點膠機械手的結(jié)構(gòu)設(shè)計,機械手,結(jié)構(gòu)設(shè)計 International Journal of Machine Tools & Manufacture 42 (2002) 11051112Study of precision micro-holes in borosilicate glass using microEDM combined with micro ultrasonic vibration machiningB.H. Yana, A.C. Wanga, C.Y. Huanga, F.Y. HuangaaDepartment of Mechanical Engineering, National Central University, Chung-Li, Taiwan 32054, ROCReceived 11 November 2001; received in revised form 10 May 2002; accepted 14 May 2002AbstractBecause of its excellent anodic bonding property and surface integrity, borosilicate glass is usually used as the substrate formicro-electro mechanical systems (MEMS). For building the communication interface, micro-holes need to be drilled on this sub-strate. However, a micro-hole with diameter below 200 m is difficult to manufacture using traditional machining processes. Tosolve this problem, a machining method that combines micro electrical-discharge machining (MEDM) and micro ultrasonic vibrationmachining (MUSM) is proposed herein for producing precise micro-holes with high aspect ratios in borosilicate glass. In theinvestigations described in this paper, a circular micro-tool was produced using the MEDM process. This tool was then used todrill a hole in glass using the MUSM process. The experiments showed that using appropriate machining parameters; the diametervariations between the entrances and exits (DVEE) could reach a value of about 2 m in micro-holes with diameters of about 150m and depths of 500 m. DVEE could be improved if an appropriate slurry concentration; ultrasonic amplitude or rotational speedwas utilized. In the roundness investigations, the machining tool rotation speed had a close relationship to the degree of micro-hole roundness. Micro-holes with a roundness value of about 2 m (the max. radius minus the min. radius) could be obtained ifthe appropriate rotational speed was employed. 2002 Elsevier Science Ltd. All rights reserved.Keywords: Borosilicate glass; Micro electrical-discharge machining; Micro ultrasonic vibration machining; Micro-hole; micro-tool; High aspect ratio1. IntroductionIn packaging MEMS-related devices, such as microvalves and micro flow sensors etc., borosilicate glass isusually used as the substrate for bonding with siliconwafers. To build up the electrical through channel andconnect the internal system between the silicon waferand environment, micro-holes are drilled in the glass sur-face before they are bonded. Drilling these micro-holesis difficult with traditional machining processes. Thereare several methods used to manufacture micro-holes,MEDM 1,3,5,8,9,11, excimer laser drilling 6, LIGA2,10, electrochemical discharge machining (ECDM)12 and MUSM 4,7 etc. Because of the different work-ing mechanisms, the results produced by these methodsare distinct.Correspondingauthor.Tel.:+886-3-4267353;fax:+886-3-4254501.E-mail address: bhyencc.ncu.edu.tw (B.H. Yan).0890-6955/02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.PII: S0890-6955(02)00061-5For example, in MEDM, a micro-hole with a diameterof 160 m and depth of 380 m could be drilled within2 min 5. MEDM can be used to manufacture only con-ductive material and the recast layer on a machined sur-face, containing craters and micro cracks will cause poorsurface and size accuracy. Laser micro machining tech-nology can be used to fabricate a hole under diameterof 4 m 6. However, laser beam machining causesdeterioration and micro-cracks on the machined surface.The LIGA technique has been found suitable for produc-ing three-dimensional microstructures with micro-holesin metal, polymers and ceramics 2,10. However, theLIGA method affects the configuration precision inmicro-hole machining with high aspect ratios because oflight diffraction (such as X-ray). ECDM can improve thematerial removal rate (MRR) and surface roughness to1.5 mm/min and 0.08 m, respectively 12, but as withchemical etching, the walls of the micro-holes will beover etched due to the ECDM process. MUSM has beenproved successful in hard and brittle material. Masuzawaet al. demonstrated that micro-holes as small as 5 m1106B.H. Yan et al. / International Journal of Machine Tools & Manufacture 42 (2002) 11051112(depth 10 m) and 3-D micro machines could be createdbycombiningwireelectricaldischargegrinding(WEDG), MEDM and MUSM 4,7. Because MUSMrelies on the micro mechanical forces to removematerial, for micro-hole machining with high aspectratios, small changes in the mechanical forces can havea significant consequence on manufacturing stability andprecision. Furthermore, the mechanical forces dominatedthe MUSM parameters. In depth studies have not beenconducted on how these parameters affect the manufac-turing stability and accuracy.A machining method combining MEDM and MUSMhas been designed to finish micro-holes with high aspectratios. During the entire machining process, the micro-tool was remained in the same fixture, so tool eccen-tricity problems were avoided. For avoiding the micro-tool oscillating or breaking during the manufacturingprocess, the ultrasonic apparatus was set up the side tovibrateworkpiece.Thisarrangementsignificantlyenhanced the micro-holes machining precision.2. Method2.1. Experimental set-upThe experiment equipment consisted of an EDMmachine, a four-axis control system and an ultrasonicmachining unit (as shown in Fig. 1). The four-axis con-trol system was fixed onto the EDM worktable. Here theborosilicate glass or copper plate moved along the frontFig. 1.The configuration of MEDM and MUSM apparatus: U, ultra-sonic vibration equipment; OS, optical scale; OP, optical scale counter;X, Y and Z, motors for x-, y- and z-axes movement; Cu (copper plate),EDM electrode; H, rotating chuck holder; t, micro-tool; D, computercontrolled display; C, motor for c-axis rotation; ID, interface circuitand motor driver; G, function generator; CPU, computer.and back direction using motor X and moved up anddown using motor Y. The micro tool was clamped intoa horizontal chuck rotated by motor C and directed leftand right by motor Z. The movement resolutions ofmotors X, Y and Z were 0.2 m, 0.2 m and 0.5 m,respectively. To enable removing the debris easily fromthe micro-holes during MUSM, the machining operationwas performed horizontally. The ultrasonic machiningunit (frequency, 30 kHz) included an electronic gener-ator, a transducer and horn-tool combination equipment.The tool was a cylindrical rod screwed onto the horn tip.A small piece of borosilicate glass was chemically gluedonto a small rectangular plate 4 attached to the tool tip(as shown in Fig. 2).2.2. MaterialsBorosilicate glass (Pyrex, Corning 7740) is a silicacomposition with excellent anodic bonding property,surface integrity, thermal properties and acid resistance.This glass has always served as the substrate for microsensors. Because borosilicate glass is hard and brittle, itis very suitable for micro-hole drilling using the MUSMmethod. For maintaining precise micro-hole sizes andshapes, the micro-tools must have high wear resistanceand rigidity. A circular tungsten carbide rod with a diam-eter of 0.3 mm was selected as the MEDM and MUSMtool 4. Micro-hole precision can be improved by usingoil as the slurry medium 1315. Silicon carbine grains,suspended in kerosene, were chosen as the workingslurry whose concentrations were 10%, 20% and 30%in the MUSM process. And the averaged abrasive sizeswere 1.2 m (about 75% particle sizes were from 0.9 to1.5 m) and 3 m (about 75% particle sizes were from2.6 to 3.4 m).2.3. Machining proceduresThe machining processes were divided into two mainparts. First, the tungsten carbide rod was fashioned intoa micro-tool using a copper plate as electrode in theMEDM step. This tool was then used with the MUSMprocedure to drill a micro-hole in the borosilicate glass.The above procedures are described in detail below:To EDM the micro-tool, the circular tungsten carbiderod was fixed at horizontal direction and rotated clock-wise. At the same time, a copper plate was fastened toa jig and moved vertically up and down automatically.The diameter of the tool was reduced by the movingplate edge (as shown in Fig. 3(a) and the EDM dielec-tric was sprayed to the working area when MEDM wasbeginning. The completed micro-tool was 2 mm longand diameter 150 m. To produce high stress concen-tration in the workpiece during MUSM, the front end ofthe micro-tool was reduced in diameter to 20 m andlength 0.2 mm. Fig. 3(b) displays the finished micro-1107B.H. Yan et al. / International Journal of Machine Tools & Manufacture 42 (2002) 11051112Fig. 2.The detail diagrams of experimental apparatus at MUSM process.Fig. 3.The micro-tool machining procedure and micro-tool finishingshapes after MEDM. (a) Using Cu electrode to fashion a micro-toolin MEDM process. (b) The micro-tool finished shape.tool. The experimental parameters for the MEDM pro-cesses are listed in Table 1.With the micro-tool in the chuck, a micro-hole wasdrilled in the glass using MUSM. To decrease theattrition of the tool at the lower level 15, the tool can-not touch the workpiece before the machining processstart, so it existed about 0.1 mm between the tool andglass surface when the machining process was begin-ning. In micro-hole fabricating procedure, the micro-feeding tool accompanying with spray slurry was util-Table 1The experimental MEDM parametersWorkpieceCircular rod of tungsten carbide (K20)ElectrodeCopperWorking fluidKeroseneRotational speed of50, 150workpiece (rpm)Polarity+ (rough); ? (finish)The open load voltage (V)100Working voltage (V)25Discharge current (A)0.95, 1.45Pulse duration (s)4, 10Off time (s)4, 10ized to manufacture the glass with ultrasonic vibration(as shown in Fig. 2). The flow rate of slurry was 450ml/min. Table 2 lists the experimental parameters for theMUSM processes.3. Experimental resultsIn addition to evaluating the size accuracy of high pre-cision micro-holes, the shape precision and surfaceroughness was estimated. Hence, the following dis-cussion is organized into three main parts: (A) diametervariation between the entrance and exit (DVEE), (B)roundness and (C) surface roughness. The factors affect-ing the precision of micro-holes include the slurry con-centration, abrasive grain size and the MUSM machin-ing parameters.3.1. DVEE of micro holesIn the MUSM processes, the micro-feeding method ofmicro-tool was applied to manufacture the micro-holes.In machining micro-holes with high aspect ratios, thetools could touch the walls of the holes for a long time,causing abrasion to the sides of the tools, or inducingirregular expansion of the micro-holes. These conditionswill reduce the accuracy of the holes. DVEE of themicro-holes is an essential element in the MUSM pro-cess. The following sections utilize several MUSM para-meters to study the DVEE forming effects.Table 2The experimental MUSM parametersUltrasonic vibration directionLongitudinalUltrasonic vibration frequency30 kHzUltrasonic vibration amplitude (m)1.2, 1.4, 1.6, 1.8,2.0, 2.2Rotational speed of micro tool (rpm)50, 100, 150, 200, 250Feed rate (m/min)6, 6.7, 7.5, 8.6, 10Concentration of slurry (wt %)10%, 20%, 30%Averaged abrasive size (m)1.2, 31108B.H. Yan et al. / International Journal of Machine Tools & Manufacture 42 (2002) 110511123.1.1. The effects of abrasive slurry concentration andgrain sizeThe abrasive slurry concentration and grain size arethe most important factors affecting MUSM machiningprecision. Abrasive with higher slurry concentrations,the material removed by the abrasive grains at the mach-ining surface will be faster than the lower slurry concen-trations. The fast material removal will reduce the fric-tion between the micro-tool front end and the micro-holewall in the machining process. The DVEE will be lowerwhen higher slurry concentrations are used. Fig. 4 dis-plays that whether the averaged abrasive size was 1.2m or 3 m, a 20% slurry concentration would producea smaller DVEE than a 10% slurry concentration. ButDVEE became larger when the slurry concentrationreached 30%. The 20% concentration provided almostdouble abrasive particles to manufacture the hole thanthe 10% concentration at an average size of 1.2 m,causing the DVEE to become smaller. However, becausethe micro-hole machining was set up in the horizontalmode, the abrasives would gather between the holeentrance and the tool, these particles were be fed intothe hole by rotating tool and ultrasonic amplitude. Butit would be hindered abrasives to enter the hole whenthe amount of particles was gathered too much betweenthe hole entrance and tool, thereby influencing theDVEE of the micro holes. At 30% slurry concentration,this situation would become clearer, so the micro-holedrilling effect significantly decreased. Moreover, theaverage grain size (3 m) was bigger than the ultrasonicamplitude (1.8 m), inducing abrasives hard to enter thehole during the MUSM. So the machining effect onDVEE was not obvious than the small grain size. Fig.4 also shows that employing the 1.2 m averaged par-ticle size created a better DVEE than the 3 m averagedparticle size. At the same concentration, the small abras-ive particles were more uniformly suspended in theFig. 4.The abrasive slurry concentration and grain size effect onDVEE through MUSM.slurry and easily entered to the hole than the large one.However, the MRR was less for each grain. Therefore,a smoother machining surface and a straighter cross sec-tion of micro-hole could be obtained, improving theDVEE of the micro-holes. To obtain finer finishingeffects, the following experiment used a 20% particleconcentration with an averaged diameter of 1.2 m.3.1.2. The effect of ultrasonic vibration amplitudesIn the USM procedures, larger machining tool ampli-tudes cause higher MRR 15,16. The machining toolmay bend in the drilling process when the ultrasonicvibration amplitudes are large. This will affect the exact-ness of the holes. This phenomenon is more apparentduring MUSM. In these experiments, ultrasonic ampli-tude was measured using a tool microscope (1000)three times (in air), and then took the averaged valueas the working amplitude. Fig. 5 presents the effect ofultrasonic vibration amplitudes on DVEE. The figureshows that the DVEE decreased with increasing ampli-tude from 1.2 m to 1.8 m. The DVEE increased whenthe amplitude increased from 1.8 m to 2.2 m. Thisindicates that the appropriate amplitudes could increasethe preciseness of the micro-holes. However, utilizingsmaller amplitudes to manufacture micro-holes wouldincrease the machining time and cause more abrasion ofthe micro-tool, producing a larger DVEE. Further, themachining time became shorter when the amplitudeswere increased. This reduced the wear on the micro-tool.A lower DVEE could therefore be found. Owing to athe slender ratio in the MUSM process, micro-toolswould be bent because of the greater amplitudes. Thisinduced irregular machining of the micro-holes (asshown in Fig. 6), making DVEE values larger. Themicro-tools could also be broken when the irregular holemachining became serious.Fig. 5.The ultrasonic vibration amplitude effect on DVEE viaMUSM.1109B.H. Yan et al. / International Journal of Machine Tools & Manufacture 42 (2002) 11051112Fig. 6.The irregular expansion of micro-hole produced by ultrasonicamplitude 2.2 m (averaged abrasive size 3 m).3.1.3. The effect of rotational speeds of micro-toolsThe rotational speed of the micro-tools is also a keyparameter affecting the micro-holes accuracy. Because arotating tool can assist the suspended particles inentering the micro-hole, the arrangement can drive theparticles to grind the hole during the MUSM process.Therefore, the DVEE of the micro-holes, produced byrotating tools, will be better if the tools are not rotated.Fig. 7 shows the effect of rotational speed on DVEE.The experiments illustrated that the DVEE becamesmaller when the rotational speed was increased from 50rpm to 150 rpm. The DVEE changed greater when therotational speed was increased from 150 rpm to 250 rpm.This revealed that the correct rotational speed couldenhance the micro-holes accuracy. The abrasive particleswere fed into the hole via the rotational tool and ultra-sonic vibration. At the same vibration mode, the numberof abrasive grains fed into the hole when the rotationalspeeds were increased at the beginning stage. The mach-ining efficiency was therefore enlarged, and micro-toolwear was reduced, so a smaller DVEE could be obtained.Fig. 7.The rotational speed effect on DVEE by MUSM.The abrasion of the tool side and hole surface will beincreased by abrasive particles when rotational speed isincreased 17. This result will be clearer with higherspeeds. Moreover, the stability of the cutting process isalso affected by high speeds. Due to these reasons,DVEE not only became large but also had obviouslychanged after 150 rpm.3.1.4. The effect of feed rate on the micro-toolsDVEE is influenced by changes in the feed rates dur-ing MUSM. In these experiments, feed rates utilized theprogram interface to control motor Z and optical scale,producing constant feed rates. Fig. 8 details the effectof feed rates on DVEE. This figure shows that DVEEwas smaller when a lower feed rate was employed.DVEE became large when a large feed rate was used.However, the gap between the micro-tool end surfaceand glass face became smaller when a larger feed ratewas used. This smaller gap induced poor slurry circu-lation. When this occurred, fewer abrasive particlesentered the gap through MUSM, inducing a not verygood working effect; the front end of the tool producedmore abrasion during this machining process. Hence, theDVEE became large. Fig. 9 shows a SEM photographof the worn micro-tool. Fig. 9(a) represents the smallcircular step at the front end of the micro-tool seriouslyabraded from a larger feed rate machining (8.6 m/min).The tool suffered less wear at the same position whena smaller feed rate was used (6 m/min), as shown inFig. 9(b).3.2. RoundnessIn the USM processes, tool rotation or not, definitelyinfluences the roundness of the holes 15,18. Toolrotation aids the suspended particles to enter the holes,thereby increasing the working efficiency of the USM.Rotating tools can also induce the particles to grind theholes, thereby improving the roundness of the holes. InFig. 8.The feed rate effect on DVEE via MUSM.1110B.H. Yan et al. / International Journal of Machine Tools & Manufacture 42 (2002) 11051112Fig. 9.The SEM of micro-tool wear after MUSM. (a) At higher feedrate of 8.6 m/min. (b) At appropriate feed rate of 6 m/min.these experiments, the roundness of exits is discussedherein, because the exits, such as nozzles, adequatelyaffect the micro holes performance. The roundness com-putation used the measuring program to gauge the SEMimages of micro-holes, taking the max. radius minus themin. radius as roundness values. Fig. 10 displays theeffect of rotational speed on the roundness of micro-holes. This figure illustrates that micro-hole roundnessFig. 10.The rotational speed effect on roundness by MUSM.was better with rotational speeds from 50 to 150 rpm.The roundness values became larger when the rotationalspeed was increased from 150 to 250 rpm. This wassimilar to the rotational speed effect on DVEE. How-ever, high rotational speed not only caused more abra-sion at the tool but also induced instability in the cuttingprocess, prompting clearly out-of-round micro-holesafter 150 rpm. In these experiments, the best roundnessvalue found in this study was about 2 m. Fig. 11presents micro-holes with acceptable entrances and exitsproduced at 150 rpm rotational speed.3.3. RoughnessIn the USM process, the rotation effect of the toolcan drive the abrasive particles to grind the hole surface.Therefore, the surface roughness value is generally lessFig. 11.The shapes of micro-holes at rotational speed of 150 rpmthrough MUSM. (a) The entrance of the micro-hole. (b) The exit ofthe micro-hole.1111B.H. Yan et al. / International Journal of Machine Tools & Manufact