Solidification of Nb-Bearing Superalloys: Part I. Reaction Sequences

Solidification of Nb-Bearing Superalloys: Part I. Reaction Sequences J.N. DuPONT, C.V. ROBINO, J.R. MICHAEL, M.R. NOTIS, and A.R. MARDER The solidification reaction sequences of experimental superalloys containing systematic variations in Fe, Nb, Si, and C were studied using differential thermal analysis (DTA) and microstructural characterization techniques. The reaction sequences responsible for microstructural development were found to be similar to those expected in the Ni-Nb-C ternary system and commercial superalloys of comparable composition. The solute-rich interdendritic liquid generally exhibited two eutectic-type reactions at the terminal stages of solidification: L ( NbC) and L ( Laves). The Nibase alloys with a high C/Nb ratio represented the only exception to this general solidification sequence. This group of alloys terminated solidification with the L ( NbC) reaction and did not exhibit the /Laves constituent. At similar levels of solute elements (Nb, Si, and C), the Fe-base alloys always formed more of the /Laves eutectic-type constituent than the corresponding Ni-base alloys. Silicon additions also increased the amount of the /Laves constituent that formed in the assolidified microstructure, while C additions promoted formation of /NbC. The influence of Nb was dependent on the C content of the alloy. When the C content was low, Nb additions generally promoted formation of /Laves, while Nb additions to alloys with high C led to formation of the /NbC constituent. The results of this work are combined with quantitative analyses for developing -Nb-C pseudoternary solidification diagrams in a companion article. I. INTRODUCTION NICKEL-BASE Nb-strengthened alloys represent a significant portion of the superalloys currently in use. Commercial examples include alloys IN625, IN706, IN718, IN903, IN909, and Thermo-Span, to name a few. These alloys are often used for components that are fabricated by solidification processes such as casting and fusion welding. They also find use in dissimilar weld applications such as claddings on steel components subjected to highly aggressive wear and corrosion conditions. [1,2] In these applications, the fusion zone composition can become significantly enriched in Fe and C due to dilution from the carbon steel substrate, and this composition modification can alter microstructural development during solidification. [3,4] Previous studies conducted on commercial superalloys [5,6] have shown that minor variations in Nb, Si, and C strongly affect the type and amount of secondary phases that form during the terminal stages of solidification. In particular, two types of secondary phases, NbC and Laves, are known to form, and the wide range of possible solidification behavior has been shown to control the susceptibility for fusion zone solidification cracking. [3,7,8] Although detailed investigations have been conducted on specific commercial alloys and dissimilar alloy combinations, no general solidification model has been developed to provide more quantitative prediction capabilities of microstructural evolution during solidification. In the current article, reaction sequences are established for J.N. DuPONT, Research Scientist and Associate Director, Energy Liaison Program, and M.R. NOTIS and A.R. MARDER, Professors, Materials Science and Engineering Department, are with Lehigh University, Bethlehem, PA 18015. C.V. ROBINO and J.R. MICHAEL, Principal Members of the Technical Staff, are with Sandia National Laboratories, Albuquerque, NM 87185. Manuscript submitted July 14, 1997. multicomponent (Fe, Ni, Cr, Nb, Si, and C) alloys. These alloys contain systematic variations in Fe, Nb, Si, and C, and simulate the wide range in solidification behavior observed in commercial alloys and Fe-enriched fusion zones of dissimilar welds made between Nb-bearing Ni-base alloys and steel substrates. In companion articles, this information is applied to develop pseudoternary solidification surfaces, [9] which have been included in a solute redistribution model for making semiquantitative predictions of microstructural evolution during solidification. [10] II. EXPERIMENTAL PROCEDURE A. Experimental Alloy Compositions A four-factor, two-level set of experimental alloys was designed to simulate the commercial compositions of interest in this study. The alloy compositions are summarized in Table I. The alloys contain factorial variations in Fe (in exchange for Ni), Nb, Si, and C at two levels. The high and low target levels of Nb, Si, and C are set as follows (all values in wt pct): 2 Nb 5, 0.02 C 0.15, and 0.10 Si 0.60. These limits were chosen to represent low and high composition values of wrought alloys and filler metals as well as composition limits that can arise in fusion welds made between nickel-base alloys and carbon steels. Several additional alloys with intermediate C contents (alloys 1.5, 3.5, 7.5, and 11.5) were also investigated. The alloys were prepared at Sandia National Laboratories by investment casting. All samples were melted and poured in vacuum to produce six bars of each composition with approximate dimensions of 170 25 6.3 mm. The ascast samples were machined into Varestraint samples of 165 25 3.2 mm dimensions. Welds on these specimens were prepared using the gas tungsten arc (GTA) process at a current of 95 amperes, 2.5-mm arc gap (10 V), and 3.3 METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, NOVEMBER 1998 2785

Table I. Alloy Compositions* Alloy Fe Ni Cr Nb Si C P S 1 1.5 2 3 3.5 4 5 6 7 7.5 8 Ni-l 9 10 11 11.5 12 13 14 15 16 Fe-1 10.49 10.75 11.12 10.70 10.39 10.72 10.84 10.88 10.70 10.82 10.80 7.71 46.03 46.69 45.38 47.38 45.28 44.55 44.05 45.40 44.47 40.87 68.53 67.95 68.20 68.11 66.80 67.60 65.79 65.22 65.53 63.93 64.96 55.03 33.56 32.80 32.80 31.05 32.39 31.24 31.93 30.03 30.89 27.89 18.90 19.21 19.12 19.02 19.29 19.08 18.98 18.89 19.30 18.54 18.90 18.23 19.31 19.70 19.53 19.64 19.89 19.63 19.52 19.54 19.45 19.69 *All values are in weight percent. ND Not determined. 1.93 2.00 1.95 1.82 1.94 1.91 5.17 4.87 4.86 4.92 4.72 17.79 1.66 1.66 1.77 1.84 1.93 4.42 4.51 4.88 4.77 10.63 0.08 0.03 0.06 0.38 0.41 0.40 0.05 0.08 0.52 0.46 0.52 1.03 0.10 0.01 0.57 0.67 0.61 0.02 0.08 0.66 0.64 0.63 0.017 0.052 0.132 0.010 0.075 0.155 0.013 0.161 0.010 0.081 0.170 0.08 0.108 0.116 0.079 0.015 0.210 0.010 0.216 0.081 0.005 0.005 0.005 0.005 0.005 ND 0.007 0.007 ND 0.002 0.001 0.010 0.007 0.009 0.007 ND 0.002 0.002 0.001 0.002 0.002 0.002 ND mm/s travel speed. The Varestraint work was conducted as part of a separate study on the weldability of these alloys, which will be reported in a future article. Two high-nb alloys (Ni-1 and Fe-1) were also added to produce alloys with high quantities of eutectic-type constituents. Small cast buttons of these two alloys were produced from high-purity powders ( 99.99 wt pct) by GTA melting under an argon atmosphere. The nominal Cr content was selected at 20 wt pct, which is typically used in many commercial alloys for good corrosion resistance. B. Differential Thermal Analysis The temperatures of eutectic-type reactions were measured using differential thermal analysis (DTA). Differential thermal analysis was conducted on a Netzsch STA 409 differential thermal analyzer using 500- to 550-mg samples. The DTA system was calibrated to within 2 C by using a pure Ni standard (melting point 1455 C). Samples were melted and solidified under flowing argon in alumina crucibles using pure Ni as the reference material. Preliminary tests were first conducted to establish the liquidus temperature of each alloy. After these preliminary tests, the liquidus and solidus temperatures were determined by heating samples slowly at a rate of 5 C/min to approximately 10 C above their liquidus temperature (the liquidus and solidus data are discussed in a companion article [9] ). The samples were then solidified at a relatively fast cooling rate of 20 C/min to determine temperatures of eutectic-type reactions, which occur under nonequilibrium solidification conditions. In agreement with previously published work on similar alloy systems, [3,5,7,14,15,19] reaction temperatures were taken as deviations from the local baseline. In relatively few cases where overlap occurred between two neighboring peaks, the reaction start temperatures were taken as the inflection point between the peaks. This identification scheme is consistent with previous DTA data on commercial alloys of similar composition. [19] All reported reaction temperatures are an average from at least three tests, and the order of testing was randomized. C. Microstructural Characterization Light optical microscopy (LOM) was conducted on DTA samples polished through 0.04- m colloidal silica and electrolytically etched in a 10 pct chromic acid 90 pct water solution at 3 V. Scanning electron microscopy was conducted on autogenous weld metals (prepared using the same preparation procedure) using a JEOL* 6300 field emission *JEOL is a trademark of Japan Electron Optics Ltd., Tokyo. gun scanning electron microscope (FEG SEM) at an accelerating voltage of 15 kv. Quantitative image analysis (QIA) was conducted on the autogenous GTA welds at locations far removed from the solidification cracks that were produced during Varestraint testing. Area fractions of total eutectic-type contents and, where possible, individual /NbC and /Laves eutectic-type constituents were measured along the centerline of each weld with at least ten SEM photomicrographs. Area fractions were assumed to be equivalent to volume fractions. Electron probe microanalysis (EPMA) was conducted on a JEOL 733 probe equipped with four wavelength dispersive spectrometers. The EPMA was conducted at an accelerating voltage of 15 kv and beam current of 20 na. All EPMA samples were mounted in thermal setting epoxy, polished flat to a 0.3- m finish using an alumina slurry, ultrasonically cleaned in acetone, and carbon coated prior to analysis. The K lines were used for Fe, Ni, Cr, and Si, while the L line was used for Nb. Raw data were reduced to weight percentages using a ZAF algorithm. [11] The crystal structure of secondary phases in welds of alloys 6 and 13 were determined by collecting backscattered electron kikuchi patterns, the details of which are described elsewhere. [12] III. RESULTS AND DISCUSSION A. Solidification Microstructures Typical solidification microstructures of the weld metals are presented in Figure 1. These SEM photomicrographs were taken in the vicinity of solidification cracks produced during Varestraint testing. In addition to the primary austenite dendrites, the alloys from one or both of two eutectictype constituents during solidification. The microstructures of DTA samples (Figure 2) and weld metals at locations remote from solidification cracks (Figure 3) showed similar features. The high-nb alloys (Ni-1 and Fe-1) exhibit these constituents in very high amounts (Figure 4). Due to the relatively slow cooling rates of the DTA specimens (0.33 C/s), the secondary phases that form in the eutectic-type constituents of these samples exhibit regions that are 2 to 5 m in diameter and can thus be analyzed by EPMA. Typical compositions of the first type of secondary phase (e.g., labeled NbC in Figure 2) present in DTA samples of several alloys are summarized in Table II. (All reported EPMA measurements are the average and standard deviation from at least five measurements.) The 2786 VOLUME 29A, NOVEMBER 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 1 SEM photomicrographs of microstructures in the vicinity of solidification cracks produced during Varestraint testing for (a) and (b) alloy 2, (c) and (d) alloy 8, and (e) and ( f ) alloy 16. Nb content of this phase varies between 83.6 and 89.5 wt pct Nb and is consistent with niobium carbide (NbC), which has a stoichiometry range of approximately 88 to 91 wt pct Nb in the binary Nb-C system. [13] The slightly lower values of Nb measured in the NbC here may be due to the presence of other metal elements (Fe, Ni, and Cr), which possibly substitute for Nb. Backscattered electron kikuchi patterns collected from this phase in the weld metal of alloy METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, NOVEMBER 1998 2787

Fig. 2 LOM photomicrographs of DTA microstructures for (a) alloy 2, (b) alloy 8, and (c) alloy 16. 6 revealed an fcc crystal structure, which is also consistent with the NbC phase. [13] Typical EPMA measurements from the secondary phase in the other eutectic-type constituent are summarized in Table III. Comparison is made to the Laves phase identified by EPMA and transmission electron microscopy in commercial superalloys by other investigators. [14,15] The Nb content of this phase varies between 29.1 and 36.4 wt pct, which is similar to the Nb content observed in the Laves phase of commercial alloys of 29.0 to 34.1 wt pct Nb. Backscattered electron kikuchi patterns collected from this phase in several weld metals revealed a hexagonal crystal structure, which is also consistent with the Laves phase. The total volume percentages of eutectic-type constituents ( /NbC /Laves) and individual volume percentages of each eutectic-type constituent are summarized in Table 2788 VOLUME 29A, NOVEMBER 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 3 (a) and (b) SEM photomicrographs of weld metal microstructure in alloy 16 far removed from solidification cracks. Fig. 4 LOM photomicrographs of (a) alloy Ni-1 and (b) alloy Fe-1. IV and shown in graphical form in Figures 5 and 6. Confident phase identifications could not be made on alloy 1. The /NbC was present in alloy 1.5 along with small particles, which may have been Laves, but positive identification was not made. In the low-nb/low-c alloys (alloys 1, 1.5, 3, 3.5, 9, and 11), the secondary phases do not always form a well-defined lamellar structure with the eutectic. Thus, only the NbC or Laves portion of the eutectic-type structure is often included in the measurement. As a result, the reported data for these alloys represent a lower bound to the actual total eutectic-type volume percentages. For total volume percentage measurements (Figure 5), Fe- and Ni-base alloys with similar levels of Nb, Si, and C are matched for comparison. The results from individual constituent measurements, where such distinctions were possible (Figure 6), are aligned to make similar comparisons. Some of the comparisons should be considered semiquantitative, since, as noted in Figure 6, NbC and Laves phases were present at levels too low to be accurately quantified in some of the alloys. There are several important effects of Fe, Nb, Si, and C displayed in Figure 5 and 6. First, at similar levels of minor alloy additions (Nb, Si, and C), the iron-base alloys generally form higher levels of total eutectic-type constituents. Measurements of the individual eutectic-type constituents (Figure 6) indicate that this response can be attributed primarily to the formation of higher amounts of /Laves in the Fe-base alloys. It is worth noting that no Laves phase exists in the simple Ni-Nb binary phase diagram, while the METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, NOVEMBER 1998 2789

Table II. Compositions of NbC in DTA Samples (Metal Elements Only)* Alloy Fe Ni Cr Nb 3.5 6 8 9 14 16 0.5 0.1 0.4 0.1 0.5 0.1 1.3 0.1 1.8 0.9 1.2 0.2 3.3 0.1 2.9 0.3 3.0 0.7 1.1 0.1 1.2 0.6 1.1 0.2 *All values are in weight percent. Table III. 1.8 0.1 1.1 0.2 1.3 0.5 1.7 0.4 1.1 0.5 0.9 0.2 83.6 0.8 87.5 2.0 86.9 1.5 87.7 0.4 89.5 0.3 88.0 0.5 Composition of Laves Phase Identified in DTA and GTA Melt Alloys* Alloy Fe Ni Cr Nb Si Co Mo Ti 5 2.8 63.1 4.9 29.1 0.1 (0.3) (0.4) (1.0) (1.3) (0.1) Ni-l 5.5 46.5 12.6 33.7 1.7 (0.2) (1.0) (0.1) (1.2) (0.2) 13 25.9 25.3 11.1 35.7 0.5 (1.5) (0.8) (0.1) (1.1) (0.1) 14 25.5 26.0 12.2 36.0 0.7 (1.5) (0.4) (1.2) (0.3) (0.1) 15 26.5 (0.9) 24.1 (1.5) 11.3 (0.1) 36.4 (0.4) 1.8 (0.2) 16 23.9 26.1 11.5 36.0 2.6 (0.4) (0.6) (0.2) (0.2) (0.1) Fe-l 25.5 23.9 12.5 36.1 2.0 T-span 18.1 18.4 2.4 34.1 1.1 25.7 1.5 Ref. 16 IN909 20.2 30.5 33.0 2.2 11.2 2.1 Ref. 17 IN718 Ref. 17 10.6 39.4 11.2 29.0 1.7 7.5 0.8 *All values are in weight percent. Fe-Nb system forms the Laves phase over a composition range of 38 to 50 wt pct Nb and a broad temperature range of 600 C to 1400 C. [16] The Cr-Nb system also forms the Laves phase over a similar composition range (45 to 52 wt pct Nb at 1000 C). [16] The Ni-Nb system forms Ni 3 Nb over the composition range of approximately 33 to 38 wt pct Nb (at 1200 C). The addition of Fe to Ni-Nb alloys is well known to promote Laves phase at the expense of the Ni 3 Nb phase. [17] Thus, Fe additions to the austenite matrix, at the expense of Ni, can be expected to promote the formation of more /Laves, as observed here. The addition of Si has a similar effect, as the amount of /Laves always increases with increasing Si within a given matrix composition (Figure 6, compare alloy 5 to 7, 6 to 8, 13 to 15, and 14 to 16). The Laves promoting tendency of Si has also been documented in superalloys. [18] The Nb additions promote higher amounts of total eutectic. This is to be expected, since each of the secondary phases that form in the eutectictype structures (NbC and Laves) are both highly enriched in Nb. At high C levels, Nb additions generally promote more /NbC (compare alloy 2 to 6, 4 to 8, 10 to 14, and 11.5 to 16). When the C and Fe levels are both high, Nb additions will increase the amounts of both the /NbC and /Laves constituents (compare alloys 10 to 14 and 11.5 to 16). Alloys with high C levels (even-numbered alloys) form large amounts of the /NbC eutectic-type constituent. The presence of NbC and Laves has been observed in a large number of commercial superalloys, including alloys IN625, [5] IN718, [19] IN909, [15] IN903, [20] IN519, [6] Thermo- Span, [14] and IN625 weld overlays. [3,4] Alloys IN718, IN909, and Thermo-Span all contain 5 wt pct Nb, with relatively low C levels of to 0.05 wt pct, and exhibit the Laves phase as the predominant secondary solidification constituent, with NbC typically forming in smaller quantities. This is very similar to the results obtained here (e.g., alloys 5, 7, 13, and 15). IN903 contains 3 wt pct Nb, 0.03 wt pct C, and 0.07 wt pct Si. At this composition, NbC is observed as the major secondary phase, with Laves forming in trace amounts. Considering the low Si and intermediate C content, it is not surprising to find NbC as the major secondary phase in this alloy. Cieslak et al. [5] conducted a detailed investigation on the effects of Si and C additions to IN625. As found in this work, Si additions promoted the Laves phase while C promoted NbC. In one case, the addition of C led to the formation of only the /NbC constituent and no Laves formed at all. A similar effect was observed in alloys 2, 3.5, and 4 in this work. The experimental alloys employed in this research simulate the wide range of behavior displayed by individual commercial alloy systems. Thus, the systematic variations in composition that are used here to develop a solidification model should be directly applicable to commercially important superalloys. B. Solidification Reaction Sequences Typical DTA solidification traces, which show the range of behavior exhibited by all the alloys, are shown in Figure 7. Figure 7(a) shows a scan acquired on alloy 3. On cooling from above the liquidus temperature, a large exothermic peak is initiated at 1408 C, which is associated with solidification of the primary austenite dendrites. (It should be noted that the on-cooling DTA scans can contain some undercooling. More accurate measurements of liquidus temperatures are reported for these alloys from heating curves elsewhere. [9] ) Although a small amount (1.2 0.7 vol pct) of eutectic-type constituent formed in this alloy, no secondary peak could be discerned in the DTA scan. This behavior was typical for alloys containing less than 3 vol pct total eutectic-type constituent (e.g., alloys 1, 1.5, 3, 5, 7, 9, and 11). At these low quantities of eutectic-type constituents, there is apparently insufficient energy released during the reaction to produce a temperature difference between the standard and sample, which can be detected by the DTA system. Thus, terminal solidification reaction temperatures could not be determined for these alloys. Figure 7(b) shows a DTA scan of alloy 4. In addition to the large exothermic peak from the primary austenite dendrites, a distinct secondary exothermic peak is initiated at 1331 C. The peak is rather broad, suggesting the corresponding reaction occurs over a wide temperature range. This type of secondary peak was observed for the alloys that contained C contents above 0.075 wt pct and eutectictype constituents above 3 vol pct (e.g., 2, 3.5, 4, 6, 7.5, 8, 10, 11.5, and 12). As listed in the QIA results (Table IV), /NbC forms as the only eutectic-type constituent in alloys 2, 3.5, and 4 and is the major constituent in alloys 6, 7.5, 8, 10, 11.5, and 12. In addition, the integrated peak area generally increases in the same order as the /NbC volume percent (Tables V and VI). Thus, these secondary 2790 VOLUME 29A, NOVEMBER 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Table IV. Phase Identification Summary and Quantitative Image Analysis Results Alloy 1 1.5 2 3 3.5 4 5 6 7 7.5 8 9 10 11 11.5 12 13 14 15 16 Phases Present? NbC* NbC NbC NbC Total Eutectic-Type Constituent (Vol Pct) /NbC (Vol Pct) /Laves (Vol Pct) 0.8 0.2 1.6 0.4 5.2 1.5 1.2 0.7 3.0 0.4 7.3 1.3 2.5 1.1 13.3 2.5 2.7 0.4 7.5 0.5 15.2 1.6 3.8 0.5 9.0 1.0 2.5 0.4 7.3 1.1 6.7 1.7 12.9 1.4 23.8 1.6 15.5 1.4 23.6 2.1 *May contain Laves also. **NbC or /NbC present in too low an amount to accurately quantify. Laves or /Laves present in too low an amount to accurately quantify. 5.2 1.5 3.0 0.4 7.3 1.3 0** 13.3 2.5 0** 5.2 0.5 13.7 1.6 0** 9.0 1.0 0** 6.6 1.1 4.4 1.7 0** 18.8 1.6 0** 17.5 2.1 0 0 0 2.5 1.1 0 2.7 0.4 2.3 0.5 1.5 0.7 3.8 0.5 0 2.5 0.4 0.7 0.1 2.3 0.8 12.9 1.4 5.0 0.6 15.5 1.4 6.1 0.7 Fig. 5 Amount of total ( /NbC /Laves) eutectic-type constituents in the experimental alloys. peaks are associated with the L ( NbC) reaction, and the first deviation from the local baseline is taken as the initiation of this reaction. The differences in reaction temperatures among the alloys result mainly from variations in their solidification paths, which are controlled by the nominal alloy composition and distribution coefficients for the solute elements (Nb, C, and Si). These details are discussed more quantitatively in companion articles. [9,10] The DTA cooling scan for alloy 13 is shown in Figure 7(c). After the large exothermic peak from formation of the dendrites, the alloy exhibits a small exothermic deviation from the local baseline at 1332 C, followed by a larger peak at 1247 C. /Laves is the major constituent formed in this alloy, while /NbC forms in much smaller quantities. Thus, the small deviation occurring after the primary peak is associated with the L ( NbC) reaction and the lower temperature peak represents the L ( Laves) reaction. Alloy 15 showed similar behavior. Alloy 16 (Figure 7(d)) exhibited a large secondary reaction peak at 1356 C from the L ( NbC) reaction ( /NbC is the major constituent in this alloy), followed by a smaller peak at 1246 C from the L ( Laves) reaction. Alloy 14 exhibited a similar DTA cooling scan. The DTA solidification scan for alloy Ni-1 is shown Figure 7(e), where a large exothermic peak is initiated at 1190 C. As shown in Figure 4(a), this alloy contains mostly /Laves, with smaller amounts of /NbC and primary austenite. Thus, the large peak initiated at 1190 C is readily matched to the L ( Laves) reaction, which is in good agreement with the temperature of 1198 C reported for IN718 by Knorovsky et al. using this reaction. [19] It is somewhat surprising to find only one peak occurring at higher temperatures in the Ni-1 cooling scan (1244 C), since both primary austenite and /NbC are present in alloy Ni-1. As discussed in more detail elsewhere, [10] this alloy is very close to the line of twofold saturation, which separates the and NbC primary phase fields. Thus, the primary L reaction and L ( NbC) eutectic-type reaction are very close to each other in terms of reaction temperature. This, coupled with the fact that the /NbC and primary both form in relatively small quantities in alloy Ni-1, precludes unequivocal matching of the deviation at 1244 C to either the L or L ( NbC) reaction. The DTA scans recorded from two separate tests on alloy Fe-1 are shown in Figure 7(f). On cooling from above the liquidus, primary dendrites form in the range of 1339 C to 1346 C. There is a small, but reproducible, deviation from the local baseline at 1325 C, which is interpreted as the L ( NbC) reaction. Solidification is terminated with the L ( Laves) reaction at 1266 C. METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, NOVEMBER 1998 2791

surprisingly, the Fe contents of the eutectic-type constituents in the Fe-base alloys are considerably higher than those in the Ni-base alloys. Considering that the melting temperature of Fe is higher than that of Ni, the higher reaction temperatures for the Fe-base alloys are to be expected. The general reaction sequence of these alloys, in which the L ( NbC) transformation occurs over a broad temperature range prior to termination of solidification by the L ( Laves) reaction occurring over a narrow temperature range, is well documented for commercial alloy systems as well. [5,15,19] Fig. 6 Amount of individual /NbC and /Laves eutectic-type constituents in (a) Ni-base alloys and (b) Fe-base alloys. Table V. Alloy 3.5 2 7.5 4 6 8 Summary of Secondary Solidification Reaction Temperatures in Ni-Base Alloys L ( NbC) Start Temperature ( C) 1322 1.4 1341.5 0.7 1305.5 0.7 1330.0 1.3 1332.0 1.4 1328.0 2.8 /NbC Amount (Vol Pct) 3.0 0.4 5.2 1.5 5.2 0.5 7.3 1.3 13.3 2.5 13.7 1.6 Integrated Peak Area ( V s/mg) 0.2 0.5 0.4 0.8 1.6 2.1 At similar levels of solute elements, the solidification reaction temperatures in the Fe-base alloys are consistently higher than those in the corresponding Ni-base alloys. The compositions of the /NbC and /Laves eutectic-type constituents in the weld metals of several alloys are summarized in Tables VII and VIII. As noted in earlier work by Knorovsky et al., [19] the eutectic-type constituents that form in welds exhibit fine structures that are suitable for obtaining an average eutectic composition measurement by EPMA. The majority of measurements were acquired on weld metals, where composition modifications during postsolidification cooling were avoided due to the high cooling rates. Several measurements were also made on the fine eutectic-type /Laves structure, which formed in DTA samples (results for alloys 13 and 16 are in Table VIII). Not C. Analogy to the Ni-Nb-C System As noted in Table IV, all the alloys in this study exhibited the NbC phase (except for alloy 1, where no confident phase identification could be made). The Laves phase was observed in all alloys except 2, 3.5, and 4 (it may be present in alloy 1.5). It is interesting to note the similarity in secondary phases formed during solidification in these alloys and those expected in the ternary Ni-Nb-C system. The liquidus projection for the Ni-Nb-C system was estimated by Stadelmaier and Fiedler, [21] and the Ni-rich corner of this diagram is reproduced in Figure 8. This diagram is only an estimate, which was obtained by observing the primary solidification phase in a number of ternary Ni-Nb-C alloys. The accuracy of the boundaries separating each primary phase field is unknown, as few alloys were investigated with compositions close to the reported boundaries. Nevertheless, the projection provides a useful starting point for more detailed analysis of the experimental alloys used in this study. The liquidus projection exhibits three primary phase fields that are of interest here:, NbC, and Ni 3 Nb. A primary C (graphite) phase field exists at high C contents, which is not of interest. As previously noted, additions of Fe, Cr, and Si to the Ni-Nb system are well known to promote Laves at the expense of Ni 3 Nb in commercial superalloys as well as the experimental alloys used in this work. Thus, by replacing Ni 3 Nb with Laves, the Ni-Nb-C liquidus projection can be utilized as a guide in developing a qualitative description of the solidification reaction sequences in these alloys. [22] Solidification begins with formation of primary dendrites, which, upon forming, reject Nb and C to the liquid. (As shown in other work, [5,9,19] the distribution coefficients for Nb and C are less than unity for these alloy systems, indicating they segregate to the liquid during solidification.) Thus, as solidification proceeds, the liquid composition moves radially away from the Ni corner, becoming progressively richer in Nb and C until the line of twofold saturation between and NbC is reached. At this point, the maximum solubility of Nb and C in the austenite matrix is exceeded and the and NbC form simultaneously from the liquid by a eutectic-type reaction as the liquid composition travels down the line of twofold saturation. This initial solidification sequence explains the broad exothermic peaks exhibited in the DTA scans, where the L ( NbC) reaction occurred over a wide temperature range. Due to its high C content ( 9.5 wt pct), the formation of NbC depletes the liquid of C while the Nb content of the liquid continues to increase. For the Ni-base alloys with low Nb and high C (alloys 2, 3.5, and 4), solidification is completed 2792 VOLUME 29A, NOVEMBER 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

(a) (b) (c) (d) (e) Fig. 7 Typical DTA solidification scans for (a) alloy 3, (b) alloy 4, (c) alloy 13, (d) alloy 16, (e) alloy Ni-1, and ( f) alloy Fe-1. (f)

Table VI. Summary of Secondary Solidification Reaction Temperatures in Fe-Base Alloys Alloy 10 11.5 12 Start Temperature ( C) 1357.5 (0.7) 1347.5 (2.1) 1348.0 (1.4) 1332.5 L ( NbC) reaction /NbC (Vol Pct) 9.0 (1.0) 6.6 (1.1) 4.4 (1.7) Peak Area ( V s/mg) Start Temperature ( C) 0.1 1249.5 13 (0.7) ND (3.5) 1360.8 18.8 3.0 1243.0 14 (2.2) (1.6) (2.1) 1289.5 ND 0.1 1256.3 15 (0.8) (1.2) 1355.2 17.5 3.3 1248.2 16 (1.7) (2.1) (4.8) ND present in quantities too low for accurate quantification. L ( Laves) reaction /Laves (Vol Pct) Peak Area ( V s/mg) 0.4 0.8 0.4 12.9 (1.4) 5.0 (0.6) 15.5 (1.4) 6.1 (0.7) 0.7 0.7 1.8 0.2 Table VII. /NbC Constituent Average Compositions Alloy/ Sample Fe, wt pct Ni, wt pct Cr, wt pct Nb, wt pct Si, wt pct Alloy 6 weld 8.4 0.2 58.7 0.7 19.5 0.2 13.3 0.8 0.1 0.02 Alloy 8 weld 8.9 0.3 58.7 0.5 19.4 0.2 12.5 0.6 0.5 0.02 Alloy 14 weld 40.2 0.7 28.5 0.7 20.3 0.3 10.9 0.7 0.1 0.02 Alloy 16 weld 40.5 0.5 28.6 0.3 19.7 0.2 10.8 0.7 0.5 0.02 *All values are in weight percent. Table VIII. /Laves Average Constituent Compositions Alloy/ Sample Fe, wt pct Ni, wt pct Cr, wt pct Nb, wt pct Si, wt pct Alloy 7 DTA Alloy 7.5 DTA Alloy 8 DTA Ni-1 GTA melt Alloy 13 weld Alloy 13 DTA Alloy 14 weld Alloy 15 weld Alloy 16 weld Alloy 16 DTA Fe-1 GTA melt 4.9 0.2 4.9 0.2 4.4 0.2 7.0 0.1 31.4 1.8 31.2 0.9 31.0 1.1 32.7 0.5 32.7 0.4 28.6 1.9 37.7 0.3 59.4 0.8 58.1 1.7 60.3 0.5 52.2 0.4 29.8 1.1 30.4 0.5 29.5 0.8 29.7 1.0 30.8 0.7 33.4 1.4 26.2 0.4 10.2 1.0 11.4 0.8 11.3 0.3 16.7 0.2 16.6 0.6 17.3 0.2 16.8 0.7 16.7 0.4 16.9 0.1 16.5 0.9 18.0 0.1 24.3 0.7 23.5 1.3 21.5 0.5 23.1 0.5 21.1 1.3 20.9 1.0 22.4 2.4 20.1 0.9 18.3 0.7 19.8 1.2 17.2 0.1 1.7 0.1 2.0 0.1 2.5 0.2 1.0 0.1 0.3 0.1 0.3 0.1 0.3 0.1 0.9 0.1 1.3 0.1 1.7 0.1 0.9 0.1 *All values are in weight percent. 2794 VOLUME 29A, NOVEMBER 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 8 Ni-rich corner of the Ni-Nb-C liquidus projection. [23] along the line of twofold saturation between and NbC, as no Laves phase was observed in these alloys. The completion of solidification prior to reaching a minimum in the liquidus surface (i.e., the proposed ternary eutectic point of four-phase L- -NbC-Laves equilibrium) is rather unique. Flemings [23] points out that, for the case of two solutes with negligible diffusion rates in the solid, the liquid composition should become progressively enriched in solute until a local minimum is reached on the liquidus surface, at which point the remaining liquid will transform to solid. The continued enrichment in liquid composition is a direct result of insignificant solute diffusion in the solid. As demonstrated in the companion article [9] the diffusivity of Nb in austenite is expected to be negligible, but the diffusion rate of C is likely to be high enough [24] to significantly alter the solidification reaction from the ideal case discussed by Flemings. This effect contributes to solidification terminating before a local minimum is reached, as analyzed in quantitative detail in a separate article. [10] The initial reaction sequence is similar for the remaining alloys, except that solidification does not terminate with the L ( NbC) reaction. Instead, the liquid composition continues to travel down the line of twofold saturation between and NbC, becoming progressively enriched in Nb (and depleted in C). According to the proposed Ni-Nb-C liquidus projection, solidification should terminate with the ternary L ( NbC Laves) reaction, where the liquidus surface is at an apparent minimum. Under this condition, the Laves and ternary and NbC phases should be intermixed. However, this type of structure is not observed in any of the alloys that form both the NbC and Laves phases (e.g., examine microstructures of alloys 8 and 16 in Figure 1 through 3). Instead, the /NbC and /Laves eutectic-type constituents are always distinctly separated. This suggests that the actual liquidus projection for the alloys in this work is more properly represented by a class II reaction. [25] In this case, the local minimum on the liquidus surface occurs where the line of twofold saturation separating the and Laves phases intersects the Ni-Nb binary side of the diagram. The solidification sequence for alloys with this type of surface would occur as follows. After primary solidification, the liquid composition travels along the line of twofold saturation separating the and NbC phases during the L ( NbC) reaction. This process continues until the class II reaction is reached, at which point the NbC stops forming as the L ( NbC) reaction is replaced by the L ( Laves) reaction. Solidification is completed at the Ni-Nb binary side of the diagram by the L ( Laves) reaction. This solidification sequence accounts for the two spatially separate /NbC and /Laves eutectictype constituents, which are observed experimentally. The Ni-Nb-C liquidus projection, together with results from DTA and microstructural characterization, provides a qualitative interpretation of the solidification behavior of these alloys. This information forms the basis for developing a model for semiquantitative prediction of microstructural evolution during solidification. Several key elements are required to construct such an analysis. These include solidification parameters related to both the alloy system and processing conditions, a liquidus surface to identify the lines of twofold saturation separating primary phase fields, and an appropriate solute redistribution model that describes solutal transport between the liquid and solid during solidification. A companion article [9] is dedicated to determining the first two elements of the solidification model (solidification parameters and diagrams representing ternary-type liquidus projections), and the modeling results are reported elsewhere. [10] IV. CONCLUSIONS The solidification reaction sequences of experimental superalloys with systematic variations in Fe, Nb, Si, and C were investigated by DTA and microstructural characterization techniques. The following conclusions can be drawn from this work. 1. The general solidification sequence of these alloys can be described by a three-step process:(1) primary L solidification in which the interdendritic liquid becomes enriched in Nb and C, followed by (2) a eutectic-type L ( NbC) reaction that depletes the interdendritic liquid of C, and (3) termination of solidification by a second eutectic-type reaction L ( Laves). This solidification reaction sequence is similar to that expected in the ternary Ni-Nb-C system and observed experimentally in many commercial alloy systems. 2. The Ni-base alloys with relatively high C ( 0.075 wt pct) and low Nb ( 2 wt pct) represent the only exception to this general solidification sequence. This group of alloys terminates solidification with the L ( NbC) reaction and does not exhibit the /Laves constituent. 3. The Fe and Si additions increase the amount of the /Laves constituent, which forms in the as-solidified microstructure, while C additions promote formation of /NbC. The influence of Nb depends on the C content of the alloy. When the C content is low, Nb additions promote formation of /Laves. When the C content is high, Nb promotes formation of the /NbC constituent. ACKNOWLEDGMENTS One author (JND) gratefully acknowledges financial support for this research from the American Welding Society METALLURGICAL AND MATERIALS TRANSACTIONS A VOLUME 29A, NOVEMBER 1998 2795

Fellowship Award. Preparation of the experimental alloys by B. Damkroger and M. Maguire, Sandia National Laboratories, is also greatly appreciated. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract No. DE-AC04-94AL85000. REFERENCES 1. R.M. Nugent: Welding J., 1986, vol. 65 (6), pp. 33-39. 2. B.F. Levin, J.N. DuPont, and A.R. Marder: in Elevated Temperature Coatings: Science and Technology, I, N.B. Dahotre, J.M. Hampikian, and J.J. Stiglich, eds., TMS, Warrendale, PA, 1995, pp. 325-40 3. J.N. DuPont: Metall. Mater. Trans. A, 1996, vol. 27A, pp. 3612-20. 4. Q.H. Zhao, Y.P. Gao, J.H. Devletian, J.M. McCarthy, and W.E. Wood: International Trends in Welding Science and Technology, Proc. 3rd Int. Conf., S.A. David and J.M. Vitek, eds. ASM, Materials Park, OH, 1992, pp. 339-43. 5. M.J. Cieslak, T.J. Headley, T. Kollie, and A.D. Romig, Jr.: Metall. Trans. A, 1988, vol. 19A, pp. 2319-31. 6. Y. Nakao, H. Ohshige, S. Koga, H. Nishihara, and J. Sugitani: J. Jpn. Welding Soc., 1982, vol. 51, pp. 989-95. 7. M.J. Cieslak: Welding J., 1991, vol. 70, pp. 49s-56s. 8. R.A. Patterson and J.O. Milewski: Welding J., 1985, vol. 64, pp. 227s- 231s. 9. J.N. DuPont, C.V. Robino, and A.R. Marder: Metall. Mater. Trans. A, 1998, vol. 29A, pp. 2797-2806. 10. J.N. DuPont, C.V. Robino, and A.R. Marder: Acta Mater., 1998, vol. 46, pp. 4781-90. 11. K.F.J. Heinrich, Microbeam Analysis, Proc. 21st Int. Conf., A.D. Romig, Jr. and W.F. Chambers, eds., Albuquerque, NM, 1986, pp. 279-80. 12. R.P. Goehner and J.R. Michael: J. Res. Nat. Inst. Standards Technol., 1996, vol. 101 (3), pp. 301-08. 13. Binary Alloy Phase Diagrams, T.B. Massalski, ed., ASM, Materials Park, OH, 1990, vol. 1, p. 863. 14. C.V. Robino, J.R. Michael, and M.J. Cieslak: Sci. Technol. Welding Joining, 1997, vol. 2, pp. 220-30. 15. M.J. Cieslak, T.J. Headley, G.A. Knorovsky, A.D. Romig, Jr., and T. Kollie: Metall. Trans. A, 1990, vol. 21A, pp. 479-88. 16. Metals Handbook, 8th ed., ASM, Metals Park, OH, 1973, vol. 8. 17. Z. Blazina and R. Trojko: J. Less-Common Met., 1986, vol. 119, pp. 297-305. 18. S.T. Wlodek: Trans. Am. Soc. Met., 1963, vol. 56, pp. 287-303. 19. G.A. Knorovsky, M.J. Cieslak, T.J. Headley, A.D. Romig, Jr., and W.F. Hammetter: Metall. Trans. A, 1989, vol. 20A, pp. 2149-58. 20. R. Nakkalil, N.L. Richards, and M.C. Chaturvedi: Metall. Trans. A, 1993, vol. 24A, pp. 1169-79. 21. H.H. Stadelmaier and M.L. Fiedler: Z. Metallkd., 1975, vol. 66 (4), pp. 224-25. 22. B. Radhakrishnan and R.G. Thompson: Metall. Trans. A, 1989, vol. 20A, pp. 2866-68. 23. M.C. Flemings: Solidification Processing, McGraw-Hill, New York, NY, 1974. 24. T.W. Clyne and W. Kurz: Metall. Trans. A, 1981, vol. 12A, pp. 965-71. 25. F.N. Rhines: Phase Diagrams in Metallurgy, R.F. Mehl and M.B. Beaver, eds., McGraw-Hill, New York, NY, 1956, pp. 175-85. 2796 VOLUME 29A, NOVEMBER 1998 METALLURGICAL AND MATERIALS TRANSACTIONS A