Seite - 66 - in Photovoltaic Materials and Electronic Devices

industriallypreparedpanels. Thus, improvements inunderstandingandquantifying thestructural transitionofSi:Hfromamorphoustonanocrystalline,asobtainedfrom single spot in situ RTSE measurements, can be applied to develop more advanced opticalmodelsandmore thoroughlyanalyzemappingSEmeasurementscollected over largerareas. We have applied SE from 0.734 to 5.88 eV to extract layer thicknesses, interface composition, and optical response in the form of complex dielectric function spectra (ε= ε1 +iε2) forallAg,ZnO,anddopedandundopedPECVDSi:Hlayers foundin substraten-i-pPVdevices. ThesestudiesbeginwithcharacterizationofZnO/AgBRs and have been applied over the near IR to UV spectral range [27,28]. The purpose of BRstructuresis toincreasetheopticalpathlengthof lightwithintheabsorberlayerof thePVdevice,whereeachphotonabsorbedhasthepotentialtogenerateelectron-hole pairs and thus electrical current. Any light not absorbed in the first pass of light throughthe absorber layer is reflected orscatteredby theBR back into theabsorber layer. Thus PV absorber layers can be made thinner or from materials with low minoritycarrierdiffusion lengths. Dueto thesubstratedependenceofSi:Hgrowth, thesameZnO/Ag BRstructureswereusedto studythegrowthevolutionof doped andundopedSi:Hrequiredforuse inn-i-pa-Si:HPVdevices. RTSEusingaglobal Σσ-minimizationanalysisprocedurehasbeenusedtotrackthebehaviorofstructural transitionsinSi:Hdepositedinthen-i-pPVdevicestructure,asfunctionsofhydrogen to reactive gas flow ratio R = [H2]/[SiH4], to produce growth evolution diagrams for undoped, p-type, and n-type layers. GlobalΣσ-minimization analysis of RTSE involves using test structural parameters, most commonly a bulk layer thickness (db)andsurfaceroughness thickness (ds), tonumericallysolve for test ε [29]. These testvaluesof εare thenusedtofitotherellipsometricspectracollectedatdifferent times when the film is relatively homogeneous. The approach is iterated in order to obtain numerically inverted εyielding the lowest spectrally and time-averaged error,σ, over themultiple timemeasurementsselected. Thenumerically inverted ε that minimizesσare then used to determine structural parameter variations over the full setofRTSEdata,withmaterial transitions identifiedeither in thestructural parameters themselvesorbyincreases intheerror function. Virtual interfaceanalysis (VIA) [30,31] has similarly been applied to track the depth profile of nc-Si:H as well as the formation and stabilization of voids throughout intrinsic layer nc-Si:H growth. In VIA, the full sample stack is not analyzed. Instead, optical properties of a pseudo-substrate are generated from ellipsometric spectra collected earlier in the deposition by numerically inverting the measurement to obtain ε using a simplified model consisting of a semi-infinite pseudo-substrate and a surface roughness layer. The effective ε for the pseudo-substrate contains information of allunderlyingmaterial(s) in thesamplestackandis thenusedas thesemi-infinite substrate for analysis of subsequent data sets. In this sense, the time derivative of 66