光刻胶薄膜中溶涨性能检测方案(石英晶体天平)

智能文字提取功能测试中

Langmuir 2006， 22， 10009-1001510009 10010 Langmuir， Vol. 22， No. 24， 2006Rao et al. Effect of Deprotection Extent on Swelling and Dissolution Regimes ofThin Polymer Films .Ashwin Rao， Shuhui Kang， Bryan D. Vogt， Vivek M. Prabhu，* Eric K. Lin，Wen-Li Wu， and M. Muthukumar Polymers Division， National Institute of Standards and Technology， Gaithersburg， Maryland 20899， andDepartment ofPolymer Science and Engineering， University of Massachusetts，Amherst， Massachusetts 01003 Received June 20， 2006. In Final Form： September 11， 2006 The response of unentangled polymer thin films to aqueous hydroxide solutions is measured as a function ofincreasing weakly acidic methacrylic acid comonomer content produced by an in situ reaction-diffusion process.Quartz crystal microbalance with energy dissipation and Fourier transform infrared spectroscopy measurements areused to identify four regimes： (I) nonswelling， (II) quasiequilibrium swelling， (III) swelling coupled with partial filmdissolution， and (IV) film dissolution. These regimes result from chemical heterogeneity in local composition of thepolymer film. The acid-catalyzed deprotection of a hydrophobic group to the methacrylic acid tends to increase thehydrophilic domain size within the film. This nanoscale structure swells in aqueous base by ionization of the methacrylicacid groups. The swollen film stability， however， is determined by the hydrophobic matrix that can act as physicalcross-links to prevent dissolution of the polyelectrolyte chains. These observations challenge current models of photoresistfilm dissolution that do not include the effects of swelling and partial film dissolution on image quality. I. Introduction The microlithography industry has successfully developed newmaterials to image sub-45-nm features through systematic changesin the chemistry and composition of the polymeric photoresistmaterials. Resolving these structures depends on the change inpolymer solubility over a narrow range of chemical compositionthat is controlled through a chemical reaction during the imagingprocess. This change in film composition is produced by anacid-catalyzed deprotection reaction that converts hydrophobicgroups to weakly acidic groups. As the concentration of acidicgroups increases， the polymer becomes soluble in an aqueoushydroxide solution (developer). One key element in the designof new photoresist polymer materials is controlling theionizationof the polymer during the dissolution process. Numerous approaches have been taken to understand pho-toresist polymer dissolution through the application of percolationmodels，2-4 the investigation of interfacial reaction kinetics， thedevelopment of a surface-etching critical ionization model，.and the combination of reaction kinetics with a critical ionizationmodel.8，9 These studies focused on the kinetics ofthe dissolutionprocess and have highlighted the influence of the polymermolecular weight and developer pH， especially with regard to ( *Corresponding author. F ax： (301)975-3928.E-mail： v prabhu@nist.gov. TPresent address： Department o f Chemical Engineering， Arizona S tateUniversity， Tempe， AZ 85287. ) ( University of Massachusetts. ) ( (1) Ito， H. Adu . Polym. Sci. 2005， 172， 37- 2 45. ) ( (2) Oertel， H. K.； Weiss， M.； Dammel， R.； Theis， J. Mic r oelectron. Eng. 1990， 11(1 -4 )，267-270. ) ( ( 3 ) Yeh， T. F. ； Shih， H. Y . ；Reiser， A . Macromol . 1992 ， 25 (20) ， 534 5 -5352.(4) Yeh， T. F .； Reiser， A.； Dammel， R . R . ； Pawlowski， G.； Roeschert， H. M acromol. 1 993， 26 ( 15)，3862-3869. ) ( (5) Hunek， B .； C u ssler， E. AIChE J. 2003， 4 8， 661-672. ) ( (6) Tsiartas， P. C .； Flanagin， L . W.； Hinsberg， W.D.； Henderson，C. L .； Sanchez， ) ( I . C.； Bonnecaze， R. T.； Willson， C . G . Macromol. 1997， 30， 4656-4664. ) ( ( 7) Burns， S. D .； S chmid， G. M.； Tsiartas， P. C . ； Willson， C. G .； Flanagin， L. J. Vac. Sci. Techno l . B 2002， 2 0(2)，537-543. ) ( (8) Hinsberg， W.； Houle， F . A.； L ee， S. W.； Ito ， H.； Kanazawa， K. Macromol. 2005，38(5)， 1 882-1898. ) ( ( 9) Houle， F . A.； Hinsberg， W. D .； Sanchez， M . I. Macromol. 2 0 02， 35(2 2 )， 8591 - 8600. ) correlations to surface roughening. These models and experimentswere developed primarily for a previous generation of photoresistmaterials that did not swell during development but dissolvedin a manner similar to an etching process. However， manyionizable polymers can swell when exposed to an aqueoussolution. For these classes of photoresist polymers， the effectof swelling on the rate of dissolution and the influence of thepolymer chemistry on swelling remain poorly characterized；swelling may occur for polymers below the entanglementmolecular weight and in the absence ofchemical cross-links.10-13These swollen film can be stable over long times， suggesting thepresence of a restoring force that limits the swelling and inhibitsdissolution. In this case， the solvent permeates the entire filmand swelling is coupled to the chain ionization.8，12 To suppressswelling， additives and inert comonomers13 are often incorporatedinto these photoresist formulations. However， this can also leadto problems in other aspects of resist design， such as transparencyand sensitivity. In this work， we investigate the response of a thin film to anaqueous base solution as a function of the average compositionofthe polymer as it changes during an acid-catalyzed deprotectionreaction. The photoacid reaction-diffusion process used inmodern photolithography produces a chemically heterogeneousthin film.9，14-18 In this process， a photoacid is generated by lightexposure from an initial random distribution of photoacid (10) Ito， H.； Wallraff， G. M.； Fender， N.； Brock， P. J.； Hinsberg， W. D.；Mahorowala， A.； Larson， C. E.； Truong， H. D.； Breyta，G.； Allen， R. D. J. Vac.Sci. Technol.B 2001， 19 (6)，2678-2684. (11) Ito， H.； Hinsberg， W.D.； Rhodes， L.； Chang， C. Proceedings of the SPIE，Advances in Resist Technology and Processing XX 2003， 5039， 70-79. (12) Prabhu， V.M.； Vogt， B. D.； Wu， W. L.；Douglas， J. F.； Lin， E. K.； Satija，S. K.； Goldfarb， D. L.； Ito， H. Langmuir 2005，21(15)，6647-6651. (13) Ito， H.； Allen， R. D.； Opitz，J.；Wallow， T. I.； Troung， H. D.； Hofer， D. C.； Varansai， P. R.； Jordhamo， G. M.； Jayaraman， S.； Vicari， R. Proceedings ofthe SPIE， Aduances in Resist Technology and Processing XVII 2000， 3999，2-12. (14) Houle， F. A.； Hinsberg， W. D.； Sanchez， M. I. J. Vac. Sci. Technol. B2004，22(2)，747-757. (15) Houle， F. A.； Hinsberg， W. D.； Morrison， M.； Sanchez， M. I； Wallraff，G.； Larson， C.； Hoffnagle， J. J. Vac. Sci. Technol. B 2000， 18 (4)，1874-1885. (16) Hinsberg， W. D.； Houle， F. A.； Sanchez， M. I.； Wallraff， G. M. IBM J.Res.Dev. 2001，45(5)，667-682. generator (PAG) within the photoresist film. At elevatedtemperature the photoacid diffuses and may catalyze hundredsof deprotection reactions15，19 by cleaving an acid-labile hydro-phobic protecting group to yield a hydrophilic weakly acidicgroup. At short reaction times， the photoacid diffuses and reactsto a limited extent， leaving a distribution of nanometer-scalediffuse reaction domains20 composed of hydrophilic groups withinthe initial hydrophobic continuum. In the case of long reactiontimes， the hydrophilic content pervades the film and the weakpolyelectrolyte dissolves via reaction ionization-dissolutionkinetics.8 In contrast to the behavior of thin films of statisticalcopolymers， the materials studied here have a locally hetero-geneous composition due to the reaction-diffusion process. Quartz crystal microbalance with energy dissipation and FourierttSransform infrared spectroscopy measurements identify fourregimes of film response for these unentangled polymer thinfilms when equilibrated with aqueous hydroxide solutions. Asexpected， fully protected polymer films do not swell due to theirhydrophobic nature， and these films will dissolve in the aqueousdeveloper solution when the polymer is fully deprotected. Atintermediate reaction levels， the local morphology of thehydrophilic domain structure provides a physicochemical frame-work that permits film swelling and swelling with partial polymerdissolution. These measurements highlight factors that are crucialto comprehend the dissolution mechanism that controls the fidelityof imaged patterns. II. Experimental Section Certain commercial equipment and materials are identified in thispaper in order to specify adequately the experimental procedure. Inno case does such identification imply recommendations by theNational Institute of Standards and Technology nor does it implythat the material or equipment identified is necessarily the bestavailable for this purpose. A. Materials. Poly(methyladamantyl methacrylate)(DuPont) withmass-average relative molecular mass of Mr，w=8800 g mol- andpolydispersity Mr，w/Mr，n= 1.18 was dissolved in cyclohexanone(Aldrich， Reagent Grade) at a concentration of 5% by mass. Thephotoacid generator triphenylsulfonium perfluorobutanesulfonate(TPS-PFBS) was added at a loading of 2% by mass of polymer.The mixture was then heated to 55 °℃ for 30 min to ensure completedissolution of the polymer. Tetramethylammonium hydroxidesolutions were prepared by diluting a 25% (by mass) stock solution(Aldrich) with deionized water purified and filtered by a Milli-Qsystem (Millipore) with final resistivity of 18 MQ cm. B. Sample Preparation. AT-cut quartz crystals (5MHz， Q-Sense)coated with silicon oxide or chromium were treated with hexa-methyldisilazane to improve the adhesion of the polymer films tothe surface. Liquid hexamethyldisilazane was syringed upon theQCM crystals， heated for 90 s on a hot plate set at 150°C， and， aftercooling to room temperature， thoroughly rinsed with toluene (Aldrich，Reagent Grade). This procedure yielded hydrophobic surfaces havingstatic water contact angles close to 90°.The polymer solution wasthen spin coated onto the QCM substrates at 209 rad/s (2000 rpm)，followed by annealing at 130°C for 60 s in a convection oven. Theaverage film thickness was approximately 120 nm. The polymer-coated substrates were exposed to a broadband 248-nm filtereddeuterium lamp (Oriel， 350 mW) for 30 s to generate the photoacid.The supported films were then baked at 90 °C for different periods ( (17) Schmid， G. M . ； S t ewart， M . D.； S i ngh， V . K.； W i llson， C. G . J. V ac. Sci.Technol. B2002，20(1)，185-190. ) ( (18) H oule， F. A .； H insberg， W. D . ； S a nchez， M. I.； H o ffnagle， J. A. J . Vac. Sci. Technol. B 2002， 20 (3)， 924-931. ) ( (19) M ckean， D. R.； S chaedeli， U.； Macdonald， S . A. J. P olym. Sci. Part A 1989， 27(12)，3927-3935. ) ( (20) Jones， R .L.； Hu， T.J. ； Lin，E. K.； Wu， W.L.；Goldfarb，D.L.； Angelopoulos， M.； T rinque， B . C.； Schmid， G. M.； Stewart， M. D.；Will s on， C. G . J. Polym. S c i.Part B 2004， 4 2 (17)，3063-3069. ) Poly(MAdMA) Poly(MAdMA-co-MAA) Figure 1. Acid-catalyzed deprotection products starting fromhomopolymer polymer to the copolymer of poly(methyladamantylmethacrylate-co-methacrylic acid). The average copolymer com-position is determined quantitatively by reflectance infrared spec-troscopy. oftime to obtain variable degrees of deprotection and then quenchedto room temperature by placing the substrates on a large aluminumblock. The deprotection reaction is an acid-catalyzed， thermallyactivated process that results in the adamantyl protecting group beingremoved to form a poly(methyladamantyl methacrylate-co-meth-acrylic acid) copolymer as shown in Figure 1. C. FTIR Spectroscopy. The deprotection level (copolymercomposition) and the amount of residual methylene adamantane inthe films on the chromium-coated quartz crystals were measured bya Bruker Equinox 55 spectrometer (Billerica， MA) with a MCTdetector in reflection mode. The VeeMax reflection optics is employedand is aligned to provide a 76° angle incidence. The entire setup issealed into the sample compartment of the FTIR spectrometer usingPlexiglas with a 10 mm sampling hole of the VeeMax system opento the atmosphere. A resolution of 4 cm-l is used and 100 scans areaveraged to improve the signal-to-noise ratio. The quantification of deprotection reaction degree is based on thebending vibration mode of CH (1360 cm-l) in the MA group ofPMAdMA. This band completely disappears and leaves a flat baselineifall the protected MA groups are reacted. The advantage ofchoosingthis band to measure the deprotection level is that it not only givesan absolute value of deprotection reaction extent but also allows usto discriminate easily between the residual MA and the protectedMA group. The quantification of the amount of residual MA isbased on the stretching vibration of H-C(=C) (3065 cm-) in thefree MA molecule. This band is isolated from other H-C(-C)vibration (usually <3000 cm-l)， which makes it possible to quantifyit accurately. Since a pure sample of MA is not available， thecalibration from IR absorbance of MA to its molar quantity cannotbe done directly. Instead， we use an indirect extrapolation methodby assuming the MA is completely trapped if the deprotection levelis close to zero. A quantitative relationship was determined to preparefilms ofknown deprotection extents， based on the above UV exposuredose and bake temperature， such that the average level of methacrylicacid content (fMAA) as a function of reaction time (t) was fMAA=0.78(1- e-0.05r). D. Small-Angle Neutron Scattering： Polymer-Solvent In-teraction Parameter. The Flory-Huggins interaction parameter(x) betweenMAdMA segments and water was estimated from small-angle neutron scattering to be 8.7 in dimensionless units of kgT at298 K， where kg is the Boltzmann constant and T is temperature.Small-angle neutron scattering was performed on the NG 3 30 minstrument at the NIST Center for Neutron Research.The randomphase approximation equations21 for polymer in solvent were fit tosemidilute solutions of PMAdMA in deuterated tetrahydrofuran(THF) with statistical segment length and y as the only fit parameters.Using the relationship between x and the solubility parameterdifference between components， which is known for protonated THF，the solubility parameter for MAdMA was determined.22 Therefore，x for MadMA-water was calculated using the known value of thesolubility parameter for water.22 This methodology emphasizes that ( (21) Hammouda， B. Adu. P o lym. Sci. 1993， 1 0 6， 87-133. ) ( 0I(22) Polymer Handbook， 4 ed.； Wiley-Interscience： Ne w York ， 1999 ； Vol. 2， p 675. ) a large chemical mismatch exists between the PMAdMA segmentsand aqueous solutions. E. Quartz Crystal Microbalance Technique. The response ofthe thin films to varied concentrations of TMAH solutions wasmeasured by a quartz crystal microbalance(Q-Sense) with dissipationmode (QCM-D). The QCM-D setup follows the changes in theresonance frequency (~5 MHz) as well as the overtones (15，25， and35 MHz) of the polymer-coated quartz crystal. After mounting thequartz crystals onto the temperature-controlled QCM flow cell(maintained at 298 K)， the polymer films were equilibrated for 5min in deionized water. The liquid flow rate was adjusted to 1.025mL/min. During this equilibration stage， changes in the crystalresponse can be attributed to viscosity and density effects from thewater23 with negligible water absorption into the film. Afterequilibrating the polymer film in water， liquid flow was switchedto introduce aqueous TMAH solutions into the cell. Changes in theresonance frequency as well as energy dissipation of the polymer-coated quartz crystals upon exposure to TMAH solutions were thenmeasured as a function oftime. The shifts in the resonance frequencyas well as energy dissipation were calculated relative to the resonancefrequency and energy dissipation of the coated crystal when immersedin water. For the polymer films used in this study， we were unableto accurately measure the shifts in the overtone frequencies due tothe rapid kinetics and large dissipative losses in highly swollenfilms. Therefore， only changes in the fundamental frequency of thequartz and the energy dissipation at this frequency will bepresented；uncertainties are calculated as the estimated standard deviation fromthe mean. In the case where the limits are smaller than the plottedsymbols， the limits are left out for clarity. The QCM technique²4 monitors the change in the mass of anoscillating quartz crystal. In a vacuum， the change in mass (Am) ofa rigid film coated onto the surface of the quartz crystal causes ashift in resonance frequency (AF=Fo -F) that can be calculatedusing the Sauerbrey relationship where Fo is the fundamental resonance frequency ofthe quartz crystaland pq and hg are the density and thickness of the quartz crystal.However， if the coated film is viscoelastic in nature and is immersedin a liquid， the change in the resonance frequency is given by24 where ω=2nF，=yn/pw， h is the thickness of the viscoelasticfilm； uj is the elastic shear modulus of the viscoelastic film， and niand ?i are the shear viscosities of the liquid and viscoelastic film，respectively. The viscoelastic nature of the coated over layer alsoresults in a viscous loss (dissipation) within the film， which is givenby24 For such a film， determination of the film thickness requires thesimultaneous measurement of the shift in the resonance frequencyas well as the energy dissipation.25，26 One can then use eqs 2 and3 to determine the thickness as well as the viscoelastic propertiesof the film. However， since there are four independent variables (p，u，n， and h) involved in the above equations， an accurate estimationof the material parameters requires the measurement ofAF and AD ( (23)Kanazawa， K. K.； Gordon， J . G. Anal. Chim. Acta 1985 ， 175，99-105.(24) Voinova， M . V.； R odahl， M.； Jonson， M.； Kasemo， B. Phys. S c r. 1 9 99， 59 (5)， 391 - 396. ) ( (25) Kanazawa，K . K . Faraday Discus s . 1 9 97， (107)，77-90. ) ( (26) White， C. C.； Schrag， J . L.J. Chem. Ph y s. 1999，111(24) ，1 1192 - 11206. ) Figure 2. Change in QCM frequency (AF) after exposure to 0.065N tetramethylammonium hydroxide solution for different averagemethacrylic acid copolymer content (fMAA). for the fundamental resonance frequency， as well as the harmonicovertones of the fundamental frequency. II.Results Polymer thin films with different average methacrylic acid(fMAA) concentrations were prepared by the in situ acid-catalyzeddeprotection reaction，27 where the hydrophobic PMAdMA isconverted into a methacrylic acid containing copolymer， as shownin Figure 1. The photoacid reaction-diffusion kinetics dependson UV exposure dose and photoacid generator (PAG) loading，which define the acid concentration， as well as photoacidchemistry and size， which determine the acidity anddiffusivity.15，16，18，28-31 To avoid complications from these factors，a constant photoacid concentration is maintained by completephotolysis of a dilute PAG concentration by UV exposure. Asystematic variation of deprotection is produced by changing thereaction time. This approach results in thin films with amorphology in chemical composition induced by photoaciddeprotection in contrast with films prepared from statisticalcopolymers. The response of these films to water and aqueous hydroxidesolutions was characterized by the time dependence ofthe QCM-Dsignal. In Figure 2， the change in resonance frequency (AF)upon exposure to a 0.065 N TMAH solution is shown versusequilibration time for an average fMAA ranging from 0 to 0.78mole fraction. The corresponding change in energy dissipation(AD) by the polymer film is shown in Figure 3. Qualitatively，a drop in the resonance frequency may be interpreted as filmswelling， while an increase in the resonance frequency indicatesfilm dissolution. Similarly， an increase in the dissipation can beinterpreted as being representative of viscous losses due to theswollen film. ( (27) Kang，S . H.； Prabhu， V.M.； Vogt， B.D . ；Lin ， E. K.； W u ， W.1.； Turnquest， K. Polymer 2006， 47(18)，6293-6302. ) (28) Stewart，M. D.； Tran， H. V.； Schmid， G. M.； Stachowiak， T. B.； Becker，D. J.； Willson， C. G. J. Vac. Sci. Technol. B 2002，20(6)，2946-2952. (29) Ablaza， S. L.； Cameron， J.F.； Xu， G. Y.； Yueh， W. J. Vac. Sci. Technol.B 2000， 18 (5)， 2543-2550. (30) Stewart， M. D.； Somervell， M. H.； Tran， H. V.；Postnikov，S. V.； Willson，C. G. Proc. SPIE 2000， 3999，665-674. (31) Wallraff，G.； Hutchinson，J.；Hinsberg， W.； Houle， F.； Seidel， P.；Johnson，R.； Oldham， W. J. Vac. Sci. Technol. B 1994， 12(6)，3857-3862. Figure 3. Change in QCM energy dissipation (AD) after exposureto 0.065 N tetramethylammonium hydroxide solution for differentaverage methacrylic acid copolymer content (fMAA). The lowest fMAA investigated were0 and 0.31， in which nomeasurable mass uptake was observed upon exposure to purewater or TMAH. From Figure 2， once the fMAA is increased to0.41， swelling occurs as evidenced by the decrease in the quartzoscillation frequency and the kinetics ofswelling plateaus withinminutes. There is no further change in frequency for times aslong as 45 min. Hence， we refer to these stable films as beingin a quasiequilibrium state. The short-time behavior is shown tohighlight the time scale of the film response. As the average fMAAis increased to 0.60， the response to the hydroxide solution hasfaster kinetics as well as a larger frequency change that isindicative of increased film swelling. The same trend is observedfor fMAA = 0.70. However， we notice that， with enhanced filmswelling， there is a slight recovery in AF leading to a shallowminimum followed by a plateau. This effect is amplified as theaverage MAA content increases， as seen from the right-handside of Figure 2 forMAA=0.74，0.76， and0.77. The film responseis to first swell and then exhibit a form of relaxation leading toa positive change in AF， consistent with partial dissolution ofthe film. This preswelling followed by dissolution providesevidence for a clear transition in film behavior as a function ofthe average film composition and processing conditions. However，in these cases， the plateau response indicates that the polymerremains in a swollen quasiequilibrium condition. Note that forfMAA=0.74， 0.76， and 0.77， the dissipation of these films atquasiequilibrium is still significant， corresponding to a highlyswollen structure. An overall positive change in AF due tosignificant mass loss is observed only at the highest/MAA=0.78. The energy dissipation measured in the QCM measurementqualitatively tracks the mechanical response of the film. At thelowest fMAA = 0.31， where no swelling is observed， the energydissipation is near zero， indicating an elastic film where theSauerbrey expression can be used to quantify film thickness.8，32，33However， as the film swells， the energy dissipated increases，reaching a level of 265×10-6 for fMAA = 0.70. Therefore， thefilms with the largest change in frequency (most swollen) displaya measurable viscous response. As the fMAA increases， the AD ( (32) Lee， S. W.； Hinsberg， W. D.； Kanazawa， K. Anal. Chem.2 0 02，74 (1 ) ，125-131. ) ( ( 33) V ogt， B. D.； L in， E . K.； Wu， W . L.； White， C . C . J. Phys. Chem. B 2004， 108 (34)，12685-12690. ) Figure 4. Polyelectrolyte quasiequilibrium phase behavior deter-mined from the long-time QCM response in frequency (AF) anddissipation (AD). Four regimes are observed： (I) nonswelling films(0 0.7). Lines serve as a guide for the eye. exhibits a maximum with time， followed by a recovery and plateauto 600 × 10-6 and 500 ×10-6 for fMAA = 0.74 and 0.76，respectively， consistent with the changes in frequency. The filmsat quasiequilibrium behave more elasticlike at higher fMAA (0.77and 0.78) with a significant decrease in AD in comparison tofilms with only slightly less MAA content. These data illustratedin Figures 2 and 3 correspond to the film’s response to 0.065 NTMAH. Changing the base concentration influences the swell-ing-dissolution response of these films. Figure 4 illustrates thequasiequilibrium AFand AD for films swollen by 0.01 and 0.26N TMAH. There is no significant shift in the type of responsefor differentfMAA；only the magnitude of the swelling is influencedby the TMAH concentration. IV. Discussion The poly(methyladamantyl methacrylate) homopolymer ishydrophobic as judged by its static water contact angle of 78°，large chemical mismatch between MAdMA and water quantifiedby the Flory-Huggins interaction parameter (x=8.7)， and lackof water or hydroxide solution uptake. However， the acid-catalyzed deprotection reaction of this polymer into a copolymerof poly(methyladamantyl methacrylate-co-methacrylic acid)introduces hydrophilic character to the polymer， leading to alarge change in physicochemical properties of the polymer filmwhen it is immersed in aqueous alkaline solutions. The range ofswelling and dissolution behavior is controlled by the averagefilm composition and spatial distribution of hydrophilic andhydrophobic groups induced by the in situ reaction (Figure 1). A. Phase Behavior. The plateau in the QCM response withtime can be used to quantify the long-time quasiequilibriumbehavior of the swelling ofthe polymer thin film. The plateauvalues in AF and AD are plotted separately in Figure 4 for twoadditional TMAH concentrations， 0.01 and 0.26N. For clarity，the 0.065 N data are not included. From inspection ofthe QCM-Ddata and the form of the kinetic curves， we observe four distinctregimes of behavior as a function of average film composition.In regime I there is no film swelling between compositions (0 0.7)， nearlycomplete dissolution is observed. The different responses of the films to the aqueous solutionare a function of the copolymer composition and the distributionof the methacrylic acid within the polymer films. Thesemethacrylic acid groups are generated through an in situ reactioninvolving the cleavage of adamantyl side groups. Consequently，the methacrylic acid distribution within the films will be dependenton the amount of photoacid generated within the film as well asthe ability of the photoacid to diffuse and react. At the UVexposure conditions employed， all the photoacid generatormolecules present in the film are activated for subsequent reactionand diffusion. At short reaction bake times， the photoacid diffusesand reacts to a limited extent， forming domains of reactedphotoresist that do not overlap.20 Therefore， under these conditionsthe polymer film may be regarded as a hydrophobic continuumwith a distribution of reacted domains composed of hydrophilicmethacrylic acid groups. Since the deprotection domains mayspan several different chains， the average film composition doesnot correspond to the composition of any particular chain. In thelow-reaction-extent regimes， we expect the domains to benonoverlapping， in concert with earlier investigations. At thisearly stage， regime I， the hydrophobic continuum of protectedpolymer prevents penetration of water. The methacrylic acidcontent is inaccessible to the solution. As observed in Figure 4，no swelling is observed at fMAA ≈ 0.3 and is independent ofwater and base concentration， supporting this concept ofinaccessible hydrophilic regions for the solution molecules. With longer reaction time， diffusion and reaction of thephotoacid molecules increase the methacrylic acid content， whicheventually forms a connected or percolated network. Thepercolated character ofthese groups provides pathways for solutetransport throughout the film， enabling the hydroxide ions toinduce film swelling by titrating the methacrylic acid groups.34-36The onset of film swelling in our system corresponds to an averagemethacrylic acid fraction greater than fMAA =0.3， a value closeto that predicted for the percolation threshold on a simple cubiclattice (pe = 0.3116). Above this percolation threshold inmethacrylic acid content， the rate of swelling increases， as shownin Figure 2， as does the magnitude of the frequency shiftcorresponding to uptake of the solution. The film swelling isenabled by the percolated distribution of deprotection domains，while the degree of swelling is controlled by the ionizable groupcontent. The swollen nature of the copolymer film is consistent withthe energy dissipation in regime II and the large changes in thevalues of AF. These data raise an interesting question about theorigin of the stability of the swollen films. For most polymergels， film stability during swelling is due to the existence ofcross-link junctions formed covalently or by physical entangle-ment of high molecular weight chains.7 However， the systemhere does not have any chemical cross-linking and the molecularweight of the polymers is below the entanglement molecularweight. The stability of our swollen films is attributed tophysicochemical junctions formed by the continuum of hydro- ( (34) Kumacheva，E.；Rharbi，Y.； Wi n nik，M. A.； Guo，L.； T a m， K. C.； Jenkins，R. D. Langmuir 1997， 13 ( 2)， 182-186. ) ( ( 35) Annable， T.； Buscall， R.； Ettelaie， R.； W h ittlestone， D. J. Rh e ol. 199 3 ， 37(4)， 695-726. ) ( (36) Petit， F.； Lliopoulos， I ； A udebert， R . J. C h im. Ph y s. Phys.-Chim. Biol. 1996， 93(5) ， 887-898. ) ( (37) F lory，P . P h ase Equ i libria. In Principles of Polymer Che m istry； Corn e llUniversity P ress： I thaca， NY， 19 53； p 5 7 7. ) Figure 5. FTIR measurements ofthe change in the film compositionand mass before and after film exposure to 0.26 N TMAH. In thenonswelling and swelling-only regime， no mass loss and remarkablyno change in composition are measured， as predicted by the QCMresponse. In regimes III and IV， a larger hydrophobic content isdirectly measured due to partial dissolution of MAA-rich poly-electrolyte chains through the percolative network. phobic (methyladamantyl) groups. The formation of thesejunctions is analogous to the aggregation of hydrophobic/hydrophilic copolymers in aqueous solution.38 The ability ofthefilm to sustain the swelling implies that the concentration ordistribution of the hydrophobic domains also is above thepercolation threshold. Therefore， the generality of these resultsshould depend on the relative hydrophobicity of the protectedto deprotected functional groups. Further， the observations of swelling will reduce the polymerconcentration near the solid substrate. One consequence is thatthe film may partially detach or delaminate during the measure-ment. This problem is solved by the hexamethyldisilazane(adhesion promoter) surface treatment. However， if partialdetachment were to occur， water would accumulate at thesubstrate， thus lowering the average viscosity of the thin filmoverlayer， leading to an increase in AD from the initial elasticfilm in contact with water. Simultaneously， a positive AF shiftwould result， due to less film mass coupled with the QCMoscillator. As the kinetics of detachment progresses， the dissipationshift would go through a maximum and return to the initial ADdue to only water coupled to the QCM oscillator， while amonotonic positive shift in AF would be observed. The presentexperiments do not show these signatures of physical detachmentnor delamination， as assured by the AD and AF responses.Moreover， independent FTIR measurements quantifying massloss confirm these results. Following on to regime II， the drop in plateau values ofAFin Figure 4 as the methacrylic acid fraction exceeds 0.5 is dueto the partial dissolution of the polyelectrolyte chains from theswollen layer. This observation was confirmed independentlyby FTIR measurements. In Figure 5， the final MAA composition(left-hand axis) and the total mass loss (right-hand axis) of thedeveloped film are plotted versus the initial composition forfilms dissolved in 0.26NTMAH. In regimes I and II， as expected，there is no change in methacrylic acid content， nor any mass lossas determined by FTIR. After quantifying the change in averagefilm composition， swelling in regime II expels only a fractionof the chains composed of high levels of MAA， leading to areduced average MAA content. Figure 6. Schematic of the deprotection morphologies of the four regimes and relationship to the average fMAA (and fMADMA) and percolationthreshold (pc). This behavior shows that the change in solubility ofthe polymerfilms occurs gradually and dissolution proceeds via the expulsionof highly charged polymer chains from a swollen matrix. Thenumber of chains expelled from the matrix increases withincreasing methacrylic acid content. These observations can beinterpreted with respect to the spatial heterogeneity in the localcopolymer composition induced by the reaction-diffusionprocess as follows： as the average MAA fraction ofthe copolymerfilm increases， chains with excess methacrylic acid contents arerendered soluble and escape the swollen matrix (regime III inFigure 4) on a time scale associated with the relaxation of thematrix or disengagement ofthe chain from the film. As the averagemethacrylic acid content increases， the relative fraction ofsolublechains increases， at the expense of the hydrophobic matrix.Therefore， the film dissolves faster than it swells， whichcorresponds to regime IV in Figure 4. This phenomenon is seenin the kinetic data plotted in Figure 2. This hypothesis of polymerchains dissolving from a matrix (regime III) is confirmed byFTIR measurements of the change in the average methacrylicacid content as well as the change in the total film mass afterexposure to a 0.26 N TMAH solution (Figure 5). A schematicof the morphology expected from these regimes is summarizedby Figure 6. Beginning with a completely protected PMAdMAfilm， photoacids deprotect local regions of the polymer.Initially，these can be envisioned as isolated fuzzy blobs20(regime I). Asthe deprotection continues， the deprotection volumes begin tooverlap， forming a percolated network (regime II)， which allowsthe film to swell. In regime III， this overlap grows， but a networkof the hydrophobic PMAdMA remains， preventing completedissolution. At high deprotection (regime IV)， the hydrophobicdomains are no longer percolated， leading to nearly completedissolution. The hydrophilic domains and their percolation aregoverned by the average composition with respect to thepercolation threshold. Most experimental data involving pho-toresist polymer dissolution correspond to regime IV， where thedeprotection morphology was not considered. In regime IV， reflectance IR as well as QCM measurementsindicate that the polymer film does not fully dissolve uponexposure to TMAH solutions. From the IR measurements， itappears that the residual layer corresponds to approximately 10-15% of the original film mass and has an average methacrylicacid composition that is close to the percolation threshold forfilm swelling. The low values of AD for this residual layer alsoconfirm its relatively unswollen nature. The origin ofthis residuallayer is not clear and deserves further study. B. Regime I： Hydroxide Concentration Effect. Figure 4highlights the effect of TMAH concentration on the extent offilm swelling within regime II. The data in Figure 4 demonstratethat the swelling of the films is larger in the more dilute TMAH Figure 7. Change in (a) frequency (AF) and (b) energy dissipation(AD) for three average methacrylic acid composition， within thequasiequilibrium swelling regime， as a function of TMAH con-centration. solution. To obtain a better understanding of this phenomenon，we studied three different copolymer compositions with meth-acrylic acid fraction ranging from 0.4 to 0.6 as a function ofTMAH concentration. The changes in AF and AD of these filmsupon equilibration are plotted in parts a and b， respectively， ofFigure 7. In the case of copolymer films containing 40%methacrylic acid， AF and AD exhibit a weak dependence onTMAH concentration， indicating that the extent of film swellingat this composition does not depend strongly upon the hydroxide concentration. However， as the average methacrylic acid contentof the film increases to 52 and 60% MAA， the dependence ofAF and AD on hydroxide concentration becomes more pro-nounced. Most noticeable， AD decreases sharply with increasingTMAH concentration after increasing substantially from purewater. The in situ titration of the methacrylic acid groups (pKaof 7.2) by TMAH is the primary source for the film swelling.At 0.01 N TMAH， the polymer films are completely ionized.Increasing the TMAH concentration results in electrostaticrepulsions within the film and is followed by the increased osmoticpressure by the TMAt counterions， leading to a decrease inswelling. Since AD~ hu-lm-2， a drop in dissipation withincreasing TMAH concentration may be interpreted as an increasein shear viscosity (n)， an increase in shear modulus (u)，a decreasein film thickness， or a combination thereof， consistent with alower extent of film swelling. Increasing the TMAH concentration leads to a diffusion ofexcess TMAH into the polymer film， resulting in screening ofthe electrostatic repulsion between the ionized groups of thepolymer chains as well as reduced osmotic driving force forswelling. This would be expected by true cross-linked gelbehavior， as observed by Horkay et a. on model charged gelswhere an increased electrolyte concentration results in a decreasedswelling ratio. In the current case， excess TMAHwill contributeto the solution ionic strength. This reduced swelling extent willincrease in the local polymer concentration， thereby increasingthe shear viscosity and elastic storage modulus. This behavioris consistent with the decrease in AD such as those observed inFigure 7b. While the direct comparison to a quantitative theory is currentlylacking， one important element that controls the response of thepolymer films to an aqueous base solution is the effective physicalcross-link density from the hydrophobic matrix. This responsestrongly depends on the fractal or percolation structure ofweaklycharged groups that would render the chains soluble in the absenceof the hydrophobic domains. These charged domains undergoionization， causing film swelling， but also induce measurablechanges in the mechanical properties. Additional experimentsthat tune the degree of chemical mismatch (hydrophobicity) as well as introduce well-defined morphologies (deprotection) suchas that examined by colloidal templating40 are needed to quantifythe effects of different parameters on this important technologicalproblem. V. Conclusion The behavior of these thin polymer films upon exposure toaqueous base solution is controlled by an in situ preparedcopolymer structure. The onset of film swelling is consistentwith a percolation framework； when the weakly acidic methacrylicacid fraction exceeds 0.3， these films swell. The swollen filmsare stable， suggestive of a quasiequilibrium state， due to thebalance between the uptake of solvent into the film and thespatial extent of physical cross-links from the distribution of thehydrophobic domains in the polymer matrix. The effect ofhydroxide concentration did not change the onset to regime II(swelling only)， consistent with the need for a percolation pathwayof hydrophilic groups before any swelling can occur. However，within regime II， in the extent of film swelling was controlledby the hydroxide concentration， with greatest swelling observedat low TMAH concentration. Regime III exhibits partialdissolution of highly charged polyelectrolyte chains. The chainsthat do not yet have sufficient hydrophilicity are trapped by thehydrophobic copolymer structure， as determined by FTIR. Thisbehavior highlights the coupling of the diffusion ofhydroxideand solvent into the film， inducing ionization and providing asufficient thermodynamic driving force for the chains to escapethe interconnected structure. These results are different fromthose for previous generations of photoresists， where swellingwas not an important factor and surface ionization was the rate-limiting step for dissolution. The development of a predictivemodel requires a minimum of including the spatial extent ofchemical heterogeneity to recover these observations on unen-tangled weakly acidic polyelectrolyte films. Acknowledgment. This work was supported by SEMATECHunder Agreement #309841 OF. The authors acknowledge Dr.William Hinsberg (IBM Almaden Research Center) for helpfultechnical discussions， Dr. Curtis Meuse (Biochemical ScienceDivision，NIST) for assistance in the reflection IR measurements，and Drs. Jim Sounik and Michael Sheehan (DuPont ElectronicPolymers) for providing the polymers used in this study. ( LA061773P ) ( (40) Liu， L.； Li， P . S.； Asher， S . A. Nature (London) 1999， 397 (6715)，141- 144. ) la CCC： $ C American Chemical SocietyPublished on Web ( Noda， T.； Morishima， Y. Macromol. ，