Available online at www.sciencedirect.com
Scripta Materialia 61 (2009) 395–398 www.elsevier.com/locate/scriptamat
Spark plasma extrusion (SPE): Prospects and potential K. Morsi,* A. El-Desouky, B. Johnson, A. Mar and S. Lanka Department of Mechanical Engineering, San Diego State University, 5500 Campanile Drive, San Diego, CA 92182, USA Received 25 March 2009; revised 14 April 2009; accepted 14 April 2009 Available online 22 April 2009
Despite the advantages of spark plasma sintering, it has so far been limited to the processing of simple shapes, due to its inherent geometric configuration. This paper discusses the prospects and potential of spark plasma ‘‘extrusion” as a process that can allow the production of extended geometries via electric-current processing. Results on the spark plasma extrusion and properties of aluminum are also discussed, showing the feasibility of this processing approach, which has major implications for the spark plasma sintering field. Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Powder processing; Powder consolidation; Extrusion; Hot working; Aluminum
Spark plasma sintering (SPS) of metallic powder involves the passage of pulsed electric current through powder while subjected to an applied pressure. Advantages of the process include sintering at significantly lower sintering temperatures, shorter sintering times, higher heating rates and much faster phase transformation kinetics than conventionally possible [1]. The process has been applied to a wide range of materials including nanopowders [2], biomaterials [3], carbon nanotube ceramic composites [4], titanium dual matrix composites [5], shape memory materials [6], intermetallics [7] and transparent ceramics [8]. It is clear that SPS will remain a very important process for some time to come. However, most research so far has been focused on the processing of simple shapes. Due to the recent intense research and the now documented remarkable advantages of SPS [9], its extension to spark plasma extrusion (SPE) has major implications, including the production of powder-based materials of extended geometries. Other implications include the potential generation of new unique microstructures due to the effect of stress-induced deformation under the influence of electric current. In addition, material recrystallization under the influence of current during SPE can lead to grain refinement, which would not normally occur in SPS. It is noteworthy that high electric current density has been shown to decrease the flow stress of metals, and work by Conrad and co-workers have focused
* Corresponding author. E-mail:
[email protected]
on examining this effect in more detail, in addition to establishing a fundamental understanding of mechanisms involved [10]. They reported that above critical current densities of 1000–10,000 A cm 2 strain rates were observed to increase by orders of magnitude. It is therefore fair to expect in SPE a reduction in extrusion temperature requirements as well as extrusion pressure at some critical current density. Apart from limited work on the electric rolling of powders [11,12], a literature search by the authors did not reveal any publications on the SPE or electric extrusion of powder-based materials. A patent search, though, did reveal that a continuous extrusion apparatus that uses electric current as a means to uniformly heat and simultaneously sinter and ‘‘continuously” extrude electrically conductive granulated materials (using ceramic extrusion containers/tooling) had in fact been patented in 1983 [13] (hence the origination of the idea must be placed there), which was before the thrust of the research on SPS [14]. However, to date the SPE of powder-based materials remains a largely virgin area from the research point of view, with many questions needing still to be answered and disseminated to the scientific community. These include: What is the effect of combined deformation and electric current activation on the recovery and recrystallization behavior of extruding materials? What is the effect of current density, extrusion speed and extrusion ratio on the activation energy for extrusion, deformation mechanisms, extrusion pressure requirements and resulting microstructures?
1359-6462/$ - see front matter Ó 2009 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.scriptamat.2009.04.026
396
K. Morsi et al. / Scripta Materialia 61 (2009) 395–398
Given that SPE will have different heat generation and dissipation sources (e.g. Joule heating, frictional heating, deformation heating, heat losses to the tooling and decreased Joule heating during the process due to loss of resistance through consolidation), what would be the temperature distribution within a spark plasma extruded/extruding sample? And how does it affect the microstructure? The processing-microstructure-properties relations also need to be fully understood for this process and material systems which include, amongst others, nanopowders, composites, nanocomposites, reacting powder systems, metallic glasses, intermetallics and powderbased biomaterials. In this short communication, we present for the first time the consolidation of aluminum powder via SPE, in which powder compacts are heated via direct electric current (DC) to the extrusion temperature and simultaneously extruded under the influence of electric current. The results prove the feasibility of the SPE process and literally open the door for the SPE processing of numerous other powder-based material systems. Commercially pure aluminum powder (Atlantic Equipment Engineers, USA) were first ball milled for 3 min under an argon atmosphere in a SPEX 8000 mixer using WC-Co balls with a ball-to-powder weight ratio of 5:1. The ball milling serves the purpose of breaking down oxide layers that have been reported to exist at the surface of the aluminum powder [15], which in our case prevented the conduction of electric current through the powder billet. The ball-milled aluminum powder was pressed in a hardened steel compaction die to produce billets of dimensions 19 mm diameter and 25 mm height with a green density of 80% of theoretical. A home-built SPE rig (Fig. 1) was used for the spark plasma extrusion of the powder billets at three extrusion onset temperatures (250, 350 and 450 °C). The electric currents applied were 400 A (current density = 129 A cm 2), 500 A (current density = 160 A cm 2) and 700 A (current density = 225 A cm 2) respectively at less than 5 V. The SPE rig was constructed in such a way as to allow conventional hot extrusion, SPE or a combination of both, making it useful as a research tool. The container, die holder and extrusion dies were all made from H13 tool steel, to allow operation under high stress and elevated temperatures. An Inconel 718 inner liner was also shrunk fitted in the H13 steel container, giving an inner liner diameter of 20 mm. The bolster was designed
with an observation window and gas inlets to allow extrusion in different environments. The rig was placed on a universal testing machine, which was used to apply the extrusion load and control the extrusion speed. The extrusion ram in SPE serves two purposes: first, to transmit potentially high load at possibly elevated temperatures, and second, to transmit electricity. WC-Co was used as the ram material as it possesses high compressive strength at elevated temperatures and relatively low electrical resistivity, as opposed to the regular graphite rams used in SPS. The ram speed was set to 1.5 mm s 1 for all SPE experiments in this study. Temperature was monitored using a K-type thermocouple embedded in the bottom of the specimen through the die opening, which would be simply pushed out on extrusion and can be re-used. All specimens were heated through the application of direct current using a power supply (0–5 V, 0–1000 A). Only 50% of the height of each billet was extruded (using an extrusion ratio of 16:1). This way, each specimen would experience two different types of electric-current processing (as will be explained later). Another advantage of SPE is that heating is localized in the specimen by using a flexible insulator ( 0.05 mm thick) between the specimen/ram arrangement and the container inner liner, which is a highly energy efficient approach to heating (an area that has not been given significant attention in the SPS literature). For microstructural observations, the specimens were cut along the central axis of the longitudinal (extrusion direction), ground and polished to a 1 lm finish. A scanning electron microscope was used to characterize the surface of the discard and the extrudate. Etching was conducted using an etchant of composition 2% NaOH + 4% Na2CO3, with the balance being H2O, for 60 s. Hardness was measured with a Vickers microhardness tester using a 500 g load. All specimens were heated to their respective extrusion temperatures via direct electric current, after which extrusion was conducted. Figure 2 shows the pressure–ram displacement curves for specimens spark plasma extruded at three different extrusion onset temperatures. All the curves show that there is an initial upsetting stage, during which the powder compacts conform to the shape of the internal diameter of the extrusion inner liner, followed by a steep rise in pressure during powder consolidation, then extrusion through the die exit. These curves represent the temperatures for the ‘‘onset” of the SPE process (i.e. the start of the ram displacement). Immediately after SPE onset, the temperature of the specimen was always observed to significantly decrease
Figure 1. (a) Schematic and (b) photograph of San Diego State University’s SPE rig.
K. Morsi et al. / Scripta Materialia 61 (2009) 395–398
Figure 2. Pressure–ram displacement curves for specimens extruded at 250, 350 and 450 °C, and a photograph of the extrudate (350 °C onset temperature) with a centimeter scale.
with an increase in the displacement of the extrusion ram. In conventional hot extrusion, it is well known that heat is generated due to plastic deformation, internal shear and friction between the deforming material and the tooling (which is normally at elevated temperatures); heat is also transferred within the billet and between the billet and the tooling [16]. Hence competing processes exist that would tend to either continuously increase or decrease the extrudate exit temperature with ram displacement during extrusion. For example, deformation and frictional heating favor a continuous increase in the material exit temperature with ram displacement. In the case of SPE, the lower temperature of our tooling (room temperature) compared to the hot billets (250, 350 and 450 °C) will favor significant heat losses from the material to the tooling and therefore a decrease in exit temperature with ram displacement during extrusion. Moreover, Joule heating is the initial heating source in SPE, which is expected to decrease with densification during extrusion due to a decrease in the electrical resistance of the powder compact. Hence this and the temperature difference between the inner hot billet and the outer colder container contributed to the observed decline in temperature with ram displacement, which consequently led to the increasing extrusion pressure observed during SPE. Figure 2 also shows that the extrusion pressure requirements increase with a decrease in extrusion onset temperature to the point where extrusion was not possible at an extrusion onset temperature of 250 °C due to the preset pressure limit. This increase in extrusion pressure requirements is due to the corresponding decline in material flow stress and possible strain hardening in the absence of recovery. A schematic of an extruded specimen is shown in Figure 3, where typical deformation and dead metal zones are identified. The material above the deformation zone is simply pushed down and pressed under current; hence the material in this location would have experienced SPS (or pressing). On the other hand, the material that passed through the deformation zone and was consolidated and extruded would have experienced SPE. Figure 3 also shows that for two extrusion temperatures (350 and 450 °C) the average microhardness values taken along the central line between 1 and 7 mm from the top surface of the discard (representing regions that had undergone SPS) are lower than the
397
Figure 3. Average Vicker’s microhardness measurements taken within different regions in specimens extruded at 350 and 450 °C (error bars represent standard deviations) and a schematic of an extruded specimen showing typical regions in the discard and extrudate.
values taken along the central line between 1 and 7 mm outside the die exit (representing material that has experienced SPE) due to better consolidation and possibly improved interfacial bonding between the particles. This finding should not be generalized, however, as it depends on a number of complex interactions, including prior powder process history (e.g. degree of cold work), extrusion speed (which in SPE can affect the temperature distribution during extrusion), container temperature, current density, dynamic recrystallization and grain growth in the extrudate at high extrusion temperatures. For both SPE and SPS regions, though, the microhardness is higher for an extrusion onset temperature of 350 °C compared to 450 °C. The microhardness of the 250 °C discard was also found to be the highest, at 93 ± 3.8 HV. Extrusion at 350 and 450 °C resulted in increasingly softer materials due to possible temperature/current-induced recovery or softening processes [17]. Scanning electron micrographs of etched and unetched regions in the extrudates and discards for samples extruded at 350 and 450 °C are shown in Figure 4. A reduction in residual porosity from the discard to the extrudate is evident, as clearly shown for materials spark plasma extruded at 450 °C. It is noted that even the material regions experiencing SPS were subjected to much higher applied pressures than have been previously published; it was therefore not surprising to observe low levels of porosity in the SPS regions. Consequently, it is important to point out that plastic flow is a nearly instantaneous densification mechanism in sintering [18], but has not received significant attention so far as a viable densification mechanism in SPS. Some cell structures are also seen in the micrographs of Figure 4. A detailed analysis will need to be conducted to understand the microstructural evolution in this unique process. A number of conclusions can be drawn from the present work: 1. SPE has been discussed in this paper for the first time for the successful fabrication of Al extrudates. 2. Under the investigated processing conditions, SPE was only possible at extrusion onset temperatures of 350 °C and higher.
398
K. Morsi et al. / Scripta Materialia 61 (2009) 395–398
Figure 4. Electron micrographs taken of etched and unetched regions in the discard and extrudate of specimens extruded at 350 and 450 °C.
3. Materials subjected to SPS conditions displayed lower hardness and higher porosity levels than those spark plasma extruded; however, both displayed a reduction in microhardness with increase in extrusion onset temperatures due to possible dynamic recovery/ softening processes. The present work has major implications for SPS and current-activated processes, and can be applied to a wide range of materials. Future work will examine the use of higher extrusion speeds to overcome the problem of reduced extrusion temperature with ram displacement, in addition to modeling efforts. Research is also underway to examine the possibility of using SPE for the fabrication of nanowires and tubes using ultra-high current densities. The authors wish to thank the National Science Foundation (Award No. 0710869) for their much appreciated support. We also thank Mr. Greg Morris, Mr. Preetam Borah and Dr. Steve Barlow for assistance. [1] J.E. Garay, U. Anselmi-Tamburini, Z.A. Munir, Acta Mater. 51 (2003) 4487–4495. [2] U. Anselmi-Tamburini, J.E. Garay, Z.A. Munir, A. Tacca, F. Maglia, G. Chiodelli, G. Spinolo, J. Mater. Res. 19 (2004) 3225. [3] Y.Gu, K.A. Khor, P. Cheang, Biomaterials 25 (2004) 4127.
[4] C. Balazsi, Z. Shen, Z. Konya, Z. Kasztovszky, F. Weber, Z. Vertsey, L.P. Biro, I. Kiricsi, P. Arato, Compos. Sci. Technol. 65 (2005) 727. [5] V.V. Patel, A. El-Desouky, J.E. Garay, K. Morsi, Mater. Sci. Eng. A 507 (1–2) (2009) 161–166. [6] Sia nemat-Nasser, Yu Su, Wei-Guo Guo, Jon Isaacs, J. Mech. Phys. solids 53 (2005) 2320. [7] V. Gauthier, F. Bernard, E. Gaffet, Z.A. Munir, J.P. Larpin, Intermetallics 9 (2001) 571. [8] R. Chaim, Z. Shen, M. Nygren, J. Mater. Res. 19 (2004) 2527. [9] Z.A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, J. Mater. Sci. 41 (2006) 763–777. [10] Hans Conrad, Mater. Sci. Eng. A 322 (2002) 100– 107. [11] I.M. Mal’tsev, Powder Metall. Met. Ceram. 44 (5–6) (2005) 222–227. [12] C. Mochizuki, M. Mikami, J. Nuclear Mater. 271–272 (1999) 508–511. [13] Klaus Lichtinghagen, US Patent No. 4420294, 1983. [14] Z.A. Munir, U. Anselmi-Tamburini, M. Ohyanagi, J. Mater. Sci. 41 (2006) 763–777. [15] G. Xie, O. Ohashi, M. Song, K. Furuya, T. Noda, Metall. Mater. Trans. 34A (2003) 3. [16] J.R. Davis, Extrusion in ASM Specialty Handbook, Al and Al-Alloys, ASM International, Materials Park, OH, 1992, p. 262. [17] H. Conrad, Mater. Sci. Eng. A287 (2000) 227–237. [18] R. German, Powder Metallurgy and Particulate Materials Processing, Metal Powder Industries Federation, Princeton, NJ, 2005, pp. 293–294.
