- Email: [email protected]
Crystallization and thermal stability of mechanically alloyed W–Ni–Fe noncrystalline materials
Materials Science and Engineering A315 (2001) 166– 173 www.elsevier.com/locate/msea
Crystallization and thermal stability of mechanically alloyed W–Ni–Fe noncrystalline materials Z. He, T.H. Courtney * Department of Materials Science and Engineering, Michigan Technological Uni6ersity, Houghton, MI 49931, USA Received 1 August 2000; received in revised form 3 January 2001
Abstract Tungsten–nickel noncrystalline alloys containing remnant nanocrystalline W particles have been synthesized by mechanical alloying (MA). Two initial compositions (50 and 75 at.% W) milled for various times have been investigated. The W content of the noncrystalline matrix depends on the MA (milling) time and the overall alloy W content. For 50 at.% W alloys, three exothermic reactions – monitored by differential scanning calorimetry – take place on continuous heating of the alloy. Crystalline W first precipitates from the noncrystalline matrix at about 825 K. This is followed by partial crystallization (to an fcc phase) of the noncrystalline matrix. At even higher temperatures, two intermetallics (NiW and, to a lesser extent, Ni4W) form from the remaining Ni–W noncrystalline matrix along with additional amounts of the fcc and W phases. Similar crystallization processes occur for 75 at.% W alloys milled for less than 15 h. However, when milled for times exceeding 15 h, only one exothermic reaction occurs in these alloys. The crystallization temperature is found to increase with increasing W content in the noncrystalline phase. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Mechanical alloying; Metallic glasses; Crystallization
1. Introduction Metallic glasses fail in tension following formation and propagation of a single shear band [1,2]. As a consequence, tensile ductility of metallic glasses is limited. Various efforts have been conducted to beneficially enhance the tensile ductility and macroscopic toughness of metallic glasses. Examples include plate and fiber composites using a metallic glass as one of their constituents, and development of two-phase crystalline –noncrystalline alloys [3 – 6]. In each circumstance, multiple shear banding is expected on the basis of the constraint the other constituent places on shear band propagation. Fine crystalline particles can also be introduced into glassy matrices by controlled precipitation from the matrix and, in some circumstances, toughness is enhanced thereby [7 – 9]. Mechanical alloying (MA) can be utilized to produce metallic glasses containing imbedded nanocrystalline particles [10]; e.g., * Corresponding author. Tel.: + 1-906-4872036; fax: +1-9064872934. E-mail address: [email protected] (T.H. Courtney).
W–Ni –Fe alloys having such a structure have been synthesized via MA of elemental powders [11]. Could such a structure be produced in bulk form, the resulting properties might be of interest. For example, high W content Ni –Fe –W alloys are potential kinetic energy penetrators. However, their use is restricted due to their high temperature flow behavior, which results in blunting of the penetrator tip. A W-based metal –metallic glass composite might behave differently. Mechanically alloyed products are powders, and must be consolidated to be used in bulk form. Such consolidation often results in crystallization of glassy materials at the (usually elevated) consolidation temperature. Identification of crystallization tendencies is required to establish processing routes that may maintain the noncrystalline structure during consolidation. We also note that crystallization processes taking place during heating of glassy metals need not always culminate in either equilibrium phases or in complete crystallization [12,13]. This suggests the possibility of generating microstructures having interesting properties by appropriate thermal manipulation during glassy metal consolidation. Regardless of the path to equi-
0921-5093/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 5 0 9 3 ( 0 1 ) 0 1 1 4 8 - 0
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173 Table 1 Fe content in the as-milled powdera
167
2. Experimental procedure
50 At.% W samples
75 At.% W–Ni samples
Milling time (h)
Fe content (at.%)
Milling time (h)
Fe content (at.%)
10 15 20 25 30 40
7.9 8.5 (*), 10.6 8.3 (*) 8.2 (*), 7.9 14.7 12.8
5 6 7 8 10 15 20
7 11 13 13 18.3 20.7 19.5
a
Obtained by XRF except for composition with (*). They were estimated by measuring the mass loss of the vial and grinding media following milling.
librium, a systematic study of the thermal stability of a specific glassy metal is required to identify promising processing routes. This paper provides the results of a study on the thermal stability of W– Ni – Fe glassy materials containing nanometer-sized crystalline W particles. In the paper, we emphasize the effects that alloy composition and MA time have on the crystallization process and its kinetics.
A few comments on the development of noncrystalline structures in mechanically alloyed Ni–W alloys are in order; details can be found in [11]. During MA of elemental mixtures of Ni and W, the latter dissolves in the Ni-rich fcc phase. If the W content is less than about 28 at.%, a supersaturated crystalline solid solution develops after sufficiently long milling times. However, for W contents exceeding about 28 at.%, the fcc phase amorphizes when the W content in it exceeds this value. With further increases in W content, remnant W particles are present in the glassy matrix. Initial tungsten contents of the Ni–W alloys discussed here range from 50 to 75 at.%. Thus, the mechanically alloyed powders have a glassy matrix containing nanocrystalline W particles. The amount of this crystalline phase is a function of overall alloy W content as well as the milling time. Elemental Ni and W powders (− 325 mesh, purity \99%), were starting materials. Samples of two compositions (50 and 75 at.% W) were mechanically alloyed in a SPEX mill at room temperature. Previous work in our laboratory has shown that the 50 at.% W sample could be amorphized by milling for 10 h. Thus, the 50 at.% W material was milled for times between 10 and 40 h. Amorphization occurs more rapidly for alloys having a higher initial W composition. For 75 at.% W
Fig. 1. (a) XRD patterns for 50 at.% W powder milled for different times. Volume fraction of crystalline W and the estimated W atomic fraction in the noncrystalline phase as a function of milling time for the (b) 50 at.% W alloy and (c) 75 at.% W alloy.
168
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
Fig. 2. DSC trace (heating rate = 10 K min − 1) for the 50 at.% W sample milled for 40 h. The arrows indicate the apparent crystallization reactions and their associated temperatures.
Fig. 3. XRD patterns illustrating phase evolution in the 50 at.% W alloy powders milled for 40 h and heated to various temperatures.
samples, 5 h of milling results in powder amorphization; therefore, 75 at.% W samples were milled for times between 5 and 20 h. Before milling, powders were loaded in a stainless steel vial in a Ar-filled glove box. 440-C stainless steel balls were used as the grinding media and the mass ratio of the balls to the powder was 9:1. In order to minimize powder sticking to the vial wall, the vial axis was rotated 180° after every hour of milling. After milling for 5 h, the vial was removed to the argon-filled glove box, and any powder attached to the vial wall was ‘decaked’. Due to mill abrasion, Fe contamination of the powders occurs [14]. The extent of this was assessed by X-ray fluorescence spectroscopy and/or by measuring the mass loss of the milling vial and grinding balls after milling: Fe contents derived in these manners are given in Table 1. It has been reported that Fe does not significantly affect the solution and amorphization kinetics during MA of Ni– W [11]. Furthermore, Fe and Ni chemically behave similarly in these alloys. For example, Fe is found only in the amorphous phase and not in the W as amorphization proceeds [11]. And, on crystallization of the amorphous powders, Fe and Ni are present in proportions to their overall composition in the fcc crystalline product and in the intermetallics that form during crystallization [15]. Even though the Fe contamination reduces the W contents of the alloys, for convenience, we refer to the
alloys as 50 at.% W and 75 at.% W alloys throughout this paper. Following milling, the powders were examined by X-ray diffraction (XRD) conducted in a Scintag diffractometer using Cu Ka radiation. The Ni–W solid solution was taken as fully amorphous when the crystallite size estimated using the Scherrer line-broadening formula was less than 5 nm. Diffraction patterns corresponding to this situation are shown later. Crystallization was monitored using differential scanning calorimetry (DSC) at various heating rates, although 10 K min − 1 was the standard heating rate employed. The DSC unit was calibrated using the melting temperatures of In, Sn, Zn, Al, and BaCO3. In addition to continuous heating to the highest temperature employed (850°C), other samples were heated to intermediate temperatures. After cooling, these were examined by XRD to determine phase evolution during heating. The volume fraction of crystalline W in the powders was also measured by XRD, employing the internal standard method; details are given in Appendix A. The microstructure was evaluated by backscattered scanning electron microscopy (SEM). As a complement to XRD, transmission electron microscopy (TEM) was also selectively utilized.
3. Results and discussion Previous work has shown that W dissolves in the Ni-rich fcc phase during MA. This increases the lattice parameter of the fcc phase, and the Ni {1 1 1} peak shifts to a lower diffraction angle. Following amorphization of the fcc phase, the XRD pattern shows a broad amorphous peak along with the diffraction peaks of residual crystalline W (Fig. 1a). No diffraction peaks of Fe can be identified in the XRD patterns, and the solubility of Fe in crystalline W is limited. Thus, as mentioned, Fe can be considered to dissolve entirely into the Ni-rich fcc and/or amorphous phases. The microstructure of the milled powder is characterized by nanocrystalline W particles (15–20 nm in diameter) embedded in the glassy matrix. Volume fractions of the residual W were estimated using the internal standard method. Results are shown in Figs. 1b and c for the different W content alloys. As expected, the amount of residual W decreases with milling time. Knowing the amount of the crystalline W, a mass balance permits estimation of the W content in the amorphous phase. Results are also shown in Fig. 1b and c.
3.1. 50 at.% W alloys Thermal stability and phase evolution during heating of 50 at.% W milled powders were examined by a
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
169
Fig. 4. Crystalline W volume fraction at different temperatures for 50 at.% W alloys milled for 40 h. The vertical lines indicate the onset temperature of each reaction.
combination of DSC and XRD. A DSC trace for the samples milled for 40 h is shown in Fig. 2. Three exothermic peaks, suggesting three reactions, are found. Detailed analysis was conducted on the 50 at.% W alloys milled for 40 h. The XRD patterns of these materials heated to different temperatures (heating rate= 10 K min − 1) are shown in Fig. 3. At 570°C, the onset temperature for the first reaction, the diffraction pattern is identical to that of the as-milled powder; bcc W peaks and a diffuse amorphous ‘‘peak’’ are found. However when the exposure temperature is increased, the W volume fraction increases. For example, it increases from about 10 vol.% (characteristic of the milled powder) at 550°C to about 20% at 600°C (a temperature between the first and second peak in the DSC curve), as shown in Fig. 4. The XRD pattern of Fig. 3 indicates that there remain two phases – crystalline W and a glassy phase – after the first reaction. The increase in crystalline W in the first reaction indicates that it precipitates from the glass during this reaction. A mass balance results in an estimate that the W content of the glassy phase following this first reaction is about 31 at.%. (This composition is still greater than both the equilibrium solubility (about 13 at.%, [16]) and the critical composition (28 at.%) required for amorphization). Fig. 5 is a TEM selected-area-diffraction pattern of the material heated to 620°C. The pattern also indicates the presence of an amorphous phase. The onset temperature of the second reaction is about 650°C. The XRD pattern of Fig. 3 shows that for the sample heated to 680°C, a crystalline fcc {1 1 1} peak is present in addition to W. This suggests crystallization of some of the glassy phase takes place during the second reaction. The W content of the fcc phase (about 19% based on the lattice parameter of the fcc phase and employing Vegard’s law) is not greatly in excess of the equilibrium solubility. However, Fig. 4 indicates that the crystalline W volume fraction hardly
changed as a result of the second reaction, and it should increase if all of the glassy phase had crystallized to a composition of 19 at.% W. On this basis, we suggest that the second reaction can be attributed to only partial crystallization of the glassy phase. The TEM diffraction pattern of Fig. 6, indicating the pres-
Fig. 5. TEM diffraction pattern of the 50 at.% W alloy milled for 40 h and then heated to 620°C. The broad diffuse ring is indicative of a noncrystalline phase.
Fig. 6. TEM diffraction pattern of the 50 at.% W alloy milled for 40 h and then heated to 680°C. In addition to crystalline diffraction spots, a diffuse ring – indicative of a noncrystalline phase – is observed.
170
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
Three phases – W (white), NiW (gray), and the fccphase (black) – are evident. The microstructural scale is very fine, on the order of 0.1 mm.
3.2. 75 at.% W alloys
Fig. 7. TEM diffraction pattern of the 50 at.% W alloy milled for 40 h and heated to 850°C. The indexing corresponds to the intermetallic Ni4W.
XRD (results shown later) of the 75 at.% W alloy show that the final crystallization products are the same in this alloy as in the 50 at.% W alloy. However, DSC indicates that the crystallization sequence of the 75 at.% W alloy depends on the milling time. Specifically, two different crystallization sequences are found as indicated in Fig. 9. For the sample milled for 10 h, the same three-stage crystallization process described above occurs. But for the sample milled for 20 h, only one exothermic peak is found in the DSC curve. The ‘‘transition’’ milling time for the different situations is about 15 h. Two samples (milled for 10 and 20 h, respectively) representing the different crystallization behaviors are compared and analyzed here.
3.2.1. 75 at.% W alloys milled for 10 h XRD patterns (Fig. 10) for the 75 at.% W powders milled for 10 h and heated to various temperatures parallel those of the 50 at.% W alloy (cf. Fig. 3). Likewise, as indicated in Fig. 9, three exothermic peaks are observed for the 75 at.% W alloy milled for 10 h.
Fig. 8. Backscattered scanning-electron micrograph of the 50 at.% W alloy milled for 40 h and heated to 850°C. Tungsten is light, NiW gray, and the Ni-rich phase is black. The microstructural scale is on the order of 0.1 mm.
ence of both a glassy and a crystalline phase, supports this view. Following the third reaction, two intermetallic phases (NiW and Ni4W) are present in addition to W and a fcc phase. From the XRD pattern of Fig. 3, NiW can be clearly identified, but the presence of Ni4W is not obvious. This could be a result of the amount of this phase being very small or because the strongest diffraction peaks of Ni4W [17] superimpose on those of the fcc phase. However, the presence of Ni4W was confirmed by TEM (Fig. 7). The amount of crystalline W also increases in the third reaction (Fig. 4). The increase might result from precipitation from the supersaturated Ni-rich fcc phase (the product of the second reaction) or from precipitation of W during crystallization of the noncrystalline phase that remained subsequent to the second reaction. The backscattered SEM micrograph of Fig. 8 illustrates the phase morphology in powders heated to 850°C.
Fig. 9. DSC traces for the 75 at.% W alloys milled for (a) 10 h and (b) 20 h. The arrows indicate the crystallization reaction(s) and their associated temperature.
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
171
Fig. 10. XRD patterns illustrating phase evolution in the 75 at.% W alloy powders milled for 10 h and heated to various temperatures.
Although the first broad low-temperature exothermic peak for the 75 at.% W alloy milled for 10 h is not striking, the amount of crystalline W is found to increase from about 25 to 40 vol.% as the temperature is increased to 650°C (Fig. 11). Fig. 11 also reveals the volume fraction of W increases (from 40 to 50 vol.%) in the third reaction. However, this volume fraction does not change during the second reaction. These observations suggest that this alloy crystallizes in rather much the same way as the 50 at.% W alloys.
3.2.2. 75 at.% W milled for 20 h For the 75 at.% W alloys milled for times longer than 15 h, DSC detects only one reaction during heating to 850°C (Fig. 9b). XRD (Fig. 12) shows that crystallization products are not present until the temperature exceeds 750°C. Fig. 13 also indicates that the W volume fraction does not increase until this temperature is exceeded. Fig. 14 plots the first reaction temperature peak as it varies with W content in the precursor glassy phase. This temperature increases significantly with W content (and with milling time) for both original alloy compositions. It is noteworthy that the first peak reaction temperature seems to correlate with the W glassy-phase composition; i.e., this temperature increases monotonically with this W composition and does not directly correlate with the overall alloy W alloy content. The peak temperatures of the final crystallization stage are shown in Fig. 15.1 The abscissa of Fig. 15a is the as-milled glassy-phase W content. When the third peak temperature is plotted vs. this composition the data more-or-less fall into two groups depending on the overall alloy W content; the third peak reaction temper1 Peak temperatures for the second reaction are not shown here because the second reaction often overlaps with the third one. This makes difficult a precise measurement of the second reaction temperature.
atures are higher for the 75 at.% W alloys than for the 50 at.% W alloys. The abscissa of Fig. 15b is the W content in the glassy phase that is precursor to the third reaction. For the alloy milled for 20 h, this W content is that in the as-milled powder since there are no crystallization reactions prior to the high-temperature one. For alloys milled for 10 h or less, the abscissa represents the W glassy phase content following the first reaction as determined employing a mass balance procedure. When plotted on this basis, Fig. 15b indicates a striking correlation between the temperature of the final crystallization reaction and the W content in the glassy phase preceding this reaction.
4. Conclusions The experimental results suggest that crystallization of glassy 50 at.% W alloys containing remnant nanocrystalline W particles synthesized via MA takes place in stages as summarized below: Stage I: Depletion of the W content of the glassy phase via precipitation of W or by growth of the preexisting crystalline W particles.
Fig. 11. Crystalline W volume fraction at different temperatures for 75 at.% W alloys milled for 10 h. The vertical lines indicate the onset temperature of each reaction. A three-stage crystallization sequence, similar to that found for the 50 at.% W alloys, is followed.
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
172
Fig. 12. XRD patterns illustrating phase evolution in the 75 at.% W alloy powders milled for 20 h and heated to various temperatures.
Fig. 13. Crystalline W volume fraction at different temperatures for 75 at.% W alloys milled for 20 h.
Stage II: Partial crystallization of the glassy phase. The crystallization product is a W supersaturated Nirich fcc phase. Stage III: Final crystallization of the glassy phase. Following Stage III, the structure consists of a fcc phase, W, and NiW. Minor amounts of Ni4W also appear to be present following this stage. For 75 at.% W alloys, the final crystallization products are the same as for the 50 at.% W alloy. Moreover, for short milling times crystallization also proceeds in three stages for the 75 at.% W alloys. However, for longer milling times (and higher W glassy-phase contents) only one crystallization event is observed.
Appendix A A method similar to the internal standard method [18] was used to determine the crystalline W volume fraction. The W {2 0 0} peak was used since no other peaks overlap with this one. The integrated intensity, I, of a specific W diffraction peak is given by I=KV/v,
Acknowledgements Professor Steven A. Hackney provided instruction and help with the TEM work. Professor Douglas Swenson and Mr. Sami Syed were of great aid in the DSC component of the study. This work was supported by the Army Research Office, Dr William Mullins grants monitor.
Fig. 14. First reaction peak-temperature vs. W content in the glassy phase in the as-milled powders. (The times shown are the milling times.) The reaction temperature appears to be directly related to the initial glassy-phase W content.
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
173
experiments were conducted with a constant irradiated volume. A similar expression holds for V1 [=6w1/ (6w1 + 6a0)= 6w1/60, where 6w1 is the volume of crystalline W in the 100 g sample to which 10 g of additional W has been added]. To obtain the volume 60, we must assume the atomic volumes in the phases are the same. This is not correct, but any error (several percent or so) is likely within the experimental error of the measured volume ratios. On this basis 60 =Mw/ zw + MNi/zNi + MFe/zFe, where Mw, MNi, and MFe are the weight percentages of W, Ni, and Fe in the initial sample, respectively. To obtain V1, 6w1 can be expressed as d6w + 6w0; thus, 6w1 + 6a0 = 60 + d6w, where d6w is the 1 volume of 10 g of pure W (=10 z − W ). Therefore, V1 [(d6w + 6w0)/(60 + d6w)] (1+ d6w/6w0) = = V0 [6w0/60] (1+d6w/60)
(A.1)
Eq. (A.1) leads to the following expression; 6w0 =
d6w . {[(V1/V0)(1+ d6w/60)] − 1}
(A.2)
Then, the volume fraction of crystalline W (V0) is obtained by dividing 6w0 by 60.
References Fig. 15. Peak temperatures for the high-temperature crystallization reaction vs. W. content (a) in the glassy phase in the as-milled powders and, (b) in the glassy phase present prior to the high-temperature reaction. (The times shown are milling times.)
where V is the volume fraction of the W and v is the absorption factor. The experimental procedure consists of two steps. First, the integrated intensity (I0 =KV0/v0) of the W {2 0 0} peak of the sample having volume fraction V0 is measured. The absorption factor used is a mass average absorption factor of W, Ni, and Fe; it is known since the sample composition is known. Ten weight per cent of pure W powder was then added to the sample and the integrated intensity (I1 =KV1/v1) of the W {2 0 0} peak for this sample is determined; the absorption factor v1 for this sample is also known. Taking the ratio I1/I0 permits V1/V0 to be determined. The volume fraction 60 is (considering a mixture of an amorphous phase and the crystalline W phase to illustrate) the ratio 6w0/(6w0 +6a0)= 6w0/60, where 6w0 is the volume of crystalline W, 6a0 the volume of the glassy phase, and 60 the total sample volume in a 100 g sample. Diffraction
.
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
[13]
[14] [15] [16] [17] [18]
L.A. Davis, S. Kavesh, J. Mater. Sci. 10 (1975) 453. C.A. Pampillo, Scr. Metall. 6 (1972) 915. Y. Leng, T.H. Courtney, Metall. Trans. A 21A (1990) 2159. Y. Leng, T.H. Courtney, J. Mater. Sci. 24 (1989) 2006. Y. Leng, T.H. Courtney, J. Mater. Sci. 26 (1991) 588. R.D. Conner, W.L. Johnson, Acta Mater. 46 (1998) 6089. Y.H. Kim, K. Hiraga, A. Inoue, T. Masumoto, Mater. Trans. JIM 35 (1994) 293. C. Fan, A. Inoue, Mater. Trans. JIM 38 (1997) 1040. A. Inoue, T. Zhang, Y.H. Kim, Mater. Trans. JIM 38 (1997) 749. J. Eckert, A. Kubler, L. Schultz, J. Appl. Phys. 85 (1999) 7112. A.O. Aning, Z. Wang, T.H. Courtney, Acta Metall. Mater. 41 (1993) 165. C.C. Koch, P.Y. Lee, in: E. Arzt, L. Schulz (Eds.), New Materials by Mechanical Alloying Techniques, DGM Informationsgesellschaft, Oberursel, 1989, p. 101. G. Veltl, B. Scholz, H.D. Kunze, in: E. Arzt, L. Schulz (Eds.), New Materials by Mechanical Alloying Techniques, DGM Informationsgesellschaft, Oberursel, 1989, p. 79. T.H. Courtney, Z. Wang, Scr. Metall. Mater. 27 (1992) 777. C.G. Mukira, Ph.D. Thesis, Michigan Technological University (1997). P. Nash (Ed.), Phase Diagrams of Binary Nickel Alloys, ASM International, Materials Park, OH, 1991, p. 369. E. Epremiam, D. Harker, Trans. AIME 185 (1949) 267. B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley, Reading, MA, 1977, p. 415.
Crystallization and thermal stability of mechanically alloyed W–Ni–Fe noncrystalline materials Z. He, T.H. Courtney * Department of Materials Science and Engineering, Michigan Technological Uni6ersity, Houghton, MI 49931, USA Received 1 August 2000; received in revised form 3 January 2001
Abstract Tungsten–nickel noncrystalline alloys containing remnant nanocrystalline W particles have been synthesized by mechanical alloying (MA). Two initial compositions (50 and 75 at.% W) milled for various times have been investigated. The W content of the noncrystalline matrix depends on the MA (milling) time and the overall alloy W content. For 50 at.% W alloys, three exothermic reactions – monitored by differential scanning calorimetry – take place on continuous heating of the alloy. Crystalline W first precipitates from the noncrystalline matrix at about 825 K. This is followed by partial crystallization (to an fcc phase) of the noncrystalline matrix. At even higher temperatures, two intermetallics (NiW and, to a lesser extent, Ni4W) form from the remaining Ni–W noncrystalline matrix along with additional amounts of the fcc and W phases. Similar crystallization processes occur for 75 at.% W alloys milled for less than 15 h. However, when milled for times exceeding 15 h, only one exothermic reaction occurs in these alloys. The crystallization temperature is found to increase with increasing W content in the noncrystalline phase. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Mechanical alloying; Metallic glasses; Crystallization
1. Introduction Metallic glasses fail in tension following formation and propagation of a single shear band [1,2]. As a consequence, tensile ductility of metallic glasses is limited. Various efforts have been conducted to beneficially enhance the tensile ductility and macroscopic toughness of metallic glasses. Examples include plate and fiber composites using a metallic glass as one of their constituents, and development of two-phase crystalline –noncrystalline alloys [3 – 6]. In each circumstance, multiple shear banding is expected on the basis of the constraint the other constituent places on shear band propagation. Fine crystalline particles can also be introduced into glassy matrices by controlled precipitation from the matrix and, in some circumstances, toughness is enhanced thereby [7 – 9]. Mechanical alloying (MA) can be utilized to produce metallic glasses containing imbedded nanocrystalline particles [10]; e.g., * Corresponding author. Tel.: + 1-906-4872036; fax: +1-9064872934. E-mail address: [email protected] (T.H. Courtney).
W–Ni –Fe alloys having such a structure have been synthesized via MA of elemental powders [11]. Could such a structure be produced in bulk form, the resulting properties might be of interest. For example, high W content Ni –Fe –W alloys are potential kinetic energy penetrators. However, their use is restricted due to their high temperature flow behavior, which results in blunting of the penetrator tip. A W-based metal –metallic glass composite might behave differently. Mechanically alloyed products are powders, and must be consolidated to be used in bulk form. Such consolidation often results in crystallization of glassy materials at the (usually elevated) consolidation temperature. Identification of crystallization tendencies is required to establish processing routes that may maintain the noncrystalline structure during consolidation. We also note that crystallization processes taking place during heating of glassy metals need not always culminate in either equilibrium phases or in complete crystallization [12,13]. This suggests the possibility of generating microstructures having interesting properties by appropriate thermal manipulation during glassy metal consolidation. Regardless of the path to equi-
0921-5093/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 1 - 5 0 9 3 ( 0 1 ) 0 1 1 4 8 - 0
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173 Table 1 Fe content in the as-milled powdera
167
2. Experimental procedure
50 At.% W samples
75 At.% W–Ni samples
Milling time (h)
Fe content (at.%)
Milling time (h)
Fe content (at.%)
10 15 20 25 30 40
7.9 8.5 (*), 10.6 8.3 (*) 8.2 (*), 7.9 14.7 12.8
5 6 7 8 10 15 20
7 11 13 13 18.3 20.7 19.5
a
Obtained by XRF except for composition with (*). They were estimated by measuring the mass loss of the vial and grinding media following milling.
librium, a systematic study of the thermal stability of a specific glassy metal is required to identify promising processing routes. This paper provides the results of a study on the thermal stability of W– Ni – Fe glassy materials containing nanometer-sized crystalline W particles. In the paper, we emphasize the effects that alloy composition and MA time have on the crystallization process and its kinetics.
A few comments on the development of noncrystalline structures in mechanically alloyed Ni–W alloys are in order; details can be found in [11]. During MA of elemental mixtures of Ni and W, the latter dissolves in the Ni-rich fcc phase. If the W content is less than about 28 at.%, a supersaturated crystalline solid solution develops after sufficiently long milling times. However, for W contents exceeding about 28 at.%, the fcc phase amorphizes when the W content in it exceeds this value. With further increases in W content, remnant W particles are present in the glassy matrix. Initial tungsten contents of the Ni–W alloys discussed here range from 50 to 75 at.%. Thus, the mechanically alloyed powders have a glassy matrix containing nanocrystalline W particles. The amount of this crystalline phase is a function of overall alloy W content as well as the milling time. Elemental Ni and W powders (− 325 mesh, purity \99%), were starting materials. Samples of two compositions (50 and 75 at.% W) were mechanically alloyed in a SPEX mill at room temperature. Previous work in our laboratory has shown that the 50 at.% W sample could be amorphized by milling for 10 h. Thus, the 50 at.% W material was milled for times between 10 and 40 h. Amorphization occurs more rapidly for alloys having a higher initial W composition. For 75 at.% W
Fig. 1. (a) XRD patterns for 50 at.% W powder milled for different times. Volume fraction of crystalline W and the estimated W atomic fraction in the noncrystalline phase as a function of milling time for the (b) 50 at.% W alloy and (c) 75 at.% W alloy.
168
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
Fig. 2. DSC trace (heating rate = 10 K min − 1) for the 50 at.% W sample milled for 40 h. The arrows indicate the apparent crystallization reactions and their associated temperatures.
Fig. 3. XRD patterns illustrating phase evolution in the 50 at.% W alloy powders milled for 40 h and heated to various temperatures.
samples, 5 h of milling results in powder amorphization; therefore, 75 at.% W samples were milled for times between 5 and 20 h. Before milling, powders were loaded in a stainless steel vial in a Ar-filled glove box. 440-C stainless steel balls were used as the grinding media and the mass ratio of the balls to the powder was 9:1. In order to minimize powder sticking to the vial wall, the vial axis was rotated 180° after every hour of milling. After milling for 5 h, the vial was removed to the argon-filled glove box, and any powder attached to the vial wall was ‘decaked’. Due to mill abrasion, Fe contamination of the powders occurs [14]. The extent of this was assessed by X-ray fluorescence spectroscopy and/or by measuring the mass loss of the milling vial and grinding balls after milling: Fe contents derived in these manners are given in Table 1. It has been reported that Fe does not significantly affect the solution and amorphization kinetics during MA of Ni– W [11]. Furthermore, Fe and Ni chemically behave similarly in these alloys. For example, Fe is found only in the amorphous phase and not in the W as amorphization proceeds [11]. And, on crystallization of the amorphous powders, Fe and Ni are present in proportions to their overall composition in the fcc crystalline product and in the intermetallics that form during crystallization [15]. Even though the Fe contamination reduces the W contents of the alloys, for convenience, we refer to the
alloys as 50 at.% W and 75 at.% W alloys throughout this paper. Following milling, the powders were examined by X-ray diffraction (XRD) conducted in a Scintag diffractometer using Cu Ka radiation. The Ni–W solid solution was taken as fully amorphous when the crystallite size estimated using the Scherrer line-broadening formula was less than 5 nm. Diffraction patterns corresponding to this situation are shown later. Crystallization was monitored using differential scanning calorimetry (DSC) at various heating rates, although 10 K min − 1 was the standard heating rate employed. The DSC unit was calibrated using the melting temperatures of In, Sn, Zn, Al, and BaCO3. In addition to continuous heating to the highest temperature employed (850°C), other samples were heated to intermediate temperatures. After cooling, these were examined by XRD to determine phase evolution during heating. The volume fraction of crystalline W in the powders was also measured by XRD, employing the internal standard method; details are given in Appendix A. The microstructure was evaluated by backscattered scanning electron microscopy (SEM). As a complement to XRD, transmission electron microscopy (TEM) was also selectively utilized.
3. Results and discussion Previous work has shown that W dissolves in the Ni-rich fcc phase during MA. This increases the lattice parameter of the fcc phase, and the Ni {1 1 1} peak shifts to a lower diffraction angle. Following amorphization of the fcc phase, the XRD pattern shows a broad amorphous peak along with the diffraction peaks of residual crystalline W (Fig. 1a). No diffraction peaks of Fe can be identified in the XRD patterns, and the solubility of Fe in crystalline W is limited. Thus, as mentioned, Fe can be considered to dissolve entirely into the Ni-rich fcc and/or amorphous phases. The microstructure of the milled powder is characterized by nanocrystalline W particles (15–20 nm in diameter) embedded in the glassy matrix. Volume fractions of the residual W were estimated using the internal standard method. Results are shown in Figs. 1b and c for the different W content alloys. As expected, the amount of residual W decreases with milling time. Knowing the amount of the crystalline W, a mass balance permits estimation of the W content in the amorphous phase. Results are also shown in Fig. 1b and c.
3.1. 50 at.% W alloys Thermal stability and phase evolution during heating of 50 at.% W milled powders were examined by a
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
169
Fig. 4. Crystalline W volume fraction at different temperatures for 50 at.% W alloys milled for 40 h. The vertical lines indicate the onset temperature of each reaction.
combination of DSC and XRD. A DSC trace for the samples milled for 40 h is shown in Fig. 2. Three exothermic peaks, suggesting three reactions, are found. Detailed analysis was conducted on the 50 at.% W alloys milled for 40 h. The XRD patterns of these materials heated to different temperatures (heating rate= 10 K min − 1) are shown in Fig. 3. At 570°C, the onset temperature for the first reaction, the diffraction pattern is identical to that of the as-milled powder; bcc W peaks and a diffuse amorphous ‘‘peak’’ are found. However when the exposure temperature is increased, the W volume fraction increases. For example, it increases from about 10 vol.% (characteristic of the milled powder) at 550°C to about 20% at 600°C (a temperature between the first and second peak in the DSC curve), as shown in Fig. 4. The XRD pattern of Fig. 3 indicates that there remain two phases – crystalline W and a glassy phase – after the first reaction. The increase in crystalline W in the first reaction indicates that it precipitates from the glass during this reaction. A mass balance results in an estimate that the W content of the glassy phase following this first reaction is about 31 at.%. (This composition is still greater than both the equilibrium solubility (about 13 at.%, [16]) and the critical composition (28 at.%) required for amorphization). Fig. 5 is a TEM selected-area-diffraction pattern of the material heated to 620°C. The pattern also indicates the presence of an amorphous phase. The onset temperature of the second reaction is about 650°C. The XRD pattern of Fig. 3 shows that for the sample heated to 680°C, a crystalline fcc {1 1 1} peak is present in addition to W. This suggests crystallization of some of the glassy phase takes place during the second reaction. The W content of the fcc phase (about 19% based on the lattice parameter of the fcc phase and employing Vegard’s law) is not greatly in excess of the equilibrium solubility. However, Fig. 4 indicates that the crystalline W volume fraction hardly
changed as a result of the second reaction, and it should increase if all of the glassy phase had crystallized to a composition of 19 at.% W. On this basis, we suggest that the second reaction can be attributed to only partial crystallization of the glassy phase. The TEM diffraction pattern of Fig. 6, indicating the pres-
Fig. 5. TEM diffraction pattern of the 50 at.% W alloy milled for 40 h and then heated to 620°C. The broad diffuse ring is indicative of a noncrystalline phase.
Fig. 6. TEM diffraction pattern of the 50 at.% W alloy milled for 40 h and then heated to 680°C. In addition to crystalline diffraction spots, a diffuse ring – indicative of a noncrystalline phase – is observed.
170
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
Three phases – W (white), NiW (gray), and the fccphase (black) – are evident. The microstructural scale is very fine, on the order of 0.1 mm.
3.2. 75 at.% W alloys
Fig. 7. TEM diffraction pattern of the 50 at.% W alloy milled for 40 h and heated to 850°C. The indexing corresponds to the intermetallic Ni4W.
XRD (results shown later) of the 75 at.% W alloy show that the final crystallization products are the same in this alloy as in the 50 at.% W alloy. However, DSC indicates that the crystallization sequence of the 75 at.% W alloy depends on the milling time. Specifically, two different crystallization sequences are found as indicated in Fig. 9. For the sample milled for 10 h, the same three-stage crystallization process described above occurs. But for the sample milled for 20 h, only one exothermic peak is found in the DSC curve. The ‘‘transition’’ milling time for the different situations is about 15 h. Two samples (milled for 10 and 20 h, respectively) representing the different crystallization behaviors are compared and analyzed here.
3.2.1. 75 at.% W alloys milled for 10 h XRD patterns (Fig. 10) for the 75 at.% W powders milled for 10 h and heated to various temperatures parallel those of the 50 at.% W alloy (cf. Fig. 3). Likewise, as indicated in Fig. 9, three exothermic peaks are observed for the 75 at.% W alloy milled for 10 h.
Fig. 8. Backscattered scanning-electron micrograph of the 50 at.% W alloy milled for 40 h and heated to 850°C. Tungsten is light, NiW gray, and the Ni-rich phase is black. The microstructural scale is on the order of 0.1 mm.
ence of both a glassy and a crystalline phase, supports this view. Following the third reaction, two intermetallic phases (NiW and Ni4W) are present in addition to W and a fcc phase. From the XRD pattern of Fig. 3, NiW can be clearly identified, but the presence of Ni4W is not obvious. This could be a result of the amount of this phase being very small or because the strongest diffraction peaks of Ni4W [17] superimpose on those of the fcc phase. However, the presence of Ni4W was confirmed by TEM (Fig. 7). The amount of crystalline W also increases in the third reaction (Fig. 4). The increase might result from precipitation from the supersaturated Ni-rich fcc phase (the product of the second reaction) or from precipitation of W during crystallization of the noncrystalline phase that remained subsequent to the second reaction. The backscattered SEM micrograph of Fig. 8 illustrates the phase morphology in powders heated to 850°C.
Fig. 9. DSC traces for the 75 at.% W alloys milled for (a) 10 h and (b) 20 h. The arrows indicate the crystallization reaction(s) and their associated temperature.
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
171
Fig. 10. XRD patterns illustrating phase evolution in the 75 at.% W alloy powders milled for 10 h and heated to various temperatures.
Although the first broad low-temperature exothermic peak for the 75 at.% W alloy milled for 10 h is not striking, the amount of crystalline W is found to increase from about 25 to 40 vol.% as the temperature is increased to 650°C (Fig. 11). Fig. 11 also reveals the volume fraction of W increases (from 40 to 50 vol.%) in the third reaction. However, this volume fraction does not change during the second reaction. These observations suggest that this alloy crystallizes in rather much the same way as the 50 at.% W alloys.
3.2.2. 75 at.% W milled for 20 h For the 75 at.% W alloys milled for times longer than 15 h, DSC detects only one reaction during heating to 850°C (Fig. 9b). XRD (Fig. 12) shows that crystallization products are not present until the temperature exceeds 750°C. Fig. 13 also indicates that the W volume fraction does not increase until this temperature is exceeded. Fig. 14 plots the first reaction temperature peak as it varies with W content in the precursor glassy phase. This temperature increases significantly with W content (and with milling time) for both original alloy compositions. It is noteworthy that the first peak reaction temperature seems to correlate with the W glassy-phase composition; i.e., this temperature increases monotonically with this W composition and does not directly correlate with the overall alloy W alloy content. The peak temperatures of the final crystallization stage are shown in Fig. 15.1 The abscissa of Fig. 15a is the as-milled glassy-phase W content. When the third peak temperature is plotted vs. this composition the data more-or-less fall into two groups depending on the overall alloy W content; the third peak reaction temper1 Peak temperatures for the second reaction are not shown here because the second reaction often overlaps with the third one. This makes difficult a precise measurement of the second reaction temperature.
atures are higher for the 75 at.% W alloys than for the 50 at.% W alloys. The abscissa of Fig. 15b is the W content in the glassy phase that is precursor to the third reaction. For the alloy milled for 20 h, this W content is that in the as-milled powder since there are no crystallization reactions prior to the high-temperature one. For alloys milled for 10 h or less, the abscissa represents the W glassy phase content following the first reaction as determined employing a mass balance procedure. When plotted on this basis, Fig. 15b indicates a striking correlation between the temperature of the final crystallization reaction and the W content in the glassy phase preceding this reaction.
4. Conclusions The experimental results suggest that crystallization of glassy 50 at.% W alloys containing remnant nanocrystalline W particles synthesized via MA takes place in stages as summarized below: Stage I: Depletion of the W content of the glassy phase via precipitation of W or by growth of the preexisting crystalline W particles.
Fig. 11. Crystalline W volume fraction at different temperatures for 75 at.% W alloys milled for 10 h. The vertical lines indicate the onset temperature of each reaction. A three-stage crystallization sequence, similar to that found for the 50 at.% W alloys, is followed.
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
172
Fig. 12. XRD patterns illustrating phase evolution in the 75 at.% W alloy powders milled for 20 h and heated to various temperatures.
Fig. 13. Crystalline W volume fraction at different temperatures for 75 at.% W alloys milled for 20 h.
Stage II: Partial crystallization of the glassy phase. The crystallization product is a W supersaturated Nirich fcc phase. Stage III: Final crystallization of the glassy phase. Following Stage III, the structure consists of a fcc phase, W, and NiW. Minor amounts of Ni4W also appear to be present following this stage. For 75 at.% W alloys, the final crystallization products are the same as for the 50 at.% W alloy. Moreover, for short milling times crystallization also proceeds in three stages for the 75 at.% W alloys. However, for longer milling times (and higher W glassy-phase contents) only one crystallization event is observed.
Appendix A A method similar to the internal standard method [18] was used to determine the crystalline W volume fraction. The W {2 0 0} peak was used since no other peaks overlap with this one. The integrated intensity, I, of a specific W diffraction peak is given by I=KV/v,
Acknowledgements Professor Steven A. Hackney provided instruction and help with the TEM work. Professor Douglas Swenson and Mr. Sami Syed were of great aid in the DSC component of the study. This work was supported by the Army Research Office, Dr William Mullins grants monitor.
Fig. 14. First reaction peak-temperature vs. W content in the glassy phase in the as-milled powders. (The times shown are the milling times.) The reaction temperature appears to be directly related to the initial glassy-phase W content.
Z. He, T.H. Courtney / Materials Science and Engineering A315 (2001) 166–173
173
experiments were conducted with a constant irradiated volume. A similar expression holds for V1 [=6w1/ (6w1 + 6a0)= 6w1/60, where 6w1 is the volume of crystalline W in the 100 g sample to which 10 g of additional W has been added]. To obtain the volume 60, we must assume the atomic volumes in the phases are the same. This is not correct, but any error (several percent or so) is likely within the experimental error of the measured volume ratios. On this basis 60 =Mw/ zw + MNi/zNi + MFe/zFe, where Mw, MNi, and MFe are the weight percentages of W, Ni, and Fe in the initial sample, respectively. To obtain V1, 6w1 can be expressed as d6w + 6w0; thus, 6w1 + 6a0 = 60 + d6w, where d6w is the 1 volume of 10 g of pure W (=10 z − W ). Therefore, V1 [(d6w + 6w0)/(60 + d6w)] (1+ d6w/6w0) = = V0 [6w0/60] (1+d6w/60)
(A.1)
Eq. (A.1) leads to the following expression; 6w0 =
d6w . {[(V1/V0)(1+ d6w/60)] − 1}
(A.2)
Then, the volume fraction of crystalline W (V0) is obtained by dividing 6w0 by 60.
References Fig. 15. Peak temperatures for the high-temperature crystallization reaction vs. W. content (a) in the glassy phase in the as-milled powders and, (b) in the glassy phase present prior to the high-temperature reaction. (The times shown are milling times.)
where V is the volume fraction of the W and v is the absorption factor. The experimental procedure consists of two steps. First, the integrated intensity (I0 =KV0/v0) of the W {2 0 0} peak of the sample having volume fraction V0 is measured. The absorption factor used is a mass average absorption factor of W, Ni, and Fe; it is known since the sample composition is known. Ten weight per cent of pure W powder was then added to the sample and the integrated intensity (I1 =KV1/v1) of the W {2 0 0} peak for this sample is determined; the absorption factor v1 for this sample is also known. Taking the ratio I1/I0 permits V1/V0 to be determined. The volume fraction 60 is (considering a mixture of an amorphous phase and the crystalline W phase to illustrate) the ratio 6w0/(6w0 +6a0)= 6w0/60, where 6w0 is the volume of crystalline W, 6a0 the volume of the glassy phase, and 60 the total sample volume in a 100 g sample. Diffraction
.
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]
[13]
[14] [15] [16] [17] [18]
L.A. Davis, S. Kavesh, J. Mater. Sci. 10 (1975) 453. C.A. Pampillo, Scr. Metall. 6 (1972) 915. Y. Leng, T.H. Courtney, Metall. Trans. A 21A (1990) 2159. Y. Leng, T.H. Courtney, J. Mater. Sci. 24 (1989) 2006. Y. Leng, T.H. Courtney, J. Mater. Sci. 26 (1991) 588. R.D. Conner, W.L. Johnson, Acta Mater. 46 (1998) 6089. Y.H. Kim, K. Hiraga, A. Inoue, T. Masumoto, Mater. Trans. JIM 35 (1994) 293. C. Fan, A. Inoue, Mater. Trans. JIM 38 (1997) 1040. A. Inoue, T. Zhang, Y.H. Kim, Mater. Trans. JIM 38 (1997) 749. J. Eckert, A. Kubler, L. Schultz, J. Appl. Phys. 85 (1999) 7112. A.O. Aning, Z. Wang, T.H. Courtney, Acta Metall. Mater. 41 (1993) 165. C.C. Koch, P.Y. Lee, in: E. Arzt, L. Schulz (Eds.), New Materials by Mechanical Alloying Techniques, DGM Informationsgesellschaft, Oberursel, 1989, p. 101. G. Veltl, B. Scholz, H.D. Kunze, in: E. Arzt, L. Schulz (Eds.), New Materials by Mechanical Alloying Techniques, DGM Informationsgesellschaft, Oberursel, 1989, p. 79. T.H. Courtney, Z. Wang, Scr. Metall. Mater. 27 (1992) 777. C.G. Mukira, Ph.D. Thesis, Michigan Technological University (1997). P. Nash (Ed.), Phase Diagrams of Binary Nickel Alloys, ASM International, Materials Park, OH, 1991, p. 369. E. Epremiam, D. Harker, Trans. AIME 185 (1949) 267. B.D. Cullity, Elements of X-Ray Diffraction, Addison-Wesley, Reading, MA, 1977, p. 415.