D0I:10.13374/i.issn1001-053x.1989.06.020 北京科技大学学报 第11卷第6期 Vo1.11No.8 1989年11月 Journal of University of Science and Technology Betjing Nov.1989 The Behavior of Inclusions during the Special Melting Processes of Superalloys' Fu Jie(停杰)**，Zhao Junhua(赵俊华)，Xu Gaoyang（徐高阳) Wang Hu:(王忠)*◆* ABSTRACT:The behavior of nitride inclusions during the VIM,VAR and ESR processes of superalloys has been investigated.The experimental results have proved that the inclusions in the original metals,even if the pure TiN power added rapidly decompose at the superalloy liquid processing temperature and the nitrides in the remelted metals are mainly precipitated from solution during the solidification process. KEY WORDS:superalloy,inclusion,special melting process The type,quantity,morphology,size and distribution of inclusions have an important effect on the properties of superalloys.The oxide and other types of inclusions potentially decrease the low cycle fatigue life of nickle-base superalloys, As a result of the recent investigationt1l,it has been found that the increase of nitrogen content,consequently,the increase of the amount and size of multi- angular MC-type carbide inclusions in GH36 (31481)obviously decrease the reduction of area,impact toughness,stress rupture life and stress rupture elonga- tion of the alloy. To control the clcanliness of superalloys,it is necessary to know the beha- vior of inclusions,first of all,to determine the inclusion state of existing during the special melting processes of superalloys.The state of existing of inclusions determines their removal mechanism,consequently,the technique for removing inclusions.The solid inclusions in the liquid metal can be removed,relying on the appoximate interfacial property of ceramic materials or molten slag adhering Manuscript Received May 1,1989 ·,Dept of Metallurgy ··+Dcpt,of Material Scicace and Engineering 501
气 第 卷第 期 年 月 北 京 科 技 大 学 学 报 以 手 。 。 ‘ 傅 杰 专 ， 赵 俊华 不厂 王 惠 徐 高 阳 ， · ， 、 · · 口 ， ， 、 、 ， ， ， 几 一 ， ， ， 一 ， ， 。 ， ， ， 丫 ， ， · ， ， 一 … DOI ：10．13374／j ．issn1001－053x．1989．06．020
to inclusions to cause adsorption of them on the crucible wall (for VIM process), filter surface (for filtration process)or in the molten slag (for ESR process). If the inclusions in the original metal decompose in the liquid metal during the special melting processes,the inclusion removal will result from the interaction between the liquid metal and vacuum atmosphere (for VIM and VAR processes) or melten slag (for ESR process).Therefore,the techaique for removing inclu- sions or improving cleanliness will be different under the two conditions mentioned above. In this paper,the behavior of inclusions during the VIM,VAR and ESR processes of superalloys has been investigated.For most superalloys,the prime target for inclusion reduction is that of nitrides,since their content is usually an order of magnitude higher than that of the oxides.The study emphasized the behavior of nitride inclusions. 1 The Decomposition of Pure TiN Powder in the Liquid Metal during the VIM Process of Superalloys In work 2,it was supposed that TiN cannot be decomposed by vacuum induction melting at pressures realistic for superalloy processing (ie.,>1um)and some technique (e.g,filtration)for removing solid TiN from the liquid super- alloy must be proposed. For confirming this viewpoint,a special experiment was arranged. The experiment was carried out in a 10kg VIM furnace.The alloys tested were the commonly used forging superalloy GH169 (In-718)and GH132 (A- 286).During the experimental process,a piece of forging bar from 30kg VIM ingot remelted in a corundum crucible at pressure of higher than 1330Pa.When the temperature of liquid metal reached 1550C,the pure TiN powder weighing 0.2 wt%of charge was added into the melt.The melt remained in the crucible at 1550C for 2 min then poured into a cast iron mold.The ingot weight was about 1.7kg. The nitrogen contents of original forging bar and VIM ingot were analysed by TN-114.The size,morphology and composition of TiN powder and the character of inclusions in the ingot were investigated by JMS-35C SEM. The nitrogen content analysis results are shown in Table 1.It can be seen that 240ppm and 236ppm nitrogen have been added into GH169 and GH132, respectively. The size of TiN powder is more than 104m,predominately,within 20-30 um.The size and morphology of TiN powder are shown in Fig.1,a. The SEM wave spectrum and the energy spectrum observation results indicated 502
， · ， ， 爪 几 ， ， · ， ， 。 声 ， · · ， ” ， ， · 一 一 五 ， · ， 。 “ 皿 一 · ， 一 · 。 ， 拼 ， ， 一 “ · 一 ， · 尸
that the TiN powder had a high purity which contained no carbon (see Fig.2, a)and other metallic elements except titanium. If the TiN powder did not decompose in the liquid metal during the VIM process,it would remain its original size and morphology in the VIM ingot, being the large inclusion.However,by observing the samples of the VIM ingot carefully,we have not found the original TiN powder in them,although the nitrogen content of the ingot is about 240 ppm higher than that of the original forging bar. Table 1 The Nitrogen Contents of Original Forging Bar and VIM Ingot (ppm) Alloy Forging Bar TiN Added Nitrogen Added VIM Ingot GH169 (1n-718) 96 2000 453 336 GH132 (A-286) 44 2000 453 280 10 10818初BPC Fig.1 Morphology of pure TiN powder (a)and Inclusions in the VIM Ingot of GH169 (In-718)(b)(um) The investigation result shows that the average size of inclusions in the VIM ingot for both GH169 and GH132 is about 2um smaller than that in the original forging bar,but the quantity larger and a number of inclusion clusters,consis- Table 2 The Composition of Inclusions in the VIM Ingot of GH169 (In-718),wt% Inclusion Ti Nb Fe Ni Cr 1008-1 6.85 72.13 4.01 11.44 5.56 2 32.17 16.10 19.38 29.21 12.14 3 53.75 12.07 7.12 17.00 9.26 -4 48.49 30.12 4.36 10.86 6.10 503
了 一 ， 。 ， ， · ， ， ， 一 一 一 笋 黔 霎 澳 羚滁聋 李 蠢 毕 、 参 纂 粉 、 牛 穷 髦移 娜酬 一 拼 拼 ， ， 一 一 ， 目二，上口，比 任丹﹄八 … 七丹乙户 上 一 一 一 一 。 。 。 。 。 。 口口上， 曰
ting of many individual particles,have been observed somewhere in the VIM ingots.The morphology of a typical inclusion cluster in the VIM ingot of GH169 is shown in Fig.1,b.These small particles contained carbon and niobium in a certain degree beside nitrogen and titanium (see Fig.2,b,and Table 2). Fig.2 Wave spectrun photos of pure TiN powder (a)and jnclusions in the VIM Ingot cf GH169 (In-718)(b) The results obtained demonstrated that the pure TiN powder added either adhered to the Al2O3 crucible wall,or decomposed in the liquid metal.From the experimental conditions and results,(ie.1060 ppm TiN powder,having a size of about 20um,completely decomposed in two minites),it was supposed that the decomposition rate of TiN powder be very high. Work 2 indicated that the TiN solubility corresponds to contents between 25 and 40 ppm N for the commonly used forging alloys.From Table 1,it can be seen that the nitrogen content in the original forging bar is higher than 40 ppm,but about 1000 ppm TiN (corresponds to about 240 ppm N)has decomposed in the liquid metals. To explain this result,the possibility of TiN powder decomposition in the liquid metal is discussed. During the VIM process,the reactions CN3-N2() (1) TiN(e)=〔Ti)+〔N) (2) simultaneously occur in the various regions.The denitrification reaction (1) occurs at the vaccum atmosphere/liquid metal interface.The nitrogen content of liquid metal in equilibrium with the partial pressure of nitrogen in vacuum atmosphere is very low,as the low nitrogen alloys (e.g.the In-718 containing 15 ppm N)can be melted in VIM furnace.The decomposition reaction (2)occurs within the melt.The nitrogen content of liquid metal in equilibrium with the 504
， · 宜 ， 一 · ， ， 。 刀 卜 一 ， 一 ， ， 召 ， ， 犷 · ， ， 。 ， 。 ， ， 、 、 、 “ 、 玄一 ‘、 “ ‘ ’ 。 〔 〕 〔 〕 。 ， 一 一 一 尸
K titanium contot,caleulated by Nmay i rory high,aithough cannot calculate the value accurately as the shortage of suifable thermodynamic data.When the overall system does nct reach the thermedynamic equilibrium, the pure TiN powder will decompese in the liquid metal. The equilibrium nitrogen content for reaction (2)decreases with decreasing the temperature of the alloy system,some new type nitride inclusions,concerning with the composition of the alloy melted,will precipitate from solution during the solidification process.As the liquid metal,in which the pure TiN powder was added,only remained in the crucible for 2 minites,the nitrogen distribution might be heterogeneous in the melt.At the place where the TiN powder just decomposed,the nitrogen content might be higher,consequently,the inclusicn cluster mentioned above intended to form. 2 The Behavior of inclusion during the VAR and ESR Processes of Superalloys The type of inclusions in the high Al,Ti alloy GH220 (311220),high Ti low Al alloy GH132,without Al,Ti alloy GH36 and the hot alloy GH169 has been studied by using optical microscope and scanning electron microscope.It was found that the main inclusions in the alloys tested are MC type inclusions inclu- ding primary carbides and nitride type inclusions.Usually,the former contains some degree of nitrogen and the latter is titanium nitride with some carbon solution (for GH132,GH220)or titanium,niobium nitride with some carbon solution (for GH169).It is to be noted that the alloy GH36 contains 700-1000 ppm nitrogen,but the main inclasions are MC type carbides containing V,Nb,C and N,no nitrides have been found in it (see Table 3). As the pure TiN powder can rapidly decompose in the liquid GH169 and GH132 at 1550C,the inclusions (carbides and nitrides)in the GH169,GH220, GH132 and GH36 should be able to decompose during the VAR and ESR pro- cesses. Table 3.The Composition of Inclusions in the ESR Ingot of GH36(3M481),wt% Inclusion Nb Cr Mo Fe Ni Mn Si 1 60.77.15.51.420.5 2.23.10.1 2 26.645.56.33.15.40.51.0 0.0 If the inclusions in the superalloys decompose during ihe VAR and ESR pro- 505
几 ， 〔 〕 一 了荞万 一 ， 一 。 · ， ，， · ， 从 、 厂 飞 一 丫 五 互。 上 ， 了 五 从 五 、 、 ， · 几 叭 了 ， 手 厂 ， 手 ， 从 · ， 。 ， 从 、 ， 。 ， 玉 ， ， 。 如、 ， ， ， ， 一 一 ， ， ， 一 一 ， ， ， ， 住 ， ， ， ， 一 上 日土，﹄ 。 。 。 。 。 。 。 。 。 。 。 。 。 扭
cesses,the size of inclusions in the solidified metal film will be smaller than that in the solidified ingot as the solidification rate of the metal film is higher than that of the remelted ingot. The remelting experiment results for the superalloys have proved the expe- ctation.The experimental conditions are as follows: GH220,the consumable electrode of 40mm dia.was melted in a 50kg VIM furnace and casted in a refractory mold.VAR ingot dia.is 65mm.The vari- ation of inclusions during the VAR process of GH220 is shown in Fig-3. GH169,the consumable electrode of 40mm dia.was melted in a 50kg VIM furnace and casted in a refractory material mold.During the VIM process,the pure TiN powder weighing 0.1 wt%of charge was added into the melt.The melt remained in the magnesia crucible at 1550C for 2 min.The consumable electrode contains 140 ppm nitrogen and some nitride clusters.ESR ingot dia. is 100mm,the slag composition is 70%CaF2+30%Al,Oa.The variation of inclusions during the ESR process of GH169 is shown in Fig.4. GH36,the consumable electrode of 40mm dia.was forged from AAM (air arc melting)+ESR forging bar of 90mm square.ESR ingot dia.is 100mm,the slag composition is 70%CaF2+30%Al2 Os. 20m Fig.3 Variation of Inclusions during the VAR process of GH220 (31I220) a-cast clectrode; b-mctal filmy c-wrought remelted metal. b 20细 Fig.4 Variation of Inclusions during the ESR process of GH169(In-718) a-cast electrode (TiN powder added) b-metal film, c-wrought remelted metal; 506
， 主 主 主 一 五 ， 一 · 且 主 一 。 ， 一 一 宜 王 。 ， 。 住 ， 。 · ， 扭 、 · 巧 ﹁ 确 位 住 一 ， 业器一 。 · 淤 称 募 咬 气 决 矿补兮 一 厂 一。 尹 美 一 一 ， 泌 黔 榔 农 一 。 豁 一 甄 一 一 、 未笋侧翔一 猫 夔弩 勇 扩巍 矜 噢氛纂 赘 脚一弓 梦 铆 飞 矛 一奋 一 瓜 矿爵中 百 一 犷 扩 万感称 终一粤谁 一 赢 涟咬一月找州减彼资汁汹月份︸雌门斗咚该份汗味套 易 声谬， 城翻节 。 卜 一 ， 一 口 ， 一 一
The variation characteristics of inclusions for GH132 and GH36 are similar to those for GH169 and GH220. The results obtained demonstrated that the inclusions in the original metals decomposed during the formation period of liquid metal droplets at the electrode tip during the VAR and ESR processes and those in the remelted ingots were mainly precipitated from solution bearing nitrogen during the solidification pro- cess of the remelted metal pool. 3 Some Comments on the Technologies Related to the Behavior of Inclusions during the Special Melting Processes of Superalloys 3.1 The Superalloy Cleanliness Evaluation by the EB-Button Test In recent years,the EB-button test has been used for the quantitative eva- luation of the inclusions of superalloys in a number of US corporationsts.This test is considered as an effective method to evaluate the superalloy cleanliness. As well known,the EB-button test is based on the floatation of inclusions from within the melt to the melt surface (form a raft.In this case,the inclusions should not decompose in the melt and keep a certain shape and size during the EB-button test process.However,the experimental results mentioned above indicaled that the inclusions including carbides,nitrides and carbonitrides in superalloys were formed during the solidification prccess.The size of inclu- sions results from the solidification rate of metals. During the EB-button test process,the inclusions in superallcys will deco- mpose in the melt,then precipitate to form new inclusions.The newly forming inclusions during the solidification process cannot completly floate to the melt surface.The solidification rate of the raft is different from that of the metal tested.Thus,the content and size of inclusions in the floating raft cannot be expressive of the real content and size of inclusions in the metal tested. Based on this point of riew,perhaps we can explain why the oxide content in the Rene'95 raft is so low (much less than 1 ppmt61)and the oxygen content and nitrogen content in the button are less than those in the samplet51. 3.2 The Liquid Metal Filtration The liquid metal filtration has been publicized recently for removing inclu- sions from the melt.The mechanism of inclusion removal is the adsorption of inclusion particles on the filter surface.The key problem is that the inclusions must be solid particles at the filtration temperature.As the nitride type inclusi- 507
， 」时 、 、 。 卜 丫 丫 卜 一 ， 一 、 ’ · · ， 一 ， 一 · ， ， 从 · · 。 呈 一 ， 。 ， 。 。 。 ， ， ， 〔 。 、 一
ons decompose in the liquid metal,the filtration efficiency will be limited.If all of the superalloys need to be filtrated is worthy of consideration.Work 7 reported "Fifteen-ton,VIM heats of wrought alloys of the turbine disk type (e. g.In-718)have been filtered.Some cleanliness improvement is observed.Ho- wever,effects on product are somewhat confouaded by secondary melting and conversion".Perhaps,it is related to the decomposition of inclusions.For removing oxide inclusions,especially in production of alloying steels,to adopt a good deoxidation method concerning with a low oxygen content dissoved is signi- ficant. 3.3 The Denitrification of Superalloys during the Spe:ial Melting Process During the primary melting (e.g.VIM)and secondary melting (e.g.VAR and ESR)processes,the nitrides and carbonitrides will rapidly decompose in the liquid metal at the melting temperature.The denitrification of superalloys occurs at the liquid metal/vacuum atmosphere interface relying on the reaction (I)for VIM and VAR processes or at the liquid metal/molten slag interface for ESR precess.No doubt,to increase the refining time at high temperature,select a suitable slag system will be beneficial to the denitrification.As the nitrogen diss)ving in the liquid metal will result in the precipitation of new inclusions during the solidification process,the control of solidification conditions must be significant for improving the superalloy cleanliness. 4 Conclusion The carbides,nitrides and carbonitrides in the original melals,even if the pure TiN powder added rapidly decomposed in the liquid metal at the superalloy melting temperaturc.The nitride type inclusions in the remelted metals are mai- nly precipitated froin s.lution during the solidification process.It is supposed that the control of slidification process is significant for improving cleanliness of superalloys. Based on the experimental results obtained,some comments on the technolo- gies relaled to the bchavior of inclusions during the special melting processes of superalloys can be given.As the inclusions decompose in the liquid metal and the new inclusions precipitate during solidification,the application arca of the EB-button test for the superalloy cleanliness evaluation and the filtration effici- ency of liquid metal will be limited.The removal mechanism of nitrides during the filtration process becomes comslex,if the nitrides can be removed. 508
， 爪 一 “ 一 ， 从 ， · · 一 一 、 一 ， 、、 · 以 ， · ， 元 ， ， 。 气妞 卜 之 ， 丁 刀 · · 了 · ， 、 、 · 、 一 。 〔 ， · ， ， 一 · 、、 · 认 ， 一 ‘ 声口 主 ， ， 。 。 蛋 · 手 · 几 。 之 ， 一 〕 、 、 一 。 · 、 ， 。 、 一 于 、 、 一 、 ，少 、 。 一 一 、 〔 ， 一 几 、、 · 、 ， 一 。 、 、 、 一 、 。 〕 ， 一 、 一
人 REFERENCES 1 Fu J,Gao L.J Vac.Sci.Technol.A,1987;5(4):2687,2692 2 Alex Mitchell.J Vac.Sci.Technol.A,1987;5(4):2672 3 Fu J,Hu Y H,et al.Acta Metallurgical,1986;22(3):B87-B104 4 Fu J,Wang H,et al.Proc.of the 7th Inter.Conf.on Vac.Metall. Special Meltings and Metallurgical Coatings,Tokyo,1982;1266 5 Domigue J A,Sutto W H.Proc.Vac.Metall.Conj.,Editor:Bhat GK, 1984;161 6 Shamblen C E,Culp S L.Proc.Vec.Metall.Conf.,Editor:Bhat G K, 1984;145 7 Domingue J A.Schweizer F A.J Vac.Sci.Technol.A,1987;5(4)t 2655 L 509
火，吸 ， 。 ‘ 人 ， ， 。 犷 。 ， ， ， · 夕 ， 一 ， ， 一 · · 一 · 夕 五 夕 了 夕 ， ， ， · · 犷 · ， ， ， 八 ， ， ， 从· · 八 了 人，， · ， 哭