A study of stacked PECVD silicon nitride films used for surface micromachined membranes

A study of stacked PECVD silicon nitride films used for surface micromachined membranes
of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A study of stacked PECVD silicon nitride  fi lms used for surfacemicromachined membranes M. Mikolaj ū nas a,d , R. Kaliasas a , M. Andrulevi č ius b , V. Grigali ū nas b , J. Baltru š aitis c , D. Vir ž onis a,d, ⁎ a  JSC   “  Minatech ”  , Savanori ų  271, LT50131 Kaunas, Lithuania b Institute of Physical Electronics, Savanori ų  271, LT50131 Kaunas, Lithuania c Department of Chemistry and Central Microscopy Research Facility, 85 Eckstein Medical Research Building, University of Iowa, Iowa City, Iowa 52242, USA d Kaunas University of Technology Panevezys Institute, Department of Electrical Engineering, Daukanto 12, LT-35212, Panev ėž   ys, Lithuania A B S T R A C TA R T I C L E I N F O  Article history: Received 23 October 2007Received in revised form 30 May 2008Accepted 20 June 2008Available online 25 June 2008 Keywords: Silicon nitrideStacked  fi lmsPECVDMembrane arraysFilm stress Silicon nitride stacked  fi lms containing three layers differing in mechanical – chemical properties aresynthesized using plasma enhanced chemical vapor deposition method from monosilane (SiH 4 ) andammonia (NH 3 ) mixture. The composition is analyzed using X-ray Photoelectron Spectroscopy and stress ismeasured using a substrate bending method. The ability to obtain stacked  fi lms with the custom tensilestress in the overall structure was demonstrated by the series of experiments. The tensile stress in the topand bottom  fi lms was obtained between 200 and 300 MPa whereas the stress in the middle  fi lm could beadjusted from compressive 60 MPa to tensile 300 MPa. Since the appropriate stress value is important inachieving required mechanical properties of the membranes, the results obtained are discussed in thecontext of surface micromachined membrane structures.© 2008 Elsevier B.V. All rights reserved. 1. Introduction Controlling stress in the thin  fi lm materials is essential in today'selectronics and micro electromechanical systems technology, since thestrainedgatetransistors[1 – 3]andvariousfreestandingmicrostructures[4 – 8] are becoming widely used in contemporary electronic devices. Inmany cases of the electronics device development material engineersare facing challenges in simultaneously optimizing the chemical,mechanical and electrical properties of the materials. Sacri fi cial releasemicromachining technique is a common tool for manufacturing of themembrane-shaped microstructures in the adjustable Fabry – Pérot opti-cal  fi lters [7], pressure transducers [9], capacitive silicon microphones [10], speakers and capacitive micromachined ultrasound transducers(cMUT) [11]. Usually these structures require high uniformity of themembranes and call for a very high etching selectivity ratio (due thesacri fi cial etching intricacies). For example, sacri fi cial release processused in cMUT fabrication [11] employs a low pressure chemical vapordeposition(LPCVD)onpolycrystallinesiliconasasacri fi cialmaterialandthe LPCVD on silicon nitride as a structural material. The sacri fi cialetchingisdonebypotassiumhydroxide(KOH)whichgivesanextremelygood etch selectivity of 400,000 to 1, as reported by the authors [11].However, there has been a report of the dif  fi culty in manufacturing themembraneslargerthan160 μ  mindiameterduetotheintrinsicstressof both(sacri fi cialand structural)materials[11]. Ergunet al.discussedthepossibility of using phosphosilicate glass interface to reduce the stressmismatchbetweenthesacri fi cialandmembranematerialthusimprov-ing the sacri fi cial release [11]. Various silicon nitride stress reductionmethods, such as using plasma enhanced chemical vapor deposition(PECVD) or inductively coupled plasma deposition [12 – 15], excesssiliconinsiliconnitride[15]ordepositiontemperaturereduction[14,15] tend to worsen the etching selectivity and mechanical and/or electricalpropertiesofresultingstructures.Dualfrequencyplasmaorhighdensityplasma CVD are well known methods of deposited silicon nitride andsiliconoxideintrinsicstressadjustment[4].However,theintrinsicstressisreducedatthepriceofdensi fi cationof the fi lm,andthisisnotalwaysfavorableintermsofelectromechanicalconversioneffectiveness:higherdensity of micro electromechanical membrane material may in fl uencethe performance, as the moving mass increases and changes dynamicsof the system [16,17]. Thermal treatment processes are capable toeffectively relax the intrinsic stress in silicon nitride thin  fi lms [18].However, highdeposition temperatures(ca.1000°C) are rarelyallowedduethediffusioneffects.Lowertemperature(ca.500°C)post-depositionthermal treatment is used to reduce PECVD silicon nitride compressivestress [19]. This process involves the mechanism of solid-state post- Thin Solid Films 516 (2008) 8788 – 8792 ⁎  Corresponding author. Kaunas University of Technology Panevezys Institute,Department of Electrical Engineering, Daukanto 12, LT-35212, Panev ėž ys, Lithuania.Tel.: +370 45 434247; fax: +370 45 516161. E-mail addresses: (M. Mikolaj ū nas), (R. Kaliasas), Andrulevi č ius), (V. Grigali ū nas), (J. Baltru š aitis), (D. Vir ž onis).0040-6090/$  –  see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.tsf.2008.06.063 Contents lists available at ScienceDirect Thin Solid Films  journal homepage:  depositiondensi fi cationofthe fi lmasdescribedinrecentstudies,wherean increase in tensile stress during the thermal treatment is demon-strated [3,20]. Even if the range of the treatment temperatures is con-siderably lower and acceptable, densi fi cation process can not bereverted towards the decrease of the tensile stress. Other post-pro-cessing methods as ion beam milling [20,21] are time consuming andthus avoidable . Thin  fi lms usually form stacks when a protective layer is present onthedevice.Thestackingofthe fi lmstendstoincreasetheoverallstressof thewholestructure[3,8].Stressadjustmentbydepositionofthestackof different thin  fi lms with appropriate stress was reported by severalauthors [9,22]. However, deeper analysis of the stack constituent pro- perties is required in order to predict the properties of resulting struc-tures. In this work we explored the possibility to gain the requiredproperties of the freestanding micro membranes by using the stacked(multilayer) thin  fi lms. These stacked  fi lms consist of a combination of thehightensilestresssiliconfree(dielectric)PECVDsiliconnitride fi lmsfor mechanical, electrical reliability and chemical passivity and the lowtensile - mid-compressive stress PECVD nitride  fi lms that reduce thehigh aggregate stress in the system. To understand the stressdevelopment mechanism the composition of the bare  fi lms wasmeasured. Finally, as an illustration of usability of stacked thin  fi lms inpractical micro membrane device manufacturing, the arrays of capacitive membrane elements with a custom aggregate stress weremanufactured with the surface micromachining technique. 2. Experimental methods  2.1. Deposition conditions The deposition of experimental  fi lms was performed in a custommadePECVDdeviceshowninFig.1.Thedetaileddescriptionofitcanbefound elsewhere [23]. Brie fl y, the gas mixture was supplied to thebottom electrode through 90 openings with the circular diameter of 2.5mm.Thesubstratewasplacedontopofthebottomelectrode,whichalso contained automatically controlled heater, capable of heating thesubstrate up to 500 °C. The output of the 13.56 MHz radio frequencygenerator wasconnectedtothewatercooledtop electrode.The plasmapowerwasmanuallyadjustedfrom 0 to 1000 W. Plasmapowerdensitywasadjustedbyvaryingthedistancebetweentheelectrodesfrom10to60mm.Threemass fl owcontrollers(AALBORG,modelGFC)maintainedthe  fl ow of the reactants and diluting gas with the within ±1.5% of therequired value to obtain ~1 atm pressure. The debit was kept constantat a selected value in 2 – 20 sccm/min range.In current experiments, single crystal silicon (100) and (111) waferswereusedassubstrates.Monosilane(SiH 4 )andammonia(NH 3 )mixture,diluted with nitrogen were used during the deposition process. Thedeposition temperature was varied from room temperature to 550 °C.The output of the 13.56 MHz radio frequency generator was set to theconstant60Wvalue.Stacked fi lmsweredepositedwithoutinterruptingthevacuum;apausebetweendepositionsofdifferentlayersrangedfrom5 min to several hours, when it was required to cool down the sample.  2.2. X-ray Photoelectron Spectroscopy All XPS experiments were performed on Kratos Axis Ultra spectro-meterequippedwiththedelayedlinedetector.Sampleswerescannedin the surface analysis chamber with a base pressure of 1×10 − 9 Torr.Aluminumanodewithpowerof225W,emissioncurrentof15mAandan accelerating voltage of 15 kV generating K α  radiation with photonenergy of 1486.6 eV was used in all experiments. A 500 mm Rowlandcircle silicon single crystal monochromator was used to obtain amonochromatic beam and an X-ray spot size of 700×300 µm. Lowenergyelectrons were used for charge compensation to neutralize thesample. Survey scans were collected using the following instrumentparameters:energyscanrangeof1200to − 5eV;passenergyof160eV;step size of 1 eV; dwell time of 200 ms High resolution spectra wereacquired in the region of interest using the following experimentalparameters: 20 to 40 eV energy window; pass energy of 20 eV; stepsize of 0.1 eV and dwell time of 1000 ms. The absolute energy scale of spectrometer was calibrated according to the Cu 2p 2/3  peak bindingenergy of 932.9 eV using an argon etched copper plate.All spectra were corrected for residual charge effects using theadventitiouscarbonC1speakat285.0eV.CasaXPSsoftwarewasusedtoprocess the XPS data [24]. Shirley-type background subtraction andtransmission corrected relative sensitivity factor (RSF) values from theKratos library were used for elemental quanti fi cation, as implementedinto CasaXPS. The components of the peaks contain a Gaussian/Loren-tzian product with 30% Lorentzian and 70% Gaussian character.  2.3. Scanning electron microscopy All SEM measurements were performed on JEOL JSM-IC25S scanningelectron microscope. Samples were prepared for SEM measurements bycoating with a thin gold layer and mounted onto a sample holder bymeansofconductivesilverpaste.Anacceleratingvoltageusedforimagingwas 25 kV. Samples were tilted with respect to the horizontal by 85°.  2.4. Stress measurements The stress determination based on a substrate bending is referredto as one of the classical approaches in many contemporary studies[7,25]. It is based on Stoney [26] formula and regarded to as  “ tensionmeasurements ”  in the srcinal paper: σ   f   ¼  E  s 6 R  1 − υ s ð Þ d 2 d  f  ;  ð 1 Þ where  E  s ,  υ s , and  d s  are Young's modulus, Poisson ratio and thickness of thesubstrate,respectively. R istheexperimentallydeterminedsubstrate'sradiusofcurvatureand d  f  isthethicknessofthe fi lm.Thisrelationshipcanbe regarded to as an average estimate of the thin  fi lm stress over therelatively large area, when the condition of   d s NN d  f   is met.Alternative methods of the stress measurement in freestandingthin  fi lm structures in some cases are used [7, 25, 27]. The digital SEM Fig. 1.  Schematic of the PECVD system. The system consists of the two electrodes andcan be heated up to 500 °C. The bottom electrode has 90 openings with the circulardiameterof 2.5mmforagas supply.Thesystem isconnectedtothe vacuum system andthe silane and ammonia tanks. See Experimental methods section for a more detaileddescription.8789 M. Mikolaj ū nas et al. / Thin Solid Films 516 (2008) 8788 – 8792  image correlation method, described by Sabate et al., currently is themost advanced method of nanoscale local stress measurement in thin fi lms [8]. Although being advantageous over the substrate bendingtechnique, above mentioned methods are still using it as a reference[8]. Additionally, they are include complicated measurement modelsand sophisticated (such as ion milling) fabrication techniques of thetest structures. Moreover, the substrate bending technique allows thestraightforward detection of the stress direction (compressive or ten-sile) without applying the additional diagnostics and measurements.As a result, we selected the substrate bending technique due to itsrelative simplicity and cost-effectiveness.Rectangular silicon (111) substrates of 30×4 mm size and 340  μ  mthickness were used for bending measurements. We assumed the tensilestress to be positive and the compressive stress negative. The substratebending in our case was an inverse of the radius of the curvature,  R , andwasmeasured bylaserinterferometerof theMichelsontype[28].Havingmeasuredthesubstratebendingvalue,1/ R ,weusedEq.(1)tocalculatethevalue of the stress. The in fl uence of the  fi lm thickness to the calculationresults was neglected since it was more than three orders of magnitudeless than the substrate thickness. Substrate thickness of all  fi lms in ourexperimentswasdeterminedusingacommerciallyavailableellipsometer.Only the substrates that satis fi ed the following requirements wereused in deposition experiments: (a) uniform initial bending along thebig substrate side, as measured with Michelson type interferometerprior to the deposition of the  fi lms (b) the curvature radius more than100 m for tensile stress measurement and (c) the curvature radiusmore than 200 m forcompressive stress measurement. Thestress wascalculated using Eq. (2), which is a modi fi ed Eq. (1): σ   f   ¼  E  s 6 1 − υ s ð Þ d s 2 d  f  1 R s −  1 R s þ  f    ;  ð 2 Þ where  R s  is the initial radius of curvature of the substrate and  R s +  f   — the radius of curvature of substrate with the  fi lm deposited on it. 3. Results and discussion  3.1. Composition of the  fi lms The fi rststepinourstudywastoexplorethesigni fi canceoftheexcesssilicontotheresidualstressinthesiliconnitride fi lm.Wevariedtheratioof silane to ammonia  fl ow ( R = Φ SiH4 / Φ NH3 ) from 0.1 to 1.2 during thePECVDdepositionprocess.Allotherparametersofthedepositionprocess,such as sample temperature and radio frequency generator powerdensity were kept constant. The chemical composition of the resulting fi lm was studied by the X-ray Photoelectron Spectroscopy (XPS). Fig. 2ashows a Si2p region XPS spectra obtained after PECVD deposition as afunction of increasing ratio of silane to ammonia  fl ow ( R = Φ SiH4 / Φ NH3 ).Three peaks, located at 103.1, 102.0 and 99.6 eV binding energy can beobservedinallexperiments.Peaksat103.1and102.0eVwereassignedtoSiO 2  bonded to some nitrogen atoms and Si 3 N 4 , respectively [29 – 31].Peak at 99.6 eV increased in intensity with the increase in silane toammonia  fl ow ratio while at the same time shifting in binding energyfrom99.6to100.2eV.Thesplittingbetweenthispeakand thatofSiO 2 is3.5eV,lowerthanthatforSiO 2 — metallicSi4.4±0.1eV [32].AsystematicshifttothehigherbindingenergiesindicatesthatthisSispeciesisaffectedbythesilanetoammonia fl owratio.Chainanietal.observedsimilarpeakat 99.0±0.1 eV and assigned it to SiO 1.6 N 0.3  species [29]. Additionally,Kaercheretal.inthecomprehensivestudyofSiN  x compoundsobservedapeak at 99.6 eV for a-Si and shift to higher binding energies with theincreasingamountofSi – Nbonds[33].AsindicatedbyKaercheretal.,theaverage chemical shift (to a higher binding energy side) per Si – N bondwas~0.62eV.Fromtheseliteraturevaluesweassignpeakat99.6eVtoametallic Si, having a small amount of bound nitrogen and/or oxygen.Additionally,inallXPSspectraofN1sregionasharppeakat397.9eV(notshown here) was indicative of a SiN  x SiN x  species [30]. Fig. 2.  (a) A Si2p region XPS spectra obtained after PECVD deposition as a function of increasing ratio of silane to ammonia  fl ow ( R = Φ SiH4 / Φ NH3 ). The Gaussian – Lorentziancomponent  fi t is shown for the  R  value of 0.7. (b) A calculated stress (according to theEq. (2)) of the bare  fi lms as a function of the chemical composition as obtained by XPSexperiments and the silane to ammonia  fl ow ratio,  R . Fig. 3.  A calculated stress (according to Eq. (2)) of the bare  fi lms as a function of thesubstrate temperature. The dependency is illustrated for two values of silane toammonia  fl ow ratio,  R , 0.7 and 1.0.8790  M. Mikolaj ū nas et al. / Thin Solid Films 516 (2008) 8788 – 8792  The relative amountofexcesssiliconwascalculatedasaratio Si2p Si /Si2p total , where Si2p Si  stands for the integrated intensity of the  fi ttedpeak of metallic Si at 99.6 eV and Si2p total  —  integrated intensity of theSi2p region. Fig. 2b shows a decrease in the calculated stress with theincrease in calculated relative amount of Si and the ratio of silane toammonia  fl ow. This unambiguously shows that stress is chemicalcomposition dependent and can be decreased by increasingthe ratio of silanetoammonia fl ow.ThepresenceofSiO 2 (approximatelyat103.1eV)was due to the residual contamination of the synthesis chamber withoxygen and surface aging effects when the excess silicon is present.Fig. 3 shows the positive relationship between the substratetemperature at the time of deposition and the  fi lm stress, when thevalue of   R  ratio is 0.7 and 1.0. Calculated stress in thin  fi lms increasesalmostlinearlywiththe increasingsubstrate temperature. Additionally,a decrease in the  R  ratio at a constant temperature corresponds to ahigher calculated stress values. A positive relationship between thesubstrate temperature during the  fi lm deposition and the  fi lm intrinsicstress also with possibility of generating negative values (compressivestress) agrees well with many publications about PECVD silicon nitrideavailable elsewhere.  3.2. Stress in stacked  fi lms Aggregate stress of a multilayer fi lmwith small total thicknesswascalculated as an arithmetic average weighted with respect toconstituent's thickness [9,34]. If   d  f  ≪ d s , the aggregate stress wouldbe calculated as follows: σ   f   ¼ ∑ i σ  i d i d  f  ;  ð 3 Þ where  σ   f   and  d  f   are the aggregate stress and total  fi lm thickness respect-ively; σ  I   and d i  are the stress and the thickness of the fi lm constituents.Two series of stacked  fi lms were deposited to test the effect of thestress compensation. Parameters of both triple layer stacks are shownin Table 1. The inner and outer layers were always deposited at thesame conditions, namely  R =0.2 and 480 °C, which resulted in dense,mechanically proof and highly insulating  fi lms. Measured aggregatestressofbothstackedcompositionsisshowninFig.4intriangularandsquare markers, while circular markers illustrate the stress in a bare fi lm with  R =1. Continuous lines represent the trends that werecalculatedusingEq.(3).Theerrorbarsshowthelevelofaninstrumen-tal con fi dence (uncertainty in astress measurement),which was eval-uated as ±20 MPa. The calculated stress for composition have a Fig. 4.  A calculated aggregate stress of the stacked  fi lms as a function of thecompensating layer deposition temperature. The dependency is illustrated for twocompositions A and B of the stacked  fi lms described in the Table 1. For comparison,a bare (non-stacked) silicon nitride  fi lm with  R =1.0 values are shown.  Table 1 Parameters of the stacked membranesComposition Constituent  fi lmsthickness,  μ  mGas ratio,  R , duringconstituent  fi lmsdeposition a Temperatureduring middle fi lm deposition, °CBottom Middle Top Bottom b Middle Top b A (0.3  μ  moverallthickness)0.1 0.1 0.1 0.2 0.2 0.2 50, 150, 275, 375and 475B (0.6  μ  moverallthickness)0.1 0.3 0.2 0.2 1.0 0.2 50 and 150Bare  fi lm(0.3  μ  m overallthickness)0.3 1.0 200, 275, 375 and475 a Gas ratio ( R ) is de fi ned as the ratio of SiH 4  to NH 3  ( R = Φ SiH4 / Φ NH3 ). See text forfurther details. b Deposited at 480 °C. Fig. 5.  Measured current as a function of applied voltage for bare 600-nm membrane( R =1.0) and two devices prepared with composition B membrane. Fig.6. SEMimagesofthestackedmembranesafterthereleaseexperiments.Themembranesin(a)arethoseofbare fi lminaTable1andpossessahightensilestress.Membranesin(b)areofcompositionBinTable1withalowtensilestress.Imagesshowthebreakingofmembranesin (a), but not in (b) after the release experiments. The scale bar in both images is 10 µm.8791 M. Mikolaj ū nas et al. / Thin Solid Films 516 (2008) 8788 – 8792  satisfactory fi ttothemeasureddata,thesameforcompositionBcanberegarded as having some deviation from the perfect  fi t.In order to test the insulating abilities of the membranes we haveperformed the voltage – current response measurements.Karl Suss PM5probe station was used in these measurements. Before the measure-ments a 100 nm thick Au  fi lm was deposited over the membranes andthen patterned to form a network of the top electrodes as describedelsewhere [4,5,11,28]. Fig. 5 shows the measured current (µA) as a functionofappliedvoltagein0to20Vrangeforbare600nmmembrane( R =1.0) and two devices fabricated withcomposition B membranes (B1and B2). The data obtained for a bare fi lm membrane are shown on theleft axis, whereas those of a composition B membrane (two separatedevices,B1and B2)are plottedontherightaxis.A sharpincreaseinthecurrentisregisteredforthebare fi lmmembranesat15V,whereasthereis no increase in the current for the composition B  fi lm membranes.Since our goal was only to evaluate overall conductivity of the devicestructures,wedidnotconsiderthemechanismofconductivityinvolved.The currents measured for devices B1 and B2  fl uctuated at the noiselevel of the current measurement. Therefore it can be concluded thatcomposition B membranes had good insulating abilities, at least in therange 0 to 20 V, when compared to those of the bare  fi lms.  3.3. Sacri  fi cial release experiments in membranes In addition to the stress measurements we also performed themembrane sacri fi cial release experiments. A thermally evaporatedcopper(Cu,300nmthickness)wasusedasasacri fi cialmaterialforthesurface release, etched by Cr 2 O 3 :H 2 O:H 2 SO 4  [4,5,11,28]. Sacri fi ciallayer was patterned by a contact optical (UV) lithography technique(de fi ning  n membranesof 30 μ  m inradius, 4 μ  mwideetchingbiasand10×10  μ  m area for etching holes). The structural layer, consisting of 0.6  μ  m thick bare PECVD deposited silicon nitride or stacked PECVDdeposited silicon nitride  fi lms was deposited. After structural layerdeposition the etching holes were opened with reactive ion etchingduring the next optical lithography step.Scanning electron microscopy images in Fig. 6 show two cases of membrane sacri fi cial release experiments. The image in Fig. 6a cor-responds to the casewhen densesilicon nitridewith high tensile stress(217 MPa) was used as a membrane material (a bare  fi lm in a Table 1).This caused membranes to break and bend up during the release. Theimage in Fig. 6b was taken after the release of stacked membranes(compositionBinTable1)witheffective tensilestressof60MPa.Inthiscase membranes preserved the integrity. This illustrates the usefulnessof the stacked  fi lms with adjustable mechanical properties in a manu-facturing of the micromachined membrane structures. 4. Conclusions Stackingthe fi lmsofdifferentmechanicalandchemicalpropertiesisadvantageous in making arrays of membranes with surface micro-machining technique. It allows obtaining structures with desirablephysical properties thus improving the quality and performance of thestructure. Here we demonstrated that the stacked silicon nitride  fi lmsdepositedusingPECVDmethodcontainanunboundedsiliconrichlayerwhichcompensatesfortheintrinsicstress.These fi lmsaremorereliableduring the surface micromachining than the bare PECVD silicon nitride fi lm. Moreover, greater reliability is achieved without the risk of degrading other properties of the structure, such as chemical inertia orelectrical resistivity.  Acknowledgements This research is supported by the Lithuanian State Science andStudy foundation. It is also part of the research under the grant fromEU structural funds, contract No BPD2004-ERPF-3.1.7-06-06/0029.Authors also thank the technical staff of the Institute of PhysicalElectronics at the Kaunas University of Technology for the assistanceand support and Dr. Kenneth Moore, Director of Central MicroscopyResearch Facility at the University of Iowa, for providing necessaryinstrument time. References [1] J.L. Hoyt, H.M. Nayfeh, S. Eguchi, I. Aberg, G. Xia, T. Drake, E.A. Fitzgerald, D.A.Antoniadis,TechnicalDigest — InternationalElectronDevicesMeeting,2002,p.23.[2] V. Zubkov, M. Balseanu, L.-Q. Xia, H. M'Saad, in: S. Wagner, V. Chu, J. Harry, A.Atwater, K. Yamamoto, H.-W. Zan (Eds.), Amorphous and Polycrystalline Thin-FilmSilicon Science and Technology —  2006, Warrendale, PA, U.S.A., Materials ResearchSociety Symposium Proceedings, vol. 910, 2007, p. 473, 2007.[3] M. Belyansky, N. Klymko, A. Madan, A. Mallikarjunan, Y. Li, A. Chakravarti, S.Deshpande,A.Domenicucci,S.Bedell,E.Adams,J.Cof  fi n,L.Tai,S.-P.Sun,J.Widodo,C.-W. Lai, in: P.R. Besser, A.J. McKerrow, F. Iacopi, C.P. Wong, J. Vlassak (Eds.),Materials, Technology and Reliability of Advanced Interconnects  —  2005, Warren-dale,PA,U.S.A,MaterialsResearchSociety SymposiumProceedings,vol.863,2005,2005, B7.9.1/O11.9.1.[4] E. Cianci, A. Schina, A. Minotti, S. Quaresima, V. Foglietti, Sens. Actuators, A 127(2006) 80.[5] A.Caronti,G.Caliano,R.Carotenuto,A.Savoia,M.Pappalardo, E.Cianci,V.Foglietti,Microelectron. J. 37 (2006) 770.[6] P. Muralt, N. Ledermann, J. Paborowski, A. Barzegar, S. Gentil, B. Belgacem, S.Petitgrand, A. Bosseboeuf, N. Setter, IEEE T. Ultrason. Ferr. 52 (2005) 2276.[7] H. Huang, K. Winchester, Y. Liu, X.Z. Hu, C.A. Musca, J.M. Dell, L. Faraone, J. Micromech. Microeng.15 (2005) 608.[8] N. Sabate, I. Gracia, J. Santander, L. Fonseca, E. Figueras, C. Cane, J.R. Morante, Sens.Actuators, A 125 (2006) 260.[9] D. Resnik, U. Aljancic, D. Vrtacnik, M. Mozek, S. Amon, Vacuum 80 (2005) 236.[10] D. Ozevin, S.P. Pessiki, D.W. Greve, I.J. Oppenheim, IEEE Sens. J. (2005) 4.[11] A.S. Ergun, H. Yongli, Z. Xuefeng, O. Oralkan, G.G. Yarahoglu, B.T. Khuri-Yakub, IEEET. Ultrason. Ferr. 52 (2005) 2242.[12] P.L. Ong, J. Wei, F.E.H. Tay, C. Iliescu, J. Phys.: Conf. Ser. 34 (2006) 1742.[13] L.d.S. Zambom, R.D. Mansano, Vacuum 71 (2003) 439.[14] H. Kim, Y. Lee, Y. Ra, G.P. Li, J. Yota, ECS Trans. 6 (2007) 531.[15] H. Huang, K.J. Winchester, A. Suvorova, B.R. Lawn, Y. Liu, X.Z. Hu, J.M. Dell, L.Faraone, Mater. Sci. Eng., A 435 – 436 (2006) 453.[16] K.K.M. Manoj, S. Goravar, V. Kommareddy, Y. Fan, J. Barshinger, M. Andrew, X.F.Wang, in: D.O. Thompson, D.E. Chimenti (Eds.), Review of Progress in QuantitativeNondestructive Evaluation, Vol. 26a and 26b, Portland, Oregon, U.S.A., July 30 – August 4, 2006, AIP Conference Proceedings 894 (2006) 870.[17] G. Vanagas,V. Sinkevicius, M. Mikolajunas, D. Virzonis, V. Grigaliunas, Proceedingsof the 6th International Conference on Vibroengineering, Kaunas, Lithuania,October 12 – 14 2006, p. 203.[18] P. Morin, E. Martinez, F. Wacquant, J.L. Regolini, in: T.E. Buchheit, A.M. Minor, R.Spolenak, K. Takashima (Eds.), Thin Films  —  Stressesand Mechanical PropertiesXI,Warrendale, PA, U.S.A., Materials Research Society Symposium Proceedings,vol. 875, 2005, p. 437, 2005.[19] E. Cianci, L. Visigalli, V. Foglietti, G. Caliano, M. Pappalardo, Microelectron. Eng.67 – 68 (2003) 602.[20] T.G. Bifano, H.T. Johnson, P. Bierden, R.K. Mali, J. Microelectromech. Sys.11 (2002)592.[21] R.G. Elliman, T.D.M. Weijers-Dall, M.G. Spooner, T.-H. Kim, A.R. Wilkinson, S. Huth,V. Tobias, Nucl. Instrum. Methods Phys. Res., Sect. B 257 (2007) 554.[22] J.M. Olson, Mater. Sci. Semicond. Process. 5 (2002) 51.[23] D. Vir ž onis, V. Sinkevi č ius, V. Grigali ū nas, T. Sigitas, K. Remigijus, Mater. Sci.(Medziagotyra) 12 (2006) 97.[24] N. Fairley, CasaXPS, 1999 – 2007[25] M.Martyniuk,J.Antoszewski,C.A.Musca,J.M.Dell,L.Faraone,Smart Mater.Struct.15 (2006) S29.[26] G.G. Stoney, Proc. R. Soc. A82 (1909) 172.[27] G.G. Yaralioglu, A.S. Ergun, B. Bayram, T. Marentis, B.T. Khuri-Yakub, UltrasonicsSymposium, 2001 IEEE, 2, October 7 – 10 2001, p. 953.[28] M. Mikolaj ū nas, D. Vir ž onis, V. Grigali ū nas, S. Tamulevi č ius, R. Kaliasas, Mater. Sci.(Medziagotyra) 13 (2007) 14.[29] A. Chainani, S.K. Nema, P. Kikani, P.I. John, J. Phys. D: Appl. Phys. 35 (2002) L44.[30] M.C. Poon, C.W. Kok, H. Wong, P.J. Chan, Thin Solid Films 462 – 463 (2004) 42.[31] N. Jehanathan, Y. Liu, B. Walmsley, J. Dell, M. Saunders, J. Appl. Phys. 100 (2006)123516/1.[32] G.E. McGuire, in: R.F. Bunshah (Ed.), Handbook of Deposition Technologies forFilms and Coatings, Noyes, Park Ridge, N. J, 1994, p. 763.[33] R. Kaercher, L. Ley, R.L. Johnson, Phys. Rev. B: Condens. Matter 30 (1984) 1896.[34] L.B. Freund, S. Suresh, Thin  fi lm materials: stress, defect formation, and surfaceevolution, Cambridge University Press, Cambridge, UK, 2004.8792  M. Mikolaj ū nas et al. / Thin Solid Films 516 (2008) 8788 – 8792
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks