（9）High-speed turning experiments on metal matrix composites .pdf

High-speed turning experiments on metalmatrix compositesL. Iulianoa, L. Settineriaand A. Gattob,*aPolitecnico di Torino, Dipartimento di Sistemi di Produzione ed EconomiadellAzienda, C. so Duca degli Abruzzi, 24-10129 Torino, ItalybUniversita di AnconaDipartimento di Meccanica, Via Brecce Bianche-60131 Ancona,Italy(Received 1 October 1997; 12 May 1998)The hard abrasive ceramic component which increases the mechanical characteristics of metal matrix composites(MMC) causes quick wear and premature tool failure in the machining operations. The aim of the paper is tocompare the behaviour of high rake angle carbide tools with their diamond coated versions in high-speedmachining of an Al2O3Al 6061 MMC. The influence of the cutting parameters, in particular cutting feed andspeed, on tool wear and surface finish has been investigated. The higher abrasion resistance of the coatings resultsin increased tool life performances and different chip formation mechanisms. q 1998 Published by ElsevierScience Ltd. All rights reserved.(Keywords: metal matrix composites (MMCs); high-speed machining; tool wear)INTRODUCTIONThe benefit of using composite materials, and the cause oftheir increasing adoption is to be looked for in the advantageof attaining property combinations, that can result in anumber of service benefits. Among these are: increasedstrength, decreased weight, higher service temperature,improved wear resistance and higher elastic module.The main advantage of composites lies in the taylorabilityof their mechanical and physical properties to meet specificdesign criteria.In recent years, a new generation of materials known asmetal matrix composites (MMC)1,2has been developed, inorder to satisfy the demand for a high strength andtoughness material, capable of operating effectively underadverse conditions.The development of these materials began in the 1960swith the introduction of boron, graphite and aramide fibres.Work on MMCs resulted in several boron/aluminiumMMCs parts. However, interest in MMCs diminished inthe early 1970s as polymer matrix composites became thedominant material. The introduction of new fibre reinforce-ment materials in the late 1980s encouraged once again thenew development of MMCs.These materials are generally formed by a matrix which canbe made of any suitable metal (aluminium, magnesium, titaniumand some superalloys being the most popular), reinforced withSiC or Al2O3in the form of continuous or discontinuousfibres, whiskers or particles of the ceramic material.These advanced composites are considered excellentcandidates for high temperature applications. The specificmass of most MMCs is about one third that of steel, and thespecific strength and stiffness of these materials is quite high3. These properties are important in automotive andaerospace applications because of the potential for largereductions in weightup to 25%. Furthermore, their hightemperature strength retention is also an important feature,making them suitable candidate materials for use inautomotive and aircraft-engine applications.The properties that make MMCs appealing to engineeringdesigners can, however, present a major challenge whenattempting to machine these materials because of theirbrittle behaviour and high hardness.For instance, the presence of hard Al2O3particles inAl2O3/Al-alloy MMC materials involves considerabledifficulties when machining these materials using conven-tional methods such as turning, drilling, milling and sawing,etc., due to the high wear rate47. This is not surprisinggiven that Al2O3forms the basis of many cutting-toolmaterials. Therefore, machining MMCs using these con-ventional methods often involves frequent and expensivetool changes and therefore increased job-completion times.Turning, milling and drilling of MMCs, therefore, requiresthe use of carbide, diamond or hard-nitride-coated tools.Even then machining times tend to be two to four timesgreater than those for unreinforced matrix material becauseof increased tool wear, the necessity to use reducedComposites Part A 29A (1998) 150115091359-835X/98/$ - see front matterq 1998 Published by Elsevier Science Ltd.All rights reserved.PII: S1359-835X(98)00105-51501* Corresponding author. Tel.: +39-11-5647230; fax: +39-11-5647299;e-mail: lsettineathena.polito.itfeed-rates, and the need to achieve good surface finish,which is required by the brittle behaviour of the metal-matrix composites810.The difficulties associated with the machining of MMCsmust be minimised if these materials are to be used moreextensively. Therefore, the machining of MMCs is nowconsidered to be one of the most interesting areas ofmanufacturing science requiring urgent attention. SinceMMCs are relatively new materials, comprehensivemachinability data have yet to be established and this hasaroused some research interest5,6,813.In this paper the results of some investigation are outlined,involving the high-speed machining of a 10% particulateAl2O3/6061 T6 aluminium matrix composite by conventionallathe-turning. Carbide and diamond-coated carbide tools ofthe same geometry were used for the tests and measurementsof tool wear and part-surface finish were carried out. Testswere performed over a range of cutting speeds and feed rates.SEM observations of the chips, as well as the wear resultshave enabled assessments to be made of the mechanisms ofchip formation at different cutting speeds and feeds.The authors have previously reported13the results ofinvestigations on a (Ti,Al)N coating to improve tool life.However, these results indicated that the coating would notbe economically feasible, due to the small advantage in toolduration compared with the relatively higher costs.EXPERIMENTAL SET-UPTurning tests were performed on a numerically controlledvertical lathe which has a stiff, sturdy spindle with specialpurpose roller bearings, allowing spindle speeds to reach5000 rpm. The spindle is powered by a DC motor rated80 kW in continuous duty, with a peak power of 120 kW.Commercial uncoated and CVD diamond-coated carbidetools in the form of triangular indexable inserts were used,the geometry of these inserts being typical for the machiningof aluminium alloys. Polycrystalline diamond tools were nottested due to their high cost compared with the uncoatedcarbide tools. A tool holder with a lead angle x 918 andinclination angle of l 08 was used.Tests were carried on in continuous cutting on a 10%particulate Al2O3/6061 T6 aluminium matrix alloy, whosecomposition is reported in Table 1.InFigure 1 thehomogeneous reinforcing Al2O3particle distribution isshown, parallel and perpendicular to the extrusion direction.Following preliminary tests, the cutting conditions werechosen to avoid catastrophic failure. The selected cuttingconditions are reported in Table 2. Overall, 12 tests for eachtool type were carried out.After the machining of 150 3 103mm3and verifying thatthe pattern of tool wear was uniform, the flank wear VBwasmeasured according to the ANSI/ASME B94 55M stan-dards. A SEM picture of the cutting edge area of a worncarbide tool is shown in Figure 2 to show the uniformpattern of the flank wear.High speed turning experiments on metal matrix composites: L. Iuliano et al.1502Table 1 Chemical composition of the 6061 aluminium matrixSi Fe Cu Mn Mg Cr Zn Ti Ni Al% weight 0.59 0.15 0.27 0.004 1.04 0.09 0.002 0.006 0.001 bal.Figure 1 Al2O3particle distribution parallel (a) and perpendicular (b) to the extrusion directionTable 2 Cutting conditionsInserts Uncoated carbide grade ISO KDiamond coated carbide grade ISO KRake angle (g) 208Nose radius (r) 0.4 mmMachined volume (Y) 150 3 103mm3Depth of cut (a)2mSpeed (Vt) (630, 800, 1000, 1260) m/minFeed (f) (0.03, 0.10, 0.17) mm/revCooling noneSurface roughness Rawas measured on the machinedsurfaces by a Hommel T1000 stylus instrument. Measure-ments were made in four positions spaced at v 908 aroundthe workpiece to obtain medium values and theirdistributions.Chips and tools were prepared metallographically andexamined by SEM microscope using both secondaryelectrons (SE) and backscattered electrons (BSE). EDAXsemi-quantitative analysis was also carried out.RESULTS AND DISCUSSIONThe tool wear VBand the average surface roughness Ravalues obtained for each parameter couple Vtf usinguncoated and diamond-coated tools were statisticallyanalysed using multiple linear regression. The independentvariables are feed, cutting speed and their selected secondorder terms. The models, explaining a fair percentage oftotal variation, fitted the observed responses in terms ofmain factor and second terms. The significance of themodels and the parameters were checked using varianceanalysis; the percentage of variability explained by bothmodels is greater than 80%.The analysis criteria did not allow us to exclude theeffects of some variables because they merely pinpointedwhich associations were the stronger. It should be under-lined that the validity of the models, for predictive purposes,did not exceed the relevant sample space defined by thecombination of the tested independent variables. Notice thatthe preliminary analysis of the experimental data excludedthat the behaviour of tool duration versus cutting speedagrees with Taylors law. Statistical analysis showed thedata to fit a second order polynomial model.Tool wearIn Figures 3 and 4 the wear behaviour of the uncoated anddiamond-coated tools is shown. The main affecting factorsare: feed, square feed, square cutting speed and cuttingspeed times feed, according to the following models:VB 0:33 4:63 3 f 23 3 10 83 V2t 24:74 3 f2 0:002 3 Vt3 ffor the uncoated tools, andVB 0:11 0:95 3 f 2:3 3 10 83 V2t 3:82 3 f2 0:0003 3 Vt3 ffor the diamond-coated tools.1503High speed turning experiments on metal matrix composites: L. Iuliano et al.Figure 2 SEM picture of the cutting edge area of a worn uncoated carbidetool, Vt 630 m/min, f 0.03 mm/revFigure 3 Computed model of the flank wear VB(mm) for uncoated carbide tools versus cutting speed and feed. The stars indicate the experimental resultsover which the model has been computed. Model used: VB 0:33 4:63 3 f 23 3 10 83 V2t 24:74 3 f2 0:002 3 Vt3 f . Statistical parameter: R2,Adj 0.92In fact it can be observed that, for a given removedmaterial volume, minimum flank wear is achieved with aparticular combination of low cutting speed and relativelyhigh feed. The combination of low feed and high cuttingspeed is, on the other hand, not desirable. From acomparison of the two pictures it can be noticed that thediamond-coated tools show, as expected, a much higherwear resistance and the minimum for the flank wear isachieved at higher cutting feed and speed. In Figure 5 a greyzone is shown, representing the parameters combinationarea that can be explored safely only with diamond-coatedcutting tools.The higher wear resistance of the coated tools is due tothe higher abrasion resistance of the diamond coatings withrespect to the carbide.Surface finishThe analysis of surface roughness Ravalues clearlyshows that the roughness does not depend on the positionangle v, this fact being due to the uniform distribution of theAl2O3reinforcement on the matrix.The four Ravalues measured for each of the machiningtests performed with the uncoated tools were processed toHigh speed turning experiments on metal matrix composites: L. Iuliano et al.1504Figure 4 Computed model of the flank wear VB(mm) for diamond-coated carbide tools versus cutting speed and feed. The stars indicate the experimentalresults over which the model has been computed. Model used: VB 0:11 0:95 3 f 2:3 3 10 83 V2t 3:82 3 f2 0:0003 3 Vt3 f . Statistical parameter:R2,Adj 0.91Figure 5 The feedspeed combinations falling into the grey area can only be safely explored with the diamond-coated cutting toolsobtain the average values Ramwhich were used in thestatistical analysis, which yielded the following model:Ram0:36 0:0013 3 Vt 18:40 3 fIn Figure 6 the surface roughness behaviour of the uncoatedtools is shown. Second order terms of cutting speed and feedand their product have no significant effects, therefore thedominant factors are cutting speed and feed.The parts machined with the diamond-coated tools didnot show any significant correlation between the cuttingparameters and the surface roughness.Chip formationThe tool, chip and workpiece surfaces were observed bySEM, using SE and BSE techniques.From the analysis of the chips images it can be observedthat, at f 0.03 mm/rev, a detachment between the matrixand reinforcing particles occurs. This phenomenon, inde-pendent of the cutting tool material, becomes more evidentas the cutting speed increases (Figure 7).At higher feedrates, the effect of the cutting speed is lessstrong. From the SEM observations of the lower surfaces ofchips machined at f 0.17 mm/rev, no differences can benoticed between Vt 630 m/min and Vt 1260 m/min(Figure 8).At a cutting speed of 630 m/min, some scraps parallel tothe flow speed have been noticed, which become moreevident as cutting speed and feedrate increase. At the endof the scraps, a reinforcing particle partially emergingfrom the matrix can often be observed. A simpleexplanation of this could be that the reinforcing particlesare stuck on the tool rake and engrave the scratches on thechip, until they sink, at least partially, in the chip matrix1505High speed turning experiments on metal matrix composites: L. Iuliano et al.Figure 6 Computed model of the average surface roughness Ram(mm) for uncoated carbide tools versus cutting speed and feed. The stars indicate theexperimental results over which the model has been computed. Model used: Ram0:36 0:0013 3 Vt 18:40 3 f . Statistical parameter: R2,Adj 0.83Figure 7 Lower surfaces of chips cut with the diamond-coated tools; the emerging particles are visible: (a) Vt 630 m/min, f 0.03 mm/rev; (b) Vt1260 m/min, f 0.03 mm/revandarecarriedawayfromthetoolrake.Toverifythis,some longitudinal sections of the chips have been cut andobserved (Figure 9).From the section images it is evident that the reinforcingparticles tend to sink and pile up in the matrix along theshear planes described by the Pijspanen model (Figure 10).When enough particles have piled up to prevent anotherparticle from sinking, this last one sticks out partially fromthe matrix. The effect is stronger in chips obtained at highercutting speeds. The higher temperatures developed at highercutting speeds ease the transportation of the particle alongthe shear planes. Furthermore, the shear planes becomemore irregular and the number of scratches and ridgesvisible on the lower surfaces increases.The same behaviour is induced by an increase in cuttingfeed: the distance between the shear plane grows (Figures1113).As regards the influence of the coatings, it can be saidthat, under unchanged machining parameters, withdiamond coatings, the distance between the shearplanes is higher than the one obtained using a carbidetool and that the chips lower surfaces are moreHigh speed turning experiments on metal matrix composites: L. Iuliano et al.1506Figure 8 Lower surfaces of chips cut with the diamond-coated tools; the piling-up of the particles can be seen: (a) Vt 630 m/min, f 0.17 mm/rev; (b) Vt1260 m/min, f 0.17 mm/revFigure 9 Longitudinal chip sections ( 3 50) machined with uncoated tools: (a) Vt 630 m/min, f 0.03 mm/rev; (b) Vt 1000 m/min,f 0.03 mm/revFigure 10 Model of dispersion of reinforcing particles along the shearplaneshomogeneous and regular when machined with a coatedtool. This might be explained by the lower frictioncoefficient of the coatings1416.Worn cutting toolsThe thermal conductivity of the diamond coatings isabout 1020 times higher than that of the cemented carbide(10002000 W/mK vs. 100 W/mK). However, along thegrowth direction its value is twice that measured perpendi-cularly to the growth direction. This fact promotes the hightemperature gradient at the interface between the coatingand the substrate and relatively low temperature in the chip16: there is a great difference between Al2O3reinforcementhardness value and the soft matrix one, so that when the harddiamond tool edge comes into contact with the hard1507High speed turning experiments on metal matrix composites: L. Iuliano et al.Figure 11 Longitudinal sections of chips ( 3 50) machined with diamond-coated tools at f 0.17 mm/rev: (a) Vt 630 m/min; (b) Vt 1000 m/min; (c)Vt 1260 m/min; the piling-up of the particles can be clearly observedFigure 12 Shear planes on the upper surface of chips cut with the diamond-coated tools: (a) Vt 630 m/min, f 0.03 mm/rev; (b) Vt 630 m/min,f 0.17 mm/revFigure 13 Shear planes on the upper surface of chips machined with uncoated tools: (a) Vt 630 m/min, f 0.03 mm/rev; (b) Vt 630 m/min,f 0.17particles these are moved; the same effect has beendescribed for PCD tools in Ref.9. The observed chipformation mechanism is not much different from the onedescribed by the authors in Ref.13for (Ti,Al)N coated tools,but here higher localised forces are required to move theparticles in the matrix due to the lower chip temperature.The effects of these localised forces and thermal load can beobserved on the diamond coated tools after machining(Figure 14). A large material adhesion area (white zone) canbe noticed on the tool flank; the extension of such an areaincreases with the cutting speed. The adhesion areadecreases as the feedrate increases. Immediately underneaththe adhesion zone a dark line is evident, parallel to thecutting edge in which the tool coating has been removed.Another area in which the coating has disappeared can beseen on the tool rake, close to the cutting edge. It must betaken into account that adhesion between diamond coatingand its substrate still continues to be the critical factor,particularly in case of a carbide substrate17.CONCLUSIVE REMARKSThe following conclusions can b