Seite - 74 - in Photovoltaic Materials and Electronic Devices

2.7˘0.3 were determined from the Drude oscillator parameters of the ZnO + Ag interface. These values indicate that when compared to bulk Ag, the interface is less conductivedue to incorporation ofhigher resistivityundoped ZnOand potentially more disordered as suggested by the lower scattering time. Over this spectral range, ε for ZnO was initially fit using two CPPB oscillators, ε8, and a zero-broadened Sellmeieroscillator [49] representedby: εpEq“ ApE2n´E2q (6) where Aand En represent theamplitudeandresonanceenergy, respectively. the  Ag  surface  roughness  in  this  work  is  76.9  Å  as  compared  to  that  obtained  in  our  parametric  analysis  of  108  Å  [28].  Our  parametric  value  slightly  overestimates  the  prediction,  however  in  Dahal  et  al.  [28]  the  samples  with  similar  Ag  surface  roughness,  25–30  Å,  also  has  an  interface  thickness  of  75–110  Å  which  are  greater  than  the  linear  prediction.  Figure  2  shows  that  the  spectra  in  ε  obtained  for  the  ZnO  +  Ag  interface  is  optically  different  than  Ag  and  ZnO  alone  and  can  be  modeled  by  a  Lorentz  oscillator  and  a  Drude  oscillator  in  the  near  IR  to  near  UV  range  (0.734  to  5  eV)  with  ε   =  1.  The  ZnO  +  Ag  interface  exhibits  a  clear  localized  particle  plasmon  absorption  feature  which  can  be  modeled  using  a  Lorentz  oscillator  with  a  resonance  energy  at  2.83     0.01  eV  [27,48].  A  resistivity  of  3.7     0.5  ×  10−5   cm  and  a  scattering  time  of  2.7     0.3  were  determined  from  the  Drude  oscillator  param ters  of  the  ZnO  +  A   int rface.  T se  values  indicate  that  when  compared  to  bulk  Ag,  the  interface  is  less  conductive  due  to  incorporati n  of  higher  resistivity  undoped  ZnO  and  potentially  more  disordered  as  suggested  by  the  lower  scattering  time.  Over  this  spectral  range,  ε  for  ZnO  was  initially  fit  using  two  CPPB  oscillators,  ε ,  and  a  zero‐broadened  Sellmeier  oscillator  [49]  represented  by:  2 2( ) ( )   n AE E E  (6) where  A  and  En  represent  the  amplitude  and  resonance  energy,  respectively.    Figure  2.  Spectra  in  ε  (arrow  pointing  left  for  ε1  axis,  arrow  pointing  right  for  ε2  axis)  from  0.734  to    5.0  eV  for  the  108  ±  1  Å  thick  ZnO  +  Ag  interface  layer  parameterized  with  a  Lorentz  and  a  Drude  oscillator  with  parameters  listed  in  Table  3. Figure2. Spectrainε (arrowpointingleft forε1 axis,arrowpointingrightforε2 axis) from0.734to5.0eVfor the108˘1ÅthickZnO+Aginterface layerparameterized withaLorentzandaDrudeoscillatorwithparameters listed inTable3. 3.1.2. PhononModes inZnO The analysis was extended to the IR by fitting parameters defining ε for ZnO only and fixing those defining ε for Ag and the ZnO + Ag interface as well as the interface layer thickness. This analysis approach was chosen because free carrier absorptionrepresentedbytheDrudefeaturedominates the IRresponseofAgand the ZnO + Ag interface layers and is already established from near IR to UV spectral range analysis. A common parameterization of ε for the ZnO was applied for the datacollectedfromthetwoinstrumentswithspectral ranges from0.04 to0.734eV and 0.734 to 5.0 eV, respectively, although the bulk ZnO layer thickness was allowed to vary for the ellipsometric spectra collected from each respective instrument to account for measurement on different spots over the sample surface. A common surface roughness thickness between the two sets of measured spectra was obtained, as this effect will vary less with non-uniformity than the overall bulk layer thickness. Figure 3 shows ε for ZnO represented by a combination of CPPB oscillators for 74