- Email: [email protected]
Internal friction in a new kind of metal matrix composites
Materials Science and Engineering A 442 (2006) 429–432
Internal friction in a new kind of metal matrix composites J. San Juan a,c,∗ , M.L. N´o b,c a
Dpt. Fisica Materia Condensada, Universidad del Pa´ıs Vasco, Facultad de Ciencia y Tecnolog´ıa, Apdo. 644-48080, Bilbao, Spain b Dpt. F´ısica Aplicada II, Universidad del Pa´ıs Vasco, Facultad de Ciencia y Tecnolog´ıa, Apdo. 644-48080, Bilbao, Spain c Instituto de S´ıntesis y Estudio de Materiales, Universidad del Pa´ıs Vasco, Facultad de Ciencia y Tecnolog´ıa, Apdo. 644-48080, Bilbao, Spain Received 6 September 2005; received in revised form 18 April 2006; accepted 18 April 2006
Abstract We have developed a new kind of metal matrix composites, based on powders of Cu–Al–Ni shape memory alloys (SMAs) surrounded by an indium matrix, specifically designed to exhibit high mechanical damping. The damping properties have been characterized by mechanical spectroscopy as a function of temperature between 150 and 400 K, frequency between 3 × 10−3 and 3 Hz, and strain amplitude between 5 × 10−6 and 10−4 . The material exhibits, in some range of temperature, internal friction as high as 0.54. The extremely high damping is discussed in the light of the microstructure of the material, which has been characterized in parallel. © 2006 Elsevier B.V. All rights reserved. Keywords: High damping; Shape memory alloys; Metal matrix composites; Martensitic transformation
1. Introduction High damping materials are useful and attract nowadays scientific and technological interest. High damping materials can reduce acoustic pollution, improve the speed and precision of machining tools, increase the life of aeronautical components and protect buildings against earthquake, for instance. Traditionally, the materials exhibiting high damping are polymeric materials, owing to their viscoelastic behaviour [1]. However, polymers have in general a low elastic modulus and this is a drawback for the design of high damping materials for structural applications. The merit index for the design of structural damping is the product of the elastic modulus (stiffness) E and the damping coefficient tan φ [2], which should be optimized. According to this principle, several kinds of high damping metallic materials, called HIDAMETS, have been developed [3]; metals have higher modulus and strength than polymers. Among various metallic materials, one of the alloy families that exhibit high damping is the shape memory alloys (SMAs) [4]. These alloys undergo a reversible thermo-elastic martensitic transformation from the high temperature phase, called austenite, to the low temperature phase, called martensite, which can be induced by cooling or by applying a mechanical stress. The ∗
Corresponding author. Tel.: +34 94 601 2478; fax: +34 94 601 3500. E-mail address: [email protected] (J.S. Juan).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.04.146
interfaces between martensite and austenite are mobile during the transformation, as well as those in the martensite phase, and under mechanical stress move and absorb mechanical energy, which is the origin of the high damping [5]. Copper-based SMAs exhibit, in general, higher damping coefficient than the Ti–Ni SMA commercially used in most applications. Nevertheless, bulk SMAs does not exhibit high enough damping. We consider, therefore, the possibility of producing a new kind of high damping SMA metal matrix composites. In the last decade, we have developed a production technology of copper-based SMA by powder metallurgy [6,7]. This method allows a good control of the transformation temperatures through the composition control [8,9]. In this paper we present some fundamental aspects about this kind of materials, as well as their damping behaviour at various frequencies and strain amplitudes. 2. Materials and techniques While the method of production of this new kind of composites, is presented elsewhere [10], here we describe briefly their microstructure. Powders of Cu–13.1% Al–3.1% Ni (wt.%) were produced by gas atomisation with argon [6]. In this particular work, powders with sizes between 25 and 50 m were selected and degassed under a vacuum, inside a tubular mould, at 130 ◦ C and 1 Pa for 6 h. Then, at 190 ◦ C a metal matrix of indium was
430
J.S. Juan, M.L. N´o / Materials Science and Engineering A 442 (2006) 429–432
Fig. 1. Optical micrograph of the metal matrix composite obtained by Nomarsky interferential contrast. The indium matrix surrounds Cu–Al–Ni powder particles, showing some surface relief from the martensite variants.
posite has been obtained from the internal friction spectra during cooling and heating; the integrated area of the internal friction peak is proportional to the transformed volume fraction [11]. There is only a slight shift towards higher temperature, as well as a slight broadening of the transformation cycle of the composite. The shift in temperature is associated with the general evolution of the transformation cycle in Cu–Al–Ni during ageing after quenching. The original powders were quenched, while the powders of the composite were aged at 190 ◦ C for about 1 h. The damping behaviour of the composite has been studied by mechanical spectroscopy in a forced torsion pendulum, described elsewhere [12], measuring tan φ; φ is the lag angle between the strain and the stress, and tan φ represents the dissipated energy by unit of volume and is called internal friction. Internal friction has been measured as a function of temperature at various frequencies between 3 and 0.01 Hz and at various strain amplitudes. In all cases the spectra as a function of temperature have been measured under linear heating–cooling at a rate of 60 K/h. 3. Results and discussion
infiltrated at a pressure of 3 × 105 Pa. Fig. 1 is an optical micrograph obtained by Nomarsky interferential contrast in a LEICA DMRXA optical microscope. It shows a general view of the microstructure of the composite, in which the indium matrix surrounds the Cu–Al–Ni alloy particles. The particles undergo the martensitic transformation after the production of the composite, as if the composite exhibits such a transformation. This fact is illustrated in Fig. 2 showing the martensitic transformation cycle, measured by differential scanning calorimetry (DSC) in the original atomised powders, in comparison with the cycle measured in the composite. The transformation cycle of the com-
Fig. 2. Transformed volume fraction during martensitic transformation of the original Cu–Al–Ni powders, measured by DSC (rhombi), in comparison with those of the composite prepared from the same powders, evaluated by integrating the internal friction spectrum (circles).
In Fig. 3, we present the internal friction spectra and the associated modulus variation of the composite during cooling and heating. We show the results for three frequencies, 3 Hz (Fig. 3a), 0.3 Hz (Fig. 3b) and 0.03 Hz (Fig. 3c), all with a strain amplitude of ε = 10−5 . In all the cases an internal friction peak appears during cooling, which is associated with the forward martensitic transformation, and another during heating, which is associated with the reverse martensitic transformation. Both peaks are accompanied by a fall in the modulus, which is linked to the softening of the elastic constants, undergone by the Cu–Al–Ni particles during the martensitic transformation. An increasing background is observed, which is due to the metallic matrix of indium, whose melting point is at 429.6 K. Attention has to be paid to the different levels of internal friction at the three frequencies: internal friction increases with decreasing frequency, as predicted theoretically [11,13]. In Fig. 4 the internal friction spectrum during the forward transformation of the composite is compared with that of the bulk Cu–Al–Ni alloy prepared by powder metallurgy after compaction by hot isostatic pressing and hot rolling [7]. Both spectra have been measured in the same experimental conditions: at a frequency of 1 Hz, a strain amplitude of ε = 10−5 and a cooling rate of 60 K/h. The bulk material exhibits the transformation at higher temperature, owing to the internal stresses created during the powder metallurgy processing. Second, the composite exhibits a broader transformation peak and higher internal friction background. The broadening of the peak in the composite can be understood if we consider that, during the transformation of the Cu–Al–Ni particles, some surface relief should appear as a consequence of the transformation shearing. However, the indium matrix surrounding the particles will oppose to the relief modifications and will perform some stress against the shearing, which is responsible for the delay in the transformation of the particles. Besides, when the shearing relief due to the martensite variants finally appears, it will produce some local plastic deformation on the
J.S. Juan, M.L. N´o / Materials Science and Engineering A 442 (2006) 429–432
431
Fig. 4. Internal friction measured with the same experimental conditions in the bulk material obtained from Cu–Al–Ni powders (rhombi) and in the composite prepared from the same powders (circles).
soft matrix of indium, which will generate another secondary contribution to the damping due to the mobility of the dislocations. A similar situation should take place at low temperatures, in the martensitic phase, when the applied external stress induce the motion of the martensite interfaces to re-orient the variants with respect to stress. Indeed, in such a case the surface relief should also be modified, and the matrix will be locally deformed under the oscillating stress. This local strain will be delayed from the applied stress, being responsible for a local contribution to
Fig. 3. Internal friction and the associated modulus of the composite measured during cooling and heating at three frequencies: 3 Hz (a), 0.3 Hz (b) and 0.03 Hz (c). Fig. 5. Internal friction spectra of the composite during the forward transformation (cooling) at three frequencies: 0.1 Hz (squares), 0.03 Hz (triangles) and 0.01 Hz (circles), and those of a bulk compacted material (rhombi).
432
J.S. Juan, M.L. N´o / Materials Science and Engineering A 442 (2006) 429–432
damping owing to the moving dislocations around the particles. This behaviour is very well illustrated by Fig. 5, which shows the internal friction spectra of the composite at three frequencies 0.1 Hz (squares), 0.03 Hz (rhombi) and 0.01 Hz (circles), in comparison with the spectrum of the bulk material prepared from the same powders. The damping is enhanced in all the temperature range, and especially at the peak, reaching tan φ = 0.54, which is exceptionally high as the damping of a metallic material. The low temperature background is also increased in the temperature range in which the particles are transformed into martensite phase. This increase is stronger than normally observed for indium [14]. 4. Summary and conclusions We have developed a new kind of metal matrix composites based on powders of shape memory alloys, specifically designed for high damping. Possible origins of the extremely high damping peak and the high level of the low temperature internal friction background are discussed. The coupling of motion of the martensite interfaces, of the Cu–Al–Ni SMA particles, with the soft surrounding matrix produces an enhancement of the local damping, generating an amplifying effect of the damping of the composite. We conclude that the newly developed metal matrix composites produced by copper-based SMA powders surrounded by a soft matrix introduce a new concept for designing high damping metallic materials. The extremely high damping of these new composites demonstrate that they are very promising materials for high damping applications.
Acknowledgements The authors thank the Spanish Ministry of Science and Education (project MAT 2004-03166), and the programme ETORTEK of the Basque Government (project ACTIMAT) for the financial support. References [1] R.S. Lakes, Viscoelastic Solids, CRC Press, Boca Raton, USA, 1999. [2] Y.C. Wang, M. Ludwigson, R.S. Lakes, Mater. Sci. Eng. A 370 (2004) 41–49. [3] R. Schaller, Mater. Sci. Forum 366–368 (2001) 621–631. [4] K. Otsuka, C.M. Wayman (Eds.), Shape Memory Materials, Cambridge University Press, Cambridge, 1998. [5] J. San Juan, M.L. N´o, J. Alloys Compd. 355 (2003) 65–71. [6] J. San Juan, R.B. P´erez-Saez, V. Recarte, M.L. N´o, G. Caruana, M. Lieblich, O. Ruano, J. Phys. IV C8 (1995) 919–924. [7] R.B. P´erez-Saez, V. Recarte, M.L. N´o, O. Ruano, J. San Juan, Adv. Eng. Mater. 2 (2000) 49–53. [8] V. Recarte, R.B. Perez-Saez, E.H. Bocanegra, M.L. N´o, J. San Juan, Metall. Mater. Trans. 33A (2002) 2581–2591. [9] A. Ibarra, P.P. Rodriguez, V. Recarte, J.I. Perez-Landazabal, M.L. N´o, J. San Juan, Mater. Sci. Eng. A 370 (2004) 492–496. [10] J. San Juan, M.L. N´o, Second International Symposium on High Damping Materials, Kyoto, September 2005. [11] R.B. Perez-Saez, V. Recarte, M.L. N´o, J. San Juan, Phys. Rev. B 54 (1998) 5684–5692. [12] I. Gutierrez-Urrutia, M.L. N´o, E. Carre˜no-Morelli, B. Guisolan, R. Scaller, J. San Juan, Mater. Sci. Eng. A 370 (2004) 435–439. [13] J. Van Humbeeck, Mater. Sci. Forum 366–368 (2001) 382–415. [14] L.S. Cook, R.S. Lakes, Scr. Metall. Mater. 32 (1995) 773–777.
Internal friction in a new kind of metal matrix composites J. San Juan a,c,∗ , M.L. N´o b,c a
Dpt. Fisica Materia Condensada, Universidad del Pa´ıs Vasco, Facultad de Ciencia y Tecnolog´ıa, Apdo. 644-48080, Bilbao, Spain b Dpt. F´ısica Aplicada II, Universidad del Pa´ıs Vasco, Facultad de Ciencia y Tecnolog´ıa, Apdo. 644-48080, Bilbao, Spain c Instituto de S´ıntesis y Estudio de Materiales, Universidad del Pa´ıs Vasco, Facultad de Ciencia y Tecnolog´ıa, Apdo. 644-48080, Bilbao, Spain Received 6 September 2005; received in revised form 18 April 2006; accepted 18 April 2006
Abstract We have developed a new kind of metal matrix composites, based on powders of Cu–Al–Ni shape memory alloys (SMAs) surrounded by an indium matrix, specifically designed to exhibit high mechanical damping. The damping properties have been characterized by mechanical spectroscopy as a function of temperature between 150 and 400 K, frequency between 3 × 10−3 and 3 Hz, and strain amplitude between 5 × 10−6 and 10−4 . The material exhibits, in some range of temperature, internal friction as high as 0.54. The extremely high damping is discussed in the light of the microstructure of the material, which has been characterized in parallel. © 2006 Elsevier B.V. All rights reserved. Keywords: High damping; Shape memory alloys; Metal matrix composites; Martensitic transformation
1. Introduction High damping materials are useful and attract nowadays scientific and technological interest. High damping materials can reduce acoustic pollution, improve the speed and precision of machining tools, increase the life of aeronautical components and protect buildings against earthquake, for instance. Traditionally, the materials exhibiting high damping are polymeric materials, owing to their viscoelastic behaviour [1]. However, polymers have in general a low elastic modulus and this is a drawback for the design of high damping materials for structural applications. The merit index for the design of structural damping is the product of the elastic modulus (stiffness) E and the damping coefficient tan φ [2], which should be optimized. According to this principle, several kinds of high damping metallic materials, called HIDAMETS, have been developed [3]; metals have higher modulus and strength than polymers. Among various metallic materials, one of the alloy families that exhibit high damping is the shape memory alloys (SMAs) [4]. These alloys undergo a reversible thermo-elastic martensitic transformation from the high temperature phase, called austenite, to the low temperature phase, called martensite, which can be induced by cooling or by applying a mechanical stress. The ∗
Corresponding author. Tel.: +34 94 601 2478; fax: +34 94 601 3500. E-mail address: [email protected] (J.S. Juan).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.04.146
interfaces between martensite and austenite are mobile during the transformation, as well as those in the martensite phase, and under mechanical stress move and absorb mechanical energy, which is the origin of the high damping [5]. Copper-based SMAs exhibit, in general, higher damping coefficient than the Ti–Ni SMA commercially used in most applications. Nevertheless, bulk SMAs does not exhibit high enough damping. We consider, therefore, the possibility of producing a new kind of high damping SMA metal matrix composites. In the last decade, we have developed a production technology of copper-based SMA by powder metallurgy [6,7]. This method allows a good control of the transformation temperatures through the composition control [8,9]. In this paper we present some fundamental aspects about this kind of materials, as well as their damping behaviour at various frequencies and strain amplitudes. 2. Materials and techniques While the method of production of this new kind of composites, is presented elsewhere [10], here we describe briefly their microstructure. Powders of Cu–13.1% Al–3.1% Ni (wt.%) were produced by gas atomisation with argon [6]. In this particular work, powders with sizes between 25 and 50 m were selected and degassed under a vacuum, inside a tubular mould, at 130 ◦ C and 1 Pa for 6 h. Then, at 190 ◦ C a metal matrix of indium was
430
J.S. Juan, M.L. N´o / Materials Science and Engineering A 442 (2006) 429–432
Fig. 1. Optical micrograph of the metal matrix composite obtained by Nomarsky interferential contrast. The indium matrix surrounds Cu–Al–Ni powder particles, showing some surface relief from the martensite variants.
posite has been obtained from the internal friction spectra during cooling and heating; the integrated area of the internal friction peak is proportional to the transformed volume fraction [11]. There is only a slight shift towards higher temperature, as well as a slight broadening of the transformation cycle of the composite. The shift in temperature is associated with the general evolution of the transformation cycle in Cu–Al–Ni during ageing after quenching. The original powders were quenched, while the powders of the composite were aged at 190 ◦ C for about 1 h. The damping behaviour of the composite has been studied by mechanical spectroscopy in a forced torsion pendulum, described elsewhere [12], measuring tan φ; φ is the lag angle between the strain and the stress, and tan φ represents the dissipated energy by unit of volume and is called internal friction. Internal friction has been measured as a function of temperature at various frequencies between 3 and 0.01 Hz and at various strain amplitudes. In all cases the spectra as a function of temperature have been measured under linear heating–cooling at a rate of 60 K/h. 3. Results and discussion
infiltrated at a pressure of 3 × 105 Pa. Fig. 1 is an optical micrograph obtained by Nomarsky interferential contrast in a LEICA DMRXA optical microscope. It shows a general view of the microstructure of the composite, in which the indium matrix surrounds the Cu–Al–Ni alloy particles. The particles undergo the martensitic transformation after the production of the composite, as if the composite exhibits such a transformation. This fact is illustrated in Fig. 2 showing the martensitic transformation cycle, measured by differential scanning calorimetry (DSC) in the original atomised powders, in comparison with the cycle measured in the composite. The transformation cycle of the com-
Fig. 2. Transformed volume fraction during martensitic transformation of the original Cu–Al–Ni powders, measured by DSC (rhombi), in comparison with those of the composite prepared from the same powders, evaluated by integrating the internal friction spectrum (circles).
In Fig. 3, we present the internal friction spectra and the associated modulus variation of the composite during cooling and heating. We show the results for three frequencies, 3 Hz (Fig. 3a), 0.3 Hz (Fig. 3b) and 0.03 Hz (Fig. 3c), all with a strain amplitude of ε = 10−5 . In all the cases an internal friction peak appears during cooling, which is associated with the forward martensitic transformation, and another during heating, which is associated with the reverse martensitic transformation. Both peaks are accompanied by a fall in the modulus, which is linked to the softening of the elastic constants, undergone by the Cu–Al–Ni particles during the martensitic transformation. An increasing background is observed, which is due to the metallic matrix of indium, whose melting point is at 429.6 K. Attention has to be paid to the different levels of internal friction at the three frequencies: internal friction increases with decreasing frequency, as predicted theoretically [11,13]. In Fig. 4 the internal friction spectrum during the forward transformation of the composite is compared with that of the bulk Cu–Al–Ni alloy prepared by powder metallurgy after compaction by hot isostatic pressing and hot rolling [7]. Both spectra have been measured in the same experimental conditions: at a frequency of 1 Hz, a strain amplitude of ε = 10−5 and a cooling rate of 60 K/h. The bulk material exhibits the transformation at higher temperature, owing to the internal stresses created during the powder metallurgy processing. Second, the composite exhibits a broader transformation peak and higher internal friction background. The broadening of the peak in the composite can be understood if we consider that, during the transformation of the Cu–Al–Ni particles, some surface relief should appear as a consequence of the transformation shearing. However, the indium matrix surrounding the particles will oppose to the relief modifications and will perform some stress against the shearing, which is responsible for the delay in the transformation of the particles. Besides, when the shearing relief due to the martensite variants finally appears, it will produce some local plastic deformation on the
J.S. Juan, M.L. N´o / Materials Science and Engineering A 442 (2006) 429–432
431
Fig. 4. Internal friction measured with the same experimental conditions in the bulk material obtained from Cu–Al–Ni powders (rhombi) and in the composite prepared from the same powders (circles).
soft matrix of indium, which will generate another secondary contribution to the damping due to the mobility of the dislocations. A similar situation should take place at low temperatures, in the martensitic phase, when the applied external stress induce the motion of the martensite interfaces to re-orient the variants with respect to stress. Indeed, in such a case the surface relief should also be modified, and the matrix will be locally deformed under the oscillating stress. This local strain will be delayed from the applied stress, being responsible for a local contribution to
Fig. 3. Internal friction and the associated modulus of the composite measured during cooling and heating at three frequencies: 3 Hz (a), 0.3 Hz (b) and 0.03 Hz (c). Fig. 5. Internal friction spectra of the composite during the forward transformation (cooling) at three frequencies: 0.1 Hz (squares), 0.03 Hz (triangles) and 0.01 Hz (circles), and those of a bulk compacted material (rhombi).
432
J.S. Juan, M.L. N´o / Materials Science and Engineering A 442 (2006) 429–432
damping owing to the moving dislocations around the particles. This behaviour is very well illustrated by Fig. 5, which shows the internal friction spectra of the composite at three frequencies 0.1 Hz (squares), 0.03 Hz (rhombi) and 0.01 Hz (circles), in comparison with the spectrum of the bulk material prepared from the same powders. The damping is enhanced in all the temperature range, and especially at the peak, reaching tan φ = 0.54, which is exceptionally high as the damping of a metallic material. The low temperature background is also increased in the temperature range in which the particles are transformed into martensite phase. This increase is stronger than normally observed for indium [14]. 4. Summary and conclusions We have developed a new kind of metal matrix composites based on powders of shape memory alloys, specifically designed for high damping. Possible origins of the extremely high damping peak and the high level of the low temperature internal friction background are discussed. The coupling of motion of the martensite interfaces, of the Cu–Al–Ni SMA particles, with the soft surrounding matrix produces an enhancement of the local damping, generating an amplifying effect of the damping of the composite. We conclude that the newly developed metal matrix composites produced by copper-based SMA powders surrounded by a soft matrix introduce a new concept for designing high damping metallic materials. The extremely high damping of these new composites demonstrate that they are very promising materials for high damping applications.
Acknowledgements The authors thank the Spanish Ministry of Science and Education (project MAT 2004-03166), and the programme ETORTEK of the Basque Government (project ACTIMAT) for the financial support. References [1] R.S. Lakes, Viscoelastic Solids, CRC Press, Boca Raton, USA, 1999. [2] Y.C. Wang, M. Ludwigson, R.S. Lakes, Mater. Sci. Eng. A 370 (2004) 41–49. [3] R. Schaller, Mater. Sci. Forum 366–368 (2001) 621–631. [4] K. Otsuka, C.M. Wayman (Eds.), Shape Memory Materials, Cambridge University Press, Cambridge, 1998. [5] J. San Juan, M.L. N´o, J. Alloys Compd. 355 (2003) 65–71. [6] J. San Juan, R.B. P´erez-Saez, V. Recarte, M.L. N´o, G. Caruana, M. Lieblich, O. Ruano, J. Phys. IV C8 (1995) 919–924. [7] R.B. P´erez-Saez, V. Recarte, M.L. N´o, O. Ruano, J. San Juan, Adv. Eng. Mater. 2 (2000) 49–53. [8] V. Recarte, R.B. Perez-Saez, E.H. Bocanegra, M.L. N´o, J. San Juan, Metall. Mater. Trans. 33A (2002) 2581–2591. [9] A. Ibarra, P.P. Rodriguez, V. Recarte, J.I. Perez-Landazabal, M.L. N´o, J. San Juan, Mater. Sci. Eng. A 370 (2004) 492–496. [10] J. San Juan, M.L. N´o, Second International Symposium on High Damping Materials, Kyoto, September 2005. [11] R.B. Perez-Saez, V. Recarte, M.L. N´o, J. San Juan, Phys. Rev. B 54 (1998) 5684–5692. [12] I. Gutierrez-Urrutia, M.L. N´o, E. Carre˜no-Morelli, B. Guisolan, R. Scaller, J. San Juan, Mater. Sci. Eng. A 370 (2004) 435–439. [13] J. Van Humbeeck, Mater. Sci. Forum 366–368 (2001) 382–415. [14] L.S. Cook, R.S. Lakes, Scr. Metall. Mater. 32 (1995) 773–777.