a ...-O ..Aolui 1.1A.
‘- 5‘.‘ '. 3"... .03... 4.0‘.‘ .C 4 . . .
.. AborcLA» r. 3??4u.y}*.‘.. .’. ......orof A 3“ IA! .A .
...-51.: .33... ...-[.3913 .... ...... v.2 :28...!-:-..- on; 2...: ...Au . .AA .-
$90.10}. .3. I ...: . )0... 169.?- .t#’. .. A 28.. .3. 0'. [9.0.0. -' . .. A: o..3.r..A"\Au‘or pl .‘3 ... A r A IIAIA A. 44 .l
3-10....5oco‘Act no; A.\ .u“...‘h . .),A a: - .0. o "I’ c .SIAA. ol“.x'§9 A r.-3Lvo.Iso..oA .. .— ... .9 flwAIH. .l“ A.
...-..."..F‘fl- ... ..9?!.. ...-2‘... . ‘. . .33...... s! 9.! ... .- .21....ol 3 .33.. .ll....3.AII!....Z-.w!u ..L 18‘... ’m‘.“ it. . A -
fi :0..." -.\ .O n .0..." 30" A- .. ’3‘; .. A. ....“l ... .9 .. 3‘. 90:3. .9... ."$ (...-0L1, 1‘s. A 5.. C... x . II o «A .l‘ ‘ahg 1‘
”ma!!- ....-. .70. .93...» .......:0A.v.v..u3 . L .. ...... 31.... .... 0&1; 3. a... .. ...... . .. A... .. .A R~w¢0 4.x .3. o .Ik. ... '5]! a. ... 1.3!,“ L! hail o. v- A .
. . ...-A... ... 3... .... ... ....8. ....-.A: ...-1. - .:A .0... ... .3211}: .Q .... . .... .1...- . .9. .... . . 1 A .... ”...-6.. .. . . oh.- ..zlh-l: Z. QAA-vg It A AA: - A 1:493: {- A
cut” .0 .. 0.... . ......I...32=.A. ...; ... .....«I- 1.... 1..., . A: 2.9... . As. 2.1.99.3; 279.. .Cixsru K \ts it 2 A. A. A A A. .15.. p .3, {EX-21;.1: ......
.. A . .. it! so .7... A ...-...! A... ......I ...... .. ..o .52.... I 1:... «3* e 3.30. @123 NttC AI... .3 ..nu ... f at .A a... 5.. 1 4 A..A Hid”...- .v ..rA-Avtpwu 31¢ ...-
A . n 3 v: .3. 4 A." . 0.. :- ... AAA AN A 2 ......Ino..ilc:-.. .. .. ......c: .o RIA L 56‘. .u . . .t ...... _ 330A! . If. Sifw 3. . t ..N. 01> .A l. . 1L1... ..1 7 1 ...-1...... E;
.A no.6 A.-. 3 i Z)... 2...)...10. .3... n. 35*)... P... .I l! A ..v. 3.3.... .. I... ....119462; .... LA; 53 .1. cQu-Lu. 23.48%}. ... . . ”A “q nwvl Ha. u..-\1.< $1.37... SIN? .....¥¢.<€.1)
A . I3..- 5.-!o-..v!..9..g.ll . A.r< ...: ...... ... .... . ....tf.... A?! a .1 .58.... A .Jrv.a1..-.A .... 82': . Au]; .. A I. . V.- hfi I 0%.. I !. .- .. A19. 0.39.1 .4831 a.
.Q. . ......o. ‘11... cv‘3.‘q.. ...-...... .8lt..zt$A.. A... ...-9. TI... . a“! . . . QIJA.8'D:\! I -..'.O.~AA alt. .. .. 9’ Au .1.“ 3 . (... v r .- .. L; .
.1 ...... -.Xo‘advé): . Ar. A ...... ...........:;9~ ...... A. . -. ......t. ....- .O.. ...-.1... . , ..--39...: 2. ... .IraIAIAA.A9A.aA...ACc .... . .6» .fl A -, . ‘5‘... 1.3 ...-AA . 4.. A A:
gut... . a _ AAA-v... (-.... J. A‘ .... 31:9... A. . 3.0 o n A. ...-1...; . .333 cc o ..A..£-.. c ...... a a». v.7. .... .. .........._. A .I-..- .3... 9.16.."I.nu|§. .. .Ahal ..‘X . T. tJA-c ¢ xiv-17 AA}..-
.tlAAA 81:1.ntArtowA“; Univ-...: 79... '10 . . 3.... o... v . A‘.t.\..$;...r ... .....A!:.. It... .. A... . {...-:26. .:|,.? _ 5...... 7:13... 490:3... ...4 ...S. ..ct . .2 . Al. ,
3P 1:! . I: 3-3...- i..t.fif..- A AA. ....-|19.;A..na ...... 0.03.... ...-...:tslc I .73.... . «A 02... -..... , .I. 1 .. £15 ...-.32.“... any... .1! A. ...”. 3a .
vV-Aaoc I 0: Alt-5:3,... p... :1 .87.!)- I ...-d .... A... s ...-3.1.3.13. .......o.l.\.... . .. ..t‘..r:sl . 79...... .c A . 3:5! no .8. .19 .
Y......I\..m.....-1397:2359... ...-.21 0.. o: A’s)... 0:8.- .A.,'..AA.9L .Q a. 7...... ..s... 2|.... A. .33.. a: o: C: (IA .58.. 4 ..
no... .-"P....:Is. \ 9-9.9!1...l ...-o ... I. A .- ...: ..433s9..I? .I .u.‘ go. . . .. . .2. 1 \w M § .
. c 3...: .....o' o . .- ..o. Pu.Jfl.do.\fin..-..nvhu.x¢t. .38. A3“... in. “at! s :2
A . A. d A #5 b 1. ... . i O ...
1.. ...vl.-.la.rr“f.io: tab - .us V
A . s ’40,

   

                               
  

 

‘ g .-

5999....03 f$ .- . v .9“. .

g. ...-......tiux 3.1V. YoIJ-Q-3A_.A\.3.oc .

, I...l-5..35...8.' .- .w....o‘£..uuu‘3. A. .95. .... ...9)..1AIAI.Q§A.. 99.3....

1.50.2.4 .A...A ... .0... . so... A A... . .... :8. .A\.. .13.... A2 .... ... ..A .09 . .l.

. 41A ... ‘§ .v‘ 71’. .0. v A. Aflpv-‘P’p . .9... OAI' oéh’v two“... 1.. c i .5 .AA 9". TIA-.0..- .Ao: \ A ....lp .

51:122.”: fl. . tic... .u (JV. .. ... 3,9. A: ASA“?! 8......11. ... .. . .9121... 1.. i .3. .. ...... ... 3’. A. ... o. ... cl. .... . .0 .

...-u 6‘2... 9 .21.. . A a. . . :0... A}... 6.. .. .IAckftoo. f. .v .A-..O..§..I. . ’IOJN.‘ .0 ...-£913.91.- .. $.E0t 34... 30.:- .v6# 3’. ...-l. 9.02....

‘R’A: I. ....o ‘ .. '8 S . tut-C. {oh-'1! A .199 Av .21It J. .39.! 0!... Us ‘0’ .40.... t. 4.. ...... A A! . . .3 ...-...- AA.. 1». . 'Q A .A. - .6. A c ya}... 0‘3! oils. $8.30!...s‘. .30.

. .. . .. ”3“. a .. ... .0... .IA ... ...Alo...f. . 0.305.... .43....‘..I....9. l..‘.A-4: .- . no 3?... ... 09...: .... ... .l . .. r: .l.. O... .- .- no . Pbm.‘ .81.... t .. 081...!0. ...!
.‘rv‘. ‘9Afi 3 ... A J‘s-’00? 2 e. .5... .. ' ...)Q..... 0217...... .18.. . .... c To: . 79. 3!.- . ... 4. . . . 3a A. ....v". .A . ... 20.... A'b..‘t - .. «.....A IIS‘ 7
... ’4. . rat! I ....o u - .. ...?I .l ...-3.7.. £31533! .5?! in 7.... .W. $.o.!§-8.f:9'2 ... t .l. . ...: . 0.! ... v .... If! . ol...f-.t...w1..l. I... 1-! ..vTflrti

£3.39... .s ,3 393...... .A 9A. 0.... .- ...! . A (go... ... .3... . A 31.. A. .A...o:ld3 .. r AA.v..At1. . . ...o. A‘... .. .. . .....A‘... . , .. ...... . s»... (...-o! «A . x .A

...-s 0‘ 3.2.... .93 ...I ...-2...??? .35....) latct: ft? 10.3 21.. a: I. . 1 . 31.9.3! A . . .A . ...-3:33.; .AI\_.:..A.. . ..orx . ... 8 tiékv e l:
o....‘lfo.!.ao 1. ...qqA-A.Y!-A.A|- ...-...}... 11A... ... .. ...-$1.25... . . ... . ...A: .1..0A...’vc_o A .31."...9.\ . .05.. .A .

. 06...]: A.— .-..S .. ... it. ...; 3......3'Av’izoltuf 01': 3 9’... B. ... 0.8.0.. . . old 8'350".w08 A.)

...... cv.’ .. . .. . A 0.... 0-..‘1-v.‘ ...3 .r ...(o o In —A ...}... u 3.0... . A. ..IA .. c .3. {-0.01’0 Q... \ co
8 t... 9A.. 3....95‘. ."..to:. c ...-.0. .. "cloak-A”— Ao’aa-o .32."...9 '0. I.

. ... In . A A ...! 4 .QA A h V A.
. w l...3l..:..? 19‘ am zinc-.03. ..51 L?

... A .1 I. u

:55?! t u .3
oir. 7A. \. ...L.» 7:5. ."‘.
..u ‘.Dto..£...; .. . a
.1..£.oon‘8..o.t..ou.l .2. .
.. ...-......10 r80.‘..| c . .t... .
054. r .. .:.'A ..A. .... .
(IQ; 0:...0.;§_.:'

    

-. 3.0. u
in o... A}, “tag \J‘So. - . . .
‘I A,“ ..o ... .9!) $.40... tn 3. ... 1an .
.1 o¢....\)O-u u 0.... (p? o A. (’6‘. .0 D.§boi"l *
A L. 9.. .A.’ 50A: ...-... 0.01. A... .15.. ... I...-
‘ . '0- I..’ O . ‘K’l‘...: .
.... .u .w . a..\.d..:.-.V. $5....
:0.- .. .... llurot. .7
n A... A I .A

J.

 
     

 

 

           

.
vriuw‘t
A . 3A .
I... . . 1 .3 ....
In. a _ .J‘ .v 3" siti; .‘J .
- 8 . . ...Sobo... .1; ’9. OVA... 35¢: .... ...-.1. {O ...fia‘oh13
- La A.” 3.1.. A .....A .3". . . t..$o4‘ A... . .I A. .392. .vl ......rvu: .Afi... Iiftl ... 90.... ..
.. .l. ......rc: . AS?! SI... ...-3... .../g .....155 ‘JNJgdfisit .ct .. ......r I... N ...-0.80.! .1. . A .
...-A... .1 .A .l ...r ‘A‘P'r 1.00.. ..J ......Oo’. YEW-‘5. 0.. :3 A! . L Q .32....qCAc‘rro ...-...... AAQP..9.Y. ..- .....O¢ 21?: o. -. .06 ...). . .
hp .. «‘3'. A . as; 'YIIICAQF ‘5. 3.!1». . J“... .A A. ..A. ”aw“ A\ a O Q. .4- us...‘ 1.3.1.9.! u p .. .o u '6 A - ’ u .1 1... .A. 6 .. .. .3545. ...-ALI- 51330 93. c ..o 2 v. «I U. I.
A. 9.3.23.3}! ......al .10: .A .11.»...ALO'F .lr - r1..ig . ..oA-Ao...2. .28.... VI!!! .A . . ... ...-.7 .. ...: 11.3291... .0? A ....o ...:A t .A At .-¢.. Isl»... . ... 3...!)- ....
61; lo».“\ 5 .....0. Alnol. .‘ t . A... C.- . 3818.“? ..l on “-13-. .93.... .99. onxlv. :3 A $300.. ..1 .A. o .10 I)! . . .1991... in? ... .931. :1 p A . 9.3.3.- ... . A 9A. 0.... a . .3!!! .I . .9‘.A.\r- '2‘ .
. . A. A A 03’ o ‘I 190.03%. AI... 3. ...)... $2.9: .. .... Pool... 39...... ch .‘3..o. .- .-o. '9..-.‘ :- AAASI... "AA. .A. . «Er‘u I... . .91.... .1“. 1.5.55 4-. ~89 .115 0.3"..3Ak‘2szl 0-
o... ....A .21.! A I... 3.... ... Aft-$3.5! ..a... .... 1......5-5 .o . A... . . 9.... ...:31. . .. :90. ...-r. .... 3r...» .. A .16....A... wit . ... v. 12... . . 5.5.5.5.... ..tv!L-.
n on. ..AJ¢«D\*1‘.‘:=-:-J-u1m . .e. .0. A'x.‘ir.o. q-u.$a .. "RA 3. Ali; Q -A 0.. A. ..Q‘a’1uoo.v.:7£.. A? . . ......o. A... . .1... A. .....D‘v’ l .’ ..00" . 0A....‘l..§o-o§v,4b do”! ..r 5‘ g
.0.- I. ..‘xa‘é. ‘3. ”a. ...-J: A o .o N J . l. 1:... ....9 .. :1: ... ....«33‘. I. 3 .l: . A. A! . ..‘ul‘ I39: {3...}... 3).... 1‘}... A v... .o 1.6301! . .....qb‘ 3.. -A.,..\I.Ao. ..AA- .1...
a .I 4.5...‘3‘50b3lr A .9 .- A... fi?‘ . .. A. 1!...122 ... .6 0.5.... .... tn. NA '5. ...... tin .... C ...: .3 9.. (1......0A .A v.1. .57.... 1. osfic .tx! ...;r .... no. A it; ...-......AaAk‘.
A113}! 3o??? 0... 3.31::331 AA... .0) L‘.A...no.r8t,!!..!7 9 2- no. ..f.vi.. y. ‘2'. ...... o A. .82.... 9 ... «It... .. .... .....‘fc ...’..8 .*-'1$..4 .l’: .A .59.... 4 .....‘l . ......A ‘21.. “19,02‘ .
O . .Ii’ ... ..--.0 l. to A A: A ... . .. 3.129.901... . .0 A... 031:? A 51...... .A .31. .....r:: .I . o .. .Ju...Al ...s ...I! .... a)... ... o .....c. 1... 59.99... git‘b‘
. Av . . A)? ..ll .8 38.... ... rt: q ’39.. . ...... (Ii. 33:... .A. ... A.- .. ... if... 1.... So... ...!!1 . 4... :1! . ....UC3 ...)a-Akttltxiovw»
: AP. a. .4 . :1 A ......a! i' A .o '0. 2 g If... u.$4..l.oq.'~otv A. ‘12.- ..7: A .10. ...ttucatzt; ...I'AAJZAo-A'ld v .r.§.~ .tle*o\
. I . .' ......rr 1.? .09.... ..A. ......A . . «3A....-A .o .5 7.... :2.‘ .l—l..llsv.fc4}.’ ,..\.¢iz .1... ‘c-..$}A‘ .9
A. .5 .LI. ...{.o!0:..0. ..A. . . 3.0.41: . ..1:b..t~...!. 1A...A...A!.. A.._.A...o!-.o..-... :2th. .... c!
i.) l I); o. 6.0.1. ...! 32.30... “....v :3......~‘ .AO~A$‘.r.A W i
A..A.A0h .... .Jt.‘ tfi.i«fl A
. .-.:w‘rlv ig.‘
3‘

 

.. ... lit-Cit. .. ....
r o 1J0! .Q..o.-$ .. _ .
ft.‘ no. A‘ul‘! ,A. 5.433.? 1.. AI . o.
.. A
. 4.2.6.: ‘A‘o'3.; not. .0
. o .. .... V... {J.1t—,. ......Am: .. .A
. .. '5. IIL‘,‘ .\ ...? ~ .3. 9.. 45.1.... i
. ..A ....A at. 1! ... u at
LouI'" A3...’ ....A‘o’tt. ’t‘.

PM-fiAAn.h...n.A.-.. u... ...-3...; ..
.. a....infinaiiunw....A.-.A-A..u....:.”Author-Ar“... i... A. .
A A '1. I. .... .IO- .uai... . ..hn-lro...v.(lu..v..ca.. 0‘ . - 0.. .A . A
...-.933... '17 ......IA-HA...A_ Hn‘.-~a.u‘( L... ...... r3.“I.’.JL". AL.
A .

   

O... .
J. %r
.1‘i‘ . 5V.-
. t9 ..rb‘fll.
o.v..la\fi‘ .
«CY-u...- £371..
u! .
A. . 9
I‘ll A .....,A\ a. a: .0. ....» T
.A ...A .u A .3. ..

        
  

 

 

     
        

     

    
  

.«Mlyt ¥ ..‘I’\ A...
A u o A a
i... .... 4L... . a... 9.9... ...-:1... .43..
‘QAI. h of. .v .49. A... {uiq 1...,PHVA. w. ... m" ... t ”I." .. .... coal-A ‘13
u «“210... AA ......AtuA $IX J _§s..u. 1%.... ..05. Frid . ’2'. .. c A...- ... o L636. ...-7. . .. ... A. .1 .ok
. hr“...- rc‘vl 5.3.. .. . ’t u z)! [Azoc‘tvi 0.... . ..J.. - IQ-..o'331'. 3‘ . to: o. I.- .n‘ r...—. ...: A .
. . fishufilA icing-akin... ....sn..é..n...rm...uz.. . in»... :2...- Au ...-.... -... Eran; ALE. 2......
Al . .. O A .. o . .. A . r a. . . . .
.00. {v.Ih'tl _. . 3......JA 933:5..‘2. o.d144..fl‘!.ro.o. )Au ...”... usn.AXH.A.AS..Ioo-.t. ...4. .. n . v .- .. ...-3... 0...: A ..o... .A..:..A. ... A .o- [A .3. .u 2 . .
oo “0.33.: .A l A av. 00:5 .. I A) 0A 3.3.. ... .sta Atatuo- . o . .... .2. ...-21... 9A.. .. . . . .. . O o ...-......I! ...... AAA-1.8.42. A .1... ...: ...-...- VIA—13.... . A 3.17:, ...: .-
..10‘!‘ A. 7i a ....l. 911...: ......«orof'r-N I. 5“... ..Q .. ... ...”... . v. .. . It A. 9 o... . o .13.. o s. o... .A ...) _. .. o. .0.».. ... ... ...! A. .4 .Qoon 51%? .-.C‘A‘fiifi.?§...
.— A.qflb. it cu‘AI fie. ......ll .0. on . . 4 A» 5...... .A .. . LfiPfL 0.? t coil-33.. .Ilc 6:33.. .o 2... Q .1 0 2713...: ...IrV. '..toA. .0:......Aoo:r O .\:- ...... o?.‘. 3...? o.‘
7...! ‘t‘. 3.-.!!!9. v'. . ”to”. out. ~ ...-0. ... . .-".A- .8 ...) .15... pan...) .3...cTA..Av0-. ... . ' .0. ‘1! . LI. .... ... \. ... . A. .... . ...-...... 3"‘ onf’. ... >095; ‘56.“.- Ac."r
.. ~ A ...: we). a... v (A... .53... .A o -..:l‘tbouv. A ‘6 A" T... v . ... 0 ..23..- . A! 2...... ...! A . 3., 3 A -...A 1%.; \."«ho¢..031128:13.lko‘§.
... 9.- von .V...) t .11.?! ...”.J ".0 A!» .0 .A A O. ... A A0!" . ..- 8"23vc0 1" x . . 0-; .. ... . ...??f. A. .91:- ....‘.‘.~ ......AAQAS3OE-mpuiou! 0
:Q ... ‘vi-OA’AA CY . 3 . . A. 1......IA . 3! ...? AA. A ..I.Q¢. .. .5 . .....‘uvcl Alt 5...... bio-1.0!. ‘Lbcflg
. . 30...... 2... . - .00.!!! 1!! ... v. t... .. .0 3.923...‘ .91..
. ...-9:01.32. ‘3 ..

   

. . WOA‘JU i..r.r. A 3:43 . .

   

   

. .F. Q o' I .0
.153 .. r .. 13¢. J 3.. A - . ,
. A r. AOIJA‘; AP” «11.3.15? .AAC.” .34.!) 1‘ 0. 3.3.1.6. .
. AL}. 3 J. . .tQAA-ofiwu. ‘xl. ......hu.A ml .AIAA. o C .
A a. n .- IO 5.! I -, . Q
i ... b ...-30. « n.&........9tf $.00...) «2 ...
V . . ... ... 37...... ... ol-.’ ..4 .AI- .3. . A! v
. . ...-...“. .2. ”J ....} luv-...“...
. 4 Taco: .- A

     

      

         
      

            

                            
                 

    
     

   
                           

  

     

 
 

   

.“\. Ann-"1. .
...-..a .... 2.335.353. . 20?... .A,.... .
3231-. “AAA. «AL-5A.... iii“. i1 ...-ASK. . 630 A. «u-
. .. . c. 1 if ’40: o ‘ A ‘-
ANKJSRARO. . .wfii Whit... .o bio); .Va 1 .1 .A ...“: 0‘: .m . .5... A r1... . It: ~
om”. rwfi. ‘. 3". .3: A '0‘. “lb. Agmfla ‘c ywt. “:g ‘v: 5 ”no I Uni) ...-lmlnuh‘o ‘ no: - o ~..6....l~. - .I ”A "l .-
... \ ‘A1 A‘ ‘1‘; b O 1‘. . .... A . A .n . . .A o. I AIL? T . ...!
ooQA ..'}\~nu o. 040...“. A\u a... ”a: . _ 1‘." AM" “dumb, .9J-Hfi 'r‘nvov'”. huh”. hat.“ 0 Aswan-O" "IA. "rub." k3”). ..".a.!¢".¢o‘ ... A” .v ”.9" .... a... . .. ... ~ . v v.».‘.. a.
n‘l. .0-34 C A c 03.. .1 A. ‘. 1 I .. o .A .. . “to. o 5"”. a A u .A . ..Q‘IA .... .1‘..IA| - ....l.’ .60.... III 30.0.... ....1: (...-'1'; o... A tux... A I. .r .7 5.60.79! . q u ..
C ..A‘A P .13 :1... .9. A . z... A _ ... ‘88.... ... .....A I. A... .. A Jal- . 4 o «LA-A. . . .r. _ .. A v. a
. t A . . 8.2!. “-3- . . .0... 1.. - AA.) .0 fiwfi. A “I A3.«u: U . J .0. unto. val-con”, 9' 3.0.3.? .o 4’ o .8. .n .1... ,. ~- 1 AA .(A. . :5 J4 .00 1.0.. 34...”?! n." l .1} ifs-..AJAF. v.1... 0.9V. . .....1.'!
J. . 1 n 71“.. u .13, A1. . l .0. hut:- .’f €9,384. lag-A... It? ‘A. 13. . . EA... ,1 {A ‘AA . .... . A ...? 9 32¢ din-9... . . .8 . ... o . . . ... . AA AA. . . . 40.... .
«3:.AAA.\1.A.L;A40.1IIo-Au i.»..... . ...: h 76...... .....T _ ...-.515. -..tux‘aa; 5?... A. .NAJQ...3 ’1 A!“ toe-3m? to". ”nu-Ar.v:rof-.42l .. Alf... {...-2.9.199 Lice... ......o A. .. .... ...... ... A I.
.- A, . . . . . A. A o. ..v I..- . AA A, o. . Q. o I . nl‘ . A. 110...... v; .o A: .A. . .l! A I: .. ...... .o '- . o. A. o. A. . . .. . A .. v o -
. .7695. 8..., i Av 3.. A! cam-LA ‘52 “fawning :2). Aunt-J" ..Awl:o; Ari-a. .... 0A... 9. . .. A “tots. A: .6! ‘- 1. N... .i....a,3£u.0 ‘._t.}...~uufilitak.o. ... ... ”I: 4.. :A. t .9... . . ”A. ..It. #8 ...-... . .3!..¥K..o.. rm. ...... .Y.
.....Q... . t: {5.33.90.20 In in . 5.9 £0.39... .... . 311763.. '3'“... ... O. A»... ‘15.. A... :0 (..‘110! OAIAQ. ...-.... v .....v ..- .1 ...-...... A. ... .1 re... .. .....‘A...! 1. -.:t. ...A ... . 1. £2... ... Ema-$1! . IV A .... l .1 ...:9I 19A.
QonkatznAofizu AAAA~§ .* iltiéwu. 1‘90... u! 3 ADI-a A... E ...-i. . . A 0- 19-0! .tl .IIL- 0.3. 303:... A .. 0.. .P. - o. ...A 15...»?19 -..: .0?! ,. . AA ..A .,.- . . A . .A.. .99....933 .32 .. IA): ......c..t. ...... Apr $.i.,\7 It. :....~IO).A’" a\:..(
.... AA A} .73.)... Epcot. .30 .5. 23.!) ... . a)? A A 3gb... .2? 9.. on... r... 6' t v.91... ‘ .16.. ’1. ... .... 9 .. ply... . A. . 30...... c A ... A.I.: ...... .18..x...\ ..1. 02...... ..§.. . .- . ......on ...: 8.:- .. ‘3‘... 2131.. t . A... he,»
- .4 ”d ....b ‘ a" .uf. ... A . ...“? .‘o~ ’2‘. ...-‘7 .1993. 5.}: 3.. r: 0.01-3.10. \ffi .. ‘.o..l Au. .2 0.. - 'n.\. ..‘u:. . F. - ...... .v A 1...: N v- a. 3': . noo'...l . A... . .. Co... ... ...... ~At=.“ .7 cm i a...
. C. . 3L $.40... .. t! ..O. . 3.. 5.3.11.3}! A. 2t‘rllvl3r- ....1, ...! .. 1.. N. Al. ’1. ......Ao... .99 A}... .n .A - A 11.. . u}! .3. , . 26.235... . .. . ...... .... 1.8 A. t; 39:... ...: .40: 3.3.
v .1. «...-b . 1:31 av ~31! r. . . . . "finding. ‘3. A p‘. . . l .14.. .C. PD. 3... H.- ITM.‘ 3‘“ .. Xu.-.9A.¢Aro 53? 2.1.; 1...".09 ‘- ..3JAI..! ...-.3... . . ._ . fift: . . 1.... ... ... A...- ...-a... .0. 1.1.1.... -. £042... cc} 0......
...-1...“.- Y”: ..-. ... a.» . Abr E? ...-r2 1.”. v t... boat... ... .... . .A.‘ -.l A a 5...... .. 13-... 1 AI). .... .15.. ..u .‘AAA ... A.“ .A. . .. _ . : ...... . ..AA..1¢;.9 ... .- . . .15.-..29AIAA .7 . A .kl km. 3.... 000 .
.l .30. A:—.bwAA..baD.L-I.v-.A(,.. . .. 5.? ..A 3 g _ . A ..3‘2 ... $1....» «Iva»... . .- ..ul-A O. t3 8......3178. .. lib-v ‘1 at. L. ..3 .. .3: A . .o A ...... . . . . . . . . 0- b . ._ .IC. rt..$_ .3 ’53....2 1... ’ viii...» i324 ‘1... .. . o.
.. . .A A An: AI”... . «x... .or 0.. 133mv.» .- $0.33,; U. vol .21 a . 'AaAA‘rouo. ......‘ic‘aoul... AA. \ o u 3.... a A ......“502‘3 1.3—,(37u0 .. 4L 3.. \J‘oafiat .. . . - .. . . .. . ....‘.......vr{:‘ ... 12‘ A! 1.95 ”ft . .W‘A. . A. 5...": 3
... 2.8... L...- .. ... .I..A ..vAI ... no! cu .QA. .. o!- 31-I A. .0. 3. 501.... ..v ..y . .30. - .1l.x£.o. Iona: . ....l. A.) .3. u .o. . ... ... .. ... ... . 3!. .8. 0. I! .. . . . . . . _ . - ....A ...... 3:2.” .A..l..€.-“.. o A. .....oov . .5. a
...l: 0 . .1 (.153 . 4 v. P... 4‘ I! . v. I” A n. .«oA $25 5y...1 P. n‘zIAmMO... c AltoAntot coo-Ist. If. ...A..Ao. ...i‘... . ......A .....{1 .- ... .v... ..ISoT. r.- . . . .. . .. . . , .. .93..- A¢$ ...-3:». .. 4...?! «It»? ...: quc. . :usr
if..4....._( {AIAIA .... A 0.. “to .... _ .h ...-0......“ hi... 3. 4.2 v2 r.(J..tu Aatlnt .. C - .... $6800....) u. Ia: ....A Al "A1: . .... ... . .. Ozlo . . . . A. A .. ... (r. ......A- ....O.!.¢....IA\1:. ...9..l . .- it‘ntvfiaJn...
- 6 A13]. ... 40..“vvu . ... 9H"... “3'! 3... ..~.1!..Q.o.g...o’lo.‘:. :33‘ .33.... . {8.1- “A A Aunt-$9.... O... IJio 4:21... A. .... I! .t‘.’... ...v .. . . . ...... . . O...» . ' ..1- if ?L¢.-.o Iiiliflototil A «....
a 1 1 . ...“. "r‘lfiio silo}. {gun-... § 1‘... .5835”. «41:1. 3 Q. 3.”. .o . .‘&.A;.0333.A ...Vv .. . 33?...t I»... l A. :30! ... ...i‘. . .. . ......-.~"v.t.... ... ...: ....u......' .‘l. 5’; .3 .- .... {9.1.7}... . ..a’n-I‘o‘!
. A 5.»: .1713..- .ixuz.‘ .1. _.\.. . . ..IAQQ!.-..~ .A. .. 4!. .9 ... .. .. A . .. ...: A: o 1.52.... .A ....J. A. .... a. _ .. AAOA . U411tl} . l a... ‘$81
or . .6 1A“ A )0... 3...... ’- ..- va.qJ.-‘DMAC .0... .n . . ... . ...c-KAZLAA .Afil. . a! . . .. I' . .Xo. .5.....v.|.. A. ...? . ..pcan. 0 ' K“
3"... . .38..»no ‘ . . .. a. A .\ ...: AA: KI... A ....A . A. Q .- ...... .
a A o ..O A :1...H Avg-fag...“ .50». o... A. . i-o ... ...aollm. . Ac... 91 to. ... 1.113.... ....O'..O.'.~.’ .
« ..fll. .. 0!. $2. ... .\ . to A. ...... 3.9.9.1....331.
.35.. A... QAAIwioAAo‘v .0. .9015 .Jo..).’ A 9A...
. who... . .SA...1.. 3"A
- A a. ....19. o. . 6.‘ ....
0 ’9.

.1!" C. .
A .0. :3, .. A...
.....L 73...... .m. ...,- .A.-...-..A-......_. . n. _ ...

. A. A {.01. ... “ 3;..5fi‘ . V A .t

1.1.11: ... 13.x» .v .§!.. In
,. AN b. Q...Juf “"00; 1‘ 0.0- 3.... 9‘0. -
~ . . A o. of. . r .v . 0'
. .- ‘I...“ .4.-.- .O.:AI|. .
L ... .v JAVA“? .-

  

c ... 39:2

    
 

A “o.l..a . .A
0’. ‘1. AA: .
.1910. 5:... . . For

    

- r 5!. I g a “o.
.- .Tc .2.- 3- ”...-... :1. A .1... A
.0. .A z . v LA. Q .
«on. a"..A' A’AI 1"?“ I. A ...!"IOO~N.-O-u:ol hr‘ul .V“... ‘avrdvh‘h vb“. *0 on: I A‘... -.....
I» ‘ I. | y o . . . I
2‘92Hl‘1x. tl....3r.. .- .t"....6..ol3¢rn:.fi ..a.’...th'- ..l‘.» ..c’vu- 1A.. 3...‘ ..rl slnéc I ...-A: c‘..o»...0 ..J . b; ‘2
.A-AFSIA.u 3:3... AalutATA: - . ......A ..nA u . ..3... ...»AAazl...v:o. .\ ......c o- ...... A... . .... .
..AP A. 0 . 12.0.. A .. ”.15“ AA ...33. AU“ .9... ‘Lu! ...-i Y! A. . ..8 . . ....HAU‘..\J.A.rh.\. huh... {riwnov-HA {K .....AAN. ”“4"”. $325....“ . A. . 10.3... ...... A OBI... ...- . . 8 AA .A
. . .‘I‘ .. O c . . A A. u .. ... . A d u‘ to . A _ ~A. .. . . .A ...
‘3 u. 5:. .... you-0110...... Alp. ’puc ...:AnwA-UKPNn attonunfir“ 10mm”: .. .A‘r.~ .A ...”.- Vfiamuo. “Au... “stylnod-t. o!- o3~.-m‘._.o . . I. n 5' 1.1“. A ... ‘3 A‘QBC...‘A. A. A.quA.f.do. A339! .I’a'
AA . 9...: A'V.‘.Qolt| 63.11.. - 1.3.03.153 3......- 4...‘ Ar 6.. ... r. ...: ...f . A . c s‘ to... . sfit. .50... .c». . . 2.2;!)- .o. A. .7! . Avg. 0 . . .55. o
. . 0.10.1.1. 0......- . .4 i.“ . A :‘Iv O ..VA . a AA - o ‘lvev. .- 21!. . ...AI‘ . .f... ..L...‘ -A.... 3.9!... .7 $00.).-
..‘3-0 EA...- SA‘.‘.¢§DA Iuv .. ......‘A‘ A... . .....(v . .A. A '— :l..._l ...-... ... ...-on. A‘;
2-] .‘P I. A! 3". Act-... A J“ ...-1...... ....1'1 V v ... .. A913}. ...!- . 1. .
3036.3:3.‘ A 2‘... 0 ....VA a .... o. ‘0 .. ...}: ...l.-OI..2.A~.. to. l . .
13" $.34...- o .. . .... .. .. . ... .v: . . . . A .‘o.. )L
.. . ...»). A“.O . I . . . .....'v 7...,»‘....?’.
A . ..A ' v- .’.o’. swat».
_ . . . ABC-I. . 1’...- . .

on. .. vvmuuuvwx (.3... L. .... .
r ‘ O uOV \A'. .0 6“...I...A‘. 3.3.. A A‘o’... It... . .A . ....O o D 16?.
. A. 0.0! ... M. ).‘.u~r‘. .....- .A.»..BU.J1= ...; A. (1“...
. do. .9. ow... ‘8‘ A . “2...... JR
.0

 

'0‘
..i.v¢.0i ...-03‘.
.. - Q- ..-

       

 

                                
     

        
    

     
          

 

               

  
      

     

  

   

                   
             

 
   
      

   

  

 

 

 

 

 

                     

 

  

     
   

  
 
  
  
 
    

 

                                    

   
  
 
 

     

.
v 3;.
v. .0; 3'. «4000” II. ...-1v v. .0
. . a A. 0A -. a a
.o . ...-b 1.... ..‘V ‘2 553....‘3 «I... .1 0. sv.‘-Ao.ov..{00
ADA 3. . vb '--.OTO 1-. o. ...AAA. .' ‘1. .2. .41 '04.. $.10 ....‘5; 3 V: .A- u
r A! ...». .9... ....‘2.’oo.o.. 3:193: 31.... ... {... “3 .... . .t 1.3!... ... AA . . u A . . A A
1. I . . 1(- a.. L $.’. A. ..A ~.. ‘AACIA Oil |.% I 0... AA! 0. ' .AAJ 1.. Ana. .IJ'J. .... .3: o... v . guru“: .... so. 4..
A: .9 ,¥iA3 3“ 67 .... .... AVA Aafiot, on. . I 5?..2‘04 ...“.OI. '99 ... ...1 to urn-....- o.-?AIA A. l v‘A A. . u . , .
... I. ‘1 1". .. - . .A A u!- A’..... V41..... .1 \ 33A A. .v .0... at! ...). n. ~H~v¢...;o.. .1... .....fl.‘ -... cg". . .. c
A 90 Q! o .. (q 19‘.\1 .. v .3...‘ “‘91 «Al Aol.lo.-yo..‘.o. (A. c .A ale-v 1“”) o. .83).! 40 29....» 'A .A I . A. A. . . . A .
... ...! 'n...‘8?.i A... A. '. AA). A. A. v. ...A... . -Al. O. [0’ I . V3 . to“ .311. n ‘1. It A O‘A 3A. . At; . .olr l..'vv . .. ’13.“.3: Acc .. . I u .8... v‘
.o ... .V.v J" . I. o .v whoa-A... off; .10.... rA'.‘ c) 3"}... .OC. .. J...- v...o’l. I...a.v..AI...0Q.-v oz... 3 c93‘: A ..,AA.....J:! .. t. ... DEV ...... 9.. ..A 0 Jo... .0. 1 O . A v A .
. u A....r\. .. . A .o..v A. . A...) x...- «03.3%.... A. ... 0.9.3.35... ...“..alc -.!..c. \AA.*:.-.....-...A AAA .8: 0.0!..53AA. A. r.- A . 3.03.30... .9». I .0 v0- .. ...... Pt! . 31.3...9
Al AlA . $0.. 1H1. L..A.hAu .0... .l .06. .7 A ..6A. o...‘....‘.‘.~.$ ¢ .6. . 2-3.... ...: g ...: .3 «03.1" .... a... 4. ..O - A ...... : , . v at . . 0.. . 0.. .‘3 .... n . . o. . . ... ... I .... .. .... .. ... .. ...?
a. A. A Aow- . A~ ... - ..1. . £3134 V .. 0‘. .AOAvA: .11.? . ... A. .A- .\ *9... .1 . I... 1 ...-.34.}. ... . ... ... ... . . 9’ ..AA.’... ..A. . - ...- ..11..... .A... ... ... ...: ..A.... A. ... A .. ..‘wl.v....e: . ....
\. I... no 0‘. . l A A'A. _ 3.5.1111‘I . b 0 1.”! .A ‘ 90.03.. 2.3 ...! "o‘. K.‘ A .o‘co .01.; o :3... - ...... Q! E . . . ... A o . coo A ......QQ A. . o .. i o v . v A. c . .. . ..4 A. C . ...! A. )‘p’.
.3 3.0. ....-. .4 ...“luoo... .0 ...!L?‘ .3!(.\ .s_\. ao‘ XIC...0...¢.AIA A $3.. ...I ’5‘: ...)IA . I . Int-{Ju- .. .321... . . . . . . ...... .le .... O. 2 .A -A!.\ ... . .- .. . ..v .
{.12 3 Movn . ... A v.. ......i .. i .. .‘~ .. A. -..... u . 5?. . c .A. . ....o o 1...; 9. ......Ai . (co-...)...- . . 0.3.32! . . co. 4”. 01...: .A . .3 30. t v .. . A. .
7 A. A. .1... I. J 91.8.}. I {30... hi...“ Scarf r21 15.11:): ... . fizz-.196. .3 \z ...-.1. ...-... .335... u .....l S. a J. . put... 200... A... . 1 ...... ... V. a. . .9: .
1... .- 19c...“ . ._ o. 6%.... ......1 . .3123 . i=9 ..)-\... .31.. ....l‘. 1-33:-.. 19.7.2?" .A .0 .A 9.8. A. . ...:I. ......» .. 1...! . . ..- A... A. .... .' ...... o . . ...: .
.. 4 so \3 ..A .3. o “h. 0. 14. 6.1;... . A. . . 5.1."!- ‘1. )3. .5..t ... ...)...3A... tu no. ... A‘Et’gvt‘; . - .05.... 10A. - . . .. utép ...-.03.... 3'55 ...-.9. . I. . 142.....5. . o .. . o... ...? . . 9 A . .
v... Q v A. ta..o1. ..‘O AV, .t-c \- . n‘ .‘o o‘).1-.1-{ 3.1. v . ..1knP‘).O‘.. cl, t.'-b.h;~..s.’.‘o.fii.a .- -n .96. 9D... 0“A‘ O ..l . A . . ‘1. n‘- .09 ‘1 n . A . v. A....J.. 0. 099A. - c b . . A A l p. o . A 0 v‘.“‘ . .....I.'I:A|-l.o ‘ ..
. .O-QA .v . r A .V 3.. .0... a. .o . . A‘s . o. . A ..v.. .....n..3 3.2.6 A. AA. .. . ....AA..!A.C.A.IQ. also...» . . ... . .319: .. . . “LtA ...t .. .0... .99: ... . . . . . . 2... ...».A? ... fr ...... t:0.rs$...1f..0\a .1 nd‘...’ 4::
. .Q 1. . .1) z)... 0- war A2... ..“I .33..m.fi.o .3. Vol.3). .23 ... l.-J.t..|1lh-OJAoO-3..r . ‘81....‘5-9.‘3.t u.) 0.. . . r A to ... ... p: . . 2.. .... A O 5.! ... ....o . . . . . A . . . . ..-A'\£93...... AIA. A‘s AAA. -. A. H, ...-«in- . 0
v i. .A 0A co: .0... A o - . v A . . o .0 t A A. . (#6.’ A.-. ...! 11!... . ... A . .I ....A.¢At¢. A to. t... .. A A. J . I. v... . . .. A... ...v .... ... a. A .... .. A . . ,‘ . . .A . . . .r‘.’..;. A... :9
. .-....AJ..._\ .5 A - O... .. A.._A ‘- I it! ...... oc‘ .... .53.... 3.. ...-AA 2....901 - 01‘. .24 Agni-3.. 11‘... LA; .37 ..~ 9!. ,. .4. 9. ... . :5? u . 3. C. o. .2... 0...... ..l .. 000 . .. . . A . . Zia-A &‘
..lvou _ A. A t ..- v ‘6 .l - Q. - a: A ....v} .4- AA ...-....A .3 .. .. u‘ . ...o. ...? v .A.vO-'o. .00..\ .93. . 2 .IT 0 .. A.-. o. . .3! In; .. 31:93. . A.-. x o ....- . . At. 0
. 5. .n. O . l. A! ...... .o I ...t A u 29.... F .1. (IVA? 161...? ..wt ...b\ “Ax-...") ’3\.!.n..¢-.. .J.Y!.r 1.22,... A. Av”! v... .1337... .. .0 1. ...... 7.3. . A I .0. ... ....-tl . .... 3‘- i...
. . ..-- . .. .. ... ....A .. ... AA A. a: 0-. . . . ...: . :1. . A. A. no I? ...... A A 5...... . I $..rt.\.'. . .o . .o .....Ao... ...JQOJ . . . ...! - ...: ...-o ...
.. ‘4‘. .A . .v . AA 0 . A. O. ‘I 1 . .
A . Ail, u. . .o . .Av... s... 1 5-... .. . _ a 0...... . it.” A. 'o .8? .... ..QA..\AA.CA...! ‘3! . t... .5 5.; . u. :0.- - KAI t. A. ... I. . 6A . . 2 ..A. A z . .A ..C . . . 5A.
A O A .- .IM. .to. Woo. r . .....V: 1 o .c ....I...rhlu¢oo§ 65.1.. a!” A... A; A... .o. .. (... 3.3... ......Oc..|.a .V... . A . . .1 c .AI.. .9? A. 9.3... I. . ... , .. -3. 13.0.. . A! . ..
1‘A..:. L!’ ... ... - ..Q A .2 k0...$¢ fit. A. if .A ... ... A. cf..|. o2L .11... ..r. 31.1.. .V ....A I... ......c‘... . 2 o ..A. t .270. 5 AAA. .0. T ...? A... -. :2. .
~91 0 ..A . lo. ......9 A) .v 71.7... .13... 931 -. 0.0.5.1.... ..-\.b 9 ....-.3.-. 09“- A‘...A\..“.. ‘xct‘uAS‘L .l .0813: . I no. ........A....A A. 2 he: a .. . . . .- .. o. .d A3131? 0A... . . .....
...... l— '1... .... ... a a; . Ll.'... ltd. 9A: 0. . .o :1. -1? 2-33,..83‘». ... ...... .( 49... :1 o ...: A . . 51......- , .A. A . .. Al. ... ‘l. . . o . . . A . . ... .. . .. A A .- . . . A
A. o... ‘n. m . 3..a.f.u.§.o ‘Ifi‘ IAI- . TA ....o .L op .‘J a I 3‘... .ur (A. - . A A. :' O’J..o., AIIv to: . . . .. . . ‘ .w .A. :AuuuA...‘ :- .A :9,- v A ...-3‘ . . . . . . . . . . . .
n.... 0 I. v.- ..I... I. A .A . 6.»? Co. .... .11.: (.20. 3.....- . . . . 3......0 . . o A. . u .. r .. . ...! .9 . .0.)O..I~1..,a... .. . . A . .
... O . 0.. o. n 4. A ... \ .A.. VWJ. ‘ .025. .. . A-ISAA‘ \ A 3" . A . . . A A. 3.3 I; ...3 t!..lo .... :- 9. .... ... . . .. A :0 .A. . A . ..933 u 9 A .A. . . . . L2... . I I .
u a ”LuflvAOL... u.- hi. ’.6 .. . $32.6? .....aoov. A...A .30.... 97. E»; 11%.... . it... . a . .... ..«AA Lr.s......o.-....o....o..{ .0 o It .A.! ....r-.. . ......‘0-1....1.... ...! ....3. A .Msz .. . o>v.\no..:&.2.¢ v.1..l‘unhtl-‘l. A
. . A - o. A .u A. A 6 .. o .o . A . . . o. .. A .... ... . . . A. .A - . 0‘:
a. d .J. «A 1. A} l a. . . v . .Oflflt 4“. . 1 ’55 n. ...J 4.. A A... no.4 ... o :- 3: A . 3...... . I. . .... . . .. . . h . . ‘< ..f... . ouAroIA > .I‘voaz. 3.0“..DU n v 0 v . . At:0.vuv.figi‘;.0. s
o- A. .F. to... .‘.ur ‘0‘v..T-c£c - r A.-993Ataon‘ . § A '9 . .A v. 3-... a ..l )‘n‘cdo ...... .v .5 ‘ A!
4 A .o .o .....“C‘ I. o. o v.-.4IA . ... Alo-Y. ...A. A ’0. I. oo~v a. o. o . . .. . ... .. ...? ...Q. ..0 .r- .. I; r. A . A. on. .9... .I .solo 0. . u.‘..¢o .. \rqlv‘oy 9-...‘0 1‘! a i ,
70. 0‘ w . f- . .f? 0 L17... .3650: ~ ... a 5);... $872.... to . .... ..‘\. . .. . ...-04: I... on . :33]... a! .....o. - .3. I (...-D. .4 A o‘$.0vl':“ Do
‘u o u (“A .. ...)” .../1‘. .16.... . .(i iv .. QY‘J‘ . .... A 3..-} ... . v. . ... y . . . . - ...... . 3A2. . ... o .7 ...AA‘; .... ...». A; ... .. . ;\‘$‘:A‘§.Yb 3!...
. ...-o. . _ A! At. . ‘1. .Yuotfi .. . E. I. M433 an} Y.J. C urn. . .11....“05. “AA. (I: . . . . i .. . . . o A. ...”..A. .u u. c. ...O. . . o v.3 .‘Ol _A.. c... a I.” Amo- «u A SCAN A.) .... ..oA . . 1. V 3 . .EAI:-.‘Ib u
- . . .....AA. :- 2A . ., ... ... v... .3: 9‘. o A. b- o It. ... o . ... .
‘o. .....r _ . F. :3. r“ .v All. ..DL A! it... . I «f. 33.9 10.2.9104 AI . . .- .. 3. . 9:91.» .. .. . 13 . ... A... ... I. . . .c . A... .1 .. I. f.. .1. ‘Q ......v‘thh .uovu ...-...: . ‘2”... 3.35... 9. . 0%; c
‘. “a 0...... 'A\ . All.) on 1“.qu 4 W! n. ....1 . t. .- - A‘A‘A‘I' ... ... . <..Av.fl.h)‘ol.v.o. 0.0 ... ...-3...... Ac. .8. . 4...." .0 D. 5.! 9’3... .. .. Jv‘A AYQ.’A.P. JASI“’CA . l I. O. A
I ‘4 Oar . _ ..\.c.r . LL}; 0“ . \ol 0’: 1-. .... 1 AA-Aioof; , . . . ...-...: 0.. :’ .olubfi-GA . 70.. \r «A A. ... .I. 3). . . if .. C Q. I .0 ...3"1‘o.\ I. 1'.- _ 2 a . .
A”. .0. .. A: . A-. mJ... .‘.‘.o_o A ‘ a . A . . .'L. 4 c . - ol.’ 1. . A. .. AA--. 62:. . . . . ...; A. . .A . 1- Av} . .vo¢-o..o.s. I. C:.. 0 A Ion. n .Y? .- ... an." - ..‘;', A... .. I. v .91 a . .0... . $192.0!
”V c .. 6... ....Y.. A .A v "T .. {...IAIAETQ . ..x. 103...! .- 0“. ... .. .... ...-A .... . . ... . . t... .0. :fn. .3. : A. .. o. p .. . . 9 A. 7 . ... . . 2.22.. A... I. A AA; 0' . ...-.2 ..A ..d! Aunt-v9 .
59".“. r¢oh.au.4.... ....vA. {if}. .. .tl .. . . r .. .... ...-...... .. ...-a]... .JIQA._ AL. ""9011 ..9. Von-.... 13!”. AM AT? A. .. h. A it!!! ...-9.: ...}! 0 LC. . . ...-Loi‘. . b n A . a... L VLifi’DS!
. . A .- .A . , J A .00. . A. ...-1.0). . . . you v . A. . .. 4 ... .A A . .6 :0 . a. I. a. o . A
. ...) .1.) . ... who-u: . -..: .A ...rrul . v:-....V, . . . ... a 9 :35... . . A . I... ...-..r 51.. 3 IS: a .. . .4 . .A.wo . . of 31.8.... A... . J Nuns“... 7. canAAoflvula. '1‘~.P:wA\50AuI¥\
W‘..4...o . .Vv“ .A. oHOr. 1015. .0 ‘0 4 rLD‘ .....A.A ‘- c“( ““350. ‘33:. . _ . . . .c . ...-no. t... . 49.51....m» do. .v. Au‘ _ ... ...l o "...“: on ,5 A.) A ' A'v~. .....u. . .31 ..O)... o. a" ‘ id“. to g
. 00A . . .0. .o- J o u . -. . A .. o . - ..... - o. . . A. .0 .5 f ’0 ‘0' ’3‘;
'd..AAA.. <L$w UK. A .. .A-x o. 2.1:. . YA: . ...... . . .. - iv. J. o A. I; A .- A .. :o a... ... ... . u 9) clot-(tip... .. 9.... . A ..k' 00. A 3. "Auk-1 4. out ...! 1|." )3 .
.AA A . A . v, A ...! . c i... 3.3. 1.. . .....OA . ..von. .... _ .A I! 3... . . A l A314...IC - 5“” ..hi.
".... ..«A ...th-JAKAp-u-G. . .- ._.. J. .....L...; 3.... ..."...A- ...... ks“... ......“ ... .....; .A . A A :........-.....-z..:-.A.: . ...»..Lbns
. . .[A - .. . A . .A . . (A. ....- . .... A. j . - . .l. o ..~ .. . . I. r .. .. I .QO... . . .
. ..ID Vs l‘v‘ .n ... fi‘tlaf .....F 4< . 9' ‘04. 2:9... .14 .v§.- 0:) ‘v. A ~' . .A 1‘ g 0. A. .. Avovl..lnn..-.-:...u’- .A A. i! a n. ‘m .A 5:. o- .o. ....A.o..}..rn;a‘o AA I’M: ....A 9 . . .- . A l a .“MOH’
T‘s '. I 0 . c .16. ... 23... a In... “...-1. A ... . .I.... .. . ..w...d.o. ..A? .... ...- A . ......)J. . cu... ._ -..... . .AA.‘_. .. Aer...‘ 7... AAA. ...... . .... . . ‘. 0041‘.
.A . .IoA... no. . ..A ..A‘A.... Lo .1: t, ... . . v: A. . . A . ‘A ......It!.. 8 A... ..Au.n......}. ... .. {.22 ‘1...» .... ...-...: 0.... . .l. . . . .. A o 3... . .... A. A....... l. .... utm‘ ...... .33.... 3,216. . . .A .
(1.... =0 ... ,4 ..A 1.13.... ..9: ..V . - .vlo. .c . .. .0. .... .... ' .... 5. 1!?! _ O .9~ A. . ... . . .. . .. i...l-.a.:i... L 0 . O. . .- .1. . A . Is. . . ..1 ...... ..-.-.. ...-AA. La -. A . A .1 go“. ‘. g
at, o urflth . ... 01..- . - : ‘7 .\.qr.oa ...... .- . _. I 0 .~ 701:. ..‘a‘. . . . ...: A ...: .A...‘ 0‘. “An" .- .I .. s‘Nohi.0..r to. 39.3.1... ..5. ..X’ . . .33297‘c... ._ . o O t‘.AQ'OA3‘-lll-‘Iu .o-"A a»
. o. o... I . . .. A. I . q n o .. u. 0 . A A . ‘5- 3... f
.v. A. LO‘ " “.... vhf." . w . .d§_.un4.v’so‘.w'~1¢A.-.A|: . , . .....o. '..J. o. O. ._ .. . c .... . ... .l A .....t." ..n.ln..u”u.r......ba. . .lq‘9(lnuosd.\}.'tl¥%o o.
3. f . A .00? . .v... . p“ . . . .0...“ o. . a. ... . ...1 . . , AAA . .0. \o 0:3.“ u... ... v. ,5 . u . . . , c .0 To 9. .6 .. 9A1. 1‘ IL?“ A -.hv ,‘bOCE ‘3‘.
It. 0. . J A. - I H’ .‘1... ...k‘... . .l 01.. A 2.3:... ..A . .. . . ..OA ull....o‘.:...9.. . . ......c . . IA: .v . f i . ..‘A In. a . .3... . 0 v. C :1 A, M... Asa?" Av.
15.. A We... ,AA‘. 0..¢ol..r ... .a . -.. O )A v a I L. . .....i a .... AA. ... 21.1. (:2, 7.2.5.2.. tfl ...... .... ...I .3. ... . . 5...... 0-. ......A91 A «to. . 3... iii!” firtoA-L
. A. ”no. I1 CA .. .tAAI.‘ V .1 . A . =1)... 03”.... (...-....S. - ~\ .u ... §..AA.A.. . A ,. . Adv. . A. AI: .... ... ... pp. A... AA! 6. ...2\... 2.15.0303 1.. : . t .
3.... A. . 1 b . VIA . A J r M . A A. A. V. to. ._ ... . L..o~1\.‘o3o.o A. .u f ...... 1. Au ‘0. .9... ...L 1....04I. p i0 .. A . . .t A. . .. . . - . v... . Po... .I-t. 4 1111A. I DIG-‘3‘I-O‘u’9ivial
..r L‘. 1k... a... A.. 3. 5". 0.0. . . .A c . A . . . ..- hAA .0- ... ...! 4:..Axo... A.s.r.. A... .I.I. v . 0. It. A lot-I . . . . . .....1 . 2 0.. v .. . . . A .. . .32... I... ..{I . . (”is 3": O”.h"‘l.il
1a]. . v.5 A ‘HNP AA A 0.4, .3. .. o. A . .\ “.~.:~U.i:..b.. o ...-3‘ .....‘Vs. . f A.... .... .. lol. . .u 3‘60”: . . 00: In 3...... . ..u u \...o.!...§..-. .. . A . . . _ . . lln(7.: ... JV a; c. " 7‘1“»! . D A1} ’51 on
. . . .A . k A Q... o .A ”.A6 . l3.v!._- . ..Koi... .X...cv..-.Y. .1. .... . . .. 91A . - 192...... 'f‘... .. I. ’36.. ...-A; A A. n. .. .. . . . .. . . 1A....Z A .. . ....... . . . ... _
.4. ....A'. 90.10“" find.“ A. vQL. W. As. I—WA. «.33 .3. - . . .I‘SJCVWQ . .‘m V! ..J .. .90. A . VI...“ ...-3.0.3": .11. to 0.... i... In. It: lull...» 0.. fl... A.. . a A. ”A: och..\L ... I A A.0.le~ .- 0. . . . .... .l' v... A. 6 . A . _. . . . . . . . . . . v. 0"”...."C‘9bu’-n.dd."1wv.’1.lvi3;lfl"s fhfifihnfiu’bo Mng‘?
. o . ... .. - I . . o. .. .41... ‘AO... A... .30. ..5 ...-i... .Io Z-l... .- - .. v... x- a ..I xbrA v . 0..- co. . . . . ..A ...... Irv... . o o. ... c. . . ... . o. . . . ._ .- A .0 A .. A
'7 A ...! A .0! p Ir. "J..- . .0)! . P... . ..19 a, .19.. 3A. \5‘.“ . . n I .... 1...... QA.'.A: .OAI...0 o ...!” ‘A. 3. A 1|. .181) ...t .0... . it .11... .... A I ...l t 1-- ....-. .v.....:. o, .l r . .... ..A . . ...bAl ... A... .- .. A .AIAPA... “Hutu-......“ W...” ”p‘uuu‘nnu .‘ln'Anaun'Jm ..dfll”..u“i’oi;tr“t
321...}... ... -...: ....r ...-.- .. ....-A. 52.93 .35. n..... . J.- .nu... . ... ....5-.n.1hnu...._ .. . ... ...-... $1...." .... .1.- .. .... A..." ... . A... A... - . . ._ - .. .A...-...x...i..z.§, .,.!-.....-....-.<A...L. xxx-AI. ...snya..1::AiAA.A1fi...-.A.Aa.-A.Ah
. .. D A. .. .A . s : A r o A O. . .v. I. . - -I .... .. U .5 . oi - .o . u . A . . A I-. . 5v . 0.! t... .- . . V c. ' s c .... . o. o .0. o .o A A . 3 .A .l. ‘oc A . v . . I . .
. thou . v A a . . '.c . Ob. 4.31... A.) ....»orc/ L993}: .-«.A<‘A.~A . v Q A 1!... . 'o .. ... ‘. owi. 3o. ... .0. - 5.- _-. . . ... 300AJI? . A .v Q .... 4. ‘i ... 1 A AA ...-1...: .9 ...... 2. A A . . . . . . . . ... . .. A c0.\ . 88.... IN“. . “US... “.9 .3t...\0<§.h1".. .Ialvo Nth-In...” Jltufio‘l A}. $it$vbi§t.‘lylo
. .0 .~.A[ . “0.07. .l I. A. .. .0!- . all 9 ....- . 47A. .. A u... . . ‘A. . u . .. . . .A ... ..fi. ... .1. . IA. - A O .3. ... . .. i Q . .... . ... nu. . pl. . no. . L .- .. . A . .. .. A. o‘ . o 3 o o. . .... : .é.... ..w ‘AAAO 0....."AI3-‘1‘52-A‘I. . 0.! .3”... “A“... on»? ..Ia . ‘..'\O.AA.‘
.LAACrH. 6:!“ r1! .. a“. .AIA! .DAA. . .31.... ... .A . 19.]. AH. ....)- AQV YAA ..uros. ... . _' ...-...”... ....” . 9..."... . ..A .I .. . o.-'O..o..: . A . ... A - 0.89“? ... o ..l. 8‘... . 1,: p 3. . A. . A. t . .. A. A .. ... . . .... . 3...: 239500....A0. . A A . \A}..Qu9« a” 39“..-“ IN.‘ 4M3,- 9IO¥§ilt IAVNWI
. n A .A ...L‘o ... . I. .... ' Ar'.’ 0. 0*- v. .A- .0. .9 In.) .. v: .. I .. A. . .. o A; .A. c \ ... .l... . c AA ...... - . A . .. v . u ...? p... .. .. ... . _. ... .. h . . . , .o .‘ A: a. .. A . . A. . .. n - 0A.: A. .. .. . . V I .. .K ... ..u. t ’
€0.Q..~.—Amtv...01..-. . . 0:. azflAl... - 0.. 1.0.3.3.. . ....VI «...-Y 2.1.-. BA .53.“. . A . z. . o... r? ...-.9... .v . .2 . . . . .. v . .... 1A.. .A A .... ..v: .... .Ar... .... . Ar A . . .. 33...; AA. A... ... .... .A . . .. : A A A . . . . 1 A. A. A .A. ...... -..... . .. ...... . thug-HAW“...- .A. JAG A. H...» “DJ-...! Ania it’s)!!!
.o \n_.A| AA .‘...I‘O. . o. no I ‘1 A . . T (v.7! (. O .. a o 0.... A. .... . .... ...-...... . Agni-:1... . A A.. A e - . . .. ... . . .. 0.. ... . ..AA. _. A v .....T... .v .. . . ...s‘... v} ... . _. . .. . I: ...-f... . . . ..3...‘ t. A. .- I... 3' ! ‘ilrgi
. A .o A ). .9“ Alv p53... .3. .- t r. b'! QAA. L o. .I' . .‘hl. . . . X3... ....o ...: . . .A A... . o ... ... Q... . A\. ...AA .... . . 9. V. . .1 . ....A\ A...v~A.s A $4 0.0:“. A ..--o . A . on AVA-fizlici‘il- .
. ... OI I“ .A . h . ...iugv Cu .. ‘. A 00. . I"... o... 9:... O. ... . 0" . ...l .. O ... ... v. .. : .. . v ... .... .o . A 3. 0.0... . . . 9""... .A in .....IA... . . . .VHAIAAQ A. yeti-O. 6... .{115 . Q.‘O‘§‘2‘I I‘.‘
o .oo'. L‘ 5 A v. ' 3‘ (-..IO aQA '10! 00.. ..l v. ...... It? pl 9‘0. 0. A. .‘ .. .I . 0 0'. DA A .. .... .v. ...o .... u o a o .. ... .. . o . v .. . .s . . . ' AA... ’2 . 3. 'LI ...“.H-xuhl‘uf Ii... loqo‘c!1’m.\.uu’|‘oAA-‘A Vfi A
..lva'.‘ onA‘quho - . ,yo. 3; . 1‘; ..~ . . o o . 1‘. ... . 0,. Lc.‘ ... A n Dr .0. .. ..V c A. . o . . ... 1.....9‘. - A. o- I ... A .L .cAA. o. A . Q ...A |.. o 5“. ..oA . C At, 9.... A I130. «1".Vluh dorfio’ibifinn’. ’0
.Arw » A. .8: 19.6.82 .00. A ...... .. .1633: 0 . - .. 'UJH .. .ud 4010.. c.- . .V.» ”I ... . ..c 35.. 1A... .... ...... A! u p... .. A ”A . ~ . A... 2.. .A . ....-...._ 0...”. . . A. 2-1313. -..OA‘uiozioAiu: :19»... ”z ’{rAA'I‘pnl’iA ,A
v A ..|.. ....I. \- . . - . .3? .0. 7 X ... o. I . 3... o .o .. .. A. . A I A - . 1.. A- 9. .. .. .... I :A‘1A‘Ao..... ‘A; I o .0. . o. D}-
As... .1?......-......- ...-2 9.3.3.-.. . .. . ..f ... - .. A. .. .........-.A_... .. -, . .1 - .-.... :A: .. .. . ....- A. . - A . .1... 2.3.. ..- .: A Anibfi-ui «km-n.1, A ...-3.1.
vA I. . I.. . . . A . . A o . .n I pro. . . A .. .....Y ...-0.. . .... . A. . .. .A . ...... 0.. . .. v! .. . tl. .v .. . . . .
...- .. 1.. .A uvr.fl.0 ... 2"... 4. P- "05.77.. or . 4.." . .JV ....-." . A .o. . . . .. . v. 1.. K... ...-o 1' 3 . -'oh.. ‘0 I. ... Joe. .5 . ..p. _ A . 1... . ... .u c. 3 . .- wont-obnlpoil11 V bun u ‘0 a. ot'I‘. ta.‘0~§|’o~.1. . ‘.l‘v‘ ‘I it)
....NAJ’ “I . 6.0..“ 0;»: ... . «rt 2.- _ o A»... AA . . .. A .I. ... a} ...-o A . . :3. .. . A . A . .. .. o . A .AA. A .. . .A .. it... . . t . .. :9 A $390... .A’ AAIAr...olIa-. b h: 3- Olin-‘1‘ ‘I 'l.
. I. A...‘ r. . , ...: A .A i F, . ... .. .... J...: In; ...! ..r.r . 9.1!: J _. . . .... 1.... ... ... 1A.. .... I53... ...... A. . .. . . . .... :59... . 8?: . AA 1.. .. Q... ... ._ ... ¢. .. .A . \. ‘52. fl... .21... 3: o 30.1.1)...‘ L .- . i.‘ i‘t‘o'! AI 03.!
A A.>.no-.c..‘.> '“Ofi‘s ” vv ,.A‘A1.O. ,.-l..A .0.- . .. a .. A A r .. .A ...r I... a U A 056. I I . . .. . .‘ AA _‘ _ v . _. - . . to... 3 .4... .. . ...s. d . ..- o ._ . A - .I? .v ~ ‘1 so ...I. .. VAN—V. . At «AU... ou§ 6- A. 1 ...-.... 0". ‘01-."0..C‘HPIA‘.I.’\I‘ .
. $.rv VA. , . . . 0“ ....AV .c\ . .0. 5-“..Aflf ...... J}. . 3.5.3.. -1L . 5-. t..“n..lo . J 3 ... -. C I. A ..l \ ...! o. . .. . AI . . AA... ... ..I ‘ ... , . AA ..A 5......» 3.4. ‘30. ...-Q IVS-Al. A5‘v-u nan"?- fic’i it... '5‘9‘1? A.
7. r 3%- ...n..--...! .4 AV: 1‘! 1.. v. .9 AAA .—..II . . l . . AA- A ... .. .-...A... A. f t... . .. 1'. .. .. ..A A. .. 9.. . . . . sq, ... ... .. .. ..v. .1.AA. . . . . n.(. .A A V‘A.$.!..C'!§O 38 .)A.”1".’.~.lt£!.la(a$ I II
a. o _ 0. .Id avvoA 3-13: ....ioO. 32... c.. 1.3.... . .00 I. A . ... 3 .. .... -.A. l.. . Av . I. o... .... . .. I: . ... . A . v .0 v. 3.1.... A I1!Au‘.\..lc -'!¢.‘O.A" 9' A ”0" O! I
..‘P. . . .3. 56‘... .9... . ...in .A- v). ...u ..1. C v ... 3.. $329.: ., ..Q; I. .....I. A A . A . o . -. .... A c . .. .....A o. . .v .0 . . .. ...: ..o A 10. o . ... A1... .onl!QQ 3 tttz.ht.o .I taint-ht}. ! 44.2 ...-Avie.“ .0“
+0"... .0# f3 I. I.’ 1.. .o 7.. Iv 5‘ ‘.1|.:‘< (:78...- o. . . . A. no; If. A .r ... . . .o. . . e. . .o . .A . a. . . o . . . . . . v ....r. v .A .o ... .00: A. {(.O‘t....l‘\.oA"-¢¢.o ‘3 «A. nufl‘zq ....O D. 00‘”
c.1304]. VA .- ...:f} I I. . .b .. ...... . I. u. u .1. ... ...... .....o... . . . 0’ o .0. u I 4.. A to... o I . A u o ‘A ub..| . .... c u . ,0 n n . . .. . .. ..Al . 91‘ Cr. -01 5‘. ol ‘0‘, A .0. ‘K It.-. o‘gpiggh
AL. .0 lunar” ”v.14. A._A....l. ... o .. A LP‘.Q~.A..uJ .. . . . A. ._o. A: 5.... - Z ... ... o ......J... ......i... A! .u o. 111v .A1 . . A o .. A .... . . A .03 v. s . A A v .. ’A . A.- 9.’ .1 ~ 0-. ' A? A1 ...'| .. A§ ...lo.AA.v. .....ICDNAAHJHWIJVou-o... . h
. . A}. .I. «9.... . t... .. 0‘, ... ...... '2‘ 3 .o o. .A.. .. .... ...-A. .l A... . .. . AZ. .- .... . v A u . I . é.
— u .2» . 3......” onus..- . ..MA. ... Iv.o.. U, Q. A “A: ”on . ...; .r... 5.. .AO. 0| 1 .1 PA .. . . . h?" . ._ .l ..- . A. .... .. .S _. . . .3 . A. A! . ...- "!an d... I. o.u¢.o~...fl!1 ..h.-.~..A..¢ru1h!¢6 AL...
A. . \051 . or A“ w. 0. .kvVQML\AO.. . .r. . .4“: c 1“. A; o.r ‘ .. 1.. A. . A Ali... . ...... . 0.151 AQA. . . . . s .A .»\| . A. - v. .... A a! ‘5 n ..n ... A V. .1 1 Ag..vh.¢9 N..¥‘av."-..A‘..OA4 A
.5“ ’6’} on AL... V . . ......1L1. ....v ... o. ...l: \_AI...o . . A. .. A . . at... AA. .. ... .79 A 0 ...... . o: , .o . . . A. . .o - A A v! A Q . ...- ... 2' ... ..A-....:.vIbo.c ... .... Pl"$‘..t I *o‘bt
. to. I. 70... U . s ...-.1... . . TI. 3 A .99. .0 a ..O ... . ... A .A .lA-oootiv .. . .9 .c o I... .. ....A ....l A‘ ~ . ‘ A. "A L A I... to. D. . S: A». v . - ,7. w . A I Q . .A. ... ~ AAQS..S 9“! .5900- ‘«‘b. ‘VA
49.7.. “~AIAO AA, ... .t .1... ......t | .3. A 9 ...-A»! .3 O. . A ... . . . o ...?) c. ..7.‘ .. A 4 . .. .c .. . ... .Il.r Ah .5 . g o . .A . o. w. A. u .. . . r .... .. .. . . A.. .. . 3 9... . {'355-ZA. . a”. "(1., L...
Q.’ . orb-{O A A. 9.1. . .. . .1. .A ...-D .v - :13... 2 .... ~ 9.. Isl. .Sool . A -. .... . .A‘. . . A . . . . . . . .... . ... 0 . . . . ... .0. . . .6! .l 6 A .3. 3.5 AOA 1? .1...- onfln: A, .3 ...-“l . . Qli
....1...._0 3.x 4‘ O . A.‘ ....r . (3.1 . A! 3.... .5. ...! ... It... .. JI A...AA:. v A ,f 2 X...VA .1... .. luA .. ... . . . . iv... . . . A .. . . .. 4630... .qa'x . .1... ....tho . y’L.Qh.-A-O.
A 1‘7”‘ .u I. I... . $.12. . - .. . .. .03. . .A .. .. .. . . £1 ‘ A . 0. ~ .A . ...!‘o. v. Os. . t... .t . .... A u . . _ A nu... I a. 0.. ‘A .- ... 0.. r . . - . ._ A a . .. . A. A . . . . . .. Au: .' ... A” A... Al,“- oAi-u. A. gnu.“
v t. by) A... - fl 0 . A .. A ‘0 I c- . . ..A . A. v ._ ... .1... .' . -. Q. . AA .~ .. o. 3...}. ...-,1... o A . . - . :I. .l a 9. . .0 A o A A. ... . . - . . ul...... A.‘ . . 0. .\5V ’A... A | I \ 1. '.
.4oouv|.Yv b. O o . 0.. .C 1.... I Lg... . ..v a V.... 11.... . A . .- A.lr. A. A ~»c. A. I At. . .. VA ... A o. 0.... - A .1 4 A u. ... ...’A ‘.v I? c¢ .c O o p. . .1. o .v .v ..... . IQ... .... .0. - . .. . . . . AAA... ‘. ..i'A-l‘l‘. A I. A A” no .- I . 4“.\0
. .u v . A; . A.-.. a. .... . ..- ..... \.:.... A AA . A ..‘51 .AA. . . .- .. ...-A. .. 3‘ AI A .o . v .. . . . A. . . .5. A . . I: . w... 2 . . . . .A. : . . ...: A: A ... ...... .. ...-A: HAS-...: «n Y. ...... AD...” 1.»
O'A'. ,3 n. t ’1. n.’ # . A O .fi7‘. . I A... I. . _._ - . 1‘ . A) O .A o.) l. .. .. 11‘ 5.. . a 9A. A o .‘ .... . I! ...? A... . .A .l. o 3.; 3c .0 A... .. *0. .... ... ... >5 u A 0 ...Z A. u c. v... .... ... A.Aa .... «3' 3-3:. 9A. . CIA. . -...CII. fl. «Knot-0k}. A A!
..c... . . l ... .l' ‘31 .. ..A ..v..-A.-A .v... A A. .... A 3 Al...’ v .‘oou In .4...IA.. . .. <‘A 5- 0. ...... ... o . .0 A l - A .0 A. . ... ... . - ... . .. .. #2.. A I . , . 5 f A. ... ...A 7‘. .. O. 5‘8 II.‘ .....s...-‘.Alu.o.l\ .... .0. ‘C" 3‘"?
"L . as. A4 . A L‘AF. A .1 .A.f...\-u: A13... . .. ... A A... . - 7%.: A ... ... . . . :1... 1 I. ..o. 3.....luo ... 1-1.3.3. . . .. .... ..voo. . ... .A, A ...A . ... . u. - fl... ...-... A....... ..A 9... .. . .. .. 3.. .. ...... . .... _ ...-Io}... vt... «A A . S..- A: It... 3 .9}. 1! ......
.51 “a... “IQ“...V q... 169.. A o .l ...A .4! A a... .\’ . A Iool A on. ..., .... ‘KA I. O. .... .. y A ... . .. \ . 2... ... .. .30.... . ..A. .. .... ...o u... ..£~.A v A A. .. .. 10.. iii; . ...-’ovaoOAQ . 7.. {IA ..‘AO. .
.‘Q ..A ... on Am5»: buy-fl..- . “A... .l I V. .I c . . .. {Aw . ..v u .. I .c . A... . .p . .. . . . I. . 3 A ... ..I .... ... 7. 99.13.? A . on. . . _ Ah... .. . .. .p o l. .. L .. u I. .a . . o9 - .. .o . .1 . . .... .‘ A . . . ...—"flout h!‘.~ crotnuvg\
' .VALAALA __‘.I DOME. .1 .0. .u.. v. .I «‘Uuoat _ i .....IJ ...“. 20-. I .0. r ..l 0 -c A: IA 6‘ . AA. - o : A. . A b I .. . .... .. .. a. A . a.... A . v. .u I... A. .‘. v ..AI.I.~AA o o . o. l .. AA... A A .. . A . A. ._ .v 4 "I‘O.AO-'A..’.r \A. A ... ... . .Anr:
A f..£.fi. :1. 'I .1 .v.‘.:. .‘VI.’ 6 o . A 3 A A s.“ o . .. 0A.. ... o 1.1. .. VA ole-......” . A'. .. s. A. . ..A ‘nl... .... o . .. v. . . 3A.. .. A. .. . . A . v . .90 >r|....c .1. ... I... ...I. . . A A: ...l . . . 1:.0 9.. 19.,- A s IVY o 4“.
_ .HAL- .. .. A A; A . A a 1 - \A - . I. II. . . - . o . . A. to. AA .. . -. . .‘f .. . IA ..- A . . . . A- . . . . A . A. A
A ....r. A... 5.0.. F... .fl- ‘.:A ..H ... ... .0 . . - .rfiul o ’5 1 o . A o o .A l. o A . 0- o g . o w.‘ v IAV O. A .-0 Ar . .0 'A .9 o 8 A. . A. . .. 'A . V AA .0” . . ... . v . . o. . I Q .o .s .A ‘0‘! . .30.].‘u-u . T oduflo Inc-.0 IUD}. IPADJoAOIrAcfavA I.HC"A’0
....n..._,.:...h...- .A.. ._ ..-..Z....:. A .. 3.1.... -. .. .. ...-...- : ...r .A . .A -.... . -... -. .. ... . . . .. . 1...- .. . -. .- -... .. ...-2.1:... -..- -.u.....A-.y-.A
.. v a.“ c A... C.- An’. Ant pt): .0. Q A _. . . . .A.‘ . 0‘ .. s...» [As A. 3 .1 A . .....x A .A A 03.. A . .A 0.. A - o . .... 03...... o r? o.‘......1 . . . . c‘. .- s. . . A. . . 0.. 0 A .AA A CI. . . A. a? 2’0. A o .4..\. Am. . . A. u . t§L-Lh.AM.A a
u A U0 A H .D . :6. ..r q .... 0!: ...-11" A; A»- A,6: . 1.. .... . .. v2.3.1... v0. . .. . A [.1 to 9. A . . . l A .. .. . A ...-0.5 .. .. A ... . . .Q A. c . . . . . . . A. ...: a A ...... n3...- \.9 .. ...". A4...
‘ 03.51 .. \M A .L‘. .0. Q .AJQO.OQ.. ‘JAA..HQO . . .I .AI. ....lAA.... . . . IO. .0' cl A. coo.t..«\.l..| A. a. V I . at! o o. . . '0‘. .0. A to. ... ob. .. (.09... O oat... Z a . .a A A.'.I-ho.'§ 1,.
st 0 I A‘- .A.. .0... ..AAAA. .11 .Al I. .. .fi ...~ A .-....3. .v-A! Ao ._ .. .o . - A}... . \J»...A ....Ao’.‘ Al A .A ..t. .. - A AA. .AAA , . A ..n ~. .. a o b. .9! ...!- ..O..hv. .IFQ. 0.4.“ 1.3.5:...an
5 O .. A AA... . o. ..I A A . . v .O-A . .ol . .- 3... .- p A. .. p .u 7 . 2 A .o A . ..0 1-3. A. In . . . .v on . u a A A. u . .. A. s . . c ... .0 A . 5'.O A. A... 0 K, .1. . :03... 7!.‘3 A
, . . D. ..1 .A .. .v d .... A a. 0.. .‘ r. o A ..v. A .‘L t ... can .A o A o I 0A0 - i. .. . I . -. . Y Iz.o'. v‘ . 0A.]; v'Antv‘. ntv’. on: ‘0
.0 . . 0. o . PIA . .A - . . .o - ...A v A O tn... . A .v o A ..':AA .Ov. . A. x '0 o. ... :IA9..0! .13....
. .A. A . A . . .. A.. o. I u I .0 I no. A. . .P AI .. ... c A. 6.7V..I.O.Ovd... ...... A-ooAA‘iAIIU.
- .. ‘- .oA . ... ratvo A A ...V U A ..u... .A. ..5‘905.1A 5:?- .I'II.O.A-I‘IIO-
0: . A .A .... A u. A. 0A.! ....2. m-‘AAOAAAA-aol.‘I.-‘\VA It. A
o . 0.009 '0 ..s.‘ ’0’) O A. .AIIAAQ..~¢
- I . A . A v I?! A
r , _. .... . .. _
. .... a S . A. A . 1 o n A .
A. A ...Hxa- a. A Era-«wt. z. ... . w A. A . L... A A- A .. - . . .... _ . . A .1. A.
.. . . . .... ... .,-\..... . .., . A I
up”... .a' .u 6.39"! 0 ~ vlffiowt ff fo- Q 54”. .fi.n-A.&. “raviqfif. Jvhmtrvox ANAL.....AA~H . ..KT. .4 ...-..s...»:....... A .. .. A .(J~.1_..A 1.... . 0.. ‘AIAJ'VJA... «d ..\.\a . ' \
. . .... .. .. 2.. .51.... A.. 1- .. 7 .A.. ...,A.A..... ... .....:.. .A,....1.....-r._...r Aid.
A o . VA. A .....A. A. r. A.t71.. ... ... . u ..JH. .....k ..A.. ...L r352. .... .A: A . ...t 5a... .3. .. 3.....Af..-c.... 7...: . n d...“ ...
. -1 . . A ... . . I A q. A ...r L 2 .At....‘.\.i ... ., xi . A V. 27" ... . Fat .... r
I . .
A. | All .. v0. .\ ~1t.‘-...¢1 ... o -|-. . v \.
.III. 'I' Ill 0 . .o A .
..III E
III ' -

 

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

NANOPARTICLE ASSEMBLY IN POLYMER BASED SOLAR
CELLS

presented by

Jonathan W Kiel

has been accepted towards fulfillment
of the requirements for the

Ph D degree In Chemical Engineerinl

 

T 14M

Major Professor’ 3 Signature

Ol/w/ZOIO

Date

 

MSU is an Affirmative ActiorVEquaI Opportunity Employer

-A_l-A-A-;-0-0-C-0-0-!-I-I'l-I-l-I-.-.-O-O-o--t-l-I-I-I_l-o-I-.-

 

on-I-o--n-o-o-n-u-I-I-I-n-n-.--—»— —- —

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProjIAoc&PresJClRC/DateDue.indd

NANOPARTICLE ASSEMBLY IN POLYMER BASED SOLAR CELLS
By

Jonathan W Kiel

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Chemical Engineering

2010

ABSTRACT
NANOPARTICLE ASSEMBLY IN POLYMER BASED SOLAR CELLS
By
Jonathan W Kiel

Polymer solar cells have become an intense focus of research due to their promise as a
lightweight, flexible and inexpensive alternative to traditional, inorganic solar cells. A
typical polymer solar cell consists of a light-absorbing polymer and electron-accepting
fullerene derivative, both of which have poor electrical transport properties compared to
inorganic materials. Due to the poor electrical properties the performance of these
devices is dictated by their nanoscale morphology. In this work we report a detailed
characterization study of the internal structure of a traditional poly(3—hexylthophene)
(P3HT) : [6,6]-phenyl-C61-butyric acid methyl ester (PCBM) solar cell.

Phase Sensitive Neutron Reflectivity (PSNR) was used to map the vertical profile
of PCBM through the active layer of a solar cell. From these results we found that
directly after fabrication a large amount of PCBM agglomerates at the substrate of a
device and near, but not at the metal electrode; an opposite structure than desired.
Thermal annealing caused some of the PCBM to migrate to the metal electrode,
improving device performance.

Small Angle Neutron Scattering was used to determine the size of PCBM clusters
both before and afier thermal annealingr It is widely known that thermal annealing
improves device performance but annealing also increases total device

photoluminescence, a competing efl‘ect. We show that upon annealing the PCBM

particles cluster together creating better pathways for electron conduction but this
clustering creates larger void space for increased photoluminescence.

We have created nanorough substrates that enhance overall device performance.
These substrates increased the total interfacial area between the conducting oxide and
active layer and improved light harvesting within the devices. These effects acted
together to more than double the efficiency of a standard device.

Finally, an overview of nanoparticle assembly within polystyrene, poly(methyl
methacrylate) and poly(3-hexylthiophene) is presented. We demonstrate a substantial
amount of control over the assembly of nanoparticles at different interfaces within thin

films by changing either the stabilizing ligand of the nanoparticle or the polymer matrix.

ACKNOWLEDGEMENTS

This dissertation was produced over a period of 4 years and 10 months at two
universities, Michigan State University and the University of Delaware, with a great deal
or work being performed at the National Institute of Standards and Testing Center for
Neutron Research. With three institutions heavily involved in this work it is difficult to
adequately acknowledge everyone who has helped me as I completed this work.

My advisor, Michael Mackay, deserves the first acknowledgement. Beginning a
new area of research for the lab and moving universities has been difficult but rewarding.
Michael was there through all of it and I am a better scientist for having worked under
Michael for my PhD. Professor Phil Duxbury was like a second advisor for me and I
understand the basics of polymer solar cells because of him. Phil’s door was always open
and I am grateful to him for his help and guidance.

Dr. Brian Kirby and Dr. Chuck Majkrzak of NIST have been instrumental in a
great deal of the work presented in this dissertation and are two of the most enthusiastic
and helpful people I have met. I am lucky to have worked with them and learned from
them. All of the electron microscope images in this work are a result of Dr. Alicia
Pastor’s teaching and assistance. I am grateful for her help and friendship during my
years in graduate school.

I could not have completed graduate school without the help and guidance of all
of the Mackay lab members who aided me so much in my work. First and foremost Jon
Seppala and Erica Tseng. Jon, Erica and I began grad school together and moved to
Delaware together; I could not have asked for better fiiends or colleagues during this

journey. The former group members who taught me many of the techniques I used

iv

throughout my PhD; Melissa Yaklin, Tiffany Bohnsack, Dave Bohnsack, RS Krishnan
and Anish Tuteja deserve a great amount of thanks as do the newer group members who
brought enthusiasm at a much needed time; Brett Guarlnick, JJ Wie, Hao Shen and
Wenluan Zheng.

Without Bakhtyar Ali and Susan Huang at the University of Delaware I would not
have been able to make polymer solar cells nor understand much of the basics of solar
cell principles. I am honored to have worked with them and consider them fiiends.

My friends, both at MSU and UDel, are some of the most important people in my
life. Their friendship cannot be looked upon highly enough.

Finally, my family. My parents Bernie and Bill, sister Lynn, Brother-in-Law Joe,
niece Addison and, of course, Hannah, are hands down my favorite people in the world.

Their support and enthusiasm through this process has been amazing.

TABLE OF CONTENTS

 

 

 

 

LIST OF TABLES - - - -- - - - - - -- -- - - - ---IX
LIST OF FIGURES - - - - - - - - - ..... X
LIST OF SYMBOLS AND ABBREVIATIONS - - XIX
CHAPTER 1: INTRODUCTION -- 1
1.1 Motivation ................................................................................................................ 1
1.2 Thesis Outline .......................................................................................................... 5

CHAPTER 2: POLYMER BASED SOLAR CELLS: BACKGROUND AND

 

OPERATING PRINCIPLES 7
2.1 Introduction .............................................................................................................. 7
2.2 Background .............................................................................................................. 8
2.3 Solar Cell Operating Principles ............................................................................. l4
2. 3. 1 Equivalent Circuit Diagram ........................................................................... I 4
2.3.2 Testing Solar Cells .......................................................................................... I 5
2.3.3 Short Circuit Current ...................................................................................... 18
2.3.4 Open Circuit Voltage ...................................................................................... 20
2.3.5 Fill Factor ....................................................................................................... 21
2. 3. 6 Efi‘iciency ........................................................................................................ 22
2.4 Polymer Solar Cell Devices ................................................................................... 22

CHAPTER 3: NANOPARTICLE CONCENTRATION PROFILE IN POLYMER

 

BASED SOLAR CELLS 31
3.1 Introduction ............................................................................................................ 3 1
3.2 Neutron Refelctometry ........................................................................................... 37
3.2.1 Specular Neutron Reflectometry ..................................................................... 3 7
3.2.2 Phase Sensitive Neutron Reflectometry .......................................................... 38
3.3 Experimental Methods ........................................................................................... 39
3.3.1 Neutron Reflectometry Samples ...................................................................... 39
3.3.2 Phase Sensitive Neutron Reflectometry Measurements .................................. 41
3. 3. 3 Density Measurements .................................................................................... 43
3. 3. 4 Solar Cell Fabrication and Testing ................................................................ 43
3.4 Results and Discussions ......................................................................................... 44
3.4.1 Pure Component Films ................................................................................... 44
3.4.2 Solar Cell Mimics ........................................................................................... 48
3.5 Conclusions ............................................................................................................ 58

vi

CHAPTER 4: PCBM AGGLOMERATION IN POLYMER BASED SOLAR

 

CELLS 60
4.1 Introduction ............................................................................................................... 60
4.2 Experimental Section ............................................................................................. 73
4. 2. I Neutron Scattering .......................................................................................... 73
4. 2. 3 X -ray Scattering .............................................................................................. 74
4.2.4 Photoluminescence Measurements ................................................................. 75
4. 2.5 Solar Cells ...................................................................................................... 75

CHAPTER 5: CREATION OF ROUGH SURFACES TO ENHANCE SOLAR

 

 

CELL PERFORMANCE - - 76
5. 1 Introduction ............................................................................................................ 76
5.2 Rough Surface Creation ......................................................................................... 78
5.3 Solar Cells on Rough Surfaces .............................................................................. 81
5.4 Spectroscopic Measurements ................................................................................. 84
5.5 Discussion Section ................................................................................................. 90
5.6 Experimental Section ............................................................................................. 92
CHAPTER 6: NAN OPARTICLE ASSEMBLY IN THIN POLYMER FILMS ..... 93
6.1 Introduction ............................................................................................................ 93
6.2 Background ............................................................................................................ 94

6.2.2 Dewetting Inhibition ....................................................................................... 94

6. 2. 3 Nanoparticle Interfacial Assembly ................................................................. 97
6.3 Results and Discussions ....................................................................................... 100

6.3.1 Eflect of Changing Substrate ........................................................................ 100

6. 3. 2 Assembly of Higher Refi'active Index Nanoparticles .................................... 101

6. 3. 3 Nanoparticle assembly in P3HT ................................................................... 105
6.4 Conclusions .......................................................................................................... 1 10
6.5 Experimental ........................................................................................................ 1 11

6. 5. I Dewetting Experiments ................................................................................. I I I

6. 5. 2 Transmission Electron Microscopy .............................................................. I 12

6. 5. 3 Thermal Analysis .......................................................................................... I 14

6. 5. 4 Atomic Force Microscopy ............................................................................ I I 4
CHAPTER 7: CONCLUSIONS ....... 115
7.1 Future Work ......................................................................................................... 1 16
APPENDIX A: HOW TO MAKE POLYMER BASED SOLAR CELLS ............. 118
Al Patterned Slide Preperation .................................................................................. 118
A2 Slide Cleaning ...................................................................................................... 123
A3 PEDOTzPSS Application ..................................................................................... 123
A.4 Active Layer Application ..................................................................................... 125
A5 Electrode Deposition ............................................................................................ 126

vii

APPENDIX B: POLYMER FILM THICKNESSES -- ..... 129

 

B. l Polystyrene ........................................................................................................... 1 30
B.2 Poly(methyl methacrylate) ................................................................................... 13 1
B3 Poly(3-hexyl thiophene) ....................................................................................... 132
B4 Poly(dimethyl siloxane) ....................................................................................... 133
B .5 Polyisobutene ....................................................................................................... 1 34
APPENDIX C: IDEALITY FACTOR PLOTS - 135
CI P3HT: PCBM device on a pure PEDOT: PSS substrate with no LiF between the Al
and active layer. .............................................................................................................. 135
C2 P3HT:PCBM device on a PEDOTzPSS substrate with 10 mg/ml SiOz particles
dispersed within the PEDOTzPSS solution. .................................................................... 138
C3 P3HT:PCBM device on a PEDOTzPSS substrate with 20 mg/ml SiOz particles
dispersed within the PEDOTzPSS solution. .................................................................... 141
C4 P3HT:PCBM device on a PEDOTzPSS substrate with 40 mg/ml Si02 particles
dispersed within the PEDOTzPSS solution ..................................................................... 144
REFERENCES - - - - - - - 147

 

viii

LIST OF TABLES

Table 2.1: Basic parameters for solar cells as cast and annealed both with and without a 1
nm LiF layer between the Al electrode and the active layer ............................................. 24

Table 2.2: Series (R5) and Shunt (RSH) resistances of the four devices shown in Figure
2.8. Annealing and adding a layer of LiF both improve device performance, which is
indicated by a decrease in Rs. However, these processing condition changes do not seem
to affect RSH. ..................................................................................................................... 25

Table 3.1: PCBM volume percentages at the air interface, substrate interface and
averaged over a 10 nm depth at both interfaces for comparison to the XPS data of Xu et
al."'1 The two spin coating speeds, 800 and 2500 RPM, were used for comparison to the
Slow and Fast evaporation conditions used by Xu et al. The concentrations for the
sample prepared at 2500 RPM was determined by simultaneous fitting of the two profiles
(Sim. Fit) as well as with PSNR while the 800 RPM concentrations were determined
solely by simultaneous fitting. The data are from Figure. 3.5. ........................................ 55

Table 4.1: As cast and annealed solar cell properties and their relative
photoluminescence obtained by integrating both curves between the wavelengths studied
........................................................................................................................................... 63

Table 4.2: Comparison of PCBM agglomerate size and polydispersity, spacing and
P3 HT crystallite size. Fitting the neutron scattering data resulted in the concentration of
the PCBM agglomerates and the amount of PCBM in the surrounding polyrner-rich phase
(matrix). These concentrations were added together to get the total PCBM concentration
which, based on the solution concentration, should be 47%. ........................................... 68

Table 5.1: Solar cell properties of the devices shown in Figure 5.3. They symbols next
to each cell title correspond to the device J-V curves in Figure 5.3. ................................ 81

Table 5.2: Total PEDOTzPSS surface area increase and surface coverage change upon
addition of Si02 particles into a film of PEDOTzPSS. ..................................................... 84

LIST OF FIGURES

Figure 2.1: The molecular structure of PCBM and P3HT ............................................... 11

Figure 2.2: (a) The structure of a typical polymer based solar cell comprised of P3HT
and PCBM and (b) its ideal band diagram. The lithium fluoride layer is lefi out of the
band diagram because its conduction and valence bands (1.0 and 14 eV, respectively)20
do not match with either PCBM or A1. Much research has shown the addition of a thin
layer of LiF between the back contact of a cell and the active layer greatly improves
performance but the reason for this enhancement is still unknown. ................................. 13

Figure 2.3: Equivalent circuit diagram of an illuminated solar cell. ............................... 14
Figure 2.4: Solar spectrum for AM 0, Am 1.5 Global and AM 1.5 Direct sunlight ........ 16

Figure 2.5: Geometric description of how AM 0, AM 1.0 and AM 1.5 spectra impact
earth and the atmosphere. ................................................................................................. 17

Figure 2.6: Typical I-V curve obtained from testing a solar cell. Upon illumination a
solar cell is defined mainly by three major points from the I-V curve, the open circuit
voltage (Voc) the short circuit current (15c) and the fill factor (FF). ................................ 18

Figure 2.7: Theoretical maximum current density (15c) obtainable versus band gap.21 .. 19

Figure 2.8: Current Density versus Voltage plot for solar cells as cast and after annealing
for 20 minutes at 140°C. Closed symbols denote annealed devices and open symbols
denote as cast devices. The squares indicate a device was fabricated with a 1 nm layer of
LiF between the active layer and aluminum and the circles indicate no LiF was used. 23

Figure 2.9: Illuminated and dark currents of a typical P3HT:PCBM solar cell fabricated
with a 1 nm LiF layer between the Al electrode and active layer ..................................... 27

Figure 2.10: Derivative of current versus voltage (dJ/dV) of both the illuminated and
dark currents of a typical P3HT:PCBM solar cell fabricated with a 1 nm LiF layer
between the A1 electrode and active layer. The dark current shows good diode behavior
which is seen with the flat dJ/dV profile at low V but the illuminated device shows
substantial voltage dependant current values .................................................................... 28

Figure 2.11: Plot of dV/dJ versus J'1 to determine the ideality factor of an annealed
P3HT:PCBM solar cell with LiF as an interlayer. The main body of the plot displays all
the data obtained for values of J > O and plots the straight line at low values of J. The
inset shows a magnified view of the higher current region (i.e. far from Voc). The values
of Rs reported were obtained from each best fit line. They differ because RS is
determined from the dV/dJ intercept, which will change depending on the part of the
dataset chosen for a best fit line. ....................................................................................... 29

Figure 3.1: Idealized energy diagram of a P3HT:PCBM solar cell, showing ideal
alignment of the molecular orbitals and work functions, has the PCBM component of the
cell in contact with the aluminum electrode and the P3HT component in contact with the
hole conducting layer of PEDOTzPSS. ............................................................................. 33

Figure 3.2: Defocused transmission electron microscopy cross sectional images of
polymer samples prepared by ultramicrotomy shows little difference for a variety of
materials. A 1:1 by weight PCBM:P3HT blend (a) and pure polystyrene (b) both have
lighter and darker regions demonstrating the difficulty in interpreting such images. ...... 37

Figure 3.3: A cross sectional image of the wet cell holder used for Phase Sensitive
Neutron Reflectivity measurements. The path of the neutron beam enters through a
silicon ‘fronting’, encounters the film being studied and exits through to the ‘backing’
reservoir, which was either air or D20 for this experiment. ............................................. 42

Figure 3.4: (a) Reflectance (R) times wave vector (q) to the fourth power versus wave
vector for a pure P3HT film spin coated at 2500 RPMs with associated best fits and (b)
the SLD and corresponding density profile of the best fit lines. A varying density profile
(black line) fits the data much better than the constant density profile (gray line),
indicating the density of the P3HT film varies slightly as a function of depth. ............... 45

Figure 3.5: (a) Reflectance (R) times wave vector (q) to the fourth power versus wave
vector for a pure P3HT film spin coated at 2500 RPMs and annealed at 140°C for 20
minutes with associated best fits and (b) the SLD and corresponding density profile of the
best fit lines. A varying density profile (black line) fits the data much better than the
constant density profile (gray line), indicating the density of the P3HT film varies slightly
as a function of depth. A noticeable increase in density is seen at the substrate. ............ 46

Figure 3.6: Reflectance (R) times wave vector (Q) to the fourth power versus wave
vector for pure PCBM. The solid dark line is the best fit line to the data, corresponding
to a density of 1.25 g/cc and the grey dashed line is the calculated reflectance for a
PCBM film with a density of 1.5 g/cc. ............................................................................. 47

xi

Figure 3.7: Neutron reflectivity uses the natural contrast between P3HT and PCBM to
determine the scattering length density (SLD) and eventually the concentration profile.
(a) Reflectivity multiplied by wave vector to the fourth power (RQ’) versus wave vector
for the sample spin coated at 2500 RPM using either an air (triangles) or D20 (circles)
backing and associated best fits found though simultaneous fitting of both datasets. The
deuterium allows characterization to larger wave vector and more accurate determination
of the SLD profile. (b) The best fits of the reflectivity data for the SLD profile were
found via the PSNR calculations as well as simultaneous fitting of both data sets (Sim.
Fit-Air & Sim. Fit D20) and the Parratt formalism for the air backed sample (Parratt Fit-
Air). An expansion of the distance axis shows (c) the SLD near the substrate (vertical
dotted line) where the SLD bump due to silicon dioxide in the simultaneous fit is clearly
visible and (d) the SLD near the air interface (vertical dotted lines) which isquite similar
to the SLD of pure P3HT (horizontal dotted line) within the first 2 nm. The sample was
unannealed and characterized directly after spin coating. ................................................ 50

Figure 3.8: (3) Composite reflectivity data sets obtained for the Slow Grown film and fits
which were determined simultaneously and (b) the Slow Grown film SLD profile
obtained by simultaneous fitting of the corresponding data. As the SLD profiles were
simultaneously fitted the profiles overlap until the backing layer was reached and they
differ with the air backing film going to SLD = oiclotA'2 and the D20 backing film
going to an SLD = 6.2x10 A'z. ....................................................................................... 51

Figure 3.9: The concentration determined from neutron reflectivity shows a large amount
of PCBM near both interfaces. (a) Spin coating at either 800 or 2500 RPM yields very
similar concentration profiles, determined from the SLD profile, with slight differences
near the interfaces (see inset). The film thicknesses were 140 nm (800 RPM) and 90 nm
(2500 RPM) and used to normalize the distance through the film. (b) A cartoon of a
possible PCBM/P3HT morphology demonstrating higher PCBM concentration near the
interfaces with a large P3HT concentration within 3 nm of the air interface (see Fig. 3.1).
........................................................................................................................................... 53

Figure 3.10: The PCBM concentration profile changes upon annealing by broadening the
peaks near the interfaces which certainly influences solar cell performance. (a) The
PCBM concentration versus normalized distance graph for a PCBM/P3HT blend spin
coated onto a PEDOTzPSS layer shows peak concentrations near the interfaces for the
unannealed (84 nm thick) and annealed (75 nm thick) cases. The PCBM concentration
peaks widen at the interfaces and the PCBM concentration increases near the air interface
after annealing. The Parratt formalism was used in all cases to invert the reflectivity
profiles and “Substrate” indicates the top of the PEDOT/PSS layer. (b) The current
density - voltage curve changes between the unannealed and annealed conditions for a
solar cell and is certainly, in part, due to the change in the concentration profile. The film
thicknesses were 90 nm for both cases and the efficiency, fill factor and open circuit
voltage were 1.7%, 0.44 and 0.5 V, respectively, for the annealed cell. .......................... 57

xii

Figure 4.1: Comparison of thermal annealing effects on photoluminescence and solar
cell performance. Thermal annealing a 1:1 by weight P3HT:PCBM solar cell at 140°C for
20 minutes increases the total amount of photoluminescence (a), indicating a much higher
degree of geminate recombination. However, this annealing also produces a much
improved device (b) with device efficiency increasing from 1.2% to 2.9% ..................... 62
Figure 4.2: Plot of intensity (1) versus wave vector (q) for both the annealed and
unannealed samples of 1:1 by weight ratio of PCBM and P3HT as well as annealed and
unannealed samples of pure P3HT. Annealing pure P3HT has no noticeable effect on the
scattering but annealing the mixture shows a clear increase in scattering, indicating a
substantial change in PCBM agglomerate size. The solid lines labeled “Best Fits” are
from a polydisperse Schulz model having hard sphere interactions as described in the
text ..................................................................................................................................... 67

Figure 4.3: Cartoon of morphology change after thermal annealing. The PCBM particles
are well dispersed after spincoating and a poor charge transport network results due to
lose interconnectivity of the PCBM aggregates. The PCBM and P3HT form better
transport networks after annealing through a coarsening of the structure with a
subsequent increase in spacing between PCBM aggregates. These two effects yield an
increase in efficiency through better conductive pathways but simultaneously increase
photoluminescence due to greater interparticle gaps. ....................................................... 69

Figure 4.4: Grazing-Incidence X-ray Diffraction (GIXRD) plot of intensity versus wave
vector (q) for an as cast and annealed film of a 1:1 by weight mixture of P3HT and
PCBM. The peak narrowing and decreased FWHM of the annealed sample indicates an
increase in P3HT crystallite coherence length from 9 to 16 nm. ...................................... 71

Figure 5.1: Top down and 3D AF M images of (a,c) pure PEDOT:PSS and PEDOT:PSS
films containing (b,d) 10 mg/ml, (e,g) 20 mg/ml and (f,h) 40 mg/ml Si02 particles in the
initial solution. Increasing the solution concentration of SiOz increases the overall
surface coverage and roughness of the PEDOT:PSS coated films. .................................. 79

Figure 5.2: Top down and 3D AF M images of a 1:1 by weight film of P3HT and PCBM
on top of PEDOT:PSS with varying concentrations of SiOz in the initial solution. (a,c)
pure PEDOT:PSS and PEDOT:PSS films containing (b,d) 10 mg/ml, (c,e) 20 mg/ml and
(d,f) 40 mg/ml SiOz particles in the initial solution. Increasing the solution concentration
of Si02 increases the overall surface coverage and roughness of the PEDOT:PSS coated
films. ................................................................................................................................. 80

Figure 5.3: Current density (J) versus voltage (V) plot of solar cells having an increasing

amount of SiOz particles dispersed within the PEDOT:PSS layer. The pure PEDOT:PSS
device produces the worst cell but upon addition of SiOz in the PEDOT:PSS layer the
efficiency, short circuit current and fill factor all increase. .............................................. 82

xiii

Figure 5.4: Absorbance spectrum of PEDOT:PSS layers on glass with increasing
amounts of SiOz dispersed within. A larger amount of SiOz dispersed within the
PEDOT:PSS layer shows a higher light absorbance, which indicates an increase in
scattered light from the SiOz particles. The inset shows scattering efficiency versus
wavelength for Mie scattering theoretical calculations across 120 nm spheres. The
scattering increase according to Mie theory directly correlates with the increase in
observed absorbance in the rough PEDOT:PSS layers suggesting Mie scattering from the
120 nm SiOzspheres is the cause for the increase absorbance. ......................................... 86

Figure 5.5: Diagram of proposed scattering from rough surfaces with SiOz inclusions.
For a pure PEDOT:PSS film (a) a percentage of the incident light is absorbed by the layer
and the rest passes through into the detector. For a rough PEDOT:PSS film with SiOz
inclusions (b) the same proportion of the incident light is absorbed by the film but a
proportion of the non-absorbed light is scattered away from the detector. ...................... 87

Figure 5.6: Absorption spectrum of 1:1 by weight P3HT:PCBM films on top of
PEDOT:PSS films with various concentrations of Si02 particles within the film. An
increase in the amount of Si02 particles within the PEDOT:PSS film increases the overall
absorption, but the increase is small compared to the increases seen in the device
properties ........................................................................................................................... 89

Figure 5.7: Photoluminescence data of 1:1 by weight P3HT:PCBM films on top of
PEDOT:PSS films with various concentrations of Si02 particles dispersed within. An
increase in the amount of SiOz particles within the PEDOT:PSS film total
photoluminescence to a much greater degree than the observed solar cell properties or
absorbance ......................................................................................................................... 90

Figure 6.1: Time resolved optical micrographs (a-f) of thermal annealing a 75 nm thick
film of pure 115 kDa PS on a silicon wafer Silanized with Sigmacote ®. The large
pictures show the onset and continued dewetting of the PS film over the course of 50
minutes and the insets show a similar 75 nm thick film of PS with 1 monolayer
equivalent 60 kDa PS nanoparticles dispersed within, which completely inhibit
dewetting. .......................................................................................................................... 96

Figure 6.2: TEM image of a four layer system of polystyrene and o-CdSe nanoparticles
on a PET substrate created in two spin coats. A mixture of 140 kDa crosslinkable PS
with nanoparticles was spin coated onto a PET substrate, annealed at 170°C for 12 hours,
cross linked at 220°C for 20 minutes and repeated. ........................................................ 101

xiv

Figure 6.3: Cross sectional TEM images of polystyrene films with pyridine coated
cadmium selenide nanoparticles dispersed within. (a) A four layer system of PS and p-
CdSe nanoparticles created by spin coating the PS/p-CdSe mixture, annealing the film
followed by cross-linking the polymer and then repeating the process. (b) Cross sectional
image of the PS/p-CdSe film directly after spin coating. ............................................... 102

Figure 6.4: Cross sectional images of oleic acid coated CdSe nanoparticles assembled at
the substrate of thin PMMA films. Due to the difference in refractive index of PMMA
from PS the o-CdSe particles assemble at the substrate compared to the air interface as
shown in previous work”. The assembly is very robust and occurs in (a) thin films less
than 100 nm, and (b) films approaching 400 nm in thickness. ....................................... 103

Figure 6.5: A thin film of polystyrene with both o-CdSe and p-CdSe nanoparticles
dispersed within. After annealing above the glass transition temperature of PS the o-
CdSe particles migrate to the air interface while the p-CdSe particles migrate to the
substrate. To the left and right of the TEM image are representations of the CdSe
particles with their respective stabilizing ligands. .......................................................... 104

Figure 6.6: Plot of thermal gravimetric analysis (left axis) and differential scanning
calorimitry (right axis) of pure P3HT. The onset of melting and degredation are almost
identical at about 200°C (horizontal dashed line) indicating thermally processing P3HT
for assembling nanoparticles may not be feasible. ......................................................... 106

Figure 6.7: Cross sectional TEM images of films of (a) pure P3HT with dispersed p-
CdSe particles assembling on the top of the film and (b) a solar cell mimic film consisting
of a 1:1 by weight mixture of P3HT and PCBM with a small volume fraction of p-CdSe
particles dispersed within. Upon heating to 250°C in a nitrogen environment the CdSe
particles assemble into the middle of the film, presumably on top of a layer of PCBM that
has assembled at the substrate and extends upwards to roughly the middle of the film
based on the volume fraction of PCBM added. .............................................................. 108

Figure A1: A cartoon of the mask used to make the majority of the devices used in this
work and the associated ITO pattern it eventually creates. The dashed square is the
outline of a slide the mask will cover. ............................................................................ 121

Figure A2: Top down and cross sectional views of applying PEDOT:PSS onto patterned
ITO slides and the subsequent removal of the layer from the edges to allow for good
electrical contact during testing. ..................................................................................... 124

Figure A3: Top down and cross sectional views of applying the active layer onto
patterned ITO slides with a PEDOT:PSS layer and the subsequent removal of the layer
from the edges to allow for good electrical contact during testing. ................................ 126

Figure A4: Top down and cross sectional views of deposited electrode. The dashed
squares indicate where the active area of the four solar cells created fi'om this masking
pattern. ............................................................................................................................ 127

Figure A5: Process map of the 6 steps (starting from clean ITO slides) needed to make a
polymer based solar cell on glass slides using a transparent electrode, in this case indium
tin oxide (ITO). ............................................................................................................... 128

Figure Bl: Concentration versus height data for 75 kDa polystyrene. Both toluene and
pyridine were used as solvents and spinning speeds were 2000 and 5000 RPMS. Closed
symbols represent height data from ellipsometry and open symbols represent height data
obtained from X-ray reflectivity. Both Erica Tseng and Wenluan Zhang aided in
obtaining the PS height data. .......................................................................................... 130

Figure B2: Concentration versus height data of 72.6 kDa and 49.2 kDa poly(methyl
methacrylate) in toluene and chlorobenzene. All films were spin coated at 5000 RPMs
and all data is from ellipsometry. .................................................................................... 131

Figure B3: Concentration versus height data of P3 HT in chlorobenzene spin coated at
5000 RPMS. Height data was obtained from scratching a film and measuring the step
height of the scratch. ....................................................................................................... 132

Figure B4: Concentration versus height of PDMS in toluene spin coated at 5000 RPMS.
The PDMS used in this study was cross-linkable PDMS from Dow, Sylgard® 184, that
can be cross-linked within 30 minutes at 60°C. Height data was obtained from
ellipsometry ..................................................................................................................... 133

Figure B5: Concentration versus height of PIB in toluene spin coated at 5000 RPMS.
The PIB used in this study was low molecular weight and the polymer was liquid at room
temperature. Height data was obtained from ellipsometry. ........................................... 134

Figure Cl: J-V curve of a P3HT:PCBM device on pure PEDOT:PSS with no LiF layer
between the active layer and aluminum electrode. The series resistance reported is taken
from the slope of the J-V curve at the open circuit voltage. ........................................... 135

Figure C2: dJ/dV versus V plot of a P3HT:PCBM device on pure PEDOT:PSS with no
LiF layer between the active layer and aluminum electrode. The voltage dependence of
the illuminated current shows that traditional methods to evaluate the device must be
based on the dark curve, which has minimal dependence on the voltage ....................... 136

_ Figure C3: dV/dJ versus 1'1 plot of a P3HT:PCBM device in the dark on pure
PEDOT:PSS with no LiF layer between the active layer and aluminum electrode ........ 137

Figure C4: J-V curve of a P3HT:PCBM device on PEDOT:PSS with 10 mg/ml of SiOz
particles dispersed within and no LiF layer between the active layer and aluminum
electrode. The series resistance reported is taken from the slope of the J-V curve at the
open circuit voltage. ........................................................................................................ 138

Figure C5: dJ/dV versus V plot of a P3HT:PCBM device on PEDOT:PSS with 10 mg/ml
of Si02 particles dispersed within and no LiF layer between the active layer and
aluminum electrode. The voltage dependence of the illuminated current shows that
traditional methods to evaluate the device must be based on the dark curve, which has
minimal dependence on the voltage. ............................................................................... 139

Figure C6: dV/dJ versus J'l plot of a P3HT:PCBM with 10 mg/ml of Si02 particles
dispersed within the PEDOT:PSS and no LiF layer between the active layer and
aluminum electrode. ........................................................................................................ 140

Figure C7: J-V curve of a P3HT:PCBM device on PEDOT:PSS with 20 mg/ml of Si02
particles dispersed within and no LiF layer between the active layer and aluminum
electrode. The series resistance reported is taken from the slope of the J-V curve at the
open circuit voltage. ........................................................................................................ 141

Figure C8: dJ/dV versus V plot of a P3HT:PCBM device on PEDOT:PSS with 20 mg/ml
of Si02 particles dispersed within and no LiF layer between the active layer and
aluminum electrode. The voltage dependence of the illuminated current shows that
traditional methods to evaluate the device must be based on the dark curve, which has
minimal dependence on the voltage ................................................................................ 142

Figure C9: dV/dJ versus J'l plot of a P3HT:PCBM with 20 mg/ml of SiOz particles
dispersed within the PEDOT:PSS and no LiF layer between the active layer and
aluminum electrode. ........................................................................................................ 143

Figure C10: J-V curve of a P3HT:PCBM device on PEDOT:PSS with 40 mg/ml of SiOz
particles dispersed within and no LiF layer between the active layer and aluminum

xvii

electrode. The series resistance reported is taken from the Slope of the J-V curve at the
open circuit voltage. ........................................................................................................ 144

Figure C11: dJ/dV versus V plot of a P3HT:PCBM device on PEDOT:PSS with 40
mg/ml of Si02 particles dispersed within and no LiF layer between the active layer and
aluminum electrode. The voltage dependence of the illuminated current Shows that
traditional methods to evaluate the device must be based on the dark curve, which has
minimal dependence on the voltage. ............................................................................... 145

Figure C12: dV/dJ versus J'l plot of a P3HT:PCBM with 40 mg/ml of Si02 particles
dispersed within the PEDOT:PSS and no LiF layer between the active layer and
aluminum electrode. ........................................................................................................ 146

xviii

3»

Q

3

>®EDJtfiP80

i>
—
N
0)

AFM

LIST OF SYMBOLS AND ABBREVIATIONS

Angstrom

absorption coefficient (unites are inverse centimeters) (cm'l)
dielectric constant of a material

dielectric permittivity of a vacuum

Wavelength (units are Angstroms or nanometers) (A or nrn)
Pi

charge carrier mobility

Efficiency

Maximum packing fraction of particles

Volume fraction of particles

Scattering angle

Hamaker constant .

Hamaker constant ordering for a tri-layer system of three components
Atomic force microscope

Aluminum

Aluminum Oxide

Cubic centimeters

Cadmium selenide

Cadmium selenide nanoparticles coated with an oleic acid stabilizing layer
Cadmium selenide nanoparticles coated with a pyridine stabilizing layer

Degrees Celsius

Carbon Dioxide
Diameter

Diffusion coefficient of an electron
Diffusion coefficient of a hole

Deuterium oxide

Dynamic light scattering
Differential Scanning Calorimetry
Derivative of term X

Energy

External quantum efficiency
Electron volts

Fill factor

Grams

Gigawatt

xix

HF
HOMO

Edge to edge distance between particles
hydrofluoric acid

Highest occupied molecular orbital
Current (units are arnperes) (A)

Light intensity

Illuminated current (units are arnperes) (A)
Incident light intensity

Dark current, also called saturation current (amperes) (A)
Internal Quantum Efficiency

Indium tin oxide

Current Density

Short circuit current density (milliamperes/square centimeter) (mA/cmz)
Boltzmann’s constant (electron volts/temperature) (eV/K)

Diffusion length of an electron

Diffusion length of a hole

Lithium fluoride

Lowest unoccupied molecular orbital
milligram

milliliter

Molecular Weight

Refiactive index

Non-ideality coefficient in the diode equation

Intrinsic carrier concentration
Acceptor dopant concentration

Donor dopant concentration

Near edge x-ray absorption fine structure spectroscopy
nanometer

Polydispersity Index

poly(ethylene terephthalate)
polyisobutene

Photoluminesence
poly(3-hexylthiophene)

poly(ethylene dioxythiophene)
[6,6]-phenyl-C61-butyric acid methyl ester
poly(methyl methacrylate)

Phase sensitive neutron reflectivity
polystyrene

polystyrene sulfonate

XX

Wave vector, generally used for reflectivity (inverse Angstroms) (A'l)
Wave vector, generally used for SANS (inverse Angstroms) (A'l)
Charge of an elemental particle, either an electron or hole (eV)
Reflectivity or Reflectance

Revolutions per minute

Series resistance (ohms) (Q)

Shunt resistance (ohms) (0)
Silicon

Silicon dioxide or silica

Small angle neutron scattering
Scanning electron microscope

Scattering length density (unit are inverse Angstroms squared) (A'z)
Temperature

Degradation temperature

Glass transition temperature
Transmission electron microscope
Terrawatt

Thermal Gravimetric Analysis
Ultraviolet & Visible spectrum spectroscopy
Voltage

Open circuit voltage (units are volts) (V)
Watt

X-ray photon spectroscopy

X-ray diffraction

vertical distance through a film or sample

xxi

CHAPTER 1: INTRODUCTION

1.1 Motivation

The increasing globallenergy demand is one of society’s greatest engineering problems.
Our society and way of life depend on having an abundant energy supply and the
worldwide demand for energy is increasing at a rapid rate. In 2001 the total world energy
consumption was roughly 12 terawatts (TW) and in 2008 consumption had increased to

over 15 TW. Current predictions indicate a worldwide energy demand of roughly 40 TW
by 2050, almost triple our current usage.l Approximately 85% of today’s energy is
created by burning fossil fuels; mainly coal but a significant amount of oil and natural gas
as well. A large coal powered electricity plant produces about 1 gigawatt of energy,
which means meeting the worlds increasing energy demands requires bringing over 10
coal powered electricity plants online every week for the next 40 years.

Coupled with increasing worldwide energy demand is securing an energy source

for our future. The United States imports roughly 60% of all its petroleum and over 50%

of those imports come from politically unstable countries.2 With unstable politics,

increased pricing, the emerging economies of India and China and the ever growing
demand for more energy, finding an energy source not dependent on foreign suppliers is
of high importance for our country.

Alongside this increase in energy demand and concern over a stable energy

supply is an increase in carbon dioxide (C02) emissions from the massive amount of

burning fossil fuels. C02 is a well known greenhouse gas, meaning it absorbs infrared

radiation from the Earth, not allowing the radiation to escape from the atmosphere.

Greenhouse gasses are vital to life on Earth because they keep surface temperatures

hospitable, however, a substantial amount of data exists showing a correlation between
C02 emissions and increasing temperatures on Earth. Global warming is a hot issue in

both science and politics today with critics on both sides of the issue. This dissertation

will not discuss whether global warming is real and if C02 is truly a contributor to global

temperature rise but few people will argue that reducing C02 emissions will have a

negative impact on the global environment.

To meet increasing energy demands without emitting such a massive amount of

C02 and other pollutants from burning fossil fuels, 3 switch to renewable energy sources

must begin. Many renewable sources exist including biomass, wind, geothermal, tidal
and solar and all of these are being investigated to some extant or another as possible
alternatives to fossil fuels. Nuclear is also a potential option but, due to a great deal of
public fear and governmental regulations, it is a relatively small source of power
worldwide outside of France.

As for the truly renewable sources, wind, hydroelectric and biomass provide
roughly 2 TW of total worldwide energy combined. Wind farms are becoming more
prevalent today because of their relative ease of installation and large, commercial scale
integration that makes them more cost competitive than the others. Biomass, which is
mainly used as a fuel for heat, is receiving a great deal of attention for its potential to be
converted to a liquid fuel such as ethanol and geothermal, due in large part to increases in

drilling technology, has some backing as a potential energy supplier.

Despite the great deal of attention these energy sources are receiving, their total
energy potential is rather limited. Lewis and Nocera point out that all of these energy

sources, biomass, hydroelectric, wind, geothermal and tidal have substantial limits on

their practicality.l In fact, the total potential of all of these energy sources combined,

assuming complete harnessing of all plausible energy from each source, amounts to less
than 18 TW. This is a substantial amount of energy but obtaining all 18 TW still leaves
the world roughly 7 TW short of the projected 40 TW demand in 2050.

Essentially an infinite supply of energy, the sun provides a constant 120,000 TW
incident on the earth with over 600 TW assumed to be harvestable.l With such a massive
amount of energy irradiating the Earth, solar energy is an obvious renewable source to
aid in the ever increasing energy demands.

Despite its seemingly infinite supply, solar energy is an insignificant fraction of
the world’s energy. Due to its large required area, low efficiencies and the fact that only
about 8 hours of the day produce a substantial amount of energy in any one location,
solar energy has proven to be uneconomical in today’s world. In 2008 the total
worldwide production of solar cells was 5.6 GW. This means that at peak sunlight all of
the cells combined could produce 5.6 GW of electricity, less than 0.01% of the world’s
energy supply and as these devices only operate at peak sunlight for a short period each
day the actually energy output is significantly less.

Solar cells made from inorganic materials have been around for many years and
were first developed into a commercial product in 1942 by Bell Labs. Despite this long
history they remain an insignificant fraction of the world’s energy supply due to their

high cost compared to fossil fuel energy. Until recently, most solar cells were made from

single crystal Silicon ingots that were cut and placed into large modules. These devices
have decent efficiencies, ~12-15%, but are costly due to the large amount of labor
involved as well as the high cost of creating high quality single crystal silicon. More
recently thin film devices, such as cadmium teleuride (CdTe) and copper indium gallium
selenide (CIGS) devices have seen a massive increase in research and commercialization.
Thin film devices have the benefit of being faster and cheaper to produce because
continuous processing such as evaporation and sputtering can be used as opposed to
cutting slices of individual ingots. While faster, these thin film modules still are energy
intensive, relatively slow to manufacture and require expensive and sometimes exotic raw
materials.

A better material for solar cells, in terms of speed and low manufacturing costs is
plastic. Polymers are cheap, easy to make on very large scales and can be produced in a
roll to roll process extremely efficiently. Optically active polymers, otherwise known as
semiconducting or highly conjugated polymers, have very high absorption coefficients
meaning they can be made very thin, roughly 100 nm, while still absorbing most of the
incident light within their absorption spectrum. Having such small films means much
less raw materials are needed. Polymer based devices are roughly 100 nm thick
compared to roughly 100 microns for a crystalline silicon cell. This difference in
thickness correlates to almost 3 orders of magnitude less material needed than silicon
based devices.

Polymer based solar cells are relatively new devices with the first realistic

polymer based device being produced in 1995.3 Unfortunately they are plagued with

many intrinsic problems such as poor electrical conductivity, limited absorption

spectrurns, high band gaps and short life times. The goal of this dissertation is to add to
the overall understanding of polymer based solar cells, specifically by developing

improved characterization techniques to probe the internal morphology of these devices.

1.2 Thesis Outline

The general focus of this dissertation is to fully describe the internal structure of polymer
based solar cells. Much research has gone into these devices over the past 15 years but a
good description of the internal size scales and distribution of the polymer and
nanoparticle used in these devices has not previously been available. Previous work has
shown the charge transport within these devices is governed by length scales on the order
of 10 nm and this work focuses on characterizing the devices at that length scale.

Chapter 2 provides an overview of work performed on polymer based solar cells
and a brief description of how these devices work. The chapter provides a basic
understanding of the terms and experiments used to study solar cells and presents some of
the characteristics of the devices.

Chapter 3 is a description of the vertical profile of PCBM (the nanoparticle used
in most polymer based solar cells) throughout the device. Most research in this field
focused on the characteristics of the polymer and little consideration had been given to
the location of PCBM with the devices. Determining the location of PCBM within these
films is difficult because electron contrast between the nanoparticle and polymer is
minimal. Neutron scattering, due to its natural contrast between the two components
provides a great method to probe the locations of these components. This chapter

describes the technique of neutron reflectometry, a more detailed probing method of

phase sensitive neturon reflectometry and discusses the structure formed after spin
coating and thermally annealing these devices.

Chapter 4 focuses on using another neutron scattering technique, small angle
neutron scattering (SANS) to determine the size of PCBM clusters formed upon spin
coating and after thermal annealing. These solar cells use a very high volume fraction of
nanoparticle, ~48%, making interparticle interactions and agglomeration of the particles
extremely likely. This chapter describes the size of PCBM clusters formed upon creation
of these devices and the effect of these clusters on device performance. Additionally,
Chapter 4 answers a longstanding question in the field of how an annealed device will
increase overall efficiency while also increasing overall photoluminescence.

Chapter 5 focuses on making nanorough substrates for fabrication of these
devices to enhance overall device performance. Incorporating silica nanoparticles into
the hole blocking layer of these devices increases overall efficiency through a
combination of increased contact area, light scattering and other synergistic interactions.

Chapter 6 describes the characterization work done with nanoparticles dispersed
in amorphous polymers in an attempt to better understand the forces dominating
nanoparticle dispersion and assembly within solar cells. The amorphous
polymer/nanoparticle characterization work is then extended into conjugated polymers
used in solar cells.

Chapter 7 summarizes the main conclusions of this dissertation and discusses
possibilities for future work.

Chapter 8 is a literature review of much of the important literature to this

dissertation and the field of polymer solar cells in general.

CHAPTER 2: POLYMER BASED SOLAR CELLS: BACKGROUND AND

OPERATING PRINCIPLES

2.1 Introduction
All solar cells consist of a light absorbing material where, upon light absorption,

an electron is promoted from a ground state to an excited state. This promoted electron
leaves behind a positively charged empty space called a hole, which is treated as if it
were a particle with a mass, charge and mobility. Collectively this electron-hole pair is
referred to as an exciton.

An exciton is a quasiparticle bound together by a strong Coulombic interaction.
As the Coulombic interaction of two particles varies inversely with the dielectric constant
of their surrounding medium, particles in a lower dielectric media, such as a polymer
matrix, will have much greater binding energy than particles in a higher dielectric
material, such as an inorganic semiconductor. The binding energy of an exciton can be

expressed as4

4
mred e

_ 2(47reeo)2 It?

(2.1)

where E is the binding energy between the electron and hole, mred is the reduced mass of

the exciton defined by mred = me*mh*/(me* + mh*) with m; and m}: being the electron

and hole masses, respectively, e is the charge of the particle, e is the dielectric constant of

the material, so is the dielectric permittivity of free space and h is Planck’s constant

divided by 211:, one can see the total interaction of an exciton is governed by the dielectric
constant of the medium. Traditional solar cells made from inorganic materials such as

silicon or cadmium telluride have high dielectric constants ranging between 10 and 15,

which produce excitons with energies around 0.015 eV.5 The thermal voltage at room

temperature, kT/q, where q is the charge on an electron, k is Boltzmann’s constant and T
is the temperature is about 0.025 eV. This means that any exciton created in a high
dielectric constant material will most likely dissociate at room temperature. Highly
dissociable excitons such as these are called Wannier—Mott excitons.

Conjugated polymers, however, have dieclectric constants ranging between about
2-3, and create excitons with binding energies between 0.4 and l eV.6’ 7 An exciton with
such a high binding energy, referred to as a F renkel exciton, is confined to the molecular
unit in which it was created and requires an interface with a favorable energy band offset
to dissociate. This high exciton binding energy limits the time an exciton will remain in

an excited state which leads to a low diffusion length, one of the most damaging features

of polymer based solar cells.

2.2 Background

Dissociation of the exciton into free electrons and holes requires counteracting the

Coulombic force holding the particles together. In 1992 Saricifici et al.8 showed the
addition of C60 to thin films of poly[2-methoxy-5-(2'-ethyl-hexyloxy)-1,4-phenylene

vinylene] (MEH-PPV) will quench photoluminescence. The C60 particles acted as

8

electron acceptors, overcoming the exciton binding energy and separating the charge

carriers. However, due to the low solubility of C60, good pathways to the electrodes

could not be formed and most dissociated charge carriers were unable to reach their

respective electrodes, preventing these films from being good devices. In 1995 Yu et al.3
overcame the charge extraction problem by blending [6,6]-phenyl-C61-butyric acid

methyl ester (PCBM), a more soluble derivative of C60, with MEH-PPV to create a solar

cell of about 1.5% efficiency. Due to increased solubility a much larger amount of
PCBM can be added to a thin film, forming an interpenetrating network of electron
donors and acceptors called a bulk heterojunction. When an exciton is created by
photoabsorption in a bulk heterojunction it is much more likely to encounter an acceptor,
PCBM in this case, that has a pathway to an electrode.

Polymers have intrinsically low and sluggish charge transport properties with

electron and hole mobilities many orders of magnitude lower than those of inorganic

9, 10

materials and hence have exciton diffusion lengths of about 5-10 nm. An exciton

diffusion length of this magnitude in a thin film of about 200 nm suggests a comb-like

architecture for the active layer, with 200 nm long tines spaced at 10 nm intervals.”

Thus, the device’s form is critical since there should be a clear path to each electrode
after exciton dissociation with no inoperative material. Fabricating such a. structure is
certainly a challenge. While the bulk heterojunction does allow for both exciton
dissociation and charge transport to electrodes, it is not optimized and certainly not a well

characterized morphology.

Improving bulk heterojunction cell efficiency has been difficult, requiring
comprehensive, empirical device manufacturing and subsequent testing. Processing

parameters have been extensively studied with the aim of manufacturing a more suitable

device. Brabec et al. '2 showed that changing the solvent from toluene to chlorobenzene

for a device of Poly[2-methoxy-5-(3’,7'-dimethyloctyloxy)-l,4-phenylenevinylene]
(MDMO-PPV):PCBM almost doubled device efficiency. The cells spin coated from
cholorbenzene had the same open circuit voltage as the toluene cells but had over twice
the short circuit current and nearly identical absorption properties. Since the absorption
properties of each cell were the same, the doubling in efficiency indicated some internal
morphological change that enhanced charge carrier transport. Original theories on this
work suggested that changing from toluene to chlorobenzene increased the solubility of
PCBM and created a more homogenous distribution of acceptors throughout the active
layer. While this theory is in part true, more recently groups have shown that a slower

drying solvent will enhance polymer crystallization, increasing the charge transport of the

. . . . I
material and allowing more efficrent charge extraction 3.

Yang Yang’s group at UCLA has done extensive work showing both thermalM’ 15

and ‘solvent’ annealing13 of the active layer greatly increases polymer crystallization and

therefore device properties. Spin coating a cell until most of the solvent is removed and

then annealing the device at 120-1500C, which is well below the melting temperature of

the material, adds enough thermal energy to enhance polymer crystallinity and improve
charge transport. Spin coating for a short time, less than is required to remove all

residual solvent from the device, and then letting the device dry slowly, termed ‘solvent’

10

annealing in the polymer photovoltaic literature, allows a greater degree of crystallization
and even higher device efficiencies. In addition to solvent choice and fabrication,

increased molecular weight can also have a profound impact on device performance by

increasing hole mobilities by orders of magnitude.16

In recent years groups have moved from PPV derivatives to
poly(3-hexylthiophene) (P3HT), shown in Figure 2.1, because of P3HT’s increased
crystallinity and better charge transport properties, which correlate into better performing
devices. Because of the improved performance and the wealth of published data on
P3HT we have chosen to study it in this dissertation. Furthermore, morphological
characterization of this system has provided diverse, contradictory concentration profiles
within the film not allowing a clear relation between structure and performance to be
made. Crystallinity is obviously an important factor in device efficiency, but, given the

extremely low exciton diffusion lengths, PCBM location is also vital to device properties.

/ \

 

P3HT

Figure 2.1: The molecular structure of PCBM and P3HT

11

The best P3HT:PCBM devices to date are fabricated in a 1:1 weight ratio as this
ratio has been determined empirically to be the best ratio for achieving high

efficiencies.17

The two components are dispersed in a solvent, generally a chlorinated
solvent such as chlorobenzene, and spin coated onto a substrate. This spin coating
creates a thin film with the two components mixed together in a bulk heterojunction.
Ideally, the bulk heterojunction is an interpenetrating mixture of the polymer and the
nanoparticle, however, little is known of the actual make-up and morphology of the bulk
heterojunction and a large portion of this thesis focuses on characterizing the internal
structure of a P3HT:PCBM blend.

The P3HT:PCBM active layer is sandwiched between two electrodes, a
transparent electrode and an evaporated metal. The top electrode is a transparent
conducting oxide, generally indium tin oxide (ITO) coated with the hole transporting
macromolecular salt poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)
(PEDOT:PSS), although different transparent electrodes and hole conducting layers are
occasionally used in the literature. As there is virtually no built in electric field in a
polymer solar cell, such as in traditional p-n junction inorganic devices, PEDOT:PSS is
used to block electron recombination at the ITO surface. Without a PEDOT:PSS layer,
excitons created near the ITO surface are free to dissociate into the ITO as opposed to a
PCBM molecule, creating a dramatic loss in current and very poor devices.

The back contact of these devices is generally a thin layer of aluminum, about 80-
100 nm, and often a 1 nm layer of lithium fluoride (LiF) or a 20 nm layer of calcium is

deposited between the Al and the active layer. There are currently a number of

hypotheses in the literature as to why a buffer layer is needed between the polymer and

12

the aluminum electrode; including a better work function match between calcium and the
polymer is obtained,18 a small amount of lithium diffusing into the polymer to enhance
charge transport,19 or, as our group currently believes, the buffer layer absorbs much of
the energy of the evaporated Al as it condenses on the surface of the device, protecting
the polymer from degradation. The last conjecture is based on preliminary results and

will be tested further in the near future.

 

a

/

Aluminum

 

 

  
 

 

Polymer/Nanoparticle
Layer

PEDOT:PSS

ITO

 

 

 

 

 

 

 

GLASS

 

 

 

Figure 2.2: (a) The structure of a typical polymer based solar cell comprised of P3HT
and PCBM and (b) its ideal band diagram. The lithium fluoride layer is left out of the
band diagram because its conduction and valence bands (1.0 and 14 eV, respectively)20
do not match with either PCBM or Al. Much research has shown the addition of a thin
layer of LiF between the back contact of a cell and the active layer greatly improves
performance but the reason for this enhancement is still unknown.

Proper alignment of the molecular orbitals and work functions of the cell
components is required for adequate device performance. Figure 2.2a is a cartoon of a
typical P3HT:PCBM solar cell and the associated idealized energy diagram, Figure 2.2b.

The energy diagram depicts the highest occupied molecular orbitals (HOMO) and lowest

13

unoccupied molecular orbitals (LUMO) of the organic materials and the work firnctions
of the inorganic electrodes. From the energy diagram we can see P3HT must be in
contact with the PEDOT:PSS layer and PCBM must be in contact with the metal

electrode.

2.3 Solar Cell Operating Principles

2.3.1 Equivalent Circuit Diagram

A solar cell is diode that will create an electrical current upon illumination. To describe
the current flow of a solar cell an equivalent circuit diagram can be drawn, shown in

Figure 2.3. An ideal solar cell circuit diagram would consist only of the illuminated

current, IL, and the diode current, 11), but fabricated solar cells are not ideal and contain
both a series and shunt resistance, Rs and Rs“, respectively. The Rs in any solar cell is
the resistance of flow through the device and should be minimized. A high Rs causes a

solar cell to have a lower overall current output. The RSH is the resistance to flow

backwards through the diode, otherwise known as leakage current, and should be kept as

high as possible. A low RSH can result in lower total voltage or current for a solar cell.

 

 

Figure 2.3: Equivalent circuit diagram of an illuminated solar cell.

14

A solar cell shown in the above circuit diagram can be described by the equation

(V+IRE)
1 = 10 (eq nkr — 1) — “3+”‘S—l (2.2)
SH

where I is the current, 10 is the dark, or saturation current, q is the charge of an electron, k

is Boltzmann’s constant, n is the ideality factor and Rs and RSH are the series and shunt

resistances, respectively. The ideality factor, n, is a term inserted into the diode equation
to account for the non-ideality of a device. It can be used as a probe into the types of

recombination in a solar cell, which will be discussed later.
2.3.2 Testing Solar Cells

Ideally, solar cells are tested in direct sunlight. However, as sunlight intensity
varies based on a number of factors most testing is performed in a controlled environment
using a solar simulator. Most solar simulators today consist of a xenon arc lamp and a
filter to closely match the solar spectrum. Figure 2.4 is the actual solar spectrum for AM

0, AM 1.5 Global and AM 1.5 Direct conditions.

15

 

 

2.0 h -
I — AM 0
A l“ —— AM 1.5 Global
NE 15 ... '. ‘ .............. AM 1.5 Direct _.
2 ll ‘0
E ll
I 1.0 I i' -
C
2 l l
C
.— 0.5 'l . y . _.
0.0 l J int—......

 

 

 

1,000 2,000 3,000 4,000
Wavelength (nm)

Figure 2.4: Solar spectrum for AM 0, Am 1.5 Global and AM 1.5 Direct sunlight

AM refers to Air Mass and is a standard used in all solar cell testing. AM 0,
Figure 2.4 black line, is the sun spectrum directly incident on the Earth’s atmosphere,
similar to a black body emitting at 5500 K. Once the sun’s light passes into the
atmosphere certain wavelengths are absorbed due mainly to water, carbon dioxide and
ozone. Thus, the solar spectrum arriving at Earth’s surface is slightly modified compared
to the one at the edge of the atmosphere. AM 1.5 Direct (Figure 2.4 dotted line) is the
light incident on a device after it has passed through the atmosphere and AM 1.5 Global
(Figure 2.4 grey line) refers to the incident light on a device plus any diffuse light
scattered from the surroundings into the cell. AM 1.5 Global is the spectrum usually

used in testing.

16

The incident angle of the sun compared to its zenith can slightly affect the
spectrum and greatly affect the intensity. If the sun is at its zenith and passes directly
through the atmosphere to the ground it is said to have passed through 1 Air Mass unit,
known as AM 1. The standard unit for solar testing is AM 1.5. In other words, the

sunlight has passed through 1.5 times more atmosphere than it would have if the sun was
at its zenith. This is equivalent to an angle of 48.20, as shown in Figure 2.5.

Geometrically, this is defined as

— (2.3)

where AM is the Air Mass unit and 0 is the angle measured from the zenith of the sun.

 

Figure 2.5: Geometric description of how AM 0, AM 1.0 and AM 1.5 spectra impact
earth and the atmosphere.

The standard illumination intensity in solar testing is 1000 W/mz. A solar cell is

connected to a voltage/current sourcemeter, illuminated with 1000 W/m2 of simulated

17

sunlight and a potential is applied to the device ramping from negative to positive bias
while the corresponding current is measured. The same procedure is then applied to the
device in the dark. The resulting current (1) versus voltage (V) plot defines the main
characteristics of the solar cell, shown in Figure 2.6. Often times current density (J) is
used, which is the total current produced divided by the area of the cell, allowing for

adequate comparison between devices.

I

(current)

  

l
V I
Dark Current____..-n-i"' VOC
JPV

(voltage)

 

 

      

 

Illuminated Current

Figure 2.6: Typical I-V curve obtained from testing a solar cell. Upon illumination a
solar cell is defined mainly by three major points from the I-V curve, the open circuit

voltage (V 0C) the short circuit current (13(3) and the fill factor (FF).

2.3.3 Short Circuit Current

The short circuit current, 13C, is the total current from a cell upon illumination when no

voltage is applied. Essentially it is the total number of electrons coming out of a device.

18

The maximum Isc obtainable from a device is determined by integrating the spectrum of

incident light, generally taken as a blackbody emitter at 5500K, from the high energy
edge of the spectrum to the band gap of the material. With the assumption that every

incident photon creates one harvestable electron, the maximum theoretical current can be
found. Figure 2.7 is a plot of the max JSC versus material band gap. For reference, the
maximum current density obtainable from a silicon device (band gap of 1.1 eV) is

roughly 44 mA/cmz, whereas the maximum Jsc obtainable for a solar cell based on P3HT

(band gap of 1.9 eV) is only about 18 mA/cmz.

 

70 IIITIITTTITTTFIIIIIIITITITIIIIIIIIIIIII

60

50

40

30

20

10

. 2
Max Current DenSIty (mA/cm )

TIII'TIII'ITIIITTITIIITIIWIIIIIII

lllllllllllllllllLllllllllllllllll

 

 

 

Band Gap (eV)

Figure 2.7: Theoretical maximum current density (Jsc) obtainable versus band gap.21

19

2.3.4 Open Circuit Voltage

The open circuit voltage (Voc) is the maximum voltage obtainable from a solar cell,

which occurs at a current of zero. The theoretical maximum VOC of a device is the band

gap of the absorber material. By setting I = 0 in the illuminated diode equation (Eq: 2.2)

and solving for voltage we obtain
nkT I
Voc = —ln (—" + 1) (2.4)
q 10

showing that the Voc of a device is determined by the ratio of illuminated current to the

dark current and that the maximum Voc occurs in conjunction with the lowest possible

dark current. The dark current for a p-n heterojunction is defined by

2 2
qD n- (10 n-
LeNA LhND

where q is the charge of an electron, De and Dh are the electron and hole diffusion

coefficients, respectively, defined as D = kTp/q with It being the particle charge mobility,

NA and ND are the n and p dopant concentrations, 11; is the intrinsic charge carrier
concentration, Le and Lh are the charge carrier diffusion lengths and NA and ND are the

acceptor and donor concentrations, respectively.

20

It is beyond the scope of this dissertation to discuss how the parameters mi, N A

and ND transfer from a doped inorganic device to an undoped polymer film, but a

reasonable approximation is that they will be some constant for a given polymer. Thus,

10 can be reasonably represented as

10 = (C —+ Ch —) (2.6)

where Ce and Ch represent the polymer based equivalent of the ratio of intrinsic charge

carrier concentration to electron and hole dopants. The terms D/L have units of length

per time, a velocity, and are directly related to recombination. With either a large

diffusion coefficient or, more realistically, a small diffusion length, 10 will become

larger, which results directly in a lower Voc as described by equation 2.4.

2.3.5 Fill Factor

The fill factor (FF) of a device is a measure of how ideal the solar cell is. For any

photovoltaic device the maximum current and voltage obtainable are the Jsc and VOC,
respectively. However, the maximum power obtainable from a device occurs at the
maximum power point (PM), which is the point where the product of the voltage and

current is a maximum (P=IV). The fill factor is then a ratio of the theoretical PM, the

product of JSC and VOC, and the actual PM. The FF is defined by

21

V I
FF: ”M (M)
Voclsc

 

with Vm and Im being the maximum voltage and current power points, respectively.

2.3.6 Efficiency

The total efficiency of a device is defined as the ratio of the power out of a cell and the

power incident on the cell

_ PMAX
IN

 

where n is the efficiency of the device, PIN is the light intensity incident on the device,
typically 1000 W/mz) and PM Ax is the maximum electrical power produced. PMAX can

be expressed as the product of the Voc, 13c and FF.

P MAX = VoclscFF (2-9)

2.4 Polymer Solar Cell Devices
Figure 2.8 is a plot of 4 different P3HT:PCBM solar cells with the parameters of each
shown in Table 2.1. Two devices were tested directly after spin coating with one device

having only an Al back electrode while the other had a ~1 nm layer of LiF between the

22

active layer and the Al. The other two devices were tested after annealing for 20 minutes

at 140°C, one having only an Al electrode and the other having the LiF interlayer.

Clearly the annealed devices are much better with improved 1], 13¢ and FF and the

devices with LiF, as discussed above,’°’ '9

perform even better yet. Some of the
morphological reasons for the increased performance upon annealing are discussed in the

ensuing chapters, but a fair amount of information can be gained about these devices

from the J-V data.

 

I

-e— As Cast AI only' ' I

O "-B-ASCastw/LiF -
+ AnnealAl only . f
-I- Annealw/LiF :

-
-2~ - -
- v
v

-
v

    
   

0

0

 

 

Current Density (mA/cmz)
i.

 

 

40 i L I I
-o.2 0.0 0.2 0.4 0.6
Voltage(V)

Figure 2.8: Current Density versus Voltage plot for solar cells as cast and after annealing

for 20 minutes at 140°C. Closed symbols denote annealed devices and open symbols

denote as cast devices. The squares indicate a device was fabricated with a 1 nm layer of
LiF between the active layer and aluminum and the circles indicate no LiF was used.

23

The ch of these devices increases from about 2 to 9 mA/cm2 and there are many

reports in the literature of groups achieving even higher J SC values with similar systems

22-24

due to Slightly modified processing conditions. In Figure 2.7 we showed the

maximum Jsc for a device with a band gap of 1.9 eV, that of P3HT, is about 18 mA/cmz.
However, based on calculations incorporating the absorbance of the electrodes and the
internal quantum efficiency (IQE) of the active layer, the ratio of charge carriers
generated over the incident photons incident on a device,25 the maximtun J SC obtainable
for a P3HT solar cell is 15.4 mA/cm2 at an IQE of 100%. While reports of high IQE

devices exist,2° based on light absorbance of the electrodes and the limited absorbance

spectrum of polymers a more reasonable estimate of IQE is around 80%, which

corresponds to a maximum J SC of about 11 mA/cmz, a very achievable value as indicated

from many previous results.”24

 

Device (“1173112) Voc (V) FF (%) 11 WI)
As Cast, No A] 2.2 0.65 32 0.46
Annealed, no A1 7.9 0.49 44 1.7
As Cast w/LiF 4.2 0.62 45 1.2
Annealed w/LiF 9.2 0.57 54 2.9

 

Table 2.1: Basic parameters for solar cells as cast and annealed both with and without a 1
nm LiF layer between the Al electrode and the active layer.

The Voc of these devices is between 0.49 and .65 V, which is much lower than

the theoretical maximum of the band gap of P3HT, 1.9 V. This limit will never be fully

obtainable with current polymers because of the low dielectric constant and

24

corresponding high exciton binding energy. Once an exciton is created its binding energy
must be overcome to dissociate the electron and hole and this binding energy is lost in the

dissociation step. However, we can see from the devices above that a value of 0.65 V is

obtainable, as that is the VOC of the as cast device with no LiF, but annealing the devices

causes a noticeable drop in Voco Currently the reason for this drop in VOC as device

efficiency is improved is unknown, but it could be due to better alignment of the TI:-
orbitals in the conjugate backbone of the polymer or related to recombination within the
devices. As discussed in great detail in Chapter 4 as device performance is improved an

increase in photoluminescence (PL) is observed. While the reason for this PL increase is

described in Chapter 4, quantifying its affect on Voc is difficult. Equations 2.4 and 2.5
relate Voc with charge carrier recombination but a full evaluation of recombination on

Voc is not the point of this dissertation.

 

Device Rs ((2 cm2) Rs“ (0 cm2)
As Cast, no Al 249 523
Annealed, no Al 55 545
As Cast w/LiF 25 556
Annealed W/LiF 16 476

 

Table 2.2: Series (Rs) and Shunt (RSH) resistances of the four devices shown in Figure
2.8. Annealing and adding a layer of LiF both improve device performance, which is

indicated by a decrease in Rs. However, these processing condition changes do not seem
to affect RSH-

25

Both R5 and RSH, Table 2.2, can be estimated by taking the slopes of the J-V
curve at VOC and ch, respectively. Annealing the devices and adding LiF, both which

improve device performance, lower Rs but do not seem to affect RSH. Seemingly, the

annealing procedure can change the series resistance, which indicates an improvement of
the charge transport properties of the active layer, but do not change the diode behavior
of the device, which is indicated through the shunt resistance. These results suggest that
improvement of the diode behavior of the device, i.e. lowering leakage current, requires
improved materials as opposed to improved device morphologies.

The ideality factor, n, of a device can also be determined to aid in describing the

behavior of a device. In inorganic devices n can be determined by plotting Voc as a

function of Jsc as the light intensity on a device is varied. However, this can only be

done in the case of high fill factor devices where there is little voltage dependence on the
illuminated current, that is the current produced at lower voltages before the upturn

resulting from the diode behavior of a device. Good inorganic devices have illuminated

currents nearly equal to JSC up to voltages of ~80% that of Voc. However, there is a

substantial voltage dependence on the light generated current within the types of devices
used in this dissertation, as can be seen in the curvature at low voltages in Figure 2.8. A
value of n can still be obtained, though, by using the dark current of these devices.

Figure 2.9 is a J-V plot of the annealed device with LiF in Figure 2.8 shown along
with its dark current curve and Figure 2.10 is the derivative of current with respect to
voltage (dJ/dV) of both curves. Polymer based devices have a high voltage dependent

light generated current, but the plot of dJ/dV versus V shows this more obviously.

26

Clearly the illuminated curve will not be suitable for finding a value of II, but the dark
curve is appropriate. Figure 2.11 is a plot of dV/dJ versus .1.1 for the device tested in the
dark, which was used to find the value of n. This plot is obvious after differentiating the

diode equation, eq. 2.2 and setting Jsc=0 for the dark curve of a device. With a

traditional device this plot will give a straight line with slope nkT/q and intercept of R3
for values of J > 0. This device, as all the P3HT:PCBM devices studied in this work,
shows a straight line at values of I close to Voc but the data curves downward at high

values of J.

 

Current Density (mA/cmz)

 

 

 

 

 

Voltage

Figure 2.9: Illuminated and dark currents of a typical P3HT:PCBM solar cell fabricated
with a 1 nm LiF layer between the Al electrode and active layer.

27

 

6 _ o Light g
0 Dark

5 — .9 —

4 — _

dJ/dV (mA / v cmz)

 

 

 

-0.4 -0.2 0.0 0.2
Voltage (V)

Figure 2.10: Derivative of current versus voltage (dJ/dV) of both the illuminated and
dark currents of a typical P3HT:PCBM solar cell fabricated with a 1 nm LiF layer
between the Al electrode and active layer. The dark current shows good diode behavior
which is seen with the flat dJ/dV profile at low V but the illuminated device shows
substantial voltage dependant current values.

28

 

 

   
  
  
 

 

 

 

 

 

 

05 _0-012 T 3 RS = 6.5 Q cm2_
0.011 — 0 ° -
A 0.010 - —
E 0.5 — / a
0.009 — -
N\ 04 0.008 _ —
E ' _0.007 I ' ' _
o 0.02 0.04 0.06 0.08 0.
Z 0 3 _ _
3
E
> _ _.
'0 0-2 n = 1.2
_ 2
0.1— RS-8.9§2cm _
0.0g 1 l l
0 5 10 15 20
-1 2
J (mAl cm )

Figure 2.11: Plot of dV/dJ versus J'1 to determine the ideality factor of an annealed

P3HT:PCBM solar cell with LiF as an interlayer. The main body of the plot displays all
the data obtained for values of J > 0 and plots the straight line at low values of J. The

inset shows a magnified view of the higher current region (i.e. far from Voc). The values

of Rs reported were obtained from each best fit line. They differ because RS is

determined from the dV/dJ intercept, which will change depending on the part of the
dataset chosen for a best fit line.

At values close to VOC, n = 1.2 and moving to higher current values, and hence

higher voltages, the value of It increases to 1.8. It is difficult to ascertain what the value
of n corresponds to in terms of recombination or device properties as most explanations
that exist are based on heuristics for inorganic devices. Yet, one possibility is that these

devices suffer from some form of double diode behavior. As the voltage is increased

above VOC an increase in dV/dJ is seen (the downward curve in dV/dJ vs J'l), indicating

29

more current is able to pass through the device. A double diode could cause such
behavior and evidence for this double diode exists in the structure of these devices, which
is discussed in detail in Chapter 3. After spin coating these devices, a thick layer of
mostly PCBM is found at the substrate. This layer of PCBM could act as a separate
diode to the active layer and produce this type of voltage dependence on n. The other
three devices shown in Figure 2.8 Show the same downturn of dV/dJ at high voltages but
the curves are more pronounced. They are Shown in Appendix C.

Regardless of the actual effects that cause this change of n with respect to voltage,
it is obvious the physics of these devices is complex and still unknown. A more detailed
study of the ideality factor and the conditions that affect it are certainly warranted,
although this testing may prove difficult because of the low fill factors and high voltage

dependence of the light generated current.

30

CHAPTER 3: NANOPARTICLE CONCENTRATION PROFILE IN POLYMER

BASED SOLAR CELLS

3.1 Introduction

Polymer-based solar cells have the potential to supplement the ever increasing societal
energy demands in a cost effective manner due to their low cost and ease of manufacture.
In addition, the device components are readily available consisting of a semiconducting
polymer and a nanoparticle, in our case this is a C60 fullerene derivative.
Commercialization will not be feasible, however, until device performance reaches an
efficiency level approaching that of inorganic cells, ~10%.” Their performance though,
demands control of the nanoscale morphology for the roughly 100 nm thick polymer -
nan0particle active layer as a result of the polymer’s physical properties.

For example, the polymer’s dielectric constant is relatively small creating a strong
Coulombic binding energy of the photogenerated exciton dictating a short lifetime and
small diffusion length. Poor electrical conductivity of these polymers also contributes to
the diffusion length of roughly 5-10 nm.9 Such a length requires the exciton to reach a
dissociation point, the polymer - fullerene interface, to break into its components and
generate free charge carriers. Otherwise it rapidly recombines and produces useless heat
or light.

Furthermore, the morphology must be optimized on different length scales. The
smaller exciton diffusion length requires the system to have polymer rich regions that are
of order 10 nm wide. However, a roughly 100 nm device thickness is required to
maximize light absorption based on the material’s absorptivity. This places strict

requirements on the internal morphology with an ideal device having polymer strands 10

31

nm wide and 100 nm long sandwiched between fullerene regions in a co-continuous
manner allowing both phases direct access to the appropriate electrode.

This morphology must be developed in simple processing steps otherwise the
rationale promoting this energy source is defunct. However, morphological
characterization is challenging and relationships combining processing, morphology and
performance have not been developed to guide device manufacture.

Previously, improved device performance in these types of solar cells involved
much experimentation by empirically changing material properties such as polymer
regioregularity and molecular weight,16 as well as fabrication variables such as solvent

choice,12 spin coating times13 and annealing.”

All of these processing conditions have
been shown to relate directly to the crystallinity of the polymer component of the cell,13
where improved polymer crystallinity provides greater hole mobilities through the
structure, thereby improving device performance.14

While improved crystallinity and the corresponding hole transport properties are
crucial to device performance other factors certainly affect their performance as
discussed above. Traditional polymer-based solar cells are Spin coated or printed from
solution providing an undefined and kinetically trapped mixture of nanoparticles and
polymer. The vast majority of these devices today are created fi'om solutions containing
a roughly 1:1 by weight mixture of polymer and nanoparticle making a homogenous
dispersion of the two components very unlikely due to solubility differences in the

solvent and the likelihood that they are not soluble with each other at such a

concentration.

32

Here we use a mixture of poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl-C61-
butyric acid methyl ester (PCBM) as a standard system to study polymer-based solar
cells. As seen in the energy diagram of an idealized P3HT:PCBM cell, Figure 3.1, good
electrical transport out of the cell only occurs if there is PCBM at the metal electrode and
P3HT at the transparent indium tin oxide (ITO) electrode. Based on the ideal energy
diagram and the short exciton diffusion length, the co-continuous comb like structure
mentioned above would be ideal,ll but this structure would be extremely difficult to make
requiring lithographic methods that would eliminate the potential cost benefits of solution
processable devices. To maintain the fabrication and cost benefits of solution
processable devices, a casting or printing method seems inevitable for commercialization

but determining the PCBM concentration profile within the devices has proven difficult.

 

3.2 eV

 

3.7 eV

 

4.2 eV

 

4.8 eV

 

 

5.0 CV 51 CV

 

 

 

 

 

6.1eV
ITO PEDOT P3HT PCBM Al

Figure 3.1: Idealized energy diagram of a P3HT:PCBM solar cell, showing ideal
alignment of the molecular orbitals and work functions, has the PCBM component of the
cell in contact with the aluminum electrode and the P3HT component in contact with the
hole conducting layer of PEDOT:PSS.

33

A 1:1 blend of these two components in unlikely to provide a well dispersed
mixture at the molecular level. In a previous study we estimated the upper limit of C60
fullerene miscibility in polystyrene was of order 2 vol%.27 This was done by using
experimental results and a simple F lory-Huggins theory to determine an interaction
energy of ~0.02 eV between the fullerene molecule and the monomer group, a reasonable
value. An energy of this order may exist for the system at hand and complete miscibility
is not expected at such a large concentration used in solar cells. Furthermore, the

interparticle, edge-to-edge distance (H) is of order 0.05 nm at this volume fraction ([1)

promoting phase separation. This can be estimated via H/D = [Dm/ 13]” 3 — 1 where D is

the particle diameter (~0.7 nm for C6028) and [3m is the maximum, random packing

fraction (~0.64).29
Even if a blend is miscible in the bulk, a homogeneous mixture still may not be
present in a thin film due to a delicate balance of surface (enthalpic) and entropic

forces.3°’3°

Our previous study demonstrated spin coating produced a layer rich in
fullerenes next to the substrate while none were present elsewhere in the film.37
Processing a film in this way induced segregation since it is known fullerenes are much
less soluble in the spin coating solvent than the polymer.” 39 So, they follow the solvent
concentration gradient to the substrate making a gel-like layer rich in nanoparticles at that
interface.37 Interestingly, as predicted by fluids — density functional theory calculations
performed with hard spheres,3°’ 40’ 4' the (3,0 nanoparticles should be uniformly dispersed
throughout the film which was obtained after following a complicated processing

procedure.42 Of course these films only contained ~ 2 vol% nanoparticles which is much

less than that used here. Regardless, even at 2 vol% the value of H is 1.5 nm and despite

34

a percolated network not being present may still be useful for solar cells because of their
relative closeness.

Despite PCBM particles being more soluble in spin coating solvents,43 and
possibly the polymer itself, one cannot expect them to be well dispersed in the polymer
film at the extreme concentration used in polymer-based photovoltaic devices. Because
of this a balance of forces that creates a non-equilibrium structure is expected,
necessitating careful characterization to understand device performance.

A number of groups have used a variety of techniques in structural

50

characterization“ including: scanning electron microscopy (SEM),51 transmission

5° ellipsometry52 and X-ray reflectivity,47 with

electron microscopy (TEM),49'
contradictory or conflicting results. SEM, TEM and X-rays all scatter from electrons
and, due to the high amount of carbon, all produce quite low contrast between the two
components. X-ray reflectivity has been shown to be an unsuitable method to
characterize the morphology within a polymer-based solar cell since no component
contrast exists.47 The previous SEM study used a different system, MDMO-PPV and
PCBM, and inVolves breaking the film, which induces undesirable defects. In addition,
TEM images of randomly oriented, nanoscale, low contrast materials has been shown to
be ineffective.“ 53’ 5“

Defocused TEM can be used to measure the microstructure of materials with
regular variation of morphology if the components segregate into large enough regions,
say ~10 nm and greater.“ 4° However, to do this with less ordered materials is a

challenge. TEM tomography may be used to characterize the system at hand, P3HT and

PCBM, since P3HT crystallites introduce contrast within the system,48 but there is

35

minimal, if any contrast between amorphous P3HT and PCBM. It is possible to use
ellipsometry, yet, this is a difficult technique to apply to our system, although some
interesting results were found revealing an excess of PCBM at the substrate.52

X—ray photoelectron spectroscopy55 (XPS) and near edge X—ray absorption fine
structure spectroscopy (NEXAFS)5° results, which use electron orbitals and band
structure to determine elemental compositions, show high concentrations of PCBM at an
Si02 substrate directly after Spin coating. These methods are an excellent way to
measure amounts of material near interfaces where charge is removed from a device and
their results are consistent with the results discussed here despite the differences in
sample preparation methods. However, these methods do not provide concentration
versus depth information and more information could be gained on internal device
structure.

Following a recent defocused TEM study, which suggests a uniform dispersion of
PCBM,50 we Show an example of defocused TEM micrographs for a 1:1 blend of P3HT
and PCBM and a pure polystyrene film, both sectioned by rnicrotomy, in Figures 2a and

44’ 54 and calls to

2b. This pictorially reveals the challenge in analyzing such images
question the above results. Another technique, which uses TEM and scatters from the
P3HT crystallites, shows there are more P3HT crystallites at the substrate,48 while
ellipsometry reveals there is more PCBM at the substrate.52 Clearly there is need to

carefully characterize P3HT-based solar cells since their performance is dictated by their

morphology.

36

Figure 3.2: Defocused transmission electron microscopy cross sectional images of
polymer samples prepared by ultramicrotomy Shows little difference for a variety of
materials. A 1:1 by weight PCBM:P3HT blend (a) and pure polystyrene (b) both have
lighter and darker regions demonstrating the difficulty in interpreting such images.

3.2 Neutron Refelctometry

3.2.1 Specular Neutron Reflectometry

We recognize that C60 fullerenes have a large neutron scattering length density (SLD)

difference to polystyrene,” 42

and by using simple techniques it is possible to estimate
that P3HT should have significant contrast from PCBM. Thus, neutron reflectivity is
well suited to characterize the active layer, especially given that it has excellent
resolution down to the nanometer scale.

Neutron reflectivity makes use of the nuclear cross section of the film components
to determine a depth dependant scattering length density (SLD) and hence the film
composition profile. In addition, a large sample volume is probed to yield a true
analytical average. Calculation of the individual component SLDs is accomplished by

summing the individual nuclear cross section of all the atoms in one molecule and

dividing by the molar volume, which is proportional to density,

37

SLD = Zffi (3.1)

where bc is the bound coherent scattering length of the ith of n atoms in a molecule and

the molecular volume, Vm, equals the molecular weight divided by the density.

Neutron reflectometry can be a difficult experiment because while one SLD
profile can only produce one set of reflectivity data the inverse is generally not true.57
Because only the amplitude of the scattered neutrons is measured in a reflectivity
experiment, the phase information in the scattered neutrons is lost. This loss in phase
information results in one set of reflectivity data being able to produce multiple SLD
solution profiles during fitting. Additionally, when fitting with the Parratt formalism58 a
least squares minimization is used to determine an overall profile, which can provide
multiple ‘correct’ solutions to the data of films more complex than those consisting of
only pure components. While retaining the phase information in one experiment is not
possible, the phase information can be retrieved by varying the surrounding media during

multiple reflectivity experiments.

3.2.2 Phase Sensitive Neutron Reflectometry

Phase-sensitive neutron reflectometry (PSNR) methods make it possible, through
the use of variable reference films or substrates, to uniquely determine the complex
reflection amplitude of an adjacent, ‘unknown,’ layered structure. This, in turn, can be

directly inverted to yield its corresponding, unambiguous scattering length density (SLD)

38

profile. An accessible review of neutron reflectometry (NR) techniques in general, and
PSNR in particular, is given by Berk et al.59 There are two categories of reference layers
applicable in PSNR, layers of finite thickness with variable SLD°°’ °l or surrounding
media°2 of adjustable SLD. Either the fronting or backing medium may serve as the
supporting substrate for a thin film system of interest and the reference layer may also be
either the fronting or backing medium. The ultimate sensitivity is achieved by
performing phase-sensitive NR measurements for each of a number of different labelings
of a component part. The work discussed here used a surrounding backing media with

adjustable SLD, which will be discussed further in the Experimental Section.

3.3 Experimental Methods

3.3.1 Neutron Reflectometry Samples

To make the PSNR samples, both P3HT and PCBM were weighed in a 1:1 by weight
ratio in air. The polymer - nanoparticle mixture was then transferred to a nitrogen glove
box and chlorobenzene was added such that a total solution concentration of 30 mg/ml
was obtained (15 mg P3HT and 15 mg PCBM per 1 ml of solvent). The solution was
stirred for 3 days inside the glove box and then filtered through a 200 nm PTFE filter.
Two samples for the PSNR tests were spin coated onto 75.2 mm diameter, 5 mm thick
polished silicon slabs; one at 2500 RPMs for 60 seconds, denoted ‘Fast Grown’ and one

at 800 RPMs for 180 seconds, denoted ‘Slow Grown.’ The solution was first heated to

50°C before spin coating for the Slow Grown sample to achieve a flat active layer.

Visual inspection of both of these un-annealed samples showed minor defects, with the

total defect area amounting to a relatively small fraction of the total area illuminated by

39

the incident neutron beam. Thus, based on this optical criterion, these two samples were
deemed to be of potentially sufficient in-plane homogeneity to warrant further study by
PSNR.

Unfortunately, large agglomerates of PCBM form during thermal annealing of
P3HT:PCBM films on silicon substrates63 so annealed samples could not be studied on
the same wafers. Instead, silicon wafers were coated with a thin layer of the
macromolecular salt poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)
(PEDOT:PSS) upon which the active layer was Spin coated. PEDOT:PSS is a common
hole conducting material used in organic solar cells to prevent surface recombination and
increase device performance. Coating the Si wafers with PEDOT:PSS prevents the
PCBM from massively phase separating upon annealing and allows for a good
comparison to an actual polymer based solar cell. One of the drawbacks of using

PEDOT:PSS, however, is that it is water soluble and makes PSNR experiments very

difficult. The PSNR experiments described below use deuterium oxide (D20) as the

backing medium due to its high SLD and water-like properties. PEDOT:PSS,

unfortunately, dissolves readily in D20 so using D20 as a backing medium ruins any

sample.

The pure P3HT and PCBM samples were created following the same weighing
procedure using only pure components. Pure P3HT samples were spin coated at 2500
RPMs and the pure PCBM was spin coated at 2000 RPM, both at a concentration of 20
mg/ml in chlorobenzene. In this case though the silicon wafers were 50.8 mm in

diameter and 1 mm thick. The larger silicon slabs were not needed for these tests as the

40

pure components did not require PSNR calculations nor the sample holder described
below due to the low complexity of the films.

All samples were transported from the University of Delaware to NIST in a
nitrogen filled vacuum dessicator to ensure no oxygen degradation. However,
subsequent reflectivity measurements showed no change in SLD or film profile after

many days exposure of the films to oxygen.

3.3.2 Phase Sensitive Neutron Reflectometry Measurements

PSNR measurements were performed on the NGl polarized neutron reflectometer at
NIST with an incident wavelength of 4.75A and 1.5% or less spread. A wet cell holder,
Figure. 3, allowed for easy changing of the backing media. Wafers were inserted into the

wet cell, aligned on the reflectometer and data was taken with air as the backing media.

After the air backing data was gathered, degassed D20 was slowly injected into the wet

cell as the new backing media and a new data set was taken. No realigning was

performed upon addition of the D20 but rocking curves were periodically taken during

the course of both the air and D20 backing experiments to ensure no misalignment

occurred.

41

 

Figure 3.3: A cross sectional image of the wet cell holder used for Phase Sensitive
Neutron Reflectivity measurements. The path of the neutron beam enters through a
silicon ‘fronting’, encounters the film being studied and exits through to the ‘backing’

reservoir, which was either air or D20 for this experiment.

For all of the measurements the slit apertures were opened with increasing Q,
once sufficiently above the critical angle region, so as to maintain an approximately
constant ‘footprint’ of illumination on the sample surface. This footprint measured about
25 mm in height and 40 mm in length along the horizontal. Because of the large silicon
substrates, 75.2 mm diameter, no footprint correction was needed for data reduction of
the PSNR samples. A slight footprint adjustment was needed for the low Q values on the
pure component samples because those measurements were taken on 50.8 mm diameter
substrates.

The reflectivity was obtained by subtracting background from the measured

reflected intensity and subsequently dividing by the incident intensity appr0priate for the

42

set of aperture widths at a given value of Q. Given the beam time available to us, it was

possible to collect reflectivity data for the Fast Grown sample with both air and water

references out to a Q of nearly 0.2 A]. For the Slow Grown sample, it was possible to

collect reflectivity data out to 0.2 A'1 for the D20 backing but only to about 0.1 A-1 for

the air backing.

3.3.3 Density Measurements

Density measurements were performed by mixing 20 wt% PCBM with polystyrene and

precipitating the blend from solution in a non-solvent. After filtering and drying in a

vacuum oven for 5 days, the mixture was pressed into a pellet at 170°C for 2 hr under

vacuum and then measured using a pycnometer.

3.3.4 Solar Cell Fabrication and Testing

Solar cells were prepared using the same 1:1 by weight blend of P3HT and PCBM (30

mg ml") in chlorobenzene (1 ml) as in the reflectivity samples except ITO coated glass
slides with a sheet resistance of 8-12 Q/cm'2 (Delta Technologies) were used instead
silicon wafers. For more accurate testing and creation of multiple devices on one
substrate, the ITO was patterned using an HCl etch before device fabrication. The etched
substrates were cleaned by ultrasonication in acetone and then isopropyl alcohol for 15

minutes each. Then A PEDOT:PSS (H.C. Starck, Clevios 4083) solution in a 4:1

volumetric ratio of PEDOT:PSS stock solution to water was spin coated onto each

43

substrate at 3000 RPMS for 30 seconds and then the substrates were dried in an oven at

130°C for 30 minutes.

The coated substrates were then transferred into a nitrogen glove box where the
active layer was spin coated at 2500 RPMS for 60 s. The coated substrates were

transferred to a thermal evaporator and an aluminum back electrode was thermally

deposited under a vacuum of 5x10.7 torr at a rate of 0.2 nm sec]. The devices were tested

using a 150 W Oriel class A solar simulator under AM 1.5 conditions at 1 sun
illumination as determined by a calibrated reference solar cell traceable to the National

Renewable Energy Laboratory. Glass slides (25.4 x 25.4 mm) were first coated with a
roughly 25 nm layer of PEDOT:PSS, which were baked at 130°C in air for 30 minutes
and then transferred to a nitrogen glove box. The ‘P3HT:PCBM solution was spin coated
onto the PEDOT:PSS coated substrates and the annealed samples were heated at 140°C

for 20 minutes.

3.4 Results and Discussions

3.4.1 Pure Component Films

Figure 4a is reflectivity data plotted as reflectance (R) times wave vector (Q) to the fourth

power versus wave vector and associated best fits to the data for a pure P3HT film spin

. 4 .
coated at 2500. The convention RQ 18 used because reflectance decreases at a rate of

Q'4 for systems with sharp interfaces at high Q. Shown in the figure are two fits to the

data, one fit for a constant density of P3HT through the entire film (gray line) and the

other for variable density throughout the film with a slightly higher density at the

substrate that decays slightly as it approaches the air interface (black line). Although the
fits are not substantially different, the varied density model fits the data noticeably better.
The density profile of each fit, shown in Figure 4b, is realistic for both fits as reported
densities of P3HT range from 1.1 to 1.3 g/cc,64 but the roughness in the varied model is
more realistic. In the variable density profile the P3HT density starts at roughly 1.17 g/cc

and drops to slightly below 1.0 g/cc at the air interface. The average density from this

profile is 1.15 g/cc, which corresponds to an SLD of 0.74>< 10.6 A'2 for as cast P3HT.

1 l l i

 

 

 

 

 

 

 

 

 

 

 

1-0 I ' I ' I ' I
O AsCastP3HT Data 1 _ 14
— Constant Density Fit . '
— Increased Substrate : 0-8 " r- 1.2
Density Fit
1 “-‘ - 1.0
< 0.6 r-
1 ‘90 - 0.8
1 ‘-
. . , ‘ 0.4— . . - 0.6
a. “tam, q —ConstantDensrty Flt
"3 " “I'll "l" — Increased Substrate — o_4
"' __ Dens' Fit
. I I 0.2 W _ 0.2
I r I I I . I r I r L 0.0
0.02 0.04 0.06 0.08 0 200 400 600
Q (K1) <<- TOP Distance through film (A) BOTTOM—>>

Figure 3.4: (a) Reflectance (R) times wave vector (q) to the fourth power versus wave
vector for a pure P3HT film spin coated at 2500 RPMS with associated best fits and (b)
the SLD and corresponding density profile of the best fit lines. A varying density profile

(black line) fits the data much better than the constant density profile (gray line),
indicating the density of the P3 HT film varies slightly as a function of depth.

Annealing the P3HT films, as shown in Figure 5b, noticeable changes the density
profile such that it is larger near the substrate. As shown in Figure 5b, the pure P3HT
fihn is much more dense at the substrate, roughly 1.4 g/cc, but quickly drops to about

1.15 g/cc. This higher density at the substrate corresponds to previous work performed

45

by van Bavel et al. where they show a noticeable increase in P3HT crystallinity at an Si

substrate.65 Based on the both the as cast and annealed fihns we use a P3HT density of

1.15 g/cc and the corresponding SLD of 0.74>< 10'6 A'2 for the remainder of this work.

 

 

 

 

 

 

 

 

 

 

I 1 a 1 3 1-0 a I I I _ 1.6
. o Annealed P3HT Data 0 8 _ 21.4
— Constant Density Fit ' / __ 1 2
“ — Increased Substrate N '
A _ a: Density Fit ' _
‘7 9 °< 0.6 — 1.0
“S, 1° “ ‘9
Va I g ‘0- "' 0.8
m I ‘: . h 1... q 04 '— — 0.6
i s. i y: p. '5'}; If; — Constant Density Profile
' .r , "3.1 i; I“: ‘ . 02 l‘ — Increased Substrate '0-4
10'1° ‘, i ' ll '1 1" Density Profile _ 0.2
i 1 l l -3 00 l l l 00
20 4O 60 80x10 0 200 400 600
Q (A") <<-- TOP Distance Through Film (A) BOTTOM->>

Figure 3.5: (a) Reflectance (R) times wave vector (q) to the fourth power versus wave
vector for a pure P3HT film spin coated at 2500 RPMS and annealed at 140°C for 20
minutes with associated best fits and (b) the SLD and corresponding density profile of the
best fit lines. A varying density profile (black line) fits the data much better than the
constant density profile (gray line), indicating the density of the P3HT film varies slightly
as a function of depth. A noticeable increase in density is seen at the substrate.

Reflectance data from films spin coated at 800 RPMS showed similar results to
the 2500 RPM films although, due to the thickness of the films, the data is much noisier.
The 800 RPM films were approximately 130 nm thick whereas the 2500 RPM films were
only 60-70 nm thick. As the thickness of a film increases its reflectance fiinges are
moved closer together and total reflectance intensity drops, decreasing the signal to noise

ratio making thicker films more difficult to fit.

46

Reflectance versus wave vector data and best fit lines for a film of pure PCBM is
shown in Figure 6, which corresponds to a PCBM SLD of 3.6><10.6 A.2 and density of

1.25 g/cc. This is a noticeably lower value than that reported in the literature for an FCC
C60 crystal (1.68 g/cc)°° and slightly lower than a previously reported value for bulk
PCBM (~1.5 g/cc),67 however, this value is in reference to a private communication. The

dashed line in Figure 6 shows the calculated reflectance for a PCBM film with a density

 

 

 

 

 

 

 

ofl.5 g/cc.
A I
0 Data
10'8 r I \ — Density of 1.25 g/ml -_
7. .x \ - - DenSIty of 1.5 g/ml .
' \ A _
5’ I \ / ~
4L ' I 1 \.
A " . I :'\ l . ..... _.I
1: 3 .. ' ‘ . " . '
f“ l. ‘. I I, ’
c 2' ‘ i II h“ ..
m 1!. ‘ ‘ I 7.
10'9 ..... | ......... in; 1
7 i . "g“ j
6 5.: a
5 I " I I I ‘

 

 

0.02 0.04 0.06 0.08
0(A”)

Figure 3.6: Reflectance (R) times wave vector (Q) to the fourth power versus wave
vector for pure PCBM. The solid dark line is the best fit line to the data, corresponding
to a density of 1.25 g/cc and the grey dashed line is the calculated reflectance for a

PCBM film with a density of 1.5 g/cc.

47

To further investigate this discontinuity between the reported PCBM density and
our results we used Archimedes principle, described above in the Experimental Section,
which compares the dry weight of an object to its weight in water and found a PCBM
density of 1.33 g/cc. For all neutron scattering work in this dissertation we use a PCBM
density of 1.3 g/cc, a P3HT density of 1.15 g/cc and the associated SLD’s to model the
reflectivity profiles and the SANS data in later chapters. Slight variation of the PCBM

density within the above limits does not change our conclusions.

3.4.2 Solar Cell Mimics

Reflectivity multiplied by wave vector to the fourth power (RQ’) is plotted as a
function of Q in Fig. Sa for the film spin coated at 2500 RPM with both air and D20
backings. We use this parameter pairing, RQ’, since reflectivity decays as Q'4 for
systems with sharp interfaces as well as for ease of data viewing. These fits Show an
excellent representation of the data out to a Q of almost 0.19 A". Corresponding SLD
plots of the simultaneous fit, the exact solution from PSNR and the Parratt58 method
applied to the air backed sample are shown in Fig. 5b. The agreement is remarkable,
both quantitatively and qualitatively. A small SLD peak is seen at the substrate of the
simultaneous fit solution, which is a native silicon dioxide layer imposed as a constraint
during the simultaneous fitting (Fig. Sc). One knows the layer is present but is too small
to be seen in the PSNR due to the limits of Q space investigated. The divergence at the
air interface of the simultaneous fit is due to the difference in backing media, D20 versus

air (Fig. 5d).

48

A difference between the PSNR and dual fitting profiles can be seen, most
notably at the substrate interface. In principle, this PSNR solution is unique for the
reflectivity data collected over the entire range of Q and in the absence of statistical
uncertainty in the data. In practice, statistical uncertainty in and truncation of the
reflectivity data sets at some maximum, finite value of Q, introducing a corresponding
degree of uncertainty in the SLD profile.68 The SLD density profiles obtained by direct
inversion and simultaneous fitting of the same two composite reflectivity data sets must,
therefore, be equivalent within the extent allowed by truncation and statistical
uncertainty.

The differences in the SLD profiles observed in Figure 7b are, in fact, a measure
of the confidence in the profile. The reason the simultaneously fit for the SLD appears
more accurate in some details, such as in the resolution of an oxide layer at the silicon /
polymer interface, is that in the modeling it is possible to incorporate additional
information known about the film profile from other sources. The direct inversion
process, on the other hand, does not allow for this possibility. In other words, the direct
PSNR method tells us all that is possible to know from the truncated and statistically
uncertain reflectivity data alone but the dual fitting allows us to incorporate known
parameters within the film.

We were unable to obtain a high enough Q space for the air backing case to
perform the PSNR calculations on the 800 RPM sample, but Simultaneous fitting of both
data sets shows a very similar PCBM profile between the 2500 RPM and 800 RPM case,
Figure 8. An increase in PCBM concentration at the substrate and near the air interface

was found.

49

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

3 I I I I I I r I I
.82 a o 020
10 __ j m u a Air
4 . . . 1 L. i . I - ... 7‘; ' I“ «I
YA 2 — . % I m , 0° Q 0- n." 1.0-Ti“ E c.‘
$10-9 . I I . ~- 1
v . _ . 5‘ ;
C3 11. =5 .. , | I as 3:: :
0: 4 -. r a, Ill" , , l- r , , p . .l-
2 I . r: n ninth l. w III ~ I :
-10 ' '.' - ti '1 fr" ' I i'lllr u '
10 ’ '~.i ‘1! I" II ‘.I'
4 l l I l . . .. l ..................... l . .
0.02 0.04 0.06 0.08 0.10 0.12 0 14 0.16 0.18
0(A")
3.5 3.5 r
g 2.5e
2.5
....A % 2.0+ :
.3“ 2-0 1.5 2 I
53 -4 0 4 8 12
5 1,5 14 Distance (nm)
(7|) +PSNR q," - I ' I
1.0 ._ +Sim.Fit-Air ~< 1-2 : -
-I- Sim. Fit-D20 “Po 1.0 :
0 5 _ + Parratt Fit-Air C 03 i
' o
.1 0.6 . _
00 I I I l I ‘0 0.4 r : r :
' 0 20 40 60 80 840 860 880 900
Substrate Distance (nm) Air Distance (nm)

Figure 3.7: Neutron reflectivity uses the natural contrast between P3HT and PCBM to
determine the scattering length density (SLD) and eventually the concentration profile.
(a) Reflectivity multiplied by wave vector to the fourth power (RQ’) versus wave vector
for the sample spin coated at 2500 RPM using either an air (triangles) or D20 (circles)
backing and associated best fits found though simultaneous fitting of both datasets. The
deuterium allows characterization to larger wave vector and more accurate determination
of the SLD profile. (b) The best fits of the reflectivity data for the SLD profile were
found via the PSNR calculations as well as simultaneous fitting of both data sets (Sim.
Fit-Air & Sim. Fit D20) and the Parratt formalism for the air backed sample (Parratt F it-
Air). An expansion of the distance axis shows (c) the SLD near the substrate (vertical
dotted line) where the SLD bump due to silicon dioxide in the simultaneous fit is clearly
visible and (d) the SLD near the air interface (vertical dotted lines) which is quite similar
to the SLD of pure P3HT (horizontal dotted line) within the first 2 nm. The sample was
unannealed and characterized directly after spin coating.

50

Normalizing the thicknesses of each film, by multiplying Q with thickness,
though, shows astonishing agreement between the two spin coating conditions as seen in
Figure 3.9a. Here we have plotted the volumetric percentage of PCBM for just the active
layer of each case, 2500 RPM and 800 RPM, showing agreement between the coating
conditions as well as fitting procedure. A cartoon of possible morphologies is shown in

Figure 3% based on this concentration profile.

 

 

 
     
    

 

 

 

 

 

 

 

 

6x10‘ L ' ' ' ' ' L
.8 ' NA — Air Backing Profile b j
10 F :55 5 g - - - DZO Backing Profile I_
E g i
h I
YA -9 g 4 - r
-< 10 0 '
vv 5 3 :_
U)
8 , g .
_I 2 '
10"° I .3 .'
CT} I
I A Air Backing I g 1
’ o 020; Backing;5 I 0 I L l
2 8 9o 01 0 400 800 1200
CJ:(A ) Distance through film (A)

Figure 3.8: (a) Composite reflectivity data sets obtained for the Slow Grown film and fits
which were determined simultaneously and (b) the Slow Grown film SLD profile
obtained by simultaneous fitting of the corresponding data. As the SLD profiles were
simultaneously fitted the profiles overlap until the backing layer was reached and they
differ with the air backing film going to SLD = 0x10'°A'2 and the D20 backing film
going to an SLD = 6.2x10’° A".

Before discussing specifics of the profiles, a comment must be made concerning
the concentration profile. The pure component SLD’s discussed above were used to

convert the SLD profile to concentration by a volumetric rule of mixtures. Although the

51

mass balance was not enforced in the simultaneous fitting routine, and certainly not in the
PSNR calculations, by integrating the concentration curve through the film we found the
average concentration of PCBM was precisely conserved with the PSNR calculations and
within ~ 5 vol% of the expected value, 48 vol%, for the other cases. This is a significant
confirmation of the fitting procedure.

A higher SLD indicates a larger amount of PCBM, so the SLD peaks found in
each film are directly related to PCBM agglomerating near each interface. Concentration

is directly related to the SLD and can be found using a simple equation

_ SLDMeasured-SLDPBHT

— (3.2)
PCBM SLDPCBM-SLDpsHT

where D is the volume fraction of PCBM and SLDpCBM and SLDp3m~ are the SLD’s of
the two pure components determined in separate experiments (see Fig. 2). The maximum
concentration of PCBM at the substrate and near the air interface is 74% and 65% by
volume, respectively. A slight difference can be seen between the PSNR direct inversion
technique and the simultaneous fitting, particularly near the substrate, however, this does
not significantly diminish conclusions we can gather from this data. For example, the
energy diagram, shown in Figure 3.1, indicates that PCBM should be at the air interface
(the metal electrode) and not at the substrate (the conducting oxide). Clearly, both
characterization techniques demonstrate that a high concentration of PCBM is at an

undesirable location.

52

 

j r F l

Unannealed
+ 800 RPM
-l- 2500 RPM

        
 
  

I: "a;

O n
0.94 0.96 0.98 1.00
l i |

Volume Percent PCBM

  

 

 

l
0.0 0.2 0.4 0.6 0.8 1.0
Normalized Distance

 

 

 

I PCBM :73 P3HT

..—J

Figure 3.9: The concentration determined from neutron reflectivity shows a large amount
of PCBM near both interfaces. (a) Spin coating at either 800 or 2500 RPM yields very
similar concentration profiles, determined from the SLD profile, with slight differences
near the interfaces (see inset). The film thicknesses were 140 nm (800 RPM) and 90 nm
(2500 RPM) and used to normalize the distance through the film. (b) A cartoon of a
possible PCBM/P3HT morphology demonstrating higher PCBM concentration near the
interfaces with a large P3HT concentration within 3 nm of the air interface (see Fig. 3.1).

53

The PCBM nanoparticles clearly concentrate at the substrate and near the air
interface as shown in Figure 9. This inhomogeneous dispersion appears to be
independent of spin coating speed, i.e.: fast versus slow growth conditions, and we
believe results from the balance of a number of different forces. We hypothesize the
PCBM particles migrate to the solvent rich area near the solid substrate during spin
coating, because of their limited solubility, where they eventually become trapped.37’ 42
However, it is believed the spin coating speed is high enough in both cases to induce a
rapid solvent evaporation rate to convect the PCBM nanoparticles towards the air
interface. While convective flow forces the PCBM towards the air, the
nanoparticle/polymer system will tend to assemble according to its optimum dielectric
ordering.32 Since the refractive index of P3HT is approximately 1.969 and that of C60 is
about 2.0,70 which we assume is close to that of PCBM, this will induce a preference for
P3HT at the air interface. Of course, polymer crystallites formed in the process will
affect the PCBM concentration profile as will nanoparticle/polymer entropic effects we
discussed above, although this is conjecture at this point. Suffice to say the process and
its influence on the resulting morphology is complicated and yields an interesting
concentration profile.

I." in Table 1 demonstrating

Our results are compared to those of Xu et a
agreement between both studies. Using results from simultaneous fitting of the
reflectivity profiles with an air or deuterium backing (Sim. Fit) we find approximately 75
vol% PCBM at the substrate interface while there is much less, ~ 0 — 5 vol%, at the air

interface. These concentrations are essentially independent of spin coating speed,

although one could argue there is more PCBM at the air interface for the lower spin

54

coating speed. To remove some of the arbitrariness in defining the interface, we
averaged over a length scale of 10 nm, which allows comparison to the X-ray
photoelectron spectroscopy (XPS) results of Xu et al., since this technique typically
probes this distance in the sample. Here clear agreement is seen between all samples and
characterization techniques. On average there is 34 d: 4 vol% near the air interface (~ 10

nm) while there is 59 :t 8 vol% near the substrate interface.

 

 

Interface 800 RPM 2500 RPM 2500 RPM Slow Fast
Sim. Fit Sim. Fit PSNR Xu et a1 Xu et al.

Air 8% 0.2% 0.2% - -

Substrate 75% 77% 65% - -

Air (10 nm) 28% 37% 37% 34% 34%

Substrate (10 nm) 70% 65% 53% 56% 53%

 

 

Table 3.1: PCBM volume percentages at the air interface, substrate interface and
averaged over a 10 nm depth at both interfaces for comparison to the XPS data of Xu et
al." The two spin coating speeds, 800 and 2500 RPM, were used for comparison to the
Slow and F ast evaporation conditions used by Xu et al. The concentrations for the
sample prepared at 2500 RPM was determined by simultaneous fitting of the two profiles
(Sim. Fit) as well as with PSNR while the 800 RPM concentrations were determined
solely by simultaneous fitting. The data are from Figure. 3.5.

Additionally, we note Germack et al.56 performed NEXAF S, which has a
penetration depth slightly less than XPS, on the Si02 substrate interface of a similarly
fabricated bulk heterojunction device. Their results show an interfacial concentration of
82 vol% PCBM at the substrate in strong agreement with our results.

The rather large PCBM concentration at the solid substrate is clearly undesirable,
since, typically, holes are conducted through it. Ideally an inverted concentration profile

should be present, described in Figure 1, with a larger PCBM concentration at the air

55

interface since it is an electron transporter. In addition, the layer of almost pure P3HT at
the air interface is likewise detrimental to cell performance.

We mimicked a solar cell on a silicon wafer substrate by spin coating a 30 nm
thick layer of PEDOT:PSS onto it and then, subsequently, a 90 nm thick layer of
PCBM:P3HT at 2500 PRM. It is clear that surface energy can affect the interfacial
concentration,56 so, the lower energy PEDOT:PSS was used to ensure prOper comparison
to an active solar cell. As shown in Figure 10a, a similar concentration profile was found
to our prior results for films that were not annealed. We note in this case the profile was
determined with the Parratt technique with an air interface, as discussed above the fit
results followed the mass balance.

The film was annealed at 140°C for 20 min. where it was found the concentration
profile changes. Interestingly, the amount of PCBM increased at the air interface which
is clearly desirable for a more efficient solar cell. However, very little change in the
concentration occurs at the substrate which will not promote more efficient function. The
concentration peaks broaden slightly, however, gross change in the concentration is not
possible since the melting temperature is much higher, ~220°C. Rather it is more likely
crystal refinement occurs to exclude PCBM which is forced to the air interface. This is
not energetically favorable as it does not follow optimal dielectric ordering and is
certainly a metastable state induced by the processing procedure.

A working solar cell was fabricated on an ITO covered glass and tested. As seen
by others, the cell performance substantially increases after annealing (Figure 3.7b) and

could certainly be affected by an increase in crystallite size and number after annealing.

56

However, our study introduces another possible factor through an increase in PCBM

concentration at the air layer (back electrode).

 

 

 

   

 

 

    

 

 

 

 

 

 

100
b
2 80 NA 0
m E
U o
O. 2 _2 ...
E 60 g 0'"
a:
E 2 4 —
a, 40 3
g “E
E, 20 _ 2500 RPM §-6 _ 2500 RPM -
+ Unannealed o -o- Unannealed
-l- Annealed -I- Annealed
0 1 l l 1 -8 l . I . I
0.0 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6
Substrate Normalized Distance Air Voltage (V)

Figure 3.10: The PCBM concentration profile changes upon annealing by broadening the
peaks near the interfaces which certainly influences solar cell performance. (a) The
PCBM concentration versus normalized distance graph for a PCBM/P3HT blend spin
coated onto a PEDOT:PSS layer shows peak concentrations near the interfaces for the
unannealed (84 nm thick) and annealed (75 nm thick) cases. The PCBM concentration
peaks widen at the interfaces and the PCBM concentration increases near the air interface
after annealing. The Parratt formalism was used in all cases to invert the reflectivity
profiles and “Substrate” indicates the top of the PEDOT/PSS layer. (b) The current
density — voltage curve changes between the unannealed and annealed conditions for a
solar cell and is certainly, in part, due to the change in the concentration profile. The film
thicknesses were 90 nm for both cases and the efficiency, fill factor and open circuit
voltage were 1.7%, 0.44 and 0.5 V, respectively, for the annealed cell.

Indeed, based solely on the concentration profile, it would appear that since the
concentration does not substantially change at the substrate, the solar cell performance is
dictated through an increase in the PCBM concentration at the back electrode. Of course,

other factors most likely come into play although we suggest a change of the

57

concentration, particularly near the back electrode, is a certain factor to increase

performance.

3.5 Conclusions

We have fabricated polymer-based solar cells in this study and used the large
neutron scattering length density difference between PCBM and P3HT to characterize the
devices’ morphology. The key phenomenon that must be addressed in a photovoltaic
device of this sort is, does the processing introduce a nanostructured morphology to
enhance exciton dissociation? Indeed the physics of the device indicates the need for
PCBM to be present at the metal interface and P3HT to be in contact with the
PEDOT:PSS layer on the conductive oxide. Our reflectometry study yields the
concentration profile of the acceptor and donor components, which was measured
precisely and without question as to its accuracy.

Since the equilibrium concentration profile in this device should be two layers of
completely phase separated components, any profile that is developed will be non-
equilibrium. This places an extraordinary burden on manufacture to find the optimum
processing parameters that optimize device performance. To date, empirical heuristics
were developed to fabricate devices (usually) relying solely on a single metric, that of the
photovoltaic efficiency, to guide subsequent studies.

Here we have determined the concentration profile is complicated and varies
through the film thickness. One processing parameter which positively affects device
performance is annealing which we find changes the profile, however, not significantly.

Yet, the interfacial concentration near the back electrode is advantageously changed

58

promoting our conclusion that subtle variations in the morphology are key to a more
satisfactory device. In the future we will use other neutron scattering geometries to
determine the size scale of structures which will certainly aid development of this

technology.

59

CHAPTER 4: PCBM AGGLOMERATION IN POLYMER BASED SOLAR

CELLS

4.1 Introduction
With the promise of low manufacturing costs, high production throughput and flexible
devices, polymer based solar cells are currently being rapidly developed and evaluated.
A device of roughly 1:1 by mass blend of poly(3-hexylthiophene) (P3HT) and phenyl-
C61-butyric acid methyl ester (PCBM) has many literature reports of efficiencies within
the 4-5% range.” 26 However, the potential low cost and ubiquitous applications of a
polymer based photovoltaic device capable of only 5% incident light to electricity
conversion in a laboratory setting is unlikely to become a commercially viable product.

Increasing the efficiency of these devices is difficult because their performance is
governed by the nanoscale morphology of the thin (~ 100-200 nm) active layer, because a
Coulombically bound exciton, or electron-hole pair, is formed upon light absorption. The
interaction energy for two particles varies inversely with the dielectric constant of the
surrounding medium to the second power, so, particles in a lower dielectric media, such
as those in a polymer matrix, have a greater binding energy than particles in a higher
dielectric material, such as an inorganic semiconductor. This effect limits the time an
electron-hole pair will stay in an excited state and therefore shortens the exciton diffusion
length in low dielectric media, one of the most damaging characteristics of polymer based
solar cells, and performance is directly linked to morphological details.

Of course, a high binding energy could be overcome by synthesizing polymers
with much higher dielectric constants, a difficult endeavor, or by astutely controlling the

thin film nanoscale morphology so that all created excitons will reach a dissociation

60

interface. A number of groups have measured the exciton diffirsion length of P3HT,
reporting values ranging from 2 to 8 nm.”‘ 73 Based on this diffusion length and the
roughly 200 nm active layer thickness, a comb-like structure would be ideal, having
PCBM columns separated by ~10 nm spanning the entire film thickness.11 Fabricating
such a device, however, would require expensive and time consuming lithographic
techniques that have yet to be developed and would eliminate much of the cost savings
associated with solution processing. Thus, an optimized solution processable bulk
heterojunction must be used to create a cost effective polymer based solar cell.

Initial work on improving device performance was focused on manufacturing
techniques such as choice of solvent,74 thermal annealing,ls solvent annealing[3 and

changing the polymer to electron acceptor ratio.75

These processing variations have
resulted in improved devices, but a complete understanding of what promotes
improvement is not truly understood. A number of groups have shown correlations

'3’ 76’ 77 which leads to

between processing parameters and enhanced P3HT crystallinity,
improved device performance from an augmentation in hole transport.78 While
improving the crystal structure of P3HT, and therefore hole transport, is widely accepted
to enhance device performance, there is a limit to the total improvement that can be
gained via this strategy. For example, as hole mobility is increased, by using a greater
proportion of P3HT, a corresponding drop in electron mobility is seen due to the lack of
electron transporter available which adversely affects overall perfomance.79 Clearly the

device properties are dictated by the polymer and nanoparticle ratios and their mixing

necessitating careful evaluation of the morphology.

61

Yet, even with this simple objective in mind the device properties are not always
straightforward to interpret. Experimental results revealed that increased P3HT
crystallinity, promoted through solvent annealing, increased device performance but also
increased photoluminescence (PL),'3 a counterintuitive phenomenon. If an exciton does
not dissociate it will recombine, generally radiatively, reducing overall device
performance due to less charge separation. Here we focus on characterizing the structure
to resolve this conundrum and provide detailed structural information leading to an

explanation of device performance.

 

 

 

 

 

 

 

 

 

 

 

1.5x106 "E
o

... 2
a .E,
a ‘3
2 5
9 o
C uo-o
— C
2

S

0

. . . _ -. _ 0F J l I
600 650 700 750 800 850 0.0 0.2 0.4 0.6
Wavelength (nm) Voltage (V)

Figure 4.1: Comparison of thermal annealing effects on photoluminescence and solar

cell performance. Thermal annealing a 1:1 by weight P3HT:PCBM solar cell at 140°C for

20 minutes increases the total amount of photoluminescence (a), indicating a much higher
degree of geminate recombination. However, this annealing also produces a much
improved device (b) with device efficiency increasing from 1.2% to 2.9%.

In Figure 4.1 we show PL measurements of two devices and the corresponding
current density (J) vs voltage (V) curves. All cell properties are shown in Table 4.1. Both

were spin coated at 2500 RPMs from chlorobenzene with one device tested as cast and

62

the other tested after thermal annealing for 20 minutes at 140°C. The as cast film

provides less photoluminescence indicating a high degree of exciton separation compared
to the annealed sample in terms of light emitting recombination. However, the annealed
sample is a substantially better solar cell with improved short circuit current, fill factor
and efficiency as clearly shown in the figure and table. The best device, seemingly,
should have the least amount of photoluminescence, but our experimental evidence, in

agreement with previous studies, shows this is not the case.

 

As Cast Annealed

 

Efficiency [%] 1.2 2.9
JSC [mA/cmz] 4.2 9.2
Voc [V] 0.62 0.57
FF [%] 45 54
Relative PL 1 3

 

Table 4.1: As cast and annealed solar cell properties and their relative
photoluminescence obtained by integrating both curves between the wavelengths studied.

To investigate why superior performing cells have increased photoluminescence a
detailed understanding of the P3HT:PCBM bulk heterojunction overall morphology is
investigated. Previous researchers have studied the vertical concentration profile of
polymer based solar cells using Phase Sensitive Neutron Reflectometry (PSNR),80 Near
Edge X-ray Absorption Fine Structure Spectroscopy (NEXAFS),56 ellipsometryfland X-
ray Photoelectron Spectroscopy (XPS).7| In our previous work, using PSNR,80 we
showed the vertical phase segregation of PCBM in a P3HT matrix was opposite to that
desired according to energy band level alignment. PCBM concentrated at the substrate

interface and near but not at the air interface where the aluminum counter-electrode was

63

deposited. A large amount of PCBM agglomerating at the conductive oxide is clearly
undesirable based on an energy level alignment argument and suggests why ‘inverted’

devices have shown good performance.” 83

Upon annealing, however, the PCBM
accumulates at the air interface creating an improved a pathway for electrons to leave the
device through the deposited metal electrode. This PCBM migration upon thermal
annealing is a clear indicator that the PCBM is in a kinetically trapped state upon spin
coating and has some mobility within the device. In addition, previous work63
considering annealing studies on quartz show PCBM agglomerating to large, micron
sized domains; another indication of PCBM’s mobility within the thin films, which is an
important aspect to our present study.

Neutron reflectivity is used to measure out of plane density variations within
films, and hence overall vertical composition, but cannot be used to elucidate any
information of overall particle agglomerate size scale or dispersion within a film.
Clustering of PCBM in these devices was characterized with small angle neutron

scattering (SANS) on thin films of P3HT and PCBM in a 1:1 by weight ratio. Neutron

scattering is uniquely suited to this system since there is a large, natural neutron

scattering length density (SLD) difference between P3HT (SLDP3HT = 0.74 and6 M)

and PCBM (SLDPCBM = 3.7 x10.6 A2). Both SLD values were determined

experimentally in our previous study80 and correspond well to theoretical calculations.
SANS has been used for many years as an extremely powerful tool for

investigating nanoscale structures within polymers. Size regimes of particles,84 radii of

gyrations85 and even less defined structures such as fractals86 can be characterized in great

detail. Investigating thin films of polymers is a less used, yet, still extremely powerful

64

application of SANS. Jones et al. demonstrated that thin films could be effectively
investigated with SANS by stacking multiple films together in their studies of interfacial
confinement on the radius of gyration of deuterated polymers.” 88

We performed SANS on solar cell mimics of two samples, both spin coated at

2500 RPMS on PEDOT:PSS, to characterize the phase separated morphology. One

sample was tested as cast and the other after annealing for 20 minutes at 140°C. Since

the films were so thin, 15 or more were stacked together to provide adequate scattering
intensity. Both neutron and x-ray scattering were used to measure individual film
thicknesses, which allowed calibrating each run to an absolute intensity scale.

The PEDOT:PSS layer was used for multiple reasons. Much work has shown that

changing the surface energy of a substrate can change the distribution of PCBM and, in

the case of thermal annealing, cause PCBM to cluster together on an SiOz substrate.

Using a PEDOT:PSS bottom layer assured no clustering of PCBM upon annealing.63
Additionally, to gain a high enough count rate and reduce the signal to noise ratio, a
number of films needed to be stacked on top of each other. This could be accomplished
by stacking the films spin coated on individual PEDOT:PSS coated wafers, which would
have totaled 3.75 mm in thickness, and is clearly not desirable. Additionally, the
PEDOT:PSS layer would have added roughly 30% more hydrogen to the system, which
would have contributed significantly to the incoherent scattering background. Instead,
the films were floated from the wafer in a water bath and stacked on top of one another to
form a multilayered sample. As PEDOT:PSS is water soluble, the floating procedure was

relatively simple. Previous neutron reflectivity experiments by us using this technique,

65

which is a sensitive test to processing induced defects, revealed excellent results,
indicating no adverse effects from this procedure.89

A plot of intensity (I(q)) versus wave vector (q) for the as cast and annealed pure
P3HT samples is shown in Figure 2. The upturn seen in all the scattering profiles at low
q is attributed to the surface roughness of the individual films88 and only data outside this
regime are considered. Clearly, both solar cell mimics show scattering plateaus as q

tends to zero (1(0), before the upturn) indicating an appreciable increase in scatterer size

evident from90

1(0) = ¢ V mi (4.1)

where ¢ is the particle volume fraction, V, the scatterer (particle agglomerate) volume and
Ap the difference in the scatterer and surrounding matrix scattering length densities
(SLD). Immediately one can discern the scatterer size increases by a factor of ~5 by
taking the ratio of the two plateau values.

A more thorough analysis can be made by modeling the PCBM agglomerates as
polydisperse Schultz spheres having hard sphere interactions that are surrounded by a
matrix of P3HT and solubilized PCBM. While it is unlikely the agglomerates are truly
spherical, this assumption provides a reasonable approximation to the system which we
justify a posteriori through consistency of the interpretation. The model contained a total
of four unknown fitting parameters: scatterer volume fraction, scatterer mean radius,
scatterer polydispersity and matrix SLD. To further the interpretation we made another
(standard) assumption that the SLD of the two components in the matrix could be scaled

with their volume fraction. In other words, an increase in matrix SLD would directly

66

indicate a higher concentration of dispersed, high SLD particles (PCBM). Comparison of

the fit to the data is good, as shown in Figure 4.2, with the fit parameters given in Table

 

 

 

4.2.
DI I I I I I I I I I I I I Th1
1000
* D D I ; 2: r. n‘
”g“
FA 100 + 8““
E +9 !
3 o ‘
5‘ +‘ 0.. ‘
.(T) 1 0 '6‘ - , -- .r. .
"J0"..- ("'03 .
c ...... , \ ‘ ‘
E + .4; ‘6‘“
1 _ Mixture Annealed *d' g 03:3‘
50 Mixture As Cast «5 ... ... .3}:
:+ Pure P3HT Annealed‘ o;°.f. 4», -=.- I“
L0 Pure P3HT As Cast . .
01 lBe§tFllt§llll 4 ? r 1 111:],
' 2 3 4 5 6 2 3 4 5 6
0 01 0.1

e (A")

Figure 4.2: Plot of intensity (I) versus wave vector (q) for both the annealed and
unannealed samples of 1:1 by weight ratio of PCBM and P3HT as well as annealed and
unannealed samples of pure P3HT. Annealing pure P3HT has no noticeable effect on the
scattering but annealing the mixture shows a clear increase in scattering, indicating a
substantial change in PCBM agglomerate size. The solid lines labeled “Best Fits” are
from a polydisperse Schulz model having hard sphere interactions as described in the
text.

Annealing has the effect of increasing the PCBM agglomerate size and volume
fraction suggesting a coarsening of the phase separated morphology. Further, it appears
that PCBM is soluble to ~ 16 vol% concentration in the P3HT matrix, which agrees well

with a recent study91 where the authors show a solubility of ~ 20 vol% for the same

67

system using a different technique. Taking the sum of PCBM agglomerate volume
fraction with that in the matrix we arrive at a total PCBM concentration of ~ 50 vol%
which is close to the expected value of 47 vol%. This lends some assurance that our
fitting technique is valid since thePCBM agglomerate volume fraction and matrix SLD

were both allowed to be unrestricted parameters.

 

As Cast Annealed

 

(PIESM agglomerate radius 3.3:: 5.9 i 2.4
PCBM agglomerate vol% 18% 35%
PCBM vol % in matrix 33% 16%
PCBM vol% in total 52% 51%

Distance between
agglomerates (nm)

P3HT crystallite size (nm) 9.2 16.3

1511.2 4,912.0

 

Table 4.2: Comparison of PCBM agglomerate size and polydispersity, spacing and
P3HT crystallite size. Fitting the neutron scattering data resulted in the concentration of
the PCBM agglomerates and the amount of PCBM in the surrounding polymer-rich phase
(matrix). These concentrations were added together to get the total PCBM concentration
which, based on the solution concentration, should be 47%.

Perhaps the most important characteristic of the morphology, at least in terms of

the PL measurements, is the distance between the agglomerates, H, estimated by H/D z

[1/ [HI/3 — 1, where D is the agglomerate mean diameter (we write [l/l:l]l/3 rather than
[Um/EHIB, where Um is the maximum packing fraction, since the agglomerates are

polydisperse in size and so would pack very efficiently making Elm much larger). The

68

values are listed in Table 2 and we find the distance increases to ~ 5 nm from ~ 1.5 nm

72 73 - . -
’ the increase in PL 18

after annealing. Since the exciton diffusion length is ~ 2 to 8 nm
completely consistent with the increase in distance between interfaces. However, the

efficiency also increases and we suspect the PCBM agglomerate connectivity becomes

more robust allowing greater current flow which we show in Figure 4.3.

As cast _ Annealed

      

 

 

Figure 4.3: Cartoon of morphology change after thermal annealing. The PCBM particles
are well dispersed after spincoating and a poor charge transport network results due to
lose interconnectivity of the PCBM aggregates. The PCBM and P3HT form better
transport networks after annealing through a coarsening of the structure with a
subsequent increase in spacing between PCBM aggregates. These two effects yield an
increase in efficiency through better conductive pathways but simultaneously increase
photoluminescence due to greater interparticle gaps.

In making this cartoon representation of the morphology for the figure we have
incorporated our previous neutron reflectivity results80 demonstrating enhanced PCBM
concentrations at the two interfaces and an approximately constant concentration in the
center. Our suggestion, based on this study, was that an increase in PCBM concentration
near the aluminum counter-electrode after annealing enhanced the efficiency. This is

certainly true, however, the present neutron scattering results suggest that structure

69

coarsening is a dominant factor providing better conductive pathways to the electrodes
while simultaneously explaining the increased PL.

Device efficiency is also influenced by P3HT crystallinity and we investigated the
crystallite size using grazing incidence x-ray diffraction (GIXRD) on the films. Others
have previously performed GIXRD on P3HT:PCBM films to show that after annealing,

76’ 77 an increase of

both through exposure with solvent vapor'3 or by thermal means,
coherence length of the P3HT crystalline domain occurs. A GIXRD plot of both an as
cast and annealed device, shown in Figure 4.4, demonstrate a similar increase in
coherence length after thermal armealing through the narrowing of the characteristic

(100) P3HT crystalline peak at q = 0.3 85A'l (5.45°) (Intensity in arbitrary units). We use

Scherrer’s relation to interpret our results

D _ 0.91
hkl — BMW“ (9)

 

(4.2)

where thl is the crystalline coherence length (the (100) crystal in this case), A is the

incident wavelength (1.54 A for the Cu ka source used in this work), 9th is the Full
Width Half Max (FWHM) of the W crystalline peak in radians and 0 is the peak
position. From this simple analysis we find the as cast film has a length scale (thI) of

roughly 9 nm, which increases to approximately 16 nm upon annealing. An increase in
P3HT crystallite size is desired since higher crystallinity P3HT provides better
conduction properties through perfection of the conductive network, however, an

increased crystallite size simultaneously increases the distance an exciton must travel to

70

reach a dissociation point, a competing effect. Indeed this must also account for the
increased PL in the annealed samples, acting in conjunction with the increased separation
distance between PCBM agglomerates and resulting in an obvious enhancement in

geminate recombination.

 

500 l l l I r l
O Annealed
— As Cast
400
13‘
E
3
9 300
.42
(I)
C
.9 200
E
100

 

 

 

 

0.4 0.6 0.8_ 1.0 1.2 1.4
(NA)

Figure 4.4: Grazing-Incidence X-ray Diffraction (GIXRD) plot of intensity versus wave
vector (q) for an as cast and annealed film of a 1:1 by weight mixture of P3HT and
PCBM. The peak narrowing and decreased FWHM of the annealed sample indicates an
increase in P3HT crystallite coherence length from 9 to 16 nm.

We would like to note that while a decrease in FWHM of XRD data is often
reported for these systems as an increased crystallite size, this is not necessarily true. An
increase in diffraction peak intensity with a corresponding decrease in FWHM indicates
an increase in crystallite size, a reorientation of the crystallites or, as we suspect here, a

combination of both. Previous reports on similar systems show both an increase in

71

crystallite size coupled with a slight reorientation of the polymer,”94

and we suspect the
same phenomenon is occurring here.

The coupling of PCBM agglomerates’ size and spacing and P3HT crystallite size
demonstrates the correlated interplay between the two active layer components. It infers
a theoretical maximum efficiency of a polymer/nanoparticle device at some intermediate
between complete dispersion and complete phase segregation. Monte Carlo simulations
of this exact effect were performed by Watkins et al. by varying the size regimes and

5 They showed that extremely well

dispersion of an electron acceptor/donor system.9
dispersed systems and highly phase separated systems produced poor devices due to the
limits on charge carrier extraction and exciton dissociation, respectively. An
intermediate system, however, produced the best devices from a good balance between
charge carrier extraction and exciton dissociation.

Indeed we may have an intermediately mixed system at hand afier annealing and
the structure appears to approach the length scales required for good device performance
albeit the PCBM agglomerates could be placed closer together to minimize PL.
Decreasing PCBM separation, though, seems to come at the cost of poor charge carrier
networks or decreased hole mobility. However, one should note that there is ~ 15 vol%
PCBM dissolved in the P3HT rich phase between the agglomerates. While it is not clear
where they are located, although they are probably within amorphous regions
surrounding the crystallites, they could act as traps and be a factor in recombination. In
addition, in our previous study,80 we found there was 75 vol% PCBM trapped at the

transparent conductive oxide side of the device which did not change composition upon

annealing, which is clearly undesirable. So, whether a device fabricated with these two

72

or similar components could be made substantially more efficient is not clear and
suggests the interplay between kinetically trapped structures and the path to
thermodynamic equilibrium is key to manufacture of the appropriate device dictating

careful morphological measurement be performed.

4.2 Experimental Section

4.2.1 Neutron Scattering

To mimic actual solar cell devices the SANS samples were prepared by spin coating

roughly 40 nm films of PEDOT:PSS onto silicon wafers, which were then baked at

130°C in air for 30 minutes. The coated wafers were transferred into a nitrogen filled

glove box and a 1:1 by weight solution of P3HT and PCBM in chlorobenzene was spin
coated onto the PEDOT:PSS coated wafers at 2500 RPMS for 60 seconds. The P3HT
(Rieke Metals) and PCBM (Nano-C) were dissolved at a total concentration of 30 mg ml‘

1 (15 mg P3HT and 15 mg PCBM per 1 ml of chlorobenzene). Annealing was performed

at 140°C for 20 minutes inside the glove box. The wafers were then removed from the

glove box and each film was floated off in water and then picked up by a double side
polished Si (111) wafer (25.4 mm diameter). Each film was dried in air and then a
successive film was floated into water and picked up in the same fashion. This process
was repeated 17 times for the annealed samples and 15 times for the unannealed samples
using a total of 5 wafers in each case. So, for the annealed case, 3 double sided polished
Si (111) had 5 films stacked on top of them and 2 had 6 films stacked on top. For the

unannealed case all 5 wafers had 5 films stacked on top. Pure samples were made in an

73

identical manner except the pure films were spin coated from a 30 mg/ml solution of pure
P3HT. This large number of films was needed in order to both obtain a high enough
signal to noise ratio as well as reduce counting time to a reasonable value.

All neutron scattering was performed at the National Institute for Standards and
Technology Center for Neutron Scattering on the NG3 SANS instrument. Three detector
distances were used; 2, 6.8 and 13.2 meters, all using a wavelength of 8 A. Although the
intensity of a neutron beam decreases as k4 for a neutron source, the Si (111) wafers
required the higher wavelength. Si (111) wafers can give rise to double diffraction peaks
when stacking them together,87 and hence the lower intensity 8 A neutrons were used.
Data reduction and evaluation was performed using the NCNR SANS packages SANS
Reduction v5 and SANS Analysis v3.96 To maximize total counts, all samples were

placed under vacuum during testing.

4.2.3 X-ray Scattering

Samples used for x-ray scattering were prepared using the same 1:1 by weight blend of

P3HT and PCBM (30 mg ml'l) in chlorobenzene (1 ml) as above. Glass slides (25.4 x

25.4 mm) were first coated with a roughly 25 nm layer of PEDOT:PSS, which were

baked at 130°C in air for 30 minutes and then transferred to a nitrogen glove box. The
P3HT:PCBM solution was spin coated onto the PEDOT:PSS coated substrates and the
annealed samples were heated at 140°C for 20 minutes.

All X-ray measurements were performed on a Rigaku Ultima IV multipurpose

XRD system housed at the University of Delaware. GIXRD was performed using a

74

standard XRD sample holder at a fixed incident angle of 0.6 degrees in 20 mode
sweeping the detector from 2 to 8 degrees. The 1.54 A wavelength copper kor source was

set at an accelerating voltage of 40V and a current of 44 mA for all experiments.

4.2.4 Photoluminescence Measurements

Photoluminescence measurement were performed on a Horiba Jobin Yvon F luoromax-4
spectrofuorometer with an incident wavelength of 550 nm using identical samples to the

X-ray measurements.

4.2.5 Solar Cells

Solar Cells were prepared in the same way as the X-ray and PL measurement samples
except ITO coated glass slides with a sheet resistance of 8-12 Q cm‘2 (Delta

Technologies) were used instead of pure glass slides. For more accurate testing and
creation of multiple devices on one substrate, the ITO was patterned using an HCl etch
before device fabrication. After spin coating of the active layer and annealing, the coated
substrates were transferred to a thermal evaporator and a 1 nm layer of lithium floride

(LiF) and subsequently an aluminum back electrode were both thermally deposited under

a vacuum of 5x10'7 torr at a rate of 0.2 nm sec]. The devices were tested using a 150 W

Oriel class A solar simulator under AM 1.5 conditions at 1 sun illumination as
determined by a calibrated reference solar cell traceable to the National Renewable

Energy Laboratory.

75

CHAPTER 5: CREATION OF ROUGH SURFACES TO ENHANCE SOLAR

CELL PERFORMANCE

5.1 Introduction

Much of the focus of this dissertation has been on describing the internal morphology of
P3HT:PCBM solar cells and correlating the morphology with performance. A great deal
of control over nanoparticle location in amorphous polymers such as PS and PMMA has
been demonstrated and is discussed in Chapter 6, but controlling the internal structure of
a P3HT:PCBM solar cell is more difficult. With the crystallinity and degradation
temperatures of P3HT being so close together, making vast morphological changes by
melting the polymer is not possible as devices annealed near or above the melting
temperature of P3HT have been found to be exceptionally poor. Similarly, with such
high volume fractions of PCBM needed to make a device, solubility limitations of both
the polymer and nanoparticle dominate the structures formed upon film casting
preventing controlled segregation to an interface. Proper choice of processing conditions

'5’ 76' 77 and controlling polymer molecular

such as solvent choice,74 annealing methods
weight16 can all enhance device performance, but, due to the nature of solution casting
these devices, control of nanoparticle location with these methods is limited.

As discussed in Chapter 3, the vertical PCBM concentration profile formed
directly after spin coating is nearly opposite to that desired for an ideal device. Groups
have developed ‘inverted’ devices where the cathode and anode are flipped either by

using electrodes with different work functions82 or physically depositing the traditional

electrodes, Al and ITO, on opposite interfaces.83 These inverted devices are comparable

76

with traditional ones and eliminate some of the charge transport problems associated with
the dense PCBM layer present at the substrate but they add complicated processing steps
or require more obscure materials making fabrication more difficult and potentially more
costly.

A simple yet robust method to enhance the overall morphology of these solar cells
is desired but neutron scattering results show we have only limited control over the

internal structure afier a film has been cast. With this in mind we looked to previous

work in our lab where we created nanorough substrates by adding 120 nm silica (SiOz)

particles into thin films of polystyrene.97 The particles were larger than the thickness of
the pure polymer film and produced large protrusions, increasing overall roughness and
apparent surface area of the substrate. A rougher substrate in a polymer solar cell will
increase the contact area between the PEDOT:PSS and active layer, possibly allowing for
increased exciton separation and potentially leading to the ultimate goal of producing a
co-continuous comb-like structure of donor and acceptor materials.ll Lithographic
methods of producing comb-like structures have been developed98 but these methods are
time consuming and cannot yet achieve the roughly 10 nm diffusion length scale of a
polymer solar cell exciton.

In addition to the enhanced interfacial area, the particles could act as scattering
centers to the incident light, increasing the path length of light within the device and
enhancing overall absorption. Light trapping methods have been used in inorganic solar
cells for many years, but these methods generally involve using surface modified glass to
change the direction of incident light, where the surface modifications are 10’s of

microns in size, much greater than the wavelength of incident light. Recently, however,

77

aluminum oxide (A1203) particles were used to enhance light trapping in a silicon

nanowire based solar cell.99 In this work the roughly 900 nm A1203 particles were

randomly distributed throughout a forest of silicon nanowires and increased the overall
light absorption from 80.8% to 84.4%. This result is noteworthy because the particles
were of the same order in size as the wavelength of the absorbed light, indicating
absorption enhancement is possible without using structures larger than the wavelength

of incident light.

5.2 Rough Surface Creation
In hopes of creating a surface that will aid in increasing both the overall surface

area of the active layer and absorption through light scattering, we created ‘rough’

substrates by adding 120 nm SiOz particles into the PEDOT:PSS layer of a traditional

solar cell. The SiOz particles are coated with a citrate layer, making them soluble in

water and easily dispersible in a solution of PEDOT:PSS. Figure 5.1 is a compilation of

AFM images of PEDOT:PSS layers on silicon substrates containing increasing amounts
of SI02 particles dispersed within the PEDOT:PSS. A layer of pure PEDOT:PSS (12

mg/ml PEDOT:PSS/water solution spin coated at 3000 RPMS) is relatively smooth but

both the roughness and surface coverage of SiOz particles increase as the concentration

within the 12 mg/ml PEDOT:PSS solution is increased from 10 mg/ml to 40 mg/ml.

78

 

208.00nm

252.00nm

Figure 5.1: Top down and 3D AFM images of (a,c) pure PEDOT:PSS and PEDOT:PSS
films containing (b,d) 10 mg/ml, (e,g) 20 mg/ml and (f,h) 40 mg/ml Si02 particles in the

initial solution. Increasing the solution concentration of Si02 increases the overall
surface coverage and roughness of the PEDOT:PSS coated films.

79

 

140.00nm

 

Figure 5.2: Top down and 3D AF M images of a 1:1 by weight film of P3HT and PCBM

on top of PEDOT:PSS with varying concentrations of SiOz in the initial solution. (a,c)
pure PEDOT:PSS and PEDOT:PSS films containing (b,d) 10 mg/ml, (c,e) 20 mg/ml and

(d,f) 40 mg/ml SiOz particles in the initial solution. Increasing the solution concentration

of Si02 increases the overall surface coverage and roughness of the PEDOT:PSS coated
films.

.80

Similarly, as seen in Figure 5.2, spin coating an active layer of 1:1 by weight

P3HT:PCBM on top of the rough surface and annealing it for 20 minutes at 140°C shows

a substantial roughness is still present on the top of the film.

5.3 Solar Cells on Rough Surfaces

Clearly, adding SiOz particles to the PEDOT:PSS layer adds a substantial amount

of roughness and surface area to the active layer. Each of these PEDOT:PSS layers were
tested as substrates within solar cells and, shown in Figure 5.3, a substantial increase in
device performance with the rough surfaces compared to the pure PEDOT:PSS film
results. Figure 5.3 is a plot of current density (J) versus voltage (V) for solar cells spin

coated on top of a PEDOT:PSS layer spin coated from solutions containing 0, 10, 20 and

40 mg/ml of 120 nm SiOz spheres, respectively. Cell properties are shown in Table 5.1.

Pure PEDOT- 10 mg/ml. 20 mg/mIA 40 mini]-

 

 

 

 

 

Jsc (mA/cmz) 5.0 6.1 7.4 6.7
Voc (V) 0.55 0.56 0.55 0.54
FF (%) 47 52 58 58

n (%) 1.3 1.8 2.3 2.0

 

 

 

 

 

 

Table 5.1: Solar cell properties of the devices shown in Figure 5.3. They symbols next
to each cell title correspond to the device J -V curves in Figure 5.3.

81

 

    
  
    

L ..... ...........................................
— Pure PEDOT:PSS
+ 10 mg/ml Si02

_2 _ + 20 mg/ml SiOz
+ 40 mg/ml Si02

O

 

Current Density (mA/cmz)

 

 

0.0 0.2 0.4 0.6
Voltage (V)

Figure 5.3: Current density (J) versus voltage (V) plot of solar cells having an increasing
amount of SiOz particles dispersed within the PEDOT:PSS layer. The pure PEDOT:PSS

device produces the worst cell but upon addition of Si02 in the PEDOT:PSS layer the
efficiency, short circuit current and fill factor all increase.

A solar cell on top of a pure PEDOT:PSS layer produces an unremarkable cell

with a relatively low efficiency, short circuit current and fill factor. Adding SiOz into the

PEDOT:PSS layer, however, noticeably increases all of these properties, with a
maximum ch, FF, and 11 found in the device with a PEDOT:PSS layer spin coated from

solution containing 20 mg/ml SiOz.

This increase in device performance could, at least somewhat, be related to the

increase in surface area of the film and hence increased contact area between the

82

PEDOT:PSS layer and active layer. However, the increase in surface area is a direct
result of adding more Si02 on the substrate which decreases the amount of substrate

exposed directly to PEDOT:PSS. While an increase in surface area is theoretically

desired for increased charge transfer, completely covering the susbtrate with SiOz

particles would prevent most of the holes from being extracted from the PEDOT:PSS by

the conducting oxide electrode.

The total surface area increase and SiOz surface coverage is shown in Table 5.2.

While the overall area does increase with increasing Si02 concentrations, the improved

surface area does not equal the improvement in device performance. Short circuit current
increases roughly 25-50% for the devices with $102 but the maximum surface area
increase of the PEDOT:PSS film is only 16%. Furthermore, the best cell produced here,
the 20 mg/ml device, does not see the largest surface area enhancement. One reason for
this could be that as Si02 concentration is increased the exposed ITO decreases. For the
20 mg/ml there is still 65% of the ITO exposed to the PEDOT:PSS film. For the 40
mg/ml case, however, a relatively small increase in overall surface area from the 20
mg/ml device (13% to 16% relative to a pure PEDOT:PSS film) is coupled with a
substantial decrease in exposed ITO (65% exposed to 35% exposed). This large drop in

available ITO for charge transfer could be the reason a maximum performing device was

found at a concentration of 20 mg/ml.

83

Pure 10 mg/ml 20 mg/ml 40 mg/ml

 

 

PEDOT:PSS SiOz SiOz SiOz
PEDOT:PSS
Surface Area 0% 7% 13% 16%
Increase
S102 Surface 0% 20% 35% 65%
Coverage

 

 

 

 

 

 

 

Table 5.2: Total PEDOT:PSS surface area increase and surface coverage change upon
addition of SiOz particles into a film of PEDOT:PSS.

Clearly the addition of SiOz improves device performance but the increase in
overall surface area does not account for the total improvement. To further explain the
mechanism of performance enhancement we investigated if the Si02 particles were

scattering incident light and increasing overall exciton production.

5.4 Spectroscopic Measurements
To investigate whether the particles do enhance light scattering, and hence
absorption, we performed absorption spectroscopy measurements on the PEDOT:PSS

layers. Figure 5.4 is a plot of absorbance versus wavelength for a pure PEDOT:PSS film,

all three SiOz concentrations studied and a pure glass substrate are shown for

comparison. The absorbance of pure glass was subtracted, via a reference sample, for all
data shown. As can be seen in the figure, pure PEDOT:PSS begins to absorb light

around 350 nm but does not absorb substantially until below 300 nm. Addition of the

SiOz particles, however, shows absorbance beginning at 600 nm, a substantially lower

energy wavelength than pure PEDOT:PSS will absorb. The 10 and 20 mg/ml samples
show nearly identical absorption but the 40 mg/ml sample has significantly enhanced
absorption compared to the others.

From the absorption plots of a glass substrate and that of pure PEDOT:PSS we

can reasonably assume that the films containing SiOz particles are not actually absorbing

more light, so the SiOz particles must be scattering the light such that it does not

propagate to the detector of the spectrometer. Shown pictorially in Figure 5.5, some of

the incident light on the PEDOT:PSS film (Figure 5.5a) that is above the band gap will be

absorbed by the film while the rest passes through. For the samples containing SiOz

(Figure 5.5b) the same amount of light is absorbed by the film, but some of the light is
scattered as it contacts the particles and never reaches the detector. Hence, an apparent
increase in absorption is seen when, in actuality, the rough films merely scatter the light,
preventing it from reaching the detector. Increasing the number of particles on the
surface increases the number of scatterers and the total light scattering, which is apparent

in Figure 5.4.

85

 

 

 

 

 

 
 
    
 

 

 

 

0.4 I A. . .
i e\: l T I
. 00.30 _ j
i E 0.20 ~ H
S 0-3 ‘t €0.10:- aj "
1 0.00 ' I ‘
f; : ‘0 300 400 500 600
2 0 2 _i Wavelength (nm)
5 i — Pure PEDOT:PSS
8 '. - - - Glass
2 ', + 10 mglml SiO2
0.1 ' + 20 mglml SiO2 -
—l- 40 mglml SiO2
0.0 1

 

300 400 500 600 700

Wavelength (nm)

Figure 5.4: Absorbance spectrum of PEDOT:PSS layers on glass with increasing
amounts of Si02 dispersed within. A larger amount of Si02 dispersed within the
PEDOT:PSS layer shows a higher light absorbance, which indicates an increase in
scattered light from the SiOz particles. The inset shows scattering efficiency versus
wavelength for Mie scattering theoretical calculations across 120 nm spheres. The
scattering increase according to Mie theory directly correlates with the increase in
observed absorbance in the rough PEDOT:PSS layers suggesting Mie scattering from the
120 nm SiOgspheres is the cause for the increase absorbance.

86

Pure PEDOT:PSS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

333‘s
:3 ’ g
E 9 8
TE.’ ’ '°
0
.5

Rough PEDOT:/y
a .
.2” :5
2". a) 8
5 1'3
..o '0
'8
.5 \

 

 

Figure 5.5. Diagram of proposed scattering from rough surfaces with SiOz inclusions.
For a pure PEDOT:PSS film (a) a percentage of the incident light is absorbed by the layer

and the rest passes through into the detector. For a rough PEDOT:PSS film with SiOz

inclusions (b) the same proportion of the incident light is absorbed by the film but a
proportion of the non-absorbed light is scattered away from the detector.

To confirm the hypothesis that the particles are scattering light from the detector
we looked to Mie scattering theory. Mie scattering is an extension of Rayleigh scattering
that is used describe the scattering effects of particles on incident light. Rayleigh theory
describes the scattering of light when a particle is much smaller than the incident
wavelength but breaks down when the particle size reaches about 10% of the wavelength.
Mie theory describes scattering that occurs from particles that are of the same size order
of incident light. The overall scattering efficiency, i.e. the fraction of incident light on a
given area that will be scattered by the particles, is shown as the inset in Figure 5.4. The

scattering efficiency increases nearly identically with the increase in observed absorbance

87

for the rough PEDOT:PSS layers indicating that Mie scattering is the reason for the
increased observed absorbance. An experimental confirmation of this theory requires an
integrating sphere in conjunction with a UV-Vis absorption spectrometer, but no
integrating sphere was available.

The intensity of light as a function of distance within a material can be determined

by the equation

I : 108—a2 (5-1)

where I is the light intensity at a distance 2 into the material, 10 is the incident light

intensity and or is the absorption coefficient. The solar cell films studied here are about
90 nm thick and roughly 50% by volume P3HT. Using a value of 1.75 x105 cm.1 for (1'00

and assuming the absorption properties do not change substantially for P3HT when it is
in a mixture with PCBM, only about 60% of the incident light within the absorption
spectrum of P3HT will be absorbed. Even if the entire 90 nm film was P3HT only 80%
of the incident light would be absorbed, so regardless of the assumptions made for the
absorption properties of P3HT in a mixture, all of the light is not being harvested by the

active layer.

Despite the excess scattering provided by the SiOz particles and the less than

100% absorption by the active layer, absorption characteristics of P3HT:PCBM films

coated on the rough PEDOT:PSS layers show minimal absorption increases compared to

a pure PEDOT:PSS device, see Figure 5.6. The devices with SiOz showed a 25-50%

88

increase in short circuit current but the increase in absorption was only 3-8% over the

entire absorption spectrum, not nearly enough to account for the J sc increase.

 

    

Intensity (A.U.)

— Pure PEDOT:PSS Ann
+ 10 mglml SiO2 Ann

.. — --I— 20 mglml SiO2 Ann
L + 4q mglml SIOZ Ann L

300 400 500 600 700

 

 

Wavelength (nm)

Figure 5.6: Absorption spectrum of 1:1 by weight P3HT:PCBM films on top of
PEDOT:PSS fihns with various concentrations of SiOz particles within the film. An

increase in the amount of SiOz particles within the PEDOT:PSS film increases the overall
absorption, but the increase is small compared to the increases seen in the device
properties.

Although the absorbance spectrum of a P3HT:PCBM film changes very little
when spin coated on these rough surfaces, the PL increases dramatically, Figure 5.7: The
absorption for 10, 20 and 40 mg/ml samples increases 3%, 5% and 8%, respectively, but
the PL increases by 40%, 67% and 80% for same samples. Based on the SANS results

discussed in the Chapter 4 we know that an increase in PL can indicate excitons are being

89

produced farther from PCBM molecules and we must conclude this same phenomenon is

occurring here.

 

I l l
— Pure PEDOT:PSS
+ 10 mglml SiO2 n
-l— 20 mglml SiO2
+ 40 mglml SiO2

O

N

O

X

.3

O

I
x

 

 

 

 

’7‘ it
2 1.5 _ ' \, —
a. l c
3- / x
.5 \
C
o 1-0 ‘ a A
a f’

0.5 — g n

\
\ ... 4‘
l l l
600 650 700 750 800
wavelength (nm)

Figure 5.7: Photoluminescence data of 1:1 by weight P3HT:PCBM films on top of
PEDOT:PSS films with various concentrations of SiOz particles dispersed within. An

increase in the amount of Si02 particles within the PEDOT:PSS film total

photoluminescence to a much greater degree than the observed solar cell properties or
absorbance.

5.5 Discussion Section
Clearly the addition of a rough surface by inclusion of Si02 particles into the

PEDOT:PSS layer creates a complex balance of phenomena contributing to solar cell
performance. The rough surface creates higher interfacial area between the PEDOT:PSS

and the active layer as well as the metal electrode and the active layer, which may

90

increase overall charge separation at the interfaces. Furthermore, we know the formation

of the active layer morphology is dominated by the limited solubility of PCBM. Creating
a rougher surface with the Si02 particles could create regions within the film that trap

solvent from evaporating which would effectively trap more PCBM particles within these
regions of higher solvent concentrations. This PCBM trapping could expose more of the
PEDOT:PSS layer to P3HT, allowing for improved hole conduction. Trapping more
PCBM could also explain the PL increase with greater surface roughness. As a larger
amount of PCBM is pulled out of the film the distance between PCBM molecules will
increase, lowering the chance an exciton will find a PCBM and dissociate.

Another possible explanation as to why the rough surface could provide a better
device is from an increase in vertical alignment of P3HT crystallites within the device.
Previous work has shown that P3HT prefers to align horizontally with the substrate upon
annealing?"94 but optimal transport of charge carriers out of a device requires the
polymer to be aligned vertically. By adding large protrusions in the solar cell, the P3HT
can align vertically along the protrusions instead of aligning horizontally along the
substrate.

Unfortunately, characterizing the PCBM morphology and P3HT alignment in a
system such as this is rather difficult. Accurately determining PCBM cluster size
requires the use of neutrons, but performing small angle neutron scattering on a 4
component system without being able to contrast match would be nearly impossible.

Performing XRD measurements on these films to determine the orientation and
alignment of the P3HT is possible, but the SiOz particles would make traditional GD(RD,

which is usually used in thin film work, not straightforward. Properly aligning and

91

rotating the samples could become a time consuming and difficult task. Regardless of the
difficulties in further characterizing the morphology of these devices, it is clear that

roughening the surface with a light scattering particle does enhance device performance.
In addition, there is an apparent optimal concentration of 20 mg/ml SiOz particles in the

PEDOT:PSS solution.

5.6 Experimental Section

PEDOT:PSS films with SiOz particles dispersed within were prepared as follows. 120
nm SiOz particles (Bangs labs) were added to a stock solution of PEDOT:PSS (H.C.

Stark) such that the final solution would contain 10, 20 or 40 mg of SiOz particles per

total ml of solution. The solution were then diluted with ultrapure water from the stock
concentration of 15 mg/ml down to 12 mg PEDOT:PSS/ml water. The resulting
solutions were sonicated for 30 minutes to break up any aggregates, filtered through a
200 nm PTFE filter and then spin coated at 3000 RPMS for 60 seconds. The pure

PEDOT:PSS films were 23 nm in thickness as determined by neutron and x-ray

reflectometry. The films were then dried in an oven for 30 minutes at 130°C and

characterized or transferred to a glove box to be coated with a 1:1 by weight solution of
P3HT:PCBM. Solar cell fabrication, UV-Vis and PL measurements were all performed

as described in Chapters 3 and 4. No LiF layer was used for the solar cells reported here.

92

CHAPTER 6: NANOPARTICLE ASSEMBLY IN THIN POLYMER FILMS

6.1 Introduction

Most of this dissertation has focused on describing and characterizing polymer based
solar cells, but a great deal of initial work focused on characterizing inorganic
nanoparticle assembly in thin films. One of the initial goals of this project was to
developed methods to control nanoparticle location within polymer based solar cells, but
we quickly discovered a more thorough characterization of the internal nanoparticle
dispersion was needed before any type of control could be applied. This chapter
discusses the work done on assembling inorganic nanoparticles in amorphous polymers
and extends somewhat into P3HT:PCBM systems.

Polymer coatings are ubiquitous in society. Used in paints, adhesives,
antireflective coatings, electronic applications and a variety of other forms, polymer
coatings have been around since the development of synthetic polymers in the 19305. As
the development of nanotechnology has grown so has research into thin polymer films,

that is, films around 10 to 100 nm in thickness. Films with such small thicknesses can

. . . . . 101
have umque properties for research or commercral applications such as, sensors,

. . 102 . . . . 103 .
photovoltalcs,3 transrstors, light emitting diodes and a host of others. While many

of the benefits of bulk polymer coatings are retained in thin film applications, the thin

films have the added benefits of being lighter, cheaper and often times, being optical

. 104
active.

As a demonstration of the uniqueness of thin films on a basic level, films with

thicknesses smaller than the wavelength of light appear to be different colors depending

93

on their thickness. For example, a 1 nm thick layer of silicon dioxide (SiOz) grown on a

silicon (Si) wafer is essentially transparent but a 100 nm thick layer of SiOz appears

reddish-purple. Similarly, polystyrene films show a color change as they increase in
thickness going from transparent to gold to light blue. An equivalent demonstration of
doubling or tripling the thickness of a coating, say red paint, does nothing to alter its

color because its thickness is much greater than the wavelength of light.

6.2 Background

6.2.2 Dewetting Inhibition

One of the most important considerations of thin films, specifically polymers, is
their ability to wet a substrate. To not dewet a substrate upon deposition or after
annealing, a polymer film must have a lower total energy than the substrate upon which it
is deposited. When using high energy surfaces such as metals dewetting is not a
problem, but lower energy substrates such as glass or polymers can cause dewetting
depending on the energies of the system. Additionally, surfaces with protrusions or
larger particles (ie: very rough surfaces) are much more likely to cause a deposited
polymer layer to dewet than a smooth surface.

Dewetting occurs for two reasons; a defect, such as a large dust particle, acts as a
nucleation site or thermal fluctuations in a film allow a nanosized hole to reach the
substrate begins the dewetting process. A film will only dewet from a surface if the
substrate supporting the film has a lower energy than the film itself. This is described by

Young’s equation

94

7 -7
Cos 6 = ——5" 5" (6.1)
71.11

where 0 is the angle between a drop of liquid and its supporting surface and 'ygv, 731, and

YLV are the interfacial surface tensions between the substrate and vapor, the substrate and
liquid and the liquid and vapor, respectively. In a controlled setting most of the above

variables can be known quite accurately and surface energies can be calculated.

A large body of work has been developed showing many types of nanoparticles”-

33’ 101 and dendrimers105 added to polymer films inhibit dewetting. Krishnan et al.

showed the addition of one monolayer of cadmium selenide (CdSe) nanoparticles32 or

9

cross linked polystyrene nanoparticles3 33 inhibit dewetting of polystyrene on a
Silanized substrate after thermal annealing. Similarly, Holmes et al. showed dewetting

inhibition with C60101 after solvent annealing for many hours. The majority of these
works showed that dewetting was completely inhibited only when the equivalent of a full
monolayer or more of nanoparticles were added to a film.

Figure 6.1 is a compilation of time resolved optical micrographs of thermally

annealing a 75 nm thick film of 115 kDa polystyrene (PS) spin coated on a Silanized
silicon wafer. The large pictures are a pure PS film being annealed at 170°C, well above
the Tg of PS (Tg = 104°C) and each inset is an identical PS film with one monolayer

equivalent of 60 kDa intramolecularly crosslinked PS nanoparticles dispersed within.106

95

Details describing the silanization process and dewetting experiment are explained in the
Experimental Section. As can be seen in Figure 6.1, thermal annealing induces a
dewetting of the film from the Silanized surface, which compares well with reports in the

literature mentioned above. The Silanized surface has a surface energy of 28 mJ/m2 89

whereas the PS surface energy is about 36 to 42 mJ/m2,1°7’ ‘08 and hence the higher

energy film dewets upon annealing.

 

Figure 6.1: Time resolved optical micrographs (a-f) of thermal annealing a 75 nm thick
film of pure 115 kDa PS on a silicon wafer Silanized with Sigmacote ®. The large
pictures show the onset and continued dewetting of the PS fihn over the course of 50
minutes and the insets show a similar 75 nm thick film of PS with 1 monolayer
equivalent 60 kDa PS nanoparticles dispersed within, which completely inhibit
dewetting.

96

6.2.3 Nanoparticle Interfacial Assembly

To explain why nanoparticles inhibit dewetting, neutron reflectivity was used to
investigate the location of the nanoparticles within the thin films. Krishnan et al. showed
that upon spin coating a layer of deuterated polystyrene with cross linked polystyrene
nanoparticles the polymer and nanoparticles were equally distributed throughout the thin

31, 33
m.

fil Thermal annealing, however, caused all the PS nanoparticles to assemble at

the substrate of the film. As both components of this system are polystyrene there is
essentially no heat of mixing between the two components so that entropy is the only
driving force acting on the system. By assembling at the substrate the PS nanoparticles

lower the overall system entropy by maximizing the degrees of freedom of polymer

motion.32 Neutron reflectivity work by Yaklin et al.89 showed that C60 will assemble at
the substrate directly after spin coating.89 This assembly is attributed to the extremely

low solubility of C60 in toluene, the spin coating solvent for this work, and that the C60

crash out of the film as the solvent evaporates.

Despite the difference in assembly forces attributed to both systems, each system
inhibited dewetting equally due to the ‘carpet’ of nanoparticles assembled on the
substrate. Each work discussed above showed dewetting is severely inhibited once a
monolayer of nanoparticles covers the substrate surface. An estimate of the coverage of

nanoparticles at the substrate is

6 = (7a) (15 (6.2)

97

where 0 is the fractional surface coverage of the nanoparticles, h is the height of the film,
a is the particle radius and (1) is the bulk volume fraction of nanoparticles. As thermal

fluctuations occur within the film the polymer can never fully expose the low energy
substrate due to the presence of the nanoparticle layer. Since the nanoparticles form a
monolayer that completely covers the substrate they are in a jammed state preventing
separation. The nanoparticle assembly force is so large it will prevent dewetting of

polystyrene on a very low energy surface for days and even weeks of being held well

above its glass transition temperature (Tg).3 1

Further work was carried out by Krishnan et al. with the assembly of cadmium

selenide (CdSe) nanoparticles. As mentioned above, CdSe nanoparticles will prevent

dewetting and, unlike the carbon based nanoparticles of PS and C60, have a very large Z

contrast with PS. Transmission electron microscopy (TEM) images of thin films of PS

and CdSe nanoparticles showed the CdSe were actually assembling at the air interface of

the films instead of the substrate.32 This assembly can be described by considering the

Hamaker constant (A) for a trilayer system consisiting of three components; air (1), PS
(2) and CdSe particles (3). Using refractive index, the dominant term in the trilayer
Hamaker constant, the overall sign can be determined and hence the projected assembly.

Using the equation,

A132~I7li " 77%le175 — ’75] (6-3)

98

where 77,; is the refractive index of the individual components 1, 2 and 3. Using values of
1.0, 1.59 and 1.54 for air, polystyrene and the CdSe particles, respectively, a negative

value is found for an ordering of A] 32, which is air, CdSe, PS.

The CdSe particles used in this study were coated with oleic acid, a solublizing
agent roughly 2.5 nm long when fully extended. To calculate the particle’s total
refractive index a volumetric average was performed with the 2 nm CdSe (n = 2.8) core
and 2.2 nm shell (71 = 1.4) giving a total particle refractive index of n = 1.54. This
assembly, as was demonstrated by Krishnan et al., can be used to make multilayered
films by assembling the nanoparticles at the air interface of a polymer layer, crosslinking
the polymer and repeating the procedure.

The situation studied by Krishnan was unique in that the oleic acid coated CdSe
particles (o-CdSe), due mainly to the large volume of the particle taken up by the low
refractive index layer, had a lower refractive index then PS, which has a relatively high
refractive index for a polymer. According to the trilayer assembly theory, a particle will
assembly at the substrate provided its refractive index is greater than that of the polymer
film. Further investigation into this phenomenon was accomplished using three methods;
changing the substrate material, using a higher refractive index particle dispersed in a PS

film and a using lower refractive index polymer film with the same o-CdSe particles.

99

6.3 Results and Discussions

6.3.1 Effect of Changing Substrate

As an alternative substrate to a silicon a PET substrate was used. When
considering the assembly of nanoparticles in films using the Hamaker constant
approximation a trilayer system is used that takes into account the polymer, the

nanoparticle and the air but the substrate is ignored. All of the cases studied by Krishnan
used an SiOz, which has a refractive index of n = 1.46. This is lower than the two
components, PS and o-CdSe, but no o-CdSe particles are seen migrating to the substrate
to shield the PS from the Si02 indicating that neglecting the substrate is an acceptable

assumption.
Further investigation into the substrate was performed by using a PET substrate.

PET has a refractive index of 1.58, which is between that of PS and o-CdSe and a good
comparison to SiOz substrate. Figure 6.2 is a transmission electron microscope (TEM)

image of crosslinkable PS and o-CdSe nanoparticles spin coated onto PET, crosslinked
and repeated. As can be seen in the image, the expected assembly occurs with the

nanoparticles migrating to the air interface in each of the two layers.

100

 

Figure 6.2: TEM image of a four layer system of polystyrene and o-CdSe nanoparticles
on a PET substrate created in two spin coats. A mixture of 140 kDa crosslinkable PS

with nanoparticles was spin coated onto a PET substrate, annealed at 170°C for 12 hours,

cross linked at 220°C for 20 minutes and repeated.

6.3.2 Assembly of Higher Refractive Index Nanoparticles

To investigate the assembly of a higher refractive index particle a CdSe
nanoparticle with a pyridine coating was used. Because pyridine is a very small
molecule, roughly 0.5 nm in length, the majority of the particle is high refractive index
CdSe. A 2.0 nm CdSe core coated with pyridine (p-CdSe) has a refractive index of n =

2.2, creating a negative Hamaker constant for the assembly of air, PS, p-CdSe. TEM

101

investigations into this assembly show exactly what the theory predicts, an assembly of
p-CdSe particles at the substrate. Figure 6.3a is a TEM image of system consisting of
p-CdSe nanoparticles assembled at the substrate covered by a layer of crosslinkable PS
and a repeat coating on top. Some p-CdSe are trapped between the layers due mainly to

the speed at which the polystyrene was crosslinked.

 

Figure 6.3: Cross sectional TEM images of polystyrene fihns with pyridine coated
cadmium selenide nanoparticles dispersed within. (a) A four layer system of PS and p-
CdSe nanoparticles created by spin coating the PS/p-CdSe mixture, annealing the film
followed by cross-linking the polymer and then repeating the process. (b) Cross sectional
image of the PS/p-CdSe film directly after spin coating.

Because the pyridine molecule is so small and is less soluble than oleic acid in
toluene, the solvent used here, a larger amount of nanoparticle agglomeration can occur
either in solution or upon deposition. As the particles agglomerate into clusters their

effective diffusivity is lowered, requiring more time to migrate to the substrate upon

thermal annealing. Figure 6.3b is a cross sectional image of a film before cross-linking

102

the PS. Some agglomerates can be seen but the particles are fairly well dispersed

thoughout the thin film.

 

Figure 6.4: Cross sectional images of oleic acid coated CdSe nanoparticles assembled at
the substrate of thin PMMA films. Due to the difference in refractive index of PMMA
from PS the o-CdSe particles assemble at the substrate compared to the air interface as
shown in previous work”. The assembly is very robust and occurs in (a) thin films less
than 100 nm, and (b) fihns approaching 400 nm in thickness.

Investigation into the assembly of o-CdSe in a lower refractive index was
performed with poly (methylmethacrylate) (PMMA), which has a refractive index of
1.49. A negative Hamaker constant, indicating a driving force towards assembly, is
found when the particles order as air (n = l), PMMA (n = 1.49), o-CdSe (n = 1.54). TEM
images, shown in Figures 6.4a and 6.4b, show this assembly does indeed occur. A thin
layer of o-CdSe particles can be seen at the substrate of the PMMA film with virtually no

particles trapped within. The assembly is robust and complete in films less than 100 nm,

103

Figure 6.43, and approaching 400 nm, Figure 6.4b, indicating a relatively large length
scale over which this assembly occurs.

This technique can be further applied to assemble multiple particles in one film.
As shown in Figure 6.5, both p-CdSe and o-CdSe particles assemble at their respective
interfaces of a PS thin film; o-CdSe at the air and p-CdSe at the substrate. By merely
changing the stabilizing ligand attached to the nanoparticle core it is possible to assemble
these particles at either interface of a film. For a functional nanoparticle, such as the
quantum dots used here, interfacial assembly can be extremely beneficial in aligning
energy levels of different electrically active materials for use in photovoltaics, light

emitting diodes or transistors.

 

Figure 6.5: A thin fihn of polystyrene with both o-CdSe and p-CdSe nanoparticles
dispersed within. After annealing above the glass transition temperature of PS the o-
CdSe particles migrate to the air interface while the p-CdSe particles migrate to the
substrate. To the left and right of the TEM image are representations of the CdSe
particles with their respective stabilizing ligands.

104

6.3.3 Nanoparticle assembly in P3HT

With the hopes of assembling functional nanoparticles at the interfaces of
electrically active polymer devices, specifically polymer based solar cells, both p-CdSe
and o—CdSe nanoparticles were dispersed in poly (3-hexylthiophene) (P3HT), a common
polymer in polymer based solar cells. Many difficulties were encountered in these
experiments dealing mainly with the crystallinity of P3HT. All previous experiments in

thin film assembly were performed with either PS or PMMA. Both are glassy polymers
that easily dissolve in organic solvents, both have a Tg well below their respective
decomposition temperature and both can be purchased with very low polydispersity

indices (PDI) at a variety of molecular weights (MW). P3HT is not as ideal. The best

P3HT commercially available at the time of this thesis had a PDI of about 2.5 with a

molecular weight average of about 35,000 kg/mol. While the Tg of P3HT, Figure 6.6, is
relatively low, about 72°C, it is highly crystalline with a melting point very near its

degradation point, about 213°C. As P3HT is highly crystalline any nanoparticle

assembly in a P3HT film would require melting the polymer. As can be seen in the
comparison of the TGA and DSC, Figure 6.6, the onset of P3HT degradation occurs just
as the crystallites begin to melt. With decomposition occurring right as the polymer
melts, assembling nanoparticles within a P3HT matrix for the purpose of device

applications, at least using thermal methods, will most likely produce inactive devices.

105

 

 

 

 

 

100.0 I l I I _ 00
l’. '
s‘ I
3*”: 99.9 — ‘s : q -01
H I
r: I
8 i g;
I— I a-o-
til.) 99 8 - 3
4—0 0
g, E
g 99.7 a ,
— 080 of P3HT (right axis) l
---- TGA of P3HT (left axis) :
996 l l l l

 

50 100 150 200
Temperature (C)

Figure 6.6: Plot of thermal gravimetric analysis (left axis) and differential scanning
calorimitry (right axis) of pure P3HT. The onset of melting and degredation are almost

identical at about 200°C (horizontal dashed line) indicating thermally processing P3 HT
for assembling nanoparticles may not be feasible.

Despite the degradation occurring at the melting point of P3HT, some work was
performed on these films to investigate if nanoparticle assembly would even occur within
thin films of the melted polymer. Obtaining high quality, or any, cross sectional TEM
images of P3HT fihns with or without nanoparticles proved extremely difficult. The
method for obtaining these images, described fully in the Experimental Section of this
chapter, rarely produced any images and many were of very poor or broken films. These

difficulties most likely resulted from the P3HT being highly crystalline and forming some

106

type of bond between itself and the substrate making removal of the film from the
substrate very difficult. A focused ion beam (FIB) would be ideal for cutting thin film
sections and obtaining cross sectional images of these films, but no FIB was available at
either Michigan State University or the University of Delaware. The main reason for
showing the images below is to describe preliminary work performed on the nanoparticle
assembly within P3HT and offer a basis for possible future work for other graduate
students.

Figure 6.7a is a cross sectional TEM image of P3HT with p-CdSe nanoparticles
directly after spin coating. All of the p-CdSe nanoparticles are seen to assemble at the air
interface, identified by the gold layer sputtered on top for viewing. P3HT has a higher
refractive index than most polymers so the assembly of the particles, which have a
refractive index of about 2.2, at the air interface is not entirely surprising. What is
surprising is that the particles assembled at the interface without any annealing. In this
experiment the P3HT film was spin coated from chlorobenzene at 2500 RPMs. The
slower speed and higher boiling point chlorobenzene caused the film to dry much more
slowly than films spun from toluene at higher RPMS, allowing the particles a chance to
assemble before the film was completely dry. This is similar to solvent annealing the
film in the manner described by Yaklin et al.89 except the fihn here was allowed to dry
slowly after spin coating as opposed to being placed in a solvent atmosphere. Attempts
to follow the solvent annealing method from Yaklin et a1. proved very difficult as the
completely dry P3HT films seemed to uptake no solvent even after days of exposure to

atmospheres of toluene, chlorobenzene and chloroform. Seemingly, the crystalline nature

107

of the P3HT prevents solvent uptake from a saturated atmosphere, even when the solvent

vapor fraction was increased through heating.

 

Figure 6.7: Cross sectional TEM images of films of (a) pure P3HT with dispersed p-
CdSe particles assembling on the top of the film and (b) a solar cell mimic film consisting
of a 1:1 by weight mixture of P3HT and PCBM with a small volume fraction of p-CdSe

particles dispersed within. Upon heating to 250°C in a nitrogen environment the CdSe

particles assemble into the middle of the film, presumably on top of a layer of PCBM that

has assembled at the substrate and extends upwards to roughly the middle of the film
based on the volume fiaction of PCBM added.

Figure 6.7b shows a cross section of a film of P3HT, p—CdSe nanoparticles and
the fullerene derivative [6,6]-phenyl-C51-butyric acid methyl ester (PCBM), essentially a

polymer based solar cell mimic. Due to difficulties in obtaining these pictures, an image

of the as cast film was unable to be obtained. The film was heated to 250°C, well over its

melting point, in a nitrogen environment. Quite surprisingly, the CdSe particles assemble

almost directly in the middle of the fihn. Before discussing possible reasons for the

108

assembly of CdSe in the middle of the film it must be noted that heating P3HT above its
melting point most likely causes some type of degradation. This degradation cannot be
fully accounted for from the study performed here, but, as can be seen from the image, a
stable film is still maintained.

As mentioned above, the film is a polymer based solar cell mimic and as such, the
ratio of P3HT and PCBM in the film is 1:1 by weight (47% by volume PCBM), with a
small volume fraction of p-CdSe particles. We hypothesize that as the film is spin coated

the p-CdSe particles assemble at the air interface, as in the pure P3HT sample. After

heating to 2500C, however, the pyridine layer coating the CdSe core is removed, leaving

behind only the high refractive index core. TGA results show the pyridine layer begins to

be removed from the CdSe core at 150°C, annealing at 250°C will certainly remove all

attached pyridine. With a refractive index of 5.8 the pure CdSe particles will migrate to
the substrate, as dictated by the Hamaker constant ordering. However, the PCBM, which
cannot be distinguished from P3HT in TEM images, seems to have formed a layer

comprising the bottom half of the film forcing the CdSe particles to assemble on top.

Thus, the actual ordering of the film, after annealing at 250°C is air, P3HT, CdSe,

PCBM, substrate. The results presented in Chapter 3 on the location of PCBM in this
films are a good verification of this hypothesis.

Another possible explanation of this assembly could be that the P3HT degrades to
a point where its refractive index is lower than the p-CdSe particles, forcing the particles
downward into the film. Regardless of the actual driving force for the CdSe particles, a
PCBM layer present at the bottom of the film is highly likely and most likely the case for
CdSe assembly in the middle of the film.

109

On a related note, much of the initial discussion in this chapter dealt with

dewetting of polymers but there is no real mention of dewetting work done on P3HT.

This is because P3HT would not dewet silicon, Si02 or Silanized substrates. Presumably,

the high crystallinity of P3HT forms such a strong network that thermal fluctuations are
not energetic enough to penetrate the crystalline network and exposed the lower energy
surface. Additionally, while P3HT has a relatively high energy backbone consisting of
repeat units with sulfur atoms, it also has extremely low energy hexyl side chains. The
presence of the low energy side chains, depending on their orientation within the thin
film, could be enough to offset the high energy backbone and allow the P3HT film to

remain stable on a low energy substrate.

6.4 Conclusions

Polymer thin films have many potential commercial applications, especially if the
location of functional nanoparticles added to the films can be controlled. Control over
the assembly of functional particles such as CdSe in glassy polymers such as PS and
PMMA can be achieved through modification of the surface stabilizing ligand.
Modification of this ligand changes the effective refractive index and determines the
interface of the fihn where the particle will assemble.

Assembling nanoparticles in crystalline polymers such as P3HT is also seemingly
feasible given higher boiling point solvents and using a modified solvent annealing
approach. One of the major difficulties in assembling nanoparticles within conjugated

polymer films seems to be determining the location of all particles through either TEM or

110

other methods. Because of the difficulties in preparing samples for TEM, other methods

such as reflectivity and ellipsometry may be more suitable for such characterization.

6.5 Experimental

6.5.1 Dewetting Experiments

Dewetting experiments were performed by floating polystyrene films onto either
Silanized silicon wafers or PET substrates. Polystyrene films were made by mixing 115
kDa PS (Polymer Source) with toluene and placing the solution on a rocking mixer for 24
hours. Refer to Appendix A for concentration versus thickness curves. For solutions
with nanoparticles, 60 kDa intramolecularly crosslinked PS nanoparticles were added to
the overall solution in a volumetric ratio that would provide one monolayer of surface
coverage of the nanoparticles. A cleaved mica surface (Ted Pella) was flooded with the
polystyrene solution and spin coated for 60 seconds at 5000 RPMS. After spin coating
the films were floated off into water and picked up by a Silanized silicon wafer.
Silanization was performed using the commercially available silanizing agent

Sigmacote ® (Sigma Aldrich) according to the manufacturer’s specifications. Briefly, a

roughly 1x1 cm2 silicon wafer (wafers were purchased as 100 mm diameter wafers from

Wafer World and cleaved to size using a razor blade) was flooded with Sigmacote ®
solution. After 30 seconds of exposure the solution was washed with hexanes and dried

under nitrogen. After floating the PS films onto the Silanized wafers the film and

substrates were placed in a 40°C vacuum oven for 24 hours to remove any residual water.

111

The substrates were then placed in a hot stage under an optical microscope, heated to
170°C and images were taken at predetermined intervals.

All dewetting experiments performed on PET substrates were prepared exactly

the same as those performed on wafers except no silanization was needed.

6.5.2 Transmission Electron Microscopy

Films prepared for cross sectional microscopy images were prepared in the same way as
the dewetting experiments with the appropriate polymers and nanoparticles added.
Specifications and sources for all of the materials are as followed: Crosslinkable PS was
synthesized by Craig Hawker’s group at UCSB, 60 kDa PMMA (Polymer Source),
PCBM (Nano-C), 35 kDa P3HT (Rieke Metals) and the o-CdSe and p-CdSe
nanoparticles were synthesized by Dr. Michael Wong’s group at Rice University.

All cross sectional images obtained using PS or PMMA were created by first spin
coating at 5000 RPMS for one minute a solution of polymer and nanoparticle dispersed in

toluene onto a silicon wafer with a 100 nm oxide layer grown on top. After casting the

films were annealed well above the polymers respective Tg (generally 170°C) and then

crosslinked at 220°C when applicable. For the multilayered films a second coating of

polymer and nanoparticle was added on top and the process repeated. After all coatings

and annealings the SiOz wafers were placed into a small vial containing 8% hydrofluoric

acid, which etches away the SiOz and leaves the film suspended above the Si substrate.

After etching a gold layer of about 30 nm was sputtered onto each film, which acts as a

112

marker to find and view the thin film. Without the gold layer the films would be virtually
impossible to find.

Once the gold has been sputtered Polybed 812 embedding resin is dropped onto

the films and cured in a 60°C oven for 24 hours. The resin coated substrates are then

dropped into liquid nitrogen, which cracks the fihn and removes the resin from the Si
wafer. Generally, but not always, the film adheres to the-resin and is removed from the
wafer. The thin strip of film containing resin is then cut into small pieces and placed into
embedding blocks for microtoming, covered again with Polybed 812 and cured for -24
hours. Once the blocks are cured the films were microtomed using a diamond knife on
an RMC Ultrarnicrotome into sections roughly 80 nm thick, floated onto copper grids and
viewed with a J EOL 1000 TEM at an accelerating voltage of 100 keV.

Cross sections of P3HT containing films were prepared slightly different as the
HF etch destroyed the P3HT films. Instead of spin coating on an SiOz coated Si wafer
the P3HT containing films were spin coated onto Si wafers coated with poly(ethylene
dioxythiophene) polystyrene sulfonate (PEDOT:PSS), a water soluble macromolecular
salt used as a hole conduction layer in organic solar cells and organic light emitting

diodes. After spin coating onto the PEDOT:PSS layers the films were floated off into

water and picked up with pieces of cured Polybed 812. The resin and films were then

dried in a 40°C vacuum oven overnight before a gold layer was sputtered on top. The

resin pieces were then cut and placed into embedding blocks and the procedure described

above was followed for viewing the samples.

113

6.5.3 Thermal Analysis

All differential scanning calorimetry was performed with a TA instruments Q1000
modulated DSC and all thermal gravimetric analysis was performed with a TA

instruments Q5 00.

6.5.4 Atomic Force Microscopy

All Atomic Force Microscopy was performed with a Pacific Nanotech Nano R AFM in

close contact mode with silicon tips.

114

CHAPTER 7: CONCLUSIONS

Polymer solar cells are a possible source for inexpensive renewable energy, but their
performance is dictated by structures on the nanscale. For the field of polymer solar cells
and organic electronics as a whole to push forward, high quality analysis of the internal
structure of these thin films is very important. In this work we have shown that neutron
scattering is a robust and important tool for studying polymer based solar cells. Neutron
reflectometry, through the use of phase sensitive neutron reflectivity, is a highly sensitive
measurement technique that has aided in successfully mapping the vertical PCBM
composition of P3HT:PCBM solar cells. We have found that after spin coating there is a
high concentration of PCBM at the substrate, and high concentration of PCBM near but
not at the metal electrode interface. PCBM at the substrate of a device is highly
undesirable as it prevents holes from leaving the device. Upon thermal annealing the
PCBM migrates towards the metal electrode, creating more efficient pathways for
electrons to leave the device.

Small angle neutron scattering (SANS) has shown that PCBM agglomerates into
clusters spread through the active layer upon thermal annealing. These clusters form
better charge transport networks allowing for more efficient electron harvesting but also
create more space between PCBM molecules. This increased spacing reduces the chance
a photogenerated exciton will find a dissociation point and increases photoluminescence.
The SANS data presented here explained this phenomenon of increased

photoluminescence with an increase in device performance.

115

In addition to characterizing these devices more accurately, we have created
nanorough surfaces that enhance the performance of these solar cells. The rough surfaces
increase contact area between the PEDOT:PSS layer and the active layer allowing for a
greater amount of charge transport. The Si02 particles that create the rough surfaces act

as light scatterers and slightly increase device absorption.

7.1 Future Work

Clearly there is still much need to further characterize these solar cells. We have a basic
understanding of the vertical profile of PCBM in a device, but the devices tested in this
work were based on a system that provided a quality sample for neutron reflectivity, not a
sample that provided the best performing solar cell. More work is needed to correlate the
highest performing devices with internal morphology. Phase sensitive neutron
reflectivity could be used to study these devices but much smaller samples and therefore
longer count times would be needed to do this. The formation of clusters within these
devices can also be further studied through SANS and these tests are independent of film
roughness; hence any device can be tested in SANS.

The devices fabricated on nanorough surfaces require more extensive
characterization as a full understanding of why they improve has not been developed.
We currently theorize that the P3HT is aligned more vertically in the device and therefore
provides more efficient charge transport, but this is a difficult aspect to study due to the

roughness of the film.

116

Additionally, it is clear a P3HT based device will not be commercialized because
it cannot achieve a high enough efficiency to be cost effective. New polymers are being

developed that produce higher efficiency cells and these devices need to be characterized.

117

APPENDIX A: HOW TO MAKE POLYMER BASED SOLAR CELLS

Polymer solar cells are not too difficult to make but require care in terms of cleaning and
timing to get a good and reproducible device. Depending on fabrication procedures,
typical devices can be made in 4-6 hours depending on the pump down time of the
evaporator and provided patterened slides are ready and solution is made. Over half a
day is needed for patterning and at least two days, preferably three, of stirring/dissolving

are needed for solution preparation.

A.l Patterned Slide Preperation
Polymer based solar cells have very low conductivities and, therefore, must be fabricated
using a transparent electrode for the top contact. Inorganic solar cells have high
conductivities allowing photoinduced charge carriers to travel long distances to small
electrodes, or busbars, on the top of the cell. All polymer based cells are fabricated
starting from the top of the cell, the part where light passes though first, and built
‘downward’ to the bottom of the cell. All cells used in this work were created on 1” glass
slides coated with indium tin oxide (ITO). All slides used in this work were purchased
from Delta Technologies (delta-technologies.com), although other suppliers exist, and
varied in resistivity from 8-15 ohms/cmz.

Contacting devices properly requires a patterned ITO slide, which can be
accomplished by some form of masking and etching. ITO is etched rather easily by a

solution consisting of a 60:40 volumetric ratio of hydrochloric acid (HCl) in water at

60°C. Immersion of ITO coated glass into this solution for 60 seconds will completely

118

etch away any exposed ITO. An easy way to remember this recipe is 60:60:60. That is,

60% HCl heated to 60°C and an immersion time of 60 seconds.

The most common and precise method to perform this etch is to use
photolithography, a method used in the semiconductor industry for etching oxides and
other coatings to achieve very precise patterns and small length scales. A full description
of how to use photolithography for etching ITO is described below, which is the method
used to make roughly half of the patterned slides used in this thesis. However, as precise
features are not required to make these devices, an easier method to patterning is to
simply cover the areas of the slide with transparent tape, such as Scotch® brand tape, and
to etch the slides in HCl as described above. Using Scotch® tape as the mask for etching
is much faster and still produces slides of decent quality. Residual adhesive left over
from the tape after the etch can be removed easily with a solvent. Methylethylketone
(MEK) works very well, and then the cleaning procedure described below can be
followed.

The main design used in this work allowed for 4 devices to be fabricated on one

slide with a total cell area of approximately 0.2 cm2.

Photolithography is accomplished in 7 steps:

1) Photoresist application
The photoresist used here was S1818 from Rhom and Hass, which is a positive
photoresist. A positive photoresist will break down upon exposure to UV light

and be easily washed away. An ITO coated slide is placed on a vacuum chuck of

119

2)

3)

a spin coater and a drop of photoresist is placed on top. The photoresist does not
need to cover the entire slide before spinning. Slides are spun at 2000 RPM for

15 seconds, which creates a film that is on the order of microns thick.

Pre—bake

Once all slides have been coated with photoresist they are placed in an oven at
about 90°C and baked for about 20 minutes. The temperature and time are
approximate as this bake is used only to remove residual solvent from the

photoresist.

UV-exposure

After the pre—bake all slides are placed under a photolithographic mask and
exposed to UV light for 20 seconds. UV light breaks down the photoresist and
allows it to be washed away with a developer solution. Exposed areas of the
slides will have the photoresist removed and areas covered by the mask will retain
the photoresist. The photoresist used for most of the work here was created by
printing multiple copies of a transparency with the desired pattern and then gluing
them on top of each other. Figure A1 below is a picture of the mask used that
accommodates 4 slides and the pattern of ITO it creates. The dotted line in the

upper left hadn comer of the mask shows where one slide would sit.

120

 

,-

 

 

 

I I Fl."1
war
__.I

 

Figure A1: A cartoon of the mask used to make the majority of the devices used in this
work and the associated ITO pattern it eventually creates. The dashed square is the
outline of a slide the mask will cover.

4)

5)

Developing

Once all the slides have been exposed to UV light, the irradiated photoresist needs
to be washed away to expose the ITO to be etched. This is done by placing the
slides in a solution of developer and water (3 parts developer to 1 part water) for
about 1 minute. The total developer time depends on how much water is added
but one can look at the slides and tell if they have been fully developed because

the underlying pattern is exposed.

Hard Bake

Upon completion of developing, the slides are placed in an oven at 120°C for 60

minutes to completely bake the photoresist. This hard bake will prevent the

photoresist from washing away during the etching procedure.

121

6)

7)

3)

Etching
Etching is the process of removing the exposed ITO from the glass to create a

pattern. This is done by placing the slides in a solution of 60% HCl (40% H20) at

60°C for 60 seconds. Exposure for too long will begin to remove the ITO

protected by the photoresist by etching under, but too short of an exposure will
leave trace amounts of ITO on the glass, causing highly resistive pathways
between different cells. Each slide must be measured with a voltage meter before
the final step or the slide could be worthless. If a small amount of current can
leak between etched components the slide must be placed in the etch for a longer

time.

Stripping
The final step is to remove the remaining photoresist by placing the slides in a
striping agent. Time spent in the stripper is not important provided all the

photoresist is removed. A few minutes is suffice.

Final Check

All devices need to be checked after completion of the photolithography to see
there are no connections between different contact areas of the cell. If a pathway
exists from one ITO area to another it is most likely along the edge of the glass
slide and be caused from slippage of the mask during UV exposure. This is very
common. To fix this a glass scribe can be run across the connection until it is

scratched through and no longer conductive.

122

A.2 Slide Cleaning

Proper cleaning of the patterned slides is essential to making devices. Without clean
surfaces the spin coated films will not be uniform and the charge carriers will encounter
many barriers at the PEDOT:PSS / ITO interface. Only slides being used to make
devices that day should be cleaned as cleaning and storing slides will not work. The
devices should be sonicated first in acetone for 15 minutes and then in isopropanol for 15
minutes. Some groups first sonicate in soapy DI water, which was occasionally done in
this work, but this requires a great amount of rinsing before the acetone cleaning. If the
slides were not rinsed properly the soap would occasionally leave a small film on certain
areas of a slide preventing good PEDOT:PSS contact. Hence, using soapy water as a pre-
clean could cause as much harm as good. Additionally, the acetone and isopropanol

could be reused many times before discarding.

A.3 PEDOT:PSS Application
PEDOT:PSS (poly(3,4-ethylenedioxythiophene:polystyrene sulfonate) is an electron
blocking, or hole conducting, layer that covers the ITO electrode. This layer aids in
preventing recombination. Once the slides are cleaned they should be briefly dried with
nitrogen and the directly coated with PEDOT:PSS.

The PEDOT:PSS used in this work was Clevios 4083P, medium conductivity
grade and was stored in a refrigerator until use. The solution must be heated to room
temperature before spin coating. All solutions were first diluted with water in a ration of

4 ml pure PEDOT:PSS stock solution to 1 ml of Millipore water. The solution was

123

deposited on the ITO slides from a 3 ml syringe while being filtered through a 200 nm
Nylon filter. All slides were completely flooded with the PEDOT:PSS solution and the
spin coated for 60 seconds at 3000 RPM and then a second coating was deposited
through the syringe filter and again spun at the same conditions. Two coatings are spin
coated to reduce the chance of having pin holes through the layer. This procedure results
in a film of a roughly 25 nm thick layer of PEDOT:PSS.

After spin coating a small, strip of the PEDOT:PSS film must be removed from
the entire edge of the wafer, exposing the contact areas of the slides. This is done by
wetting a cotton applicator and wiping the edges of each slide. After wiping off the
edges of the slides they are all placed in an oven at 130°C and baked for 20 minutes.
Then, a vacuum is pulled on the oven and the slides are baked for another 20 minutes
under vacuum. Upon completion of baking they are removed, placed in holders and
transferred directly to the glove box. Figure A2 is a cartoon of the PEDOT:PSS

application process from the top and cross sectional views.

 

 

 

I} I?

 

 

 

 

 

 

 

 

 

 

 

 

l I l 7 L I

Figure A2: Top down and cross sectional views of applying PEDOT:PSS onto patterned
ITO slides and the subsequent removal of the layer from the edges to allow for good
electrical contact during testing.

124

A.4 Active Layer Application
Application of the active layer can have many variables from solution concentration,
solvent choice, post fabrication annealing or drying or a number of other parameters. The
most important aspect of active layer application is to be sure the solution composition is
known and to allow the active layer to dry completely. When using high boiling point
solvents, such as dichlorobenzene, many hours of vacuum may be needed to remove all
the available solvent from the active layer.

To make a solution, both polymer and nanoparticle can be weighed and placed in
a vial outside the glove box. Once the appropriate amounts are added, the vials are
returned to the glove box, with caps loose as to remove any oxygen during transfer, and
the desired solvent is added. Anhydrous solvents should always be used and the
solutions should be stirred for a sufficiently long time before spin coating. Solutions

from cholorbenzene require at least 2 if not 3 days of stirring and solutions using

dichlorobenzene need at least two days of stirring and heating up to 50°C. Heating all

solutions may shorten total dissolution time but no work has been done in this thesis to
determine the effects, if any, on how heating effects device performance.

The most common concentration used in this work was a 30 mg/ml solution (15
mg polymer and 15 mg of PCBM to 1 ml of solvent), but many concentrations can be
used depending on thickness and polymerznanoparticle ratio desired. Spin speeds from
500 RPM to 2500 RPM are common. Generally the lower spinning speeds produce more
efficient devices and the higher spinning speeds produce flatter films that are desirable

for a variety of characterization techniques, namely reflectivity.

125

Once the active layer has been created, the same wiping of the film from the
contacts must be performed. This is done by wetting a cotton applicator in either
chloroform or chlorobenzene and wiping in the same manner as was used with the
PEDOT:PSS films. More care must be taken with these devices as they must be wiped
inside the glove box and that can be difficult while wearing bulky gloves. Figure A3 is a

cartoon of the active layer application process.

 

 

 

   

up

 

 

 

 

 

 

 

 

g, ...:

I
l
"' .‘nifi‘r 'u-i'f -
I'v‘mxtflfm
_ l
Z‘i'fii-infl '-. _

 

 

Figure A3: Top down and cross sectional views of applying the active layer onto
patterned ITO slides with a PEDOT:PSS layer and the subsequent removal of the layer
from the edges to allow for good electrical contact during testing.

A.5 Electrode Deposition

Once the active layer has been deposited, any post deposition annealing has been
performed and the edges have been cleaned of any film, the slides are inserted into the
thermal evaporator. Set-up is performed according to the manufactures instructions and
whatever electrode is to be studied is evaporated onto the devices. A mask is chosen
according to the ITO patterned onto the slides. For some devices used in this work an
lithium fluoride layer about 1 nm was first evaporated and then an aluminum layer of
about 80 nm followed. Other devices only had an aluminum layer. Due to the intense

heat of the filament, a sufficiently large distance between the devices and filament is

126

desired and a relatively small rate is also desired. Most devices in this work used an
aluminum deposition rate of approximately 0.2 nm/sec. The metal electrode covers part
of one patterend ITO strip and extends outward to cover an patterened ITO island near
the edge of the cell. This island is used because when a probe contacts the metal
electrode extension it can still make a good contact even if the metal is scratched firlly
away directly around the probe. Because the ITO is in direct contact with the metal, the
probe only needs to touch the ITO and is therefore contacting the metal electrode. Figure
A4 is a cartoon describing the placement of the metal electrode and the created fully

active solar cells.

 

 

 

 

 

 

 

 

 

L 2.). a" .3.
'l .‘ l

is- _.. M- :3]

 

 

 

 

Figure A4: Top down and cross sectional views of deposited electrode. The dashed
squares indicate where the active area of the four solar cells created from this masking
pattern.

127

 

 

 

 

at w» ->

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l l l l l I l l
ITO Coated Patterned ITO Apply pEDor;p55 Remove
Glass Slide Coated Glass Slide PEDOT:PSS

_- . _ . g'; F .. g 1 .27..” -:--‘fi_=ir
ED [1] ., _.§i -

. --.- .. :4.» J _ gm] F .
.. "I. ,

 

 

l l l l H

Apply bottom Remove Apply active
contact active layer layer

 

 

Figure A5: Process map of the 6 steps (starting from clean ITO slides) needed to make a

polymer based solar cell on glass slides using a transparent electrode, in this case indium
tin oxide (ITO).

128

APPENDIX B: POLYMER FILM THICKNESSES

Spin coating is an extremely popular laboratory scale method to create thin films of
polymers. Although it is not industrially applicable for deposition of polymers, ink jet
printing or a modified dip coating will be much more commercially applicable, its ease of
use and ability to get extremely smooth films make it an excellent research tool. This
purpose of this appendix is to provide plots of concentrations versus thickness for the

various polymers at various spin coating speeds as a reference for future work.

The majority of the following polymers were spin coated at a speed of 5000 RPMS fi'om
toluene. If no solvent or spinning speed is noted,“ those are the values. Many different
methods of determining film thicknesses was used including ellipsometry, AF M, X-ray
reflectivity and neutron reflectivity. If no method for determining the height is
mentioned, it was ellipsometry. Also, all substrates used in the following plots were Si
wafers with a native oxide grown on top and in every case the oxide layer was accounted
for to the best of the ability of the test. Error bars only appear on the P3HT plot because
it was the only polymer where AF M was used to obtain height information, which is a
noisy method for determining height. Because the error bars for ellipsometry and both

reflectivity methods are much smaller than the symbols on the plots they do not appear.

129

B.l Polystyrene

 

 

 

 

 

 

 

 

30 —< . I - ; -' g - ‘ ..
A
— A
E 25 - . ‘ on—
E) :
E e L 0‘ CI
v 20— = _ A -
.5 .
...;
m 15 —~ 0 . l a E] -
4‘: o til
C s
8 10— 0 amp ' t
c A Toluene, 5000 RPM
0 A o Pyridine, 5000RPM
0 5L ‘ .an , El Toluene,5000 RPM_

0938» o Toluene, 2000 RPM
0 G] . l . l r l .
0 40 80 120 160

Thickness (nm)

Figure B1: Concentration versus height data for 75 kDa polystyrene. Both toluene and
pyridine were used as solvents and spinning speeds were 2000 and 5000 RPMS. Closed
symbols represent height data from ellipsometry and open symbols represent height data
obtained from X—ray reflectivity. Both Erica Tseng and Wenluan Zhang aided in
obtaining the PS height data.

130

B.2 Poly(methyl methacrylate)

 

 

 

 

 

16 I I 1 l l T
+ 72.6 kDa PMMA in Chlorobenzene
+ 49.2 kDa PMMA in Toluene 7
c 14 T + 72.6 kDa PMMA in Toluene ' _
E5 1
E
v
c:
.S! 10" ‘
...—a
g
c: 8- p
a)
‘c’
o 6 - -
C)
4 " . ; _
l J l l .i l

 

 

 

1O 15 20 25 30 35
Thickness (nm)

Figure B2: Concentration versus height data of 72.6 kDa and 49.2 kDa poly(methyl
methacrylate) in toluene and chlorobenzene. All films were spin coated at 5000 RPMS
and all data is from ellipsometry.

131

B.3 Poly(3-hexyl thiophene)

 

 

 

 

 

T I
[+ P3HT in Chlorobezene

.3
.5.

.3
N
I

Concentration (mg/ml)
00 ES
I l

 

 

 

l I l l
5 10 15 20 25 30

 

Thickness (nm)

Figure B3: Concentration versus height data of P3 HT in chlorobenzene spin coated at
5000 RPMS. Height data was obtained from scratching a film and measuring the step
height of the scratch.

132

 

B.4 Poly(dimethyl siloxane)

 

 

 

 

 

 

 

 

 

l l l
14 — +PDMS in Toluene _
E
a, 12 —- 1
£5,
.5 10 — —
g
8 8 - ~
0
S
0 6 — _
4 l l l
10 20 30 4O 50

Thickness (nm)

Figure B4: Concentration versus height of PDMS in toluene spin coated at 5000 RPMS.
The PDMS used in this study was cross-linkable PDMS from Dow, Sylgard® 184, that

can be cross-linked within 30 minutes at 600C. Height data was obtained from

ellipsometry.

133

B.5 Polyisobutene

 

 

 

 

 

 

F l l l
25 "" + Polyisobutene in toluene **** ‘
C
E
\o) 20 - a
E
v
C _ _
.9 15
...-o
02 1O - 1
0
8
o 5 _ ._
0 I I I I

 

 

 

20 40 60 80
Thickness (nm)

Figure B5: Concentration versus height of PIB in toluene spin coated at 5000 RPMS.
The PIB used in this study was low molecular weight and the polymer was liquid at room
temperature. Height data was obtained from ellipsometry.

134

APPENDIX C: IDEALITY FACTOR PLOTS

C.l P3HT:PCBM device on a pure PEDOT:PSS substrate with no LiF between
the Al and active layer.

 

I I I I I I
15 —
+ Light
—e— Dark
10 — a
5 — _

 

 

Current Density (mA/cmz)

 

 

 

 

 

-0.4 0.0 0.4 0.8
Voltage (V)

Figure Cl: J-V curve of a P3HT:PCBM device on pure PEDOT:PSS with no LiF layer

between the active layer and aluminum electrode. The series resistance reported is taken
from the slope of the J -V curve at the open circuit voltage.

135

dJ/dV (mA cm2 I V)

 

l l
6—
4—
go
-0
2" . -...o
9..
P- .

 

 

 

0.0
Voltage (V)

 

Figure C2: dJ/dV versus V plot of a P3HT:PCBM device on pure PEDOT:PSS with no
LiF layer between the active layer and aluminum electrode. The voltage dependence of
the illuminated current shows that traditional methods to evaluate the device must be
based on the dark curve, which has minimal dependence on the voltage.

136

 

   
    

 

 

 

A = 1.6
A 0'10 _ R8 = 32 Q cm2 —
<
E
N‘ 0.08 — _.
E
0
Z
a 0.06 — _
'2
> 0°
'C
0.04 — ° -+
A = 3.2
R3 = 18.6 :2 cm2
0.02 . - '
0.0 0.5 1.0 1.5 2.0

.1-1 (cm2 / mA)

Figure C3: dV/dJ versus J'l plot of a P3HT:PCBM device in the dark on pure
PEDOT:PSS with no LiF layer between the active layer and aluminum electrode.

137

C.2 P3HT:PCBM device on a PEDOT:PSS substrate with 10 mglml SiOz
particles dispersed within the PEDOT:PSS solution.

 

 

 

 

30 I T I I I l I
I + Light
: -e— dark
NE .
o 20 .- i h
l
E .
v I
g I
w 10 -- I ~-
C l
8 ' 2
... Rs = 24 Q cm
C .
e l
s o . - -------- -
O u
I
I
l
-10 i . | . A . J
-0.4 0.0 0.4 0.8
Voltage (V)

Figure C4: J-V curve of a P3HT:PCBM device on PEDOT:PSS with 10 mg/ml of Si02

particles dispersed within and no LiF layer between the active layer and aluminum

electrode. The series resistance reported is taken from the slope of the J-V curve at the
open circuit voltage.

138

 

dJ/dV (mA / v cmz)

 

 

 

 

-O.4 -0.2 0.0 0.2 0.4

Voltage (V)

Figure C5: dJ/dV versus V plot of a P3HT:PCBM device on PEDOT:PSS with 10 mglml
of SiOz particles dispersed within and no LiF layer between the active layer and

aluminum electrode. The voltage dependence of the illuminated current shows that
traditional methods to evaluate the device must be based on the dark curve, which has
minimal dependence on the voltage.

139

 

80x10‘3 1 I

    

 

 

 

| I
A = 1.4
_ 2
A so _ Rs — 22 0 cm
<
E
"E
Z
5
2
>
'° 20 — —
A = 3.6
R8 = 9 (2 cm2
0 l I l l
0.0 0.2 0.4 0.6 0.8 1.0

J" (cm2 I mA)

Figure C6: dV/dJ versus J.1 plot of a P3HT:PCBM with 10 mg/ml of SiOz particles

dispersed within the PEDOT:PSS and no LiF layer between the active layer and
aluminum electrode.

140

C.3 P3HT:PCBM device on a PEDOT:PSS substrate with 20 mglml SiOz
particles dispersed within the PEDOT:PSS solution.

 

 

 

 

I ' I . I ' I
I + Light
40 — : -B— Dark
NA I
g I
I
E 30 :
V I
g 20 — I —
a) I
C I
8 I
10 — 2 q
“E Rs = 15.6 0 cm
g :
3 DE ------ —
O I
I
-10 — . -
I I l . l . I
-0 4 0.0 0 4 0 8
Voltage (V)

Figure C7: J-V curve of a P3HT:PCBM device on PEDOT:PSS with 20 mg/ml of SiOz

particles dispersed within and no LiF layer between the active layer and aluminum
electrode. The series resistance reported is taken from the slope of the J-V curve at the
open circuit voltage.

141

 

I Light I
[3 Dark

dJ/dV (mA / v cmz)

 

 

 

 

Voltage (V)

Figure C8: dJ/dV versus V plot of a P3HT:PCBM device on PEDOT:PSS with 20 mg/ml

of Si02 particles dispersed within and no LiF layer between the active layer and

aluminum electrode. The voltage dependence of the illuminated current shows that
traditional methods to evaluate the device must be based on the dark curve, which has
minimal dependence on the voltage.

142

020 I I I I I I

 

 

E 0.15 —
\
N
S
0.10 2
>
B
s W/ D
'° .5— _

[:1
D A=2.6

f5] Rs = 6 Q cm2
0.00 I I I I I

0 1 2 3 4 5 6 7
.I'1 (cm2 / mA)

 

\

 

—

Figure C9: dV/dJ versus J.l plot of a P3HT:PCBM with 20 mg/ml of Si02 particles

dispersed within the PEDOT:PSS and no LiF layer between the active layer and
aluminum electrode.

143

C.4 P3HT:PCBM device on a PEDOT:PSS substrate with 40 mglml SiOz
particles dispersed within the PEDOT:PSS solution.

    

 

40

30

20

1O

 

. D
- - - - - - - - - - - - - - - - - - - - - -
D

Current Density (mA/cmz)

 

 

-o.4 0.0 0.4 0.8
Voltage (V)

Figure C10: J-V curve of a P3HT:PCBM device on PEDOT:PSS with 40 mg/ml of SiOz

particles dispersed within and no LiF layer between the active layer and aluminum
electrode. The series resistance reported is taken from the slope of the J-V curve at the
open circuit voltage.

144

 

A Light A
A Dark
4 — ..
M
A
3 - ‘ _

dJ/dV (mA / v cm2)

 

 

 

 

Voltage

Figure C11: dJ/dV versus V plot of a P3HT:PCBM device on PEDOT:PSS with 40

mg/ml of SiOz particles dispersed within and no LiF layer between the active layer and

aluminum electrode. The voltage dependence of the illuminated current shows that
traditional methods to evaluate the device must be based on the dark curve, which has
minimal dependence on the voltage.

145

 

0.040 I l l I

  
    
 

0.035 — A = 1-5

R8 = 9.6 Q cm
0.030 —

0.025 -

0.020 P

dV/dJ (v cm2 / mA)

0.015 -

 
 

A = 2.4
R5 = 7.6 Q cm2 7

0.005 J ' ' '
0.0 0.2 0.4 0.6 0.8

0'1 (cm2 / mA)

0.010

 

 

 

Figure C12: dV/dJ versus J.1 plot of a P3HT:PCBM with 40 mglml of SiOz particles

dispersed within the PEDOT:PSS and no LiF layer between the active layer and
aluminum electrode.

146

10.

11.

12.

13.

14.

15.

REFERENCES

N. S. Lewis and D. G. Nocera, Proceedings of the National Academy of the
Sciences, 2006, 103, 15731.

www.eia.doe.gov.

G. Yu, J. Gao, J. C. Hummelen, F. Wudl and A. J. Heeger, Science, 1995, 270,
1789-1791.

P. Wurfel, Physics of Solar Cells: From Principles to Basic Concepts, Wiley-
VCH, Weinheim, 2009.

M. A. Green, Silicon Solar cells: Advanced principles and practice, Centre for
Photovoltaic Devices and Systems, Sydney, 1995.

R. N. Marks, J. J. M. Halls, D. D. C. Bradley, R. H. Friend and A. B. Holmes,
Journal of Physics-Condensed Matter, 1994, 6, 1379-1394.

Y. X. Liu, M. A. Summers, C. Edder, J. M. J. Frechet and M. D. McGehee,
Advanced Materials, 2005, 17, 2960.

N. S. Saricifici, L. Smilowitz, A. J. Heeger and F. Wudl, Science, 1992, 258,
1474-1476. '

D. E. Markov, E. Amsterdam, P. W. M. Blom, A. B. Sieval and J. C. Hummelen,
Journal of Physical Chemistry A, 2005, 109, 5266-5274.

J. J. M. Halls, K. Pichler, R. H. Friend, S. C. Moratti and A. B. Holmes, Applied
Physics Letters, 1996, 68, 3120-3122.

K. M. Coakley and M. D. McGehee, Chemistry of Materials, 2004, 16, 4533-
4542.

C. J. Brabec, N. S. Saricifici and J. C. Hummelen, Advanced Functional
Materials, 2001, 11, 1526.

G. Li, Y. Yao, H. Yang, V. Shrotriya, G. Yang and Y. Yang, Advanced
Functional Materials, 2007, 17, 1636-1644.

G. Li, V. Shrotriya, J. S. Huang, Y. Yao, T. Moriarty, K. Emery and Y. Yang,
Nature Materials, 2005, 4, 864-868.

G. Li, V. Shrotriya, Y. Yao and Y. Yang, Journal of Applied Physics, 2005, 98.

147

 

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

R. J. Kline, M. D. McGehee, E. N. Kadnikova, J. S. Liu, J. M. J. Frechet and M.
F. Toney, Macromolecules, 2005, 38, 3312-3319.

Y. Kim, s. A. Choulis, J. Nelson, D. D. C. Bradley, 3. Cook and J. R. Durrant,
2005.

J. F. Zhu, F. Bebensee, W. Hieringer, W. Zhao, J. H. Baricuatro, J. A. Farmer, Y.
Bai, H. P. Steinruck, J. M. Gottfried and C. T. Campbell, J. Am. Chem. Soc.,
2009, 131, 13498-13507.

C. J. Brabec, S. E. Shaheen, C. Winder, N. S. Sariciftci and P. Denk, Applied
Physics Letters, 2002, 80, 1288-1290.

Y. Kim, Nanotechnology, 2008, 19.

C. Honsberg and S. Bowden, http://pvcdrom.pveducation.org.

M. D. Irwin, B. Buchholz, A. W. Hains, R. P. H. Chang and T. J. Marks,
Proceedings of the National Academy of Sciences of the United States of America,

2008, 105, 2783-2787.

M. S. White, D. C. Olson, S. E. Shaheen, N. Kopidakis and D. S. Ginley, Applied
Physics Letters, 2006, 89.

Y. Kim, S. Cook, S. M. Tuladhar, S. A. Choulis, J. Nelson, J. R. Durrant, D. D. C.
Bradley, M. Giles, I. McCulloch, C. S. Ha and M. Ree, Nature Materials, 2006, 5,
197-203.

G. Dennler, M. C. Scharber and C. J. Brabec, Advanced Materials, 2009, 21,
1323-1338.

S. H. Park, A. Roy, S. Beaupre, S. Cho, N. Coates, J. S. Moon, D. Moses, M.
Leclerc, K. Lee and A. J. Heeger, Nature Photonics, 2009, 3, 297-U295.

M. E. Mackay, A. Tuteja, P. M. Duxbury, C. J. Hawker, B. Van Horn, Z. B.
Guan, G. H. Chen and R. S. Krishnan, Science, 2006, 311, 1740-1743.

Y. B. Melnichenko, G. D. Wignall, R. N. Compton and G. Bakale, J Chem Phys,
1999, 111, 4724-4728.

M. E. Mackay, T. T. Dao, A. Tuteja, D. L. Ho, B. Van Horn, H. C. Kim and C. J.
Hawker, Nature Mater., 2003, 2, 762-766.

E. S. McGarrity, A. L. Frischknecht, L. J. D. Frink and M. E. Mackay, Physical
Review Letters, 2007, 99, 4.

148

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

R. S. Krishnan, M. E. Mackay, P. M. Duxbury, C. J. Hawker, S. Asokan, M. S.
Wong, R. Goyette and P. Thiyagarajan, Journal of Physics-Condensed Matter,
2007, 19.

R. S. Krishnan, M. E. Mackay, P. M. Duxbury, A. Pastor, C. J. Hawker, B. Van
Horn, S. Asokan and M. S. Wong, Nano Letters, 2007, 7, 484-489.

R. S. Krishnan, M. E. Mackay, C. J. Hawker and B. Van Horn, Langmuir, 2005,
21, 5770-5776.

S. Gupta, Q. L. Zhang, T. Emrick, A. C. Balazs and T. P. Russell, Nature
Materials, 2006, 5, 229-233.

Y. Lin, A. Boker, J. B. He, K. Sill, H. Q. Xiang, C. Abetz, X. F. Li, J. Wang, T.
Emrick, S. Long, Q. Wang, A. Balazs and T. P. Russell, Nature, 2005, 434, 55-
59.

Y. Lin, H. Skaff, T. Emrick, A. D. Dinsmore and T. P. Russell, Science, 2003,
299, 226-229.

M. A. Holmes, M. E. Mackay and R. K. Giunta, J. Nanopart. Res, 2007, 9, 753-
763.

R. S. Ruoff, R. Malhotra, D. L. Huestis, D. S. Tse and D. C. Lorents, Nature,
1993, 362, 140-141.

R. S. Ruoff, D. S. Tse, R. Malhotra and D. C. Lorents, J. Phys. Chem, 1993, 97,
3379-3383.

B. S. McGarrity, P. M. Duxbury, M. E. Mackay and A. L. Frischknecht,
Macromolecules, 2008, 41, 5952-5954.

B. S. McGarrity, A. L. Frischknecht and M. E. Mackay, Journal of Chemical
Physics, 2008, 128.

M. A. Yaklin, P. M. Duxbury and M. E. Mackay, Soft Matter, 2008, 4, 2441 -
2447.

P. A. Troshin, H. Hoppe, J. Renz, M. Egginger, J. Y. Mayorova, A. E.
Goryochev, A. S. Peregudov, R. N. Lyubovskaya, G. Gobsch, N. S. Sariciftci and
V. F. Razumov, Adv F unct Mater, 2009, 19, 779-788.

E. J. Roche and E. L. Thomas, Polymer, 1981, 22, 333-341.

D. L. Handlin and E. L. Thomas, J Mat Sci Lett, 1984, 3, 137-140.

149

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

D. ,L. Handlin and E. L. Thomas, Macromolecules, 1983, 16, 1514-1525.

J. W. Andreasen, S. A. Gevorgyan, C. M. Schleputz and F. C. Krebs, Solar
Energy Mat Solar Cells, 2008, 92, 793-798.

S. S. van Bavel, E. Sourty, G. de With and J. Loos, Nano Letters, 2009, 9, 507-
513.

J. Y. Kim, K. Lee, N. E. Coates, D. Moses, T. Q. Nguyen, M. Dante and A. J.
Heeger, Science, 2007, 317, 222-225.

J. S. Moon, J. K. Lee, S. N. Cho, J. Y. Byun and A. J. Heeger, Nano Letters,
2009, 9, 230-234.

H. Hoppe, M. Niggemann, C. Winder, J. Kraut, R. Hiesgen, A. Hinsch, D.
Meissner and N. S. Saricifici, Adv Funct Mater, 2004, 14, 1005-1011.

M. Campoy-Quiles, T. Ferenczi, T. Agostinelli, P. G. Etchegoin, Y. Kim, T. D.
Anthopoulos, P. N. Stavrinou, D. D. C. Bradley and J. Nelson, Nature Materials,
2008, 7, 158-164.

D. L. Handlin, E. L. Thomas and W. J. Macknight, Macromolecules, 1981, 14,
795-801.

E. L. Thomas and E. J. Roche, Polymer, 1979, 20, 1413-1422.

Z. Xu, L. M. Chen, G. Yang, C. H. Huang, J. Hou, Y. Wu, G. Li, C. S. Hsu and
Y. Yang, in Advanced Functional Materials, Editon edn., 2009, vol. 19.

D. S. Germack, C. K. Chan, B. H. Hamadani, L. J. Richter, D. A. Fischer, D. J.
Gundlach and D. M. DeLongchamp, Applied Physics Letters, 2009, 94, 3.

C. F. Majkrzak, N. F. Berk, V. Silin and C. W. Meuse, Phyisca B, 2000, 283, 248-
252.

L. G. Parratt, Physical Review, 1954, 95, 359-369.

C. F. Majkrzak, N. F. Berk and U. A. Perez-Salas, Langmuir, 2003, 19, 7796-
7810.

C. F. Majkrzak and N. F. Berk, Physical Review B, 1995, 52, 10827-10830.

V. O. Dehaan, A. A. Vanwell, S. Adenwalla and G. P. Felcher, Physical Review
B, 1995, 52, 10831-10833.

150

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

C. F. Majkrzak and N. F. Berk, Physical Review B, 1998, 58, 15416.

M. Campoy-Quiles, T. Ferenczi, T. Agostinelli, P. G. Etchegoin, Y. Kim, T. D.
Anthopoulos, P. N. Stavrinou, D. D. C. Bradley and J. Nelson, Nature Materials,
2008, 7, 158-164.

T. J. Prosa, M. J. Winokur, J. Moulton, P. Smith and A. J. Heeger,
Macromolecules, 1992, 25, 4364-4372.

S. van Bavel, E. Sourty, G. de With, K. Frolic and J. Loos, Macromolecules,
2009, 42, 7396-7403.

P. A. Heiney, J. E. Fischer, A. R. Mcghie, W. J. Romanow, A. M. Denenstein, J.
P. Mccauley, A. B. Smith and D. E. Cox, Phys. Rev. Lett., 1991, 66, 2911-2914.

W. Geens, T. Martens, J. Poortmans, T. Aemouts, J. Manca, L. Lutsen, P.
Heremans, S. Borghs, R. Mertens and D. Vanderzande, Thin Solid Films, 2004,
451-52, 498-502.

N. F. Berk and C. F. Majkrzak, Langmuir, 2009, 25, 4132-4144.

D. McBranch, M. Sinclair, A. Hays, D. Moses and A. J. Heeger, Syn Metals,
1992, 49, 147-157.

D. Datta, V. Tripathi, P. Gogoi, S. Banerjee and S. Kumar, Thin Solid Films,
2008, 516, 7237-7240.

Z. Xu, L. M. Chen, G. Yang, C. H. Huang, J. Hou, Y. Wu, G. Li, C. S. Hsu and
Y. Yang, Advanced Functional Materials, 2009, 19.

P. E. Shaw, A. Ruseckas and I. D. W. Samuel, Advanced Materials, 2008, 20,
3516-+.

J. E. Kroeze, T. J. Savenije, M. J. W. Verrneulen and J. M. Warman, Journal of
Physical Chemistry B, 2003, 107, 7696-7705.

S. E. Shaheen, C. J. Brabec, N. S. Saricifici, F. Padinger, T. Fromherz and J. C.
Hummelen, Applied Physics Letters, 2001, 78, 841-843.

D. Chirvase, J. Parisi, J. C. Hummelen and V. Dyakonov, Nanotechnology, 2004,
15, 1317-1323.

T. Erb, U. Zhokhavets, G. Gobsch, S. Raleva, B. Stuhn, P. Schilinsky, C. Waldauf
and C. J. Brabec, Advanced Functional Materials, 2005, 15, 1193-1196.

151

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

91.

M. Y. Chiu, U. S. Jeng, C. H. Su, K. S. Liang and K. H. Wei, Advanced
Materials, 2008, 20, 2573-+.

V. D. Mihailetchi, H. X. Xie, B. de Boer, L. J. A. Koster and P. W. M. Blom,
Advanced Functional Materials, 2006, 16, 699-708.

M. Y. Chiu, U. S. Jeng, M. S. Su and K. H. Wei, Macromolecules, 2010, 43, 428-
432.

J. W. Kiel, B. J. Kirby, C. F. Majkrzak, B. B. Maranville and M. E. Mackay, Soft
Matter, 2010, 6, 641-646.

D. S. Germack, C. K. Chan, R. J. Kline, D. A. Fischer, D. J. Gundlach, M. F.
Toney, L. J. Richter and D. M. DeLongchamp, Macromolecules, 2009, In Press.

C. Waldauf, M. Morana, P. Denk, P. Schilinsky, K. Coakley, S. A. Choulis and C.
J. Brabec, Applied Physics Letters, 2006, 89.

F. C. Krebs, S. A. Gevorgyan and J. Alstrup, Journal of Materials Chemistry,
2009, 19, 5442-5451.

A. P. R. Eberle, N. J. Wagner, B. Akgun and S. K. Satija, Langmuir, 26, 3003-
3007.

A. Tuteja, M. E. Mackay, C. J. Hawker and B. Van Horn, Macromolecules, 2005,
38, 8000-8011.

P. Akcora, S. K. Kumar, J. Moll, S. Lewis, L. S. Schadler, Y. Li, B. C.
Benicewicz, A. Sandy, S. Narayanan, J. Illavsky, P. Thiyagarajan, R. H. Colby
and J. F. Douglas, Macromolecules, 2010, 43, 1003-1010.

R. L. Jones, S. K. Kumar, D. L. Ho, R. M. Briber and T. P. Russell, Nature, 1999,
400, 146-149.

R. L. Jones, S. K. Kumar, D. L. Ho, R. M. Briber and T. P. Russell,
Macromolecules, 2001, 34, 559-567.

M. A. Yaklin, P. M. Duxbury and M. E. Mackay, Soft Matter, 2008, 4, 2441-
2447.

J. S. Higgins and H. C. Benoit, Polymers and Neutron Scattering, Clarendon
Press, Oxford, 2002.

M. Yin, N. Henry, K. Xiao, J. Ankner and M. Dadum, Submitted

152

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

D. M. DeLongchamp, R. J. Kline, Y. Jung, D. S. Germack, E. K. Lin, A. J. Moad,
L. J. Richter, M. F. Toney, M. Heeney and I. McCulloch, Acs Nano, 2009, 3, 780-
787.

L. H. Jimison, A. Salleo, M. L. Chabinyc, D. P. Bernstein and M. F. Toney,
Physical Review B, 2008, 78.

S. Hugger, R. Thomann, T. Heinzel and T. Thurn-Albrecht, Colloid Polym. Sci,
2004, 282, 932-938.

P. K. Watkins, A. B. Walker and G. L. B. Verschoor, Nano Letters, 2005, 5,
1814-1818.

S. R. Kline, Journal of Applied Crystallography, 2006, 39, 895.

T. C. Tseng, E. S. McGarrity, J. W. Kiel, P. M. Duxbury, M. E. Mackay, A. L.
Frischknecht, S. Asokan and M. S. Wong, Soft Matter, 6, 1533-1538.

M. S. Kim, J. S. Kim, J. C. Cho, M. Shtein, L. J. Guo and J. Kim, Applied Physics
Letters, 2007, 90.

M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-Evans, M. C.
Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs, N. S. Lewis and H. A.
Atwater, Nature Materials, 9.

M. Al-Ibrahim, H. K. Roth, M. Schroedner, A. Konkin, U. Zhokhavets, G.
Gobsch, P. Scharfi‘ and S. Sensfuss, Organic Electronics, 2005, 6, 65-77.

M. A. Holmes, M. E. Mackay and R. K. Giunta, Journal of Nanoparticle
Research, 2007, 9, 753-763.

H. Sirringhaus, P. J. Brown, R. H. Friend, M. M. Nielsen, K. Bechgaard, B. M.
W. Langeveld-Voss, A. J. H. Spiering, R. A. J. Janssen, E. W. Meijer, P. Herwig
and D. M. de Leeuw, Nature, 1999, 401, 685-688.

J. H. Burroughes, D. D. C. Bradley, A. R. Brown, R. N. Marks, K. Mackay, R. H.
Friend, P. L. Burns and A. B. Holmes, Nature, 1990, 347, 539-541.

H. Sirringhaus, N. Tessler and R. H. Friend, Science, 1998, 280, 1741-1744.

M. E. Mackay, Y. Hong, M. Jeong, S. Hong, T. P. Russell, C. J. Hawker, R.
Vestberg and J. F. Douglas, Langmuir, 2002, 18, 1877-1882.

B. Hatth, B. Van Horn, V. Y. Lee, D. S. Germack, C. P. Gonzales, R. D. Miller
and C. J. Hawker, J. Am. Chem. Soc., 2002, 124, 8653-8660.

153

107. L. H. Lee, J. Appl. Polym. Sci, 1968, 12, 719-&.

108. C. J. Vanoss, R. J. Good and M. K. Chaudhury, Journal of Colloid and Interface
Science, 1986, 111, 378-390.

154

 

"7111111111111113111111 1111) 11111“