Study of machining induced surface defects and its effect on fatigue performance of AZ91/15%SiCp metal matrix composite

Nishita Anandan, M.Ramulu

Department of Mechanical Engineering, University of Washington, Seattle, WA 98195, USA

Abstract The quality of surface generated in a peripheral milling of AZ91/SiCp/15% for varying machining conditions and its effect on the fatigue performance are investigated in this study.The machined surface quality was evaluated through roughness measurements and SEM micrographs of the machined surface.Tensile tests were performed to measure the mechanical properties of the composite.Subsequently,fatigue life of milled specimens was measured through axial fatigue tests at four loading conditions.Optical and SEM/EDS micrographs of the fractured surface were studied to identify the crack initiation site and propagation mechanism.Specimens machined at a lower feed rate of 0.1mm/rev was found to have excellent surface finis and consequently higher fatigue life.At 0.3mm/rev, the presence of feed marks and other surface defects resulted in a drastic decrease in fatigue life.Five distinct regions were identifie on the fractured surface, particle fracture along and perpendicular to the surface, voids in the matrix due to particle debonding and pull out and typical ductile failure of matrix with embedded SiC particles.

Keywords: Magnesium composite; Machined surface; Surface integrity; Fatigue.

1.Introduction

Metal matrix composites(MMCs) consists of light weight metallic matrix incorporating various reinforcing phases.Reinforcing aluminum and magnesium alloys with ceramics such as SiC,B4C and Al2O3resulted in MMCs with increased specifi strength and stiffness,improved creep resistance,thermal shock resistance, corrosion resistance and reduction in thermal elongation [1,2].However, owing to the high hardness of these reinforcements surface damage is induced while machining.When the cutting edge interacts with the reinforcement particles, they may be pulled out or pushed into the matrix forming a void on the surface, are dragged along the surface producing ridges, or the particles may fracture and matrix cracking occurs owing to the limited ductility [3].The damages on the machined surface act as stress concentrators and have a detrimental effect on fatigue life.Therefore, in order to avail the outstanding potential of metal matrix composites for structural components, it is imperative to study their behavior under cyclic loading.The fatigue performance of aluminum composites is reported by several studies [4,5],while only very few studies were performed on magnesium composites [6,7].Magnesium composites are very promising in aerospace and biomedical industries owing to the matrix being the lightest metal with a density of 1.74g/cm3and having highly tailorable mechanical properties through the addition of ceramics [8,9].In addition, magnesium alloys possess high dimensional stability, excellent damping capacity and recyclability [10].

Fatigue behavior of MMCs are quite complex due its dependence on the manufacturing processes used, reinforcement particles, metallic matrix and the interface between the particles and the matrix [11].The composites are usually produced through squeeze casting process,spray co-deposition of matrix and particles and powder metallurgy techniques, that are conducted at high temperatures and mismatch in coeffi cient of thermal expansion between the matrix and the reinforcements induces thermal residual strain in the composite as it cools down to the room temperature.The large residual strain results in an increase in dislocation density in the matrix[12-14].As a result, cyclic softening behavior was exhibited by composites similar to the cold worked metal with high dislocation density [15].

Fig.1.Schematic of steps involved in specimen fabrication.All dimensions are in mm.

In addition to the dislocation densities,the pores and inclusions introduced in the composites during the manufacturing process also have a significan effect [16].Cast composites were observed to have lower fatigue life due to the presence of porosity and inclusion in addition to the microstructural defects like particle clusters and non- uniform particle distribution in comparison to forged and extruded composites.Secondary manufacturing processes like forging and rolling help in breaking up the agglomeration, homogenizing the distribution, eliminating the porosity and improving the interfacial bonding [5].

Only limited studies have reported the effect of machining induced damage on fatigue life of composites [17,18].The machining induced surface damage were studied as a function of particle size and machining conditions.The damage in the form of particle fracture and pull out were reported to increase with the particle size 18].Consequently,the fatigue life improved with decreasing particle size.The fatigue life was found to depend significantl on the surface and sub-surface quality generated, with coarser surface finis exhibiting a significantl lower life.In addition, the deformation due to machining resulted in surface softening, reducing the fatigue life[17].Therefore, it is imperative to study the surface and sub surface damage induced while machining on the fatigue life to produce reliable components.

The purpose of this study is to experimentally investigate the machined surface topography and the effect of milling induced surface defects on the fatigue behavior of extruded AZ91/SiCp/15%.The surface topology was studied in terms of surface roughness parameters and inspection of machined surface through SEM micrographs.SEM fractography was performed to identify the crack initiation and fracture mechanism of the matrix and the SiC reinforcement particles.

Table 1Cutting tool and machining conditions.

Fig.2.SEM micrograph of AZ91-T5/SiCp/15%.

Fig.3.Stress strain curve for AZ91/SiCp/15%.

2.Experimental set up and procedure

2.1.Material

The heat treated magnesium composite rods of 55mm diameter were supplied by Pacifi Northwest Laboratory.The material used in this investigation was vacuum cast into 203mm diameter billets and then extruded into rods of fina diameter 55mm and then subjected to T5 heat treatment.The composite consists of AZ91-T5 magnesium matrix reinforced with 15vol% SiC particles.

2.2.Specimen geometry and fabrication

The sequence of steps involved in the fabrication of the specimen are shown in Fig.1.The extruded composite was sectioned into 5mm thick disks using waterjet.The disk was then face milled to reduce the thickness to 3mm, shown in Fig.1b, Fig.1c and d show the rectangular specimen with 9.04mm wide gauge section machined using waterjet.Peripheral milling was used to reduce the width of the gauge section to 6mm by removing 1.52mm on either side.The machining conditions used for face milling and peripheral milling operations are given in Table 1.The geometry of the rectangular miniature specimen with dimensions is shown in Fig.1e.The radius of curvature in the specimen resulted in a stress concentration factor of 1.4 [19].

The face milled surface and the peripherally milled surfaces are referred to as S1 and S2 respectively.PCD with small axial depth of cut was used for the face milling process to ensure minimum surface damage on S1.Owing to the limitation of the composite rod dimension, miniature specimens were developed and tested in this study.A total of 35 specimen were fabricated in this experimental study using the machining conditions presented in Table 1.All the dimensions of the machined specimen after face milling of width and peripheral milling of specimen edges and their surface roughness were examined and quantified Peripherally milled and face milled surfaces were analyzed through the measurement of 2D surface roughness profile using Mahr surface profilomete.The surface roughness was quantifie in terms of average surface roughness,Ra, root mean square roughness,Rq, peak to valley height roughness,Ry, and ten point average roughness,Rzfor a profil height distribution z over a profil length L are define by equations 1 through 4 as:

Fig.5.SEM micrograph of surfaces peripherally milled edge (thickness)surface and face milled fla surface (width).

In addition, stereo microscope and FEI XL30 SEM/EDS were used to obtain the optical and SEM micrographs of the fractured surfaces crack initiation and propagation.

2.3.Tensile and fatigue testing

Tensile properties were measured using the INSTRON 5585H electro-mechanical test frame.The tests were performed on compact specimens machined using abrasive waterjet to achieve the geometry.Crosshead speed was fi ed at 1.27mm/min.Tensile test was performed to observe the behavior of the composite under monotonic loading.A minimum of three tests were conducted to fin the mechanical properties.

The axial fatigue tests were performed using INSTRON 8511 20kN servo-hydraulic test frame at 10Hz.Constant stress amplitude tension-tension tests were performed with stress ratio,R=0.1 at four stress levels,0.8Sut,0.6Sut,0.4Sutand 0.25Sut, whereSutis the ultimate tensile strength of the material to obtain the number of cycles to failure.The cycles to failure were measured and recorded through LABVIEW interface.The fatigue tests were run up to fi e million cycles.If the fracture had not occurred by then,the test was stopped, and infinit life was assumed for the specimen.

Table 2Average surface Roughness of face milled surface and peripherally milled surfaces.

Fig.6.SEM micrographs of the machined surface showing voids(A), matrix cracking(B), particle agglomeration(C).

Fig.7.Maximum stress ratio vs Number of cycles to failure for machined magnesium composites, Sut=360 MPa.

Further, the optical and SEM micrographs were also utilized to examine the fractured surface to identify mechanisms of fracture.EDS was employed to distinguish the matrix and particle fracture on the surface.

3.Results and discussion

The size and distribution of SiC particles in the magnesium matrix were evaluated through analysis of optical and scanning electron micrographs obtained after polishing the as received composite up to 1μm.The average diameter of SiC particles was found to be 3μm with a standard deviation of 1μm and a volume fraction of 15% SiC particles.Typical microstructure of the composite is shown in Fig.2.

3.1.Mechanical properties

A typical stress strain curve for the composite is shown in Fig.3.The material exhibited an average ultimate tensile strength (UTS) of 360MPa, 0.2% yield strength of 330MPa and Young’s modulus of 80GPa.The composite did not exhibit any necking before fracture, leading to limited ductility with average fina strain of 0.0073.

Fig.8.SEM fractography of tensile specimen a) at low magnificatio of 40x, b) at 1000x, c) at 4000x, d) and e) at 16000x.A: SiC fractured along the plane with cleavage lines, B: Particle fracture perpendicular to the surface, C: dimples formed by the matrix.

Fig.9.Optical micrographs of typical fractured surfaces, a) 1000 RPM,0.1mm/rev, 0.6Sut, b) 1000 RPM, 0.1mm/rev, 0.8Sut, red arrows point the crack initiation site.S1:Face milled and S2:Peripherally milled surface.(For interpretation of the references to color in this figur legend, the reader is referred to the web version of this article.)

The limited ductility of the magnesium composite can be attributed to further reduction in ductility of the magnesium alloy due to the presence of SiC that hinder the motion of slip planes.Similar behavior was reported by Luo in studying the mechanical behavior of AZ91 matrix composites.The authors attributed the differences in ductility between the pure matrix and composite to the higher strain hardening rate observed in the composite [20].Additionally, the fracture mechanism was identifie by studying the SEM micrographs of the fractures surface, discussed in Section 3.4.

The UTS obtained from the tensile test was used to determine the stresses used in fatigue testing.Owing to the high sensitivity of fatigue performance on the surface topology,quality of the machined surfaces was investigated through surface profil measurements and verifie through SEM micrographs.Further,the machining induced defects were investigated through SEM micrographs, discussed in the following section.

3.2.Surface topography of machined surface

The surface quality of the machined surface was studied through two-dimensional surface roughness profil measurements along the gauge section through contact profilometr .Machining induced damages on the surface are detrimental to the life of a component.The roughness profile of the machined surface show the deviations from the nominal surface that can act as stress concentration sites.Typical roughness profile of the peripheral and face milled surfaces are shown in Fig.4.Good surface finis was obtained for all the machining conditions, observed in Fig.4 and through SEM micrographs in Fig.5.Surface profile of all the specimens were obtained and the average surface roughness parameters with standard deviations for face milled and peripheral milled surfaces are given in Table 2.

Fig.10.SEM micrographs of crack initiation site, arrow indicates the crack origin.

The SEM micrographs of specimen machined at 0.1mm/rev did not show any feed marks on the surface at 67x, observed in Fig.5a and b.The SiC particles were found to be scattered on the surface.At 0.3mm/rev, prominent feed marks were observed on the surface.Although,extensive scattering of SiC particles were not observed at this condition,the feed marks contribute to the increased surface roughness,shown in Fig.5c and d.The feed marks were also clearly visible on the face milled surface, shown in Fig.5e.

Upon further magnification surface defects such as voids due to particle pull out, matrix cracking and particle agglomeration were observed on the surface, shown in Fig.6.The matrix cracking can be attributed to the lower ductility of the matrix, and due to hard reinforcement particles dragged along the surface.These defects on the surface act as stress concentrators resulting in early onset of crack initiation and propagation.

3.3.Fatigue performance

Fatigue performance of the machined coupons were evaluated by conducting tension-tension fatigue tests to develop the S-N curve.The test was stopped at fi e million cycles,if the specimen had not failed by then, it was considered to have infinit life.The maximum stress ratio and number of cycles to failure for all the machining conditions is plotted in Fig.7.

The fatigue life of the machined specimen machined at 0.1mm/rev was found to be similar irrespective of the spindle speed.This can be attributed to the smoother surface fin ish with average roughness of 0.3μm at this feed rate.At 0.3mm/rev, the roughness increased to about 1μm.This increase in surface roughness was reflecte in the reduction of fatigue life of specimen machined at this feed rate.Although,at 0.3mm/rev, the surface roughness parameters are about the same for the both the spindle speeds,the specimens machined at 1000 RPM was found to have much lower fatigue life than specimens machined at 3000 RPM.

3.4.Fractography

Tensile fractured surfaces were examined to investigate fracture mechanism under monotonic loads.Fig.8 show the typical tensile fractured surface.The SiC particles were fractured along the surface exhibiting cleavage lines typical of brittle fracture, identifie through point A in Fig.8b and c respectively.In addition, the SiC particles were also observed to be cracked

perpendicular to the fractured surface, shown (Fig.8c and e) identifie as point B.The matrix was found to form dimples typical of ductile fractures, point C as shown in Fig.8d.Particle debonding was not observed extensively in this case owing to the high wettability of magnesium matrix with reinforcements [21].Similar observations were made by Luo, in comparing the crack propagation in pure AZ91 and AZ91/SiCp composite [20].

Additionally, the fractured surfaces of specimen under fatigue loading was also investigated.Optical micrographs of the fractured surface were observed to identify the crack initiation sites on the surface.The cracks were clearly found to initiate from the defects on the peripheral milled surface.

Fig.9 shows the optical micrograph of fractured surface of specimen machined at 0.1mm/rev and 1000 RPM and the arrows indicate the crack initiation site.The crack initiation was typically found to occur on the peripheral milled surface(S2).At a load of 0.6Sut, three distinct regions can be identifie from the micrographs, the crack initiation site indicated by arrows, followed by smooth region corresponding to crack propagation and the rough region corresponding to fracture from overloading.However, at a high load of 0.8Sut,fractured surface did not show a significan crack propagation region.The crack initiation site was further investigated through SEM micrographs, shown in Fig.9.

Fig.11.SEM micrographs obtained at 0.1mm/rev, 1000 RPM, 0.6Sut, a)Fractured surface at 2000x, region E: ductile fracture of the matrix with smaller particles embedded, b) magnifie view at 4000x, region A: cleavage lines typical for brittle fracture, B: particle fracture perpendicular to the fractured surface, C: voids due to particle pull out, c and d) EDS maps of (a)showing Si and Mg respectively, e) Fractured surface at 8000x, region D:matrix cracking, f) EDS map of (e) showing Mg.

Characteristic inward lines were found to be emanating from the crack initiation site, shown in Fig.10.Upon further magnification a crack was found to originate from the surface.The presence of machining induced defects such as sharp valleys and the presence of particle agglomerations,shown in Fig.6, act as stress concentrators initiating cracks.

The crack propagation mechanism through the matrix and the SiC particles were investigated through SEM micrographs of crack growth region,shown in Fig.11.Five distinct regions could be identifie , (1) particle fracture along the surface,showing cleavage lines typical for brittle fracture, region A,(2) Particle fracture perpendicular to the surface, shown in region B, (3) voids due to debonding and particle pull out,region C, (4) matrix cracking, region D, can be attributed to overloading, (5) ductile failure of matrix with embedded SiC particles, region E.Most of the SiC particles were found to be embedded in the matrix and voids represent the particles pulled out from the matrix as in regions C and E.Particle fracture was not extensively observed in this composite due to the smaller size, with fractured particles being 10μm or larger.

4.Conclusion

Surface quality of the milled magnesium composites were evaluated through roughness measurements and inspection of machined surface SEM micrographs and its subsequent effect on their fatigue behavior was observed.

·Specimens machined at 0.1mm/rev exhibited a good surface finish withRa=0.29 and 0.37 μm at 1000 and 3000 RPM.At a higher feed rate,Rawas close to 1 μm.SEM micrographs indicated the presence of feed marks at higher feed rate, resulting in higher surface roughness.Surface defects such as voids and matrix cracking were observed on the surface.Consequently, specimen machined at 0.1mm/rev exhibited the highest fatigue life.

·Crack initiation site, propagation and fina fracture region could be identifie through optical micrographs.The crack was found to initiate from the peripherally milled surface.The extent of these regions was found to depend on the loading condition.

·Additionally, the crack propagation mechanism was identifie by investigating the SEM micrographs of the fractured surface.Particle fracture, matrix cracking, voids due to particle debonding, ductile failure of the matrix were observed.Owing to the smaller size of the reinforcement particles, particle fracture was not observed extensively.The smaller particles were embedded within the matrix or pulled out from the matrix as the crack propagated.

Acknowledgments

We sincerely thank Mr.Curt Lavender of Pacifi Northwest Laboratories for providing the material for this investigation.This research work was financiall supported through Boeing Pennell Professorship funds.Authors sincerely acknowledge the Department of mechanical engineering support throughout this investigation.