze-ments between sub-strate and the solidified braze alloy. However, the diffusivity of B is much higher than that of other ele-ments. It will cause nonsymmetrical mass transport during interdiffusion. As demonstrated inFig. 7, some Kirkendall porosity (about 1–2 microns) close to the braze is observed. The Kirkendall effect describes that the vacancy will be formed if the rate of interdiffusion is not balanced[36]. Boron is a very small atom, and it can diffuse much faster than Al, Si, Cr, Ni and Fe atoms with a larger atomic radius. Therefore, boron can fast diffuse into substrate driven by concentration gradient, but much slower diffusion rates of Al and Fe atoms into the solidified braze are expected. Consequently, vacancy will be formed after heat treatment.
4. Conclusion
A novel approach in joining Fe3Al by infrared
vacuum brazing using BNi-2 as the brazing filler alloy is performed in the study. The species and morphology of phases in the brazed joint are extensively examined. Many transient phases, including: (Ni,Fe)3Al, (Ni,Fe)3(Si,Al),
(Fe,Ni,Cr)3B and BCr, were observed after infrared
brazing. With increasing the homogenization time of the brazement at 1000 C, the stoichiometry of both
(Ni,Fe)3Al and (Ni,Fe)3(Si,Al) phase has changed into
(Ni,Fe)2Al and (Ni,Fe)2(Si,Al). Most of BCr and
(Fe,Cr,Ni)3B phases can be dissolved into the Fe3Al
matrix during homogenization of the brazement, so the amounts of the chromium boride and (Fe,Ni,Cr)3B
phases are decreased with the time increment of homo-genization. Finally, a single-phase joint can be obtained if sufficient time is applied in homogenization treatment of the brazement.
Unlike traditional furnace brazing, the interdiffusion between the braze alloy and base metal is very limited during infrared brazing due to its rapid thermal cycle. The subsequent homogenization treatment of braze-ment results in interdiffusion of elebraze-ments between the substrate and the solidified braze alloy. Since the dif-fusivity of B is much higher than that of other ele-ments, it will cause nonsymmetrical mass transport during interdiffusion. Therefore, fine Kirkendall poros-ity (about 1–2 microns) close to the braze is observed in the experiment.
Acknowledgements
The authors gratefully acknowledge the financial support from National Science Council (NSC), Repub-lic of China, under the grants NSC 89-2216-E002-002. EPMA analysis by Ms. Shu-Yueh Tsai in NSC Instru-ment Center, National Tsing Hua University, Hsinchu, Taiwan, is also gratefully acknowledged.
References
[1] Liu CT, George EP, Maziasz PJ, Schneibel JH. Mat Sci Eng A-Struct A258 1998:84.
[2] Stoloff NS. Mat Sci Eng A-Struct A258 1998:1.
[3] Stoloff NS, Sikka VK. Physical metallurgy and processing of intermetallic compounds. New York, NY: Chapman and Hall; 1996.
[4] Davies RG, Stoloff NS. Acta Metall 1963;11:1187.
[5] Liu CT. High-temperature order intermetallic alloys. In: Baker I, Darolia R, Whittenberger JD, Yoo MH, editors. V. MRS Symp. Proc. 288. Pittsburgh, USA: MRS; 1993. p. 3.
[6] Agarwal A, Dahotre NB. Scripta Mater 2000;42:493. [7] Deevi SC. Intermetallics 2000;8:679.
[8] Sundar RS, Baligidad RG, Prasad YVRK. Mat Sci Eng A-Struct A258 1998:219.
[9] Blackford JR, Buckley RA, Jones H, Sellars CM. J Mater Sci 1998;33:4417.
[10] Hajaligol MR, Deevi SC, Sikka VK. Mat Sci Eng A-Struct A258 1998:249.
[11] Chakraborty SP, Sharma IG, Bose DK. J Alloy Compd 1998; 280:255.
[12] McKamey CG, Maziasz PJ. Intermetallics 1998;6:303. [13] Baligidad RG, Prakash U, Rao VR. ISIJ Int 1996;36:1448.
[14] Subramanian R, Schneibel JH. JOM-J Min Met Mat S 1997; 49:50.
[15] Agwarwal A, Akhtar MJ, Balasubramaniam R. J Mater Sci 1996; 31:5207.
[16] Deevi SC, Sikka VK, Liu CT. Prog Mater Sci 1997;42:177. [17] Mistler RE, Sikka VK, Scorey CR, McKernan JE, Hajaligol MR.
Mat Sci Eng A-Struct A258 1998:258.
[18] Blackford JR, Buckley RA, Jones H, Sellars CM. Scripta Mater 1996;34:721.
[19] Balasubramaniam R. J Alloy Compd 253- 1997;4:148. [20] Balasubramaniam R. Scripta Mater 1996;34:127.
[21] Jung SB, Minamino Y, Yamane T, Saji S. J Mater Sci Lett 1993; 12:1684.
[22] Fasching AA, Edwards GR, David SA. Sci Technol Weld Joi 1997;2:167.
[23] Fasching AA, Ash DI, Edwards GR, David SA. Scripta Metall 1995;32:389.
[24] Shiue RK, Wu SK. O JM, Wang JY Metall Mater Trans A 2000; 31A:2527.
[25] Lee SJ, Wu SK, Lin RY. Acta Mater 1998;46:1283. [26] Lee SJ, Wu SK, Lin RY. Acta Mater 1998;46:1297. [27] Yang TY, Wu SK, Shiue RK. Intermetallics 2001;9:341. [28] Olson DL, Siewert TA, Liu S, Edwards GR. ASM handbook
(Vol. 6). Materials Park, OH: ASM International; 1993. [29] Lee RE. Scanning electron microscopy and X-ray microanalysis.
1993
[30] Massalski TB. Binary alloy phase diagrams. Materials Park, OH: ASM International; 1990.
[31] Villars P, Prince A, Okamoto H. Handbook of ternary alloy phase diagrams. Metals Park, OH: ASM International; 1995. [32] Kubachewski O, Alcock CB, Spencer PJ. Materials
thermo-chemistry. New York, NY: Pergamon Press; 1993.
[33] Shyam A, Suwas S, Bhargava S. Prakt Metallogr 1997;34:264. [34] Gale WF, Wallach ER. J Mater Sci 1992;27:5653.
[35] Hasaka M, Morimura T, Kondo S-I. Scripta Metall Mater 1993; 29:967.
[36] Humpston G, Jacobson DM. Principles of soldering and brazing. Metals Park, OH: ASM International; 1993.
![Fig. 4. Liquidus projection of B–Cr–Ni phase diagram [31].](https://thumb-ap.123doks.com/thumbv2/9libinfo/8669079.195806/5.892.126.772.352.1077/fig-liquidus-projection-of-cr-ni-phase-diagram.webp)
