Creep Resistance depending on Particle Reinforcement Size of Al-Alloys produced by Powder Metallurgy
Bernd Bauer and Guillermo Requena
Institute of Materials Science and Technology
Vienna University of Technology
Marcela Lieblich
National Centre for Metallurgical Research, CENIM-CSIC
Spanish Research Council, Madrid
Abstract
The Young’s Modulus, the high temperature strength and the tensile creep behaviour are investigated for unreinforced and particle reinforced AA2124 and AA6061 matrices with 25vol% of SiC particles with different sizes. The materials were produced via powder metallurgy and subsequent extrusion. The blending of the metallic powders and the SiC particles was carried out by wet blending (WB) and ball milling (BM). The SiC particles size in WB samples seems less affected by the blending operation, but the bigger particles are affected by the extrusion process. The BM process fractures the SiC particles leading to sub-µm SiC particles and a significant amount of oxide dispersoids is incorporated.
The Young’s Modulus is increased by about 50% compared to the unreinforced alloys due to the introduction of the stiffer SiC particles regardless of the initial particle size. All BM samples and only SiC particles <5µm in WB samples provide increased hot tensile strength. The stationary creep rate of composites produced by BM is two orders of magnitude lower than that of the WB composites and the unreinforced matrices. This enhancement of creep resistance of BM composites is achieved by the presence of dispersoids as a result of BM. The creep resistance of the WB composites decreases the larger the particle size acting as dislocation sources. A similar behaviour is observed for the high temperature strength for which the dispersoids in the BM composites increase the strength, while the larger SiC particles decrease the strength with respect to the matrices.
Introduction
Metal matrix composites (MMCs) are obtained when a continuous metallic matrix is reinforced by means of a secondary stiffer phase, usually a ceramic [[1]]. The reinforcement of metals can have many different objectives such as an increase in specific mechanical properties like creep resistance, tensile strength, wear resistance and Young’s Modulus [1]. Accordingly, the requirements for the reinforcement phases are low density, thermal stability, high compression strength, high tensile strength and high Young’s Modulus [[2]].
Powder metallurgy (PM) techniques are solid phase processes that can be applied to produce MMCs. These typically involve discontinuous reinforcement, due to the ease of mixing and blending. The ceramic and metal powders are mixed, cold compacted and hot pressed. The compacted and pressed material then typically undergoes a secondary operation such as extrusion. Powder processing is used to fabricate primarily particle- or whisker-reinforced MMCs. The matrix and the reinforcement powders are blended to produce a homogeneous distribution [2]. The blending can be carried out dry or in liquid suspension. This is usually followed by cold compaction, canning, degassing and high temperature consolidation. The relative size of metal and ceramic particles has been identified as significant for homogeneous blending of powders to produce particulate reinforced metals (PRM) [[3]].
According to [[4]] many properties of powders– such as powder flow, apparent density, ejection stress, delubrication behaviour as well as dimensional change and mechanical strength of extruded parts – are quite sensitive to even small changes in particle size distribution and to fluctuations in the concentrations of components within a powder mixture. In the present study the dependence of the high temperature strength, the Young’s Modulus as well as the creep resistance of Al-based PRMs produced by PM are investigated for different mixing routes and particle sizes. Regarding the creep resistance controversial results were reported during the last decades. On the one hand, there should be an increase of creep resistance of Al alloys due to the addition of SiC particles [[5], [6]], whereas it was also reported that the SiC particles degrades the creep resistance [[7], [8]].
Experimental Procedures
The Materials
Two Al-based matrices reinforced with 25vol% of SiC particles of different sizes were produced by PM. The matrices in the present study are AA2124 (AlCu4) and AA6061 (AlMg1SiCu) alloys. The chemical compositions in Table 1 were determined by Inductively Coupled Plasma (ICP) - analysis [[9]]. The matrix powder grains have sizes <75µm. Three different SiC particle sizes (denominated <5µm, <10µm and <25µm) were used to reinforce the matrices. The shape of the utilized SiC particles is irregular and with sharp edges. The initial SiC particle sizes can be seen in Table 2 [[10]]. Hereby, the designations F360, F600 and F1000 correspond to the particle sizes <25µm, <10µm and <5µm, respectively. The SiC powders were provided by ESK-GmbH [10]. Three values are presented to characterize the powders: 1) the P97-value is the 97% percentile that means that 97% of the particle number is below this value, 2) the median and 3) the P6-value which indicates that 6% of the particle number is below this value.
The powders of the matrices and the SiC particulates were blended using two different routes: low energy mixing was carried out within cyclohexane for a period of 24 hours and the liquid was subsequently evaporated. Hereafter, this process is called wet blending (WB). During WB the powder particles are neither deformed nor fractured. The blending time of 24 hours was chosen because of the relatively high size ratio between the matrix powder particles and the SiC particles (3-15:1). The smaller the ratio the easier the blending [[11]]. Taking into account the different SiC sizes and the fixed blending time of 24 hours it was found empirically that the optimal amount of blended powder is 100g and 200g in the given process. SiC particles with sizes <10µm were not used for the production of PRM by WB.
High energy ball milling (BM) was the second blending technique. It was applied using a planetary milling device. A period of 4 hours was necessary to obtain a homogeneous blending of powders as determined by microscopic investigations of the blended powders after 2, 4, 6, 8 and 10 hours of BM. The rotating velocity of the containers was 200 rpm. An amount of 60g of powder mixture of matrix and SiC particles was blended in each of the containers. 12 steel balls with a diameter of 14.4mm and 31g weight were placed in each container to blend the alloy and the SiC powders. The powder agglomerates are repeatedly deformed, fractured and cold welded during BM [4, [12]].
Both types of blended powders were cold compacted within Al tubes and subsequently extruded. Non lubricated forward extrusion [[13]] of the blended powders was conducted by means of a die with a diameter of 11.8mm at 450°C. The inert gas fusion method was used to determine the oxygen of the materials. The determination was carried out in the Institute of Chemical Technologies and Analytics [[14]] using a LECO device. The principle of operation is based on fusion of a sample in a high-purity graphite crucible at temperatures up to 3000°C in an inert gas such as helium. The oxygen in the sample reacts with the carbon from the crucible to form carbon monoxide (CO) and/or carbon dioxide (CO2), and these are detected using infrared (IR) detection. The results are given in Figure 1. The BM matrix and PRM contain about 2-3 times more oxygen (i.e. oxides) than the unreinforced mixed powder compacts and the WB PRM, respectively. The WB PRM contains about the same fraction of oxygen as the unreinforced BM matrix. The oxygen content of the BM PRM with the small particles as ingredients is about 40% higher than that of the BM PRM with the bigger ones.
Prior to testing all the materials were heat treated at 495°C for 20 minutes with subsequent water quenching and overaged at 300°C for 1 hour in order to stabilize the precipitates by overaging. This treatment is designated as T4S where S means stabilisation. Table 3 shows all the investigated materials.
Matrix Alloy / Si [wt%] / Fe [wt%] / Cu [wt%] / Mn [wt%] / Mg [wt%] / Cr [wt%] / Ni, Zn, TiPM 2124 / 0.03 / 0.08 / 3.9 / 0.6 / 1.45 / ≤ 0.01 / ≤ 0.01
PM 6061 / 0.45 / 0.15 / 0.27 / ≤ 0.01 / 0.96 / 0.16 / -
Table 1 Chemical composition of the matrices
Table 2 Initial SiC-particle sizes [10]
Figure 1 Oxygen fraction of the unreinforced 2124 without and with BM, WB 2124/SiC<5µm/25p, BM 2124/SiC<5µm/25p, BM 2124/SiC<25µm/25p
Matrix alloy / SiC [vol.%] / SiC size [µm] class / Production route / Designation2124 / 0 / - / Extrusion / 2124 Matrix without BM
2124 / 0 / - / BM + extrusion / BM 2124 Matrix
6061 / 0 / - / BM + extrusion / BM 6061 Matrix
2124 / 25 / <5 / BM + extrusion / BM 2124/SiC<5µm/25p
<10 / BM 2124/SiC<10µm/25p
<25 / BM 2124/SiC<25µm/25p
<5 / WB + extrusion / WB 2124/SiC<5µm/25p
<25 / WB 2124/SiC<25µm/25p
6061 / 25 / <5 / WB + extrusion / WB 6061/SiC<5µm/25p
<25 / WB 6061/SiC<25µm/25p
Table 3 List of processed and investigated materials
Light Optical Microscopy (LOM)
The samples were cut and embedded in order to study the microstructure in the longitudinal and perpendicular directions with respect to extrusion. Images taken by LOM were analysed by means of the software Axio Vision 4.7.1 in order to determine the distribution of the SiC particles and the presence of porosity. Only the amount and distribution of SiC particles for the WB PRMs (>1µm) were analysed by means of LOM because of the limited resolution. The existence of smaller ones can be clearly observed but not analysed in terms of distribution and size.
Field Emission Gun – Scanning Electron Microscopy (FEG-SEM)
FEG-SEM was utilized to analyse the sub-µm SiC particles which could not be quantified by LOM as well as the SiC particles >1µm for the BM PRMs. Images taken by FEG-SEM in CENIM, Madrid, were used to analyse the SiC particles >1µm of the BM material. The applied device was a FEG-SEM with detectors for secondary electrons as well as backscattered electrons achieving a resolution of 1.5 nm (15kV) or 5nm (1kV). Other FEG-SEM images were taken by means of FEI QUANTA 200 FEG-SEM device provided by University Service for Transmission Elektron Microscopy (USTEM) at the Vienna University of Technology [[15]] in order to investigate the SiC particles <1µm. The observed particles were identified by energy dispersive microanalysis. Negligible differences were found for the SiC particles <1µm between the longitudinal and the perpendicular directions. Therefore, only the images taken in longitudinal direction were analysed.
EBSD measurements were carried out at the USTEM for all the materials in order to obtain their initial grain size by means of FEI QUANTA 200 FEG.
Transmission Electron Microscopy (TEM)
The samples of the unreinforced matrices were prepared by electro polishing using a TenuPol-5 provided by Struers. The ion milling of the WB PRM was done by means of a Precision Ion Polishing System (PIPS). The composition of the sub-µm particles was analysed by energy dispersive X-ray spectroscopy (EDX) in a FEI TECNAI G20. TEM investigations as well as the samples preparation were conducted by USTEM at the Vienna University of Technology [15].
Dynamical Mechanical Analysis (DMA)
The Young’s Modulus of all the investigated materials was measured as a function of the temperature by means of DMA in order to evaluate the quality of the produced materials regarding porosity and the processing routes. The investigations were conducted with a TA Instrument DMA 2980 equipment in air atmosphere using a 3-point-bending clamp with a frequency of 1Hz, amplitude of 40µm and in a temperature ranging from room temperature to 300°C. The thermal stability is given by ±0.1°C above 50°C and ±1°C below 50°C. The samples’ geometry was 55 x 4 x 2 mm³ (Length x Width x Thickness). The accuracy of the results is within ±3%.
High temperature (HT) tensile tests
Hot tensile tests were carried out for all the investigated materials by means of a Zwick Z250 universal test machine equipped with a heating chamber heated by electrical resistance. The tensile tests were conducted in air atmosphere at a temperature of 300°C. The temperature was measured using 2 thermocouples, one on the upper end of the sample and another on the lower end of the sample.
The resulting temperature gradient was less than 7°C. The sample geometry is the same as for the isothermal tensile creep tests.
Isothermal tensile creep tests
The tensile creep test rigs were loaded by weights on a lever system and equipped with a digital temperature control system. The rigs allow to carry out simultaneously 10 tensile creep tests at temperatures up to 800°C [[16]]. The creep tests were carried out at 300°C in air.
The temperature of the samples was measured using two thermocouples one at the bottom and another at the top of the sample. The observed temperature gradients were smaller than 5°C. The strain during the tests was measured using linear variable displacement transducers (LVDT) provided by Micro-Epsilon [[17]] with a sensitivity of 56mV/V and a resolution of 0.01µm. The measured signals were then amplified and recorded using a data acquisition system made up of a PC and external amplifiers model Spider8 provided by Hottinger Baldwin Meßtechnik G.m.b.H. [[18]]. The materials were tested under constant load conditions to determine the stationary creep rate. Once the stationary stage was reached the load was increased in order to continue the test to the corresponding stationary stage. The applied loads ranged from 15MPa to 70MPa resulting in test periods between 0.5h and 8000h.