2.626/2.

627  Fall 2011 –  Fundamentals  of  Photovoltaics  

Quiz  #1  –  Due  9:30am  Session  #10  

Quiz  Instructions  

• The  undergraduate  version  of  this  quiz  (2.627)  consists  of  four  (4)  multipart  
questions  for  a  point  total  of  80.  

• The  graduate  version  of  this  quiz  (2.626)  consists  of  five  (5)  mutlipart  
questions  for  a  point  total  of  100.  

• This  is  a  take-­‐home  examination.  The  quiz  will  be  handed  out  on  Friday  
October  7th,  and  is  due  completed  at  9:30am  on  Session  10.  

• This  is  an  individual  effort.  Each  student  is  expected  to  hand  in  her/his  own  
exam.  You  are  not  allowed  to  discuss  the  quiz  with  each  other.  

• All  resources  (written,  verbal,  and  otherwise)  are  at  your  disposal.  (The  only  
exception  is  that  you  are  not  to  talk,  discuss,  or  otherwise  share  information  
pertaining  to  this  quiz  with  other  people.)  

• Quizzes  may  be  submitted  online.  

• Quizzes  arriving  within  24  hours  after  9:30am  on  10/13,  either  in  person  or  
online,  will  incur  a  25%  late  penalty.  Quizzes  arriving  24  hours  after  9:30am  
10/13  will  not  be  accepted,  barring  extraordinary  circumstances.  

  1  
1.  Class  Project  [15  pts]  

Peruse   the  list  of  2011  class  projects  online  and  pick  three  that  capture  
your   interest.   Write   a   short   paragraph   for   each,   describing   why   your   background  
and   interests   are   ideally   matched   to   that   project.   Class   instructors   will   use   this  
information  to  form  project  teams,  so  be  sure  to  pick  wisely!  

  2  
2.  Better  Estimates  for  PV  Output  [25  pt.]  

In  this  course,  we  often  use  the  following  “peak  hours”  formula  to  calculate  the  daily  
energy  output  from  a  PV  array,  given  the  solar  insolation:  
Prated × S
Daily Energy Output =
ΦAM 1.5G
Prated = Rated Peak Power of PV Panel
⎡ kWh ⎤  
S = Insolation ⎢ 2 ⎥
⎣ m − day ⎦
⎡ W ⎤
ΦAM 1.5G = Radiative Power Flux at peak sun ⎢ 2 ⎥
⎣ m ⎦

This  formula  is  equivalent  to  saying  that  a  panel  will  receive  S/ΦAM1.5G  peak  hours  of  
sun   per   day.   In   other   words,   if   Boston   receives   4   kWh/m2/day,   our   formula   will  
€ assume   that   our   panels   will   see   4   hours   of   sun   that   has   an   intensity   of   1kW/m2.  
However,   this   overestimates   the   amount   of   power   output   from   our   PV   array  
because   the   efficiency   of   a   solar   cell   decreases   at   lower   illumination   levels.  
Therefore,   we   should   really   account   for   these   lower   efficiencies   to   get   a   better  
estimate  for  the  power  output.  One  possibility  is  to  model  the  radiative  power  flux  
as  a  sine  wave  instead  of  a  square  wave:  

   
To   obtain   our  more   accurate   power  output  estimate,  we  need  to  estimate  how  the  
efficiency  scales  with  illumination  intensity.  Recall  that  our  efficiency  scales  with  the  
product   of   the   Jsc   and   Voc   (we’ll   assume   that   FF   doesn’t   change   with   illumination  
intensity.)  

  3  
a) If   we   assume   that   our   Jsc   is   equal   to   our   illumination   current,   how   does   Jsc  
scale   with   our   radiative   power   flux,   Φ.   In   other   words,   is   our   Jsc(Φ)   function  
linear,   quadratic,   exponential,   logarithmic,   etc?   Please   explain   you   reasoning.  
[7  pt.]  
b) Now   that   we   have   an   estimate   for   Jsc,   we   can   use   our   equation   for   Voc   to  
estimate  how  it  scales  with  Φ:  
⎛ J (Φ) ⎞
VOC ∝ ln⎜⎜ sc
+1⎟⎟  
⎝ J sc (ΦAM 1.5G ) ⎠
Using   your   answer   from   part   (a),   please   write   out   an   expression   for   how   Voc  
scales  with  Φ.  [7  pt.]  
c) Given   your   answer   in   (a)   and   (b)   above,   write   out   a   formula   for   how  
efficiency   scales   € with   Φ.   Your   answer   should   also   include   some  
normalization   factor,   ΦAM1.5G,   that   denotes   the   peak   sun   radiation   flux   (as  
shown  in  part  (b).)  Again  we  are  assuming  that  FF  remains  constant.  [3  pt.]  
d) Now   that   we   have   an   estimate   for   how   the   efficiency   scales   with   the   sun’s  
radiative  power  flux,  estimate  the   fractional   decrease   in   daily  energy   output  
if   we   assume   a   sine   wave   function   for   Φ   as   opposed   to   a   square   wave.   To   get  
full  credit,  please  indicate  the  following:  
i. The  equations  you  used  to  estimate  both  your  “peak  hours”  and  “sine-­‐
wave”   solar   radiation   flux   as   a   function   of   time.   Your   “sine   wave  
function   should   have   a   peak   radiative   power   flux   equal   to   one   sun  
illumination.  (Hint:  if  you   integrate  the  two  curves  for  Φ,   you  should  
get   the   same   insolation   for   both!   Please   take   care   in   creating   your  
equations!)   Your   solution   should   valid   for   any   value   of   insolation.   [4  
pt.]  
ii. The   equation  used   to   estimate   the   ratio  of   daily  energy   output   from   a  
PV  panel  that  sees  a  “sine  wave”  temporal  profile  vs.  a  “peak  hour”  (or  
square  wave)  temporal  profile  for  Φ.  [4  pt.]  
 
 
 

  4  
3.  Metal  Contact  /  Semiconductor  Energy  Band  Diagram    [20  pts]  

This   question   asks   you   to   draw   an   energy   band   diagram   for   a   metal   contact   /  
semiconductor  junction.  Consider  the  metal  and  semiconductor  below,  separated  by  
an   imaginary   barrier,   with   vacuum   energy   (Evac),   conduction   band   energy   (EC),  
valence  band  energy  (EV),  Fermi  energy  (EF),  electron  affinity  (χ),  and  work  function  
(ΦM)  defined:  

 
a. To   draw   a   metal   contact/semicondcutor   energy   band   diagram,   first   line   up  
the   vacuum   levels   on   each   side   of   the   “imaginary   barrier”   between   the   two  
materials,   as   shown   above.   Notice   how   the   Fermi   levels   don’t   line   up.   (No  
action  required)  

b. Shift   the   semiconductor’s   energy   band   diagram   up   or   down   until   the   Fermi  
energy   matches   that   of   the   metal   (make   sure   all   horizontal   lines   move  
together   –  EV,  EC,  EF,  Evac).  Plot  this  configuration  on  an  energy  band  diagram,  
labeling  all  variables  as  shown  on  the  figure  above.  [5  pts]  

c. Once   the   imaginary   barrier   is   removed,   free   carriers   will   move   across   the  
junction   until   equilibrium   is   established   (diffusion   and   drift   currents).   The  
build-­‐up  of  charge  at  the  interface  within  the  semiconductor  will  cause  a  field  
to   form,   which   will   cause   the   potential   energy   of   free   carriers   to   change  
within   the   semiconductor   as   one   nears   the   junction.   As   a   function   of   “x”  
across  the  semiconductor-­‐metal  junction,  plot:  (i)  the  charge  distribution;  (ii)  
the  resulting   electric   field;   (iii)   the  potential  energy  of  free  carriers;  and  (iv)  
the   electron   energy.   Label   axes.   (Hint:   You   may   wish   to   follow   the   general  
approach   outlined   in   Lecture   5,   slides   19   onward.)   Congratulations,   you’ve  
now  sketched  the  energy  band  diagram  of  the  contact!  [5  pts]  

d. This   ideal   case   assumes   zero   density   of   interface   states.   If   interface   states  
exist   between   the   metal   and   the   semiconductor,   describe   qualitatively   in   a  
sentence  how  you  would  expect  the  energy  band  diagram  to  change?  [5  pts]  

e. Will   this   junction   have   a   linear   (ohmic)   or   exponential   (Schottky)   current-­‐

  5  
4.  Recombination  in  Solar  Cell  Materials  [20  pts]  

You   work   at   the   PVClass   solar   company,   a   crystalline   silicon   solar   cell   company   that  
buys   wafers   from   other   companies   and   then   processes   them   into   solar   cells   in-­‐
house,   selling   finished   cells   to   customers.   Your   company   makes   cells   with   n-­‐type  
emitters   and   p-­‐type   bases.   You  have   recently   been   making   major  improvements   to  
your   solar   cell   fabrication   process   in   an   attempt   to   cut   costs   and   stay   alive   in   this  
competitive   solar   market.  Your  two   most  recent   changes  have   been:   (1)  increasing  
the  deposition  rate  of  your  silicon  nitride  (SiNx)  antireflection  coating  (ARC),  and  (2)  
switching   your   silicon   wafer   supplier,   which   can   provide   wafers   at   a   much   lower  
cost.   Unfortunately,   one   of   theses   changes   to   the   cell   fab   line   has   decreased   your  
short-­‐circuit  current  dramatically.  Your  boss  claims  that  this  is  due  to  increased  SRH  
recombination   due   to  iron  contamination   found   in  the  new   wafers,   while   you  claim  
changing  the  SiNx  deposition  rate  has  greatly  increased  surface  recombination.  Let’s  
try  to  figure  out  who  is  right.  

a. You   decide   to   measure   the   minority   carrier   lifetime   of   two   wafers   of   different  
thicknesses.   Both   wafers   have   SiNx   ARC   deposited   on   both   sides.   You   get   the  
following  lifetime  results:  

Thickness   Measured  Lifetime,  τbulk  (μs)  

(μm)  

180   12.5  

500   22.4  

You  make  the  assumption  that  the  lifetime  is  only  limited  by  two  major  factors:  
recombination   due   to   iron   contaminants   and   surface   recombination.   Given   this  
assumption,  please   calculate:  (1)  the  surface   recombination   velocity  and  (2)  the  
bulk   minority   carrier   lifetime   limited   by   iron   (τSRH).   You   may   assume   that   the  
diffusivity  of  electrons  is  36  cm2/sec.  (Hint:  See  lecture  8  notes,  slide  43,  to  learn  
more  about  decoupling  bulk  and  surface  lifetime  effects.)  [10  pts.]  

b. Using   the   τSRH   you   found   in   (a),   estimate   the   iron   concentration   in   the   silicon.  
You   may   assume   that   all   iron   atoms   are   interstitials   (Fei.)   You   also   recall   from  
the   PV   class   you   took   in   grad   school   that   the   capture   cross   section   for   Fei   in  
silicon   is   4x10-­‐14cm2  for  electrons,   and   that   the   thermal   velocity   for   electrons   in  
silicon   is   calculated   by   v th = kT *   ,   where   me*   is   the   effective   mass   of   an  
me
electron   in   silicon.   You   may   assume   that   the   lifetime   measurements   were  
performed  in  the  low-­‐injection  regime.  [10  pts]  
€

  6  
GRADUATE  STUDENTS  ONLY  

5.  Measuring  Diffusion  Length  [20  pt.]  

In   Problem   Set   3,  we   learned   how   to  calculate   EQE/IQE   for   a  particular  wavelength  
of   light   knowing   the   bulk   diffusion   length.   Again,   this   formula   is   only   valid   for  
wavelengths   that   have   an   absorption   depth,   1/α,   that   is   much   shorter   than   the  
device   thickness   and   much   longer   than   the   emitter   depth,   such   that   the   diffusion  
lengths   of   photogenerated   carriers   are   not   unduly   affected   by   the   front   or   back  
surface  fields.  
−1
⎛ 1 ⎞
IQE = ⎜⎜1 + ⎟⎟  
⎝ α L diff ⎠

Below   are   two   EQE   curves   taken   from   different   locations   on   the   same  
multicrystalline  solar  cell,  as  well  as  the  reflectivity  of  front  surface  (which  is  coated  
with  an  ARC).  Region  1  is  full  of  many  small  grains  (and  grain  boundaries,  i.e.,  crystal  
€ 2   is   in   the  middle  of  a  large   grain   on   the   cell  (largely   devoid   of  
defects),  and  Region  
defects).   Notice   the   EQE   of   Region   1   (the   defect-­‐rich   region)   is   generally   lower   than  
Region  2  (the  defect-­‐free  region).  

 
a) The  values  for  EQE  for  both  regions  and  the  reflectivity  are  posted  online.  
Please  use  the  posted  data  to  plot  IQE  over  the  full  wavelength  range.  [3  pt.]  
b) On   the   graph   you   calculated   in   part   (a),   please   identify   the   three  
characteristic  regions  of  the  IQE  data:  [8  pt.]  

  7  
i. Wavelength  range  governed  by  front  surface  effects  
ii. Wavelength  range  governed  by  bulk  recombination  
iii. Wavelength   range   governed   by   back-­‐surface   effects   and   non-­‐
absorption.  
c) With   some  careful  manipulation  of  the  equation  shown  above,  we  obtain  the  
following  formula:    
 
−1 −1 −1
IQE = 1+ α L diff  
 
From   this   formula,   we   see   that   if   we   plot   IQE-­‐1  against  α-­‐1,   then   the   slope   of  
that   curve   is   equal   to   the   inverse   of   the   diffusion   length.   Note   that   the   values  
for  α  can  be  found  online  and  are  assumed  to  be  that  of  crystalline  silicon.  
i. Plot   IQE-­‐1  vs.  α-­‐1  for  both  Region  1  and  Region  2.  Please  only  plot  the  
range   of   wavelengths   where   the   IQE   response   is   governed   by   bulk  
recombination.  [3  pt.]  
ii. From  these  plots,  estimate  the  diffusion  length  and  report  their  values  
in  microns.  [3  pt.]  
d) Given   the   information   about   the   two   regions,   what   recombination  
mechanism   is   most  likely  responsible   for   the   difference   in  diffusion   lengths?  
[3  pt.]  
 

  8  
MIT OpenCourseWare
http://ocw.mit.edu

2.627 / 2.626 Fundamentals of Photovoltaics

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.