1、TECHNICAL PAPERDevelopment of an injection molding tool for complex microfluidicgeometriesEmil J.GeigerDieudonne A.MairFrank SvecAlbert P.PisanoReceived:2 June 2011/Accepted:15 July 2011/Published online:28 July 2011?Springer-Verlag 2011AbstractThis paper will track the design and results ofan injec
2、tion molding tool developed to manufacturemicrofluidic chips.The mold design and injection moldingprocess was complicated by the presence of integratedcapillary fluidic interconnects.We determined that designof the runner and gate system responsible for deliveringmolten plastic to the cavity had a s
3、ignificant impact on thequality of parts produced by the mold and the size of theprocess window.Numerical results confirm our findingsthat reducing gate lengths and increasing part thicknessdramatically improved the filling profile and loweredinjection pressures by 37%.Finally,the influence of gatel
4、ocation on part shrinkage is analyzed and discussed.1 IntroductionInjection molding is one of the least expensive manufac-turing technologies for the mass production of microfluidicdevices(Becker and Ga rtner 2008;Fiorini and Chiu 2005;Heckele and Schomburg 2004).This has been a drivingreason behind
5、 many of the publications that report usinginjection molding technology for the fabrication of a wide-range of microdevices,including microfluidic devices(Attia et al.2009;Chen et al.2010;Kim et al.2006;Leeet al.2005;Su et al.2004;Tosello et al.2010).However,these papers generally focus on the desig
6、n of the devicewith limited or no discussion of the design of the mold.One reason for this omission is that the molds represent asignificant cost,and therefore conducting experimentsrelated to mold design are expensive.Furthermore,mostcommercial injection molders design their molds fromextensive exp
7、erience or with expensive simulation soft-ware,both of which are beyond the reach of most academicresearchers.While there are resources available for con-ventional injection molding design(Rosato et al.2000),most microfluidic researchers are often left pursuing anexpensive and time consuming trial a
8、nd error approachwhen designing a microinjection mold for the first time.This paper will examine two revisions of an injectionmolding tool we developed.The resulting parts have beenthe basis for several publications over the past few years(Chen et al.2008;Geiger et al.2010;Mair et al.2007;Mairet al.
9、2006;Mair et al.2009;Stachowiak et al.2007).Themold design was complicated by the inclusion of integratedcapillary fluidic interconnects.While a thorough injectionmold design guide is beyond the scope of this paper,ourintent is to provide some insight into the injection moldingdesign process for the
10、 microfluidic community as theyapproach designing their own molds.2 Original injection mold toolThe first mold was designed to address the limitations of alegacy mold,shown in Fig.1,that existed before the workE.J.Geiger(&)Department of Mechanical Engineering,University of Nevada,Reno,NV 89557,USAe-
11、mail:ejgunr.eduD.A.MairExponent,149 Commonwealth Drive,Menlo Park,CA 94025,USAF.SvecMolecular Foundry,Lawrence Berkeley National Laboratory,Berkeley,CA 94720,USAA.P.PisanoDepartment of Mechanical Engineering,University of California,Berkeley,CA 94720,USA123Microsyst Technol(2011)17:15371540DOI 10.10
12、07/s00542-011-1323-xpresented here.Due to the high pressures needed to com-pletely fill the mold and the gate location,the parts pro-duced were thicker in the middle than the ends.Anyreduction in pressure led to incomplete filling of the moldrather than flatter parts.Hence,a primary goal of the init
13、ialdesign was to improve thickness uniformity along the partlength and width.To this end,the runners and gates weredesigned to deliver the polymer melt at multiple locations.The gate length was constrained by the geometry of themicromold insert,which was simply clamped between thetwo sides of the in
14、jection mold tool.Along with thisdesign,other features were added to the mold.Mostnotably,capillary fluidic interconnects were integrated withthe part.This allowed us to reduce the amount of postprocessing necessary for working devices(Mair et al.2006).Additionally,the mold was designed as a familym
15、old so that both halves of the chip were molded at thesame time,even though they had vastly different geome-tries and volumes.The main features of the mold aredetailed in Fig.2.While it was possible to successfully mold usablemicrofluidic devices with this initial design,the moldsuffered from partic
16、ular weaknesses.First,the parts con-tained high levels of internal residual stresses near thegates.These stresses would result in significant partshrinkage at the gates upon heating during the thermalbonding process.Essentially,the bonding process allowedthe highly stressed polymer to relax,deforming the parts.Second,while the parts were flatter and more consistent,the parts were surprisingly thicker at locations further fromthe sprue.Third,we observed shrinkage at the integratedports,which were
