Amorphization of metallic titanium by ball milling was presented. With the introduction of continuous pickup of impurities, hexagonally close-packed (hcp) titanium transformed gradually into an amorphous phase without experiencing any intermediate stage of forming a detectable metastable compound phase. The crystallization temperature of the obtained Ti metal glassy phase is about 640 K. The total concentration of the impurities (oxygen, nitrogen, iron, etc.) in the final product of the milled powders that was obtained after 60 h of milling was 10.85 at.%. The amorphization of metallic titanium may account for the combined effects of the pickup oxygen impurity in small amount and the Gibbs–Thompson effect.
Mechanical alloying (MA), a highly developed synthesis method of metastable materials, has been widely used to produce nanocrystalline and quasi-crystalline materials, intermetallics, solid solutions, amorphous alloys, etc.1, 2 In the process of MA, metal powders would be inevitably more or less contaminated. Contamination of the metal powders can be traced to (i) impurity of the starting powders, (ii) milling equipment (vial and grinding medium), (iii) atmospheric impurities in handling of the powders, and (iv) milling atmosphere. The impurity can lead to formation of nitrides in metallic Ti and Zr,3 amorphization of Ti-Zr, Ni60Nb40, and Nb25Sn powders.4, 5, 6 Meanwhile, the impurity, such as O, N, and Fe, can give rise to distinct differences in the behavior and products of crystallization for milled amorphous alloys, lattice parameter of milled crystalline phases, and viscosity of prepared amorphous alloys.5, 6, 7
Among all the impurities, the effect of oxygen on structural transformation of starting powders during MA has attracted great attention due to its chemical reactivity. Unfortunately, the effect of oxygen on structural transformation of starting powders has been documented mostly in binary and multicomponents alloy systems.1, 2, 5, 6, 7 It was reported that formation of metal glassy phase did not take place in TiAl24Nb11 alloy with 4.8 wt% O and in Ti6Al alloy with 44.8 at.% O.1 Koch et al. prepared amorphous Ni60Nb40 alloys with 3.4 wt% O and without O, and Nb25Sn amorphous alloy containing 11.3 to ~23.3 at.% O by MA.5, 6 In such cases, oxygen influences on structural transformation of some base metallic element may be concealed or weakened by those from other metallic elements and/or the other impurities. Recently, Lucks et al. found that amorphization of metallic molybdenum did not occur with the introduction of about 33 at.% O.8 More recently, Manna et al. studied the effect of oxygen on structural transformation of metallic Ti9; they did not detect any metal glassy phase even under 16.5 at.% O. Hence, the exact influence of oxygen impurity on the formation of a metal glassy phase in binary alloy systems is not very clear. From a theoretical viewpoint, it is also of interest to know whether a metal glassy phase can be formed in a single metallic element under the effect of oxygen. Perhaps, proper oxygen concentration has not been found for amorphization of some single metallic element by MA to date.
In the present work, amorphization of metallic titanium with 7.44 at.% O (2.65 wt% O) was obtained by MA in a milling atmosphere of purified argon, and the powders were exposed to air for a fixed time at regular time intervals. The aim of the research is to explore the effect of oxygen impurity in small amount on structural transformation of the starting powders and the minimum impurity levels needed for amorphization of metallic Ti.
Elemental powders of titanium (>99.1%) with a particle size of about 50 μm were put into a stainless steel vial (QM-2SP20) together with stainless steel balls in diameters of 15, 10, and 6 mm, respectively, the weight ratio of which was 1:3:1. The ball-to-powder weight ratio was about 12:1. The vial and the lid were sealed using an O-ring with a circular cross section. MA was performed in a high-energy planetary ball mill (QM-2SP20, made by apparatus factory of Nanjing University) at a rotation speed of 4.3 s−1 under a purified argon gas atmosphere (99.999%, 0.5 MPa). Before introducing argon gas, the vial was evacuated for 60 min by a diffusion pump (about 10−3 Pa). For the purpose of removing powders from the vial, the milling process was interrupted at 1, 2, 4, 6, 8, 10, 15, and 20 h, respectively. After 20 h, the milling process was stopped every 5 h to cool the vial to room temperature, but only at every 10 h were the powders removed. The temperature of the exterior wall of the vial at every end of 5 h was measured to about 413 K. It is worth noting that when removing the powders, they were directly exposed to air for 20 min. By doing so, the powders can pick up oxygen impurities. To confirm the reproducibility of the present results, the MA experiments were performed three times under the same conditions.
The structural transformation with the milling time was confirmed by x-ray diffraction (XRD) with Cu Kα radiation (1.5406 Å). From three diffraction peaks of Ti (100), (002), and (101), the average grain size (dc) was determined from the peak broadening analysis using the equation (B cos θ = 0.9λ/dc + η sin θ) after elimination of the respective influences of lattice strain and instrumental error.1 The thermal physical properties of the milled powders were examined using NETSCH STA449C differential scanning calorimetry (DSC; Germany). A JEM 2010 transmission electron microscopy (TEM; JEOL, Japan) was used to prove the amorphous nature of the 60 h milled glassy sample. The Fe, Ni, and Cr pickups during MA were analyzed by colorimetry and WFX-1E3 atomic absorption spectrophotometer (Third Analysis Apparatus Factory, Shanghai). The gas contamination contents of O and N were determined by TC600 Nitrogen/Oxygen Determinator (LECO Co., US) with an uncertainty of 0.025 ppm. In situ high-temperature XRD measurements of the 60 h milled glassy sample were performed by a D/max-2500/pc XRD (Rigaku Corp., Japan) in a high-temperature oven under a protective atmosphere of high purified argon. Each run of the three same experiments consists of heating to 873 K from room temperature at 20 K/min. From 473 to 873 K, an in situ XRD pattern was recorded every 20 K at constant temperature with an uncertainty of ±10 K to observe the phase transformation process. Also, the spark plasma sintering (SPS) method,10 a relatively new technique of materials treatment, was used to study the phase transformation process of the 60 h milled glassy sample by a 320MKII SPS system (Sumitomo Coal Mining Co.).
Figure 1 shows the XRD patterns of the Ti powders after ball milling for different times. The intensities of the diffracted peaks decreased significantly with increasing ball-milling time, indicating the presence of a large amount of defects in the milled powders. Meanwhile, the peaks become broader with increasing milling time, suggesting a continuous decrease in the grain size of the powders. There is no significant shift in the position of the major x-ray peaks as a function of the milling time. After 10 h, a broad peak with a maximum value of 2θ = 38.66° as shown in Fig. 1 appears, which is characteristic of an amorphous phase. After 60 h of the milling time, the sample is almost completely amorphous. With a further increase of the milling time, crystallization would take place in the amorphous sample.
The average grain size in diameter, dc, of the ball-milled Ti powders is plotted against the milling time in Fig. 2. Increasing the milling time leads to a dramatic decrease in dc. It reaches a minimum size of 10 nm after 10 h of the milling time (Fig. 2). It is worth mentioning that no remarkable changes in the grain size could be detected after this stage of milling, indicating that the grain sizes reached the steady-state value of 8 nm. The minimum of dc for the milled Ti agrees well with the milled Fe, which is 8 ± 1 nm after 60 h milling.11 According to the model proposed in Ref. 12, the excess free volume ΔVF of the atoms at the grain boundaries can be used to describe the degree of lattice strain in nanocrystalline. The dimensionless value of ΔVF can be expressed as ΔVF = [(dc + h/2)2 − d2c]/d2c, where h (= 1 nm) is the thickness of the grain boundary plane assumed to be independent of dc.12 As does the variation of dc, ΔVF increases gradually until dc reduces to 8.8 nm (15 h) and does not change significantly with the further decrease of dc until 30 h, as shown in Fig. 2.
Figure 3 shows the DSC traces of the milled Ti powders at different milling times. The 10 h milled Ti sample has two endothermic peaks within 550 to ~680 K and 800 to ~1060 K, respectively. The former is probably associated with the relaxation of mechanical stresses of the powders. As indicated in Figs. 4(e) and 5, the latter is attributed to grain growth of hexagonally close-packed (hcp) Ti. As the milling time increases, the intensities of the former decrease gradually and vanish after 60 h of the milling time. A new exothermic peak appears within 630 to ~720 K in the sample that was milled for 10 h [Fig. 3(a)], indicating formation of an amorphous phase. With the milling time, the area of the exothermic peak increases gradually, suggesting an increased amount of the amorphous phase. At 60 h of the milling time, the area reaches the maximum, further confirming that the milled sample is fully amorphous, as seen in Figs. 1 and 3. The crystallization temperature Tx (defined as the start exothermic transformation temperature of the DSC trace) of the 60 h milled amorphous sample is about 640 K, which is at least 100 K lower than those of multicomponent Ti-based amorphous alloys prepared by melt quenching and MA.13, 14 The crystallization heat of the Ti metal glassy sample is 152 J/g. Figure 4 shows in situ high-temperature XRD patterns of the Ti metal glassy sample under continuous heating. It is found that the Ti metal glassy sample transforms back to hcp Ti primarily at 593 K. The Tx observed from in situ XRD patterns is 47 K lower than that determined from DSC [Figs. 3(c) and 4(c)]. The lower Tx may be due to the temperature uncertainty and thermal treatment effect of longer time resulted from recording in situ XRD patterns below real Tx.
Generally, amorphization of single metallic Ti is impossible energetically and kinetically by melt quenching and MA, and even under high temperature and high pressure.15, 16 In the present case, the amorphization of metallic Ti may be related to the pickup impurity. Table I presents the results of chemical analysis for the starting Ti powders and the 60 h milled powders. After exposure to air many times, O and N contents of the 60 h milled Ti powders are 7.44 and 1.63 at.%. Meanwhile, Fe, Ni, and Cr pickups from vial and grinding medium is about 1.44 at.%. Because of the trace amounts, the effect of Ni, Cr, Cl, C, and Si on the amorphization of metallic Ti can be ignored. Fe is helpful for the amorphization of metallic Ti, and it can form amorphous FeTi2 phase with Ti.17 As interstitial atoms and stabilizer of hexagonally close-packed (hcp) Ti, minority N cannot exert significant influences on the amorphization of metallic Ti according to Ti–N phase diagram.18 As a result, the most important impurity atom, in our case, O, may play a key role on the amorphization of metallic Ti.
It is well accepted that the diffusional asymmetry and the negative heat of mixing are favorable for forming binary amorphous alloys by MA.19 Furthermore, if the accumulated energy is high enough to reach the free energy corresponding to the amorphous state, amorphization of the binary alloy system with positive heat of mixing is also possible.20 Except for the consumed Ti for the amorphization of Fe, the ratio of O to the remaining Ti in at.% is 8.6%. Although the Ti–O phase diagram at 8.6 at.% O below 673 K is uncertain (Fig. 5),21 we can try to use the phase diagram to explain the amorphization of metallic Ti in the present work, which in turn can verify the deduction in the Ti–O phase diagram. According to the Ti–O phase diagram at 8.6 at.% O,21 the system has a possibility of coexistence of αTi (hcp Ti) and α″ (Ti3O ordered phases) between 473 and ~673 K, as indicated by line segment “AB” in Fig. 5. Since the measured temperature of the exterior wall of the vial reaches to about 413 K at interrupting of every 5 h, it is possible that the temperature of the powders during MA has exceeded 473 K due to temperature difference between the interior and exterior of the vial. If it is true, the system during MA of the later several 5 h of the milling time would locate in the two-phase zone of αTi and α″. Because the volume of an αTi atom is 35.30 and the volume of an α″ unit cell is no less than 211.98,22, 23 their radius difference reaches 81.8%. This meets the often-cited empirical rule for bulk amorphous alloy formation on large atomic size difference above 12% and the diffusional asymmetry for forming binary amorphous alloys by MA.19, 24 The possibility of coexistence of αTi and α″ phase in Ti–O phase diagram and the rationality of our explanation about the amorphization of metallic Ti can be verified indirectly by (i) the good agreement between the Tx determined from DSC measurement for the 60 h milled glassy sample [Fig. 3(c)] and the transformation temperature of αTi at 8.6 at.% O in the Ti–O phase diagram (as indicated by line segment “BC” in Fig. 5), (ii) the primarily precipitated phase of hcp Ti observed from in situ XRD patterns [Fig. 4(c)], and (iii) the coexisting of hcp Ti and Ti3O phases after SPS treatment for the 60 h milled glassy sample [Fig. 4(f)].23
On the other hand, the amorphization of metallic Ti in our case can be explained energetically. During MA, the absorbed O on the surface layer are pumped by the successive collisions into deeper surface layers and dissolved into lattice of Ti as an interstitial atom by interdiffusion.25 According to a free-energy diagram of Ti–O and empirical principle of forming an amorphous phase during melt quenching, the introduced O decreases free energy of the system and favors the amorphization of metallic Ti.24, 26 Additionally, due to the decrease of dc, increase of ΔVF, and accumulation of structural defects, lattice strain would intensify and strain energy would increase, which contributes to the increase in free energy of the system and drives the crystal-to-amorphous transformation due to the Gibbs–Thompson effect.27 For the present case, at 30 h of the milling time, the dc decreases to 8.2 nm and the ΔVF increases consequently to 0.13; the corresponding negative (from core to periphery) hydrostatic pressure (−Ph) on the grain boundary of nanocrystalline Ti is about 5.9 GPa.9, 12 The combined effect of significant increases in −Ph and ΔVa, which is the volume difference between an hcp Ti and an amorphous Ti, is equivalent to strain energy increment.9 Because volume change between crystalline solid and amorphous alloy can be adopted as 1%,28 ΔVa is determined to be about 1.8 × 10−31 m3/atom. So, the strain energy increment is −Ph × ΔVa = 1.1 × 10−21 J/atom, which is near the theoretically predicted enthalpy change in structural transformation of a hcp Ti → fcc Ti.29 The increased strain energy may exceed the free energy needed for amorphization of the Ti–O system, which consequently induced the amorphization of metallic Ti in our case.
The present 8.6 at.% O induced amorphization of metallic Ti is different from Ref. 9, which reported that although adsorbed O content from the air reaches about 16.5 at.% O, hcp Ti transformed partly into face-centered cubic (fcc) Ti after 150 h milling. Considering that the system in Ref. 9 locates in the zone of a single α″ phase in the Ti–O phase diagram (Fig. 5), amorphization of metallic Ti did not take place due to not meeting the diffusional asymmetry for forming amorphous alloys by MA.19
The synthesis of metal Ti glassy phase with 7.44 at.% O was realized by MA in our work. Special physical mechanism of SPS exerted on powders caused the coexistence of hcp Ti and Ti3O in the SPSed metal Ti glassy sample, which did not emerge in the in situ XRD patterns. To our best knowledge, this is the first time that an O-containing metal glassy phase with a single metallic element is synthesized by ball milling of pure metal powders. Because an amorphous phase with a single metallic element cannot be now obtained even under high temperature and high pressure,15, 30 the amorphization of metallic Ti in our study may provide some insights into understanding the nature of amorphous transformation and interactions between solids and gases. In addition, our experimental results may verify indirectly the possibility of coexistence of αTi and α″ phase in the Ti–O phase diagram.
In conclusion, the oxygen-induced amorphization of metallic titanium by ball milling was obtained by exposing the powders to air at selected intervals. The obtained Ti metal glassy phase possesses the crystallization temperature of about 640 K and oxygen impurity content of 7.44 at.%. The present experimental results demonstrated the important effects of introduced oxygen impurity in small amounts on structural transformation of the starting powders during MA.
C. Suryanarayana: Mechanical alloying and milling. Prog. Mater. Sci. 46, 1 2001
M. Sherif El-Eskandarany: Mechanical Alloying for Fabrication of Advanced Engineering Materials, edited by M.S. El-Eskandarany Noyes Publications/William Andrew Publishing, New York 2001 1
A. Calka: Formation of titanium and zirconium nitrides by mechanical alloying. Appl. Phys. Lett. 59, 1568 1991
K. Aoki, A. Memezawa T. Masumoto: Nitrogen-induced amorphization of Ti–Zr powders during mechanical alloying. Appl. Phys. Lett. 61, 1037 1992
C.C. Koch, O.B. Cavin, C.G. McKamey J.O. Scarbroughy: Preparation of “amorphous” Ni60Nb40 by mechanical alloying. Appl. Phys. Lett. 43, 1017 1983
M.S. Kim C.C. Koch: Structural development during mechanical alloying of crystalline niobium and tin powders. J. Appl. Phys. 62, 3450 1987
A. Kübler, J. Eckert, A. Gebert L. Schultz: Influence of oxygen on the viscosity of Zr-Al-Cu-Ni metallic glasses in the undercooled liquid region. J. Appl. Phys. 83, 3438 1998
I. Lucks, P. Lamparter E.J. Mittemeijer: Uptake of iron, oxygen and nitrogen in molybdenum during ball milling. Acta Mater. 49, 2419 2001
I. Manna, P.P. Chattopadhyay, P. Nandi, F. Banhart H-J. Fecht: Formation of face-centered-cubic titanium by mechanical attrition. J. Appl. Phys. 93, 1520 2003
M. Tokita: Trends in advanced SPS spark plasma sintering system and technology. J. Soc. Powder Technol. Jpn. 30, 790 1993
Y.H. Zhao, H.W. Sheng K. Lu: Microstructure evolution and thermal properties in nanocrystalline Fe during mechanical attrition. Acta Mater. 49, 365 2001
P.P. Chattopadhyay, P.M.G. Nambissan, S.K. Pabi I. Manna: Polymorphic bcc to fcc transformation of nanocrystalline niobium studied by positron annihilation. Phys. Rev. B 63, 054107 2001
G. He, J. Eckert M. Hagiwara: Glass-forming ability and crystallization behavior of Ti–Cu–Ni–Sn–M (M = Zr, Mo, and Ta) metallic glasses. J. Appl. Phys. 95, 1816 2004
L.C. Zhang, J. Xu E. Ma: Mechanically alloyed amorphous Ti50(Cu4.5Ni4.5)44-XAlXSi4B2 alloy with supercooled liquid region. J. Mater. Res. 17, 1743 2002
T. Hattori, H. Saitoh, H. Kaneko, Y. Okajima, K. Aoki W. Utsumi: Does bulk metallic glass of elemental Zr and Ti exist? Phys. Rev. Lett. 96, 255504 2006
C. Yang, R.P. Liu, Z.J. Zhan, L.L. Sun, J. Zhang, Z.Z. Gong W.K. Wang: Formation of ZrTiCuNiBe bulk metallic glass by shock-wave quenching. Appl. Phys. Lett. 87, 051904 2005
Binary Alloy Phase Diagrams, edited by S. Nagasaki and M. Hirabayashi in Japanese, AGNE Gijutsu Center, Co., Ltd., Tokyo, Japan 2002 translated into Chinese by A.S. Liu Metallurgical Industry Press, Beijing, China 2004 160
Binary Alloy Phase Diagrams, translated into Chinese by A.S. Liu (Ref. 17) 315
E. Hellstern L. Schultz: Amorphization of transition metal Zr alloys by mechanical alloying. Appl. Phys. Lett. 48, 124 1986
K. Sakurai, Y. Yamada, M. Ito, C.H. Lee, T. Fukunaga U. Mizutani: Observation of solid-state amorphization in the immiscible system Cu-Ta. Appl. Phys. Lett. 57, 2660 1990
Binary Alloy Phase Diagrams, translated into Chinese by A.S. Liu (Ref. 17) 320
JCPDS File No. 44-1294
JCPDS File Nos. 72-1806, 73-1117, 73-1583, and 76-1644, 1998
A. Inoue: Stabilization of metallic supercooled liquid and bulk amorphous alloys. Acta Mater. 48, 279 2000
R. Schultz, M. Trudeau, J.Y. Huot A.V. Neste: Interdiffusion during the formation of amorphous alloys by mechanical alloying. Phys. Rev. Lett. 62, 2849 1989
W. Mo, G.Z. Deng F.C. Luo: Titanium Metallurgy Metallurgical Industry Press, Beijing, China 2004 36
R.B. Schwarz C.C. Koch: Formation of amorphous alloys by the mechanical alloying of crystalline powders of pure metals and powders of intermetallics. Appl. Phys. Lett. 49, 146 1986
C. Nagel, K. Rätzke, E. Schmidtke, J. Wolff, U. Geyer F. Faupel: Free-volume changes in the bulk metallic glass Zr46.7Ti8.3Cu7.5Ni10Be27.5 and the undercooled liquid. Phys. Rev. B. 57, 10224 1998
L. Kaufman: Phase Stability in Metals and Alloys, edited by P.S. Rudman, J. Stringer, and R.I. Jaffe McGraw-Hill, London, British 1966 125
Y. Wang, Y.Z. Fang, T. Kikegawa, C. Lathe, K. Saksl, H. Franz, J.R. Schneider, L. Gerward, F.M. Wu, J.F. Liu J.Z. Jiang: Amorphouslike diffraction pattern in solid metallic titanium. Phys. Rev. Lett. 95, 155501 2005
This work was supported by the National Science Fund of China for Distinguished Young Scholar (No. 50325516), the Science and Technology Program of Guangdong Province (No. 59872024), China Postdoctoral Science Foundation (No. 20060390198), and Postdoctoral Innovation Foundation of South China University of Technology (No. 05243). The authors are very grateful to Prof. S. Li, Dr. L.Z. Ouyang, and Prof. Y.Z. Liu for their helpful discussions and to Miss X.P. Zeng for her technical assistances in the XRD measurements.
About this article
Cite this article
Li, Y., Yang, C., Chen, W. et al. Oxygen-induced amorphization of metallic titanium by ball milling. Journal of Materials Research 22, 1927–1932 (2007). https://doi.org/10.1557/jmr.2007.0243