Alloy Design of Ti Alloys Using Ubiquitous Alloying Elements and Characteristics of Their Levitation-Melted Alloys

Materials Transactions, Vol. 51, No. 4 (2010) pp. 740 to 748 #2010 Japan Foundry Engineering Society

Kazuhiro Matsugi, Takuro Endo, Yong-Bum Choi and Gen Sasaki

Department of Mechanical Materials Engineering, Hiroshima University, Higashi-Hiroshima 739-8527, Japan

The + type Ti-5.5Al-2Fe and type Ti-2.5Fe-2Mn-2Zr alloys have been theoretically designed, for the modification of Ti-6Al-4V and the achievement of the high tensile strength more than 1000 MPa at the solution treatment state, respectively, using ubiquitous alloying elements in order to establish the strategic method for suppressing utilization of rare metals. The utilization of the cold crucible levitation melting (CCLM) is very useful for the production of ingots, because titanium is very chemically reactive at high temperature. The experimental alloys with high purity and without contaminations from a crucible were prepared, and the homogeneous melt was also achieved by the diffusion mixing effect of CCLM. The microstructure, phase stability, strength, corrosion-resistance and workable properties of the design Ti-5.5Al-2Fe alloy, were comparable to those of Ti-6Al-4V. In contrast, the solution heat treated Ti-2.5Fe-2Mn-2Zr alloy showed the tensile strength of 1200 MPa, and the 1.3 times increase in the specific strength compared with Ti-15Mo-5Zr-3Al. The alloy design can be successfully carried out even using ubiquitous alloying elements by the d-electrons concept, which leads to the establishment of one method for the strategic utilization of rare metals. [doi:10.2320/matertrans.F-M2010801]

(Received March 26, 2009; Accepted January 8, 2010; Published March 3, 2010) Keywords: rare metals, ubiquitous alloying elements, alloy design, d-electrons concept, titanium alloys, environmentally friendly materials

1. Introduction the contaminating behavior of ordinary materials (oxides, borides, silisides, sulfides, nitrides, fluorides, Mo3Al and W) Recently, a variety of new titanium alloys for aerospace, could never be totally stopped. It has been reported that Y2O3 chemical, biomedical and welfare applications1–3) have been stabilized with 8 to 15 mass% of Ti has the excellent developed over the world. The principally -stabilizing performance in suppression for reaction between molten elements in titanium alloys are rare ones due to their low titanium and crucibles, although these crucibles are very 19) abundances in the earth’s crust. In contrast, titanium is con- expensive. In contrast, interaction between ZrO2-, Al2O3- sidered to be a ubiquitous element since it has the tenth or SiC-crucibles and molten titanium was investigated and highest Clarke number of all the elements, but it is classified discussed in the point of thermodynamic calculation.15) as a rare metal because its refinement process is more Moreover, utilization of crucibles made of CaO is useful environmentally damaging than the processes used to refine for melting Ti and its alloys, because of their refining effects iron and aluminum. It is important for the strategy of titanium such as de-oxidation, de-sulfurization and de-nitrification.20) alloys that high performance alloys are developed using CaO crucibles are very expensive and the handling is difficult ubiquitous elements such as iron, aluminum and interstitial due to their hydration. Vacuum induction and arc skull ones, etc.4–6) However, in most cases methods for the devel- melting processes both are used to prevent the molten metal opment are largely dependent on the trial-and-error experi- contamination. These processes possessed, however, very ments and some empirical rules. Therefore, the development low energy efficiency and great difficulty in obtaining is so inefficient and also expensive. In order to save cost and sufficient superheats generally needed for a better molten time necessary for alloy development, more fundamental metal homogenization.21) approaches based on the solid theory are needed. Therefore, Titanium and its alloying elements are difficult to be the d-electrons concept proposed by Morinaga7) is applied combined uniformly in composition as a solid alloy using to design of titanium alloys using ubiquitous metallic usual furnaces like arc or induction melting furnaces, because elements, in order to establish the strategic method for molten Ti is very chemically reactive at high temperature. suppressing utilization of rare metals. The high performance To resolve these problems, utilization of the cold crucible metallic materials such as Ni, Ti, Al, Mg and Fe-based alloys, levitation melting (CCLM) is very useful. CCLM in a high etc8–13) had been developed by the d-electrons concept. frequency induction furnace has both upper and lower The properties of titanium alloys depend greatly on the electric coils which were utilized for heating and levitation exact chemical composition, processing history and small- of molten metals as a major function, respectively.22–24) The ness of undesirably dissolved elements including contami- molten metal is levitated by the eddy current in the melting nants such as oxygen and carbon.14) The penetration of crucible which is water cooled. The alloys can be melted contaminants occurs basically during production and proc- under untouched condition between melt and melting essing of alloys.15) Commercial production process usually crucible, which leads to no contaminant from melting involves induction melting of alloys under heavy vacuum. A crucible. Moreover, titanium alloys with uniform composi- major source of contaminants is refractory melting crucibles, tion can be produced independently from the difference in which needs to be carefully chosen. Numerous investiga- specific gravity or melting point between titanium and its tors16–18) have tried to solve above problems, but they have alloying elements, by the diffusion mixing effect of strong generally not been able to obtain a satisfying result, because stirring due to an electromagnetic force. Alloy Design of Ti Alloys Using Ubiquitous Alloying Elements and Characteristics of Their Levitation-Melted Alloys 741

Copper melting crucible K Atmosphere controlling Melting Cooling Segment Slit T/ (a) 300 Solidification Cooling water 2050~2150 melting crucib 2000~2100 200 Current in le Temperature ,

/kW (b) W

Levitation of melt 85-90kW 80

Upper electric current Upper electric coil for heating 38-40kW Repulsion 35 Upper coil

Lower coil Electric power, 5 ctric coil 1x10 (c) 9x104 Pa on RP 9%

Lower electric current Lower ele p/Pa 1x10 DP for levitati Hole for pouring melt Ar with 99.9 3x10-3

Eddy current Pressure, Time, t/s

Fig. 2 Profile of (a) temperature in molten metal, (b) electric power in upper and lower coils and (c) pressure in atmosphere of the levitation Fig. 1 Schematic illustration of principle of levitation melting, and the melting process. Abscissaand ordinate are represented with arbitrary melting crucible for the melting of alloys and solidification of melts after scales. switching off electric power.

The d-electrons concept was applied to design of titanium gradient along the length (17 mm) of samples for the alloys using ubiquitous metallic elements, in order to measurement of specific resisitivity was about 5 K. Density establish the strategic method for suppressing utilization measurement using a high density liquid was performed by of rare metals. Some properties were investigated for the Archimedes’ method. design alloys produced by the CCLM technique. Tensile tests were conducted on the sheet-shaped specimen with gauge lengths of 2 4 28 mm, at room temperature 2. Experimental Procedures under an initial strain rate of 2:0 104 s1 in air. Both immersion tests for the evaluation of hot corrosion resistance All ingots of design alloys in this study were prepared from were conducted on the specimens with 1:5 5 10 mm at raw materials of pure Ti, Fe, Al, V, Zr, Mo or Fe-20.8Mn 573 and 923 K using the molten salts of NaNO3-6.6 mol%- alloy with 99.7, 99.9, 99.9, 99.7, 99.6, 99.0 or 99.0 mass%, NaCl and Na2SO4-45 mol%NaCl, respectively. Rockwell respectively, by CCLM under the atmosphere of argon gas hardness number on the C scale was measured at room with purity of 99.99%. The principle of levitation melting is temperature. illustrated in Fig. 1. Each element was inserted in the copper Charpy impact tests were conducted at 293 and 623 K with melting crucible with 150 cm3 consisting of 24 segments. an impact speed of 2.57 m/s in air using an instrumented The alloys can be melted under untouched condition between Charpy impact testing machine of 49 J capacity, where a melt and melting crucible, which leads to no contaminant swing angle of the pendulum of 90 degree was selected for from the melting crucible. Figure 2 shows profiles of tests under the low blow condition. The Charpy impact temperature in molten metal, electric power in upper and value was obtained from the difference in initial and final lower coils and pressure in atmosphere of a furnace-chamber. heights of the swinging pendulum. Dynamic maximum load The temperature was directly measured by insertion of was determined from the recorded load-deflection curves thermocouples in molten metals. The higher temperatures of using a computer device in the instrumented Charpy impact 2050–2150 and 2000–2100 K depending on the melting point machine system.25) An imapct specimen with a width of of alloying elements, were kept for 300 and 200 s in melting 10.0 mm, thickness of 2.7 mm and length of 50.0 mm was cut process, for complete melting of raw materials and enough from the fixed position in ingots by EDM. The notch in the mixing of molten metals, respectively. Molten metals were specimen showed a slot of 0.2 mm radius and 2 mm depth. solidified in the copper melting crucible after switching off electric power after the melting process. 3. Alloy Design for the Alloy with High Tensile All ingots were solution heat treated in an argon stream Strength and water-quenched. A part of ingots was aged after the solution heat treatment. The condition of heat treatment for The development has been accelerated owing to much each alloy is described in details in Chapter 5. The specific demand for type titanium alloys for aerospace applications. resisitivity was simultaneously measured from room temper- Two calculated parameters are mainly utilized in the d- ature to one above transus by the standard four probe electrons concept.7,9,10) The one is the d-orbital energy level d.c. method in air using a computer-controlled equipment. (Md) of alloying transition elements, and the other is the The size of samples was 1 1 17 mm. The temperature bond order (Bo) that is a measure of the covalent bond 742 K. Matsugi, T. Endo, Y.-B. Choi and G. Sasaki

Table 1 List of Bo and Md values in bcc Ti. 2.84 Mo W Nb Ta Elements Md/Bo Elements Md/Bo 2.82 Ti 2.447/2.79 Zr 2.934/3.086 V 1.872/2.805 4d Nb 2.424/3.099 2.80 Zr Cr V Hf

Cr 1.478/2.779 Mo 1.961/3.063 t Mn 2.78 Ti Mn 1.194/2.723 Hf 2.957/3.11 Bo Fe 3d Co Sn Fe 0.969/2.651 5d Ta 2.531/3.114 2.76 Si Co 0.807/2.529 W 2.072/3.125 Ni 2.74 Ni 0.724/2.412 Al 2.2/2.426 Others Cu Al Cu 0.567/2.114 Sn 2.1/2.283 2.72 2.25 2.30 2.35 2.40 2.45 2.50

Mdt

Fig. 3 Mdt-Bot line of Ti-M binary alloys. The vectors represent the location of Ti-10 mass%M.

2.89

Tensile strength : 600MPa 2.84 800MPa 1000MPa

R t 2.5Fe-2Mn-2Zr 2.79 Pure Ti Bo

type alloys type α+β +2.5Fe β β / +2Zr +2Mn

2.74 Phase boundaryCommercially between β type alloys: (52.5-92.5)Ti-(0-3)Sn-(0-15)Mo-(0-7.1)Zr-(0-3.5)Al- (0-22)V-(0-6)Cr-(0-35.3)Nb-(0-13)Ta 2.69 2.25 2.30 2.35 2.40 2.45 2.50 Mdt

Fig. 4 Mdt-Bot map of showing the location of commercial type alloys and contour lines of the ultimate tensile strength at solution heat state.

strength between atoms. The values of the parameters for As shown in Fig. 3, the alloy position moves in the Mdt- each element which were calculated on the MTi14 cluster Bot diagram as the alloy composition varies. For example, model (M: alloying elements) in the case of bcc Ti, are listed for a Ti-V binary alloy it moves to the left (the lower Mdt in Table 1.9,10) In order to understand the alloying effect region) with increasing V content. The vector drawn on the of each element in Ti, the Mdt-Bot diagram was prepared for Ti-V line represents the position of a Ti-10 mass%V alloy in Ti-10 mass%M binary alloys, and shown in Fig. 3. For the diagram. Such a vector varies in direction and magnitude, alloys, the average values of Md and Bo are defined by taking depending on alloying elements, as shown in the diagram. the compositional average, and Mdt and Bot are denoted The Bot and Mdt values obtained for bcc Ti are employed using eqs. (1) and (2). for alloy design in this paper, since both parameters are X relatively insensitive to the crystal structure.9,10) Bot ¼ XiðBoÞi ð1Þ X Thirty conventional type titanium alloys are plotted in the Mdt-Bot map as shown in Fig. 4. The phase boundary of Mdt ¼ XiðMdÞi ð2Þ in this figure is duplicated from that drawn in classification 9) Here, Xi is the atomic fraction of component i in the alloy, of commercial alloys into the , + and types. The ðMdÞi and ðBoÞi are the Md and Bo values for component i, contour lines showing the ultimate tensile strength level at respectively. The values of ðMdÞi and ðBoÞi are listed in the solution heat treatment state are also indicated by dotted Table 1. The summation extends over the components, lines. The ultimate tensile strength at the room temperature i ¼ 1; 2; ...; n. shows the maximum at the Mdt value about 2.40 ev and the Alloy Design of Ti Alloys Using Ubiquitous Alloying Elements and Characteristics of Their Levitation-Melted Alloys 743

Bot value of 2.80. Thus, a target region for alloy design can similar direction, which results in the vector from ‹ to › in be specified concretely on the Mdt-Bot diagram. Once a the map. Such a change is actually seen in the modification specific Mdt-Bot region and a specific alloy system are set in from Ti-6Al-4V to the design alloy of Ti-5.5Al-2Fe, as the map, the corresponding alloy composition is simply shown by the vector from ‹ to fi showing the sum of each determined following the rule of the vector sum.26) The alloying vector of Fe and Al in Fig. 5. The composition (Ti- titanium alloys with the high strength above 1000 MPa is 5.5Al-2Fe) of design alloy in this study is similar to that (Ti- designed using this Mdt-Bot map and also taking into account 5Al-2.5Fe) of the commercial alloy for biomedical applica- selection of ubiquitous elements. The addition of 2.5 mass% tion. It is suggested that this alloy design makes selection (2.2 mol%)Fe into pure Ti shifts the alloy position from ‹ to of the promising alloy more efficient and accurate compared ›, using the alloying vectors as shown in Fig. 3. The further to the currently used methods standing on several empirical addition of 2 mass% (1.8 mol%)Mn changes the position rules and many trial-and error experiments. from › to fi. Such a step is repeated until the position showing the maximum strength level of more than 1000 MPa, 5. Characteristics of Design Alloys Produced by CCLM which results in final attainment of the alloy position fl by a few amount of addition in 2 mass% (1.1 mol%) Zr. The 5.1 Microstructures and effects of CCLM design alloy has the composition of Ti-2.5Fe-2Mn-2Zr The transus was shown at 1073 and 1233 K, for design (mass%), and this position is indicated by the vector from alloys, Ti-2.5Fe-2Mn-2Zr and Ti-5.5Al-2Fe, respectively. ‹ to fl showing the sum of each alloying vector. The Ti- The transus of Ti-5.5Al-2Fe with + phase at room 2.5Fe-2Mn-2Zr alloy has the values of 2.789 and 2.398 on the temperature was close to that of Ti-6Al-4V, which meant the Bot and Mdt, respectively. The alloy position of design alloy successful modification of Ti-6Al-4V. In contrast, the is not so far to that (R) of commercial Ti-15Mo-5Zr-3Al transus of Ti-2.5Fe-2Mn-2Zr was 35 K higher than that of alloy. Hereafter, unless otherwise noted, alloy compositions Ti-15Mo-5Zr-3Al, which corresponded to the estimation of 7) are referred to in mass per cent. that in the Bot-Mdt map. The values of density of the + type Ti-5.5Al-2Fe and 4. Alloy Modification of Ti-6Al-4V Ti-6Al-4V alloys were 4.35 and 4.36 g/cm3, respectively. In contrast, the values of density of the type Ti-2.5Fe-2Mn- Forty five commercial titanium alloys are classified into 2Zr and Ti-15Mo-5Zr-3Al alloys were 4.60 and 4.95 g/cm3, the , + and types according to the phases existing in the respectively. alloy, as shown in Fig. 5.9,10) It is apparent that the three types Microstructures under the conditions of as-cast, solution of alloys are clearly separated in this map. The alloying heat treatment or aging after solution treatment, show in vector is a good indication for alloy modification, as Fig. 6 for + type design and reference alloys. For Ti- described in Chapter 3. In order to keep the alloy position 5.5Al-2Fe, the transformation from the phase to the 0 on the Mdt-Bot map nearly unchanged with the modification, martensite with HCP is partially caused by the solution heat alloy compositions may be adjusted among the elements treatment at 1253 K for 5.4 ks, which results in micro- of having the similar vector direction. The modification of structure consisting of the needle-like shaped 0 phase in the popular Ti-6Al-4V may be carried out using ubiquitous prior particle. The fine phase is precipitated around the elements, on the basis of the alloying vectors on the Mdt-Bot primary phase and in the prior particle, after aging map. There were similar vector directions among V, Cr, Mn treatment at 843 K for 21.6 ks. As the typical microstructure and Fe, or Sn, Si and Al, as shown in Fig. 3. For example, the of Ti-6Al-4V, that of the aging state is shown in Fig. 6(d). V reduction may require the Cr, Mn or Fe addition in There are the same microstructures in the each heat treatment modification of Ti-6Al-4V, as their vectors are pointed to the state between design and reference alloys, Ti-5.5Al-2Fe and Ti-6Al-4V, respectively. Figure 7 shows the microstructures of type alloys at the solution heat treatment state. Ti-2.5Fe- 2.84 2Mn-2Zr shows the needle-like shaped martensite phase β alloy which was transformed from prior phase by the solution 2.82 α + β alloy heat treatment at 1093 K above transus for 1.8 ks. In α alloy contrast, Ti-15Mo-5Zr-3Al shows the phase even by the 2.80 solution treatment at 1058 K above transus for 1.8 ks, which +2Fe t PureTi agrees with the result of X-ray diffraction analysis. Ti-Fe and

Bo 2.78 5.5Al Ti-Mn binary systems are classiﬁed into the -eutectoid type + 5.5Al-2Fe in three typical phase diagrams. The Fe or Mn containing 2.76 Target: Ti-6Al-4V alloys show the martensite start temperatures above the 7) 2.74 (73-92)Ti-(0-11)Sn- room temperature. Further, the Zr addition in alloys leads (0-15)Mo- (0-6)Zr-(0-8)Al- to the stabilization of the phase. The microstructure of (0-13)V-(0-11)Cr-(0-8)Mn 2.72 Ti-2.5Fe-2Mn-2Zr agrees with the behavior of martensite 2.25 2.30 2.35 2.40 2.45 transformation by Fe and Mn additions, as mentioned above. Md It is clear on the basis of the analysis by EPMA that the t strong segregation of alloying elements was not observed

Fig. 5 Mdt-Bot map showing the phase stability of commercially titanium even at as-cast state for experimental alloys, which means alloys. the better homogenization of molten metal by the diﬀusion 744 K. Matsugi, T. Endo, Y.-B. Choi and G. Sasaki

(a) (b)

150 µ m 150 µ m

Fig. 6 Optical micrographs showing typical microstructures under the conditions of (a) as-cast, (b) solution heat treatment (ST) and (c) ST + aging for Ti-5.5Al-2Fe, and (d) ST + aging for Ti-6Al-4V.

(a) (b)

µ 150 µ m 150 m

Fig. 7 Optical micrographs showing typical microstructures at the solution heat treatment state for (a) Ti-2.5Fe-2Mn-2Zr and (b) Ti- 15Mo-5Zr-3Al. mixing eﬀect of strong stirring due to an electromagnetic solution heat treated alloys in this study. In contrast, Table 2 force in CCLM, even for addition of Fe, Mn, Zr or Mo with lists chemical compositions of impurities in Ti-5.5Al-2Fe higher densities or higher melting points, compared with produced by CCLM. Contents of impurities such as oxygen, other techniques such as induction and arc skull meltings.21) carbon and nitrogen in the Ti-5.5Al-2Fe prepared by the The relation between homogeneous structure and the eﬀect CCLM, were 0.046, 0.009 and 0.004 mass%, respectively. of CCLM is also described in Section 5.2. It is reported that The contents of gaseous impurities of oxygen and nitrogen much segregation of Fe was observed Ti-Fe system alloys, were lower than those (oxygen and nitrogen: 0.069 and compared with Ti-V and Ti-Mo system alloys.27) However, 0.006 mass%, respectively) in the raw materials, because of the concentration of Fe was also homogeneous on the highly vacuum level as shown in Fig. 2. Moreover, it is found Alloy Design of Ti Alloys Using Ubiquitous Alloying Elements and Characteristics of Their Levitation-Melted Alloys 745

Table 2 Chemical compositions (mass%) of impurities in Ti-5.5Al-2Fe 1400 produced by CCLM. 1200 CONSiB H 1000 0.009 0.046 0.004 0.001 0.001 0.0007 800 σ 600 400 1200 Stress, /MPa As –cast sample (a) 200 Solution heat treated sample 1000 0 0426 800 /MPa ε Failure Strain, (%) σ 600 Fig. 9 Stress-strain curves of the as-cast and solution heat treated Ti-2.5Fe- 400 2Mn-2Zr alloy. Stress, Ti-5.5Al-2Fe 200 Ti-6Al-4V

0 n 0 1.0 2.0 3.0 4.0 5.0 0.10 Strain, ε (%) 0.08 1200 0.06 (b) 1000 0.04 Failure 0.02 as-cast, st, st+aging /MPa 800

Work hardening coefficient, 0 600 Ti-5.5Al-2Fe Ti-6Al-4V Ti-2.5Fe-2Mn-2Zr Alloys

Stress, 400 Ti-5.5Al-2Fe Fig. 10 Work hardening coefficient (n value) of Ti-5.5Al-2Fe, Ti-6Al-4V 200 Ti-6Al-4V and Ti-2.5Fe-2Mn-2Zr in each heat treatment conditions. 0 0 1.0 2.0 3.0 4.0 5.0 strength indicated in Fig. 4. The reference Ti-15Mo-5Zr-3Al ε Strain, (%) alloy showed the tensile strength of 1000 MPa at the solution heat treatment state, which resulted in the decrease of the Fig. 8 Stress-strain curves of Ti-5.5Al-2Fe and Ti-6Al-4V alloys under (a) the solution heat treatment (ST) and (b) ST+aging conditions. value (202 = 1000 MPa/4.95) in specific strength, compared with that (261 = 1200 MPa/4.60) of the design Ti-2.5Fe- 2Mn-2Zr alloy. The values in Young’s modulus of Ti-2.5Fe- from the contents of impurities that the cleanly molten metals 2Mn-2Zr were estimated to be 97 GPa from the stress-strain were created by utilization of CCLM without the reaction curve at the solution treatment state. between the molten metal and water-cooled copper crucible, The strain hardening coefficient (n value) was obtained although the affinity of Ti with oxygen, carbon and nitrogen from the stress-strain curves for each heat treatment states of was strong. Ti-5.5Al-2Fe, Ti-6Al-4V and Ti-2.5Fe-2Mn-2Zr, in order to estimate the stretch-formability of their alloys. The similar n 5.2 Tensile and hardness properties values were shown at each heat treatment state between Ti- Tensile tests were conducted on some solution treated or 5.5Al-2Fe and Ti-6Al-4V of + type alloys, as shown in aged specimens of design and reference alloys. Figure 8 Fig. 10. It is reported that the decrease of oxygen and iron shows the true stress-strain curves obtained from Ti-5.5Al- content in alloys leads to improvement on ductility and 2Fe and Ti-6Al-4V of + type alloys. The same behavior in workability.28) There was same level on oxygen content the stress-strain curves was shown depending on the solution between Ti-5.5Al-2Fe and Ti-6Al-4V produced by CCLM, treatment and aging states between both the + type alloys. using the same raw materials, as listed in Table 2. It is found A little improvement on flow stress was observed by the that the same level in the stretch-formability was shown in aging after the solution treatment, for Ti-6Al-4V. The values both alloys, as well as the type Ti-2.5Fe-2Mn-2Zr alloy. of Young’s modulus were estimated to be 108 and 110 GPa There appears to be good correlation between the n value and from the stress-strain curves for Ti-5.5Al-2Fe and Ti-6Al- microstructure, and the higher n value can be obtained by 4V, respectively. microstructural control. The same microstructure was also The as-cast and solution heat-treated Ti-2.5Fe-2Mn-2Zr shown in both alloys, as shown in Fig. 6. It is considered that showed the finally fracture-stress or -elongation of approx- both alloys with + type show the same level in the pahse imately 1120 and 1200 MPa or 5.2 and 3.5%, respectively, as stability, because they have same value in Bot and Mdt, shown in Fig. 9. This value on the tensile strength agrees which may lead to same degree in solid solution strengthen- with the estimated one from the contour lines of the tensile ing of both the and phases. 746 K. Matsugi, T. Endo, Y.-B. Choi and G. Sasaki

The same values in Rockwell hardness were measured C) depending on the heat treatment states between Ti-5.5Al-2Fe R 50 and Ti-6Al-4V of + type alloys, as shown in Fig. 11. In contrast, the type Ti-2.5Fe-2Mn-2Zr alloy showed higher 40 value of hardness at the solution heat treatment state, compared with Ti-15Mo-5Zr-3Al. The constant value of 30 hardness was measured even for as-cast specimens of all experimental alloys, because of the molten metal was 20 completely diffusion-mixed by effect of CCLM and rapidly Ti-5.5Al-2Fe solidified in cooling. Ti-6Al-4V 10 Ti-2.5Fe-2Mn-2Zr Ti-15Mo-5Zr-3Al 5.3 Impact properties 0 The load-deflection curves of the as-cast Ti-5.5Al-2Fe and As-cast ST ST Ti-2.5Fe-2Mn-2Zr alloys are shown in Fig. 12, as the typical (Solution treatment) +aging curves of impact tests. The similar shape was shown in the Hardness of Rockwell on C scale(H Heat treatment conditions curves depending on test temperatures, regardless of the kinds of alloys. The values of dynamic maximum load (Pm), Fig. 11 Results of Rockwell hardness testing.

(a) (b) 5.0 5.0 /kN /kN P P Load, Load,

0 0 0 2.0 4.0 0 2.0 4.0 6.0 8.0 Deflection, D/10-3m Deflection, D/10-3m (c) (d) 5.0 5.0 /kN /kN P P Load, Load,

0 0 0 2.0 4.0 0 2.0 4.0 6.0 8.0 Deflection, D/10-3m Deflection, D/10-3m

Fig. 12 Load-deﬂection curves obtained from (a), (b) Ti-5.5Al-2Fe and (c), (d) Ti-2.5Fe-2Mn-2Zr alloys. Test temperatures are (a), (c) 293, and (b), (d) 623 K. Alloy Design of Ti Alloys Using Ubiquitous Alloying Elements and Characteristics of Their Levitation-Melted Alloys 747

Table 3 Some values obtained from the load-deﬂection curves shown in Fig. 12.

Alloys Temp [K] Pm [kN] Et [J] Ei [J] Ep [J] Ei=Et Ep=Et 293 4.7 9.5 3.3 6.1 0.35 0.65 Ti-5.5Al-2Fe 623 3.3 12.1 2.7 9.4 0.22 0.78 293 4.9 8.6 3.1 5.5 0.36 0.64 Ti-6Al-4V 623 3.4 12.4 3.8 8.7 0.31 0.69 293 4.3 11.6 3.9 7.7 0.34 0.66 Ti-2.5Fe-2Mn-2Zr 623 3.2 12.8 5.1 7.8 0.40 0.60 Ti-15Mo-5Zr-3Al 293 4.8 12.3 3.8 8.5 0.31 0.69

crack initiation energy (Ei), crack propagation energy (Ep) i 2.0 and total absorb energy (Et ¼ Ei þ Ep) are listed in Table 3. W/W ∆ Ti-5.5Al-2Fe The same value in Pm obtained at 293 K was shown between Ti-6Al-4V both the + type alloys. In contrast, the increase of total 1.5 Ti-2.5Fe-2Mn-2Zr Ti-15Mo-5Zr-3Al deﬂection and the decrease of Pm were observed at high temperature of 623 K, because of softening of alloys. The 1.0 contribution degree of the crack-initiation and -propagation energies to the total absorbed energy was also listed in Table 3. The degree of contribution in the crack-propagation 0.5 energy was estimated to be 0.64–0.69 at 293 K, which meant same behavior in the load-deﬂection curves, regardless of 0 0 5 10 15 20 25 30 35 the kinds of alloys. Further, the same value in Et obtained

Ratio of weight loss to initial weight, Immersion time,t/ks at 623 K was shown among the Ti-5.5Al-2Fe, Ti-6Al-4V and Ti-2.5Fe-2Mn-2Zr alloys. However, there was different Fig. 13 Relation between the immersion time and ratio of weight loss to degree of the contribution in the crack-initiation or initial weight obtained from immersion test in Na2SO4-45 mol% NaCl -propagation energy between + and alloys tested at molten salt at 923 K. 623 K. The value of contribution-degree for the crack- propagation energy increased and decreased at the high temperature, compared with the room temperature, for the 2.80 Ti-15Mo-5Zr-3Al + type Ti-5.5Al-2Fe or Ti-6Al-4V, and type Ti-2.5Fe- 2Mn-2Zr alloys, respectively. Ti-2.5Fe-2Mn-2Zr The impact tests were conducted on as-cast alloys. The t 2.78 same load-deflection curves or no-scatter values were Bo obtained in same alloys, which lead to the achievement of 2.76 Ti-6Al-4V the homogeneous microstructure without the segregation of Fe, Mo and so on, because of the diffusion mixing effect Ti-5.5Al-2Fe 2.74 in CCLM. 0 0.2 0.4 0.6 0.8 1.0 1.2 ∆ Ratio of weight loss to initial weight, W/Wi 5.4 Hot corrosion resistance Fig. 14 Relation between the Bot and ratio of weight loss to initial weight The value of the weight loss was null in the NaNO3- obtained from the immersion test for 21.6 ks in Na2SO4-45 mol% NaCl NaCl immersion tests for 252.9 ks, regardless of the kinds molten salt at 923 K. of alloys. Figure 13 shows the results of the Na2SO4-NaCl immersion tests. The ratio of weight loss to the initial weight of result shown in Fig. 14 indicates that an increase in the bond specimens increased as immersion period increased, in all strength between atoms leads to high corrosion resistance. alloys. There were not remarkable different values in this This agrees with the result obtained from the active corrosion 30) ratio throughout the immersion period of 0.3 to 3.3 ks, for rate of Ti alloys in 10%H2SO4 at 343 K. Ti-5.5Al-2Fe and Ti-6Al-4V of + type alloys. Figure 14 shows the relation between Bot of alloys and the ratio of 5.5 Design of ubiquitous Ti alloys by d-electrons concept weight loss to the initial weight after the immersion period The modification of Ti-6Al-4V was carried out on the basis of 21.6 ks. There was a good correlation between this ratio of the alloying vectors on the Mdt-Bot map indicating the and Bot.AsBot increased in alloys, this ratio decreased phase boundaries of , + and types. The design Ti- linearly, although there were a few data. It has been found 5.5Al-2Fe with ubiquitous alloying elements had same Bot that various physical properties could be interpreted in term and Mdt values as Ti-6Al-4V. The microstructure, phase of Bo. For example, Bo correlated well with activation stability, strength, corrosion-resistance and workable proper- energies for impurity-diffusion of transition elements in - ties of the design Ti-5.5Al-2Fe alloy, were comparable to 29) Ti. The active oxidation may be interpreted by Bot. The those of Ti-6Al-4V. 748 K. Matsugi, T. Endo, Y.-B. Choi and G. Sasaki

In order to design the high strength alloy with tensile 2) M. Niinomi: Recent Biocompatibility Metallic Materials, Strucutural strength more than 1000 MPa at the solution heat treatment Biomaterials for the 21st Century, ed. by M. Niinomi, T. Okabe, E. M. Taleff, D. R. Lesuer and H. E. Lippard, (TMS, Warrendale, 2001) state, Ti-2.5Fe-2Mn-2Zr was alloy-designed in the region pp. 3–14. of target on the basis of the alloying vectors on the Mdt-Bot 3) M. Ikeda: Materia Japan 41 (2002) 539–542. map indicating the contour lines of the strength level at 4) K. Matsugi, T. Endo, Y.-B. Choi and G. Sasaki: J. Japan Inst. Metals 72 room temperature. This Ti-2.5Fe-2Mn-2Zr alloy showed the (2008) 935–941. ultimate tensile strength of 1200 MPa on the solution heat 5) T. Ando, T. Tsuchiyama and S. Takagi: J. Japan Inst. Metals 72 (2008) 949–954. treated state. There was the 1.3 times increase in the specific 6) M. Nakai, M. Niinomi, T. Akahori, H. Ishikawa and M. Ogawa: J. strength on Ti-2.5Fe-2Mn-2Zr, compared with the commer- Japan Inst. Metals 72 (2008) 960–964. cial Ti-15Mo-5Zr-3Al. 7) M. Morinaga, N. Yukawa and H. Adachi: Tetsu-to-Hagane 72 (1986) The alloy design can be successfully carried out even using 555–562. ubiquitous alloying elements by the d-electrons concept, 8) K. Matsugi, Y. Murata, M. Morinaga and N. Yukawa: Mater. Sci. Eng. A 172 (1993) 101–110. which leads to the establishment of one method for the 9) M. Morinaga, N. Yukawa, T. Maya, K. Sone and H. Adachi: Proc. 6th strategic utilization of rare metals. World Conf. on Titanium, (1988) pp. 1601–1606. 10) M. Morinaga, J. Saito and M. Morishita: J. Japan Inst. Light Metals 42 6. Conclusions (1992) 614–621. 11) M. Morinaga, Y. Murata and H. Ezaki: Proc. Int. Symp. ‘Material Chemistry in Nuclear Environment’, (Tsukuba, March 1992) pp. 241– (1) The Ti-5.5Al-2Fe and Ti-2.5Fe-2Mn-2Zr alloys have 252. been designed, for the modification of Ti-6Al-4V and 12) R. Ninomiya, H. Yukawa and M. Morinaga: J. Japan Inst. Light Metals the achievement of the high tensile strength more than 44 (1994) 171–177. 1000 MPa at the solution treatment state, respectively, 13) K. Matsugi, H. Mamiya, Y. B. Choi, G. Sasaki, O. Yanagisawa and using ubiquitous alloying elements in order to establish H. Kuramoto: Int. J. Cast Metals Res. 21 (2008) 156–161. 14) Material Properties Handbook, Titanium Handbook, ed. by T. W. the strategic method for suppressing utilization of rare During, (ASM, Metal Park, OH, ASM, 1994). metals. 15) S. K. Sadrnezhaad and S. B. Raz: Metall. Trans. 36B (2005) 395–403. (2) The cleanly molten metals with lower contents of O, C, 16) C. Berthod, C. Weber, W. M. Thompson, H. O. Bielstein and M. A. and N, were created by utilization of CCLM without Schwartz: J. Am. Ceram. Soc. 40 (1975) 363–373. the reaction between the molten metal and water-cooled 17) M. Garfinkle and H. M. Davis: Trans. ASM 58 (1965) 520–530. 18) D. R. Schuyler and J. A. Petrusha: Casting of Low Melting Titanium copper crucible. Alloys, Vacuum Metallurgy, (Princeton, NJ, Science Press, 1977) (3) The microstructure, phase stability, strength, corrosion- pp. 475–503. resistance and workable properties of the design Ti- 19) R. L. Saha, T. K. Nandy, R. D. K. Misra and K. T. Jacob: Metall. Trans. 5.5Al-2Fe alloy, were comparable to those of Ti-6Al- 21B (1990) 559–66. 4V. 20) T. Degawa: Bull. Japan Institute of Metals 27 (1988) 466–473. 21) J. P. Kuang, R. A. Harding and J. Campbell: Mater. Sci. Technol. 16 (4) The Ti-2.5Fe-2Mn-2Zr alloy showed the ultimate (2000) 1007–1016. tensile strength of 1200 MPa, which meant the 1.3 22) H. Tadano, T. Take, M. Fujita and S. Hayashi: Proc. 2nd Int. Conf. on times higher specific strength compared with Ti-15Mo- Electromagnetic Processing of Materials 1 (1997) pp. 377–382. 5Zr-3Al. Ti-2.5Fe-2Mn-2Zr with lower Bot showed 23) T. Volkmann, W. Loser and D. M. Herlach: Metall. Trans. 28A (1997) poor corrosion resistance, compared with Ti-15Mo- 461–469. 24) S. Asai: J. Iron Steel Inst. Japan 7 (2002) 44–48. 5Zr-3Al with higher Bot, which meant good correlation 25) T. Kobayashi: Eng. Frac. Mech. 19 (1984) 49–65. between corrosion resistance and Bot. 26) K. Matsugi, Y. Murata, M. Morinaga and N. Yukawa: Proc. 7th Int (5) The alloy design can be successfully carried out even Conf. on Superalloys, ed. by S. D. Antolovich, R. W. Stusrud, R. A. using ubiquitous alloying elements by the d-electrons MacKay, D. L. Anton, T. Khan, R. D. Kissinger and D. L. Klarstrom, concept, which leads to the establishment of one (TMS, Warrendale, 1992) pp. 307–316. 27) K. Ikeda, M. Ogawa and M. Ueda: J. Japan Inst. Metals 72 (2008) 975– method for the strategic utilization of rare metals. 982. 28) M. Fukuda: Kobe Steel Engineering Report 50 (2000) 46–49. REFERENCES 29) M. Morinaga, N. Yukawa and H. Adachi: J. Iron Steel Inst. Japan 71 (1985) 1441–1451. 1) A. Suzuki, M. Ogawa and T. Shimizu: Electric Furnace Steel 75 (2004) 30) M. Morishita, Y. Ashida, M. Chikuda, M. Morinaga, N. Yukawa and 127–131. H. Adachi: ISIJ Int. 31 (1991) 890–896.