Available online at www.sciencedirect.comTransactions of罵ScienceDirectNonferrous MetalsSociety of ChinaEL SEVIER PressTrans. Nonferrous Met. Soc. China 20(2010) 980-986www.nmsc.cnDynamic measurement on infiltration process andformation mechanism of infiltration frontQI Le-hua(齊樂(lè)華)，XU Rui(徐瑞), sU Li-zheng(蘇力爭(zhēng)), ZHOU ji-ming(周計(jì)明), GUAN Jun-tao(關(guān)俊濤)School of Mechatronics, Northwestern Polytechnical University, Xi'an 710072, ChinaReceived 3 August 2009; accepted 2 February 2010Abstract: An infiltration measurement device was developed to research the infiltration process of molten AZ9ID magnesium alloyinto the Al2O3 short fbre preform. The variation of relationship between the heights of measuring points and the time for moltenalloy to reach the measuring points was ilustrated. The effect of infilration process parameters on the infiltration front was analyzed.It is found that pressure and pouring temperature are the most important factors which affect the infiltration velocity and compositequality. Furthermore, considering the infuence of temperature field, an infiltration model of molten AZ91D into the short fibrepreform was constructed on the basis of experimental results and Darcy's Law. The analysis shows that the results predicted by thismodel are consistent with the experimental results.Key words: infiltration measurement; infiltration front; temperature field; infltration modelexample, TURNER et al[1] investigated the injection1 Introductionprocess of glass fibre reinforced plastic by visualizingthe flling process.YOKOI et al[12-14] analyzed theAs an indispensable step in the fabricationshifing phenomenon of melt forefront duringtechnologies of metal matrix composites (MMC), themulti-cavity injection molding processliquid metal infilration process in porous preform, suchvisualization mould. HE et al[15-16] used a suitableas squeeze casting, vacuum infiltration, variable pressureliquid and filler particles at room temperature to simulateinfiltration and liquid infiltration- extrusion, havelow-pressure infiltration process of high- temperatureattracted research interest due to significantly effectivealuminum melt in porous media in an organic transparentimprovement in the properties of composite productsdevice. However, the methods are difficult to reflect the[1-3]. Much research work was published on theactual infitration process due to the high temperatureinvestigation of the infiltration process of liquid metalcondition and the opacity of the preform during liquidinto porous preform[4- 6]. SCHULZ and KAUFMANNinfiltration processes for MMC fabrication. LONG et[7], and MATSUNAGA et al[8] designed an infiltrational[17] and EARDLEY et a[18] obtained the changingdevice to control the process parameters of gas pressureinformation of temperature during the infiltration processinfilration. MORTENSEN et a[9 - -10] described fluicwith thermocouples which were inserted into the preform.flow and heat transfer during the liquid metal infiltrationIt should be noted that the porous structure of preform isprocess and a kinetics model was presented. But fewdamaged in this way. As a result, the infiltrationstudies were reported on the experimental research ofregularity will be changed.intermnal infilration process.In this work, the measurement device for negativeThe methodology for the experimental investigationpressure infilration was developed. It acquires theof infiltration process mainly includes two ways atchange of temperature information during the infiltrationpresent. One is a direct observation by visualizationprocess of liquid magnesium alloy into Al2O3 preform bymould and transparent fiber and the other adopts the wayutilizing a self- designed sleeve. The dynamic evolutionof physical simulation to analyze low-pressureof infiltration nrcece was vinalized by analyzing theinfiltration process based on the similarity principle. Fortemp中國(guó)煤化工dering the influenceFoundntion item: Project(50575185) supported by the National Natual Science FoTYHCNMHGoporedbytheDoteFoundation of Northwestern Polytechnical UniversityCorresponding author: QI Le-hua; Tel: +86. 29-88460447; E-mail: qilehua@nwpu.cdu.cnDOI: 10.1016/S1003-6326(09)60245-4QI Le-hua, et al/Trans. Nonferrous Met. Soc. China 20(2010) 980-986981of temperature field, a flow model of molten magnesiumalloy was constructed on the basis of experimentalnresults and Darcy's law.2 Experimental2.1 Experimental materialAZ9ID magnesium alloy was used in thexperiments as matrix， and the compositions andproperties are given in Table 1. The compositions andproperties of short Al2O3 fiber for preparing preform areshown in Table 2.Wet forming process was used to prepare preform.Firstly, proper amounts of binder and dispersant wereadded into the fibre solution. Secondly, Al2O3 fberdispersion was poured into the forming die and drainedFig.2 Macrotructure of Al2O3 preform prepared by wetunder pressure. Finally, the preform was formed. Theforming processprocess is ilustrated in Fig.1.0- -1 MPa by adjusting the inlet valve, so that the liquidContainerStirring rodForming diemagnesium alloy infiltrates into the Al2O3 fbrous, Fibre。Bladeiterpreform. Temperature control system is used to meltAZ91D magnesium alloy and preheat short Al2O3 fbrepreform. The degree of vacuum in infiltration process is(aensured by the vacuum system. A high-precisionPunchthermocouple with a diameter of 1 mm is used in thetemperature measurement system to obtain thetemperature information during infiltration process.口(c(d)Thermocouples are inserted into the stainless steel tubeFig.1 Schematic diagram of process of preparing preformand contact closely with the inclined plane of sleevethrough the hole located on the positioning cover. TheAfter mould pressing, the preform was dried for 24distribution of the located hole on the positioning coverh at room temperature, and subsequently dried in an ovenis sbown in Fig.3(b). Temperature variations during theat 250 C for 4 h. Finally, short Al2O3 fibre preforminfiltration process of liquid AZ91D alloy into porouswith certain compressive strength was obtained, and thepreform were obtained from nine high-precisionmorphology is shown in Fig.2.thermocouples on the inclined plane of sleeve.2.2 Experimental device2.3 Experimental methodFig.3(a) shows the schematic diagram of theThe inclined plane of sleeve contacts with thedeveloped infilration measurement device. It ispreform in the infiltration process. When liquidassembled by infiltration systems, pressure controlmagnesium alloy reached the inclined plane of preform,system, temperature control system, vacuum system andtemperature changed on the corresponding location oftemperature measurement system. The pressure controlthe inclined plane. This information can be captured bysystem precisely governs the pressure of crucible withinusing the high-precision thermocouple. Due to the axialTable 1 Compositions and properties of AZ9ID magnesium alayw/%_Deasity/(gcm)Solidus'CLiquidus/CAL_ZnMnMg8.5-9.50.45-0.90.17-0.3≤0.0SBal.1.81470595中國(guó)煤化工Table 2 Compositions and properties of Al2O3 fibreComponentPhaseLength/umMHCNMH GStrengthMPaSiO2<20%, Al203>80%a-Al2O3>503-103-51130982QI Le-hua, et al/Trans. Nonferrous Met. Soc. China 20(2010) 980- -986Collctionsystem.50mm43.33 mmStainless tubePositioningVacuum昏36.67 mmpumping.30mm' systemfitration5- -. 23.33 mm/ cavity16.67 mm10mmStainless steelcrucible) mm。Temperaturecontrollerd45 mmGasholdera)Fig.4 Schematic distributions of measuring points in Al2O3preform-0-3 Results and discussion3.1 Analysis on infiltration process of liquidmagnesium alloy into Al2O3 preformRelationships between heights of measuring pointsand the time that melted alloy reaches the monitoredpoints are ilustrated in Fig.5, which were drawn fromthe temperature records of thermocouples inserted in theinclined plane of preform along the axial and radialdirection. It was shown that magnesium alloy meltb)moved fast along the path from point C to D, and it tookfig.3 Scheme of experimental facility and thermocouplelonger time to infiltrate from point F to G at the samecondition of 0.4 MPa and 800 C. This means that thelocations: (a) Infiltration measurerment device; (b) Vertical viewinstable infiltration occurred at the above mentionedof positioning covercondition. Fortunately, stable infiltration can be obtainedgradient of preform, the changing temperatureby increasing pressure and decreasing melt temperature,which can be seen from Figs.5(b)- -(c). In other words,information reflected the infiltration process of poroustime for infiltration from point C to D is nearly equal topreform by molten magnesium metal. The variation ofthat from point F to Grelationship between the beights of measuring points andFig.5 shows that the molten magnesium firstlythe time of corresponding height was obtained based onreached point K in preform edge, and then reached pointanalyzing the temperature information.B, finally the core region E, where the temperature wasFig.4 shows the distribution of the measuring pointsthe lowest. This was because the preheating temperatureon the inclined plane of prefomm. The temperatureof preform decreased from the edge point K to the coreinformation of infilration process was recorded frompoint E during the preheating process, as shown in Fig.6.nine measuring points, six of which are in axial directionThe cooling rate of the liquid magnesium alloy in theand four are in radial direction. The change ofpreform increased owing to the decrease of prcheatingtemperature information on the height of preform can betemperature in the core position E, which resulted in theobtained at the six measuring points in axial gradient.viscosity and the viscositv resistance of liquidTake point K as the starting position of infiltration, thmagn中國(guó)煤化工to slow down thevariation of relationship between the height of measuringflowiCNMH(owever, the flowingpoints and the time to reach the corresponding heightvelocform was fast due towith different process parameters was obtained.the high preheating temperature. On the other hand, partQI Le-hua, et alTrans. Nonferrous Met. Soc. China 20(2010) 980- -9869832.s (a)560K5502.0D1.51.5301.0C520E」0.5- B1■510-051015202530354045500551015Height of measuring point/mmInfltrationtime/s3.5| (b)EFig.6 Experimentally measured temperature of preform in3.0radial direction2.5\FThe average velocity of corresponding infilrationheights was calculated by the curves of measuringheights and the time to reach monitored points shown inFig.5. Due to the small interval distance of axialthermocouples, average velocity was used to substituteinstantancous infiltration velocity for analysis. Fig.7shows the ftting curve between height of measuring05 : 101520253035 40 45points and average infitration velocity under differentinfilration pressures. As shown in Fig.7, instable flowand large velocity fluctuation occurred under lowpressure and high temperature. The fluctuation ofvelocity reached its maximum when the pressure was 0.44DyMPa and the pouring temperature was 800 C. With theincrease of pressure and the decrease of temperature, the3}fluctuation of velocity became smaller and reached its”minimum under the pressure of 0.6 MPa and the pouring2-\Gtemperature of 760 C, which meant that the infiltrationbecame stable. The infiltration velocity became smallwith the decrease of pouring temperature under thepressure of 0.6 MPa. Through the above analysis, it isBconcluded that the infiltration became stable with the050方202503404increase of pressure and the decrease of temperature. Theinfiltration velocity became fast with the increase ofFig.5 Curves of measuring points heights and time for alloypouring temperature. Fig.8 shows the selected coremelt to reach monitored points at diferent conditions: (a) 0.4location (as indicated by point A) on the sample for SEMMPa, 800 C; (b) 0.5 MPa, 780 C; (c) 0.6 MPa, 760 Canalysis to investigate the infiltration quality withof the liquid alloy was solidified during rapid cooling bydifferent process parameters. As shown in Fig9, largethe fiber when the preheating temperature of infiltrationamount of micro-pores with a size of about 100 umprocess was lower. Thus, the flow channel of the liquidexisted in the composite under the pressure of 0.4 MPamagnesium alloy was reduced significantly or blockedand the pouring temperature of 800 C. Based on thecompletely, so that the required infiltration pressureanalysis of Fig.7, large velocity fluctuation occurredincreased or even completely prevented furtherunde中國(guó)煤化工nperature, and thisinfiltration of liquid magnesium alloy. As a result, theinvol: infiltration process,liquid magnesium reached the edge region first where thewhiclYHC NM H Gmicro-pores in thetemperature was high, and then arrived to the corecomposite. The infilration became stable with thelocation where the temperature was lower.increase of pressuire and the decrease of temperature,984QI Le-hua, et al/Trans. Nonferrous Met. Soc. China 20(2010) 980-9860'●一0.6MPa,760C一0.5 MPa, 780 C- 0.6 MPa, 780 C■一0.4MPa,800C25}0t5|0葉.16 1820222426283032Height of measuring poin/mmFig.7 Relationship between height of measuring points andinfiltration velocity" 100um |Fig.9 SEM images of selected sample position: (間) 0.4 MPa,10111213141516171819800 C; (b) 0.5 MPa, 780 C; (c) 0.6 MPa, 760 CFig.8 Selected position for SEM analysis in infilration sampletherefore, the number of micro-pores in the compositedecreased and the pores became smaller. Themicrostructure and fibre distribution also becameuniform. No significant porosity defects were found inFiber,the composite when the pressure was up to 0.6 MPa andpreformthe pouring temperature was 760 C.tX3.2 Formation mechanism of infltration front ofRegion of_infiltrationAl2O3 preform by liquid magnesium alloyToIn order to analyze the formation mechanism of theMolten_infiltration front, an infiltration model according to themetal111111111119practical situation was created which was based on theinfluence of temperature field during the infiltrationprocess of liquid magnesium alloy into Al2O3 preform.Fig.10 Schematic diagram of infiltration model of moltenFig.10 shows the principle of this model whichAZ91中國(guó)煤化工divides the infiltration process into three regions, that is,liquid metal region， infilration region and non-the p,HCNMHGinfiltration region. The liquid magnesium infiltrates intoThe Reynolds number of Inquid magnesium alloythe fibrous preform under pressure from the bottom ofinfiltrating in the Al2O3 preform was 0.456 on the basisQI Le-hua, et al/Trans. Nonferrous Met. Soc. China 20(2010) 980986985of calculation, which is less than 5. The flow of moltensimplifed and small resistance is neglected, such as pemetal in the preform is in accordance with theConsidering or<0.2, the changing in the height ofapplication scope of Darcy's Law. It is assumed that thethe liquid metal is ignored. On the assumption that ho+liquid metal is an incompressible homogeneous fluid.ppx=ho, the variation relationship of the kinetic viscosityThe dependence of the infilration velocity u, on thecofficient with the changing temperature is expressed inpressure gradient can be described as[19]:Eq.(6), including the influence of temperature field.u=kAP(1)|Ax-=p.+Butwhere k and are the permeability ofthe perform and the32μ(1-q4)+viscosity of the liquid metal, respectively; and 4p is thenr(6)pressure gradient.| B=-pgh +2rv cosθ-PAccording to the infilration process of moltenmetal shown in Fig. 10, Darcy's law can be expressed as:|u= 7o exp(kSP:R(2)μx(l-qp)In Eq.(5), R is gas constant, T is absolutewhere φr is the volume fraction of fibres. Here thetemperature, 7o and E are material constants. For“bundles of capillaries" model is used. Where Or is themagnesium alloy, n7o is 45 and E is 30 500 MPa[22- -23].when Pa is constant, x-0 and t0, the solution tovolume fraction of fibre. In this case, the permeability kEq(6) isand E P are expressed by [20]:x=[2(P。+B)p(7){=m2A32(3)In the present experiment, the volume fraction of(Ep:=ps - pg(h+x)+Pp+P.+P-Pvfiber preform is 10%, the effective capillary tube radiusof the fiber preform is 8 pum, and surface tension of thewhere p is the density, g is the aceleration of gravity, hmolten magnesium n=0.6 Nm. Introduce theseis the infiltration height at 0 s, x is the infiltration heightparameters into Eq.(7) and ignore the influence of P gho,atts, n and r are the porosity of the fibre preform and thethe relationship among infiltration pressure Pa, pouringeffective capillary tube radius of the fbre perform,temperature T and infiltration height x is obtained whenrespectively, and r is calculated 7.5 μm using the firstthe infiltration time is 2.5 s, as shown in Fig.11. In ordermodel in Ref.[21]. Ps is the external pressure applied toto verify the accuracy of the model, predicted values arethe liquid metal, Pp is the additional pressure caused bycompared with experimental values. The deviationviscous resistance, P =8yμux/r2 according to the Poiseuillebetween simulation value and experiment value of theLaw, Pc is the capillary pressure at the infilration front,infiltration height is about 9%. The analysis shows thatPe is the tip resistance, po= pu2/4, and Pv is the pressurethe predicted results of infiltration front by this modelof infiltration front.have a good agreement with the experimental results. InThe aillary pressure Pe is given by the fllowingaddition, it can be seen from Fig.11 that there is aYong-Kelvin equation:minimum pressure during the infiltration process, which2γw cosθb一0.5MPa, 780 C 0.037mm(4)e一0.6MPa, 760 C, 0.041 mmwhere n is the surface tension of the liquid metal and θ昌60~is the contact angle between the liquid metal and the40-prefom material.20-Considering h+x=ho+Pr x, urdx/d, and introducingit into Eq.(2), Eq.(5) is obtained:了456「32μux=(1-qr)+8μx ]dx_中國(guó)煤化Imtion pesuera+4() =Ln2yw cosθYHCNMHGPa - pg(hy+ Px)+- Pv(5)Fig,11 Coupled relationship among infltration pressure PwIn order to simplify the solution, the model ispouring temperature T and infiltration depth x986QI Le-bua, et al/Trans. Nonferrous Met. Soc. China 20(2010) 980 -986is the threshold infiltration pressure. The infiltration7] SCHULz R KAUFMANN H. Fabrication cAl-inserts for local reinforcement of autonotive components [C].process is unable to be carried out when the appliedISATA, 32nd ISATA Symposium Proceedings: Materials for Energypressure is less than the threshold infiltration pressure.-Efective Vehicles. UK, 19: 185-192.8] MATSUNAGA T, OGATA K, HATAYAMA T, SHINOZAKI K,4 ConclusionsYOSHIDA M. Erect of acoustie cavitation on ease of ifiltration ofmolten aluminum alloys into carbon fibre bundles using ulrasonic1) The infiltration becomes stable and theinfltration method [0. Composites, 2007, 38(1): 771-778.9] MORTENSEN A. MASUR L, CORNIE J, FLEMINGS M C.fluctuation of velocity becomes smaller with the increaseInfitration of fibrous preforms by a pure metal: Part 1. Theory [].of pressure and the decrease of pouring temperature. TheMeallurgical Transactions, 1991, 204(1): 2535- -2547.infiltration velocity becomes fast with the increase of[10] MICHAUDA v, MORTENSEN A. On measuring wtabilily inpouring temperature.inflration processing [J]. Scripta Materi, 2007, 56(10): 859-862.2) The number of micro-pores in the composite[11] TURNER D z, HJELMSTAD K D, LAFAVE J M. Three-decreases and the pores become smaller with the increasedimensional flow visualization experiment of an RTM injection for aGFRP cuff 'mold [小Composite Structures. 2006, 76(4): 352- 361.of pressure and the decrease of pouring temperature.[12] YANG W M, YOKOI H Visual analysis of the flow behavior of coreMicrostructure and fibre distribution become morematerial in a fork portion of plastic sandwich injction molding小uniform without significant porosity defects in thePolymer Testing, 2003, 22(); 37-43.composite when the pressure is up to 0.6 MPa and the[13] JIANG Kai-yu, YOKOI H. Experiment study on fling imbalancepouring temperature is 760 C.phenomenon of melt front during injcion molding process [0]Chinese Joumal of Mechanical Engineering, 2009, 45(2): 294-300.3) Considering the effect of temperature field, a(in Chinse)flow model of molten AZ91D into the short fiber[14] CHEN Jjing-bo, SHEN Chang-yu, YOKOI H. Experimental study onpreform was constructed to describe tformationflling imbalance in mulicavity ijection mold [U]. Chinese Jourmalmechanism of the infltration front. The deviationof Mechanical Engineering, 2007, 43(10): 170-174. (in Chinese)between simulation value and experiment value of the[15] HE De-ping, HE Si-yuan, GU Fei-fi, HUANG Ke. The infilrationinfiltration height is about 9%。process of A-alloy -melt's into porous medium and porous aluminumalloy [I]. Science io China (Series B: Chemistry), 2009, 39(1): 22-30.(in Chinese)References[16] LIU Bing, HE Si-yuan, HE De-ping. The two-dimension tube modelof Aally-me's infltration in porous medium []. Chinese Jounal[1] KAPTAY G The threshold pressure of infilration into fbrousof Materials Research. 2006, 20(4): 389- 393. (in Chinese)preforms normal to the fbres' axes [0 Composites Science and[17] LONG s, ZHANG z, FLOWER H M. Characerzation of liquidTechnology, 2008, 68(1); 228- -237.metal infiltation of a chopped fbre preform aided by extemnal2] QI Le-bua, zHOU Ji-ming, su Li-cheng, OUYANG Hai-bo, upressure-1. Visualization of the flow behavior of aluminium melt in aHe-jun. Fabrication of CwMg composies using exnusion directyfbre preform [n]. Acta Metal Mater, 1995, 4319);: 3489-3498.following vacuum infiltration-Part 2: Fomming process study [小][18] EARDLEY E S, FLOWER H M. Infiltration and solidifcation ofSolid State Pbenomena, 2008, 141/143: 91-96.commercial purity aluminium matix composies [U]. Materials and3] HU Lian-xi YANG Yi-wen, LUO Sboujing. xU Xin-ying. KineticsEngineering A, 2003, 359: 303-312.of infiltration of liquid alumninum into alumina fibrous prefom D[19] HU Liam-xi YANG Yi-wen, LUO Shou-jing XU Xin-ying.The Chinese Journal of Nofrous Metals, 1998, 8(81): 75-79. (inInvestigation on the kinetics of infiltation of liquid aluminium intChinese)an alumina fibrous preform [0] Jourmal of Materials Processing[4] QI Le-bua, CUI Peiling LI He-jun. Numerical simulation ofTechnology, 1999 9423): 227- 230.infiltration process in liquid extrusion for fabrication of Al2O/LY12[20] BEAR J. Dynamics of fuids in porous media [M] New York:composites (D]. The Chinese Joumal of Nonferrous Metals, 2000,Elsevier, 1972.10(1): 141-144. (in Chinese)[21] LUO Shou-jing. Liquid inflration extrusion of composites [M].5] OUYANG Hai-bo, L He-jun, QI Le-bua, L Zhengjia, FANG Ting,Bejing: Mallurgical Industry Press, 2002: 118-120. (in Chinese)WEI Jian-feng. Fabrication of short carbon fbre preforms coated[22] SMITHELLS C J. Smihells metals reference book [M]. 礎(chǔ)Ed.with pyrocatbo/siC for liquid metal ifilration 門Journal ofMaterials Science 2008, 43(13): 4618 -4624.[23] GARCIA-CORDOVILLAC E L, NARCISO J. Pressure ifitration6] QI Le-bua, suU Li-zheng, JIANG Chun-xiao, ZHOU Ji-mning, LIof packed ceramic particulates by liquid metals [0 Acta Matralia,He-jun. Research on the numerical simulation of1999, 47(18): 4461-4479.liquid- infiltration-extrusion process for composites based on the(Edited by FANG Jing-hua)rigid-viscopliastic FEM [D Materials Science and Engineering A,2007, 45455: 608 -613.中國(guó)煤化工MYHCNMHG