SensorsandActuatorsA234(2015)23–33
ContentslistsavailableatScienceDirect
Sensors
and
Actuators
A:
Physical
jo u r n al hom e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / s n a
Development
of
a
piezoelectric-driven
miniature
pump
for
biomedical
applications
H.K.
Ma
a,∗,
R.H.
Chen
a,
Y.H.
Hsu
baDepartmentofMechanicalEngineering,NationalTaiwanUniversity,Taipei,Taiwan
bTheInstituteofAppliedMechanics,NationalTaiwanUniversity,Taiwan
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received5June2015
Receivedinrevisedform24July2015
Accepted10August2015
Availableonline22August2015
Keywords:
Miniaturepump
Piezoelectricactuator
Biomedicalfluidhandling
a
b
s
t
r
a
c
t
Miniaturepumpingsystemiswidelyusedinmechanicalandbio-medicalfields,andinthisregard, extensiveresearchesarebeingconductedinrelatedapplications.Thispaperreportsanovel piezoelectric-drivenminiaturepumpthatcanachievehighflowratethroughacombinationofpiezoelectric-actuator andpumpingchamberwithinternalribstructures.Themajorfeaturesoftheproposedminiaturepump areself-primingandhighflowrateatlowfrequencyrange.Flowratesatdifferentfrequencieswere mea-suredunderdifferentpiezoelectric-actuatorthicknessesandfluidviscosities.Inaddition,thecorrelation betweentheactuatordisplacementandthepumpingefficiencyatdifferentfrequencieswas investi-gatedanddiscussed.Highflowratesofupto196ml/minand141ml/minwereachievedforwaterand bloodmimickingfluid,respectively.Itwasachievedbyusingapiezoelectricactuatorwitha0.2mmthick piezoelectriclayeranda0.25mmthickbrassplate,andapumpingchamberwith0.05mmembedded flow-guidingribstructuresforpumpingefficiencyimprovement.
©2015ElsevierB.V.Allrightsreserved.
1. Introduction
Miniature pumps are widely used in many fields, including biology,chemistry,medicaltreatment,andcoolingsystems.In gen-eral,miniaturepumpscanbecategorizedintotwotypes,dynamic drivenand mechanical displacement.Dynamic pumping,which usually utilizes interactions of fluid with an electric or mag-neticfield toexertdrivingforcesonthefluidcontinuously, are typicallyrepresentedbyelectro-hydrodynamic[1]and magneto-hydrodynamic[2]pumps.Adisplacementpump,ontheotherhand, usesamovingmechanicalparttochangethevolumeofachamber todrivefluidwithinthechamber.Amongvariousdisplacement pumps, reciprocating micropumps are typically designed with compactsizesandarethussuitableforintegratingwithvarious fluidsystems.
Basedonactuationmechanisms,thereciprocatingmicropumps canbecategorizedasfollows—piezoelectric[3],electrostatic[4], pneumatic[5],electromagnetic[6],thermopneumatic[7],shape memoryalloy(SMA)[8]andultrasonic[9].Amongtheseactuation mechanisms,thepiezoelectric-actuatorshaveseveraladvantages, suchaslargeactuationforce,shortresponsetime,highreliability,
∗ Correspondingauthor.Fax:+886223632644.
E-mailaddress:[email protected](H.K.Ma).
simplestructure,smallsizeandlightweight,andhavebeenwidely developedandanalyzedinthefieldof micro-electro-mechanical-systems (MEMS)[10–13].Variousstudies havebeenconducted toinvestigatethecharacteristicsofpiezoelectric-actuatedMEMS micropumps.Forexample,KimandJones[14]usedalinearstrain assumptiontopredicttheoptimalrectangularactuator-to-plate thicknessratioatdifferentratiosofYoung’smodulus.Themoment wasoptimizedbyadual-layeredpiezoelectricactuator.Yoonand Washington[15]proposedamomentbalancemethodtoanalyzea beam-shapepiezoelectricactuator.LuoandYin[16]designedand fabricatedfourtypesofmicropumpswithdifferentbimorph piezo-electricactuators and checkvalves.TruongandNguyen [17,18] presentsa lamination technique by designing two micro check valvesfabricatedbyusinga100m-thickSU-8film.Aflowrate of1to1.6ml/minwasachieved.SayarandFarouk[19]developed apiezoelectricvalve-lessmicropumpandappliedittotransport water.Theeffectsofinlet-outletportanglesandtheoverallpump sizeontheflowratewereinvestigated.
For bio-fluidicsand coolingapplications,a high flow rateis highly desirable.Cantwell etal. [20] proposeda low-cost,high flowrateelectromagneticmicropump,achieving170ml/min.Ma etal.[21,22]interfacedone-sidedpiezoelectricmicropumpswith aheat-dissipationsystemandanoptimalflowrateof140ml/min wasachievedwitha pumpbackpressureof 6.4kPa. Despiteof thehighflowrateachievedbypriorresearches,sofartherehave
http://dx.doi.org/10.1016/j.sna.2015.08.003
Fig.1. Structureof(a)theminiaturepumpassembly(b)thecheckvalve.
beennocommercialized,highflow-ratemicropumpssuitablefor integratingwithbio-medicalorcoolingsystems.
Theaimof thepresentstudy istodevelop and characterize ahighflowrateminiaturepumpsuitableforcommercialusein bio-medicalandwarming/coolingapplications.Typical commer-ciallyavailablebloodtransferdeviceshaveflowratesrangingfrom 80ml/min to200ml/min. Forexample, BelmontBuddy®LiteTM
ACprovides80ml/minflowrate,3MTMRangerTMSystem(model
245) has a standard flow rate of 150ml/min, and GE Health-care enFlow systemoffers a standard flow rate is 200ml/min. Ourproposedminiaturepumpscanmeetthestandardflowrate for this kindof blood transferapplications. For the biomedical applications,acircularchamberstructurewaschosentoachieve easy assembly ofthe miniaturepumps. Bothwater and blood-mimicking fluidwere usedto investigatetheinfluence of fluid viscosityonflowrate.Tooptimizetheperformanceofthe minia-ture pump, a series of rib structures were designed based on flow-fieldsimulation withinthe miniature pumpchamber and theeffectsoftheribstructuresontheflowratewerealso inves-tigated.Threetypesofminiaturepumpswithdifferentchamber structuresandpiezoelectric-actuatorthicknessesweredesigned, fabricatedandmeasuredatdifferentfrequencies.Theminiature pump performance was evaluated and discussed based onthe measured flow rates under different pumping structures and settings.
2. Developmentofthepiezoelectric-drivenpump
2.1. Miniaturepumpstructureandflowmechanismdesign
Thebasicstructureofthecircularpiezoelectricminiaturepump is shown in Fig. 1(a). The external dimension of the pump is 50mm×50mm×12mm. Thecase of theminiature pump was madeby machining a Poly-methyl-methacrylate (PMMA)block withacomputernumericalcontrolled(CNC)millingmachine. Com-merciallyavailablepiezoelectriccoatedbrassdiskswereusedin thisstudy.Leadzirconatetitanate(PZT)wasusedtoserveasthe piezoelectriclayer.Twocheckvalvesweremadeoutofa0.09mm thicksiliconefilm,anditwascutoutbyusingaNd:YAGlaser cut-tingmachine.Theassemblywasdonebyplacingallpiecestogether layer-by-layerandthenclampedtogetherwith4screwsand sili-coneO-ringsbetweenlayers.Thespacebetweenthepiezoelectric actuatorandthelowercase formedtheminiaturepump cham-ber,withadiameterof34mmandachamberdepthd=0.45mm.A circularpiezoelectric-actuatorwithadiameterof35mmcovered thechambertoformatightseal.Thispiezoelectric-actuatordrove fluidsinandoutthechamberthroughcheckvalvesfabricatedon thelowercase.Tofurtheroptimizetheflowrate,a seriesofrib structuresweredesignedwithinthechamber.Fig.2shows differ-entribstructuredesignswithinthechamber.Thegapp,defined asthespacebetweenthepiezoelectric-actuatorandtherib struc-ture,wasonekeyfeatureoftheribstructuredesign.Twotypes
H.K.Maetal./SensorsandActuatorsA234(2015)23–33 25
Fig.2.Pumpingchambercross-sectionview(alongtheA–Blinefromthetopview)
(a)withoutribstructures(b)withribstructures.
ofribheights,0.05mmand0.25mm,werechosentoformtwo different gaps p=0.4mm and 0.2mm. For simplicity in con-structionand highrigidity,bridgetype valveswereadoptedas the design for the check valves in the piezoelectric miniature pump (Fig. 1(b)). The actuation of the check valves were cou-pledtotheoscillationofcircular piezoelectric-actuatorthrough thechamber pressure changecaused by the actuator displace-ment.
Fig.3showsthepumpingmechanismoftheminiaturepump. Thevoltage-driven oscillationof thepiezoelectric-actuator pro-duces a displacement. The displacement changes the pressure withinthechamberand thus causesthecheckvalvestomove. As the piezoelectric-actuator moves downward, the increased inner pressuremakestheinlet valvecloseand theoutletvalve open,leadingtoanoutflowtowardright.Astheactuatormoves upward,theinner pressuredecreasesandmakestheinletvalve openand theoutlet valve closed, leading toan inflow toward left.
Fig.4. Brass-basedpiezoelectricactuator:(a)schematicview(b)activatedbyAC
voltage.
2.2. Designofthepiezoelectric-actuator
Fig.4(a)shows theschematics of the piezoelectric-actuator, formedbybindingpiezoelectricmaterialwithabrassdiaphragm plate. When driving voltage was applied to the piezoelectric layer,thepiezoelectriclayerstartedoscillating,causingthebrass diaphragm to move. To ensure a vertical movement of the diaphragm,theradiusofthepiezoelectriclayershouldbesmaller thanthebrassdiaphragmradius.Thethicknessesofthe piezoelec-triclayerandthebrassdiaphragmaffectthedisplacementofthe entireactuator.InFig.4(a),thereexistsaneutralplanewithzero momentumandzerotheshearstress.Thelocationofthisneutral planecanbedeterminedbysolvingtheGoverningEquation.The followingequationcanbedeductedtorepresenttheprinciple: Fx=0=
xdA (1) −h −(h−hcu) cudA+ −(h−hcu) 0 pztdA+ 0 −(h+hcu)+hpzt pztdA=0 (2)From(2),thelocationofneutralplanecanbefound,asshown in(3) h=
1−2 pztECuh2cu +2EPzthCuhPzt+Epzth2pzt
1−2 cu 2(Ecuhcu) 1−2 pzt +Epzthpzt 1−2 cu (3) where Ecu and Epzt represent theYoung’ modulus of thebrass
diaphragmandthepiezoelectric(PZT)layer,respectively,cuand
pztarethePoissonratios,hcuisthethicknessofthebrassdisk,hpzt
isthepiezoelectriclayerthickness,andhisthedistancebetween theneutralsurfaceandthebottomofthepiezoelectriclayer.
ReferringtoFig.4(b),aperiodicallyvaryingsinusoidalelectric fieldisgeneratedwithinthepiezoelectriclayerviatheelectrical contactsonthepiezoelectriclayerandthebrassplate.Thevarying electricfieldcausesthepiezoelectricactuatortobendperiodically, andthereforepumpsthefluidflowthroughthesystemwiththe coordinationwiththebridgevalves.Aslongasthepumpingforce canovercomethefluidicresistance,inertiaandviscosity,the pump-ingflowcanbedirectedeffectively.Tounderstandtheperformance oftheminiaturepumpatdifferentpumpingfrequencies,the bend-ingvibrationofthepiezoelectricactuatorwastheoreticallyderived [23]asfollows.Forabendingplate,thelinearstraindistribution conditionsare
(Z)=pzt=cu=Z (4) cu= EcuZ 1−2 Cu (5) wherepztisthestrainofthePZT-actuator,cuisthestrainofthebrassplate,istheradiusofcurvatureofbendingplate,cu
referstothestressattheinterfacebetweenthepiezoelectriclayer andthebrassplate,Zisthedeflectiondisplacementoftheneutral surface.Fromthelinerpiezoelectricconstitutiveequations[21],the stressandstrainrelationshipforthePZT-actuatorcanbeexpressed as pzt= EpztZ 1−2 pzt
Z−Vd31 hpzt = EpztZ2 1−2 pzt − EpztVd31Z hpzt 1−2 pzt (6)whered31istheelectrical–mechanicalcouplingcoefficientinthe
z-directionandVistheappliedvoltage.Furthermore,themoment balancingequationsofthebendingstructuregives
M=−
−h hCu−h cuZcudz− hCu−h 0 pztZpztdz − 0 hCu−h+hpzt pztZpztdz (7) =− hCu−h −h Ecu 1−v2 cu z2 cudz − hcu−h+hpzt hcu−hEpztZpzt 1−
v
2 pzt − EpztV d31 hpzt 1−v
2 pzt Zpztdz=0 (8) Table1Pumpcomponentparameters.
Piezoelectricactuatorparameters Value PiezoelectriclayerYoung’smodulus,Epzt(GPa) 52 PiezoelectriclayerPoisson’sratio,pzt 0.31 Piezoelectriclayerd31(m/v) −2.10×10−10 Piezoelectriclayerdensity(kg/m3) 7.78 BrassplateYoung’smodulus,Ecu(GPa) 103 BrassplatePoisson’sratio,cu 0.33 Siliconcheckvalvedensity(kg/m3) 1800 SiliconcheckvalveYoung’smodulus,Esi(MPa) 25.5 SiliconcheckvalvePoisson’sratio,si 0.48
Eq.(7)and(8)canbefurtherdeducedto M=EpztVd31
hpzt+2hcu−2h 21−2 pzt =ECuIcu 31−2 cu + EpztIpzt 31−2 pzt
(9) Thepiezoelectric momentM canbe calculatedby usingthe parametersprovidedbyPZTvendors(listedinTable1)and sub-stitutingthemintoEq.(9).Table2liststheneutralplanepositionh andpiezoelectricmomentthicknessMunderdifferent piezoelec-triclayerthickness,calculatedfromEq.(5)through(9).Itcanbe seenthatasthethicknessofthepiezoelectriclayerdecreases,the positionoftheneutralplanehshiftstowardsthepiezoelectriclayer, thepiezoelectricmomentMdecreases,butthecurvatureradius increases.
2.3. Designofribstructureswithinthechamber
Inordertounderstandthecontributionofribstructures,finite elementsimulationwasperformed(withworkingfluid parame-tersshown in Table3)toestimate theflow resistanceandthe resultingflowfieldunderapredeterminedpressuredropapplied betweena lateral inlet and a lateraloutlet as shown in Fig.3. Athree-dimensional steady-statemodelof theminiaturepump chamber was developed with the “gap”, p, the vertical space betweentheribstructureandthechambertop,asthemain vari-able.ThesimulationmodelwasimplementedusingANSYSFluent. TheNavier-Stokesequationswereappliedasthegoverning equa-tions.SIMPLECalgorithmwasusedforpressure-velocitycoupling, andthesecond-orderupwindmethodwasusedfordiscretizing equation.Thegappchangedfrom0.45mmto0.1mm,in accor-dancewiththechangeinheightoftheribstructurefrom0mmto 0.35mm.Theassumptionsforthesimulationmodelarelistedas follows:
(1)Laminarflowforsinglephaseincompressibleliquid. (2)Three-dimensionalsteadyflow.
(3)Inletboundarycondition:constantpressure(100Pa). (4)Outletboundarycondition:constantpressure(0Pa). (5)Contributionfromgravityisignored.
(6)Non-slipperywall.
(7)Smoothtrianglemeshof20,994,227elements. (8)Iterationlimit:dimensionlessresidualsreaching10−5.
Table2
TheoreticalcalculationofPZTdeflectionpropertiesfordifferentpiezoelectricthickness.
Experimentalcondition Thickness(mm) NeutralPlane Moment Kappa
Piezoelectric 0.1 1.54×10−4 2.47×10−1 5.3×10−1
0.2 1.89×10−4 2.72×10−1 4.5×10−1
0.3 2.28×10−4 2.91×10−1 2.8×10−1
0.4 2.69×10−4 3.06×10−1 1.6×10−1
H.K.Maetal./SensorsandActuatorsA234(2015)23–33 27
Fig.5.Experimentalsetup.
Table3
Parametersfortheflowfieldsimulation.
Fluid Density(kg/m3) Viscosity(kg/m) Working
fluid
Water 1000 0.0014
Bloodmimicking 1050 0.004 Chambergap 0.1∼0.45mmwith0.05mmstep
3. Experimentalsetup
AsshowninFig.5,theproposedhighflowrateminiaturepump wasdrivenbyasinusoidalalternating-current(AC)signal gener-atedfromChroma61,501ACSourceandamplifiedto±70Vbythe PowermeterChroma66,201.Theinletandoutletofthetwopipes wereleveledtocreatea0Pabackpressureinoursystem.An MTI-2100fiber-opticmeasurementinstrumentwasusedtomeasure themaximumdisplacementofthepiezoelectricactuator.Theflow ratewasquantifiedbymeasuringtheaverageincreaseoftotal liq-uidweight(UWA-D).Theflowrateswererecordedtoanalyzethe pumpperformanceatdifferentinputfrequencies.The measure-mentwasperformedatthedrivingfrequencychangingfrom15Hz to110Hzata5Hzstep.Miniaturepumpswithdifferent param-eters,suchaspiezoelectricthickness,ribstructureandworking fluidwerefabricatedfortheseexperiments.Theresultshavebeen discussedinthefollowingsection.
4. Resultsanddiscussion
Thissectionreportsthemeasuredpropertiesoftheproposed miniaturepumpwithahighflowrate.Themeasurementresults weredividedintothreepartstoorganizethediscussion
4.1. Thecontributionofpiezoelectricactuatorperformanceto pumpefficiency
Theflowofthepiezoelectricminiaturepumpwasinitiatedby thevolumechangecausedbythedeflectionofthepiezoelectric actuator.Therefore,thedesign ofthepiezoelectric actuatorhas agreatinfluenceonitsbasicperformance.AccordingtoEqs.(3) and(9),itcanberealizedthat,whichrepresentsamplitudeof
Fig.6.Deflectionamplitudeofthepiezoelectricactuatorwithdifferent
piezoelec-triclayerthickness(drivingvoltage:±70V).
thepiezoelectricactuator,changeswithdifferentthicknessesof thepiezoelectriclayer,asshowninTable2.Tofurtherverifythe theoreticalcalculationshown inSection2,MTI-2100fiber-optic measurementwasdonetoobtainthemaximumdeflection ampli-tudesof the piezoelectric actuators withdifferentpiezoelectric layerthicknesses,withintheseriesoffabricatedmicorpumps,using waterasthepumpingfluid.Thethicknessesofpiezoelectric lay-ersinthefabricatedminiaturepumpswere0.2mm,0.3mm,and 0.4mm.Fig.6showsthefiber-opticmeasurementresults.Itcanbe foundthatthepiezoelectricactuatorwith0.2mmthick piezoelec-triclayerdisplayedthelargestdeflectionamplitudeover15–100Hz frequencyrangeinagreementwiththetheoreticalcalculation.The maximumactuatordeflectionamplitudeof0.19mmwasreached atalowvibratingfrequencyof25Hz.Fig.7furthershowsthewater flowratemeasurementresultsoftheseminiaturepumpsinthe fre-quencyrangeof15–100Hz.Theclosesimilaritybetweentheflow ratechangeandthedeflectionamplitudechangeinthe15–100Hz frequencyrangesuggestedthattheflowratewashighlyrelatedto
Fig.7.Flowratesofminiaturepumpswithdifferentpiezoelectriclayerthickness
(Drivingvoltage:±70V).
thedeflectionamplitudeoftheactuator.Theminiaturepumpwith 0.2mmpiezoelectriclayerindeedproducedahigherflowratethan theminiaturepumpswithotherpiezoelectriclayerthicknesses.
Theresonancefrequencyisonekeycharacteristicofthe pro-posedminiaturepump.AsshowninFig.6,fortheminiaturepump with0.2mmpiezoelectriclayer,thedeflectionamplitudeofthe piezoelectricactuatorshowedtwopeaksat25Hzand75Hz, sug-gestingthatthesetwofrequencieswerethebasefrequencyand thethirdoctavefrequency,respectively.Fig.6alsoshowsthatthe resonancefrequencyshiftedtohighervaluesasthethicknessof thepiezoelectriclayerincreased.Ontheotherhand,thepumping efficiencyoftheminiaturepumpwasnottotallydeterminedbythe deflectionamplitudeofthepiezoelectricactuator.Forthe minia-turepumpwith0.2mmpiezoelectriclayer,themeasuredflowrates inFig.7are161ml/minand160ml/minat25Hzand75Hz, respec-tively,whilethedeflectionamplitudesshowninFig.6are0.19mm and0.09mmat25Hzand75Hz,respectively,withwaterinthe pumpingchamber.
Toidentifythepeaksofhighflowratefoundintheexperimental result,weusedanAgilent4294APrecisionImpedanceanalyzerto measuretheresonantmodesofthepiezoelectricactuator.Because oftheexperimentalenvironment,thepipesinstalledattheinlet andoutletswereshortenedfrom20cm to5cm.Since the low-estmeasurablefrequencyofAgilent4294Awas40Hz,weuseit tostudythefrequencyresponseoftheminiaturepumpfoundat 75Hz.Fig.8(a)and(b)showsthemeasuredfrequencyresponsesof theclampedpiezoelectricdiskandafteritembeddedintothe piezo-electricminiaturepump,respectively.In Fig.8(a),itshows that edge-clampedpiezoelectricdiskhasitfirstresonancefrequency at1.5kHz.Afterassemblyandfilledwithwaterintothechamber, multipleadditionalresonantpeakswereidentifiedat65Hz,202Hz, 856Hz.Theoriginalfirstresonantpeakoftheclampedpiezoelectric diskmovedfrom1.5kHzto1.9kHzwithahigherdampingratio. Notethatthereisaresonantpeakfoundat65Hz,whichwasclose totheresonantpeakfoundintheexperimentalresultshownin Fig.8(a).AhigherresolutionscannedisalsoshowninFig.8(b).This experimentalresultdemonstratedthatadditionalresonantmodel couldbeintroducedbythedesignedpiezoelectricminiaturepump. Theyplayedanimportantroletogenerateahighflowrateinthe piezoelectricminiaturepump.Notethatthemeasuredfrequency (65Hz)waslowerthantheexperimentaldata(75Hz),itwasdue tothepipesconnectedtotheinletandoutletwasshortenedand causetheresonantfrequencyshifted.
Fig.8.Impedanceresponsesof(a)clampedpiezoelectricdisk(blackdashedline:
Impedance;reddashedline:phase)andembeddeditintotheminiaturepump(black
line:Impedance;redline:phase),and(b)theimpedanceresponseofthewhole
systemnearthefirstresonantpeak.
Fig.9.Frequencyresponseofminiaturepumpsunderthecaseofchambergap
p=0.45mmwithoutribstructurewithwater.(Forinterpretationofthereferences
tocolorinthetext,thereaderisreferredtothewebversionofthisarticle.)
Tofurtherverifythiseffect,asshowninFig.9,wecalculated thestrokevolumeperunit amplitudeofeach cyclefrom15Hz to100Hz(blueline).Itwasdonebydividingthestrokevolume (blackline) by themeasuredactuator vibratingamplitude (red line)andthecorrespondingfrequency.Thestrokevolumewas cal-culatedbydividing measuredvolumeflow ratewithrespectto operatingfrequency.Thestrokevolumeperunitamplitude repre-sentsthelevelofthephasedifferencebetweenthevibrationofthe
H.K.Maetal./SensorsandActuatorsA234(2015)23–33 29
Fig.10.Comparisonofflowfieldsimulationresults(topview)forwaterandblood-mimickingfluid(a)chamberwithoutribstructures(b)chamberwithribstructures.
actuatorandthevibrationofthecheckvalvesforeachpumping cycle.Ahighervaluemeanstheactuatorandthevalvesare vibrat-ingmore“in-phase”,andviceversa.Itcanbeclearlyobservedfrom Fig.9thatinthefrequencyrangesof15–30Hzand70–85Hz,the strokevolumeperunitvolumeishigher.Itsuggeststhatthe vibra-tionoftheactuatorandvalveshaslessphasedifference.Onthe contrary,betweenthetwo resonantfrequencies(30–70Hz),the strokevolumeperunitamplitudeisverylow,suggestingthatthe actuatorandthevalvesaremovingnearlyoutofphase.Itcauses anextremelylowpumpingefficiency.Further,arelativeflatstroke volumeperunitamplitudewasfoundatandaroundthetwohighest flowrateregions.Itimpliesthatthepiezoelectricminiaturepump mustgointoaresonantstateandahighflowratecanbeachieved atlowfrequency.
4.2. Finiteelementstudyontheeffectoftheribsstructure 4.2.1. Flowfieldsinchamberswithandwithoutribstructures
Duringtheminiaturepumpoperation,thefluidwasassumedto flowfromtheinlettotheoutlet.However,inacircularchamber withoutanyrestriction,longerflowpathsextendedtotherimof thechambermightbeformedandreducethepumpingefficiency. Thesimulationwasconductedintwosteps.First,flowinacircular chamberwithoutribstructurewassimulatedtostudythe“free” flowfieldwithinthechamber.Second,theflowfield was simu-latedwithribstructuresalignedwiththeflowpatternaddedto thechamber.Theresultsofthistwo-stepsimulationareshown inFig.10(a)and(b),respectively.Itcanbeclearlyseenthatthe ribstructurescaneffectivelyconfinetheflowintheshortestpath
Fig.11.Comparisonofflowfieldsimulationresults(cross-sectionE–F)forwaterandblood-mimickingfluidwithdifferentpumpingchamberstructure:(a)chamberwithout
ribstructures(b)chamberwithribstructures.
H.K.Maetal./SensorsandActuatorsA234(2015)23–33 31
Fig.13.Flowratesofminiaturepumpswithdifferentgapspaceswithinthechamber
forwater(drivingvoltage:±70V).
betweentheinletandtheoutlet.Fig.11furtherrevealsahighly concentratedflowbetweentheribstructuresinthecross-section viewofthechamberspacebetweentheinletandtheoutlet.
Tofurtherverifytheresultsofthesimulation,bubbleswere pumpedintothechamberandcontinuousphotosweretakento tracethe flow path of theworking fluid (water). As shown in Fig.12(b),bubbleswereeffectivelyconfinedbetweentherib struc-tures. Moreover,therandomlymoving bubblesin thechamber withoutribstructures(showninFig.12(a))hadhigher chances tostayinsidethechamber,makingtheminiaturepumpunstable. Therefore,thedesignoftheribstructuresmayalsoimprovethe bubbletolerabilityandthusthereliabilityoftheminiaturepump. 4.2.2. Effectsofgapheight
Sincetheribstructureswereproventohavepositiveeffectson flowrateenhancement,itisnecessarytodeterminetheinfluence oftheribheightontheflowefficiency.InFig.13,threeminiature pumpswithdifferentribstructureheightswerefabricatedandthe flowratesweremeasured.Thefabricationvariablewasthegapp (p=0.45mm,0.4mm,and0.2mminFig.13),thedistancebetween theloweredge ofthepiezoelectric actuatorandthe topofthe ribstructure,withthefixedchamberdepthd=0.45mm.Theflow ratecomparisoninFig.13suggeststhattheminiaturepumpwith p=0.4mmhadahigherflowratethanthepumpwithp=0.2mm andthepumpwithoutribstructures(p=0.45mm).
Thehighestflow ratewasobservedforp=0.4mmminiature pump(196ml/minat25Hzand194ml/minat75Hz).Therewas alarge21%yetconsistentincreasefromtheflowrateofthe minia-turepumpwithoutribs(161ml/minat25Hzand160ml/minat 75Hz).Ontheotherhand,thep=0.2mmminiaturepumpshowed amoderate10%flowrateincreasefromtheminiaturepump with-outribstructures,andtheflowrateincreasewassmallerthanthat oftheminiaturepumpwithp=0.4mm.Thisflowratedifference possiblyaroseduetothedifferencesintheribstructureheights. Thep=0.2mmminiaturepumphada0.25mmhighribstructure, whichwasmorethanhalfofthechamberdepthof0.45mmand probablycausedadditionalflowresistancewithinthechamber.
InFig.11,simulatedflowfieldatthecross-sectionalplane(E–F) shown intheinset isplotted for both pumpingchamber with-out(Fig.11(a))andwith(Fig.11(b))ribstructures.Differentfrom pumpingchamberwithoutribstructures,theflowpathsatedge pumpingchamberwithribstructureswerenarrowed,causingan increaseofflowresistance(60%increaseonregionsIIandIIIfor 0.05mmrib).Thishigherflowresistanceforcefluidtoflowinto and concentrate at the centralregion I. By taking the shortest pathinthecentergroove,theoverallflowratecanbeincreased
Fig.14.Simulationresultsof(a)volumeflowrate(b)flowresistancewithin
cham-berswithdifferentgapspaces.
significantly. Thiscanbefurtherverifiedby calculatingvolume flowratesfromthesimulatedresultsshowninFig.11(b).Fig.14(a) and(b)showthecalculatedresultsbasedonthefollowinggeneral equationofflowresistancebasedontheHagen-Poiseuillelaw[24]: Rhyd=
P
Q (10)
whereRhydistheflowresistance,Pisthepressuredropacross
theinletandoutlet,QisavolumeflowrateshowninFig.14(a).In Fig.14(a),wedemonstratedthatusing0.05mmhighribstructures (gapp=0.4mm)canimmediatelyincreasethevolumeflowrateto 50%more.
4.3. Effectsofdifferentworkingfluids
Inadditiontotheapplicationoftheminiaturepumpsin infu-sionjet,thisstudyintendedtoinvestigatethepotentialapplication oftheminiaturepumpsinbiomedicalinfusionbyusinga blood-mimickingfluidwithhigherviscositythanwater.Fig.15compares theflow ratesofminiaturepumpswithdifferentgapspacesfor bothwaterandtheblood-mimickingfluid.Theflowratesharply decreasedby28–38%forminiaturepumpswithdifferent cham-bergapstructuresastheworkingfluidwaschangedfromwater to theblood-mimicking fluid withhigher viscosity. The minia-turepumpwithp=0.4mmgapspace(0.05ribheight)exhibited
Fig.15.Flowratecomparisonbetweenwaterandblood-mimickingfluidfor
minia-turepumpswithdifferentgapspaceswithinthechamber(Drivingvoltage:±70V).
thehighestflowratesof141ml/minat25Hzand125ml/minat 75Hzfortheblood-mimickingfluid,consistentwiththe simula-tionresultshowninFig.14.Thissuggestedthattheribstructures alsohaveflowconcentratingeffectonblood-mimickingfluid. Com-paredwiththeflowratesof196ml/minat25Hzand194ml/minat 75Hzforwater,theincreasedfluidviscositycausedadecreasein theflowrate,via28%at25Hzand36%at75Hz.Thissignifiedthat anincreaseinviscosityreducesthepumpingefficiencyat75Hz. Therefore,forapplicationinbiomedicalinfusion,alowerpumping frequencyispreferred.
5. Conclusions
Thecurrentstudyhasproposedahighflowrate piezoelectric-actuatedminiaturepumpwithflow-concentratingribstructures deployedinsidethepumpingchamber.Aseriesofminiaturepumps madeofPMMAwithdifferentbrass-basedpiezoelectric-actuators andribstructureswerefabricatedandtestedforflowrates. Simula-tionswerealsoperformedtoverifywiththeexperimentalresults. Themajorconclusionsareasfollows:
1.Forthebrass-basedpiezoelectricactuator,thepiezoelectriclayer thicknessshowsstronginfluenceontheamplitudeofdeflection. Athinnerpiezoelectriclayercanresultinahigherdeflection.For fabricatedpiezoelectricactuators,the0.2mmthickpiezoelectric layergivesthelargestdeflectionamplitudeof0.19mmwhen pumpingwater.
2.Thefabricatedminiaturepumpsshowtworesonant frequen-ciesin therangeof 15–100Hz.Fortheminiaturepumpwith 0.2mmpiezoelectriclayer,thewaterflowratecanreachupto 161ml/minasthebasefrequencyof25Hz;ontheotherhand, underthethird-octavefrequencyof75Hz,thewaterflowrateis 160ml/min.
3.Ribstructures deployed within the pumping chamber show strongflowguidingeffectbetweentheinletandtheoutletinside thechamberandpositivelyincreasethewaterflowrateby20%. Increasingtheheightoftheribstructures,however,hasa nega-tiveeffectthewaterflowrate,possiblyduetotheincreasedflow resistanceinsidethechamber.
4.For biomedical infusion applications, blood-mimicking fluid with higher viscosity was tested. The characteristic flow responseoftheblood-mimickingfluidtothepiezoelectriclayer thicknessandtheribstructuresremainsimilartothatofwater. Thehighviscosityreducesthepumpingefficiency,itcausethe maximumflowratedecrease28%thanwater,butalot more
athigher frequency(∼75Hz). Itsuggested thata lower driv-ingfrequency(∼25Hz)workedbetterforbiomedicalinfusion applications.
Acknowledgment
ThisresearchwasfundedbytheMinistryofEconomicAffairs, R.O.C.(103-EC-17-A-19-S1-225).
AppendixA. Supplementarydata
Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion,athttp://dx.doi.org/10.1016/j.sna.2015.08.003.
References
[1]E.G.Kim,J.G.Oh,B.Choi,Astudyonthedevelopmentofacontinuous peristalticmicropumpusingmagneticfluids,Sens.Actuate.A-Phys.128 (2006)43–51.
[2]H.Moritz,M.Stubbe,J.Gimsa,Ac-field-inducedfluidpumpingin
microsystemswithasymmetrictemperaturegradients,Phys.Rev.Lett.79(2) (2009)1–10.
[3]L.C.Wu,C.J.Yang,C.Y.Chen,LowpowerconsumptionPZTactuated micropump,in:Proc.IEEEIMPACT,Taiwan,2006,pp.65–68.
[4]M.T.Mir,A.S.Ebrahim,Designandsimulationofanovelelectrostatic peristalticmicromachinedpumpfordrugdeliveryapplications,Sens.Actuate. A-Phys.117(2005)222–229.
[5]E.Meng,X.Q.Wang,H.Mak,Y.C.Tai,AChech-valvedsilicondiaphragmpump, in:Proc.13thIEEEMEMS,Japan,2000,pp.23–27.
[6]S.M.Lee,Y.D.Kuan,M.F.Sung,Designandfabricationofamagneticfluid micro-pumpforapplicationsindirectmethanolfuelcells,J.PowerSources 196(2011)7609–7615.
[7]O.C.Jeong,S.Konishi,FabricationanddrivetestofpneumaticPDMS micropump,Sens.Actuate.A-Phys.135(2007)849–856.
[8]D.Xu,L.Wang,G.F.Ding,Y.Zhou,A.B.Yu,B.H.Cai,Characteristicsand fabricationofNiTi/Simembranemicropump,Sens.Actuate.A-Phys.93(2001) 87–92.
[9]N.T.Nguyen,A.H.Meng,J.Black,R.M.White,Integratedflowsensorforin-situ measurementandcontrolofacousticstreaminginflexuralplatewavemicro pumps,Sens.Actuate.A-Phys.79(2000)115–121.
[10]C.Y.K.Chee,L.Tong,G.P.Steven,Areviewonthemodelingofpiezoelectric sensorsandactuatorsincorporatedinintelligentstructures,J.Intell.Mater. Syst.Struct.9(1998)3–19.
[11]E.Stemme,G.Stemme,Avalve-lessdiffuser/nozzlebasedfluidpump,Sens. Actuate.A-Phys.39(1993)159–167.
[12]D.L.DeVoe,A.P.Pisano,Modelingandoptimaldesignofpiezoelectric cantilevermicroactuators,J.Microelectromech.Syst.6(1997)266–270.
[13]C.J.Morris,F.K.Forster,Optimizationofacircularpiezoelectricbimorphfora micropumpdriver,J.Micromech.Microeng.10(2000)459–465.
[14]S.J.Kim,J.D.Jones,Optimaldesignofpiezoactuatorsforactivenoiseand vibrationcontrol,AIAAJ.29(1991)2047–2053.
[15]H.S.Yoon,G.Washington,Piezoceramicactuatedapertureantennae,Smart Mater.Struct.7(1998)537–542.
[16]Y.Luo,X.Yin,Z.Zhang,Vibrationperformancesofpolymericmicropump actuatedbyPbZrTiO3bimorph,MicroNanoLett.10(2013)559–562.
[17]T.Q.Truong,N.T.Nguyen,Apolymericpiezoelectricmicropumpbasedon laminationtechnology,J.Micromech.Microeng.14(2004)632–638.
[18]N.T.Nguyen,T.Q.Truong,Afullypolymericmicropumpwithpiezoelectric actuator,Sens.Actuate.B-Chem.97(2004)137–143.
[19]E.Sayar,B.Farouk,Dynamicanalysisofbulkacousticwavepiezoelectric micropumps:effectsofinlet-outletportanglesandoverallpumpsize,in: Proc.IEEEIMECE,USA,2013,pp.15–21.
[20]M.L.Cantwell,F.Amirouche,J.Citerin,Low-costhighperformancedisposable micropumpforfluidicdeliveryapplications,Sens.Actuate.A-Phys.168(2011) 187–194.
[21]H.K.Ma,B.R.Hou,C.Y.Lin,J.J.Gao,Theimprovedperformanceofone-side actuatingdiaphragmmicropumpforaliquidcoolingsystem,Int.Commun. HeatMassTransf.35(2008)957–966.
[22]H.K.Ma,B.R.Chen,C.Y.Lin,J.J.Gao,DevelopmentofanOAPCP-micropump liquidcoolingsysteminalaptop,Int.Commun.HeatMassTransf36(2009) 225–232.
[23]S.Li,S.Chen,AnalyticalanalysisofacircularPZTactuatorforvalveless micropumps,Sens.Actuate.A-Phys.104(2003)151–161.
[24]H.Bruus,Acoustofluidics1:governingequationsinmicrofluidicsmodes,Lab Chip11(2011)3742–3751.
H.K.Maetal./SensorsandActuatorsA234(2015)23–33 33
Biographies
H.K.MareceivedhisPhDdegreefromthedepartmentof
mechanicalengineering,UniversityofIllinoisatChicago,
in1985.HewasaresearchengineerforEnergyand
Envi-ronmentalResearchCo.(Irvine,USA)in1985–1987.He
thenjoinedthefacultyofthedepartmentofmechanical
engineering,NationalTaiwanUniversity,in1987.
Rong-HueiChenreceivedhisB.S.degreefromNanJeon
UniversityofScienceandTechnology,Taiwan,in2008;
andreceivedhisM.S.degreefromYuanZeUniversity,
Taiwan,in2010.HeiscurrentlyaPh.D.candidateinthe
DepartmentofMechanicalEngineering,NationalTaiwan
University,Taiwan.Heis focusedonthepiezoelectric
materialsfornovelactuators,particularlypiezoelectric
micropumps.
Dr.Yu-HsiangHsuiscurrentlyanAssistant professor
intheInstituteofAppliedMechanicsandthedirector
ofBiomechanicalMicrosystemsLaboratoryatNational
TaiwanUniversity.HereceivedhisB.S.degreein
Mechan-icalEngineeringandM.S.degreeinAppliedMechanics
from National Taiwan University in 2000 and 2002,
respectively.HethenreceivedhisM.S.degreeandPh.D.
degreeinBiomedicalEngineeringin2006and2010from
theUniversityofCaliforniaatIrvine.Hisresearch
inter-estsincludebiomedicaldevices,lab-on-a-chipsystems,
