VfllLufiaoutitfiislufi.‘ l . . 1 000‘. .0.‘ 0'0... 44 1.4. u. . 1 up 4..“- .u‘ 0 0 .Q q. I... . 6310.00.10... . . (’0 0-. o 0. 000.. . 0.. . . I. 30.2.0000: .40 370...! ‘0...40.0 1 0000001.- Y...4 0..’1 -. g ...00 .00...‘ 4 0000. . u . . . . . ,... 1.10.24. «00. -dl1l.‘1 - . 0* ....00100..00 0\.. . 0- .0 .40 ..0- ’3'0 0.0 0 10 V. 1. Dov-50 ..N!‘ 0-. 308.01%... 0.0.00 1:000...\0.001..0r..10v. :1 . . 1.... ...-......“ :3... ...-u. .....A-,..--.,.-......2..1. ....n..-..A..u....x-..-......_- . . 0 - - 00‘ . u -. c- . x . .x.‘00’.u00-.nl.00 004.000.0000‘400- '4. "04' 00.3. 0%03.1.0 ... 0% 9 .gl'0’00..'.r0 01.0100. 700140 7 0’00 000“ ““00 out. 0. ‘0... .. 4 .. ‘11. 1 I O .0 I 1 .- .o . . £190.01... 0.6.7.3.. . .0020. -. .0: ....r .... 10.0.0. . _ .2034 00..€ . .. . . ...-:4. .‘t 0.2 '.....¢-- . a , . ... (.10. . 9 ... ..40 . «x . I. ... .0. 0.0 v... ..- .u1... 90.9.00 0. .... ‘ 1004.301. .00 30.:‘30. .00.... 0 . .... .300. 000000..0-'..00000.’o 4 . , . . .. . . . 4.. .- 1‘..0...Q~.-0w‘001. ‘38 ,0 «.0300,- o.....’.-"0..'0-0. .00”. .. .01.‘0-..|. .! .240 r o. .T...-..‘.. 0. .. f . 0 3.90.! 4 . 1....1. r , 0 0- .0J.’ 0 .0. 0000 0 .00. .‘004-0 :01". .010 0.40 '3 ...0 ..a .0 0.000.... .8“.14'4 ..000- 10.0.2040. . . . . 0.0.0 0 3001.10.30 b.0-atho .1 .V10.0-‘00.:...04 . - .n.’ .. I. I‘D-10.4. 1‘ 104.. .n. - ...“..I 7m“ O.\ 1‘ .0 .. .. HQ‘. .... .... ....4'0 .001 o. 04.0. ... --00.."::1.\.0bn “I'dlt‘iiflaai- I ‘ 0 Q .00 :3. 4 .00”... 0. 00.112104. 1 ... H.000 ... 0 2:30.30} 0-... .10....500005 . . . .. . A. . . lax-.0 10.1103. 00......009 - 0 00 00. .r .0. 1.... 0.0 $0.30 . 0 0~ . . I ..uo. . .0 1 4.. 0. ...... -. . 0 .0... « ..ua 0.0. 2.1- 0%.“. ’40 “.0 0 ~ .2! 01441.11 . v? C 0-00.“a0 0-“. ll? 00.;‘04 1 00- .u ”4.. .9.0.0 #900 4.000. C‘. 0000001000000’; . A . . . . .- . .0. ..4. “£011.00 :00’0QIWH‘.YQ~0VQ. “.‘ffl.10n0"3fu.00W0,0‘0.'".I'1h001 vrqc'Jnhrsu 0 .“-4..‘ 01... “IV1 4 00.....C0 4 0 .....01LLH..« I...“ 0-600). “00 .\ 0‘...".000.I00n.\n’.” g. C O 00’. . . ... “.0” 1r“ .W «01. - l .. . A . . w . .. . ., . 71.310. .. ........ . , . . . . . ..O .. .. -. ... .. - .. 0‘. . 4d 1.. :mm. . . 900},0.m.400n”n. 4 ”.040urs.0".00u00u10 outfieuouuuwuml .000. ...u I 300.0.0J. ‘0".0n“.u2.-0.w-.4. ..v-o..u4A~u.0u.on . .. . . Q 00.. 0.1. 0‘. :90 90 :0... 01"...00..11:0.0. .0.‘.0-..34JQ.O..0.4‘. 0.00 0.0... .04. .3. .. 1-1 0. 1.0 700......1, .10.”. 1:... .. Fokflu..o.”'....0lfio.0-. 0.0 00’ 0 n . .‘4 000: 5.. “fi 0r 1.3,.)“mmulmgxfi. .0 0.4.0....00000‘350...0".9000.:01”.£”0.0 . 0001‘. .‘1‘00 .0.”— , £0.000rrr.0'0h0¥ .0... 0r30000r0'741 . .- , . . .. ...”.finofi’n‘i4on-V00qqudtflu‘b. 0.... .0000.” .14.. >.0..I.I.A.IUM0400M.V 00.0.0 ‘0.0.0...0.u010- .- -..0\0.W0.0.0.'0 .0.....u.a 0 ....u .”.0'0..Ir ““0 ...-.3000. .“00L1‘N. I-l dronv‘ unpfi.‘ I .Ifl‘vz} “4.00% 0.0 0040.15 0. ‘1‘ I: 0‘40 . I.‘K. Th” .....0. 331003004.00.4.1.un;v.da.3!ln.uotwn5huv00.0 ’Qfid-vogvpz *- .400- . ‘ 0"“..- . , . .. . . .0...v 0:400 .11...0t..vu..fii 4 . ...-r00. 0 'V. ..10 0-: .00..v..‘{ 1. :0... . 0.00003. .0. ..0. . 94.“...2f .. 4 -.. I. g . ... - . .0 . 4._1L. .- 110‘. 000, .00 ‘éhl901' . '3- . ”‘0 . : ##definr-I . t. 0 .13.... 2%.»... ... . .... ...-21.4.3.3... .... 1.1.1.1... ...-...... .... .3... ...... . . . . ..u..u...n..A...o..d.-..1.. .. ..........--. ...... . . ...-.11.. ...... «pun-Aha... ....- ..u 4...”........1hr. 1L .....- ..............uu.........._... Jinan... «...? v.9”... ......» .. .. .-, ...Q .11.... ...u: -...... . ...-...... s 1..- v.o0 ””00 24”.: $10 1200*.- .Pfhfl“..0.t~.\ 3.0.0 “oufli.000§)41h~¢d.~0.0.h00-WM«0 .14... r “0“ 0.00.00‘. . . ... a .0'0 . ..0a 0. a 10.-. 5'0 ..0 1.. 14- ‘00 .70. ......» I... 11"! 30.210.005fl. 4' .A. 4 \ .0.0. :0 I .1 .000. .00. -. a IYUr! .... .320 0’..01' .010“!.. 0.. .. 'fiflwl 0. A. i . 1.. . L01 )w 01.4...» 1 .- fl 0-10.2.0.) {Cu 00...... S 5.00- 98930... {.K.k.0.. 0.... 0.0.0.. 1.023.043.1038.- ..:. . 20:04,. ....0...0 ...-.0 00.0.9- .1 .‘...-0.1..,1..10.lz: 011.08.00.95. . .. 0.. . .A... 040...“...- 0. 1 . .L 2.0:... .09 .10... l Lr 0....-1. ..0. .0 01-0111... \ . A 1. .... .00. ...-v 0.21.01. “I 34“. , - k. .0.th- 41x .. “...-.3. .10... W ...... Egg“,- .1u.§. 07.55...- .. ......n .. 1...... ..0:.a.....-.... 13.223. ..fl.."«.-...ua.$o.v..0.¢. w. n.".nn.finh1nu...nJ1..-1..33.... ”unite-2.2. ..rfiu“... “...-z...“- uuY. ..T-1.......-1...fl-.u.. ...: nun-...“ “1.9.x......o...-:11....,....- ..n. hu “Rainy-sun...- mzh... .. .... .1. 0 .. . ,- 1-4.3.{1 0.00r-0 0 0. I. v on! 00 003 .300 .000 .0 I214£tl3 ‘ g 04,0 .‘0000-0 000 0 .... 10030.00. 0..1 . - . 1a.”! 00 4 .... . .. 0., 401.. ..r . .42.... 00‘0.1J'01 .‘.I 0.‘§... ,0. . 0.0. v.0 .' u 0 1- .u.~’0.. .. 0 30.0 -.. 00. .. .. . 20 0,0000... 8 GIN-....Ili: 0m.."0 0.0.. .1 0 1.04 ‘ .000.‘ ‘ I y . . (1 aft .0- 3....0. A! .0“ .... ...-9.00 ...... 00.0 ..0. 010.0300 .l..0!0.... €..«...-1.:.:0... ...... 3200920000 0. ..4 1. 13.0-00.5! PM. ...0.0..uR 0.1.0.4“... 0.1000120807510110.-. -. ....w 0 ... 0.0.50.5 r. 0... . 0-...0. - . ..0 0- ..1. ... ...... .0.» . 2.. 0.. . . 00.). Q. .. 0 .4 2004.0. Vo.§ .r! 0.00. . s . I 70.1.5.4. 03.. .... :10... 0040-..(1....l..5.10..40000z1.11420 ..0..\04..0-0 3-A0.5£!..0102-: .... .100... .000 .0. A . .. 70- . 0.. _ .1. .. :1... «00 0... .p .20.. . - ... . . . .- . . .. . . . .0. - ,. . .. .. ..- IQ. 1 0. t, 3 . \ o ..J . . .. I I .. . . ‘1 o . . . .. 1. n. .. 00.0.. n 0 .0 -. ... .0300 .4 03.0.0.1 ...... 9.00 0. .00.. ~ 0 : .0 1‘ 0 00. . 9.0. 30 0 . 10 30- .) . 0 . ’l .x lo. 00.0,..- .0: 4.0. 0 00.0 00.0 . . 1 . . 0.00 in .a‘htlgfluflfl 0.... 0.018“ .fluflnnt.tu0.,. 0u2.0..¢.120..-1JH.Lunuflnvu. 113...! JauJJufl: 30”.. o.Mu-.0...010.;.0u.001311.13.. a“. .... ... 08.0. .00 0.. . 0-...1..i..d.s..u ..r... 1.. ...”...r’ou... 24.00.. ...-4.01, .00 24$... . ... .. .0.....00- .. ,.P.r..0.0..0.1......01 ‘90‘0. . 1. 3.... .3...- 40040§u “f .5 .Xh .0 .\A., . .... 0 00”!- 10.00;. h. 1 cl 00%.} ...-050.0. . Vifld ,0». 0|.\~“ “...?!“000 .0 ...: ”010”,.” .54... .fi'fié..”\3.0.a o '. ......“fin” (0-.9' 1.- ..4..o~00r.0440...HP‘LMLWKflU-Urfrfidu!’ rufiroh “011“”. .‘0 '01-. ”.qrafCJIwQJ-htcdaiJWLA.u1.00a0..".‘» .tn-Wrnwvfl-VO.1.0W0h0‘”8.3”...APWJnJ-4WCJNOAU. )JVD'.OQWM0. JUN-000 v lzuofgl .1 r- . In. v V.“ .40. «(~14 ‘. 00041310040... 9.090030...’ ...?! 0. .00..) 0 I? g .0101: . . . - . . J . 0......o ...-040.. n. 0r,‘0.,0..0 .....lx:0.0,.04 ..I, ..I .4‘00 '3090-1- OPT-0” 010. 101-. v 1 '00 .040... or... ’41. . .. 1.0 . .0.- . 80000000T p g. .1 - _ 40.3.- . 0:5"!!! 500% 3.0.”. . .1130. N........1.. 01. .30.: 1.07.41.04. 0 . A .23.... 30.. Isn‘t??? ... ...-20000.0. , .30, , I. u . .0\ . ... fat-...... ... .0.......‘ 8. ...... .09.”.- ..v' 111.0. 9 .000 (.10..- .0..-. 1. ...”. .0'013-150 . 10.1 103.... . 5.00-1.3. . In... .. 000.0 0 I ..0‘ he 2 I . .0. .-\ fl .. ‘ 0 o .0 0.. .0 1-0 1.4.1.3000 Cub-b... ..0.‘?. .04 .0..1’!{H a.....‘.-lo..0l.0u w...‘.".0.. 4.00 0-1... .00 .n}..0.4u..4- 0.v .AXC L0 ... s. '0 ......I 41-00. 1.00.1.0 .00. :0 0"}: .1,I000, 0no0£ v 1.000.... .1” 00.00. . .0. 0.... 0 ... 0n-.w..l4 0. ..09 . C 0.40.0.2 .-a\ 0 .0 1.. .10. mm: d0’.§ndik 0 Ir .‘onv... "fla"!00.0.0‘4"0 cur—ht 9 5’00‘.‘ 0004.0 .01.“...I~.~0_.'. 40W”, “.“unddhowavauoutduo 0.0."000c’u-I0 .....u-a— 0N0. [.mJ-uonqo, p‘ori‘umav on...“ 1- c.3130: 1-1. ; - "- .10....4000... .0000". .. V. 00.00.. ' 031,90, .- -0-.0 . 1|. 1-.... 1 ’00. 000 1. r 3‘04“} ,0.0 1&0‘33 - no. 00000" I‘HV” l. ‘ Q 0L0 Ln! ’0 0'. ‘0 ..H'Jub..;0 L; 01.0 :00...” 1.70040010431 If. .240 .0000 .0 .03 0.. . 0.6.. Y0 .. v.»..’0...0”«0000 300.0 “WW“!H 03.0“” 0. 0000.04,”.Qu. ...-0.0.1n440100130vd .1.“ .tuu....oo..o..lav u"“|..0r£30..:h!.0fip'0anrw ".dehhq'utufiot.’ 0 «“10 . 1.01034'0lnnw 03:0 0..I0...0.....1. 0K.“ 010J:...1. v . 0 3 000.4 .rr-vp’.$.“d..0-l O 0 .- L01. . ni \Hxn- .k1 0 . n .. ..1 . .., ..., ...—0:.1 . . . . . .., A ~ . .. . .. . _. 00 batik? 1'00”!- 100.000'_ "‘00..- ‘nnt- 32"". ‘IVRVV‘IJJCf’UCI, guz‘.000nluuh #wh "0. .1”..’I0‘1‘00¢00\~u00§s ...,0 1"00004-000 015......h . 1,0 .— .0.00’.."104h”...4- 4-000 ..u’o-zrr-w 00.00001...- .0-‘ 100000.. 8-0000}... .1. 00 .1..- .1 I." \0_{.,.loo. .. 0.1"0‘... 00 I . . A .. 9).. p 10. .00. 0. . 003.10.040.00. . 300.000 ’ . 0 o 8140.0 0..., s. 0.. «.0 00.04.7900. 0.31.0.0. 0.. 3p.0.04~..01 .... r3...'..00 ...—1. ... ... 0.! 0.0...s... 04.1 .0, ”00.04.“ .5 0 f, 00p. . -..10 0 .0 1... Iodide .- 4,0” .. to... .203... 0-0 4.0..- .1: , A . . k I i g .3 03 .0... O 0 0 v 0‘.1.000.0' .000.- 03 o .0 090:»0000040. ..0 0.0.0-‘4I-I: ‘ 1“. 0100.0“.‘IIM320 03"“...- ...o....70§...0...r- 0.0”!- 427’ - .0. .050 - .1 . 043‘ .3. I 3.701340%... 0. '0 1-h000 .030. 54.0.0.1. -0..r. . .3... 00.. . , x 0 0.0 000 .cu’ 0 ‘3’." . . 4 ‘00:!3: .00 - 204. 0 000 '0bn‘roou O! 1000 0-\-.? h. ’ .‘. a L“ 0...: g 3.. 130 ...? 9’0 .04. .0‘41 {...-4.0.00 ’0W...00..0I.oo??0 5000. J... 0. . ...}... 4'\ 40000.. 0'0. I. I) . ... .1 00.1.00». ,0..10.f . . . .0 «4‘. L I. E 0000-0-31. ..- 0...0 .4 .1..-lv?.o1-0.0.3.Q10000 000i: 99.31.012.1190. 100.020.33.33 .0: 000.0440: 00.0.13. 0.. - 00.0.0 . 1:... T. 00.0. 1 [.033 11-. 30100.00: ...‘.-.0 10 . d. 0 00.1.... .. 3.. 3.... ......300. .....J....- -. .1. 1:00 ... ... .. . r! u . - . ... . ...- .LI... 00 ..00 .00....p.0 0- 4 4 0.0.73.0 . 00-04- 0-0. 01.8%000’030080 (21-7... 00.0.0.0. 0 x..- .0- €11-.. . .. . . 0.) .0; n .. 0: 009. . .1“; 0. ...- ... "1:12 0... ..uoo .- .. ..w T 010. . 31-0513}. 4%. In. . I. 0 4 . ...-0 4 1 00. m o-Lr. 1.“. .03.“..30’0. ..410 $35.0 i.n“"0...20 . .,.0 .l00lhvdulu4vcficfl-‘.1. .04. 7-3.3.0....941‘h 4 ..."1‘34r3’ . eun‘péu 2.5.43.0. 040.431.9010qu0330 0.0000... “...-4... .0 urn“? ."Y-“atorvwfiuc .J“. “4. 0-..u...H0-02'000" "W 1‘1“. as. 0-4.". 0 ....Iflro’gNQ-fi .I.0. \l.4'0...¢-..oul 1 0 .- .4 § . h 0” 041-.“ «l... .’ r401. 00001.4. ‘0. ‘0..- 0.0.0.- .0 . 0V0”. 0 I: , - .. . . . . p. .0030 .0. 00-0 .0.I.00110:.. 0...!)00Y0x-44 0-0 ,-.00.10 . 00.0.0 00.10... . ...0 . .0. o 0... .0... ‘ 4101'!) . .1 E1 é... . .. 0. 4‘0: .00.. 44890402140143... AH .000. 0. .... ...-9....1:9-:0..00 ... 0.. 0.0.4”) ...... . 1.3.... ..u .0 .. . 49... ..J . 0811......H.0u.J-,00.0 ..0..t (P 1003 390”“... .0 .906... . 0..- ... .3 . ... ...: 3 00.0 . 0.0. 40%;. +41. 0 $0 .4000) .- v . l 1. 0 .4 . .041 .....0- £00000 ._ 1...... 0.03-1%”.K.0.«¢g 0.0-0. ..L .0 ..4 0. 70.. . 00. .. N a . .1 .. . . .. . ...-.0. . . 2001.... {0.342.311 . 1.0.303. .l.0.0. ..1..00...0.40..... :- 0... .. o ,0 .100. 1. . . ”I. 0 v v . w..0..u3..1 3. 00 ...-.9315: 0... . . ..I... .... .- .... w z . ..80000.-.. .10. .u» . 0 . .000 01400 .. 0 ‘01.. .0". 0.....90.‘ 10.. . .t. ., .. 0... 0-. 4.20:). 3.3.01.0; 0. 24...... . . ...-:3 20......00 1.0.3.... 00. 30.3.7.0. .. . .... ..t l . . 0-K 0 0 .0.3¢......9...0:1.00..t. . 4. 0-31 “00. 0-0. 0h...0....0:.l...¢ .4 {um-......A .. .4 . . ... ”1040.1... 0.0. 2 0 u. ... #0.. . e... . .004. 0 0.0.0.}... 0.10:... .-..10... ...-.10 0.900.. iii-ti. .... 001.. ...? {14 -. .0: .. ...-1.0.- l ..10. ... . . , . . . .. .0 . 1. 1. 0- 00.0000 0.. ..wuflciJW014mnoglofl. 000.. ”Hum”! .. ..hrflmu$§.l.0.v?..0u010\0; .. .0 .1.‘.ua.0....u 0001...... ”Mao-:0.“ .1... .fl. ”0”. .30. 0.00 r.n..01..n. .0100 ...... 1.0.99.0. . $3.10.. ......I 99".. .001. 0010.71.- 1....‘. ....0.,..-0.0-. :010 1.07000. 3... .011 : 13.! .300090‘ . . hf ”I. .IV J... .. .210. .. 0 n3. . . 100. Jana-:89... .04.! .01...‘ 0.0.0.0304! 3 :00 .100... ”0.3.10 0- 3.6..- . . 3.0.00.- 0 .... o ....0..- .0 .. . 4...... 040.30 0.72.00.55.- . .....A . .0 0 50.02400... 0.... .. ... . x ...0: ..0 r... -. .0 .0 1 .0. 0 3.0.4.0.} A. 0\Y 1 Ann-N . .....fi ... 2.1.. . ...: .nw, -. 4.2.1....,..u-..-a-1..J ....- .62 ......-...1.... r...-......:.....r-.. , ......1...1..n.......v1. . .u ...-....u.“.a. rah-..."...nfi...gnaw-....... ...-.... ...-...,FJu-umtuu. ..."... 1......1-2... .... gnaw-1.3.... .. . . . . .5... ....- . . . . . . . 0 4. .0 u _ .0 005.0001. .0 . . 4 .a . - . . . . . . . .. ..dd...:i.wu.nnu-.d-J.fl1h.fii. . ...qu-ahaua... 43%....“ ... .e4r1k"...9-...n..-...“10...- fizz-..-... ...u..A-..3... ... :3. 5.1....» .1: .-.-1.1....- -....¢,... ...-.2911: ...; 33:3... r... ..r. . ...-Ax: . 1... . A ...... .31... ...-...... ...-4...”... {...-1.4.3.4....1452111. .. ...-1.11... “...-.414.u.,.«§»..nd.-L........a... .- «2...... ......h. .....,..“.....-...A...-u...........3...» . -3...» ......»ud ...-3mg. .- ........... flaw-...}... ...- ... u... A finch-4......mnt- -... A . . ... A . _.'00.-0,v .0! 405$”.I0000 ' 4 “.00 .ID- 00 .01 00. 00. 0R0 .00001 n o 940.” 0 .0.0 . 00.. . 0.. 0 .9'0, ,1.-.‘!. 0.09.34 I... 0.0 ~ . 0 0 .4 Q .. 0 0. ..A \ 0. ’30; 00. ‘.0.’.‘ ‘0 0 .000 .0....- 0 Q.- .QQ \ . . . . 0. ‘ . . .0 4 01.-.)...100 . .040) ...-0.0400. .fi:0. . 00.. 0.0. ”van-0V k 4 ... 00.0. 0-05 0.10 1- ”“3000300 00.03». 0.. I- 0 clan-0.02m IJ.0.I.5110 . 4.9.03.0 0. to 1:0...1ca0rw? 1.21.... 0. §-.'1..- . 0 1040:- ,0 .040_.-..05 4 -h... . 9.0.0 00.00. .. 1.. u. .0 . . . . _ . 0 l0 . t . .20. 3-! :0. 104.05. .1023»?! 30.04000... .03. .30! 00 10002... £0020... .004. ...-:20 .. . 1.1. 0 _ 11 1). 0.. ... 0..- 234.... ...-0.00.1100!» $0.0. 1.30.32.01.130: .. 0 _. 0.00.0 : on...‘_...A-..... «...-91040 .100. 3... -l1).i_ . . 0 0.4 ”0).“, fl .fl .- 1onh1000uruuwu-(Ju... .34. .0. “I'M-.ufiWI’OJWWLI-n . 40‘“? ififlh‘blfiflc‘h“fiifliflmfiflfl19 tiny, 0L0 m.04 Ju§.P-Q.1~M.“0l.li0.0rum4 ....4..J.Jowmo-:Coul.¢0.mnltn.00.u10.0.00.’0u.1...-uw. .n ...-”...”.nlbs.Lhohun.0uvu..uwm.mofi ”on 0.1.0.14...“ 4170. .0 ‘.*zr.“-..01.!0.03.0. a" .5... . . non-I .. GK I v n . . - - .. A 0.- . 0 0A . . 0 O 0.000.000. . . . .00 ... - 0 . . 0. . .p .A 1 . 1 0 . 0 . . I . . . . — ...0 .0010. 000 .00'00 0.1.0 ..pc.,’tf0 .‘73.3 0:1 Q’s-0 5‘0'..0000000\..00.000I00001h100..0000.'0.b .300... Q 0030.. cuuua‘o‘. 1 s 1 li.‘..0.0u.40 3.1.0. 40021:?! ..1 00.000.00.0 10.0.. 0.0....0000.‘ ..04 .‘0 0 . .00," . ‘01-..I-000‘001. - 70. . ~I .4 .0301 00.0.0000... 30. . 1.00 .0‘] 30" 00 400014 .3031 .00., 40‘0‘00 C. 43-000 . . .000- . 0 04. 00”.. .4. 00. ...-...... 0 n0 0 .0.(10“000.. b0- 0.!1‘1 . 1‘. ..00- . 000,000 1- [:14 ......y .000 0.10,... 0 .90.;0. 0.! $2104.... ...-0.l.)w.04 I, 10 00.0.0 4 .00 ,.£.1-...........2.1.-... 3593...... «haven-....vfiu...u....-...-..fi... --.-nuisanfiaahabzn .3510}... 1... m1...._.~u.4....uu- -..-.2... ... ., ...-....t , ....n...-...--. ...-.523... .....3 -.. . 1 . 2.. J..- .112 5.2.3.11..- - Cit-JPI... «.00... 00.00"“... 091.000) 1.5....100’0! ...!0‘1.-0 5.00:1... . 10V. .4 10-0.-.. .... 00.12.01; .. .. 0.. 0 33.7010. .4- . .030 00‘ \ 0.0, .0. . 0 1.1. 4.. .0. . 0-0. 000.4 I 0 .0 A .1000nfla0 . h it... ...-..- ....10. . ...-0.. .0000 :0 1. 0.0.1.0.... 4 0... 0.1.0... . 0.0, 21.00.... 0...? :0. :0. 97101000.. 00.1.53! 09.63... \. ... .......0...... 015.... 101.. . 00.0.0.0. - 4 cull. 000.. 0— ..vnlA. . .0 0 I: 1:44 0000*... 00.04! .0004. a . . . 9‘... ..0’1‘45 .00. .1 1400.4 '0 \ .0;,.00Q-0~0. ...-1.00.0.0! 00.0.00. ..1 1 0.0 - I"39"0? 30.0 0. . .0!!! . 01 ...: 0.100.". ... . 9. 0 0001. .240. 0.-0l4.914 I..0.Q..0-! 0. 0.4.0. 70.1. . 0 0.. . 02.1.0.1 . 0.1.010. 0.4- 1.v4:- 1-30. .0 .13... .10. 00.2.0121. v 1.2.4.1 :31. 0 . .00.. 0 ... 0. , 0t: 00 0-... 41.0110... ....0‘ .0 0. 0e? 00.0.0040. . 00.. .00.} .0 100 0 ...-.0 .00. . 1.4 .0 '0. 0.. 1.10 9040-4 00-9. .... ..I...p.....0.fiS-.1.80 41.4.9.4. ......- v . . -..... (100,-.10 {1-009-7.0.‘.'~_I'f0 410 00100.. 4., v 0... 0’1 0 71. . .. 0 r0 .0- 'Ac.1 .0052) 1....0- 0... lnr. 0 50 ‘00... . 40.0-0« .0. {0.0.0 ’10-. 4. . 43010.00. . 0 .04‘440. .00 4'01. 01000 00 2'. 05‘. _ ~0- muu. - will. . .05. .. .. . ‘0 .0091.” c o. ‘30“ 0'00'Jm004vfl-4 ..v. -0‘.... .‘ 0 0 .00 4 0‘0. 0.- ' I. . .0 S... . . -0 v .. J 0 00 .0 0.0? ’L0 mluono.00..0ll.010-.. ..0.‘ .. . I . . 0.. .‘Z-\’ 0-..001'J0 .0. 0.00.4“.0’u-000. 0,4 0. . 0‘.000-0l.”0 0-... 00.604.00.00 Liam-1.0.1:...rfir-EruéI... .- nub-......E-J 02.31.03... .3 3.0.... -. ”§.Nr0w.0mgun0.flul0.~ .. ...“ . 0a...."- ‘0. 4“ 0-4000": 09.000“ 4. 0 0” I 000A 00.06“..."{0 ’“th-‘HHOHI 01A 0 . 1 1 000 000.000... .... 0.01.110 . 0;. 00 I .0. 00. 4.0.0011 1.03000 0 1. 0 . . 1. 01.1 1.. - .. . . ...-5:... ...-2...... 9...... L...- -. 123.13.. ......fii-rnu . ...-n. .... ...-1.. :1, .. 1.-..r.............. -......A............u,l. cabana...” a 2.1... .. . ...-.....z. .- . . . 000 0000” .0. ....s.‘”.~0\000 0.0.01.0“ 0... 0:0“.0. 0. . 13.30.00: 01 400.002.01.01...- 0. 41.0.0! '0 .05. ...: 0.“ 0.0.0. ,0‘3‘11-0.~"0\£-I. 0 0- “0... 2 19.. 00. ... 0 000-0.. . A . 0 . 04 ‘ I 0.0..40’190) ‘ 0.0000. ,00.-¢P00.020:. 10.4... .0 0 .0... .60 0- ca .4 (1.0%.“- 0. .n .00 00 0 .0... 'z’aflfldh-oo: ...00 0.4mm.“ Erato”. (.1000? 100-«0..0N..0 .00....” 000 00.1.”01’0 0100‘00'4.0". 0.0 0. u x 1 ‘41 0'. I. 1.10.0.0- .0 . ., ... .0. .. ... 0.4 . 0.... 0 ,.vv00.dv0000:30 3.10.0...0uv‘00. .100 J“ ‘3. 10-10.....ufhfaf3.020.010.0“00‘ . 04 .00020 «0.1.41‘000 ..9u0; , .v. 00.0‘0..9..0 n}. . a 00‘.- I0.--«.- 0. ... 0.0. 0-0..0_..40."1...00..0 ..A.. t .0. 003 0.. .. 31.3.1.1... ... v :01 ...-III]. 0 0.. - 0.0 . . 0 0. ..0 J ‘0; 1 (3.. .724.0.0000..SI $000.. 0 0.0.00492 0.0.0.000\...00 0.... 40—0- . '0.‘ 00 0'. 9QJoa’0100 . ..4 [’0‘ ...-0.! 5.2.0.00 ‘0.}«00.0‘-0uA.05,\.-,‘00 ,-0 .- 3. ...... 1....-. ......a 3.8... .1311...- n00v-r90vu0’00‘....}.‘n‘30’i. 000'. 0' ‘0’(.0l.".""10"0.u...( 041.0?000000. hufl’000n J....00.0o-0\.0 0.40 0.00.3 1310000.“ '00: 4 {DairV‘k toil-#0000! .0 0.0.0.- 07” . . . .0. ...-5’ .0 .0001..‘¢-9 .4. I .0000 0. .0 .0 0 £30.30. 040.004.309.04. 300.000 ..JQSKI‘QOvt... ..throo. ‘01.!1010lE00 0." 40. 0. 00300144 01040 .00 L. (0(0..0110.“n.01.,9.’20.hu0‘0 0 100.4 . 0 ‘0... ‘00-... 0... 00- ...I.'a.01:4.0 'C41038-00‘3 l!!-.~0‘u0?.u1h.00.87 0.4- .0. Anna-004... d‘?-0u3flf.$....lln. 00.0. ...00014 30 ..a- 0010......“qu ”HP-011.. 0 ... .. 1 . Hr . 0070).... .... 0,011.2... 0.00.000 00.4 0.0. 1.... . .. . .91 :00 .0‘3.lt'\4-y.‘o0- 3.0.0.. 3. .. J0- . ’40 ‘0 .0...“ 01000300 ha.“ 9 0 . . ,K 00.1... 000 0."... .0- 5.!" 30000 . .0‘ 30.,30‘0 . . .. -... .. 0330‘. 2-00.... ..4 5-0.0.. 0.000.3’ . .o 1. . ... . ..4 00.0... 500 .,.wv-..I.!w!$-... $0 ... . . ...“...- . 0“.I..0 ‘fi‘ 03 Put" .710 . 00.00.. .. .. .. . ... 1.. .10. 1.000 03......4. . . , . .. ...-100.917; 3.‘\0s.0.- . ~‘h'oula. 0. L40, 1“.- 000! \\ . 00... 0'0.” '34... Igu- \‘ . JV..00.10'000«.40 . 0.0.1. c.8000.07. 144,0010’0‘40 ...) 0.0. .0007; .u‘0.0,'4330 I00. , . , 4 2.5.5.9. 0.... 0- : .. 0 0000,30: 00.... . .0 0100.. ..(l- 0 on... .. 10.20.91.401; 10.. 29.010 03.3.: A . . 09-0010. L-0I-0ru .110 00““... Ilel “00.0 .0.0:000.' g‘: 000. 000.00"...0 ,{00 .410 vuofinvm‘maduhrfi'fts‘lx:0..- ..1130”00“00.30d .H0.."W.I.9’0.-.r'00000§0“0\l"30 00. 0.0)0'm-h0u’00. I...~. ...-0:75:13 .. 0000 .. .. b.u0...K-. 0 .. .. 2V3“! 5, . ... 10.. I... .0. 4.! .....1... .0. L ....r - . p 0”. 0 .. 0000 ‘vunanhth!’ §N0§. I 0.03.9.4... Jauluou-invnrH-h5ip.‘ 7‘ 033010-0040}, l...a..-.o.«“’.§-0."o.0f ”0009*” 3000.000...’ $0100.00! .. 13.5 . . 0w . v.4 00W! .50 ..«mnn 477....0...num\a-.5139-0)-0.01‘001001.00..0. J.- u , .1 v 0. 00.00.. (filo-000134.91 0131.5”1- tad-0‘30 00-10,... ““0 30..|~m . 00.01 0 .uru-iPr-J ‘4 S .. ...1 1000. 03-.wl.l0. ...... (0.0.. .0... . .0..h0F.J~AO . .... . . z ‘- 0‘. 4 1 2 0 u . r . 0 . . . . - . . . 3.0.4.“... M“vv.."nn.0§00§wu .0 hmsvt0...a. {NH-hand. 40.0.”...hunflQ “Ho-u. Hanoi-.900. {“0902 4-0." ...-huf.‘ ’50-'30 .10090m3f400v14310004 0400-... - 9,000...I 84.90. 00.0 (b .-u...hu 0.0.0.00930'f0. ...! 0- .x . 00.0 P . .000 I. .0 . 030.000.... .04.0 ...... 1.0-0.0..100 .. u . ‘40.. .103? .\1§1.90. 0.00.120. .... . . .. .01....‘u‘...’ .0 0 \0 final!" . .0 .Ir..-.00«.0. . .330- .A‘011403831t.’ . .152910. 10.. 03.0 . 0.000. .4000. ......0‘ 0 a. 1.0 0 .. “.0 . . . . a)”. a . .07 300000.- ....0..1’0-0.o.‘0 . . 0 . 7‘- .0 00.05 ‘-‘.4. 0. .000.-. [00000 0- .0007. .0. v‘k—z’v‘lv . 0-‘0... 11.00.000.000 .....3-‘ 00v? 1*. . . ......luv...lll 0‘0... .0. 0-53.00, ’. 0 .l.- ’00., 0 4. ... 0.0.0.0 . .10.0 : 0800.10.1rd 604...). 10-.» .0 .0 . . 0 .. ..0' .963131‘ .0110...¢-2 ...... 05.0 4.00000 .0 .0 ..Ll’n ‘ ..r-hil 0:0 0.00.00 0.040 00-00.\10 O I. . 60-010 " .. .‘x‘ ‘0‘ 0.0.0....- ( .- ...14.004?30. 2...... u 0. 1. 0-00..- . 3 .00....'.00Q 1 .0 a. ...0 luv“. ”Lug-r1“: ...-"0.".01 «C 30 .2180 ...:r 4 .0 0 . . "ugh...“- >0.--.u .05..“- 0. Inns-unmiz; - 0&050-132 “0. 23-0?“3fi.§.r01. I ...?‘rhuv’v-unuwfi. .ouu.v.uoa.!0.n.4o13 ..hvztfifih. .- ’0. ... 0 - ... 0 . .00 I ,. n. ' . .4 .- 1 . n..u.-h.n-...:n...n.-.u.-.n. , u... ....ou "...-1.4“...“ 1....W...u...a!.m.§.n.“mu-fuxun-h- Jinn-rah. .. ...v..-....u..1., 4.... ...nn.15.-......u.1..--.n....-n ...-......huéb. ..1-1......-:~.1 A. .. . . . - 00.00 . ‘3 33.1.0 ll .03. I 40.\ Cd .- 0“ I. 20““ 0-. 000.4: . :0 o v . 00 - 00... 4-Vi. 0... 4...?! 8.039 9.04....00,0.0,‘u..\0. . 0.0.0!" .00.. .00.. 1 . ‘. 200‘! 2.11.0... .0..).....-£1s- ..4......ifu-uur-«00:0£fl!!l? h.‘0.10.l...1...4 0.... .00.. ..I J, .03.... 03-03.... £31.11. 3131......0WM42‘ 0.1“.0. '0. . 05,! _ -1 190,0... :4 . .- .10.-\0.I.I. .! \0...o.'...-. 03:10.1 90.....- 2.0-2.0 - :00... .00. . .101..4. 0.. . ...-.00 . 10”“ 0 . L .0. .1...9h:0r..4 .0. ...-3... 0.0.0 1&0. .0 .StI-e.0..a..0 0-. .3”? ...-0.0 f....-|0..v.l..0tv ...: . .2025... n .- . o0. 0‘ . 1.... . 1 0.0.7 .- 10... 04.1.1- ....fv... 12.00010! 0... 0... . .. f... 0.....- .r! -. .. 0. ... .0. .0: . , , - 1 .. . . . .0000... 80..§i.31 v.1 . 00 I .u 0. 1.030 \- . 1A 0. .1 . . , .nr P.0- . ..uVo-l... 0. 0 . 00. «tr-m 01“. 00-14 “ 3‘01““ ..I-. ...-I0. :0. I. 0 ’0. 3.0.30‘ .000 (0000.00 00"V...000000J.0- 000 \ .- . . . . 0001.00.00 ...00'0 ..0 .. .000... .00. . .w.‘00-..-0.\..I‘.10301. .00 1 7 .D. 0 .0. . ,2 .,0.0. 0.0-. L... . ..‘T. 0- .1: .- 30, . 1 0‘... A ...». ..- . 0 ...! .1. 7.33000 .t. 40. J?! 1.. t.:.‘..?..0.aiT00-...o£....-.0.(..0. . .90.. .00. .90 o ... ‘0. . Huh-40 .1. ...-0.000.. .0- l...-. 0.1.0..» .011... 1.40.” 00.3. 11.90.00.135 ...-.30 :0 0...... .....9: 0.1.... .0 210.110. ... :30... 4.1.0 .1 0 0)....I‘e.n. .0100 0.00... 0030’. i. ..r :00.- ‘0..0370 0.4!0‘..0443000..0.05-1:00\1 0- 001...: 1. ..0....2:....J.. 0200.60.62- 0 . .. so. I“. 05.1.. ...-010.9 . .00 03!. . 0.00.0.4 .A 0-110-.. .0. c. .01... 1.0.03. 0'..- I 1.0 . 0n . - .... 0.. ... 0.. ... 40 9 .0 a .ul. .4004... 0.4.0! {00. .01 .4‘. 00.0.3.8..0-4001‘1: 1- 000-0 .2.00....[0300-1....il,-.'.- £00100}; 1.43.33.01.46 u 3.3.0.- .,..0.. -‘. ....Lu0\.-... 0 ...-A4. ., I. ....-13. . 04 .30- l..!0 . ......‘11- .0 ‘1».101. 20.? .1103. 030.. 00.4070 . . 0,-0? 0 . 10.7.0... . ...-.1. . .. . . . . _ )0 3.. ..Muwuzx.-.um.u.21-11-nfl§. 18-h.dfi.a-1......:...fl.u.z. darn-11.... 911wna..n..fi.§uhhr .znnuk.u...u.-_-.uiz 5.1.»; . ......%1,...-...-0........-.. 4.... 4......“ 3...... yaw-$11.14... u....-....-. I... .....u......v.m.~...-. saw-...: ...-31...». ...-.21.?- ..«.-...u , .41.??- - - ... - . . . . 10...; . | . 44 0 0' 00 0.. 0. .. . . 0.0 _ 0- 1- . . , ' 400 u 1- -0.0 ‘0 0. \4 _ ,. 00.4 ... v.0 10......53001. 3&0... .0 .... 410.....10 1!!3f..\4}-\0-.o:.102.-041.05 0001-06., .0... :10 .1 ...,Wdt...¥n.0.9. ..0. ,- 0... . o . o .0104 . 3.720070 .0 . 0. .. ...-1 10:230-107.... . 0. 40. .430- I} .0003 ‘0 1.0.1.. 0.00.: ...-3009.3. 1 I. . .0 .-.-0.. O. ...-......x. ......,.v........, inn-... 1.311... stun-1.... ._ , . ,. .. 13-4. -tunnurfl-1..uu.,.b...fi-.n5.1....” -7..-........-11Isak-5.1.4.0....-.....m....a.,.-.h -............. 13.3.1.1. ... «a 5.1251,.»- dr... 1.: ...-Av- . .....1............. hfi.m43..0..fi.0|.".. “.004” .V, . .9'4020300 '- 3"... . : . . 3704-0.40..0O1..w«.1l. ...-0000‘. 33:01:30.0: 7. 001-000. ..I. .01 ..... . .Q’I.0.14.QI'10.001,0107. $4. .13.. 00.. 3110.00.90... 0. . 01.00.: . 00.4 0 6-H“. 0,. ~10 0.010.010- .000.0 . 0.. .5330," 200.“.- ‘14 0 . ..anch . 1 u. u u . . ... . . . 0.3 .n . . .0.. v. 0 0.. $9100.. 0 - i . 000.200 . ,0 . ...0..-.-04- In. . - ...-1.. 0 .001 A.. ...0- 1 0 . . . . duo...“ 3600.902»«ofuncn-Aufiauuolnwurwnvu . 0"“1443...’ 0.0.0....00-10...0. . . .0 fiflflh-‘stflt-sup‘innf .vJ-Nhruflwk...“ .H 9’... .. Hon...- V0.30” .40.. 0-. .,1 0:00:10!!! $0.303... 1‘ .. . .-.. L; p I: h.0!."0t. 0.10004 60 0 It? '93,. 0. “to 0410.10.01. '30, 0%,- I _ 01 .H-,.?‘..§. 400.101 000(000 237‘" .1 1o . _ . .. . 3.00....‘0100: -14.. .0 00.00000; 04'... (la ‘ To. .- ...-... .- .4,00.'.0.'0 "00110.€0.’0..-L.§0-0. .000 . .. .00 \c 10.000.0-1100050 .07370100 24000 ..0 ... ... . fig 0. 0- ....4 100.0 0.0. .4‘000000' ..0 gnufludu‘ 0" 00.4003 0. ’3‘. ‘4 1 . , 00-40-00-‘0'0 3".00’ ....0.‘ A v.1.000 .0? .0.‘000"(’30 .3. '70 00 1120‘300 300-" .3. ... . .. , . , 7‘}...- .000 ‘90, 11.0300 ..- :00 a 1" 0. 0 000 ....‘4. . ..‘000 ’0‘ 0.0”}3“- .3505!) 0— .000 0.. 0 00000 00.“ 30".- 0000’0‘04‘. 00 r0070 .4 "01401 .03.. 32'. ‘00,}: 00.00 . .09 ..0. ..0f. 0‘000490..I.. 00. 10.0 . ...9-0 .. .del‘; 0990- . . .0 '. .00 00.0.0. 00.00010 0.r-0140.~ ‘.‘0.‘.01. ‘05 .1..4,l000.. ”304.3900 0. 03‘ .0140. n il’dfi 04 10. .Qfiqa-v‘ .0400. 0 . . ... 04 0 1. 0. ..0. 0.....04-(0‘...4.-I..‘J....vs.'..0o4.. 0 14-....00-737 - -40200'... 0 .01.: '0... .03 I ........011 (....I‘ . . '00-'00. ..0 ’- 10.00.00 - 30...: .. ......1 10.0.\‘ 04‘ ...40 0.30.. .3 .3"; £004 .— v 0 00-0.. ‘00 ”c.‘ I0 _ 50000.. 40.0.. $04000“...44“. 00‘00"“l¢'...0“0‘_0 10r.0A-.30000‘.0.‘.0‘00 I ‘3- Q 0.. .. 0‘40,‘ .0 . .1200.400-0 0. 00 0 .A0 . . . .. . O. 1.. 'u ...-0 . 00,1"001. 1.0.0... .0 449‘. .1, 0,-1.0‘1Q‘l‘!’ 13‘00460‘ xfld0‘lpv‘1004? u3.. 3-. :00” a... 0 84.0.0.” 00. 0 1 o .5005. I .00... . . loo-(3:11:03 0-02.14... I... 1r040000. ...-.30.:00 - t.“‘.- 8’0. I. .,\.,00 .00. 00 - 0|.‘!20 .. . 0400.0 A 00 "\310‘H0.00§,I..590. .. .n0c00r1 0dr.....’uu 00.0.0 3.0“10’4 7 0 ...Izalc-La ...-”001.00.04.00 ‘Jfioua‘. - . , . . l _ u . 0‘0 - 00)- 0. It. O ‘3 I! J \- 400‘ 10-001 0.2.0.. 0.0 ..~ .0. . . 0 . 0- .000 4... .. .00 I 1 0 0 x. J— . . ‘ 0 v .000‘00 0.n‘do¢-n0l¢§f4uito0n.. ‘AHOD‘JIIoJW 0.100.001 (““00 0100.40.00 0 01.100100. - '30..0““an‘0\0‘ .0 L”0".0”0.‘.0‘h00"\‘ 00‘. 0. 300.00.000L.’.00 o. a'.x3..00 4’0Iu0‘10. (0”.an 00.0““...- o 0 0 ’00-’20“ 40.0. ““0040 “’30.. “0.13009. “wfli‘pihdn0hv0pfi ...‘Q‘th‘ ..‘LHI.WMMM.'CW0.‘I0V .5000 ‘ “$0 0., - . - . . 04 0!, i 00100,.0‘.‘.0.0. ..4. ‘. 0. .' o "0.04 1. 0-1. 0 Q‘ . I L h“. 300.) 0130...... a 0.0 11.01. 000.01.10.00...- \!4 ...”! 0.0....‘0... I. '7... 0 lull. . u .- . w xii}? .0000 .0.- ..001- :00. 0.0 .00. . 0.101. . . . 1 1......430. .. Y .0 90-5 ‘0000‘5i ...-... .v.4-.00....- 0000130050.... . 3.9.0.3! . ...; 00.0040 , ...-0. 11”.}... ._ C . .3 11.0.40!!! .. 00h.- o‘~1..0.\.0‘-«. .0 v :- {av-03010.10 . 031.00.! ...t 0. 0 0. .. 010311. 1. 0.110 ... 9.0.0.0.! n: 0. 00‘..- IIVI .0010 0.1100- 0. .V!‘ 1 .4. 0‘..‘...,4 0.0 I 00..u000.0l.u nan." Ylier30$fluUrufl0£ 0 003B“. Calico}: 01”"..0033‘ ~.0\-0’l0-00’0u0' “GOP-unu- 0-n .0... 0\ MYJ..0¢. 070 .000.- 00-06.‘. .000? A.» 0.}..- Q 0.. 0 .’?0 .I . I. 0V.- . 7 .01 7.92:.- 0.‘00‘ I 0 0‘0.“40" 0 .‘0 0"... 50‘00 ’ o ..0- or? . 00. . ‘04».u-"0 ; n .9. .uv . ”0400.." 7‘ .00. "1IL.')1 . 9‘: ...-0.30050’01...000! $030,001.01JO0I- ..1 £0.00. ...-.00.?- .0~.g9.4 0 00300-1... T. . 0.010.310: 0’0 0 50.00 1.0.00.. 30.34.00.015... 0 113’31w’010 .00 ... 3 0000.6. 0. 003. .0. 3.03.903. .2 00n- w. .0. . .1431..- . . . , Q0 0.. .... . . 4004.000?! . .’I ‘0 0.3-... ..--1.0001...» . ..1.I.0\...r2)0...0’ . . r . . . .. 1051. o 0.0. .0. 3!. 01..0..D..0O.Y 0.0 0. ‘0 a 0054.0- .1005",I10.00 . 10- .00” ”0.0304. .3». V0. ...}.in .. “$1.9" ta . “...-00.000.060.- . 0‘4. . - 3....00001. .s‘ingtwi 0-5... Ils..lo..0.0...00v0.0 .. 0.0.71. 1.9.0 11.200..-- . , . . . 1’1 0:00-24:30. I. u . . ...-3,0 00.01090 10.0 \di- 0. 0- . 0.1tv.0.000 0.0 10A- 0‘..‘ T ...... 10.0.0 0...}. 5-0. .- 105.000. .0003 Q 31.01.0003 14.51. 4 v 0. 0000-0-32! ..-...00 0.0. 01.13009. 0.0 ... r 0.....30..101.3’...0 .0 ‘10... $31.10*. 2a.. . . . . . 0.0.1, 0 0-01. 0014 {00‘ . 01-1900 00.0.. 00 2 4.!?.3!0¢0..2.3. $05.90.: iz.‘ll“ .. , 0. 0.0 .1. .0. I .0 0017.1..01'0000u1u00 0. 434-101,}: 403 2020.40031052014-0. 00.1.2.0... 0... .07... . .00 .01.... 2.1.419 .00 I1 0400 1 I. .. .- 0 0 . 10 .. ..0 0 I» 0. ...-000... 900.4010 14.41-00.17 .9 .... ..0...’..’.0..l$070 . .00 .00- 01! 10.0 4 v 00-». :00 .53... 0.1 ..0 1.0.0 28.101w {0011.1093310 40.: .....0 .00200000 21”." 4. . , .0.._l... .' - .3000 .. r :0 - .00. 0 4.... 000.....20? .0 . 0.40:0 v01»..0 - 0 0,010.... 1.00.2000. 0.0?! , 01 ... .440- 110-01. .10?10 .80. 1200-0).I3...|0.1 . 10.. ..0. .- ..: ..0 .. 0-r00‘ 0...:0018n 2....00....00...0 0.00 ., 0 hr...- ,..90. ...-{400.100.10.- ....!00.1.4.30~ 40 .0...000."0 $0.10.. 40. . 00.060.03.03 ‘0‘?)31 3.0.10.0‘"... o. 10-040.430.10. ,C!0.0-. .41.,‘3 3...! .40. .... I 1 ‘10-‘- 00. 0 4401.... 01.. .-’ 44 0 -. 1!}‘Q0 00.10...“ V .3... 0. 3 . .r .00.. 21.00.!- ..-04 $0 131-..- ...03-0. . 3.0.3.. ..13000. 0,.l. '1 0.0.01-- .104 0.100.!”02 t... 10. 000. n. 10? 04 01.0.9 0 0.0 .0 0N0- . .4011. o 0...-.. ..QO.I-.‘.7-4.u0.).:i..0 100.00 .n’ 00. .2030. 34.0.80...0- .0, . .4-1003V001100 I’3’nr -. .00... ..A .I.0!.0.r0§'0.4. .0 0‘0 000.\0 00'. ..0 ..000.’ID;00.0.1 1‘10.u.!0’ 0"00‘ 3%. 040""fi : n . 3 ...4 0.2.0.... 0‘4.....20..N~0.|1Io. - . 0- 0: . 1- .10.00 2 0 40.040...- 0 Vi.‘.\‘l.0‘ . 0.1.0- .105! 40.0 0.0.0000 001000000. 3000!.“001‘. .1 . . .. . 030‘ ‘ ...? .0 .. .- . 0. k000i. . . . , ...§.IJ. - 4.00-0.03 .,..4 100'.-.) 01.4 J .1005” 39.04 .0000- 03.10...-0’100.’i?vf ., 1 “-00.0. - ,.00...lo.!.0 0013.0? .0401...0 80.000. 4”00.4..10S.0! . 3001:0031}: 0.09.0.1. . . . , a. .0000 19.4.0. ...-.....130105225. . . 3.0. . :3 .00.... 10.0.5.4, v . . '. ('00..'.027 .0!- 1'1’003001.0.40V. .00 40511.10 3.310108. .uuuwt1442: 04.... . 0-0- . 0‘0 . .I '0 e r00h‘0“ ~‘.000'\00I3~ .000000103 .003, 70"? 0’ .. .41. : 5.010.013.00000000304312? I . . 5.0.01 0'00. 0440000 006:“. hung}. oil-.00" 00- .. ...-0...... 0500. A 30.1.. .0710 50" . ’00,. ’n-000 000 000. 1. 30.43.000.315. . .0 9.0000! 0.00. ‘0-. . . .l 0.3'000‘0.0‘ 00". ’0:"... .00 I. u o 0 .....I 30-... 101.1 0.0.1.0000 0 .0 0.0.0779. 0.009. 000.. 0‘0 . .14 000 0 .1111 140. 05.0 dia. 00 ...-0.410.030.0-O. ‘.0 400" 00.00.0"000000.04! 0.0"0100’0 '.000.¢a00 . 0 .00 0 0|- -.' o .4... 0 A 4. '0. .01 1.! 00‘ 0 .00: I. 0 . o ... A . ..0.I.00.Ilinl;4.n1ln....s§01 I... 0-, 0-4 "0“.“ 4.0..‘0. ”’0 .10000’03'.’ 0-I0 10,. 011.000.“ ..1? . C. .3.“.t..‘§o‘$'¢013\0...01010040...0...‘0..C!.20..,-C09‘ it‘ll. 0 000.. 0 n ..‘l .0000... - 3.0 . 0... .0030. 0 -..-00.0.0 0.0.0000. 90820-010... 0.00 I. 30.. 1.0.0.1.AW04001... . .0.4.o0-..01.001I1. 33:010-113.. 0.. .01}... .0 . ...! .I.-...}. ' .00.. I... . 3’4 '40 o. .0 -.r\£\.0.r 311:0 01....0a.-0-00.0.36;..,0.0.0000.V4j.H&“804 .010. g.t'..i.1..02001.000 (”0f 0800.1”...J.4..';040- "4.. “unhuvfiuz-JVJUQHMJH "u.- ..VVQJUU. 00.1(0 - .010 .010. .00.: a .. . ‘ c0 . .5 r ‘00 A ... I100‘ 00’ 0 VI 0 I i‘0l00000 0|"0A0‘1.'Ll1'1 0 g . a . . .0. 4. . - 1. . . o w I '1 . . Q0 "000040“. "Hororuv" 11-4.“.N1300VW0000 001.100.“...‘0L0‘. 0.0.” ‘30."...h...‘ 0.00‘03‘00‘133 04000-11‘0050’0000-0; ‘0 1 7 ...1 A 3.043 0.100...I(-,0..2!0 0-10- a 2.00.04. 0 '0... 000 .. o ’9 0,... . .. . 1 . v. 301000.30... . .. . . . . . 0 0“: 4.... . 0 .0 . .0? a... 4 .31. .04. 0.2.55 0.091. I..k..0-0.0.0r.1l10.. 40:000. £0...0.Q0.07..!I:10.....01-40.02 .... 0»0..0r1“.. 3.0:“...- :nl04. 1.. ... .904..- . 10.00.- 0.L.vJ. ..OJJ0H0v90JvKuV-fl. 0-0.. .. .. ... 41.331 . M. €00.00 (“...-0 ”0000-00.!“ . a .J .0 . .10. 1 0. 0.. I W. ’ . 0- , . . 00‘. 0 0 0 00‘. 0 . . 000: ’05. 04 10.0, ’1‘. 0310:000- .. .4 0... ..00A .00, . .\ 3“ 0.. .ID:.14.0...0.0.0 . 0 . .40. 1 1 0 1.. . . 0 . ... . . . . I.“ I . . . . . . - C ““700 ”lotut§.¢r1."n lvqnu‘uflvnq0r’4‘0uvlv’040101 . “(00.00010309I1000 1.0.4L0’I?‘ .000.“ 000.1‘004-0 .0 30100.6... 0.!”015’04’L‘h.0 ia'.,.00 . v.- 000‘. 00'. 30“.". 120.37.40.00. . 0-". '.Jsl.§.l‘.-. .. ..A. 0. 1.. (0070 “0000,09'010410W‘ .50 JP 01.0.0 0‘. CC ,00. 0 0. 1 .3 . 40.. poi-0.1.03.0 0 9-00.30) +010 00.0...‘00’..04000"00.l'4 .00..-0.0.. 4 1.4.0.0. - ulk‘ '00:... 30-00..0-0-.4s4. 0- 0 0.0.- 000.’0'J.400 .0...010h-0.‘ 2” 040.... . . 00 1’ . 00?: n I, . 0.,. I. ...“. .o ..00 400 “-3.00. .03.! 0 c c.0010.o§-0 930. 01.0.1-7-‘0! 02.0.0100 f-SQ ......93..l_ 4-10.... 1.1.0.. .0 ...-4 0033054150.. 0dr ...-4 . . 00181.... c 00.1. 411“: 1 . . I! .. q a” ,Hlu..((\1. 000...... 0.0.0.! 0.001.100 000.4...43...‘ ... 0004.08.01... 7... .0137. .0... A- 0...’ ...-.010 ...-1.393.344.1111 1.01 00 ~ u. 0.00.11, 533.043.”. I 1. V4“ . ._ 1 b . $091..le 1.4.001 09.0110300301: MJQ-A.0-.-2..:-.l§ ...00.l~.3...0\.1:3r.1 2.000.. 330...»). :5: ... ...- !!! .05.... .04....01111. ...: .00 4. .. 1 1.203.»... . ...r $010.0. d. I.- rCIo...u...n.0000 .. .. l. 0.0000. 0 ...»...011-100 ..0 0 (JL, 01001.2..9. 000-.....1 3.0.1.0130... 1...”..10. 1.0.- .....0... . 0000.110... (“00.0 L..- a: 012.00 .0..." 40.33... ...1. 1u.IO...-uv .. . . .. . v4 0. 0 . . “...-[.0 .100. a A? . 00040-6 I" 0.00.... ' 14-03;}.0 ..000,'1.‘0010 00....0Vé’1 \0.‘ 'al‘I‘uc 1‘0; I...lfi , ..w ... 0.11 2‘04004...’ 4 1 0: 0-0. 4 01! ... 1 0 . ..-. 0‘ ‘5' .0". .03...) 00 4.. .0- 100 \01n-.00’.1 1 11..£01040103...I0\.0 004:3)...01-5... .01 .l. . . ..-...00fi.’ i...r£449.-0$00I.801:0 0.0. 000 .W... . ...-{.0}. K. .0; 101... . "0.7.0.0 o. . :2“ 4.10.0.1- 0.0.0 0! 0 9 00 43.07 0.0.. .3 4 30.4 . 0.0.0.01- t.” 3 10.0-00.9. .0004 03.00.314.11... 2:? . ...} I 0.3430100. .r ... 4 o . .4 0-... . . . . . - . 1 02.11.... 0.0100 .0090 0.0;).0 .01 ...90: .0300-000.0.00.?J4 .0400. .0 .0. .I.0.nl-l . 4 1 L0” 10““ .!.K.U¢oul~.u“710.00n“. 1.1 0‘. .0."..||V.-u.00.u¢0‘0.“.onu\0o..-v..... .Yo»1l.0\-.huflh...2.u.-n.2030...! 0.1.01. .... 0- . 1 ... .. - . ,. -. . . . . . . \0317010. .. 0-0 00-0 t 000 - 1 . . . . . . .1 . 1.0”: '4. f: - 1‘. .0 "UHVQ ‘OWOUAI 0&0; .“400.” 0.!“0'0-‘10‘1‘00 .0”... “4‘0 "cl-“.14. ’ 0 ‘0 . fi‘n’JLNO‘.fl‘v,‘O.’.,‘-00'Z ..0000‘0130 4'. .. 4:1 .0. do." 0 0 0h.‘ .‘00': ". 0010......0 cl fig". ..010 0 ... ‘0. 0 00.10.1010 I . . 0 09 :0... ‘00! ’30 . .. . \0 007'... 00 J L 0 0" '00,... 0’00 ‘0 '00; 0 0,0..1. 1“. 0‘ 00.000 0010 0. 0 0 00 .0000; 010.010.0'- 0 0000 4. ...- v 000 .40 0. 00‘... .. 00...10.'000,..\.0.000\‘.000~0 0.0-: I. .ll 0 0 ‘4.-\0.-0.00 ‘00... 00. - ' 4‘00... ...‘00‘40I0’0l 0.0-4-Oflfi .040. 1.0 00...,5‘1 00.10 ..00 . 0'00... 0 00‘. . l h. 0.: 0. Hu‘00 0N0. 0“ L0- 00: .0 0 (0103.301 5019001....031'30000 4 (‘0. it"... .....‘009‘0330 0000.300 1 00.00 510.101.0014: 0.0.01.- . 0 00 >00 40 0o. ’0. 0- . . 1 00 v.00 0- . I .10. 001.0 910. --.-0...... .10 10- 0.0.00 0.0.0.0 . 0 0-.0 ..3 ,. 00 400-0-V00. Cl.‘ 0 0 9’0” - 0.9... 000 .0. 000A 0' .W 0, 0.0.".0m I 3000”” 5000.0 00.01.11-050:- ‘300 05 .0 0.30 0.1030}... 04013.0 90 .. 0 -- 0.04.0000I0001‘ .. ”00 ‘00 ...... .h. ... 0. 1- .10. 0.5.0150 00.0103. 1'... 0 0 0. 0'4 o 1 0 000 000 ....’..0O40.O,|: ...? 030.1?- 000_.. 0 ..‘0 0.. 7".)940. 0 0.. 0. 4.0 7.0000 00 0“ 11:0 0.0 ‘0 0.0.3.9} ..0 fl; 4' 000-". 0 r- 300. 0:30.. ‘20 ”‘0 40 0.70;.3‘0 4-0 00.." swag-00.. 0’. 1410 0!. - I 13:01.0 000.000 00‘ .. ‘0).0'4.‘ IV. '00. L01. 0 I01 0-0..‘..0 0 ,0? .0 SI... .10. .00.: 47.0 0, 0-0 .0 ...-0 . 9.040 00....0 0.0 . 9 0-00.07 1 0: .0 .0? ‘0‘ ..‘0001'0 . VI 0 -)\5.‘-0lv,11"00. 0000...}?! v 0"“10 inh l. .04 .00, 0. \.\ 0‘0}.4'Y.)\|-.-.0440 01 .oa.0~..l’1.’w.o.\c:(0¢ .- 0‘00- I0013..‘O0.00~00' 1.00 0...’000.0 l‘la..v-0><‘10v.0010100$000 lg‘:001 1000-50000? 001.00! .1 O00...91’.o.0:\. L. 00.11.0110.-.v‘ ..01,I. 0.11.0 fog-00.410. O. .400..l00 . 010-1 0;. ‘0, 0001.0 0’0 1 v 0 4 .0 .0.“ h. 0 ...-5.0010300“, 3:00;.0LNL0 ..1...£0...0’.-0I...4.,v.004u., . .300...0000,00 :- 1114000411. 0 .....v . 0 . ... 4 40-... a... u .- 1390 0 0.03-0.11? ’IX 0 0410A .10 001 . 3.4:; C0)... I.§0...a I . I _. ...... 0. \0 0 0.0 -. 0.4.0 .---.40 0 0.1.1.0.: ., . 2’10.- ..00’? ...“??? 1.4 0......03...’ 10.04003 10.01.00.100 0. .051... .0. I0. .u0 n ......000 1 0:11.10 ...:1 000.3-‘4910 . . .. .0.§‘3.0.0§.0.. O 1.8. o. u ’0. 0- 000 .0030. .4 0 . 00 v ... c ... ...-.10.- ...0 10 0.1 .0 .0 0 A 4 ..00‘ . n 3‘023‘.‘ ‘01 ..0-0 .- O 1‘. 3P1. .0. 10.00. . 0.0\ 800.0 . 0 00. .0. ,0 1.0.0. .. 0.4... . . 00:0. 0:. “ii-fix. 1.480" Hull; 00.... 4 .93.?!"413 40.437.040.30 0010:? 1.0.0010 0. 4.404: - 3010.... ...-0000:5111. . .- r'330K0.-a....- 00... ...-Q01. 60 .010. 000. A. .0 - . .0-0-0: ...)0-.(.1020 - 0 . . 1d 0‘)...- .- [...-0. ... .100 00.... . L 01 0-. . 1 v. ..0 0.0110. ‘9... 4 00- 0. . 00.0... ‘00:“. . 0A1~.0.0\l0\0.1fl 4.0-0 03.0.00ufivv. 41".». 0000-"61.:00.0 I. 00 \. (10.01000 ....4. [3.0.1.04- 00053 3}”) ...-00:33:00- .3: .1001.0A..I?'.I:. .0. 0 ......J. .- :- ...... 0. 0.. 11 ...0040104'.1n.u:.\..0.-...100 000.0. 0. “nu-NW...“ .1. .....0..0.030. ...-«0......1J 0 . .u at. .0 .001. . 010.0: 100 ..10“ .10 ...-Nb ”0-0 0Y0...n.’0,. 0...}.w0. 0h... 1.0.02. . 3.0 . l.- n. .1. . .. .. . .0 1 ...r I 000.. 4 7-000“ 100.040...0‘ .4000 ... .‘20. 001 01.00003. 0. . o 00. Z 1. . . .4. _ . .0 . 0- - c . . v 0 n ...-0‘ 0 0 . 1 0.. 0-0 - . 1 . . 400-000 , . yu-Yfluah . 002.06.“.uuunffl0'kfifiuur h..u.$n0.o.w...0ns‘.u.notr0..”000WWHW .Uuuo0..IJcI0H-Xln'~‘fl.l.;:0f.fluhur..00!.01..”.§100.\1....0- 0-. 031 003.01....00“ . 74.00..“09010‘... . uufifIb-.l14’.qhuw.10 8.0 Wan-...ert ...-non ...-......ILILMLCUTYI. 0-,”.0. duo..,.W-VAXL....-. “-0.000"...d0.010100.10.0i- .003W-Ivtuwn-300NNV-3Uo0 1 Viroflhififi 0 0v I . . ‘0 - . - ..00 .. :00", .0 I1 . a ... ‘00.;11 1,0. 00 ..000’000.-l'.- .00. ... - .00.... 0:000":-.-.....1_A100 .0000 01‘ 0.100,. 4 00.1. . :1 1.. 4.004 .‘f ... . .. .0 . .01 . I..- - . ...0.‘ . . . ... 11 . . . . . - .v '5.) .. .' , . o a ~10 0’ . 532.21.10- -.1.-....u....1.41.....or....u.-.!.1.. .10....- ..0...:£1-..ot: tiny. “3.5101. "weir-.4...- 1101- bruins-...“ .n.-.r_u..:1.<:A-.u ..1.......-“_u....z?0..01.~... .. .. . ”...“..xufi ....0-h..6. ....Nu-bn...o.3...1.....-uu-.,1n .....d..u..1..z.rn111hw. E... .1....-a......n--u..... .....u...-.....u...n..A. . 1.. ......uW....-.-.... ......1....00\ ...-.0. . ...-“1.1.... . .303...th ...»11M..9.1s.....u.hv...hq..-_ 0.0.0.. 00- 4700-0001. 00” .0M1-.'“0‘0.4.‘h:unul .0hu00hnohuulnnlp00wwé sown-”unlubrulhfloIN-LJ."0'40-“00001H000J015- ”Jun-0, -..-I ldhvigq. h..'.n00£00¢0¢04 l .000“.0n.z .0'01... “ . 014-00 .' I... - 00 .fi‘.’00..5’00 ....Q’JS0.¢000-0."-:0 0.00 9.. 03.0.0.1" .000 0 0. A. 0000. 1-0110000110: Q 000.0,40—‘0 0‘ o. I. 003(1. 1 u . 0.0é bp.‘A 40". 1‘? 94‘"? 05 Uh a v 1 103- 0 u. 0 4. .- h...- . . ..o. ’1. 040.. 0A . . 0.1.03.0: 11.000. 0 .1. .00. o. .1. .46. 03-22002... .3: 31? 00.4 13.000... 0. ... .-.-10.. . 1 In 03.». 0013-0331090). 1 ...-... .. g. .. I0 ..0 :0 0.0 00 - . . .04 1 0. ... 10.0 0. O .3. . 3 - ..'\ 44 0.000 : 3 0-0.. 04w. .0! ‘00. .0. v 1.0.1. 0. 0.00.0000... 00:11.07? '40 0.0 0 0-.- 0 "a. 01 ..v ......01010 C bah-1.5.1.0492 2.00.00 1:0 ..0 03130.0 .4. 10.0. 101.0.- .J ..0...-.00...‘.1J....000¢ 3 ...‘0...1-...0.1. . 00. 0.4.. . .0 1.00.1... 0.020.001.04 4.. 1 - - ‘0 ...-.- 00.. . ... ... .00310231 fi. 11 10.03 1 1... 0.. ..00.‘!0 0.- - i001.00.!0.600100.0.’0‘0905 l 0055.00.09; 1.100 drum...” “0.4.0.0 o. 0 1. ...... 04 . o. ..1. 1 .0. 0.0.- 0110 5.013.... 0.13.. ..00 00-. .00 .s.‘ 0004.0... 0..0'$‘00M .0... , 10 . . 00.1 0.00.... ...-0.0.0....1-0 7. .0010 0‘0 «10'... I.- .0211 0. 5.0 40.... 0'00 . to 000.044 .4... . -004. 0 10.000 7). 5.5....010900 - , ‘1." .. .n . ....- .A... .p 1. .u. .913. 1.21.1531»: 15.3.1.1. .. ...-...... :2 ...-:12”. ...-...... 1 .2:- .: .1........fl.......-. ......)n- .......-....--....u...........-- .-.... n4“, “..«J.-..11,-sn...-...... aha-2.3.01... ...-.... 0-1.. ......- 9- duh-Lab: A 1. . . . u . ' - 00 o .0- 10 004}. ~ 0.1. 0‘ . I 0.000900100 000 .. ... . 4 0. . - 1 .. , . :0. -. .. . . .0 . . . - I . u — .u.-.....-Mc..-.1......,.. .....uv. .....A...n.-..4.r:..uu.-An. ....a......:...-«-131.....u..-....u{Judah-.... ...n-....A..h:.......:.a.. .... ...-«......- v. .11....-........v-fi.h..-.1..:.?....1.1.2.5. L .....x-1..rn ...“...1..1-.--..-..,-............«...-nu..." r- E...n!--..w1.u-.... flour... ...-1......- . I. , . .. . . 0. .’.LV- .0 . 0. 00.0.0040 . 0 \ 00-0-0 0 1.. . I 000.01 , . 0.. 0 v .. 00‘. 3 .. . I ..0’. . ... n 4 0 u . 4. - . . 0. 1_J.'.0‘0 . 0.005 .000 ----- . ......0..4 ‘0 01.4.0010; 0:07“; $0 000.00? 0 I 00. 00 0... 00:04 .40 030.. .IQI‘. . . k! ’ ubwu t0 0 010 .1"00000010700}n . I .0 ' 10 a I 0- 000 0.1.0 0.0 a .0 ‘- ‘4‘0 0 .10- . 00”. C 1.?0- ‘0 .4 0. 0.0. .0,- 0 .0 ‘10 ‘.09 .00 . . . . . . 1. . .. .0 ,2 0 .3: . 03500100000 0r? '5} ‘1‘ .0 1' .0... .... .1 .. 10 0. .. . I ... 0.10”? 44g101W00’“*-49”0.00“ux“1*m0.0n40.”flu-““0”“... “MYTWI’. J".Ufi0u«"0.oh..f000‘.fl.9Liuuoni-rsuflu .30.... 10”.)..351hi-1" 4h..000 — .1.“ 0... . coal-urn”. 0.0.00.0..." 1". 'vivr ..Lls.“O-H-cwlu01 700.. - “.0 Lin-0.1000 .. 00 . h0...‘r.g.-0‘1 . 11.1.90 ”.004000 ...-9‘0. .01 .0...‘0‘.1 0.. P.~‘.'L0..0 ...-v 1.400 . 0d ’uWVUb . . , 0" I . . . .. -. 0.1 . ’0 ... . . ‘ko .0. 0.0.0 I... . .000: .. . .. . . .- . . 0 0 .. A 0 - _ . . 10‘ A I 40- 0.! . . . ...-.3..- .v .0 n... “dildo-144.7“. .ufiunufirPnN. ...u0u.0n.-0"u..0..n...00.. ... ..LgP-.v0.3¢...9..-A.0.03.300 ... 0.0.?! ...0?’-...... 10. 010.004.. 0 01..., '10-. ...-0.... 0 ... :31 0 .10 5.21.0 - 10.....0. 0.. .0 .3 010 .4 0.37.. - .00130 0.. 0. v -. . . 0 .... \10 . 0c. . .. 0005’ 5.. 00-000,: . A .A.. . .0... ’.000..’f..000 0.0.30.0 ...: .090 . 300,0.11‘00n. -( 1... 010 . . . .000...0100.00...40-0I ,1 ..00020’54.‘0-'100!-ALI . 41. ... 0.05... . 0 '0 «‘00.!0 0 0‘\ .0 , a}! .l '- ’..1~’ ‘ 0 00:... 1 M - 0 'K‘AIO‘I 0' 00.. , 04 110.40.0000‘ 0.0 .01.! .3333-0..n?1.l 0.. ...-00.0... 44... J 310.- 1. 0.. . 0 . .vrgl..§u ...‘1‘ 0. .. 0|NI000.‘ 020 . ..l.‘ (0.0 0‘. 0 300-01070 1.0 .4 41’ 40.0.0 0.00.010. ,0 I- ‘0’. $5. .00.. 4 0 00.020000.» 0-0. $1090h 0‘F.~ . ” L00 00-. t - 00. 0100 2‘0.- 0-0 ..--4:30.100 ..0 1. .0: 4 J4 W 2b . . . ‘01-.0404§§t11.p:0 0.00. 000.0 0410!. .0...‘...030.0 . 1.. 0.0 . fill 10... V. 4010‘ 0 . 10 1. ..Jo 04 . 30.0.1.1". 2.000.. . 2 1.0“; I’ 1 ‘- 05. ...lhrwlu a. .00I.o-0. 10 ‘04... if 000:...00.’.0.410.1.0.:\I. 0000.5- 4. up. 0 09.0 0. . - . .20 ..0‘..'00PA.I. ’0. 04000. .0000.‘ 0-0014. ..aLI. 0.00 ...-0.400.. 4 a ...-00 . 04‘0.0 .0 0. - 100... 0 03.0.0 0.0 1 | .0‘ . .‘0001 a 0.1 ‘.D.‘. 1’00 .0 0 .0 ...u c I 0- . 0003;300:053. ...-...?» (’0 04.0- 0.33.000. .0 0.1.07...) .0. .w . .. 1.10.0.1...’ ‘04,.0l 10-7... 0.0.019\ ‘17,). 0.0.00.0. . ... 0 .IO. .40 n. 0. 51.. 0.00100 \- 0-4 . .110 .0. 1.5.0 ‘0 ..f'0‘.0_.. ..l .0 00 1 0 Q ...-’0. .0 0" . .. . . . 0‘10‘00370 Y..~0""\0.10 ... .-.. . . - ..100-0 .0 1 03.0.40 ‘0 01.400 .10. .0. 0. .- $0.‘0.000.'.00 ..0 00 000...; ...07 I“ ..‘1'~.‘00‘$.o 000 ‘01.."00-0’00.\’ ou'....0£§§§00 I :70 000400130 . - .. o .0 40. 00’s , . ..0 01.00 . . . ~1..l.r....3. .o ...-0 10: ...-.....00 ...-7170. 1.20 o. . - 4-1". 90-. ....Io. .3. ...-30.00.. . . \ .10 \0 3A 5.00 40-0 D0 00 0 \(‘1‘.‘§0 0- o 0. but" 0'00! 31.000nvn...l.00(03'01 3. . . 00"}... 1 .0. .00....‘3 0.0.0- 40 - ...4 ..A.VA.0 1 1 0. .4-. n- 04 04. ’00!!!...0..444..11 0.. 33.00.0000‘092000-0 15. 2’00 00.0100. 0.0500 1\0\0 00.0“. 0, .00- 0101 0 ...-0 -_ . 0. . .00 lid-“00.0.. .u0 '00 . .0. .01.. I. .10 . 20.10-00 . ... . . . [0050.1031- ..0010000’ 3' L0 1.00.1 $0.13 .00 cc... . ...-1.01.181. .‘I 0 00\0 {0.11.0.0 . 0-10 0010.701. .04. 0 0- 0. 6.0- . .’0. .4 00 K0 0| K013010110 - . . 1-,. ... .7 11.. 0n .41... .0 .0310 lily-.111” .. .. ...). . .0101. I. ... ... fi.1...0-.0....1.}0.!.. t.- .10 .. I. .21....) 0| 0 . 00..,.1000.. r... . 0.. 0204-0... .0 I.- V... . 1 0.. . ..01 3-010. ...! .0 1.....312000 0." 50.00. 0 P... ..0- (.... 00:400. 000 1. 0 410.3011. 4.. .60! «HEN 0|. v u.... 0“» ..0 “H10. 10...) tinhlhfihrann 00 .0 - ...) ...-.1504- . . 7.003.. 1- 711 I A . 0:130; 1.1.1.1..-. ..0 14.018- ...001.. .03... . . .. 0..."! 40 014. $1.00: 111. .. 0’0 0 v 00. 20.. 105-. ....0 0.. ‘3‘. 0.0. 0154. 0 .030 00.3. . - .lnlu 002‘. “3.01000 0.- 0.. O‘ln-H-Ni’a it. .0 . .43... 1-1. .2... 00-. £70.10- - .1. . i 000.00.900.00..- . .10. 1 .... 0-0. .. 1. 0... ...; 21. ... ... ... ...-250.130....zu000 1 03-1.. .0110”... 0. 03.00A..00.~.1 1- . .00....34000-5 0. . 04......090 90.0% 03“. 00. . L 0:000 ' 1'00.“ .0 0 3" 0 00"»: 4.00. ...-.4 010000; 0' - 1000‘? 0.4- 0.1... .... 00.00.0000,. 0. 00.00000: 2.»- U. ..0.‘ 0. ... 10.30.. 1.. A 1,0 40.0. 10.000. 00040... 01 .. .‘9'100‘300‘ 00.- 209700 0.00.9 ll.- ...10 9 11-0. 00.!!‘0-00‘0 A... 4.00. lit-‘0 .0 0 .10.h'000.0. 01..» “~00 1.110.. .lra‘ooi .‘04..floV 1.00.00.01.13004. . \luizi 000 0 .2 .. .-.0 1001.301 0 0300.00.00.10.’ ..lS‘I00-:..0. . ....0.\-0-1..0y-0. .1. .0 300-0. .0019.00.0.!.. [00010.4 -0 0.30. - ~‘0I! 590.1000. 190-000-9073;. .0- 0‘0 001-;.....w0.0.0 . .4 01.. nun-.flnhuaOU. ...... .300.0u.0.:00. . 0- ...7f.0.-1.40. .0300. 4 ..0. .1.- .. 30:12.1. 031.014.040.111... 01...“. .2uM‘u-xluhflu‘0sushvzqqr. 0 .00 s .1 o o ...-00..\ ......Afloufloi. .n..00u.1\V‘000..-.. ...“?0013 :00... VJ. . Q... . \ ,0 .00 i” ’5 - - . 00 0 00 a 0 .0 0 o. .0 ‘0 0 . «- ... 0 0 . 01".- ..-‘.O-D r10 00‘00.0;.1.001..h-\10".0n6f.0.|000 000.! .1... ‘09..“"0’0? ”.10.. ....ufid.fl0vu.ful.400., 0‘. . 0‘04 0 0080 ll .0u0‘.‘ 0. 0 ‘0 0009040..0.O.00 , to. v. .. .I-‘o 00. - 19. 10’. 0-00 In A... .- 000..V00‘0Iu'..1_0 0. r .- .“0.dp 030‘ . A: 00:1 0. ”uni-00’ . 100 00. . “.Q0..0~l0’. 010.0,. .0A...,00f..4... .. 0 . 000 0001 «05.?50...’ ’04 00‘2 0 7100.0 0. 1. 0.30 ..§ .— , . 00,... . .1 01-010 00 0.0.0 1011 .v . 0.0 ‘00. 0 ...-0. Nb?!- 0- .. 000’. a”.. .0 0000' . -401. 0.1.4.03. 50 . .000 , run,- «000 . .- .0 .I . 1 c 0-0 0 9 .1 . ._ 1 .... .000 0- . v00 . .1... ...-A .0 10.10901 00.0019. 0. (I. 00.0”4-1fi . .flnzuvtnuxydl. ”1000.04-31 . . A. . . .0011-- ..31u I... (- .0. sink-0.0 4.0-0 .71 .0900: 11.00.! '0..$0-.0‘01..0.0..0.‘-. 0 0- {...-3.0.010. 0.»... 00 0.000 1010011100 0.41..." I"- .00‘0‘0 6-0. .0 0’ oi:.0...1..0 . 1 .. 0.0.0. 1-. 0.....000‘00-su 0110051.. I00I40001 ... 0......101-0) 0..“ , 00 0 3010.00... 0”,. ..0..-Q.0000 0.0 .1600- ) ...-10.10.. 0.00003... 8 100 '0 Ln”?! “0.06" 004000.! 1? -. . . .. - . .. . . ., . .. . 001-..3004 .040 0-..‘000‘, . .0 004.0.) 9 . . ..SS1: .090..0 ...-0.11.09 0,... ...0.‘0 000.0 .0}. 0100. 0.0.0.010 1 .4) .m. wiuqr'qunu . . . ... 00.0. 43.00 4... ... .. . . . . ... ...-0 3.1010. 90000000....2 7 . 4 .o).l0.000000';0 4 100000-041. '. t.- 2074 ’00. ...-V . ...-.00 v00 01.....- .00 o 2. 050.1 1. 0‘-‘ .0 0 1 1- .0.0 1 I.“ PVOL“? 10.0000. 00- . , . .. 400.041t ..a ... I. 3...: .0 ...«0. ..0.. ...-01.301.901 0. . . . 1,400‘\.0‘ 0.040.440 ....4 0.. ”‘0 0 ..... 0.100.- 44 ..12000 00.0. 0 4 .0- .. d‘. ' u 1 040110. 30 . 0.0 0.0 0. ..0‘0 "00.0-.Ip. ...-”“011 00-. .7941. ‘0 b.0fi010.0.0‘0 0.01. . 0- s. 410'. 3‘0.’- 00 '-10)¢000.‘.ufu..00.$0.00 . ..0- ‘9‘ A .. “”0000 “It“! , . .0”...0..0. .3000.“ 0“ 00.0- J...‘...n70 4...... .00. 0-4 0-0 0. s ”0. Mun—00‘000Mw0 ud-lwsflVlm-nc000' h 0.0. ..0 . 0. ..olhowhu. 01 . (A. . .14 . .1000; 01.7.0000 . 0, I..| 0.1-0.0. ...-.1 0.01.0.1 1 I 1 , . . I .A. 1 4 o .- 0 ,. -o$u000hl0lfl.014..-I0I.dumd.4 30. “ouhh/fhkoflo00‘tocr N1..- 0-0-00q09... 0.31.33.034hmh... 0.0 00040.- 00-0VO.-.. ...“JI-t ..40.0,-..0 ....41- .0 cur .- 0 ...! 0 o 0 A. .0... ...-.40 .l.‘...._.0.0.u~......0 0.. 1- 0.‘ n "-0101 0 up I030. .0 "$100.00 .....0’0...0 ..00 0.0.0\00.U.1. . ... 0 n1. 0 .-0 0 . . .0, . . . . 4 1 . 4 . 1. 0 0. . ~.- . .0..\ .I 0. 4 . 00 100 . V. ’0 0 4 ... .9917. . I o 4 0 0 00 0 u 9 000., .‘09. . 4 0 . . . .3}0l.l -004l.0rfi ”Pfiwowdflwnku10-Llcktm-l\ 0'0. , . 000.4 00 .1 0. hdoufl.uh0-lo.ui.yr 0?" .‘101 ° “\QHW‘W0 40, .0-, h... A. 0.0 0' 0.01.0 '000000 14- .1‘7 000‘ .. . ~ 0.0. ..0 l r 1100‘. 0 0 .80 “0 . 0-.‘0'100-0 . . 1 . 0 .. 0.0 .... 0 . 0". mafiuuroit ..aihnxicfi ...o 040. $1.211 . .40.. 4 . . . ....‘..4':..-0...All.02a4.f« ......3. .VI-i. . 3... 3.51.810 .- .0 .0- .:o..." .0. 30le 2.10 >130 0. 0 . .. 0.....7-0cu ...-VHS.“ 0.. -'0.0...91.1.1~1.0.0 0 20:301.: 0.000.055! .0 .m? 4‘. 3011.00.0'91p4008. |.0000..000...)14. . . . - . 3.1003 ......00. L. 10.00.... ..r .. 7. . ..s..-.00. q. - . 00 .An .0“. 3....- .J- 4... ....J- . 0 0091.011}. 1 . 1 .1 ..0... 0.....10 00 - «I. .0- 0100’000u £10.03 .9 JNAIOr ...-01" 00 5’0}? 00'?“ bil‘i 1 010.0.05000 .0 1.. .04. C. 0 anvil-0......00-037-01. r..0. .7 .. .40.... . . ' ‘rt: 0- '4‘". 0 . '.\ \M H00 0 “‘00“ I 00 .0 0. ‘00 1 00 u 0... 0. r. 0‘...0\1 . .00.. u .0 . . 1 . 1 0.00 ..‘0' 010370...) .0-011. . n 0000 00 - 40.1 . . 0.... ..0 . . .- . . . . .... .1. .l. .- ..,lu'00-00.“ con..ah.1.01|.1- ‘00 tun-10.10””! uh’bl. .... “4.50. . “cure-.... L001... ..JV.\ .0 .. “...-:3. 0- . Is .0... . 0:01.010. .. ... 0.00.}. . l ...-0.1.0.2-...00 .. . o. ..00- 0 90.0.9 1.....41 I 01000....02 . 01001 7.00- ..0 0 0 u,010..0'.l0000.0 . . v '00:“.‘111H-00 0.7 110.0 ._ .0 .90.. .' ((0 ..v. . ...-0' .00, 00.0 .- .J..b,000$0.. .- :2. 1000 0 0-... 000 .1. . .3 ....‘0 . .‘ :80- 1 0‘ 03-. 0.0.1:. - . t . . .... s 1, . . \fl‘0. .00.- ... v, 1. . .Yv-o ...-)0. .J- 00 )0. . 1.0 .. .. . . o .0. 40 3'..-” 0.. 0.. a, 1.1 1. .00..v 000-010. . 1.0.1 2;- 040.100.10.10. . )0. .04 00 700- 4-0. . 1.....0... 010.9,.01..00 .1 :00... . .... 00’ 4 00 700‘- \0 .l- . .1 “40.0 0 I... .0 .1 0 .0 000.. 0 0.0! 0. 401 0 0000004...‘ 0 ~10..-1 v 0, 01-0.0 9 .. 20..“ 0”...0 C. v 0‘. . M. . .. . 1.. ~ . 140.. 0 000.10! ‘1 il! 00 . ..0....000. £40. .0; 0.3.0.. .. .0 . .. ... .....0. 3.104. .010. . 0-. .. ;. . 320.09.... 00.2.10 0 4100. . .30.. 0 0-00.! .- -. ... 0 .... .. c 0011.... L, o . 0 0..?|-b l . .0 ... 2.0 - 0. 530 .... 11.. 04.0.0; . ... .1 4. o . -1 .0 030.. m..- 0i000Vl0‘) 000... 0010.0. . .101...- . 00.0.0.3! 013.300... . €4.04. w)! .. ...- t 0.. 0.01 . o0. 6.0.0.1110 . ...0...0Ac-.....1-0 .1...- ....v. I .0 ..1-0’10. .... .0430 ,c 1?. . . . . . , . ..01 . 0-. 91:732. 4 0.1.}..- 0. .1. .0. . no. . ...- lJ.00-0.0.. .4701- 0050.0 .0 .0110- 1“. 0’00 ...-h.“ ... . ...... ..1" ...-1.....- i. .1. 1.2.1.159- ...... ...-e: -..-2..-...- . .-.. .. ... ...-.31., .. ,..... ...-.u. . , . .-..uoanLMHW-a... - ....a...«...,..uu...._..1...... .. . .. ...? ...: ...... raw-...... ...... .... nun-......vznuu-nq. .d.¢..n.n..-......R-...-1. ...-1N ... .u1...-.-.-3... 1 .. . . . . .1 ,. . . . '02- 1 .310 I ,1. 1.00.000. 10.0 00.0. I. - .. . . 1,..0; ... u 0 1...... . . .. .. . . . 0 . . ... , VOA-:1 0. ...}... v . .0}: ...-0 .o p.000. $5.0... 0 0” 300...... ...I.’ 00.“...0. 0MP}, - .. ... . . 1A9 .. 12. 0..-... . .. ...-I 00-00.... 1a....60: . .0. . ... . 0c. 0 11...... ha. 0.10.... 0 . . . 0 . .1. . ... 0:..30 301.... 0-...00?‘ 0 0000.0...‘.0-4.0..04000 0.. 0.0.114. 0 3.00“. 00p.l0 0.0“!“0400. .‘ 0-0 I! “fin-.10 O‘HfLQ-n‘ucul 01M ‘10mOO-H4Wrn'u010’” ,.0fi0h.0o\“h.01||0“ 0.4. “.0" T0“ .. 0.0.9 in“. "4zO-J!.JN0~.10 ....‘0. . ‘01. .o :0. ‘0'. od.“~\ v.0 ..‘0‘I'00430. . 14 “00r.04..‘2l. . . . . .11 u.- 0"- .0 0rmror0r0“! .00 00.0.“0-“1'...,0 1 “HGKAUQ .0.“ I ... 0.," 90:00 AMOI-‘vbNC‘OOc‘. . . . u 3“...“ J ... 1 . w . .A .01. 0 . 0 - ‘ .00 4 000- 0 0,..- .10 00 0 . ...-.0114‘0030011... . 1 .1 o. . . .. 0.. 0. 0.. . .. . . , .... .... 1 ..- -. _ A . I 1 . .1 w. uni...” ..uufi3fl.n flu..1...-.-....n:...4(- 3- ".139.- ..ufl. . urn-... ...-...... {r11 ...-.... ...-1.: -..? : ......1-..-.....L.. 2 . ... .1119...- ...... trig... ...-3.41%. . .. . . . .- 1.15.... :1........ .... «...... ...... . .33....- xn,..-.1:...u1-. dun. 13-0433... . In . ‘00 . 0 , . . . . . . . . ., . - .0. 00 v. 004-0040,. 3.0... 1 .u....0.- .1011 00, 0.0 -\ 8-. . . - ... . . . . . . .1, .0 0. 0 Q .0 1 L... 1 0 . 00 . - .3 - 0-0 0 .00. .0’003- 0 0' 00 0‘ 01.0.?)H300 4 .”...1001nm-0LI10101fl 0 ELF-110.0 VTc’."H0Mo1P0.1..~\00 .l:0.00.N00-'104.01.004.90u..u"0 “010.20.? “Mu-000... 00 . ..A 0o ...-:1. 0. ...0-004 .0"... 0.0.0. 4 00... 00 .1. .... . . . . . I 1 0- . 1 ..0 V . 'ul 2". . 09‘.'. 13,0.0‘0 0 . [0.10-1.0 0.00.01QV‘. 9100-0 .0 . v. 01 c t . 0.0. 4. 00. u 0 1.0.3.8. - (I!) ..00-5...1.....4 0.. .. 0100-40-1 ....l-....!, .-.. -..- . 25-00(3011 . .. . . . . 1.... CI 001.. 0,... 1.04.0...- 0 0.0019 1 1.. ... ... - .. 0 "0001 0.000% ...-$0.0 100(0ko .010 . . . . .. £1...1?100.§ 0......t‘ 0.1.. 00 . 010.1.- ...4.. .,'19-.L.,.!11...'§ 0 . 20.. . ... . . . - ..o .0 ...0 $0000. to . 0 0.4 :- n .10 I01 . 0 0‘ $0000.. 40 04.. 0-01..- . .. 00000 D .. . . 0-‘1000 . ...-0410.0-1‘. ...-n ..01.0 I 1.0v0 0 4 o 0 .- 0,4. .31.. 19091.... ..0.90.-.-. lt".0!l.‘4.0.21 . 03,“, . .u. ...... 11.0...- .31.- 10....-.0-3..\. I. 500.00.90.04.- .‘IQ . .0 00.0201100170; .¢-.0"\.‘9 T 0.0.x“! “-070 . . -0..3.L:A.. . .....0. L .. ‘$.v..30.1 '- .A 1. 0.0 1 11.0 .0. h. 40.00.1010.1105\13010'4‘bt r.» 0“. '0 . . . . .1 .1 .91004.0..-.7..: 0-0.1..‘9 ...-.01“. . ... . 00.1.. 000. - .0 0.01 401...! . 1 0 3039.00 40430000... 0 000 ”“01 . $111000. .4 .0109- ..\0.-o..0-00\ 0.0.1.0 . . .. .. . . . . . . . L000 .0... .00.! 0.4. '0000-00. ..1- .00 001,04V‘000i0tl50 9,0.” 9‘. .- .. 00‘ .0 140.04. - ...-'1»? o. . . . .. ... . . . . , . . . .4 I -..-400.1121. . 0 . - 0. . \o‘r0 ...-120.30.".‘20 0- 0.7.1.41 400 1.5.010 0' 2. 5-0:! .. .000 ...r. . ...-0 . 00 01.0.9. . .- . 10.. 00,1 0. It. .51..’!I.. 0' 0 a. \1 00.4.4. .‘ 00 ..0 1.. 0-0 0r 10!..‘0 0.: 03‘13 . . 10-0-0 01...... ‘1‘.0.-... 0 0 .' 0. 1. 03.0 00. 0 :1... 10 $0.310“.-. . .,d(. ‘. 0.1.1 ..*.0’0. , ”a- .0‘0vl 4.. 1. 0011.00. :0. .000.‘10A0 1 1 ‘10-\1\ '0 . 0 ..00400710 .0 £4“ 0“.0'\ 0 u 000 0.0- 0.0 ‘...‘0.0 0-0.. 0 !-. 3‘ I 0.0!... 0-0.1,I.0 .0 2.000.300 4.4 1. 0.4 ...-0 0.00.0. .04 0000 .. ...).000011‘1.1.120. 00:00.01. .. . ’0- «40.1\00.Dn . 2021.. . 0.... . 0. 00-1 10.500 0.04 - x 30 o 1 .... 000 1 1 5:0, 0(- 0.01.00: 0 01. w 0.0 2". 3-043 . . 004000 I." L. 04 1.. .0.- 0.40003 10.301010.— 01 0.0. 0.0. 0.4 0 4. 0109.4. 7 9000.119 '00 4 . 0 011’.“ 000.. ‘41,. ..4 000 .0.009’ -..-"I. . . ’0‘i0r4 . 21.13.10. 4.) .- ....J. .0.i.o . .. t 0. .0. 1-. 00 0.1.0. 01.~110.1.0.|.100..D004 . ‘¥|’).I1 ... ...40 \’.1’.000000.0. 7.900- ‘0. . ;. ... -.0 0. 4.. 010.03... 0 5700.1. 11400000110! 0.23.01} '0 0‘0 . 110041.37 03000.40 00.. TV... 01300100.. . .0. 007 10.7.4~-1... .111 0. 00-0.."000' .1. 4 40.000..000‘1..'.n.109000’..'|0.".00>00’§0; 0 . Llc .- It. . V. 01 I aw-A.'0:..c 00. 4- t. 0.4., . 0- 01 00 0. s 0.9 0-0- .. r0 Ooh-IH‘ 1‘ 0...... (000. ‘0 '20" .0‘5.‘ pd‘? .0 .0... \- CISI 1 .0 .4. . Or. 9 \- C-I‘I. .. u . 0 -14.......0. 1110. .0 . .- ....-1- x0- 0 00.000.‘\w.10.0n¢.. 1,201090311'0 0.0. 10“...“0404 ‘S-00Vti§\-\l$ ‘ 4nrlu10.I.. no. , . 0 04000“ 00.4- J Z0100) 10010.0- 00'. 0‘; . . . 0\}III.~0.02) .u.... . )1. ..!0L.4..\~ .00. - 000 01V. .0. .510‘03040 ., 3404.00‘04’ 010. 5’. . .000 At 000000010- .50 0.00 .. .40. 3.40. 0 .00.IIL.01 c . . . 10000 . . .3 . . .... .. Y .‘.‘I.-. . l . 1: . A‘T . 0 1. . . .....‘000-00 ‘01-‘0 00101000049000.120‘0000.’ 00.00007\ 94‘000.) I. 00 - 1-91-3100 100.01.000.31 0350.03510010009!401...03.00.107.30..£112in‘lx .. “ 1 .c 4! .. 1|... 0 11 LII-1:0 .030 ... 10 11-..I0 0 . .0 0 . 40030. 01000 2003!.00 1- 0. 39% ’01,"...0. 00... a 10110.3.“03230000 0.0. 0.110... ...-0. 0.0. .0 00.0.. :0 11040.0. .. .0.0.000.’ 1 0.030.010‘4‘31 00 VI- 001. A .00 00 00. 0‘00 . . .. . 11- I 01.10. 0\0... \. 00 ‘ ..0 ‘. .O‘SC).!'0'! 0.0 ‘040 1 0‘1 0 10- ‘»413" :1 .00Il00. A 1.. 9-00.04... 0000. 0 0 «(0:4 ‘400‘1‘1- .-'.I00...0. ..‘0 :- 0010’10...¢-.0.0.000.0- . .0 01;. 0.0.0. - 0. ...d...... 0 731.. 00.4.0... ...; . 000 o I..04..0.- 00.10 . 1 . a 0 0 I‘h1.41 0.- .I0 int”? 0 nu‘of 0.-001.. ’“00’QV’VIH. r“ 62,1. I0. ‘00. . .0, .7. ‘.$.I. ‘0'500 0”0 ..A 70.. ”00‘. 0.009.!- -...‘ 1H," 10 Ptuuh. “4.00" .0Hw‘..4’.-J0l.hlohnr.fld nu. .HRuH )0h.‘000“"0.0h0..04h004n.h0 4n.2..o..s..r.0‘7000h00.u.“".03-04u. . ...-00.00000: 11.00. 0.4... 0' 0 1-. 1000.0. ...-“4 1 10. N 0 .u! o 30..- . 0 00 0‘. 4 ..‘V .. ‘0 ‘ P-‘Ib -‘0.0.) 00 .310 1-. ..010031010. . ' ”u 000 III1-AerI40 ...: 04.4 a... .. .010.- .o.‘l- 2-0 .0...‘ 0.0.0.0 . . - .1..- . . ,. .0 . ...... ..l .... .30 .... 0.. - .. 0.3.110 0 l . ...... a 10.. 0. . I090...- . 1 0'1 . 0.1 000 0090.300019- 00 01.0000 . 0. 5.5 .01 1 Album...“- .-..u1|.u. ...-040..”0h900n.0..~.r~0r0 4M..1”...:. 0.11m.- MHW...00.. luv.0-0.-O’H.0100."00hvudn00u40uraffb.n,...?l1 00k J:.’\0H.Jn0... "000000.00: 21... ”.0... 0 1000.. “.0001 000.00. 1.1.3.0. (0-....\. 0 .-.-1:0 ...t n 00.00 0.0 1'. 00.1... 1 9010 - .0 I. .0... 130‘,- 0...0l1 .0 .- j. ..I. 0 . 0.".- 00300 sz£0500¢001 3-,... ”04;! ‘. ”V0 0. 0. 1.. . .0. .01 1. . V .0 :0: 1. . 1.. '1 1.11, :10 0.. (30,050101 o1..4.o\.100...-. 0.2.0.100, . 0-, 03....0110..v..0...0. . - . . 01.4-01.1...319034 o 0. 1'! v y 0 V... a I. .000 179120.000“.- . 0 -. 00. \. 000.0 ..0. . 10.0.0- 309... .. .IiQu90 A \u. 010.1100 .0 Q. fi r . 0. . .A . . ..1 n1. .. x. 100 ..1. ......04. Y 10-1‘.1..0..-..0-!0010-0..1h.4 7- 10-: 0.000.15\L!. .k03000...0,.101020 0.00-001“. ‘3I.0.I’ 03.000030 ..0 . . .- v 1| .0 .. . I.-. 1-... ...0 . .0.I—.004 1. 0.0000..1'{0.ll'0I-O.. -..!) a .1‘1 1 0- I 1 . - . . - 0 . \\1 .0 1A 1 . 0 0 , I 1 . I 1 - 0 . 0 O ‘1‘ . 1 . 0 . 1 0 I 10 0. ... 0‘ J: 0.0.00.0..0..1 ‘04.." Val 1000. .Yu... .1 1.0.0....4000-1'. 00-00%. V 01 0.00. 0 < 000.0000 3. N . 0 I . o . . . .00 11 .. 0 . 0 0 0 0 . .0 . I o 0 . 1 1 . . .0 . 00.001, 00-...01u ..01'1..J_.. 3.400.01Ww00J. 101. 2.0“"- «...-.13... . 350.40.... .0. 1. 0| .. 1.0.20.0... . .100001i00. 11408.000‘0 0.00.1710: ..‘i00'h:.\ 10.0140 001 {1001000. 1 . I! 1 1 .- O . .. - . . . . .l - .0 . I . .00-.-.4.0 0.0.|0.100..,0.. 0¥§100101‘04 .-.IJ-“.0lt70. ...101. 0 - -01 0 .1 . 1. l . .0 A - - .1.00 10. 00.0010 0- . . . 0 . . 0 .0 o 0 O .. 1 11‘" . . 0 4 . , - .01.:‘ou4 4o)... ’6‘ 0. .0.10A.§ .01 . .01150 .10» 10.11011 . .. .. 1‘0. .. .... 1 . ..\ . . o I . . 1 .0 - . . 0. - A . 1" |||1 \llllllllllllllllllll i MAL EFFECTS a: EPCX‘Y REST/Vi MP érRAle-TE/ Echy (CMPc‘sJTL-‘y presented by j‘lMung 2/40“ has been accepted towards fulfillment of the requirements for Pk . D. degree in MA TE rat/1L5 sa Ell/CE Jor professor ; MSU is an Affirmative Anion/Equal Opportunity Institution 0-12771 Michigan State _ LIBRARY Unlverslty PLACE IN RETURN BOX to remove thin checkout from your To AVOID FINES Mum on or baton dot. duo. “ DATE DUE DATE DUE DATE DUE l use 2 e. m: ‘5. WW? 5‘ 28 96* .1 7 record. 1 I l ‘. __—— #— 71 ——1 _____._‘.—- HY GROTHERMAL EFFECTS OF EPOXY RESIN S AND GRAPHITE/EPOXY COMPOSITES By J iming Zhou A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Materials Science and Mechanics 1996 rel: ab34 scar Spec Siren; deten ABSTRACT HYGROTHERMAL EFFECTS OF EPOXY RESINS AND GRAPHITE/EPOXY COMPOSITES By J irning Zhou The aim of this study is to assess the nature of the sorbed water in epoxy resin and related hygrotherrnal effects in epoxy resins and graphite/epoxy (Gr/Ep) composites. Three neat epoxy resin systems and one Gr/Ep composite were used in this study. Water absorption, desorption, and dimensional variations were investigated. Optical and scanning electron microscopy (SEM and Environmental SEM), Fourier-transform infrared spectroscopy (FTIR), differential scanning calorimetry (DSC), therrnomechanical analysis (TMA), nuclear magnetic resonance (NMR), and mechanical tests (elastic modulus, strength, failure strain, fracture energy, and fracture toughness) were employed to determine and characterize the effects of water in epoxy resins and Gr/Ep composites. The claim of two types of bound water, as related to the water-matrix interactions and bonding characteristics, are investigated and indirectly justified by spectroscopy methods in neat epoxy resins. These bonded water states are designated and distinguished as A-bonded water and F—bonded water. A-bonded water is classified as physiosorbed water. Water molecules associated with the A-bonded state diffuse into the resin matrix, break the interchain hydrogen bonds which existed initially in the epoxy resin, and form weak hydrogen bonds via Van der Waals forces. F-bonded water is suggested as so- called chernisorbed water which interacts with hydrophilic groups of the epoxy resin. The amount of the F-bonded water that exists in hygrothermally-exposed epoxy resins depends tra do< hist syst long wate base to ex! strongly on the immersion temperature and time. In addition, a third type of bound water can exist on the surface and near surface area of Gr/Ep composite. Such bound water state is noted as O-bonded water. O-bonded water is characterized by its propensity to be retained at surface defects such as cracktips and voids due to surface tension and capillary effects. Reported in this work is also in the profound effects of sorbed water on the glass transition temperature, T,. The variation of TB of epoxy in a hygrothermal environment does not simply depend alone on the water content in epoxy resin. The hygrothermal history of the materials plays a major role in the resultant T3 observed. For a given epoxy system that maintains a constant maximum water content, it was show experimentally that longer exposure time and higher exposure temperature cause higher T8. T3 is a fimction of water content, exposure time and temperature, i.e., T3 = f (%M, t, T). An interpretation based on the bonding characteristics of water molecules and epoxy network is introduced to explain the T, variation. Water sorption in epoxy resins and Gr/Ep composites exhibited both Fickian and non-Fickian difi‘usion behavior. Diffusion data showed that the time for the onset of non- Fickian behavior was inversely related to the exposure temperature. Anomalous (non- Fickian) behavior in the composite resulted from chemical modification and physical damage to the epoxy resin. Cracks, voids, and surface peeling were observed clearly by SEM and optical microscopy. T o my mfe and son iv Lu Wi- flu and the Sm facil EXCI a11d Um. ACKNOWLEDGMENTS I would like to express my sincerest thanks to my advisor Professor James P. Lucas for all of his excellent guidance, friendly manner, and professional example. Without his support and patience this work would not have been possible. I would like to thank Professor Thomas P. Bieler, Professor David S. Grummon, Professor Andre Lee, and Professor David H. Yen for serving on my advisory committee and critically reading the manuscript. Special thanks go to Mr. Michael Rich and the Composite Materials and Structures Center, Michigan State University for providing convenient usage of the facilities for material preparation, sample processing, and property testing. I appreciate the financial support provided by my advisor through the Research Excellent Fund. My deepest thanks go to my wife, my parents, and my fiiends whose love, support, and encouragement have supported me to achieve a Ph. D. degree from Michigan State University. List 0 List 0' List 01 Chapt Chapt Chapt 2.1 2.2 ‘ 2.3} TABLE OF CONTENTS List of Tables -- - ..... - ‘ - ......... - x List of Figures - ‘ ................... xii List of Abbreviations and Symbols - - - -- - - - xvii Chapter Content Brief -- - -- - 1 Chapter I Introduction - - -- - 2 Chapter II Literature Review - - - ...... - ......... - -- -- 8 2.1 The Structure of Epoxy Resin and Epoxy Composites ................................... 9 2.1.1 Polymers .......................................................................................... 9 2.1.2 Epoxy ................................................................................................ 11 2.1.3 Graphite Fiber .................................................................................... 13 2.1.4. Graphite/Epoxy Composites .............................................................. 17 2.2 Water Transport in Epoxy Resin ................................................................... 18 2.2.1 Absorption Kinetics in Epoxy Resin ................................................... 18 2.2.2 Mechanisms of Water Absorption ...................................................... 21 2.3 Hygrothermal Effects on Epoxy and Gr/Ep composites ................................ 23 2.3.1 Plasticization and T, Variation ........................................................... 23 2.3.2 Change of Mechanical Properties ...................................................... 25 2.3.3 Swelling Induced by Sorbed Water ..................................................... 27 2.3.4 Hygrothennal Degradation in Epoxy Resin and Gr/Ep Composites ..... 28 4.3, 4.41 Chapter 111 Nature of Water in Epoxy ......... 31 3.1 Experimental ................................................................................................. 33 3.1.1 Materials ............................................................................................ 33 3.1.2 Water Absorption Test ....................................................................... 37 3.1.2.1 Analysis ............................................................................... 37 3.1.2.2 Test Procedures ..................................................................... 39 3.1.3 Water Desorption Test ....................................................................... 41 3.1.4 NMR test .......................................................................................... 43 3.1.5 DSC test ............................................................................................ 48 3.2 Results and Discussions ................................................................................ 49 3.2.1 Water Absorption ............................................................................. 49 3.2.2 Water Desorption ............................................................................. 53 3.2.3 F-bonded Water ................................................................................ 61 3.2.4 A-bonded Water ................................................................................ 64 3.4 Summary ....................................................................................................... 72 Chapter IV Tg Change of Epoxy in Hygrothermal Environment ........... 72 4.1 Materials and Experimental ........................................................................... 75 4.2 Results ......................................................................................................... 77 4.2.1 Tg Change with Exposure Time ......................................................... 77 4.2.2 Tg Change with Exposure Temperatures ............................................ 84 4.2.3 Tg Change at Various Desorption Stages ........................................... 89 4.3 Discussions ....................................................... 90 4.4 Summary ....................................................................................................... 94 vii Chapter V Swelling and Mechanical Property Change of Neat Epoxy in Hygrothermal Environment -- _ -- 95 5.1 Materials and Experimental Procedure ........................................................... 97 5.2 Swelling ....................................................................................................... 98 5.3 Mechanical Property Change ........................................................................ 102 5.3.1 Mechanical Property Change in Water Saturation Stage ................... 102 5.3.2 Mechanical Property Change in Different Hygrothermal Stages ......... 104 5.4 Summary ...................................................................................................... 111 Chapter VI Diifusion of Water in Gr/Ep Composites - -- -- -- 113 6.1 Materials ..................................................................................................... 115 6.2 Experimental ................................................................................................ 117 6.3 ReSults ........................................................................................................ 117 6.4 Discussions .................................................................................................. 126 6.4.1 Moisture-Induced Material Response ................................................ 126 6.4.2 Crack/Mass-Loss Model ................................................................... 129 6.5 Summary ...................................................................................................... 134 Chapter VII Hygrothermal Effects on Gr/Ep Composites -_ - -- 136 7.1 Experimental ............................................................................................... 138 7.1.1 Material ........................................................................................... 138 7.1.2 Water Absorption and Desorption ..................................................... 138 7.1.3 Glass Transition Temperature Tests ................................................... 139 7.1.4 Infi'ared Tests .................................................................................... 139 7.1.5 Mechanical Tests ............................................................................... 140 viii Bil 7.2 Results and Discussions ............................................................................... 142 7.2.1 Water Absorption and Surface Modification ..................................... 142 7.2.2 Water Desorption ............................................................................. 143 7.2.3 Change of Glass Transition Temperature ......................................... 145 7.2.4 Mechanical Testing Results ............................................................. 151 7.2.5 Swelling and Dimensional Change ................................................... 155 7.2.6 Infrared Results ............................................................................... 157 7.3 Summary .................................................................................................... 160 Chapter VIII Conclusions and Recommendations - - - - - - 161 8.1 Conclusions ................................................................................................. 161 8.2 Recommendations ....................................................................................... 165 Bibliography _- - - - - 167 ix Ta Ta Ta' Ta] Tab Tab] Table 2.1 Table 3.1 Table 3.2 Table 3.3 Table 3.4 Table 4.1 Table 4.2 Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 LIST OF TABLES Types of bonds and their strength in polymer. (p. 12) The difl’usion related parameters of the three epoxy systems: diffusivity D at difi‘erent immersion temperatures; activation energy Q; and maximum water content M... (p. 52) The amount of retained water in the three water-saturated epoxies afier desorption at 60 °C for 1450 h. The trend is that higher immersion temperature induces greater retained water. (p. 56) Desorption diffusivity of water in the material desorption at 60 °C in a dry chamber. (p. 59) Desorption diflirsivity of water in the material desorption at 140 °C in a dry chamber. (p. 60) T, change of the three epoxies in different hygrothermal stages. (p. 89) Experimental data of the three epoxies. T, of dry epoxy T8,; density of epoxy p3; equilibrium water content Mm; volume fraction of epoxy V,; liquid and glassy thermal expansion coefficient or... and age; epoxy volumetric expansion coefficient one =3(cr.e - age); and calculated T8 of water-saturated epoxy Tm. (p. 91) Dimensional change at different hygrothermal stages. The samples were immersed in 45 °C water chambers for water absorption. Water desorption was conducted at 60 °C for 1450 h and 140 °C for 240 h. (p. 101) The influences of A~bonded water and F—bonded water to swelling. (p. 102) Tensile modulus change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. (p. 109) Flexure modulus change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. (p. 109) Tensile strength change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. (p. 110) T: T2 T2 Ta Table 5.6 Table 5.7 Table 6.1 Table 7.1 Table 7.2 Table 7.3 Flexure strength change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. (p. 110) Tensile failure strain in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. (p. 111) The diffusivity of T300/934 graphite/epoxy composite materials. (p. 119) The comparison of theT, results in the long time exposed epoxy F iberite 934 and Gr/Ep composite Fiberite T300/934. (p. 151) Three point bending test results of the 90 °C samples at difi‘erent hygrothermal stages. (p. 152) Dimensional change of the 90 °C samples at different hygrothermal stages. (p. 157) Fr Fig Fig Fig Fig Fig Fig. Fig. Fig Fig Fig, i Fig 5 Fig. 3 Fig 3 Fig. 2.1 Fig. 2.2 Fig. 2.3 Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 3.5 Fig, 3.6 Fig. 3.7 Fig. 3.8 Fig. 3.9 LIST OF FIGURES Crystallographic structure of (a) graphite and (b) turbostratic graphite. (p. 13) Structure of PAN molecule. (a) original PAN fiber and (b) stabilized PAN through oxidization. (p. 15) The mechanism for the formation of graphite ribbons in PAN fibers. (p. 16) Chemical structures of TGDDM, DGEBA, mPDA, and DDS. (p. 34) FTIR spectrum of Fiberite 934 epoxy resin after being cured at 177 °C for 2 h. The existence of 910 cm'1 means that the epoxy resin was not cured completely. (p. 35) FTIR spectrum of Fiberite 934 epoxy resin after being postcured at 190 0C for 7 h. The nonexistence of 910 cm'1 means that the epoxy resin was cured completely. (p. 36) Illustration of the change of moisture content with the square root of time. (p. 40) ESEM micrographs of My 720 epoxy resin (a) as-cured and (b) hygrothermally exposed at 90 °C for 1530 h. Both shows that there is no cracking or void on the resin surface. (p. 42) Magnetic moment separates in a magnetic field. (p. 46) Energy level separation of 1H in a magnetic field. (p. 47) Typical solid state NMR spectroscopy. (p. 47) Water absorption profiles of the three epoxy systems at difi‘erent temperatures for 1530 h. The symbols are experimental data. (p. 51) Fig. 3.10 Transverse diflirsivity as a firnction of temperature to determine activiation energy Q of the three epoxy systems. (p. 52) Fig. 3.11 Water desorption profiles of the three epoxy systems at 60 °C for 1450 h and then 140 °C 240 h. The symbols are experimental data and represent the samples with different bath temperatures in water absorption process. (p. 55) X11 Fig, Fig, Fig, . Fig. 3.12 Moy’s water desorption results. (p. 56) Fig. 3.13 Marsh’s water desorption results. (p. 578) Fig. 3.14 Xiang’s water desorption results. (p. 57) Fig. 3.15 The extrapolated absorption diffusivity at temperature of 140 °C. (p. 60) Fig. 3.16 The amount of chernisorbed water of the epoxy Fiberite 934 changes with immersion time at 90°C. (p. 61) Fig. 3.17 NMR spectra of 1H for TGDDM+DDS system desorption at 60 °C for 1450 h. The results indicate that F—bonded water is tightly bonded with epoxy network. (p. 63) Fig. 3.18 NMR spectra of lH for TGDDM+DDS system. The results indicate that the mobility of A-bonded water in epoxy is between free-water and solid states. (p. 65) Fig. 3.19 NMR spectra of Fiberite 934 show fi'ee (liquid) water changes into impeded A- bonded water with time. (a) dry resin mixed with 1% liquid water and tested. Liquid water and A-bonded water can be seen clearly. (b) Ten minutes later, liquid water disappeared and all sorbed water molecules were impeded. (p. 66) Fig. 3 .20 The calibration curve of the amount of clustered (free-state) water in water saturated epoxy resin. (p. 68) Fig. 3.21 DSC trace of water saturated Fiberite 934 resin for determining the amount of fi'ee water. (p. 69) Fig. 3.22 DSC results of bulk (liquid) water content in Fiberite 934 epoxy after immersion in distilled water at different temperatures for 2200 h. Bath temperature represents the sarnples' immersion temperatures in water absorption process. (p. 70) Fig. 3.23 A schematic conceptualization of epoxy network and bonding characteristics before and after water absorption. (p. 71) Fig. 4.1 T, change of TGDDM+DDS with exposure time, the exposure temperature was 90 °C. (p. 79) Fig. 4.2 T, change of DGEBA+mPDA with exposure time, the exposure temperature was 90 °C. (p. 80) Fig. 4.3 T, change of Fiberite934 with exposure time, the exposure temperature was 90 °C. (p. 81) xiii Fir Fig Fig Fig. Fig. Fig. Fig. . Fig. : Fig~ 5 Fig.5 Fig. 5 Fig. 5 Fig. 5. Fig. 6. Fig. 6.: Fig. 6.3 Fig. 4.4 Fig. 4.5 Fig. 4.6 Fig. 4.7 Fig. 4.8 Fig. 4.9 Fig. 5.1 Fig. 5.2 Fig. 5.3 Fig. 5.4 Fig. 5.5 Fig. 5.6 Fig. 5.7 Fig. 6.1 Fig. 6.2 Fig. 6.3 T, change with exposure time. Samples were immersed in water at 90 °C for 1530h. Upper figure is the corresponded water absorption profiles at the same temperature. (p. 82) T, change with exposure time. Samples were immersed in water at 60 °C for 1530 h. Upper figure is the corresponded water absorption profiles at the same temperature. (p. 83) T, change of TGDDM+DDS with bath temperatures after immersion in water for 1530 h. (p. 85) T, change of DGEBA+mPDA with bath temperatures after immersion in water for 1530 h. (p. 86) T, change of F iberite 934 with bath temperatures after immersion in water for 1530 h. (p. 87) A comparison of T, test results btween DSC and TMA The material is DGEBA + mPDA immersion in water for 1530 h at different temperatures. The T, values compared favorable between the two test methods. (p. 88) Scheme of experimental water absorption and desorption process. (p. 98) Dimensional change with immersion time and temperatures. (p. 99) Dimensional change with sorbed water content. (p. 101) Change of modulus at different bath temperatures. (p. 105) Change of strength at different bath temperatures. (p. 106) Change of strain at different bath temperatures. (p. 107) Explanation of a chain scission in TGDDM+DDA system. (p. 107) FTIR spectrum of the matrix resin of as-received F iberite T300/934 composite. It shows that the material is full cured. (p. 116) The weight change of T300/934 graphite/epoxy composite immersed in distilled water at difl‘erent temperatures. Solid lines represent theoretical Fickian difi‘usion and the symbols are the experimental data at different exposure temperatures. (p. 119) Width change of T300/934 graphite/epoxy composite immersed in distilled water at different temperatures. (p. 122) xiv Fi a, Fig Fig Fig Fig. 6.4 qucis Fig. 6.6 Fig. 6.7 Fig. 6.8 Fig. 6.9 Thickness change of T300/934 graphite/epoxy composite immersed in distilled water at different temperatures. (p. 122) Optical micrographs of T3 00/934 graphite/epoxy composite before and alter immersion in water for 4300 hours. a) dry specimen, no crack can be seen, b) 75°C, and c) 90°C specimens. Both of the 75 and 90°C specimens have visible cracks. (p. 123) SEM nricrographs for surface images of dry and 90°C specimens. The surface layer is primarily neat resin. The woven-like surface paten is due to the impression of the cloth texture on epoxy resin in the manufacture process. a) dry specimen, no damage on the surface and b) 90°C specimen, there is resin peeling after 4300 hours immersion in distilled water. (p. 124) SEM micrographs of the surface of 90°C specimen after 4300 hours immersion in water. a) resin peeling is high visible, b) peeling, cracks and dissolution. (p. 125) Schematic diagram of unidirectional Gr/Ep composite prepreg. (p. 130) Moisture absorption profiles of a graphite/epoxy laminate showing experimental data and theoretical prediction of Fickian diffusion. (p. 131) Fig. 6.10 The schematic diagram of typical weight change profiles of Gr/Ep composites in Fig. 7.1 Fig. 7.2 Fug'is Ifig.7!i Fig. 7.5 Fig. 7.6 water environment. (p. 132) The laminate short bar (LSB) test specimen. (p. 141) The weight change profiles of the T300/934 composite desorption at 60 °C for 1250 hrs and then at 100°C for 250 hrs. (p. 144) T, of the four saturated specimens after immersion in water for more than 9000 hrs at different temperature. a) 45 °C, b) 60 °C, c) 75 °C, d) 90 °C. (p. 146) T, of the four saturated specimens after desorption at 60 °C for 1250 hrs. a) dry specimen, b) 45 °C, 0) 60°C, d) 75 °C, e) 90 °C. (p. 147) T, of the four saturated specimens after desorption at 60 °C for 1250 hrs and then at 100°C for 250 hrs. a) dry specimen, b) 45 °C, c) 60 °C, (1) 75 °C, e) 90 °C. (p.148) The T, variations with the different immersion temperatures and desorption stages. (p. 149) Fig Fig Fig Fig Fig. 7.7 The fracture energy results of the T3 00/934 Gr/Ep composite saturated in water for difi‘erent hygrothermal conditions. (p. 153) Fig. 7.8 The effect of moisture sorption on delarnination fiacture toughness is exhibited for Gr/Ep composite laminates. (p. 154) Fig. 7.9 Fracture in the crack tip region is shown for Gr/Ep composites. (a) Fracture occurs along the fiber matrix interface. (b) Fracture is shown at the interfacial regions between to fibers. (p. 156) Fig. 7.10 Infiared spectra of Fiberite T300/934 Gr/Ep larrrinate at different hygrothermal stages. (p. 159) xvi DDA DDS DGEBA DFT DMA DSC ESEM ILSS LIST OF ABBREVIATIONS AND SYMBOLS Magnetic field Moisture Concentration Ambient moisture Concentration Mass Difl’usivity Double Cantilever Beam Dicyandiamine 4,4'-Diaminodiphenyl Sulfone Diglycidyl ether of bisphenol-A Delarnination Fracture Toughness Dynamical Mechanical Analysis Differential Scanning Calorimeter Potential Energy Environmental Scanning Electron Microscopy Fourier Transform Infiared Spectroscopy Graphite/Epoxy Composites Planck’s Constant Proton Hamiltonian Operator Spin vector Interlaminar Shear Strength xvii LSB B a}: mPDA "U PAN PMCs SEM TETA We Fracture Toughness Laminate Short Bar Spin Quantum Number Magnetization Equilibrium Water Content Total Percent of Water Gain Metaphenylene Diarnine Nuclear Magnetic Resonance Angular Momentum Polyacrylonitrile Polymer Matrix Composites Activation Energy Relative Humidity Scanning Electron Microscopy Triethylenetetramine Time Temperature Glass Transition Temperature Tetraglycidyl-4, 4'-diaminodiphenyl Methane Thermalmechanical Analysis Ultraviolet Volume Fraction of Epoxy Weight of Material Dry Weight of Material xviii «f P 9. Liquid Thermal Expansion Coefficient Glassy Thermal Expansion Coefficient Volumetric Expansion Coemcient of Epoxy Density of Epoxy Magnetic Moment of Nucleus Magneto-gyric Ratio Poisson Ratio xix This t1 status of by; introduces th: miew regard resins and Gr/ three kinds of epoxy resin an temix-Write v; are CondUCied epoxy resins hl’gTOthenna] t Fickjan an d m Provides the re Funny, COnCIUg CHAPTERS CONTENT BRIEF This thesis is organized as follows: Chapter 1 gives a statement of current research status of hygrothermal efl‘ects on epoxy resins and epoxy matrix composites and introduces the motivation and objectives of this study. Chapter 2 provides a literature review regarding the nature of water in epoxy resins and the effects of water in epoxy resins and Gr/Ep composites. Experimental water absorption and desorption behaviors in three kinds of epoxy resins are shown in Chapter 3. In this chapter the nature of water in epoxy resin and the bonding characteristics are investigated. In Chapter 4 glass transition temperature variations in different hygrothermal stages are studied. Extensive discussions are conducted regarding the relation of Tg variation and water bonding characteristics in epoxy resins. Swelling, degradation, and mechanical property change of epoxy resins in a hygrothermal environment are presented in Chapter 5. Water absorption behaviors (both Fickian and non-Fickian) in Gr/Ep composites are discussed in Chapter 6. Chapter 7 provides the results of investigations regarding hygrothermal effects on Gr/Ep composites. Finally, conclusions and recommendations are given in Chapter 8. CHAPTER I INTRODUCTION Owing to their high strength, high dimensional stability, excellent dielectric properties, and light specific weight, polymer matrix composites (PMCs), especially epoxy matrix composites; have for several decades been widely used in space, electronics, and aerospace industries [Lubin 1982, ASM 1988, Springer 1981, 1984, 1988]. In recent years, with significant improvements in processing techniques, manufacturing methods, and cost reduction, applications of graphite/epoxy (Gr/Ep) composites in such commercial markets as automotive components, sporting goods, and marine environments continue to grow. Like corrosion is for its metallic material counterpart, environmentally induced efi‘ects are of major concern for epoxy resins and epoxy matrix composites. Generally environmental factors include temperature, gas (N 0,, 80,, oxygen etc.), radiation (UV, gamma radiation etc.), moisture, and chemical dissolution [Bank 1995]. Of the most interest are the hygrothermal effects because water or moisture and temperature are the most common of all environmental factors causing performance reduction in polymers and PMCs. Hygrothennal environmental effects of epoxy resins and graphite/epoxy composites have been investigated for a long time for practical reasons [Schutte 1994]. The most prominent amongst the reasons is the fact that sorbed moisture promotes material dr reversible microcrack In 2 composites Gaithersbur molecular Ir relationship swelling, era in cOmposit. term Physica Althr Quite Some . diffirsion beh water into re Fleian_baSe( ability 0f the 3 material degradation. Some of the physical and mechanical property changes are reversible (plasticization and swelling) and some are irreversible (hydrolysis and microcracking). In a recent workshop on hygrothermal effects in polymers and polymer matrix composites (sponsored by the National Institute of Standards and Technology, NIST, Gaithersburg, MD, September, 1995), the key topics that were focused on were: i) molecular level understanding of water in polymer and polymer matrix composites, ii) the relationship of sorbed water and macroscopic hygrothermal effects such as plasticization, swelling, cracking, T, and mechanical property variation, iii) water-induced failure models in composites, especially those associated with the fiber-matrix interface; and iv) long- term physical and mechanical behavior predictions of PMCs. Although the mechanisms of water in epoxy resin have been widely studied for quite some time, many aspects are still not well understood. A primary issue is the diffusion behavior of water in epoxy resins and PMCs. Some studies on the diffusion of water into resins assumed Fickian behavior. However, there are many examples of non- Fickian-based diflirsionm [Weitsman 1995, Zhou 1995, Loos 1981, Imaz 1991]. The ability of these materials to absorb water is another important topic in this field. The water equilibrium level in epoxy resins is influenced by many factors. These include the chemical structure of the resin, the bonding characteristics of water and resin, the matrix- fiber interface constraint, microcracking, hygrothermal exposure history, and others. In general, there are two mechanistic approaches to characterizing water or moisture in epoxy resin and Gr/Ep composite. One is the free volume approach, which presumes that water diffuses into epoxy resin and resides in the available free volume. CS: rm “'8 of ; b0n coll. they Wale "fig; eXchi degra 4 Chemical interaction between water molecules and the epoxy resin is considered insignificant in this approach [Gupta 1985, Woo 1987]. The other approach is the interaction concept, which suggests that water molecules couple strongly with certain hydrophilic groups such as hydroxyls or amines in resin [Moy 1980]. Also Bellenger and Verdu [1989] suggested that water sorption in a polymer resin is essentially governed by water-polymer chemical interactions. The water equilibrium concentrations can be predicted using a simple additive law, depending essentially on the extent of intrarnolecular hydrogen bonding. An integrative model has also been developed that considers that some water molecules form hydrogen bonds with hydrophilic groups in the epoxy resin while other water molecules are retained in the free volume in the epoxy resin [Adamson 1980]. Apicella et. al. [1985] proposed a model in which three modes of sorptive behavior exist: i) bulk dissolution of water in a polymer network, ii) moisture absorption onto the surface of holes that define the excess free volume of the glassy structure, and iii) hydrogen bonding between hydrophilic groups of the polymer and water. Jelinski and his collaborators [1985] investigated the nature of the epoxy-water molecule interaction and they showed that: i) the water in epoxy resin is impeded in its movement, ii) there is no fi'ee water, iii) there is no evidence for tightly bound water, and iv) it is unlikely that the water disrupts the hydrogen-bonded network in the epoxy resin. The water molecules migrate from site to site, but such a jumping motion does not involve a specific hydrogen- exchange mechanism. Major hygrothermal effects in epoxy systems are primarily plasticization, degradation, swelling, and lowered T,. Plasticization is the most extensive of all thar clus repc 18m; 5 hygrothermal effects. Many studies have reported that mechanical property degradation is due to water absorption in epoxy resins and Gr/Ep composites. Investigations have revealed that absorbed water in epoxy acts as a plasticizer [Shen 1981, Imaz 1991, Biro 1993 Lucas 1989, 1993, 1995, Wolff 1993]. The effect of water absorption in Gr/Ep composite is manifested by a reduction in elastic modulus and by a decrease in the matrix- dominated strength properties. DeNeve and Shanahan [1993] concluded that the hygrothermal aging of epoxy leads to both plasticization of the polymer (a physical effect) and chain scission (a chemical effect). Although it is widely accepted that ~1% of sorbed water may decrease T, by 10-20 °C, the interpretation for the decrease in T, is varied. Based on the fiee volume concept of polymeric materials, many studies have shown that T, change simply depends on water content of the resin [Kelley 1960, Browning 1978, Cairns 1984]. DeIasi [1978] suggested tint water which disrupts the interchain bonds could depress T,, whereas, water that forms clusters or hydroxy-water type groupings has no measurable effect on T,. Mijovic [1985] reported that T, was also influenced not only by water concentration but also by exposure temperature. Swelling denotes volumetric dimensional change due to moisture content alone, independent of thermal expansion. Hahn [1976] showed that the sorbed water produces relatively little swelling until a critical amount of water is absorbed, and then the resin sarnple volume increases proportionally to the additional water content. Adamson [1980] considered the swelling coeflicient of water absorbed to be temperature and concentration dependent. Sorbed water molecules may either occupy free volume (causing no swelling), or interrupt interchain hydrogen bonding (causing swelling). mechanically investigations 1851ng to ex; probing meth. them with Si bChavior in ep . The air water in ep0ir desorpliOn ph, water in ep0x) borh epoxy res PTOpeny Variati The mat “mposire, 0n and metapheny diamlmdlphem 6 Although there is some general agreement among some models, considerable disagreement still persists. The key point is the nature of water in epoxy resin. The problem is if water stays in the material in the form of bulk or cluster, or bonded with hydrophilic groups, or separating in free volume. Since there is no enough knowledge on the basic behavior of water in epoxy resin the interpretation on plasticization, swelling, and degradation is varied. This is due in part to the reality that the hygrothermal effects, as a consequence of moisture absorption in polymers and PMCs, are chemically and mechanically complex. The other reason is that most of the previous scientific investigations did not apply correlated integral spectroscopic methods and mechanical testing to explore the effects of water in epoxy resin systems. Usually only one or two probing methods were employed to yield some results. Consequently there is no single theory with sufficient experimental support to explain all related phenomena of moisture behavior in epoxy resins and epoxy matrix composite systems. The aim of this study is to discern the effects and bonding characteristics of sorbed water in epoxy resin and Gr/Ep composite by evaluation of: i) water absorption and desorption phenomena in epoxy resin and Gr/Ep composites, ii) sorption mechanisms of water in epoxy resin, iii) plasticization and glass transition temperature modifications for both epoxy resin and Gr/Ep composites, iv) moisture induced swelling, and v) mechanical property variations in different hygrothermal exposure conditions. The materials used in this study involve three kinds of epoxy resins and one Gr/Ep composite. One epoxy resin is diglycidyl ether of bisphenol-A (DGEBA, Shell Epon828) and metaphenylene diamine (mPDA) epoxy system. The second is tetraglycidyl-4, 4'- diaminodiphenyl methane (TGDDM, Ciba Geigy My720) resin with 4,4'-diaminodiphenyl 7 sulfone (DDS, DuPont) hardener. The third is Fiberite 934 epoxy system which consists mainly of TGDDM and DDS with small amounts of other additives of which the types and their concentrations are proprietary. The three kinds of epoxies have been used extensively for high performance polymer matrix composites. The Gr/Ep composite is Fiberite T300/934 unidirectional laminate. In this study water absorption, desorption, and dimensional change measurements were carried out. Optical and electron microscopy (SEM and Environmental SEM), Fourier-transform infrared spectroscopy (FTIR), differential scanning calorimetry (DSC), thermomechanical analysis (TMA), nuclear magnetic resonance (NMR), and mechanical test (tensile, flexural, fracture energy, and fracture toughness) were employed to determine and characterize the properties and effects of water in epoxy resins and Gr/Ep composites. Using an integrative experimental approach, a more comprehensive understanding of the nature of water and the hygrothermally induced effects in epoxy resin and epoxy matrix composites has emerged. CHAPTER II LITERATURE REVIEW In this chapter we will review the literature on the nature of water in epoxy resin and the hygrothermal effects in epoxy resins and Gr/Ep composites. The sorption of water in PMCs and their effects on composite performance are highly complex issues. Their study involves, in the least, the disciplines of polymer science and applied mechanics, where the former focuses on molecular-level interactions and the latter is concerned with mechanical response. A historical perspective of hygrothermal effects investigation can be found in several review articles [Weitsman 1995, Lee 1993, Bank 1995, Wolff 1993, Schutte 1994]. The kinetics of water sorption in polymer has been studied for about one and a half centuries, perhaps beginning with F ick [Weitsman 1995]. Before high performance PMCs were widely used in aerospace industry, hygrothermal efl‘ects were considered to be of only secondary importance. Glass fiber composites, for example, were utilized in the boating industry. While such composites exhibit substantial sensitivity to water, these detrimental effects were overcome by over- design, with lesser concern for any possible waste in weight or performance. With the application of tight design requirements and high performance polymer matrix composites, especially Gr/Ep composites applied in aerospace industry, the interest in water sorption and its effects on polymers received new impetus in the mid 1970s. The first wave on the 9 study of hygrothermal efi‘ects of high-performance PMCs came about in the late 19708 and early 19803. Springer’s three volumes of collective works [Springer 1981, 1984, 1988] represented the studies of that period. In the late 1980s, the issue was relatively quiet. With the quick development of the applications in the use of PMCs, the second wave came about in the 19905. It appears that the wave of interest in environmental effects on PMCs may be due to new applications contemplated for these materials, which combine severe exposures and tighter design requirements. These new applications occur in the ofl‘shore oil industry, in naval submersibles, and in the automotive industry. This time research has focused on molecular level interactions of water with polymers and the mechanical response of water-sorbed PMCs [Cappelletti 1995, Etxeberria 1995, Karasek 1995, Rao 1995, Lucas 1995]. 2.1. THE STRUCTURE OF EPOXY RESIN AND EPOXY COMPOSITES 2.1.1. Polymers The fimdamental difference between polymers and other classes of materials is derived from their linear-molecular nature. Although polymers may be connected by crosslinking, their basic structure consists of a linear repeating pattern of strong bonds, in contrast to the three-dimensional structures associated with metal and ceramics. The consequence of this fundamental difference in bonding is that many of the mechanical properties depend on the weaker forces between the chains, in addition to the primary covalent bonds along the main chain axis. Without entanglement or some other forces holding chains together few useful mechanical properties would develop. Consequently, polymers are susceptible to thermal efi‘ects at relatively low temperatures. Their 10 properties largely reflect the lower magnitude of these secondary forces holding the polymer chain together. This constitutes a great manufacturing advantage because it allows processing at modest temperatures. This has led to a rapid increase in the use of polymeric materials, and it provides a driving force to better understand the influence of temperature, moisture, and other environmental variables on the final properties, since water molecules and relatively modest temperatures and can disrupt the secondary force. Also, minor changes in the number of covalent bonds along a chain have an enormous effect on the molecular weight and, hence, the mechanical properties. There are two important concepts on polymeric materials, which can help us better understanding the behavior of water in epoxy resins. One is “occupied volume” and the other is “fi'ee volume”. Occupied volume is defined as that volume occupied by the actual mass of a molecule plus the volume it occupies because of thermally dependent harmonic vibration that excludes all other molecules fi'om its domain. Occupied volume is dependent on temperature because thermally induced motion causes each molecule (or molecular segment in the case of a cross-linked polymer) to move about its equilibrium position, therefore occupying more space than its actual mass volume. If the molecules were packed so that their domains were in perfect contact, the occupied volume of the polymer at any temperature would exactly equal the sum of the actual mass volume of the molecules and their vibrational volume. The domain are not in perfect contact, however, and fi'ee volume is defined here is the difference between the measured volume of a polymer and the occupied volume. This difference is the result of “holes” or “voids” caused by packing irregularities. A critical free volume fraction at T, for most polymers is about 1/40 (2.5 :1: 0.3%) of the total volume occupied by the substance [Adarnson 1980]. 11 The usefirl mechanical properties of polymers result fiom the high molecular weight of the chains. In a highly cross-linked form, an infinite molecular weight is reached and a classification according to the mean molecular weight between crosslinks is more applicable. The types of bonds that hold the polymer together are listed in table 2.1. The strongest are covalent bonds along the main chain of the polymer, while much weaker bonds exist between the repeat units of the polymer. These secondary interactions are repeated sequentially many times and lead to substantial strengthening. In addition, there are the related frictional forces that restrict the ability of chains to untangle when a polymer is deformed. 2.1.2. Epoxy By definition, any resin containing the molecular group —(lj/—Elj .— is called epoxy resin. Epoxy is one of the polymeric materials and the resins provide the highest adhesive strength of any known polymeric material. Epoxies can thoroughly wet a wide variety of substrates with minimal shrinkage. Epoxy resins can be used to join many dissimilar materials and to form excellent-property composites. The epoxy group can bond chemically with other molecules, forming a large three- dimensional network. This process, called curing, changes a liquid resin into a solid. The most commonly used curing agents are aromatic and aliphatic amine curing agents (hardeners). Each hardener molecule contains a reactive group on each end. These permit the formation of a crosslink between epoxy molecules. An amine end-group with two hydrogens on the nitrogen (a primary amine) reacts with an epoxy molecule as follows: 0 OH / \ I HzN ------ NH 2 + Hzc-CH ------ —) H2N ----- NH - CHz—CH ..... 12 When another amine hydrogen combines with a second epoxy molecule, crosslink is formed: r (rm /0\ 9H NzH ----- —CH2—CH ..... + HzC— CH ..... —> HzN ----- l—CHz—CHn—u EH2 H . —OH After thorough crosslinking, epoxy reaches almost infinitive molecular weight and forms a brittle and stifi‘ solid. The carbon-nitrogen bond formed in crosslinking is stable against most inorganic acids and alkalis. Table 2.1 Types of bonds and their strength in polymer [Atkins 1993] Main chain covalent bonds C — C, C —N, C — 0 60-100 kcal/mole Secondary (interchain) Hydrogen bond Dipole interaction Van der Waals 6 \ / \ / O C5+ CH2 CH2 N—H°°°C=C n u | | (35+ 06 CH2 CH2 / \ / \ ~5 kcal/mole ~2 kcal/mole ~05 kcal/ 111016 l 3 2.1.3. Graphite Fibers Carbon fibers are by far the predominant high-strength, high—modulus reinforcing agent currently used in the fabrication of high-performance resin-matrix composites. The crystallographic structure of a perfect single crystal of graphite is shown in part (a) of Fig. 2.1. As can be seen, the graphite crystal is composed of many sheet-like layers of the carbon atoms which are stacked one on top of the other and separated by a distance of 3.35 A. In the plane of the sheets, the carbon atoms are linked together by very strong covalent bonds. As a result, the theoretical tensile modulus of elasticity and ultimate tensile strength of the crystal in a direction parallel to the basal plane are very high — on the order of 1000 GPa and 15 GPa, respectively. On the other hand, relatively weak Van der Waals bonds hold the sheet-like layers of carbon atoms together in a direction normal to the basal plane; consequently, the mechanical properties of the crystal in this direction are much poor. In a graphite fiber, the structure of the crystallites is not that of the perfect single crystal. Instead, the stacking arrangement of the various carbon sheet is slightly displaced, thus forming a “turbostratic” graphite shown in part (b) of Fig. 2.1. The orientation of the graphite sheets, however, is still more or less parallel to the fiber axis. (h) Fig. 2.1 Crystallographic structure of (a) graphite and (b) turbostratic graphite [Lubin 1982] 14 The current technology for producing graphite fibers generally centers on the thermal decomposition of various organic precursors. Commercially available graphite fibers are made using one of three precursors: rayon, polyacrylonitrile (PAN), and pitch. For example, the process by which PAN is converted to carbon fibers involves seven steps: fl . Spinning the PAN precursor. 2. Stretching the precursor. 3. Stabilization at 220 °C in air under tension. 4. Dehydrogenation at 400-600 °C. 5. Denitrogenation at 600-1300 °C. 6. Carbonization at 1500 °C in inert atmospheres. 7. Graphitization at 1800 - 3000 °C in inert atmospheres. The chemical structure of PAN is shown in part (a) of Fig. 2.2. Part (b) of Fig. 2.2 is the structure of the stabilized PAN. Figure 2.3 shows the mechanism for the formation of graphite ribbons in PAN fibers. Carbonization heat treatments are generally carried out in an inert atmosphere at temperatures ranging from 1000 °C to 1500 °C. It is within this temperature range that most noncarbon elements are driven the precursor fiber. At temperature less than 1000 °C, a considerable amount of gaseous products, such as methane, hydrogen cyanide, water, carbon dioxide, carbon monoxide, hydrogen, and various other hydrocarbons, are evolved fi'om the precursor fiber. The chemical composition of the fiber at 1000 °C is approximately 94% carbon and 6% nitrogen. At 1300 °C the nitrogen content of the fibers is approximately 0.3%. In general, graphitization heat treatments are specifically carried out at temperatures in excess of 15 1800 °C in order to improve the tensile modulus of elasticity of the fiber by improving the crystallite structure and the preferred orientation of the graphite—like crystallites within each individual fiber. Graphite fibers have a more highly organized three-dimensional crystalline structure than non-graphitized carbon fibers. Therefore graphite fibers have higher moduli and strength and are better thermal conductors than carbon fibers. Graphite fibers are subjected to post treatment (including surface treatments an/or application of organic sizings) in order to improve their compatibility with the resin matrix and/or their handleability. The mechanical properties, including flexural strength, interlarninar shear strength (ILSS), and mode of fracture, of a graphite-fiber-reinforced H CH CH CH /C?\/2\/2\/?\/ CH CH CH CH r r r C“*N Giant C‘\‘=N C“\x‘N (a) N N OIH “\T clrr/ \i/ \$ CH c CH CH CH CH CH CH C/ \fi/ \(lz/ ’\(l:/ \?/ \CH/ \CHz/ \fi/ c c /c\ c\ o \N/ \N NH N (b) Fig. 2.2 Structure of PAN fiber. (a) original PAN and (b) stabilized PAN fiber through oxidization [Lubin 1982]. l6 composite depend on the nature of the resin-fiber bond. Organic coatings (sizings) in an amount of 0.5 -7 wt% typically applied by the manufacturer by passing the heated fibers through a sizing bath. The sizing agents most commonly employed are polyvinyl alcohol, epoxy, polyimide , and water. These coatings are applied to both untreated and surface- treated fibers. They not only improve the handleability and abrasion resistance of the fibers, but also affect their adhesion to the matrix. 400-600 C Dehydrogenation J 600-1300 C Fig. 2.3 The mechanism for the formation of graphite ribbons in PAN fibers. 1 7 2.1.4. Graphite/Epoxy Composites The main resins used in high performance fiber composites are thermosetting polymers that can be used at high temperatures. Thermoplastic polymers are used to a lesser extent because they are subject to changes in modulus and strength with temperature, solvent resistance is often poor, and high melt viscosity makes fiber wetting and infiltration difficult. Graphite/epoxy composites consist of graphite fibers embedded in an epoxy matrix. The graphite fibers provide stifliress and strength while the epoxy matrix distributes the stress to and among the fibers and holds them together. Gr/Ep composites have higher tensile moduli, tensile strength, and much lower density than metal. 2.2. WATER TRANSPORT IN EPOXY RESIN 2.2.1. Absorption Kinetics in Epoxy Resin A problem in which the temperature and moisture distributions inside the material are to be determined is frequently referred to as the “moisture problem”. It is required to find the following parameters: 1) the moisture concentration inside the material as a firnction of position and time, 2) the total amount of moisture as a function of time, and 3) changes in the “performance” of the material as a firnction of time. Answers to problem involving changes in performance (point 3) must be obtained by testing. Answers to the point 1 and 2 can be obtained by analytical means when the following conditions are met. 18 l) The temperature inside the material approaches equilibrium much faster than the moisture concentration, and hence the energy (F lourier) and mass transfer (F ick) equations are decoupled. 2) The moisture diffusion can be described by a concentration-dependent form of Fick’s law. 3) The thermal conductivity and the mass diffusivity depend only on temperature and are independent of moisture concentration or of the stress levels inside the material. When all the foregoing assumptions are satisfied the diffusion process is said to be “Fickian”. Appropriate equations and solution procedures are given in Chapter 3.2.1 for calculating temperature and moisture distributions in case of F ickian diffusion. The behavior of water in epoxy resin is highly complex. Moisture absorption and desorption at high temperatures and relative humidity can cause voids and/or rnicrocracks in the epoxy resin. If these effects are extensive, conditions for non-Fickian transport can be induced, Thus, we can expect non-Fickian as well as Fickian transport, according to the resin properties and environmental conditions. But it is still not well understood that for a given epoxy resin, some studys show physical damage and some studys do not observe any cracking and/or microvoids. The diverse results may be associated with difl‘erence in sample preparation and in chosen experimental procedures. Brewis et al. [1980] investigated water sorption kinetics of the diglycidyl ether of bisphenol A cured with several kinds of hardener. They found that the water sorption behavior of all the samples followed a Fickian diffusion model in the temperature range fiom 25 to 100 °C. Similar results in PMCs have been reported by Rao [1995]. He 19 studied moisture diffusion characteristics of T3 00/914C unidirectional composite specimens hygrothermally conditioned at 85% RH and 70 °C. All the test specimens exhibited Fickian diffusion behavior. He concluded that for well-fabricated, void-free composites, the F ickian difl’usion model is valid [Rao 1984, 1995]. Several researchers [Moy 1980, Wong 1985, Sahlin 1991] have reported non- F ickian diffusion in tetraglycidyl 4,4’-diamino-diphenylmethane (TGDDM) resin cured with DDS. Moy and Karasz [1980] suggested that such an effect is coupled either to relaxation processes or to irreversible chemical reactions. They proposed that the water interacted with the polymer through hydrogen bonding, as verified by differential scanning calorimetry. They showed that the residual water could only be removed from the resin by exposing it to a dry atmosphere heated above 100 °C. Wong and Broutrnan [1985] concluded that the non-Fickian process observed during a first sorption cycle may be due to insuficient crosslinking. Additional crosslinking could potentially occur during this sorption process. They observed Fickian behavior for subsequent sorption cycles. Sahlin and Pepas [1991] found that TGDDM/DDS resins exposed to water at 25 °C showed non- Fickian behavior. They explained that this behavior can be exhibited when the relaxation phenomena are of the same order of magnitude of time scale as diffusion. The non- Fickian behavior observed in the water uptake curves at low sorption times for samples exposed to water at 25 °C disappeared as the sorption temperature increased. Shirrell [1977] found non-Fickian absorption anomalies for both postcured and non-postcured T300/5208 composite exposed in moisture at 82 °C. The equilibrium moisture solubility of non-postcured specimens are independent of testing temperature, while there is a definite trend toward lower equilibrium moisture content with increasing test temperature 20 for postcured specimens. Loos (1981) observed both Fickian and non-Fickian absorption process in glass fiber reinforced composites and gave a plausible explanation. Moisture rapidly entered the material and caused microcracks. As the cracks developed, material, most likely in the resin particles, was actually last. As long as the moisture gain was greater than the material loss, the weight of the specimen increased. Once the weight of the lost material exceeded the weight of the absorbed water, the weight of the specimen decreased. This is a very good prediction but there was no firrther experiment to support it. Zhou and Lucas [Zhou 1995] studied the diffusion behaviors of Fiberite 934 resin and Fiberite T300/934 composite systematicly and reported the diffirsion is Fickian. The anomalous diffusion profiles are just due to material surface cracking and mass loss. Some researchers [Loos 1979, Illinger 1980, Wright 1981] showed that sorption behavior tends to occur in two stages. Whitney and Browning [1978] observed a two- stage diffusion process in Hercules 3501-5 resin at 71 °C. They suggested time-dependent matrix cracking as the mechanism associated with the two-stage difi'usion process. Gupta et al. [1985] found that the DGEBA resin cured with metaphenylenediarnine showed Langmir-type sorption behavior at room temperature while, at higher temperatures, Henry’s Law mode was more relevant. The Langmurian mode involes the entry of water molecules into preexisting gaps while in the Henry’s Law mode; the water molecules enter through gaps created by segmental motion. Carter and Kibler [Carter 1978] suggested a Langmuir-type model to predict two-stage sorption behavior for composite resins. The model could fit anomalous uptake curves for 5208 resin exposed to several relative hurnidities. From the fact that the same parameters give equally good fits to the data at all 21 humidities, they suggested that the absorption anomaly does not result fi'om non-linear efl‘ects. Studies over the last two decades have revealed some general results: the moisture transport in composites can be described by a Fickian diffusion model with a constant diflirsion coeflicient [Shen 1976, Loos 1979, Illinger 1980, B011 1985], a concentration- dependent diffusion coeflicient [Shirrell 197 8], or a stress-dependent diflirsion coefficient. Weitsman [Weitsman 1995] summarized the effects of fluids on PMCs and gave six possible water gain profiles: 1) Fickian diffusion curve, 2) a curve corresponds to the often encountered circumstance of a continuous gradual increase in weight gain, 3) a curve represents the so-called “two-stage diffusion” behavior, 4) a curve is sometimes associated with a moving diffusion fiont, 5) a curve corresponds to a rapidly increasing moisture content within the composite, which is usually accompanied by large deformations, damage growth, material breakdown, and 6) a curve accords with weight-loss that is attributable to irreversible chemical or physical break-down of a material. Most commonly, weight-loss occurs in conjunction with hydrolysis, or the separation of side group fi'om the polymeric chains, or the dissociation of matter located at the vicinities of fiber/matrix interfaces. Zhou and Lucas [Zhou 1995] also suggested six types of potential water uptake profiles. The result are similar with Weitsman’s except the case 4 because it is associated with water-vapor induced corrosion of glass fiber. Graphite fiber is water impermeable and does not has this behavor. The detail is in Chapter 6. 2.2.2 Mechanism of Water Absorption It is very important to know the nature of water in epoxy resin at the molecular level because that firndamental knowledge is usefirl for us to understand and explain the 22 salient aspects of hygrothermal phenomena. Although the mechanism of water absorption has been investigated from the initial stage of the hygrothermal effect, to date, the state of understanding is still incomplete. Adamson [1980] investigated thermal expansion and swelling by using epoxy resin Hercules 3501-5 and Gr/Ep composites Hercules AS/3 501 and Fiberite T300/934 prepreg. He suggested that some water molecules form hydrogen bonding with hydrophilic groups in epoxy resin while other water molecules are retained in free volume of the epoxy resin. Apicella et al. [1983] studied the water sorption modes of glassy epoxy resin DGEBA cured with TETA and TGDDM cured with DDS and proposed that there are three absorption modes: (1) bulk dissolution of water in the polymer networlg (2) moisture absorption onto the surface of vacuoles which define the excess fi'ee volume of the glassy structure, and (3) hydrogen bonding between polymer hydrophilic groups and water. If the first two modes occur consecutively, a dual sorption behavior can be determined. Antoon et al. [1981] showed that the fiequency of the in- plane bending mode of sorbed water in epoxy resin lies between its frequency in liquid and flu (gaseous) water, another indication of hydrogen bonding. Browning [Browning 1983] proposed that absorbed water molecules can be combined with a functional group of a highly polar nature in cured epoxy resins as follows: Ii 0 Pi .. O. H O . .— , , K )1 :o : H o’ i H’ ..... R—N—R-m- --..- $H2-C— CH2 ----- .. I _____ R_O....--.. H H 23 Jelinski and collaborators [J elinski 1985] investigated the nature of the epoxy- water molecule interaction using quadruple echo deuterium NMR spectroscopy. They revealed that (1) the water in epoxy resin is impeded in its movement, (2) there is no free water; (3) there is no evidence for tightly bound water; and (4) it is unlikely that the water form disrupts the hydrogen-bonded network in the epoxy resin. The water molecules migrate fi'om site to site, but such a jumping motion does not involve a specific hydrogen- exchange mechanism. Using dielectric experiments, Woo and Piggott [Woo 1987] suggested that the water does not appear to be bound to polar groups in the resin or hydrogen bonding sites. They reported that there was only some clustering of water molecules in the polymer, rather than complete molecular separation. Joncock and Tudgey [1986] found that water absorption depends on the amount of flu volume in the polymer network. They showed that the contribution of polar groups in terms of their hydrogen bonding capabilities is reflected by the effect of meta- chloro, bromo, and methyl substituents on the water absorption of the modified DGEBA epoxy resins cured with DDS hardener, while substituents in the ortho position adversely affect the hydrogen bonding capabilities of amine groups and limit the extent of reaction by steric interference. Also by using O-glycidyl resin systems cured with various amounts of DDS hardener, they proved that free volume plays an important part in determining the level of water absorption. 24 2.3 HY GROTHERMAL EFFECTS ON EPOXY AND GR/EP COMPOSITES 2.3.1 Plasticization and T, Variation The hygrothermal effects of absorbed water in epoxy systems are mainly plasticization, degradation, swelling, and lowered T,. Plasticization is the least controversial issue in all hygrothermal effects. Many works reported mechanical property degradation due to water absorption in epoxy resins and Gr/Ep composites. T, variation has also been studied intensively. Investigations have revealed that absorbed water in epoxy acts as a plasticizer [Shen 1981, Imaz 1991, Biro 1993, Lucas 1993, Wolff 1993]. The effect of water absorption in Gr/Ep composite is manifested by a reduction in elastic modulus and by a decrease in the matrix-dominated strength properties. Although the basic result that 1% of sorbed water may decrease T, by 10-20 °C is widely accepted, interpretation of the T, decrease is varies. Based on the free volume concept of polymeric materials, Kelley and Bueche [1960] derived the following expression for the Tg of a plasticized system: ._ apvngp + ado _ vpfl‘sd 8 Here, or is the expansion coefficient, v is the volume fraction, and subscripts p and d represent polymer and diluent, respectively. This is a classical method for predicting the variation of T, in water-sorbed resin. Several investigators [Cairns 1984, Browning 1978] applied this equation to epoxy resin/water system and explained the dependence of T, on water content. Couchman and Karasz [197 8] extended the classical thermodynamic treatment of composition-dependent T, originally proposed by Gordon et al. [1977] Karasz and 25 coworkers [Brinke 1983, Ellis 1984] modified the equation to apply it successfirlly to epoxy resin/water systems. DeIasi [1978] suggested that water which disrupts the hydrogen bond could depress T,, whereas, water that forms cluster or hydrogen-water type groupings has no measurable effect on T,. Other experimental data [Apicella 1979, McKague 1978, DeNeve 1993, Bellenger 1989] have shown that the T, of epoxy resins or Gr/Ep composites generally decreases as the water content increases. Mijovic and Weinstein [1985] found that the depression of the glass transition temperature in a Gr/Ep composite after water absorption is strongly dependent on the temperature of the environment during water absorption. At the same absorbed water content, the T, depression was greater at higher absorption temperature. They concluded that at lower temperatures, the temperature dependence comes fi'om the water that exists predominantly in the defects at the interface and, to some extent, in the lower crosslink density matrix, while, at a higher temperature, it comes from water penetrating into the highly crosslinked regions within the resin. Zhou and Lucas [Zhou 1994] found that T, does not simply depend on water content. The T, value of a hygrothermally exposed epoxy resin depends on its exposure history. Longer exposure time and higher exposure temperature induce higher value of T, These results are quite different fi'om Mijovic’s results. The detailed discussion can be found in Chapter 4 of this dissertation. 2.3.2 Change of Mechanical Properties Water absorption causes resin plasticization concurrently with swelling and lowering of its T,. These effects usually accompany modulus changes of the material. Shen and Springer [1977] summarized the previous tensile modulus data of composites and concluded that for 0" and rt/4 laminates, there appears to be very little change in the 26 bulking moduli over the entire spectrum of water contents from dry to firlly saturated in the temperature range of 200 to 450 K, whereas, for 90° laminates, the saturated moduli decreased considerably with increases in both the water content and temperature. Using the stress-strain data of various water-treated samples of neat resin and composites, Browning et a1. [Browning 1978] conclude that, as water absorption increased and/or as temperature increased, the tensile modulus of the resin matrix and the transverse modulus of the composite decreased. Crossman et al. [Crossman 1978] determined the tensile relaxation modulus after exposure to equilibrium moisture levels. They found drastic reductions in modulus and an enhanced rate of relaxation at higher temperatures and moisture contents. Dynamic mechanical tests of epoxy resins have been used to investigate the change of molecular structure for various epoxy resin curing agent systems [Chu 1984, Apicella 1979, Kuzenko 1980, Mikols 1982] and to study hygrothermal effects on their molecular structure [Kuzenko 1980, Mikols 1982, Chu 1984]. They have also been applied to reinforced epoxy composites to investigate the effects of reinforcement [Mikols 1980, Lee 1989] and hygrothermal effects [Dynes 1979]. Kuzenko et al. [1980] investigated the dynamic material properties of TGDDM epoxy resins cured with DDS curing agent and boron trifluoride catalyst. They observed a glass transition around 234 °C; the width of the glass transition peak was proportional to the cyclic “viscosity” of the sample. They observed that a sample containing 6.4 wt% water exhibited a significant plasticizing effect in the dynamic shear modulus vs. temperature curve. They also showed a shift towards lower values and a broadened peak of the glass transition. 27 The nature of the interface region between matrix and fiber in a composite influences its mechanical properties. Spathis et al. [1984] prepared an epoxy matrix/fiber specimen and an epoxy matrix/silane—coated fiber specimen and compared the response of their dynamic mechanical properties. They found that a specimen that had weaker interfacial bonding showed lower values of storage and loss moduli and higher values of damping factor. Chua [Chua 1987] characterized the quality of the interfacial adhesion by the value of the damping factor at the T,. He observed that the dynamic modulus at the T, of the glass fiber-reinforced polyester resin increases when the amount of unreacted organosilane at the interface is increased. Banerjee et al. [Banerjee 1990] reported that the poor interfacial bonding in the untreated fiber specimens led to a large viscous dissipation at the carbon-epoxy interface and that because of a weak interface the T, of the untreated fiber specimen was slightly lower than that of the surface-treated fiber specimen. 2.3.3 Swelling Induced by Absorbed Water Swelling denotes volumetric dimensional change due to moisture content alone, independent of thermal expansion. Since water is polar, it is capable of forming hydrogen bonds with hydroxyl groups. Therefore, interchain hydrogen bonds can be disrupted to increase the intersegrnental hydrogen bond length [Kwei 1966]. Rehage and Borchard [1973] showed that water can combine with polymers in the glass state, establishing an equilibrium between the liquid and glass state. Adarnson [1980] considered the swelling efficiency of sorbed water to be temperature and concentration dependent. Sorbed water molecules may either occupy free volume causing no swelling or interrupt interchain hydrogen bonding causing swelling. Kong and Adarnson [1983] also showed the increase in the partial volume of the 28 water experimentally. Hahn [197 6] showed that the absorbed water produces relatively little swelling until a critical amount of water is absorbed, and then the resin sample volume increases proportionally to the additional water content. Moisture absorption in the fiber (i.e., carbon fiber) is negligible. Therefore, Hahn [Hahn 1976] suggested that the transverse swelling strain of the composite could be calculated using: a. = [(1+v..,)/3]d(1\r-M,) where Mo = Vo/d = vmdem/d Here, a. is the swelling strain in the transverse direction, v... is the matrix Poisson’s ratio. M is the moisture concentration, M0 is the minimum value of M at which swelling is observed, V0 is the volume fraction of voids, d is the total specific gravity, and d... is the matrix specific gravity. DeNeve and Shanahan [1993] used DGEBA and dicyandiarnine DDA resin system and concluded that the hygrothermal aging of epoxy leads both to plasticization of the polymer (a physical effect) and chain scission (a chemical effect). Caims and Adams [1984] showed that the moisture-induced linear strain of such an epoxy resin exhibits a linear relation proportionally with the amount of absorbed moisture. Gazit [197 8] showed that the weight increase of moisture was proportional to the linear dimensional change and was the same for all samples having the same reinforcements at all levels of ambient humidity. 2.3.4 Hygrothermal Degradation in Epoxy Resin and Gr/Ep Composites Degradation of composite materials in structural applications may be caused not only by mechanical loading but also by environmental exposure. Most of the environmental degradation in composites is due to their exposure at increased temperature 29 and to humidity. The absorbed water not only plasticizes the matrix resin, but it can also change the state of residual stresses to cause microvoids and/or microcracks in the resin [Wright 1980, Apicella 1979, Shirrell 1978]. Some researchers [Farra 1978, Dewirnille 1983] observed hygrothermally induced cracks in the bulk epoxy resin regions of glass fiber/DGEBA epoxy composites as well as at the interface of the glass fiber and DGEBA epoxy resin. Browning [1978] reported that one significant mechanism for the loss of elevated temperature properties in a moisture-rich environment is the formation and growth of cracks in the material where the crack growth process is aided by localized chemical chain scission at the crack tip. He also found that microcracking increases the equilibrium moisture absorption level. Shirrell et al. [1979] exposed T300/5208 composite samples to several different hygrothermal environments and examined the hygrothermally induced microcracks by scanning electron microscopy (SEM). They found that severe microcracks were observed at 82 °C and that the severity and frequency of these microcracks increases with relative humidity. They also found that postcured specimens generally formed more severe microcracks than identically exposed non-postcured specimens. Morgan and Mones [1980] concluded that sorbed moisture enhances the craze cavitation and propagation processes in amine-cured epoxy resins, whereas the initial stages of failure enhance the accessibility of moisture to sorption sites within the epoxy to a greater extent than in the later stages of failure, which involve crack propagation alone. Hahn [1987] explained that the hygrothermal degradation is the result of matrix plasticization, microvoid formation, and microcracking. Apicella and Nicolais [1985] also observed a Langmuir-type, two-step moisture diffusion in epoxy resins due to the 30 moisture—induced microcavities. Halpin [1985] observed that the rate of the apparent difl‘usion process and the magnitude of the equilibrium weight gain are accelerated by the presence of cracks and voids in a laminate. Leung and Kaelble [1980] predicted that the possible microcracks and delaminations within the composite due to the stress gradient are caused by the moisture content gradient. F edors [1980] found that some inclusions in epoxy resin systems can be dissolved to form cavities when the diffused water reaches the surface of the inclusions. The difference in the chemical potential of the water in the pure state and in solutions, which manifests itself as an osmotic pressure, tends to make the initially concentrated solution in the cavity more dilute. He suggested that this effect is the driving force for the growth of the cavity. CHAPTER III THE NATURE OF WATER IN EPOXY As noted above we know that although hygrothermal efl‘ects and associated mechanisms of epoxy have for practical reasons been investigated for a long time [Loos 1981, DeNeve 1993, Imaz 1991, Brio 1993, Lee 1993, Adarnson 1980, and Lucas 1989, 1993], the basic diffusion mode of water in epoxy resin are still not well understood. The issue to be addressed in this chapter concerns how water interacts with epoxy resins. Typical approaches that have been considered are as follows: 0 Free volume approach, which presumes that water difi‘uses into epoxy resin and resides in free volume in the form of free water or clustered water. Chemical interaction between water molecule and epoxy resin is considered insignificant for this approach [Gupta 1985]. Water does not appear to be bound to polar groups in the resin or hydrogen bonding sites. There was only some clustering of water molecules in the polymer, rather than complete molecular separation [Woo 1987]. 0 Interaction concept, which suggests that water molecules couple strongly with certain hydrophilic groups such as hydroxyl or amine in epoxy resin [Moy 1980]. 0 Combination theory, which considers water molecules exist in epoxy in the form of both free water and bound state. Some water molecules form hydrogen bonding with 31 32 hydrophilic groups in epoxy resin while other water molecules are retained in free volume [Apicella 1985 and Adarnson 1980]. The difference in interpretations of water sorption-efl‘ects is due in part to the reality that the hygrothermal environmental effects in epoxy are quite complex. The other reason is that most previous results used rather limiting testing and probing methods. Furthermore, systematic and integrated investigations have been lacking. Admittedly, there is some general agreement on some models, but, considerable disagreement still persists. There is no single theory with sufficient experimental support to account for all related phenomena. Sorbed water has significant influence on the mechanical performance of graphite/epoxy composites. Knowledge of the natural state of water molecules in epoxy resins is the most important and fundamental issue for the study of hygrothermal efl‘ects of epoxy resins and Gr/Ep composites. A review of previous literature suggests that a sophisticated approach and well designed experiments are necessary to gain knowledge and alley controversy. By investigating and understanding the nature of water in neat epoxy, it allows one to gain insight and results which can be applied subsequently to moisture sorption behavior in carbon/epoxy resin systems. The aim of this chapter is to discern the nature and bonding characteristics of sorbed water in epoxy resins. An integrated experiment is designed and carried out by: a) investigating water absorption and desorption phenomena, and b) conducting nuclear magnetic resonance (NMR) and differential scanning calorimetry (DSC) testing. Through this study a better understanding of the nature of water in epoxy resins is achieved. 33 3.1 EXPERIMENTAL 3.1.] Materials Three kinds of epoxy systems were used in this study. One is diglycidyl ether of bisphenol-A (DGEBA, Shell Epon828m) and metaphenylene diamine (mDPA) epoxy system. The second is tetraglycidyl-4,4'-diaminodiphenyl methane (TGDDM, Ciba Geigy MY720T“) resin with 4,4'-diaminodiphenyl sulfone (DDS, DuPont) hardener. The third is ICI Fiberite 934TM epoxy resin which consists mainly of TGDDM resin, DDS hardener, with small amounts of other additives of which the types and their concentrations are proprietary. The chemical structures of these green materials are shown in Fig. 3.1. The three kinds of epoxies have been used extensively for high performance polymer matrix composites. The material preparation is described as follows. DGEBA and 14.5 phr ( part per hundred resin by weight ) mPDA were heated separately to 75 °C in an oven until mPDA was melted. Then the epoxy resin and hardener were mixed thoroughly and degassed for 10 minutes. The degassed mixture was poured into a mold for curing. The curing procedure was 75 °C 2 h + 125 °C 2 h + 180 °C 8 h + 60 °C. Specimens were taken out of the oven after temperature ramped down to 60 °C. The temperature ramp rate was 1.5 °C/min.. TGDDM and 44 phr DDS were heated to 13 5 °C respectively, then, mixed and stirred until DDS was melted and a clear brown liquid was obtained. The material was degassed for 15 minutes. The curing procedure was: 135 °C (mixing and degassing) + 80 °C 1h + 100 °C 2 h + 150 °C 4 h + 200 °C 7 h + 60 °C and then the specimens were removed fi'om oven. The temperature ramp rate was 1.5 °C/min. 34 The frozen Fiberite 934 epoxy resin was put in a mold and heated to 135 °C. The material was degassed for 10 minutes at 135 °C. The curing procedure was 135 °C 2 h + 177 °C 4 h + 190 °C 6 h + 60 °C and then the specimens were removed from oven. The temperature change rate was 1.5 °C/min. CH3 Ho CH (:\—CH CHz— o‘c‘o— CH;— CH —CH2 DGEBA mPDA /0\ CH2 —CH— CH2 \ /CH2 —CH ~CH2 CH H\2-CH— CHz/ NQ CHZQ’N \CHZ —CH —CH: TGDDM 9 Fig. 3.1 Chemical structures of DGEBA, mPDA, TGDDM, and DDS. Absorbance 35 After the three kinds of materials were prepared, Fourier transform infrared (FT IR) analysis was performed to determine the degree of cure. Epoxide groups have a strong IR absorption peak at 910 cm". During the curing process, the intensity of the 910 cm'1 diminishes gradually. When the epoxy resin is cured completely, all epoxide groups are opened and the 910 cm'1 peak vanishes. If the material is not completely cured it has lower crosslinkage, mechanical properties, and T,. During the hygrothermal exposure firrther crosslink may occur and it will influence the later experimental results 08 I 1 1 1 J l l 1 1 l ’1 L P 1 1 1_j_ 1 j 1 1 1 a 1 1.. 07 - ~ 1593 ~ ' 1105 t 06 ~ 1510 i 1 1134 , i t- . t- 0-5 4 ‘282 1076 t -r H i- . {t 1 U 910 ~ 03 I T I T I r "—1 l T j T T >- 3“. 2.” recs 1200 see 400 -l Wavenumber cm Fig. 3.2 FTIR spectrum of Fiberite 934 epoxy resin after being cured at 177 °C for 2 h. The existence of 910 cm'1 means that the epoxy resin was not cured completely. Absorbance 36 08 07 '1 1. 1593 ~ . 1105 : 0" 1 1510 t i 1134 r 05 - 1282 J1076 L . t 04 q i \ i- “ 1716 r 1 U 910 ~ 03 ‘4‘ 1 fl 1 1 r u 1 r 1 fir 1 t .- Wavenumber cm’l Fig. 3.3 FTIR spectrum of Fiberite 934 epoxy resin after being postcured at 190 °C for 7 h. The nonexistence of 910 cm" means that the epoxy resin was cured completely. significantly. Figure 3.2 shows the IR spectrum of F iberite 934 epoxy resin after being cured at 177 °C for 2 h. The 910 cm'l peak has not vanished completely, which means the material is not firlly-cured. Figure 3.3 shows the IR spectrum of F iberite 934 epoxy resin after being postcured at 190 °C for 7 h. The 910 cm'1 peak has vanished, which means the 37 material is fully cured. The other two epoxy resins showed similar curing characteristics at the 910 cm'1 wavenumber. This test confirmed that all three epoxy resin systems were fully cured. 3.1.2 Water Absorption Test 3.1.2.1 Analysis Consider a plate of thickness h exposed on two sides to the same environment. The plate is taken to be infinite in the y and 2 directions so that the moisture content inside the plate varies only in the x direction and the problem is one dimensional. Initially (time t = 0) the temperature, T,, and the moisture concentration, or, inside the plate are uniform. The plate is suddenly exposed to a moisture environment in which the temperature, T,, and moisture concentration, c., are constant. The objective is to detemtine the moisture distribution, c, and the total moisture content, or, of the material as a firnction of time. It has been observed that the diffusivity, D, changes very little with the moisture content [Aug] 197 5]. At a given temperature, the problem can be described by Fick's diffirsion law [Shen 1981]. 93- 62" (3.1) at 6x2 c=ct O0 (3.2b) Where c is the concentration of the moisture, t is the diffusion time, x is a dimensional coordinate, the thickness of sample is h, c, is the moisture concentration inside the plate before the sample is exposed to a hygrothermal environment and it is unifomr, c. is the moisture concentration outside the plate after the sample is exposed to a hygrothermal 38 environment and it is a constant, and D is the diffusion coefficient (diffusivity), which is temperature dependent. The solution to Eq. 3.1 for the boundary conditions given in Eq. 3.2 was given by Jost [1960]. m . . 2 2 c c, 42 l Sin(2_/+l)7zrcexp[__(21+l)erl =1..— 3.3 c_—c,. 7rj=o(2j+l) h h2 ] ( ) Here c... is the maximum moisture concentration. The total weight of the sorbed moisture in the material is obtained by integrating Eq. 3.3 over the plate thickness. h m = gfc dx ' (3.4) 0 The result of this integration is: 8 .. exp[—(2j+l)zrr3(f7)2’)] 35 =1—7Z (2141), (.) m—m, G: mm _ m1 TI j=0 m; is the initial weight of the moisture in the material and mm is the weight of moisture in the material when the material is fully saturated, in equilibrium with its environment. Experimentally, gravimetric analysis is used to determine the percent moisture content by moisture-induced weight gain of the material. Thus, in practice the parameter of interest is the percent moisture content defined as M = 100 x (weight of moist material - weight of dry material)/(weight of dry material) = 100 x (W-WJ/Wd (3.6) by noting that W = W, + m (3.7) Where m is the weight of sorbed water in the material. Equation (3 .5) may be rearranged in the form, M. = G(M.. -M,) + M, (3.8) 39 Here M. is the percent water gain with exposure time, MIII is the maximum water percentage, and M; is initial water content. Equation 3.8 may be approximated by the expression for practically analytical use [Shen 1981]: Di 0.75 M. =M..{1—expr—7.3120 °C) were required to remove the residual moisture (Fig. 3.13). They considered that the occupied bond sites set initially at the lamination were a significant factor in the absorption process. Xiang [1993] observed thermal spike effects on moisture absorption in epoxy/carbon fiber composite. The desorption curves revealed the presence of a residual weight increase which might arise in a small concentration of unremovable residual moisture, which increased with thermal spiking temperature (Fig 3.14). This behavior inferred that some water was in some way chemically reacted in the composite presumably with the matrix. Therefore the apparent residual moisture content could be in the form of either hydrolysis products or strongly hydrogen bonded water molecules. Obviously 100-120 °C is a critical temperature range. In this study 60 and 140 °C desorption temperatures were selected. The consideration is based on 1) 60 °C is a relative low temperature which is high enough to remove ordinary (unbound) sorbed water will not remove residual (more strongly bound) water simultaneously, and 2) retained water can be removed at temperature of 140 °C without a significant influence on material’s thermal stability. One can see from Table 3 .2 that very small concentration of water in DGEBA + mPDA system could not be removed at this temperature. It is quite evident that, based on our experimental water desorption results and the previous studies, there are two types of sorbed water, as characterized by bonding, in the hygrothermal process. One type is easily desorbed from the material and is classified as a physiosorbed water or ‘A-bonded’ water. In contrast, the other type of sorbed water is more difficult to remove by desorption and, thus is classified as a strong bonded water or ‘F-bonded’ water. A-bonded water is baked out in the first desorption stage. The additional 240 h 55 1.0f f (Goon/Imps DGEBA+mPDA ~ Eiberit‘e'c‘sg 45 degree C 60 degree C .0 0° 75 degree C 90 degree C tttt .0 .o .5 CD Water desorption (nomalized) .0 N . 1 . . 1 . . 1 . 1 1 1 . 1 l r l I I - 010203040 01020304001020304 Time (square root of hour) 0.0 Fig. 3.11 Water desorption profiles of the three epoxy systems at 60 °C for 1450 h, then 140 °C 240 h. The symbols are experimental data and represent the samples with drfi‘erent bath temperatures in the water absorption process. 56 Table 3.2 The amount of retained water in the three water-saturated epoxies after desorption at 60 °C for 1450 h. The trend is that higher immersion temperature induces greater levels of retained water. Retained water Retained water Retained water Retained water of sample of sample of sample of sample bathed at 45 °C bathed at 60 °C bathed at 75 °C bathed at 90 °C then desorp. at then desorp. at then desorp. at then desorp. at 60 °C 1450 h 60 °C 1450 h 60 °C 1450 h 60 °C 1450 h (w%) (w%) (w%) (w%) TGDDM 0.29 0.32 0.34 0.44 + DDS DGEBA 0.11 0.18 0.20 0.25 + mPDA Fiberite 0.27 0.30 0.33 0.39 934 1 Water Absorption r: ‘3 CD 25. <3 3 Water Desorption O Exposure Time Exposure Time Fig. 3.12 Moy’s [1980] water desorption results. 57 Saturationat 85 °C, 50%RH §85°C 100 °C §110°C £120°C Time (hm) 4328 0"" 16 0—’12 momma) JTinu' Ina-militant Fig. 3.14 Xiang’s [1993] water desorption results. 58 desorption at 140 °C is the second desorption stage associated with the desorption of I‘- bonded water. From desorption data, previous studies have suggested the existence of the two difl'erent types of sorbed water but none persuades this matter with firrther as was done in the present study. This study shows a consistent evidence of two types of sorbed water. The rest of this chapter gives more evidence and discussion on this issue by further investigation on desorption behavior and NMR tests. Table 3.3 and 3.4 show the desorption diffusivities of the three epoxy systems. Table 3.3 gives the desorption diffusivity of the three epoxy systems with 60 °C as the desorption temperature. The method to determine the difi‘usivities is the same as for absorption diffusivity (Eq. 3.10). The data indicate that the desorption difl‘usivity (5 - 6 x 10'7 mmZ/sec) is similar to absorption diffusivity at 60 °C temperature (5 - 7 x 10'7 mmzlsec). This suggests that the state of bound water desorbed at 60 °C is the same as that initially absorbed by the material. Table 3.4 is the diffusivity of the three epoxy systems at 140 °C desorption stage. There is no 140 °C absorption diffusivity for comparison. But we know that diffirsivity depends greatly on temperature. Higher temperature results larger diffusivity. Extending the absorption difi‘usivity data, an extrapolative difiirsivity at 140 °C is obtainable. Figure 3.15 shows this result and the value of difi'usivity at 140 °C is in the order of 10" mmzlsec. Compared to the extrapolated data, the experimental desorption diffusivity is much lower (10'° - 10'7 mmzlsec) and even smaller than 90 °C absorption diflirsivity. This result provides evidence that the retained F-bonded water has a much smaller degree of mobility than A-bonded water. That is, the bound state of F-bonded and A-bonded water is dissimilar. 59 Figure 3.16 shows that the amount of F-bonded water increased with exposure time while maintaining a fixed exposure temperature (90 °C). The amount of F-bonded water depends strongly on the exposure time and temperature. Both higher immersion temperatures and longer immersion times create more F—bonded water. Such behavior of the F-bonded water leads us to consider it as chemisorbed water because chemisorption processes depend strongly on exposure time and temperature [A. W. Adarnson, 1990]. Marsh [1988] and Barrie [1984] also reported that the amount of F-bonded water depended on exposure time suggesting that longer the exposure time promoted more retained water. The small increase in the saturation level shown in Figure 3 .9 may be associated with the increase of the retained water with exposure time. Table 3.3 Desorption diffusivity of water in the material desorbed at 60 °C in a dry chamber. Sample Sample Sample Sample immersed at immersed at immersed at immersed at 45 °C then 60 °C then 75 °C then 90 °C then desorbed at desorbed at desorbed at desorbed at 60 °C 60 °C 60 °C 60 °C (mm2 5") (mm2 3") (mm2 s!) (mm2 s“) TGDDM+DDS 6.43 x 10'7 5.86 x 10'7 5.90 x 10'7 5.62 x 10" DGEBA+mPDA 6.22 x 10'7 6.27 x 10'1 6.33 x 10" 5.99 x 10'7 Fiberite 934 5.55 x 10'7 5.66 x 10'7 5.67 x 10'1 5.40 x 10'1 60 Table 3.4 Desorption diffusivity of water in the material desorbed at 140 °C in a dry chamber. Sample Sample Sample Sample immersed at immersed at immersed at immersed at 45 °C then 60 °C then 75 °C then 90 °C then desorbed at desorbed at desorbed at desorbed at 140 °C 140 °C 140 °C 140 °C (mm2 8") (mm2 5") (mm2 s'l) (mm2 s") TGDDM+DDS 1.39 x 10*5 1.67 x 10*5 2.27 x 10“ 1.79 x 10*5 DGEBA+mPDA 1.61 x 10‘7 6.86 x 10'1 9.77 x 10'7 2.01 x 10'7 Fiberite 934 1.45 x 10*5 2.35 x 10*5 2.94 x 10*5 3.06 x 10“ 4 1 0 H TGDDM+DDS M DGEBA+mPDA .5 8- _ 0’0 Fiberite 934 Q 10 = O --------- O extrapolated datum >~ : ~~~~~~~~ a) ........ -. :3 1 ------ ,. E -e 0 10 "g 140 degree C -7 1 0 I I I I I I I I I l 1 I I I 0.0024 0.0026 0.0028 0.0030 0.0032 Fig. 3.15 The extrapolated absorption diffusivity at temperature of 140 °C. 1rr (1/K) 6l 0.5 $0 - E- g '0 G.) .E (U ...-o a) - 010.0 .1.141111111l4 0 400 800 1200 1600 bath time (hours) Fig. 3.16 The amount of chemisorbed water of the epoxy Fiberite 934 changes with immersion time at 90 °C. 3.2.3 F-bonded Water To reveal the nature of the F-bonded water, a solid state NMR test for 1H was carried out. After desorption at 60 °C for 1240 h, all A-bonded water was removed and I'~bonded water was still retained in the epoxies. These samples are designated as semidried samples. Figure 3.17 shows 1H NMR results of (a) dry and (b) semidried TGDDM+DDS samples. The broad peak of the dry sample means all hydrogen or hydroxyl groups in the material bond tightly with the bone structure. The spectrum of the semidried sarnples shows no difference with that of the dry epoxy even though F-bonded water is still in the materials. One might think that the NMR may not detect such a small amount (as 0.4 w%) of water due to the sensitivity limitation of the instrument. To prove that the NMR instrument is sensitive enough to detect such small amounts retained water, 62 0.3 w% water was added in a dry sample and tested its NMR spectroscopy. The result is shown in Fig. 3.17c. An extra peak can be seen clearly in this spectrum. As we know that the formation of F-bonded water needs much longer time. The new peak is associated with the A-bonded water. These results indicate that such a small amount of water is detectable by the instrument and can be clearly identified from its broad NMR background peak. The similarity of the dry and semidried spectra simply implies that F—bonded water bonds tightly with the epoxy network. From the above experiments, it is concluded that F-bonded water is chemisorbed water in epoxy resins. Some investigators [Mijovic 1985] have insisted that the water can penetrate both the macrovoids and microvoids of the epoxies and stay there in physically bonded state. The water molecules in microvoids may be trapped and not easily removed. If there were only desorption results without other more detailed evidences, we would skeptically accept this idea. But we flirther consider the remaining water to be chemisorbed water not only because of the desorption behavior but also because of the desorption difl‘usivity, NMR experimental results, and the hygrothermal effects measurements discussed in Chapter 4 and Chapter 5. As a plasticizer, physiosorbed water decreases the glass transition temperature and degrades mechanical properties. The following two chapters will reveal that the F-bonded remained water contributes to the increase of the glass transition temperature and a slight enhancement in the mechanical strength. These results could not be explained by a simple model involving just physical trapping mechanisms. Moy [1980], Bellenger [1989], and Apicella et al[l985] found that there was a strong interaction between water and specific functional groups such as hydroxyl and 63 amine groups in the polymer resin. F-bonded water is representative of this portion of sorbed water. For a fully cured epoxy there are enough hydroxyl groups for forming bonds with sorbed water molecules. The existence of these large number of and strong polar groups makes epoxy resin be able to wet with variety of substances. Although this H20-OH bonding is not as strong as backbone carbon-carbon bonding (60 -100 Kcal/mol) [Atkins 1993], it is much higher than physical bonding energy of 0.5 - 2 kcal/mol (including Van der Waals bonding, dipole bonding, and interchain hydrogen bonding). Atkins [1990] reported that breaking hydrogen bonding of water requires energy of 2 - 20 kcal/mol [5 kcal/mol for water-water hydrogen bond]. Physiosorbed water dry sample Intensity semidried sample with 0.4 % residual water W sample with 0.3% physiosorbed water WYfiIITTTlIII11IIIIIITII[TillIlTlllllIlllllllllll] 180 160 140 120 100 80 PPm Fig. 3.17 NMR spectra of 1H for TGDDM+DDS system desorption at 60 0C for 1450 h. The results indicate that F-bonded water is tightly bonded with the epoxy network. 3.2.4 A-bonded Water As we have seen in the previous section, F-bound state water has no influence on dry sample NMR spectrum. So any NMR spectrum change of water saturated epoxy comes from A-bonded water. This change comparing with dry sample presents the characteristics of A-bonded water. Three NMR spectra of the TGDDM+DDS epoxy system are shown in Figure 3.18: a) dry epoxy, b) water—saturated epoxy, and c) dry epoxy mixed with liquid water. The peak width of sorbed water is broader than that of liquid water and narrower than that of tightly bonded hydrogen and hydroxyl group in epoxy resin. The mobility of A-bonded water in epoxy is between free-water (liquid) and solid states. Jelinski et. al. [1985] investigated the nature of the epoxy-water molecule interaction using a quadruple echo deuterium NMR spectroscopy and revealed that water in epoxy resin is impeded in its movement. Antoon et. al. [1981] showed their FTIR results and proposed that the sorbed water was held by hydrogen bonding. The process of the A-bonded water diffusing into the resin matrix can be described as follows: the molecules of A-bonded water break the interchain bonds that exist initially in the epoxy resin and form weak hydrogen bonds via Van der Waals type of force. Figure 3.19 is the NMR spectra of Fiberite 934 and it shows how quickly the free (liquid) water is absorbed and changed into impeded A-bonded water. Figure 3.19-a is a spectrum of dry resin mixed with 1% liquid water. Liquid water and a small amount of A-bonded water can be seen clearly. Just ten minutes later, liquid water disappeared and all sorbed water molecules were impeded. This result showing quickly formation of A-bonded water again gives us a strong evidence that the A-bonded water is physiosorbed water. 65 Some studies suggested that there was some bulk dissolution or clustering of water in the epoxy [Woo 1987, Apicella 1985], whereas, others believed there was no free water in the epoxy [Moy 1980, Jelinski 1985]. To resolve this issue, a DSC test was conducted. The suggestion is that, if there is some bulk or clustering water in the epoxy, when the saturated epoxy is cooled below 0 °C for a while, then the temperature is elevated above 0 °C, there will be a DSC heat absorption peak corresponding to ice melting. dry sample / Physiosorbed water water saturated sample Intensity Free water ————- L dry sample with surface adsorbed water A 2 ———— TTIIIII'IIIIIITTjIIIIrrTIIIIIIIIIITllllIIIIIl[11111]] 180 160 140 120 100 80 PPm Fig. 3.18 NMR spectra of ‘H for TGDDM+DDS system. The results indicate that the mobility of A-bonded water in epoxy is between free-water and solid states. 66 Free water ———.- ' " Physiosorbed water Intensity FIVIITVT—ITIIIITIITITTITIITTIII[IIIITTVTTIITIIIIIIITIIIIIIIIII 180 160 140 120 100 80 PPm Fig. 3.19 NMR spectra of F iberite 934 show free (liquid) water changes into impeded A- bonded water with time. (a) dry resin mixed with 1% liquid water and tested. Liquid water and A-bonded water can be seen clearly. (b) Ten minutes later, liquid water disappeared and all sorbed water molecules were impeded. The samples were cooled to -30 °C and isolated for 5 min, then the temperature was increased with ramp rate of 5 °C/min. Water of 0.5, 1, 2, 5 w% was added to the dry epoxy samples and tested by DSC. Every sample has a DSC absorption peak at 0 °C, 67 which corresponding to ice melting. A calibration curve was obtained by measure the height of each absorption peak. The calibration curve is shown in Fig. 3.20. After calibration, the amount of fiee water in the water-saturated epoxies was determined. The DGEBA+mPDA system showed no free water for all samples with difi‘erent immersion temperatures. For the TGDDM+DDS system, the 45 and 60 °C samples showed some free water (less than 0.4 w%) and the 75 and 90 °C samples showed no fiee water. Free water was found in all of the Fiberite 934 samples which were immersed at different temperatures. The DSC trace of the Fiberite 934 is given in Fig. 3.21. The amount of fiee water in Fiberite 934 epoxy after immersion in water for 2200 h is roughly determined by the height of endotherms and the results are shown in Fig. 3.22. The DSC testing results indicate that most of the physiosorbed water in epoxy is dispersed and only a very small amount (if any) of physiosorbed water is in fi'ee-water state. The reason Fiberite 934 samples show much more fi'ee water than the other two resins may be due to different additives in the resin. The schematic drawing shown in Fig. 3 .23 indicates the polymer network before and after water absorption. This schematic (Fig. 3.23) pictorially conceptualizes the bound states of water in epoxy resin based on the findings and interpretation of the study. The concept shows that before water absorption a huge number of interchain hydrogen bonds and a lot of hydrophilic hydroxyl groups exist in the material. After water saturation some water molecules break the interchain hydrogen bonds and form new Van der Waals type bonds with hydrogen. This portion of water is physiosorbed water. The effects of this interchain hydrogen bond break are plasticization and swelling. Some water molecules fornr tight bonds with polar hydroxyl groups and this portion of water is 68 chemisorbed water. Some water molecules may stay in fiee water state and has no bound with epoxy network. E 1.2_ =8 1.0: o 0.8— 5 . E 0.6— a) 1. '0': 0.4— 'c 1. is 0.2— (D _ 8- 0.0L C3.0.2 Water content (%) Fig. 3.20 The calibration curve of the amount of clustered (free-state) water in water saturated epoxy resin. 69 0 .l /... d 45°C A -r— W Z 3 ..2 3 g 60°C LL —t 4.} (0 CD I O U 75 C U7 (3 _2_ 90°C —1 \\ -3 l l T l l l l l -20 -15 -1O -5 O 5 10 15 20 25 Temperature (°C) Fig. 3.21 DSC trace of water saturated Fiberite 934 resin for determining the amount of free water. 7O Free water % 30 45 60 75 90 Bath temperature (degree C) Fig. 3.22 DSC results of fiee water content in Fiberite 934 epoxy afier immersion in distilled water at different temperatures for 2200 h. Bath temperature represents the samples' immersion temperatures in water absorption process. “Sr-H+« 71 o a . \-.; van der Waals bond hygrophilic bond mainchain hydrogen hydrophilic group phsiosorbed water chemisorbed water Fig. 3.23 A schematic conceptualization of epoxy network and bonding characteristics before and after water absorption. 72 3.3 SUMMARY Two types of bonded water are found in epoxy resins by water desorption measurement, desorption diffusivity calculation, residual water determination, time and temperature relations study of the retained water, and the NMR test on the resins at difi‘erent hygrothermal stages. Depending on the water-resin bonding characteristics, they are designated as A-bonded water and F—bonded water. The A-bonded water is a physiosorbed water. The molecules of A-bonded water diffuse into the material and break the interchain hydrogen bonds that exist initially in epoxy resin and form weak hydrogen bonds via Van der Waals type force. Most of physiosorbed water is dispersed interchain hydrogen and only a very small amount of physiosorbed water is in free-water state. The F-bonded water is described as a chemisorbed water which interacts chemically with hydrophilic groups such as hydroxyl groups in the epoxies. The amount of F-bonded water depends strongly on immersion temperature and time. The relations of the two bonded water and the hygrothermal effects such as T, and mechanical property variations in epoxy resins will be discussed in the next two chapters. CHAPTER IV Tg VARIATION OF EPOXY IN HYGROTHERMAL ENVIRONMENT The glass transition temperature (T!) is a very important material parameter for epoxy, epoxy matrix composites, and resin-based composites in general, because T8 determines the service temperature of the material. In most applications, epoxy is required to serve in a glassy state — below T,. Also, Tg’s change after the material is exposed in a hygrothermal environment reflects the degree of plasticization of the material. It is well recognized that absorbed water has a significant influence on the T, of epoxy. The water molecule in epoxy is a plasticizer. A well-know method for predicting T8 of epoxy exposed to a hygrothermal environment is the so-called ”polymer-diluent" model introduced by Kelley and Bueche [Kelley 1960]. This model is based on the semiempirical work of Williams, Landel, and Ferry [Williams 195 5]. Kelley assumed that the diffusing media in polymers does not enter the original fi'ee volume and just stays in the extra free volume created by diffirsed media. McKague [1978] was somewhat indifferent to Kelley’s assumption and introduced a modified model using as. and or... instead of or. = (011., - age) and at. = (or.w - 013“), respectively. The results were believed to be more closely related to the true T3 value of the material. Although this model has been modified the fundamental concept is the same: 73 74 T, is simply the function of water content in the material. That is more sorbed water causes lower T,. This model suggests that the value of T, is independent in exposure temperature and time. In recent years, studies have revealed that water absorption behaviors of epoxy cannot be described by simply using a free volume model [Moy 1980, Apicella 1984, Adarnson 1980, Bellenger 1989, Lelinski 1985, Barrie 1984, DeIasi 1978]. DeIasi [1978] found water in epoxy had different bonding states. He suggested that water that disrupts the interchain hydrogen bond could depress T,, whereas, water that forms cluster or hydroxyl-water type groupings has no measurable effect on T,. Mijovic and Weinstein [Mijovic 1985] found that the depression of the glass transition temperature in a Gr/Ep composite afier water absorption is strongly dependent on the temperature of the environment during water absorption. Our previous work indicated that some graphite/epoxy composites immersed in different temperatures show a very large T, difference (>35 °C), although the water uptake (content) is the same as determined by moisture induced weight gain [Zhou 1994]. The technical literature of reported T, results for a given epoxy exhibits a rather wide scatter band of data [DeNeve 1993, Wright 1981]. Often, the variation may be explained by different material preparation. But, fi'om the above experimental data [DeIasi 1978, Mijovic 1985, Zhou 1994] and by results presented in Chapter 3, hygrothermal history (exposure time and temperature efl‘ect) is expected to contribute to the variation in T, results. To investigate the variation of T, in a hygrothermal environment, a comprehensive study was conducted. Three kinds of epoxy systems used in Chapter 3 were employed in this study. T, was measured at different hygrothermal stages by 75 thermomechanical analysis (TMA). The results and discussions give us a better understanding of the T, variation of epoxy in hygrothermal environments. 4.1 MATERIALS AND EXPERIMENTAL The glassy state of polymer is a supercooled liquid state. Glass transition is a second-order phase transition. Such phase changes are characterized by slope discontinuities in a volume-versus-temperature curve, or, more generally, by finite jumps in first derivatives of volume and entropy such as expansion coefficient, specific heat, and compressibility. The temperature at which such a jump occurs is designated as T,. Generally, corresponding to these different physical responses, three methods are used to determine T,. They are differential scanning calorimetry (DSC), thermomechanical analysis (TMA), and dynamic mechanical analysis (DMA). The DSC test is based on the measurement of the specific heat change. The amount of testing sample is small and usually less than 20 mg. The problem in using this method to test a water-saturated epoxy sample is how to keep moisture in the material. Although the sample is sealed in a pan during DSC testing, some moisture can still evaporate since the material mass is small. Moreover, since the phase transition curve may not be sharp enough, in some circumstances, it is diflicult to obtain a precise transition temperature. In DMA testing, dynamic modulus is measured as a function of temperature. T, can be determined by the intersection of the two tangential lines of the modulus vs. the temperature. To prevent water desorption from the sample during a dynamic test, the sample can be coated with silicon vacuum grease (Lee 1993) or aluminum foil 76 [Chateauminois 1995]. The disadvantage of the DMA method is the requirement of a larger sample (i.e. 40 x 5 x 2 mm) and a relatively slow scanning temperature rate which may induce drying of the sample during the testing. Furthermore, the testing results may be dependent on the experimental condition, i.e. the temperature scanning rate and the strain fiequency [Chateauminois 1995]. TMA tests the expansion change of a sample through the measurement of dimensional variation vs. temperature. T, is determined by the intersection of the two tangential lines drawn along slope discontinuities of the dimensional change vs. the temperature profile. During the process of measurement, water desorption is unavoidable. But the temperature ramp is usually 10 to 20 °C/min., less sorbed water is lost compared with DSC and TMA measurements. The merits of the TMA test are quickness, a small sample (i.e. 5 x 5 mm) requirement, and high sensibility. Consequently, a precise T, result can be conveniently obtained by TMA test. The detail of principle and the methodology of determining T, by TMA were reported by Carter [1978] and McKague [1978]. In this study, all T, results are determined by TMA test at 15 °C/min. DSC is employed to reconfirm some TMA results. The materials used in this experiment are the same as that in Chapter 3. They are 1) tetraglycidyl-4, 4'-diaminodiphenyl methane (TGDDM, Ciba Geigy My7 20) resin with 4,4'-diaminodiphenyl sulfone (DDS, DuPont) hardener; 2) diglycidyl ether of bisphenol-A (DGEBA, Shell Epon828 ) and metaphenylene diamine (mPDA) epoxy system; and 3) Fiberite 934 epoxy system. The details of the material preparation process have been described in Chapter 3. 77 Specimens (5 mm x 5 mm x 1.5 mm) for T, determination were cut fi'om the three kinds of epoxies which were exposed at different hygrothermal stages (water absorption and desorption, different exposure time and temperatures). A TA 930 thermomechanical analyzer ('IMA) with 2200 data acquisition and analysis system was used to determine T,. Relatively low contact loads (500 mg load) were used for the TMA analysis. The temperature ramp rate was 15 °C/min. A TA instruments 2920 modulated DSC with 2200 data acquisition and analysis system was used to test T, and compare with TMA testing method and results. In DSC test samples were sealed in aluminum pans and the temperature ramp rate was 5 °C/min with 50 ml N2 /min gas flow in the testing chamber. 4.2 RESULTS 4.2.1 T, Change with Exposure Time The T, changes of the three epoxy resins with exposure time in the water absorption process are shown in Fig. 4.1 - 4.3. The materials were isothermally immersed in water at 90 °C for 1530 h. The results show clearly that T, increases gradually after the samples saturated by water. During this test process temperature was fixed and the water content of each material kept approximately a constant (Fig. 3 .9). The T, variation is a pure time effect. After 45 h exposure in water at 90 °C TGDDM+DDS samples were saturated and the T, Value was 121 °C. 1500 h later the T, value increased to 132 °C. For DGEBA+mPDA system the change is fi'om 113 to 124 °C. Fiberite 934 shows T, variation from 113 to 126 °C. This T, change with exposure time was also observed in the other lower exposure temperature process. Figure 4.4 shows the T, change of the three epoxy resins with exposure time at 90 °C immersion temperature and Figure 4. 5 78 shows the T, change of the three epoxy resins with exposure time at 60 °C immersion temperature. The upper figure of each T, change profile figure is the corresponding water absorption data of the three epoxies at the same immersion temperature. Both figures (Fig. 4.5 & 4.6) indicate clearly that T, decreases sharply after immersion in water. The minimum value of T, is obtained when epoxy resin just saturates. Then, T, increases with exposure time while the water content of the epoxies remains constant. The T, increase is about 15 °C after immersion in water for 1530 h. A lower immersion temperature required a longer time to saturate and this prolongs the time necessary to reach the minimum T, value for all materials. Samples with 90 °C immersion temperature have higher T, than the 60 °C samples. The 90 °C sarnples' T, profiles look like an entire shift up of the 60 °C sarnples' T, profiles for all of the three epoxies. The range of T, between the minimum value of the 60 °C sample and the maximum value of the 90 °C sample is as great as 35 °C, although they have the same water content M... 79 Overlay V1.00 TA Inst.2200 (°C) Temperature l O (U 40 (wdl afiueua 'wIU vwl Fig. 4.1 T8 change of TGDDM+DDS with exposure time, the exposure temperature was 90 °C. 80 V Overlay V1.00 TA Inst.2200 (°C) Temperature 80 l O m 50‘ 104 -10 (w”) afiueua “mic vwl Fig. 4.3 T, change of F iberite 934 with exposure time, the exposure temperature was 90 0 CL 81 Overlay V1.00 TA Inst.2200 (°C) Temperature l 0 (w”) afiueuo 'wto vwl Fig. 4.2 T8 change of DGEBA+mPDA with exposure time, the exposure temperature was 90 °C. 82 8.00 4.00 I T I I I I Water gain % Absorption, 90degree c (a) l l 1 l 1 l r Glass Transition Temperature (b) Change with Emosure Time 240 t 8 200 8 8’ 3 at l y. 160 120 0 10 20 30 40 Time (square root of hour) Fig. 4.4 T, change with exposure time. Samples were immersed in water at 90 °C for 1530h. The upper figure is the corresponding water absorption profiles at the same temperature. 83 Water gain 96 240 (b) Glass Transition Temperature Change with Exposure Time Tg (degree C) 80 1 J 1 L l I l 0 10 20 30 40 Time (square root of hour) Fig. 4.5 T, change with exposure time. Samples were immersed in water at 60 °C for 1530 h. The upper figure is the corresponding water absorption profiles at the same temperature. 84 4.2.2 T, Change with Exposure Temperatures After immersion in water for 1530 h, all materials reached same moisture saturation irrespective of the immersion temperatures (Mm = 6.8 wt% for TGDDM+DDS, Mm = 3.35 wt% for DGEBA+mPDA, and Mm = 6.95 wt% for Fiberite 93 4). T, variation with immersion temperature was measured and the results are indicated in Fig. 4.6 - 4.8. The results all show the same trend. Samples exposed at higher immersion temperature have higher T,. The variation is quite broad. For the 45 °C and 90 °C immersion samples, the range in T, is 45 °C for TGDDM+DDS; 43 °C for DGEBA+mPDA; and 39 °C for Fiberite 934. The data for Fiberite 934 resin immersion at different temperatures are the same as that for the Fiberite 300/934 graphite/epoxy composite (Chapter 7). DSC testing was conducted to compare these results with the TMA results,. Figure 4.9 shows T, values of DGEBA+mPDA by TMA and DSC testing. The data were obtained using a TA 2200 data acquisition and analysis system. The T, values obtained using DSC and TMA are clearly comparable. However, T, is much easier to be determined by TMA trace since the dimensional change is quite large and the slope discontinuities are sharp. As mentioned the glass transition is a secondary order phase transition. In DSC testing a finite jump of specific heat is observed. This change is relatively small compared to dimensional change. Moreover, in water saturated sample this change may be interfered by a water evaporation endotherrn. It is suggested that TMA testing to determine T, is a better method for water-saturated epoxy. TMA testing is quicker, more convenient, and more precise. 85 160 ay V1.00 TA Inst.2200 C O 1 14 1 Over 58°C (°C) Temperature 26°C 80 10 ___,.___...——-————— 45 °C 1 O O m 50 40" (w”) afiueua 'wru vwr Fig. 4.6 T, change of TGDDM+DDS with bath temperatures after immersion in water for 1530 h. 86 150 7 l 130 Overlay V1.00 TA Inst.2200 T (°C) Temperature 90 (wfl) afiueua 'wtu vwl Fig. 4.7 T, change of DGEBA with bath temperatures after immersion in water for 1530 h. 87 O O (U (\J 0.1.: LOU] <—1C H " <1 1,- D O \—1 > >~ QfU i-VH s—iL C.) (U o > r-t ’ O '\ 1» N V O hm A V" L) O L V CU C. 3 .1.J {U L OJ C]. E O OJ “'0 l— U q—r 0 ("U >- (*1 L0 L) O O O\ ...0 CD 0 O W V >- O I I T I I I I I ‘I' I i I T I (.0 O O O O O O O O l\ LO LO Q m N H (w”) efiueuj 'wtu vwl Fig. 4.8 T, change of F iberite 934 with bath temperatures after immersion in water for 1530 h. 88 -1I A-ad I S I O u. -3- u '0 Q) I L) U') 0-4+ 4 .5—1 1 -6 20 '3 151 3 w i O C 2 010-4 t: H '1 0 <1 E 54 .1 0- 81.801°C -5 . ~ 4 . a . a - 014 O 40 60 80 100 120 1&0 Temperature (°C) 13194.9 T, test results of DSC and TMA. The material is DGEBA + mPDA immersion in water for 1530 h at different temperatures. 89 4.2.3 T, Change at Various Desorption Stages Alter immersion in water for 1530 11, all water-saturated samples were placed in a 60 °C chamber with 0.3 L/min 100 % dry air to begin the water desorption process. Water desorption was allowed to take place over 1430 h. Although most of the sorbed water was removed at the 60 °C desorption temperature, some remained and could not be removed until the desorption temperature was raised to 140 °C (Chapter 3). The amount of retained water has been shown in Table 3.2. Table 4.1 contains T, values obtained at various experimental stages of the moisture efi'ects study: a) as-received dry, b) water- saturated, c) semidried, and d) re-dried. It is interesting to note that at the semidried condition T, is nearly recovered despite a small amount of residual water remaining in the resins. Subsequently, the samples were baked at 140 °C for an additional 240 h and all sorbed water was removed completely. Table 4.1 T, change of the three epoxies in different hygrothermal stages. T, of dry as- T, of sample T, of sample T, of sample desorbed prepared saturated at desorbed at 60 at 60 °C 1450 h & sample 90°C 1530 h °C for 1450 h“ then 140 °C 240 h“ (° C) (°C) (°C) ‘ (° C) TGDDM 251 134 250 251.5 + DDS DGEBA 173 124 173 173 + mPDA Fiberite 218 126.5 218 219 934 * Semidlied desorption condition (60 °C/1450h) ** Re—dried desorption condition (60 °C/1450h + 140 ”C/240h) 90 4.3 DISCUSSION The experimental results indicate that the T, of a water-saturated epoxy depends strongly on exposure time and temperature. When the material first reaches saturation, the depression in T, has the lowest value. With time, however, T, increases gradually. Higher immersion temperature and longer time induce greater T,. Notably, in this experiment all specimens of a given material which were immersed in different environmental chambers showed the same water content. In accordance with the free volume theory, which supposes that the T, value is determined simply by the amount (weight fraction) water in the material. Pertaining to the free volume theory, all specimens, regardless of the exposure temperatures and time, would possess the same value of T,. The difference between the experimental results and the free volume theory is evident. According to free volume theory T, value of water saturated epoxy T,,... can be calculated by using the two equations below [Kelley 1960]: __ a.V.T.. +a..(l-V.)T,.. 3““ " a,V, +a,(1-V,) (4.1) V _ l ' l+0.01Mm(p,/p,,) (4.2) Here, p, = density of dry epoxy; M. = equilibrium water content; V. = volume fiaction of epoxy; cc. and or,. are rubbery and glassy linear thermal expansion coeflicient respectively determined by TMA dimensional change trace; epoxy volumetric expansion coefficient (1.. = 3(ou. - or”); T, of water T,w = 4 °C; water expansion coefficient ctw = 4 x 10'3/°C; and density of water p. = 1 g/cm3 [Browning 1978]. All the experimental data are listed in Table 4.2. The calculated Tm. of the three epoxies are shown in the last row of Table 4.2. The difference in calculated T,.... which is 91 based on the "polymer-diluent" model and the experimental T, is as large as 40 °C (Fig. 4.1 - 4.8). This model may be applicable for some polymers in which there are no strong polar groups such as hydroxyls, however, polymer-diluent model is invalid for predicting T, of water saturated epoxy resin where a dual-mechanism behavior dominates. Table 4.2 Experimental data of the three epoxies. T, of dry epoxy T,.; density of epoxy p.; equilibrium water content Mm; volume fi'action of epoxy Ve; rubbery and glassy thermal expansion coefi'rcient cc. and cr,,; epoxy volumetric expansion coefficient orc = 3(a,. - ct“); and calculated T, of water-saturated epoxy T,.... TGDDM + DDS DGEBA + mPDA Fiberite 934 T, (°C) 251 173 218 (dry, as prepared) p. (g/cm3) 1.26 1.19 1.28 M..(%) 6.8 3.35 6.95 v. 0.9211 0.9617 0.9183 6.. (cm/cm °C) 6.7 x 10" 1.45 x 10“ 8.4 x 10‘5 05,, (cm/cm °C) 1.86 x 10'5 2.8 x 10" 2.4 x 10'5 cl.(cm3/cm3 °C) 4.5 x10'4 6.1x10" 5.4x 10" Tom (° C) (calculated) 144 120 133 Since the previous models do not provide a satisfactory explanation of T, variations observed for hygrothermally exposed epoxy resins, a new interpretation is 92 proposed in the present study. The combination of the experimental T, variation results implies that there are two types of bound water. The water molecules associated with A- bonded water (physiosorbed water) diffuse into the material and break the interchain hydrogen bonds which exist initially in epoxy resin and form weak hydrogen bonds via Van der Waals force. The consequence of the interchain hydrogen bond breakage is the increase in chain mobility and the sharp decrease in T,. As more water is absorbed — more hydrogen bond breakage occurs resulting in a precipitous drop in T,. Provided only the physiosorbed water existed in the resin materials of this study, T, would have the same value for all hygrothermally-exposed sarnples. This implies that T,, determined experimentally, will be unaffected by exposure temperature and exposure time duration. That T, only depends weight fi'action of water uptake and not on exposure history (exposure time and temperatures) suggests that water exists in a single bond state in the resin. Data from the current study refuse this. The effects of chemisorbed water (F-bonded water) on the value of T, is not so direct. We have found that (1) the amount of F-bonded water increases with immersion time (Fig. 3.16), and (2) for a fixed exposure time, higher immersion temperature creates more F-bonded water (Table 3 .2). The observation that both T, and F-bonded water increase with longer exposure time and higher temperature in the water absorption process indicates that “F-bonded water” plays an important role in T, change. A plausible mechanism is that the F-bonded water fomls a secondary crosslink with hydrophilic groups such as hydroxyl in an epoxy network [DeNeve 1993] and this crosslink density increase contributes to T, increase. Hence, the experimental T, value is influenced by a two-mechanisms process involving physiosorbed (A-bonded) water that decreases T, via 93 breakage of interchain hydrogen bonds and chemisorbed (F-bonded) water which increases T, via formation of a secondary crosslink. The T, variation can be described as follows. When the epoxy initially reaches saturation, T, reaches its lowest value (Fig 4.4 and 4.5) since the partitioning of water in the material, at this instant, is such that the amount of A—bonded water is maximum and the amount of F—bonded water is minimum. With exposure time and temperatures more F—bonded water is formed. The F-bonded water creates greater secondary crosslink density and, thus, an increase in T,. After A-bonded water is removed in the desorption process, F-bonded water is still retained in the resin and the T, is recovered completely. This phenomenon should not be explained as that the F-bonded water has no effect on T, without the presence of A- bonded water. Afier physiosorbed water is removed, interchain hydrogen bonds recover quickly and this recovery cause the reversion or restoration of T,. A possible reason why one does not observe the T, increasing caused by A-bonded water at semidried stage is that the secondary crosslink effect on T, is relatively weak when all interchain hydrogen bonds are recovered. In some cases, T, increase induced by F-bonded water at the semidried stage can be observed (Chapter 7) [Zhou 1994]. This interpretation can be applied on previous works efliciently. DeIasi [1978], Moy [1980], and DeNeve et. al. [1993] reported T, test results for epoxy resin. Their experiments were carried out at constant temperature and different relative humidity. They found that T, was lowered with water uptake in the material. The explanation is that since they tested the specimens with the same exposure temperature and time and different relative humidity, all of the specimens should have the same amount of F-bonded water because P-bonded water is controlled by exposure temperature and time. The only 94 difference is the amount of physiosorbed water. Higher relative humidity creates conditions for more physiosorbed water in the resin and, thus, lower T,. 4.4 SUMMARY The variation of T, of epoxy exposed to a hygrothermal environment can be summarized as follows: 1) the change in T, does not simply depend on the sorbed water content in epoxy resin, 2) T, depends on the hygrothermal history of the materials, and 3) for a given epoxy, longer immersion time and higher exposure temperature induce higher T,. Both physiosorbed water and chemisorbed water have influence on T,. The experimental T, value represents the combined effect of the two mechanisms. Physiosorbed water causes a decrease in T, via the breaking of original interchain hydrogen bonds and the increasing chain mobility of the epoxy. Chemisorbed water causes a T, increase due to the possible formation of a secondary crosslinking with hydrophilic groups in the epoxy resins. CHAPTER V SWELLING AND MECHANICAL PROPERTY CHANGE OF NEAT EPOXY IN HYGROTHERMAL ENVIRONMENTS In many industrial and commercial applications, the dimensional stability and mechanical property change of epoxy command considerable attention. This is particularly the case when neat resin is subjected to a hygrothermal environment which can cause the resin to swell and degrade the mechanical properties of the material. Most of the related studies have focused on water effects in Gr/Ep composites and fiber-resin interface [Biro 1993, Lee 1993, Lucas 1993, d’Aimeida 1991, Frassine 1994] and not on the neat resins. Since the effect of water absorption in epoxy-matrix composite is manifested by a decrease in the matrix-dominated properties, before a comprehensive study is conducted on Gr/Ep composites, it is necessary to characterize the swelling and mechanical property change that occur to neat epoxy in a hygrothermal environment. Swelling defines dimensional change due to moisture absorption, independent of thermal expansion. Swelling occurs when interchain hydrogen bonds are disrupted by sorbed water causing the intersegrnental hydrogen bond length to increase [Kwei 1966]. Previous investigations on resin swelling can be summarized mainly by two different approaches. One is suggested by Adarnson [1980] who considered the swelling efficiency of sorbed water to be temperature and concentration dependent. Sorbed water molecules 95 96 may either occupy fi'ee volume, causing no swelling, or may interrupt interchain hydrogen bonding to cause swelling. The other approach is suggested by Hahn [1976] and others. They found that the absorbed water produces relatively little swelling until a critical amount of water is absorbed, and then the resin sample volume increases proportionally to the additional water content. A number of investigations have revealed that sorbed water in epoxy acts as a plasticizer [Imaz 1991, Biro 1993, Lucas 1993]. After exposure in a hygrothermal environment for a period of time, all mechanical properties, including modulus, strength, and failure strain, of epoxy resin are degraded. Usually, the mechanical property variations observed are dependent on water content and exposure temperature [Browning 1978, Shen 1977]. Data on swelling and mechanical degradation are obtained by placing material in different relative humidity (RH) environments long enough for the equilibrium moisture uptake to be reached. The general results are that more sorbed water and longer exposure time induce greater mechanical property degradation. Notably, there are few reports on mechanical property variation of the materials with the same moisture uptakes but exposed to different exposure temperatures. Similarly, there is no compelling report emphasizing mechanical property variations at different hygrothermal stages including water saturated and partially water saturated stages. ‘ In this chapter, swelling and mechanical properties variations under the conditions of same water content and difi‘erent exposure temperatures were thoroughly investigated. The purpose is l) to investigate swelling behavior of epoxy samples that have same water content achieved via different exposure temperatures, 2) to show that if the two types of bonded water have difl'erent effects on the extent of swelling, and 3) to demonstrate how 97 mechanical properties vary for materials at different hygrothermal stages (i.e. dry, saturated, semiodlied, and re-dried epoxy). 5.1 MATERIALS AND EXPERIMENTAL PROCEDURE As in previous chapters, the three epoxy systems investigated were: TGDDM+DDS, DGEBA+mPDA, and Fiberite 934. Specimens used to assess dimensional change due to moisture induced swelling were fabricated to the dimension of 38 x 25.4 x 1.15 mm. Tensile test samples with dimensions of 89 x 12.7 x 1.5 mm were subsequently machined into a dogbone contoured shape with gauge length and width of 38 mm and 6.35 mm, respectively. Flexure test samples were cut in dimension of 63.5 x 6.35 x 1.5 mm. The hygrothermal exposure is the same as described in Chapter 3. The dimensional change of specimen length (38 mm dimension) were measured periodically using a micro-caliper with a 0.025 mm resolution, and, then, the percent dimensional change was determined. Each dimensional change result was an average of three samples. The measurement error is about 0.13% [(38.025-37.975)/38x100%]. Tensile and flexural testing was conducted on the samples that had been exposed for various periods of time at difi‘erent temperatures. An Instron 4302 mechanical test machine was employed to conduct mechanical tests. For the tensile test, the crosshead speed was 0.5 mm/min. For the three-point bending flexure test, the crosshead speed was 1 mm/min and the span was 25.4 mm. Each mechanical test result was an average of five samples. The standard deviation was obtained directly fi'om test report. 98 5.2 SWELLIN G A schematic of the experimental absorption and desorption process is shown in Fig. 5.1 to help understanding the different hygrothermal states. The diagram shows the hygrothermal exposure process described in Chapter 3. Before the material is placed in an environmental chamber for water absorption, the sample is referred to as a dry sample. After the material is saturated by water, the sample is referred to as a saturated sample. By the end of physiosorbed water desorption process, physiosorbed water is baked out and only chemisorbed water is retained in the epoxy. This sample is referred to as a semi- dried sample. After 140°C desorption, sorbed water is removed and the sample is referred to as a re-dried sample. dry water saturated semi-dried re-dlied '5 water absorption CYCIC water desorption cycle a; desorption desorp. g at 60 °C at 140 °C (stage 1) (stage H) X e 1530 2980 3220 Time (hour) Fig. 5.1 Schematic of experimental water absorption and desorption process. Dimensional changes of the three epoxy systems with immersion time and temperatures are shown in Fig. 5.2. The characteristics of swelling in water absorption process are: 1) swelling occurs from the beginning of the water absorption process, 2) 99 TGDDM+DDS 2.0m H 45 degree C . H 60 degree C 1.6“ P1 75 degree 0 DGEBA+mPDA ‘ Q—ro 90 degree C Dimensional Change (%) .2; 1 Fiberite 934 0 10 20 30 40 Square root of hour Fig. 5.2 Dimensional change with immersion time and temperatures. 100 swelling profile is similar with its corresponded water sorption profile (Fig. 3.9), and 3) for a given epoxy resin, specimens have similar maximum swelling regardless the exposure temperatures. Afier saturation, samples exposed at different temperatures have the same water uptake and the water induced dimensional change is the same. Hahn [Hahn 1976] showed that the absorbed water produces relatively little swelling until a critical amount of water is absorbed, and then the resin sample volume increases proportionally to the additional water content. Figure 5.3 shows an essentially linear relation of swelling and water gain. The result means that swelling is proportional to water gain. Table 5.1 indicates the results of dimensional change in different hygrothermal stages for the epoxies that were immersed in water at 45 °C for absorption and then desorption at 60 °C and 140 °C. When material is water saturated swelling reaches the maximum value. On semi-dried stage the dimensional change percentage is not zero and on the re-dried stage the dimension is recovered. This result indicates that both A-bonded water and F-bonded water influence on the dimensional change. To reveal the degree of the influences of the two bonded water on swelling, every one percent physiosorbed water and chemisorbed water induced swelling percentage is calculated respectively. All original data and calculated dada are listed in Table 5.2. These data show the influence of the two bonded waters on swelling. It is quite evident that both types of bonded water affect swelling but physiosorbed water causes greater swelling and chemisorbed water has less efl‘ect on swelling. 101 lllllll ...LAN N030 llllljl .9 co Swelling, % 0.4- , H TGDDM+DDS O/o Fiberite 934 Water gain, % Fig. 5.3 Dimensional change with sorbed water content. Table 5.1 Dimensional change at different hygrothermal stages. The samples were immersed in 45 °C water chambers for water absorption. Water desorption was conducted at 60 °C for 1450 h and 140 °C for 240 h. 934 swelling of swelling of swelling of swelling of dry sample saturated sample semidried sample redried sample (%) 1%) (%) (%) TGDDM 0 1 .829 0.044 0 + DDS DGEBA 0 0.767 0.008 -0.013 + mPDA Fiberite 0 l .928 0.041 0 102 Table 5.2 The influences of A-bonded water and F-bonded water on swelling. Material TGDDM/DDS DGEBA/mPDA Fiberite 934 Property Total sorbed water (w%) 6.8 3.35 6.95 A-bonded water (w%) 6.52 3.23 6.68 F—bonded water (w%) 0.28 0.12 0.27 Total swelling (%) 1.829 0.767 1.928 A-bonded water induced swelling 1.785 0.744 1.887 (%) F-bonded water induced swelling 0.044 0.023 0.041 (%) Ratio of A-bonded water induced 0.374 0.23 0.282 swelling/A-bonded water Ratio of F—bonded water induced 0.157 0.19 0.152 swelling/F-bonded water 5. 3 MECHANICAL PROPERTY CHANGE 5.3.1 Mechanical Property Change in Water Saturation Stage In this section we will describe mechanical property variation of water-saturated (with constant water content) epoxies with different exposure temperatures. Hygrothermal efi‘ects on tensile and flexural moduli are shown in Fig. 5.4. The moduli of Fiberite 934 decrease after immersion in water at difi‘erent temperatures for 1530 h. For Fiberite 934 the tensile and flexural moduli of water saturated sample are about 15% and 17% lower than that of the dry sample respectively. Exposure temperature (range fi'om 45 to 90 °C) shows no influence on the degradation. The DGEBA+mPDA system has much lower saturated water content (3.35%), and no notable modulus (both tensile and flexural) difference is observed between dry and saturated samples. 103 Both Fiberite 934 and DGEBA+mPDA exhibit tensile strength decrease, but Fiberite 934 shows a much greater decrease (Fig. 5.5). Before being exposed to a water environment, F iberite 934 has higher tensile and flexural strength than that of DGEBA+mPDA. After saturation both tensile and flexural strength are lower than that of DGEBA+mPDA since F iberite 934 has much higher water saturation value (more than two times). The decrease of tensile and flexural strength of F iberite 934 is about 25 - 56% and 15 - 44% lower than that of the dry sample respectively. Notably, the degradation is temperature dependent. In the temperature range of 45 - 90 °C strength decreases with exposure temperature increase. Compared with the 45 °C sample, the strength of 90 °C sample is approximately one third lower. Moisture has little influence on DGEBA+mPDA flexural strength and the tensile strength shows 15 - 20% decrease. The most significant hygrothermal effect for any kind of epoxy is the large decrease in failure strain (Fig. 5.6). The failure strain decrease is strongly exposure temperature dependent. The failure strain of 90 °C hygrothermally exposed Fiberite 934 and DGEBA+mPDA is less than half of the dry sample’s value. As observed by ESEM (Chapter 3) there is no visible crack and void in the resins after hygrothermal exposure for 1530 h. A notable change for all samples during the water sorption process is discoloration. The as-prepared samples are light brown or light yellow. After long time exposure, the color of all samples became darker and darker gradually. Higher exposure temperature resulted in darker color. The 90 °C samples almost became black. This phenomenon may be explained by chain scission. DeNeve [1993] studied water absorption in epoxy resin (DGEBA+DDA) by mechanical and IR test. He found some IR peaks changed with hygrothermal exposure process. A plausible chain scission mechanism 104 was given in Fig. 5.7. Notably, the epoxy resin he used is different fiom this study. Browning [1978] investigated Gr/Ep composites (TGDDM+DDS matrix) in high humidity environments. He pointed out that chemical chain scission of any substantial magnitude can be eliminated but very localized chemical chain scission in such areas as crack tips cannot be ruled out. Also, during water desorption process the network collapse due to water removing can cause chain scission. Owing to the chain scission caused polymer network change the mechanical properties will have significant influence and the T, value may also be influenced. But from this study the influence in T, is limited because T, can be recovered completely after the material is redried. Some epoxy shows great water absorption ability such as 6.95 % for Fiberite 934 and some shows small water absorption ability such as 3.35% for DGEBA+mPDA Hygrothermal effects on mechanical properties are quite different for different epoxies. Mechanical properties such as modulus, strength, and failure strain of epoxy that has high water absorption decrease significantly in a hygrothermal environment. In contrast, less influence on the mechanical properties can be found in the epoxy with low water absorption. Some high water absorption epoxy has higher strength in dry sample and shows lower strength than low water absorption epoxy after saturation. For the choice of material which serves in a high humidity or marine environment, the first consideration should not be the strength of a dry sample but its strength after water absorption. 5.3.2 Mechanical Property Change in Different Hygrothermal Stages The mechanical property changes in different hygrothermal stages are shown in Tables 5.3 - 5.7. The three kinds of epoxy samples were placed in 90 °C water chamber for water absorption and then desorption at 60 and 140 °C in a 100 % dry chamber. 105 g. 2100 :lry sample i— DGEBA+mPDA E; . i Q:— Fiberite 934 g 1800 _ o E _ g 1500 — 0) C - ,9. 1200 1 22 1 1 1 L A 6000 2: g. dry sample v 5000 —1 (D 2 _ -. 8 4000 — O : E I i 3 + an E 3000 — i g _ I F r— ‘E H a) 1 45 60 75 90 Bath Temperature Fig. 5.4 Change of modulus at different bath temperatures. 106 A 120.0 a: to t ‘21 1 —§— DGEBA+mPDA E r ' 7 @— Fiberite 934 a) 80.0 - C a) .- .1: (I) g .— g 40-0 dry sample 1' I n l l 1 1 Eg- a § 160.0 — :E c» _ 8 120.0 _ ‘0 ‘ 1 Ti" 1 8 80.0 _— L1- 1 n 1 l 1 1 45 60 75 90 Bath Temperature (degree C) Fig. 5.5 Change of strength at different bath temperatures. 107 —:B_— Fiberite 934 % 5 l g -C_— DGEBA+mPDA, .3 N 1 ..x O 1 O) I Tensile Strain (%) co 1 A l —__.1'j. dry sample I n I l 1 45 60 75 90 Temperature (degree C) N Fig. 5.6 Change of fracture strain after saturation at different bath temperatures. &; r R. on- r a; C/OR' N-C=N-R \N—C—NHR N/\ / \ / / CH2 0 8 OH CH2 (0 CH2 \ / -——>— \ / -—+ \ CIH (.314 CHNHR CH,R _ CH,R j CH2? 1- H\ 13“" I 0 N=C—N—R' N—C—N—R' ,_ _ ~ / \ 2 / 2 /Nl-‘ c OR CH, 0 R OH CH, C CH, CH CH Cit-{NR5 C1128 _. CHZR _l CHZR Fig 5.7 Explanation of a Chain scission in TGDDM+DDA system [DeNeve 1993]. 108 Saturated samples show lowest mechanical values in all hygrothermal stages. After desorption at 60 °C (the samples are called semi-dried) the modulus was recovered and shown to be even higher than that of dry samples (Table 5.3 and 5.4). This increase is quite evident in the DGEBA+mPDA system. The modulus was recovered completely after all sorbed water was baked out at elevated temperature (the samples are called re- dried). The effect of water on the modulus of epoxy resin is reversible. Water saturated samples show lower tensile and flexural strength. After 60 °C desorption, the flexural strength recovers completely and the tensile strength shows partial strength recovery (Table 5.5 and 5.6). When all sorbed water is desorbed, tensile strength still cannot recover firlly. The strength of the re-dried sample is slightly lower than that of semidried samples. The strength of semi-dried specimen is higher than that of redried specimen. The greatest influence of sorbed water on epoxy is the decrease in failure strain (Table 5.7). After saturation the failure strain is less than half of the dry sample's failure strain. In semi-dried stage, the fi'acture strain shows partial recovery. For re-dried samples, the failure strain is about two-thirds of the dry sample for F iberite 934 and TGDDM+DDS system and just half for the DGEBA+mPDA system, although it shows less water-induced degradation on modulus and flexure strength. The decrease is strongly temperature-dependent and irreversible. The permanent strength and failure strain degradation is considered to be due to chain scission in water environment exposure [DeNeve 1993]. 109 Table 5.3 Tensile modulus change in different hygrothermal stages. temperature of all samples was 90 °C. Water absorption Tensile Tensile modulus Tensile modulus Tensile modulus modulus of dry of saturated of semi-dried of re-dried samples samples samples samples Ma) Ma) Ma) Ma) TGDDM 1885 :l: 39 1759 i 50 1969 :t 102 1867 :l: 45 + DDS DGEBA 1448 i 90 1388 :1: 59 1652 i 15 1480 i 45 + mPDA Fiberite 1861 :t 43 1565 :t 23 1872 i- 108 2021 :1: 97 934 Table 5.4 Flexure modulus change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. 934 Flexure Flexure modulus Flexure modulus Flexure modulus modulus of dry of saturated of semi-dried of re-dried samples samples samples samples (MPa) (MPa) (MPa) (MP8) TGDDM 4096 :l: 101 3350 i 59 4097 j: 112 4145 :l: 95 + DDS DGEBA 2662 :t 99 2697 i 97 3069 :l: 98 3297 i 64 + mPDA Fiberite 4130 i125 3324 i 66 4077 i 146 4186 i 100 110 Table 5.5 Tensile strength change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. Tensile Tensile strength Tensile strength Tensile strength strength of dry of saturated of semi-dried of re-dried samples samples samples samples (MPa) (MPa) (MPa) (MPa) TGDDM 96.5 i 5.4 47.6 i 2.4 77.9 1: 9.7 73.8 :1: 6.1 + DDS DGEBA 80 :t 6.9 55.9 i 5.0 60 i 3.5 57.3 :1: 2.8 + mPDA Fiberite 93.1 i 8.1 38.6 :1: 3.7 88.3 i 16 80.0 :t 7.0 934 Table 5.6 Flexure strength change in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. 934 Flexure Flexure strength Flexure strength Flexure strength strength of dry of saturated of semi-dried of re-dried samples samples samples samples (MPQ (MPa) (MPa) (MP9) TGDDM 117: 15 86.9.t6.3 157112.54 159:8.6 + DDS DGEBA 141 :1: 8.5 139 d: 3.6 149 i 5.9 88.3 :t 2.0 + mPDA Fiberite 162 :1: 14.8 90.3 :1: 11.8 179 i 26 138 :1: 4.0 111 Table 5.7 Tensile failure strain in different hygrothermal stages. Water absorption temperature of all samples was 90 °C. Failure strain Failure strain of Failure strain of Failure strain of of dry samples saturated samples semi-dried samples re-dried samples (%) (%) (%) (%) TGDDM 6.1 :t 0.70 2.4 1: 0.13 5.0 i 0.88 4.8 i 0.50 + DDS DGEBA 10.2 i 2.1 5.2 i 0.73 4.4 :1: 0.50 5.0 i 0.25 + mPDA Fiberite 6.7 i: 0.94 2.6 :1: 0.29 6.0 i 1.30 4.9 i 0.30 934 5.4 SUMMARY Swelling in epoxy is basicly proportional to water gain. Experimental results show that both physiosorbed water and chemisorbed water have influence on swelling. But effect of chemisorbed water is smaller than that of physiosorbed water. Some epoxy shows large water absorption ability and some shows lower water absorption ability. Water environmental effects are quite different for these different epoxies. Epoxy with large water absorption ability exhibits a significant decrease of both modulus and strength when epoxy is water-saturated. Low water absorption epoxy shows no significant modulus change with a slight decrease of strength. The most significant hygrothermal efl‘ect for all kinds of epoxy is the large decrease of failure strain, a change that is irreversible. Also, the failure strain decrease is exposure temperature dependent. The cause may be due to chain scission during the hygrothermal process. In choosing an 112 epoxy that serves in high humidity or in a marine environment, the primary consideration should be the nature of water absorption. CHAPTER VI DIFFUSION OF WATER IN GRAPHITE/EPOXY COMPOSITES The results of hygrothermal effects in neat epoxy resins have been shown in previous chapters. In this chapter the behavior of water in Gr/Ep composites is investigated. The environmental effect of moisture on polymer matrix composites (PMCs) is of significant interest, particularly since mechanical and physical property modifications are generally manifestations of its presence. Many investigations regarding moisture diffusion in PMCs have been conducted [Blikstad 1986, 1988, Avaen 1988, Morii 1993, Whitney 1978, Loos 1981, Bonniau 1984, Shirrell 1978, Imaz 1991, Lucas 1993] that address various aspects of moisture-induced behavior. The non-Fickian difl'ilsion behavior of moisture has been reported in PMCs by numerous investigators [Whitney 1978, Loos 1981, Bonniau 1984, Shirrell 1978, Imaz 1991, Lucas 1993, Springer 1981]. The experimentally determined profile of moisture gained versus exposure time often deviated (higher or lower) from the theoretical Fickian curve. Some diffusion profiles indicate a continuous gradual increase in weight gain, never attaining equilibrium [Grayson 1985]. Some profiles do not directly reach an equilibrium water content M, and its profile is like a stair- shaped curve, which is called two-stage diffusion [Gupta 1985]. These phenomena were usually considered as that the epoxy undergoes a slow relaxation whose rate is much less than that of difl’usion. But 113 114 other research showed that it is because of the cracking developed in hygrothermal exposure [Browning 1978, Loos 1981]. Weight loss has been observed fiequently. Researchers believe that it is attributable to irreversible chemical or physical break-down of a material. Most commonly, weight loss occurs in conjunction with hydrolysis, or the separation of a side group from the polymeric chain, or the dissociation of matter located at the vicinities of fiber/matrix interfaces. But direct evidence is lack. Many factors can influence material’s difliision behavior. The variation may be caused by material preparation, curing process, and non-stoichiometric ratio of resin/hardener. The question is that for a stoichiometric and completely cured epoxy system the diffiision behavior is Fickian or non-Fickian. If anomalous behavior is observed how to interpret it. To better comprehend difi‘usion behavior in Gr/Ep, it is essential to pursue both extrinsic and intrinsic aspects of moisture—induced diffusion behavior. Earlier investigations measured the change in weight versus time experimentally to obtain the moisture diffusion profile, which was then compared to the theoretical Fickian diffusion curve. However, some studies did not give the original material preparation and some studies just used limited experimental methods to probe this complicate issue. No published work was found that deals extensively with weight change, dimensional change, and surface modification of Gr/Ep simultaneously. In this study, moisture absorption behavior in T3 00/934 Gr/Ep composite was investigated. Specimens were immersed in distilled water at 45, 60, 7 5, and 90 °C for more than 9000 hours. Weight gain and water- induced expansion were measured, while SEM microscopy was used to investigate the surface modifications. A crack/mass—loss model is presented which yielded a remarkably 115 good description of the behavior of graphite/epoxy composites subjected to a water environment. 6.] MATERIALS The material used in this study was unidirectional T300/934 Gr/Ep composite laminate received from ICI Fiberite. The Fiberite 934 epoxy resin consists mainly of tetraglycidyl-4,4'-diaminodiphenyl methane (TGDDM) resin, 4,4'-diaminodiphenyl sulfone (DDS) hardener, with small amounts of other additives. The chemical structures of TGDDM and DDS are in Fig. 3.1. The curing temperature of a T300/934 laminate is nominally 177 0C. The laminate layup consisted of 18 plies of unidirectional prepreg (013). They were pressed together in a vacuum bag using a cloth on the two outer surfaces that caused a resin-rich layer that had the appearance of a woven pattern. The fiber volume fraction was 66 percent as measured by the optical numeric volume fraction analysis (ONVFA) method developed by Waterbury and Drzal [1989]. The nominal axial, E", and transverse, En, moduli are 19 Msi and 1.5 Msi, respectively. Poisson's ratio, 012’ is 0.28. Typical thermal expansion coefficient values, an and an, are -0.4 x 104; °C'1 and 18 x 1046 °Cl. The matrix-dominated, transverse moisture expansion coefficient, 02,, is 20 x 10'2 %Ml, an order magnitude greater than B”, the axial moisture expansion coefficient. The matrix resin of as-received composite was tested by FTIR. The result is in Fig. 6.1. The spectrum shows that the epoxy group characteristic peak 910 wavenumber is vanished and the material is fully cured. 116 ..ItltIo.l¢lQele . .4: » 00:01:37.1. .I!.Ool o0!0t.¢.t0¢11r.9.. 3.1007. «ole ....u.w.....w..p ....rfc.vl‘0lac .. .IIIIIIaII. Wavenumber rite '1‘300/934 i ived F of as-rece resrn composite. It shows that the material is fully cured. Fig. 6.1 FTIR spectrum of the ma ' 117 6.2 EXPERIMENTAL Specimens having the dimensions of 40 x 15 x 2.3 mm were cut fi'om the laminate with a diamond saw. The edges of the specimens were subsequently ground using 600 grit abrasive paper to maintain consistently smooth edge surfaces. After conditioning at 80 °C for 6 hours to remove sorbed moisture on the surfaces and to eliminate residual stress caused by sample preparation, specimens were placed into distilled water chambers heated to constant temperatures of 45, 60, 75, and 90 °C. Specimens were weighed periodically. Once the specimens were taken out of the environmental chambers, the surface water was absorbed using a clean dry paper towel. Following the towel drying procedure, the specimens were left in room air for 15 minutes at room temperature before weighing. Then specimens were weighed by using an analytical balance with 0.01 mg resolution, and the percent weight change was determined. Compared with the immersion time, the weighing time is very short, and, because, the water difiilsivity in the material is small the weight loss during weight measurement is negligible. Water induced expansion and corresponding dimensional change were measured by a micrometer. Optical microscopy and scanning electron microscopy (SEM) were used to investigate surface modification associated with cracking, mass loss, and swelling. 6.3 RESULTS In Fig. 6.2, the weight change profiles are demonstrated for graphite/epoxy material at different temperatures. For ease of comparison, both the experimental and the theoretical water absorption curves versus the square root of time are plotted on the same figure. Solid lines represent weight gain assuming Fickian diffusion. Symbols represent 118 experimental data. Each symbol represents the average of three measurements. Because the ratio of thickness/width of the thin plate specimen is << 1, calculations of diflirsion parameters were performed using a one-dimensional approach without incurring significant error. Provided the initial temperature and internal moisture distribution of the material are uniform and the exposure temperatures are constant, the difliisivity can be expressedas D=—”—( h )2(M'2‘M" 16 M, th—JZ and h is the laminate thickness. The difliisivities of T300/934 composite are shown in 2. MIn is the equilibrium amount of sorbed water Table 6.1. As seen in Fig. 6.2 during the initial stage of moisture absorption, the experimental data match the theoretical profiles very closely for all exposure temperatures. With time, however, differences between the theoretical and the experimental data profiles are readily observed. The experimental data of the lower temperature curves provide better agreement with the Fickian diffiision curves. The data for the specimen immersed in 45 °C distilled water essentially coincide with the theoretical curve over an extended time period. The data of 60 °C and 75 °C specimens is nearly coincident with Fickian diffusion curves as the theoretical saturation level is approached. However, beyond the theoretical saturation level, the experimental data diverge fi'om the theoretical curves. The 90 °C specimen data diverge from Fickian behavior before reaching the saturation level, as the experimental weight change is greater than that of the theoretical values. After being immersed in water for 1300 hours (4.68 x 106 see), a reduction in the weight change of the 90 °C 119 1.8...,-,.,.,. % Moisture a a 1 a I a I a l 0 1000 2000 3000 4000 5000 6000 Time, Vsec Fig. 6.2 The weight change of T300/934 graphite/epoxy composite immersed in distilled water at difl‘erent temperatures. Solid lines represent theoretical Fickian difliision and the symbols are the experimental data at different exposure temperatures. Table 6.1. The difi‘usivity of T300/934 graphite/epoxy composite materials. Temperature (°C) 45 60 75 90 D (m’s") 2.87 x 10'“ 5.66 x 1043 1.43 x 10'12 3.63 x 10‘12 120 specimens was observed. Beyond 4700 hours (1.7 x 107 sec.), the weight change profile data are even lower than the theoretical saturation level. The contrast between the experimental data and the theoretical Fickian curves of Fig. 6.1 will be elaborated upon later in this study. Dimensional changes were measured periodically on specimens immersed in water at 60, 7 5, and 90 °C. The dimension along the fiber direction exhibited extremely high stability as no detectable dimensional change could be measured. This environmental stability is due to the high strength and stiffness provided by the reinforcing carbon fibers in the fiber direction. Since the carbon fiber is virtually impervious to water absorption, the longitudinal fiber direction dimension is considered invariant. Figure 6.3 shows the width change of specimens at different temperatures versus exposure time. When the maximum solubility limit is reached, no further dimensional change (moisture-induced expansion) occurs. Figure 6.4 shows the thickness change of the specimens at 60, 75 and 90 °C. The 60 °C specimen reached a stabilized dimension after 1600 hours. The 75 °C specimen reached stability after 1100 hours; and, the 90 °C specimen reached its maximum thickness change at 900 hours. After 1100 hours, the thickness change curve of the 90 °C specimen exhibited a significant downward trend. 121 1‘2 r r ' I ' I f I ' 1 1 1.0 - . ... in .000 g 0 o z 3 3 8 3 . . 2 0.8 - “ 0°C 0 o .. E: 06 . e o e e 0 ° . - Q - g 0 .... . O- o e E, 0'4 co ’ 60°C I 02 ° ° 75°C ‘ ° 0 90°C 00 l 4 l l a L a 0 2030 4000 6W0 8000 10000 Time,h Fig. 6.3 Width change of T300/934 graphite/epoxy composite immersed in distilled water at different temperatures. 1.6 . , . I , ' , I 0 1.4 .. a - 3:: 1.2 L “.0... -.- (damage) .1 -= 1.’ °9998 8 9 9 9 9 9 9 9 . U 1.0 . o . . - 8 o . . 2 0.8 a 5 . o ‘ '5 °‘6 . 60°C . 1 5Q 0-4 o 75°C - 0.2 o 90°C ' . 0.0 . M a r L 1 . l , 1 0 2000 4000 6000 8000 10000 Time,h Fig. 6.4 Thickness change of T300/934 graphite/epoxy composite immersed in distilled water at different temperatures. 122 Figure 6.5 shows optical micrographs of polished specimens at different environmental exposure conditions. Notably, the dry specimen does not have cracks, whereas the 75 °C and 90 °C specimens clearly exhibit multiple cracks after being immersed for 4300 hours. Cracks develop on the surface of specimens that contact the water. Moreover, the crack depth increases with time. At 4300 hours, the average crack depth emanating from the surface into the laminate is about 0.5 mm. Some 90 °C specimens were sectioned along the centerline after immersion in water for 4300 hours. Light-optics micrographs revealed that cracks were not present on the newly polished centerline surface section. This is in contrast to numerous cracks that are associated with the old exterior surface. Subsequently, the laminate was immersed in water for 168 hours and many cracks developed on the newly polished surface. This experiment indicated that edge cracking near the surface is promoted when the surface is in contact with the water. Moreover, the extent of internal cracking is significantly less. Figures 6.6 and 6.7 are SEM micrographs of the laminate surfaces at various test conditions. The surface layer is primarily neat resin (approximately 15 um thick). The woven-like surface pattern is due to the impression of the cloth on resin during the manufacmfing process. Surface resin peeling of the laminate can be seen clearly (Fig. 6.7a) for the 90 °C test condition after 4300 hours immersed in water. The thickness of the resin-rich surface layer becomes thinner and smoother (Fig. 6.7b). This resin loss explains why the thickness dimension of 90 °C specimens exhibited a significant downward trend with exposure time. 123 Fig. 6.5 Optical micrographs of T300/934 Gr/Ep composite before and after immersion in water for 4300 hours. a) dry specimen, no crack can be seen, b) 75 °C, and c) 90 °C specimens. Both of the 75 and 90 °C specimens have visible cracks. 124 Fig. 6.6 SEM micrographs for surface images of dry and 90°C specimens. The surface layer is primarily neat resin. The woven-like surface pattern is due to the impression of the cloth texture on epoxy resin fiom the manufacture process. a) dry specimen, no damage on the surface and b) 90°C specimen, there is resin peeling after 4300 hours immersion in distilled water. 125 Fig. 6.7 SEM micrographs of the surface of 90°C specimen after 4300 hours immersion in water. a) resin peeling is highly visible, b) peeling, cracks and dissolution. 126 6.4 DISCUSSIONS 6.4.1 Moisture-Induced Material Response The hygrothermal environmental effects and behavior of PMCs are quite complex and controversy still remains. Rao and his coworkers consider that moisture absorption in permeable fiber/polymer composites can be characterized using a Fickian diffusion model [Rao 1984]. Lee and Peppas [Lee 1993] studied water transport in Gr/Ep composites. They observed that water absorption kinetics of the TGDDM resin-based composites could be modeled by F ick's law. But the DGEBA-based samples indicated non-Fickian behavior because internal cracks were found after immersion in water at higher temperatures (@ 90 and 100 °C) [Lee 1993]. Imaz et. al. found that the moisture absorption of Gr/Ep composites was Fickian but at higher temperature there is a slight tendency to increase saturation level. This trend was attributed to a slight irreversible degradation in the material [Imaz 1991]. Loos and Springer investigated the moisture absorption behavior of many kinds of resin matrix composites. Their [Loos 1981] results show that materials obey Fickian diffusion behavior at lower temperatures and non-Fickian behavior at higher temperatures. The most plausible explanation is that under a moist high temperature environment, microcracks develop on the surface and inside the materials. As the cracks developed, material was actually lost, most likely in the form of resin particles. As long as the moisture gain was greater than the material loss, the weight of the specimen increased [Loos 1981]. Shirrell [1979] also observed similar kinds of moisture-induced surface modification. Research by Bonniau et. al. confirmed the surface mass loss and dissolution [Bonniau 1984]. Weitsman et. a1. [1993] theoretically calculated the non- Fickian moisture absorption profile and found that moisture uptake was 25% higher. In 127 their calculation [W eitsman 1993] residual hygrothermal stress significantly influenced non-Fickian behavior. In the present study, cracks, peeling, voids and surface dissolution of the Gr/Ep composite laminate were quite evident after long-time immersion in distilled water. From the previous works cracking and mass-loss have been observed. but there is no systematic study and sufficient proof to discuss this problem. In this study water absorption, dimensional change, and surface modification were investigated. Both Fickian and non-Fickian difl‘usion were observed. Exposure temperature plays a key roll for the diflirsion behavior of water in Gr/Ep composites. It is seen clearly that in the material response to the overall moisture absorption process, cracks (including surface voids) and the surface mass loss (including surface peeling and dissolving) has great influence on the apparent weight change behavior. Surface peeling and resin dissolution contribute to weight loss (a net decrease of the overall weight) of the specimen, whereas surface crack and voids between fibers trap water thus promoting weight increase. The apparent weight change profile encompasses the combined competing effects of water diffusion, physio- adsorption of water at crack tips which promote weight gain in the laminate, and surface mass loss mechanisms which contribute to gross weight reduction. The formation of cracking is considered due to the large difference in water-induced stress between the fiber and matrix phase. Graphite fiber is water impermeable. During hygrothermal exposure there is no water-induced swelling in the fiber. Owing to the great difference in moisture- induced expansion, sheer stress develops along the graphite fiber and matrix interface. In the interior of the specimen, the stress equilibrium is maintained. However, at the 128 exterior, the stress equilibrium does not exist any more. When the sheer stress exceeds the adhesive fiber-matrix bonding strength, cracks may develop. It must be noted that incompletely cured and non-stoichiometric material should not be included in this discussion. Correctly stoichiometric ratio of resin and hardener and properly curing process are of the critical factors. Non-stoichiometric material contains some uncrosslinked monomers. During long time hygrothermal exposure these monomer is easy to leach out and results water gain profile downward. If the curing process is uncompleted, even though the resin/hardener ratio is stoichiometric, low molecular weight oligrners exist in the material. Under hygrothermal conditions these relative small molecules are easy for leaching and hydrolysis resulted the diffusion curve going down. Incompletely cured epoxy has other significant influence on diffilsion profiles. Further curing may occur in hygrothermally exposure process. This crosslinking results larger free volume and, then, more water difl‘uses into the material. Persistently increasing profiles are observed. Additives are another factor to influence the difl‘usion behavior. Additives are used to improve the mechanical property of the material to meet some specific requirement or improve manufacturing process. They are usually catalyst, accelerator, and flexor. The leaching of these additives makes apparent water gain decrease too. Some may have chemical reaction with water and much more complete phenomena will be observed. 129 6.4.2 Crack/Mass-Loss Model Based on the experimental results a crack/mass-loss model has been developed. This model can be used to describe the water sorption phenomena of graphite/epoxy composite materials immersed in water at different temperatures. For the 90 °C specimen, classical diffusion processes are dominant and F ickian law is obeyed for the first 170 hours. Subsequently, cracking occurs by the mechanism described earlier. Owing to the trapped water at the crack tip, the sorbed water content increases. Therefore the weight change profile is higher than the Fickian curve for a certain period of time. After 1300 hours, a difl‘erent dimensional change behavior takes place. The width dimension remains invariant after that period of time, but the thickness exhibits quite difi‘erent behavior and is coincident with the weight change profile. SEM examinations indicate surface peeling and dissolution of the matrix phase with exposure time immersed in water. The contrast in behavior exhibited between the width and thickness of 90 °C specimen is due to different surface characteristics. A fiber-epoxy network exists on the width surface, and the fibers can hold and interlock the epoxy to prevent peeling. However, on the thickness dimension, a layer of neat epoxy resin exists (15 um thick) on the surface with no interlocking by fibers (Fig. 6.8). At 90 °C the neat resin layer peels and dissolves away with time, resulting in a decrease of the gross weight of the specimen. When mass loss becomes the controlling mechanism, an overall weight reduction in the specimen is observed. This severe cracking and surface peeling did not observed for 90 °C neat resin since there is no fiber/resin interface. The great sheer stress due to the existence of the interface is the cause of cracking and peeling. Surface hydrolysis is observed for the neat resin but the degree is not as large as Gr/Ep composite because the exposure time is quite 130 different (9000 h for Gr/Ep and 1530 for neat resin). For the 75 °C specimens, at times less than 800 hours, no cracks develop in the —'>|-l<— Fig. 6.8 Schematic diagram of unidirectional Gr/Ep composite prepreg. specimen and only Fickian diffusion exists. After 800 hours, cracks develop and the water content increases. There is no or, at best, very limited surface mass loss so the water content gained due to cracks is maintained. The behavior of the 60 °C specimens is similar to that of the 75 °C specimens. The main difference is that cracks appear at a much later time and the crack density is much lower. No crack development or surface mass loss occurs on 45 °C specimens, so its water absorption behavior/diffusion is completely 131 Fickian. Aspects of non-Fickian diffusion behavior for the graphite/epoxy material in this study is clearly depicted in Figure 6.9. The solid lines in Fig. 6.9 Fig. 6.9 Moisture absorption profiles of a graphite/epoxy laminate showing experimental I \lTime, (Nisec) data and theoretical prediction of Fickian diffilsion. . l ' I'I ' riori-Fieirian‘ III (a). . Fickfan ' ' nofioFigkian' (b) 1.6 - g microcracks! dissolution :' 1: 1 com o o ‘ 14 L o M - - Moog , .. ° . ii'.‘ of m . . microcracks . 12 - unhea— - - a ‘ q, 1 r- . 1.0 \ O - r- .4 03 ' Fick's law a ; '_ _' 0.6 .I L J '8 0.4 - - - {5, 0.2 90°C 1 75°C ‘ U 00' r l J 4 1 a r a r L . l r r I 1 a r . 1:" 0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000 GD -- 1.8 4 4 4 . . 4 . . Q I I I I I I I I I . l 11 c . . . . 1 d) ‘ E 1'6 ,- Fickian non-Fickian( )- :' F‘Ck‘m ( i 1.4 - microcracks - 1- dam :- (nodamage) 00w 0 o 1 '_ (no 88:) o .1 ' 1 .' 1 - 4 45°C a 0.0 a I a 1M1 I a I a I a I n I a I 1 I A I a I a 1 0 1000 2000 3000 4000 5000 6000 O 1000 2000 3000 4000 5000 6000 132 0.. O a b a A-0.....O 'a " fl_._ - C2 0 .o. ‘3 3 c d ...? O... O 'o O O O O O 6 f .o O Exposure Time Fig. 6.10 The schematic diagram of typical weight change profiles of Gr/Ep composites in water environment. represent the theoretical Fickian diffilsion. The symbols represent experimental data. In stage I, moisture diffusion is characteristic of Fickian behavior. Stage 11 indicates non- Fickian behavior associated with microcrack formation, particularly at the free edge of the 133 test material. In stage III, the non-Fickian behavior is characterized by gradual weight loss of the specimen resulting fi'om resin leaching and dissolution. This crack/mass-loss model can also explain the previous moisture absorption research results not only for distilled water, but also for moist air, salt water, fuel, and other liquids if cracking and/or surface modifications are involved in the weight change process. In general, a hygrothermal environment influences the absorption behavior of graphite/epoxy PMCs and is a manifestation of the models demonstrated in Figure 6.10 which is a schematic depiction for typical weight change profiles of Gr/Ep composites in difi‘erent hygrothermal conditions. If no cracking or mass loss occurs, then F ickian behavior is obeyed as noted in Fig. 6.10a [Loos 1981, Imaz 1991, Rao 1984, Ghorbel 1993]. This situation generally occurs near the temperature region of T/T g <0.25, where T is the water bath temperature. In Fig. 6.10b, there is no cracking in the initial stage of moisture absorption in the material, and Fick's law is obeyed. Subsequently, cracking occurs and the experimental data deviate fi'om Fickian behavior, resulting in increased water absorption [Loos 1981, Shirrell 1978, Imaz 1991]. In Fig. 6.10c, initially no cracks exist. Afterward, cracks develop, and then, dissolution and peeling of the resin occur [Lucas 1993]. There are two ways in which the profile of Fig. 6.10d is attainable [Loos 1981, Shirrell 1979, Lee 1992]. First, at the initial stage, there are no cracks or peeling although with time both mechanisms take place. Of these two competing mechanisms, mass loss due to surface resin peeling dominates the weight gain profile. Second, no cracks develop and the absorption curve obeys Fickian law until surface mass loss occurs. This situation can be found especially in neat resin. Figure 6.10e indicates that cracks develop quickly and, even in the linear absorption stage, the weight increase is higher than 134 that representing Fickian law behavior [Loos 1981, Shirrell 1978, Gupta 1985]. No mass loss is involved in the process. In the case of Fig. 61ij surface peeling is the dominant mechanism and the specimen rapidly loses weight [Loos 1981, Lee 1992]. The last two situations are most often observed at relatively higher temperatures (T/Tg >0. 5). 6.4 SUMMARY Water sorption in a graphite/epoxy (T3 00/93 4) composite exhibited both Fickian and non-Fickian diffusion behavior. Difi‘usion data showed that time for the onset of non- F ickian behavior was inversely related to the exposure temperature. Anomalous (non- Fickian) behavior in the composite resulted from physical damage to the epoxy resin. Cracks, voids, and surface peeling were clearly observed by SEM and optical microscopy. Moisture induced expansion of the T3 00/934 composite was measured in length (fiber direction), width, and thickness directions. Essentially no expansion due to water absorption was detected in the fiber direction dimension. Significant dimensional changes resulting fiom moisture-induced expansion were observed in the width and thickness directions of the laminate. The thickness decrease of the specimen at high temperature was associated with surface resin dissolution and peeling. A crack/mass—loss model favorably describes the phenomenological behavior of Gr/Ep composite resulting fi'om water absorption processes. At a low exposure temperature compared to Tg, there is no surface dissolution or physical damage of the material and the weight gain behavior is Fickian. With increasing exposure temperature, cracks, voids, surface peeling, and dissolution occur. Cracks can retain water which contribute to absorption behavior higher than the theoretical Fickian diffusion curve. 135 Surface peeling and dissolution contribute to reduction in the specimen weight and, consequently, the weight change profile data falls below the theoretical Fickian diffiision curve. The resultant weight change profile represents the combination of these two factors, cracking and mass loss. Depending on the mechanism that dominates the experimental data, either profile higher or lower than the theoretical Fickian curve is possible. CHAPTER VII HYGROTHERMAL EFFECTS ON GRAPHITE/EPOXY COMPOSITES Having studied the diffusion behavior of water in Gr/Ep composites, we will discuss additional hygrothermal effects in this chapter. The hygrothermal effects have been extensively studied in Gr/Ep composites [Loos 1981, DeIasi 1978, Imaz 1991, Biro 1993, Lee 1993, Lucas 1993, Adarnson 1980]. It is abundantly clear that hygrothermal condition degrades both the mechanical and physical properties of PMCs, particularly, epoxy-based composites. To date, however, due to the complexity of water interaction in epoxy resin and resin-fiber interface, there is not a single theory with suflicient experimental support, that explains the all hygrothermal phenomena in Gr/Ep composites. Also, the relationship between hygrothermal effects and the bonding characteristics of water molecules in epoxy matrix are still not fully understood. The hygrothermal effects of absorbed water in epoxy systems are mainly plasticization (lowered T,), mechanical degradation, and swelling. Investigations reveal that sorbed water in epoxy acts as a plasticizer [Shen 1981, Imaz 1991, Biro 1993, Lucas 1993, Wolfi‘ 1993]. Based on the fiee volume concept of polymeric materials, Kelley and Bueche [Kelley 1960] developed a polymer-dilute model to predict T, variation. This is a classical method for predicting the variation of T, in water-sorbed resin. Several investigators [Cairns 1984, Browning 1978] applied this model to epoxy resin/water 136 137 system and explained the dependence of T, on water content. DeIasi [1978] suggested that water that disrupts the hydrogen bond to depress T,, whereas, water that forms cluster or hydrogen-water type groupings has no measurable effect on T,. Mijovic and Weinstein [1985] found that the depression of the glass transition temperature in a Gr/Ep composite after water absorption is strongly dependent on the temperature of the environment during water absorption. At the same absorbed water content, the T, depression was greater at higher absorption temperature. The mechanical effect of sorbed water in Gr/Ep composite is manifested by a reduction in elastic modulus and by a decrease in the matrix-dominated strength properties. Shen and Springer [Shen 197 7] summarized the previous tensile modulus data of composites and concluded that for 0° and 1t/4 laminates, there appears to be very little change in the bulking muduli over the entire spectrum of water contents from dry to fillly saturated in the temperature range of 200 to 450 K, whereas, for 90° laminates, the saturated moduli decreased considerably with increases in both the water content and temperature. Using the stress-strain data of various water-treated samples of neat resin and composites, Browning et al. [197 8] conclude that, as water absorption increased and/or as temperature increased, the tensile modulus of the resin matrix and the transverse modulus of the composite decreased. Crossman et al. [1978] determined the tensile relaxation modulus after exposure to equilibrium moisture levels. They found drastic reductions in modulus and an enhanced rate of relaxation at higher temperatures and moisture contents. Study of hygrothermal effects on Gr/Ep composite is more important than on neat epoxy because the final applications of epoxy as structure material are always in the form 13 8 of Gr/Ep. The hygrothermal effects of neat epoxy have been investigated in Chapter 3 - 5. The questions in this chapter is 1) if there are two types of bonded water in Gr/Ep composites, 2) if T, changes with hygrothermal history, and 3) if swelling and mechanical change mainly depend on physiosorbed water. All these questions will be answered in this chapter. The aim of this study is to discern sorbed water induced hygrothermal effects in Gr/Ep composites by evaluation of: a) water absorption and desorption phenomena, b) dimensional changes, c) T, modifications, and d) mechanical properties variations. By using these integrated experimental approaches, a more comprehensive understanding emerges regarding the hygrothermal efi'ects in epoxy-matrix composites. 7 .1 EXPERIMENTAL 7.1.1 Material The material used in this study was unidirectional T300/934 Gr/Ep composite laminate received from ICI Fiberite”. The Fiberite 934 epoxy consists mainly of tetraglycidyl-4,4'-diaminodiphenyl methane (TGDDM) resin, 4,4'-diaminodiphenyl sulfone (DDS) hardener. The chemical structures of TGDDM and DDS has been shown in Chapter 3. More detailed material information is included in Chapter 6. 7.1.2 Water Absorption and Desorption Specimens with dimensions of 43 x 15 x 2 mm were cut from the laminate panel using a diamond blade saw. The water absorption procedure is the same as the described in material section of Chapter 6. After immersion in distilled water for more than 9000 h, all specimens held at temperatures of 45, 60, 75, and 90 °C reached firll saturation. To 139 conduct water desorption experiments, water-saturated specimens were placed into a heated chamber with slowly flowing dry air. Desorption was performed using a two-step process: (i) first, desorption took place at 60 °C for 1250 h followed by (ii) desorption at 100 °C for an additional 250 h. Each datum presented in the absorption and desorption profiles represents the average of weighing measurements of three specimens. 7.1.3 Glass Transition Temperature Tests To determine T, variations at different hygrothermal stages, a TA Instruments 930 thermomechanical analyzer with 2200 data acquisition and analysis system was used. The principle of T, by TMA has been described in Chapter 4. Specimens (5 x 5 x 2 mm) for TMA measurement were cut from the hygrothermally exposed laminates at different absorption or desorption stages. To obtain more precise TMA data, the choice of the contact load value was motivated by the suggestion of Carter and his coworkers [1978]. Consequently, a 500 mg load was used for the TMA analysis. T, values were determined by the intersection of two tangential lines of the TMA expansion curve below and above the glass transition regime [Carter 1978]. 7.1.4 Infrared Tests Infrared spectroscopy was used to determine the chemical change of the epoxy resin associated with bonding characteristics at various absorption and desorption stages. A Perkin-Elmer 1800 FTIR was used in this study. Resin powder was obtained directly from T300/934 composite by scraping with a sharp metal blade. In this study, 2 mg of epoxy powder was mixed with 200 mg of spectrum-pure grade KBr and pressed into thin, 10 mm diameter discs under a 5000 kg load. The spectral resolution was 4 cm'l and the final spectrum was averaged over 100 scans. 140 7.1.5 Mechanical Tests Flexural testing was conducted on hygrothermally exposed composite laminate. An Instron 43 02 mechanical test system was employed to conduct three point bending tests. The crosshead speed was 0.5 mm/min and the span was 25.4 mm. Flexural test specimens were smaller in size (43 x 7.5 x 2 mm) than those suggested by the standard test method ASTM D790-81. Double cantilever beam specimens (DCBs) were used to determine fiacture energy results of Gr/Ep laminates. DCB tests were conducted in ambient room conditions (23 °C and RH 65%). To facilitate testing, aluminum hinges were adhesively bonded to the notched end of the DCB specimens. Testing was performed at a grip opening rate of 5.0 x 10'4 mm/s. Fracture toughness, K1,, was determined using a modified chevron notch short bar specimen as described by ASTM 1304. In this study, the Krv fiacture toughness test specimen is designated as the laminate short bar (LSB) specimen. A desirable, usefiil feature of the chevron-notch short bar test is simplicity. In stark contrast to other fiacture toughness test methods, measurement of the fatigue precracking is not necessary. Furthermore, once the compliance calibration is determined for a specified specimen geometry, the fracture toughness values can be obtained independent of test material [Barker 1977, 1983]. Essentially, only the critical load is needed- to determine the fi'acture toughness and the provided criteria for linear elastic fiacture mechanics are satisfied. The short bar plane strain fracture toughness is determined by the expression, Ki. = AB(Paa./B3’2), where A is a constant that depends on specimen dimensions and the elastic modulus, E. The parameter A is given as f(b,B, E), where b is the instantaneous crack fi'ont width, which widens with incremental crack advance, and B is the specimen width 141 modulus, E. The parameter A is given as f(b,B, E), where b is the instantaneous crack front width, which widens with incremental crack advance, and B is the specimen width [Barker 1977]. Pat, is the critical load and B is a dimensional correction factor for slightly misdimensioned specimens. Testing was performed at room temperature in air. Tests were performed in displacement control on a F ractometer H short bar test system. The grip opening displacement rate was 5.8 x 10" mm/s. Sharp crack initiation is achieved under the influence of a high stress field occurring at the chevron notch tip. The sharp crack grows to a critical length in a stable manner. When Pen-1 occurs at the critical crack length, the fiacture toughness can be determined. Fig. 7.1 The laminate shot bar (LSB) test specimen. 142 7.2 RESULTS AND DISCUSSION 7.2.1 Water Absorption and Surface Modification Water absorption and diffusion behavior in Gr/Ep composites have been investigated in Chapter 6. In this chapter, previous water sorption results are reiterated briefly and are intended to serve as background information for firrther discussion of the hygrothermal effects in Gr/Ep composites. Examination of the water absorption profiles shows that, over a 45 - 90 °C exposure temperature range, both Fickian and non-Fickian diffusion behavior is exhibited (Fig. 6.1). Cracks, voids, and surface-resin peeling were observed (Fig. 6.4-6) by using optical and scanning electron microscopy (SEM). Anomalous (non-Fickian) moisture difl‘usion behavior resulted because of surface damage and physical modification of the epoxy resin. At low temperatures relative to T,, no significant surface resin dissolution or physical damage to the material was observed, and, thus, the weight gained as a result of water uptake is characterized by classical Fickian difiiision. With increasing exposure temperature, however, cracks, voids, surface peeling, and surface dissolution occur. Microcracks retain water and consequently contribute to absorption profiles that deviate fi'om the theoretical F ickian diffusion curves as noted by the solid-line profiles in Fig. 6.1. Material loss from the specimens by surface peeling, dissolution, and leaching contributed to weight reduction and, so, the weight change profile data fall below the theoretical F ickian diffusion curve. Consequently, the experimental weight change profile represents the combined influence of the two factors: (a) microcrack-induced excessive weight gain and (b) mitigated mass/loss-weight reduction. Depending on the dominant mass-change 143 mechanism, experimental data will exhibit either a higher or a lower water profile than the classical Fickian curve. Some other important considerations are emphasized here. It is shown in Chapter 3 that for a given material, all hygrothermally exposed samples have the same water saturation level M... In other words, if Fickian diffusion is obeyed, the water content in the composite is the same irrespective of the bath (exposure) temperature. Second, cracks and other surface defects were observed for all hygrothermally exposed samples, except for the samples exposed at 45 °C. Therefore, water diffusion behavior in the composite at 45 °C is considered to be completely Fickian. Cracktips, microcracks, and surface defects trap water by surface tension and capillary phenomena. Such trapped water is not controlled by classical Fickian diffusion mechanisms. Such trapped water is classify as @- bonded water. O-bonded water contributes to excessive (non-Fickian) weight gain. 7.2.2 Water Desorption Figure 7.2 shows water desorption profiles of 45, 60, and 75 °C for hygrothermally exposed specimens resulting from baking in a dry-atmosphere chamber at 60 °C for 1250 h and, then, at 100 °C for 250 h. The initial rapid desorption of water results fi'om basic evaporative losses of surface trapped water (O-bonded water) at cracktips and other surface and near-surface defects. Once O-bonded water sources are exhausted, the water desorption rates for all (@ 45, 60, and 75 °C) specimens approach similar values. After desorption at 60 °C for 1250 h., water, albeit a small amount, is still retained in the specimens. This retained water is not readily removed at this temperature (i.e. 60 °C) regardless of how long the desorption process continues. This phenomenon is the 144 1.8 . , . . . , . . - . 8 1.6 P 0 1 1-4 - Desorption of - _ D e - bonded water it 1.2 - .. e; O E ' u g“ 1.0 _. - 2: a ...................................... E" I: § 0.8 l- .1 a r 9: . a 6 - a 0' ’ 9r Desorption of g a x A - bonded water . E 0.4 - d‘ .. 1- m u - 4 02 %89 (n o ................... 1- Xxua a , . " " Desorption of 0.0 l- r ' bondCd - _ X Bath 45 0C water .02 _ D Bath 60 °C “9 O _, 0 Bath 75 °C -04 4 1 a r . J . l a l . -05 0.0 0.5 1.0 1.5 2.0 2.5 Time, h x1000 Fig. 7.2 The weight change profiles of the T300/934 composite desorption at 60 0C for 1250 h and then at 100 0C for 250 h. 145 same as in epoxy resin reported in Chapter 3. The desorption profiles in Fig. 7.2 also indicate that composite specimens exposed at higher immersion temperatures retain more water in the resin matrix. To completely expel the retained water, baking at 100 °C for an additional 250 h was required. From the desorption results, it is clear that two types of sorbed water, like in neat epoxy resin, exist in the hygrothermally exposed Gr/Ep composites. One type is readily desorbed and is classified as "physiosorbed" water or A- bonded water. In comparison, the other type of sorbed water is more difficult to bake out and is, thus, classified as strongly bonded water or chemisorbed F-bonded water. The A- bonded water is baked out in the second desorption stage (Fig. 7 .2). F—bonded water, however, can only be removed in the third desorption stage, which is associated with strong-coupled water molecules with matrix resin. Much more detailed information about the properties of the two bounding state of water can be found in Chapter 3. We have classified the three types of bonded water involved in the hygrothermal process in Gr/Ep composites. In the following sections, the main hygrothermal effects of Gr/Ep composites and how the hygrothermal effects are influenced by the three types of sorbed water will be discussed. 7.2.3 Change of Glass Transition Temperature The T, for specimens conditioned at varying hygrothermal stages is shown in Figs. 7.3 - 7.6. The data of Figs. 7.3 - 7.6 indicate that: i) the T, of all water saturated samples decreased dramatically compared to the dry specimen, ii) the saturated specimens hygrothermally exposed at low temperatures render lower T,, iii) the T,’s of redried specimens baked at 60 °C for 1250 h are slightly higher than the T, of as-received, 146 Dimension Change (Arbitrary Scale) r r I I I I I 20 4O 60 80 100 120 140 160 (°C) Fig. 7.3 T, of the four saturated specimens after immersion in water for more than 9000 h at different temperature. 147 Dimensional Change (Arbitrary Scale) I I I l I 25 50 100 150 200 250 275 (°C) Fig. 7.4 T, of the four saturated specimens after desorption at 60 0C for 1250 h. 148 Dimensional Change (Arbitrary Scale) I I I I I 25 50 100 150 200 250 275 (°C) Fig. 7.5 Tg of the four saturated specimens after desorption at 60 0C for 1250 h and then at 100 0C for 250 h. 149 2007 190 __t'\ J‘ 5:" 170 C ' T 4) Q l- g 160 l" Tg e 150 1. ’. o 3. _ (as-received, 175 C) E 140 .. Q P [— 130 - .l a . .....- :§ 120 :- ...Jgnuno- a 110 II: ..“¢O¢ E 100 1- ‘.,o'." g. 90 f I’.’ ---I--- Tg, at saturated after 9000b 9 80 _- -o~ Tg, after desorp. - 60 °C. 125011 70 - -O— Tg, after desorp. - 100 °C. 250h 6O . . l l n a l i 1 a r i 40 50 60 70 80 90 100 Temperature of Bath, '°C Fig. 7.6 The T, variations with the different immersion temperatures and desorption stages. dry material, and iv) T, is fully recovered after desorption at 60 °C for 1250 h and then 100 °C for 250 h. The difference in T, values between the saturated 45 and 90 °C samples is about 37 °C. These results strongly suggest that A-bonded water contributes to the reduction in T,, whereas, F-bonded water does not contribute to the reduction of T,. Moreover, F-bonded water enhanced T, slightly. DeIasi [1978], Moy [1980], and DeNeve [1993] reported on T, test results for epoxy resin. Their experiments were carried out at constant temperature and different 150 relative humidity. They found that T, was lowered with increased water uptake in epoxy resin. Previous studies [DeIasi 1978, Moy 1980, DeNeve 1993] have shown that T, and AT, in Gr/Ep composites depend only on the “amount” of absorbed water [T, = f(%M)]. According to their results, one would expect that T, values of the specimens in this study should be the same since the water contents of all the specimens are the same (1.2 w%). However, results of the present study suggest that T, does not depend simply on the sorbed water content in the resin, but also on exposure (bath) temperature and time [T, = f (%M, T, t)]. From the discussion of section 7 .2.1 we have revealed that, irrespective of the bath temperature, the sorbed water content of all hygrothermally exposed specimens reaches a similar saturation level. Furthermore, as exhibited in Fig. 7 .2, the desorption profiles plateau at different water content levels depending on the hygrothermal exposure temperature. Specimens exposed at higher temperatures render correspondingly higher plateaus. Also, higher immersion temperatures promote F-bonded type water at the expense of A-bonded water. More F—bonded water and less A-bonded water resulted in apparently higher T,’s (Fig. 7 .6). The results of this study are consistent with previous investigations [DeIasi 197 8, Moy 1980, DeNeve 1993]. As noted in earlier investigations [DeIasi 1978, Moy 1980, DeNeve 1993], specimens were tested at the same exposure temperature and time, but at difl‘erent relative humidities. Regarding this hygrothermal history, all the specimens in these studies would possess the same amount of F—bonded water, since the amount of F- bonded water is controlled by exposure temperature and time. The only difference exists in the amount of A-bonded water (physiosorbed water). Higher relative humidity imposes l 5 1 conditions for enhancement of physiosorbed water in the resin and, consequently, lower T,’s. The T, variation in Gr/Ep composites and neat epoxy resins is remarkably consistent after long time exposure in hygrothermal environments. The comparison results are given in Table 7.1. Table 7.1 The comparison of the T, results in the long time exposed epoxy Fiberite 934 and Gr/Ep composite Fiberite T3 00/934. Exposure at Exposure at Exposure at Exposure at 45 °C 60 °C 75 °C 90 °C T, of Fiberite T300/934 89 102 119 126 T, of Fiberite 934 resin 87 106 112 127 7.2.4 Mechanical Testing Results Three-point bending tests were carried out on 90 °C specimens at various absorption and desorption stages. Table 7.2 shows the flexural testing results. The saturated specimens exhibited the lowest flexural strength. After desorption at 60 °C for 1250 h, the A-bonded water was removed and the flexural strength recovered to a strength level slight higher than that of the as-received dry specimen. It should be noted, however, that F-bonded water still remains in the resin matrix at this desorption stage. Enhanced desorption at 100 °C for 250 h removed all the F-bonded water and the flexural strength reverted to that of the as-received dry specimen. 152 Fracture energy results for the saturated Gr/Ep laminate are shown in Fig. 7.7. With the increase of immersion temperature, the fracture energy tended to increase. For the 90 °C specimen, the fracture energy was even higher than that of the dry specimen. Again, it is contended that higher immersion temperatures induce additional secondary crosslinking which contributes to a slight increase of the fracture energy. Also, enhanced fracture toughness is attributable to energy absorbing mechanisms such as matrix plasticization and fiber bridging. Gr/Ep fracture toughness results determined using laminate short bar specimens (LSBs) are shown in Fig. 7.8. As indicated in Fig. 7.8, sorbed moisture degrades the delarnination fi'acture toughness (DFT) in the composite. Significant degradation in delarnination fracture toughness occurred with moisture sorption under severe hygrothermal conditions. The reduction in mode I delarnination fracture toughness is approximately 60% following hygrothermal exposure of the Gr/Ep laminate for 1.3 x 1073. It has been suggested that plasticization of the matrix can lead to an increase in delamination fracture toughness [Jordan 1987, Bradley 1985]. That the DFT should increase was attributed to the enhancement in the moisture-induced crack tip plastic zone size. The contention was that the increase in plastic zone size due to plasticization enhances plastic energy dissipation at the crack tip, thus, efl‘ectively increasing the fracture toughness of the laminate. The decrease in the laminate fracture toughness 90 °C 9000h 90 °C 9000b in Dry specimen (as-received) immersion in water water . stress 626 (MPa) - n u h b p p l— » h b 0 555555 1 aaaaaaaaaaaaaaaaa n. 154 1.6 1.4 Graphite Epoxy Composite - E i .- 7- d :5: .4 2 . a: 1 0.4 . 1 . 1 1 1 . 1 . 1 1 5 0.0 2.5 5.0 7.5 10.0 12.5 15.0 Time, sec (x106) Fig. 7 .8 The effect of moisture sorption on delarnination fracture toughness is exhibited for Gr/Ep composite laminates. observed in this study with matrix plasticization (sofiening) is in contrast to the previously published results [Law 1984, Jordan 1987, Bradley 1985]. This dichotomy can be rationalized by the realization that there is a competing effect to ability of the plasticization to dissipate crack tip strain energy and, thus, enhance fracture toughness. This competing effect is the weakening of the bulk matrix and the fiber/matrix interface phase in the laminate due to matrix plasticization by sorbed moisture. Particularly, the contribution of fiber/matrix interfacial fracture to overall crack propagation and interfacial debonding in moisture-sorbed Gr/Ep laminates was rather extensive as noted in this investigation. Morgan [1980] and others [Lucas 1989, Garg 1985, Imaz 1991] have reported several degrading effects of moisture on physical and mechanical properties of therrnoset resins. Among the effects noted is that moisture lowers the tensile strength, elongation and 155 modulus of epoxy. Also, sorbed water tended to enhance the susceptibility of the matrix to cavitation and plastic flow process. Preferential cracking along the fiber/matrix interface is exhibited in Figs. 7 .9 for a Gr/Ep laminate exposed to hygrothermal conditions. Examination of the crack tip region shows major cracking in the interfacial region and microcrack link up of the inter-matrix region with interfacial cracking. In addition to primary fiacture along the interfacial region (interface phase) significant secondary cracking is observed. The secondary cracking is caused, presumably, by residual stresses in the matrix resin as a result of plasticization by moisture. The fi'acture characteristics of Fig. 7 .9b are indicative of a weakened interface as the crack appears to meander fiom the top fiber to the lower one. Macrofi'actography reveals such cracking branching as hackles [Garg 1985, Morris 1977] due to tensile stress relief in matrix-rich side of the main crack faces. 7.2.5 Swelling and Dimensional Change Table 7.3 shows the dimensional change of the 90 °C specimen at difi‘erent absorption and desorption stages. At the saturation level, the dimensional change due to swelling reaches a maximum value. Afier desorption for 60 °C 1250 h, the dimensions of the specimens essentially recover to the as-received dry specimen dimensions even though F-bonded water still remains in the specimen at this stage. Adarnson [1980] suggested that water molecules may either occupy fi'ee volume causing no swelling or interrupt interchain hydrogen bonding causing swelling. The results in Table 7.2 indicate that A- bonded water contributes mainly to swelling, whereas, F-bonded water has no significant influence on swelling. 156 Fig. 7.9 Fracture in the crack tip region is shown for Gr/Ep composites. (a) Fracture occurs along the fiber matrix interface. (b) Fracture is shown at the interfacial regions between two fibers. 157 Table 7.3. Dimensional change of the 90 °C samples at different hygrothermal stages. Dry 90 °C 9000b 90 °C 9000b in 90 °C 9000b in water specimen (as- immersion in water then then desorb. at 60 °C received) water desorb. at 60 °C 1250b and 100 °C 250h 1250b Width(mm) 43.484 43.881 43.503 43.471 % change 0 0.9 0.043 -0.029 7.2.6 Infrared Results Although many [Moy 1980, Adarnson 1980, DeNeve 1993] believe that water can interact with epoxy resin, there is little direct infi'ared evidence to verify this claim [DeNeve 1993, Antoon 1981]. In this study, FTIR spectroscopy was conducted, and results were obtained that allowed inferences to be made concerning matrix/water interactions. Fig. 7.10 shows the infiared spectra results of several temperature and moisture-conditioned specimens. Included in Fig. 7.10 are spectra of as-received dry, long-time baked dry, and water-saturated 934 epoxy resin specimens that were exposed to different water immersion temperatures. The basic chemical compounds of the Fiberite 934 are the resin, tetraglycidyl-4,4'-diaminodiphenyl methane (TGDDM), and the hardener, 4,4'-diaminodiphenyl sulfone (DDS). There are other additives in the Fiberite 934 resin, but these chemical compounds and their concentrations are proprietary. For this reason, it is difficult to ascertain in this study all the detail chemistry of the 934 resin from [R results. However, these results still reveal some usefirl information. The 1594 158 cm'1 and 1512 cm'1 peaks represent carbon-carbon double bonding v(C=C) from the aromatic ring of TGDDM. The 1720 cm‘1 is thought to be derived from the vibration of carbon-oxygen double bonding v(C=O). However, we know there is no C=O bonding in a pure TGDDM and DDS system and, therefore, the peak 1720 cm'1 results most probably from additives in the 934 resin. The obvious change of the spectra is that the 1720 cm'1 band, which exists in the original dry specimen, decreases significantly after long-time immersion in water (9000 h) at different temperatures. As the immersion temperature rose, the 1720 cm’1 peak decreased significantly. For the 90 °C specimen, the peak nearly vanished (Fig. 7.10-a). To substantiate that the change in peak height is caused by water sorption and not by a temperature increase, a sample of dry material was temperature- aged first at 60 °C for 2250 hours and, then, for an additional 250 hours at 100 °C. Subsequently, infrared testing was conducted on this temperature-aged specimen. The results (Fig. 7.10-1) indicated that 1720 cm'1 peak did not change during the long-time, temperature-aging process, and, consequently, it was discerned that the decrease of the 1720 cm’1 peak was due to the exposure of the specimens in the hygrothermal environment. When the 90 °C saturated specimen was completely redried by desorption of water at 100 °C for 250 hours, the 1720 cm“1 peak did not recover. This finding shows that the change is irreversible and may be associated with additives or oligomers leaching out from the polymer matrix. Also, a new 1666 cm’1 band is caused by thermal aging is shown. This peak might be due to an oxidation reaction during temperature aging. 159 V O\ 2 $2! :1 01 § dry - baked at 60°C for 2500 h 8 § 100°C for 250 h 2 — l 1, dry - as-received 45°C/9000 h \(f) 60°C/9000 h 75°c19000 h L(d) \. ll (C) 90°C/9000 h ”(b) “(a) - I 2000 1500 1 160 Wavenumber (l/cm ) Fig. 7.1 1 Infrared spectra of epoxy 934 at different hygrothermal stages. 160 7 .3 SUMMARY Three types of bound water exist in moisture-saturated graphite/epoxy (T300/934) composites. Depending on the water/resin interaction and bound-state characteristics, these types are designated as A-bonded water, F-bonded water, and @-bonded water. A-bonded water is described as physiosorbed water. Molecules of A-bonded water diffuse into the material and break the interchain bonds that exist initially in epoxy resin and form weak hydrogen bonds via Van der Waals type force. Most of A-bonded water is dispersed by interchain hydrogen bonds and only a small amount of A-bonded water is in a free-water state. The net effect of interchain bond breaking is increased mobility of the polymer chains and, consequently, enhanced matrix swelling and plasticization and lowering of the T,. F—bonded water is described as a chemisorbed water that interacts chemically with certain hydrophilic groups of resin matrix. The amount of F-bonded water is strongly dependent on immersion temperature and time. F-bonded water may form secondary crosslinking with the hydrophilic groups in the resin matrix. Consequently, it is contended that F—bonded water contributes to a slight increase in the T, and matrix-dominated mechanical properties of the composites. G-bonded water is characterized by its propensity to be retained at surface defects, cracktips, and voids due to surface tension and capillary effects. ®-bonded water contributes to excessive weight gain in the water sorption profile. G-bonded water is associated with anomalous, non-Fickian diffusion behavior. CHAPTER VIII CONCLUSIONS AND RECOMMENDATIONS 8.1 CONCLUSIONS The major results of this dissertation are concluded as follows. By water desorption measurement, desorption diffusivity calculation, amount of residual water determination, and NMR test, two types of bonded water were found in water-sorbed epoxy neat resins. Depending on the water-resin bonding characteristics, these types are designated as A-bonded water and F—bonded water. The A-bonded water is a physiosorbed water. The molecules of A-bonded water diffuse into the material and break the interchain hydrogen bonds that exist initially in epoxy resin and form weak hydrogen bonds via Van der Waals-type force. Most of physiosorbed water is dispersed and impeded by interchain hydrogen and only a very small (if any) amount of the physiosorbed water is in bulk (fi'ee) state. The F—bonded water is described as a chemisorbed water that interacts chemically with certain hydrophilic functional groups of epoxies. The amount of F-bonded water evolved of the total sorbed water content in the resin depends strongly on exposure temperature and time. These two types of bonded water influence hygrothermal phenomena in epoxy resins and in Gr/Ep composite materials. In addition, a third type of bound water was found 161 162 on the sru'face and near surface area of Gr/Ep composite. It is noted as O-bonded water. G-bonded water is characterized by its propensity to be retained at surface defects such as cracktips and voids due to surface tension and capillary effects. @- bonded water contributes to excessive water uptake, and, hence, immoderate water- induced weight gain in the composite material. (%)-bonded water is associated with anomalous, non-Fickian difiirsion behavior. The variation of T, of epoxy in a hygrothermal environment can be summarized as follows: i) the change of T, does not simply depend on the water content in epoxy resin, ii) T, is associated with the hygrothermal history of the materials, and iii) for a given epoxy system that maintains a constant maximum water content, the longer exposure time and higher exposure temperature cause higher T,. An interpretation is given for T, variation. Both physiosorbed water and chemisorbed water have an influence on T,. The experimental T, value is the combined result of the two mechanisms. Physiosorbed water decreases T, and chemisorbed water tends to maintain T, at its original value and in some cases slightly increases T,, Physiosorbed water decreases mechanical properties via the breaking of interchain hydrogen bonds and by increasing of mainchain mobility. Chemisorbed water does not decrease mechanical property. In contrast, it may slightly enhance strength-related properties via the formation of secondary-bond crosslinking with hydrophilic groups in epoxy. Swelling is essentially proportional to water gain. Physiosorbed water shows a significant influence on swelling because of its ability to interchain hydrogen bonds. 163 Chemisorbed water also creates some swelling, but compared to physiosorbed water the efl‘ect is significantly smaller. Moisture-induced expansion in composite was measured in length (fiber direction), width, and thickness directions. Essentially, no expansion due to water absorption was detected in the fiber direction dimension. Significant dimensional changes resulting fi'qm moisture-induced expansion were observed in the width and thickness directions of the laminate. The thickness decrease of the specimen at high temperature was associated with surface resin dissolution and peeling- Depending on the structural chemistry and processing conditions, epoxy resin systems exhibit a wide range of water saturation levels. Plasticization effects are quite different for difl‘erent epoxy systems. High water-absorption epoxy shows a significant decrease of both modulus and strength when the epoxy resin is saturated with water. Moreover, the degradation is temperature dependent. Higher exposure temperature induces lower modulus and strength. Low water-absorption epoxy resins show no significant modulus change, but a slight decrease of strength is realized. The most significant hygrothermal effect for all epoxy systems, in general, is the large decrease of failure strain. This change is irreversible. Also, the failure strain decrease is exposure temperature dependent. The most likely cause is due to chain scission during the hygrothermal process. Moisture-induced degradation was quite evident for the Gr/Ep composite. The delamination fracture toughness of the T300/934 laminate was lowered by approximately 60 percent after sustained exposure to hygrothermal environment. Plasticization of the bulk matrix and the interface phase by sorbed water 164 influenced the reduction in fi'acture toughness. Water-induced plasticization caused resin softening and strength loss. Moreover, water tended to weaken resin in the vicinity of the fiber/matrix interface. The degrading effect of water in the Gr/Ep composite was manifested by preferential cracking along the weakened fiber/matrix interface. Persistent fi'acture mechanisms appeared to be interfacial cracking, debonding, and microcrack growth and coalescence in the plasticized matrix material. Water sorption in graphite/epoxy composites exhibited both Fickian and non-Fickian difl‘usion behavior. Diflirsion data showed that the time for the onset of non-Fickian behavior was inversely related to the exposure temperature. Anomalous (non-Fickian) behavior in the composite resulted from chemical modification and physical damage to the epoxy resin. Cracks, voids, and surface peeling were observed clearly by SEM and optical microscopy. A crack/mass loss model favorably describes the phenomenological behavior of Gr/Ep composite resulting fiom water absorption processes. At a low exposure temperature compared to T,, there is minimal surface dissolution and physical damage to the material and the weight gain behavior is Fickian. However, with increasing exposure temperature, cracks, voids, surface peeling, and dissolution occur. Cracks can retain water which contribute to absorption behavior higher than the theoretical Fickian diflirsion curve. Surface peeling and dissolution contribute to reduction in the specimen weight and, consequently, the weight change profile data falls below the theoretical Fickian diffusion curve. The resultant weight change profile represents the combination of these two factors: cracking and mass loss. Depending on the dominant mechanism, the experimental 165 moisture absorption vs. t"2 profile(curves) may either be higher or lower than the theoretical Fickian curve. 8.2 RECOMMENDATIONS Many research methods, such as desorption measurement, desorption diffusivity determination, the amount of F—bonded water change, and NMR test, have been used to reveal the nature of water in epoxy resins. All the results suggested consistently that there are two different types of bonded water involving the water sorption in epoxy resins. But some key knowledge is still unclear and more effort should be pursued. Activation energy at both physiosorbed and chemisorbed water desorption stages needs to be determined. The determination of the activation energy will give a strongest and direct proof of the presence of the two bonded water. Temperatures of 40, 50, and 60 °C are suggested to determine physiosorbed water activation energy and temperatures of 120, 130, and 140 °C are recommended to determine chemisorbed water activation energy. Also, more experimental data should be collected in chemisorbed water desorption stage so that more precise desorption diffusivity can be obtained. An epoxy system with non-amine hardener is suggested using for water absorption and desorption measurement. Chemisorbed water is supposed bonding with hydrophilic group in epoxy resins. If there is no hydrophilic group in the material residual water should not exist. At low temperature all sorbed water can be removed easily and 166 quickly. This experiment will reveal very important information and provide direct evidence that the residual water is due to hydrophilic group of the material. BIBLIOGRAPHY 167 BIBLIOGRAPHY Adarnson, A. W., Physical Chemistg of Surfaces, 5th. ed., John Wiley & Son Inc., (1990). Adarnson, M. J., J. at Materials Science, 15 (1980) 1736. Antoon, M. K. and Koenig, J. L,Wm 19(1981) 1567 Apicella, A., Nicolais, L., Astarita, G., and Drioli, E., M; 20 (1979) 1143. Apicella, A., Nicolais, L., & de Cataldis, C., J. ofMembrqpe Sci. 18 (1985) 211. Apicella, A. and Nicolais, L., Adv. Polym. Sci., 72 (1985) 69. ASM International, Engm' eered Materials Handbook vol. 1, Composites, 1988. Atkins, P. W., Ma, Oxford University Press, New York, 1990. Atkins, T. & Batich, C., MRS Byfltin. 28-9 (1993) 40. Augl, J. M. and Trabocco, R., Workshop on Durability of Composites Materials, held at Battelle Memorial Institute, Columbus, Ohio, 1975. Avena, A. and Bunsell, A. R., Commsites, 19(1988) 355. Banerjee, A, Edie, D. D., and Oagle, A. A., Proc. Tech. Conf. Amer. Soc. Compos, 5 (1990) 295. Bank, L. C. & Gentry, T. R., J. QfReirgforced Plastics and Campos” 14(1995) 559. Barker, L. M., Eng’neering Fracture Mechanics, 9 (1977) 361. Barker, L. M., Eng‘neering Fracture Mechanics. 17 (1983) 289. Barrie, J. A., Sagoo, P. S., and Johncock, P., W 18 (1984) 197. Bellenger, V. & Verdu, J ., J. of Materials Science, 24 (1989) 63. 168 Biro, D. A., Pleizier, G., and Deslandes, Y., W 46 (1993) 293. Blair, c. and Zalcrewski, J., 11111. SAMPE 19.91;, mg, 22 (1990) 918. Blikstad, M., J. W., 5 (1986) 2. Blikstad, M., Sjoblom, P. O. W., and J ohannesson, T. R., W Commsite Materials, Vol.3, G. S. Springer ed., p. 101, Technomic Publishing Co. Inc. Westport, CT, 1988. Bonniau. P. and Bunsell, A R.. W Vol. 2, G. S. Springer ed., p. 209, Technomic Publishing Co. Inc. Westport, CT, 1984. B011, D. J ., Bascom, W. D., and Motiee, B., W., 24 (1985) 253. Bradley, W. L. and Cohen, R. N., Delamination and debonding of Materifls, p 389, W. S. Johson ed., ASTM STP 876, 1985. Brewis, D. M., Comyn, J ., Shalash, R. J. A., and Tegg, J. L., Polymer. 21 (1980) 357. Brinke, G. T., Karasz, F. E., and Ellis, T. 8., 111mm, 16 (1983) 244. Browning, C. E., Polymer Engineering and Science. 18-1 (1978) 16. Browning,C..,E l1 {1'11"_11111' 11:1 @111."1'.11_1-11 W p 231 J- L Koenig and L Nicolais ed .. Plenum Press, New York, ( 1983) Cai. L W. and Weitsman, Y., W ICCM-9. Vol. 5, p 509, Woodhead Publishing Ltd, Zaragoza, Spain, 1993. Cairns, D. S. and Adams, D. F .. Waals G. S. Springer ed., p. 300, Technomic Publishing Co. Westport (1984). Cappelletti, C., Rivolta, A., and Zaffaroni, G., J. at Camp. Tech. & Refserch- 17 (1995) 107. Carter, H. C., Kibler, K. G., and Reynolds, J. D., Advgced Commsite Materials- Envirgnmental Effects, J. R. Vinson ed., p. 84, ASTM STP 658 (1978). Chateauminois, A, Chabert, B., and Vincent, L., Polym. Comp, 16 (1995) 288. Chu, H. S. and Seferis, J. C., Palmer Cornflites- 5 (1984) 124. 169 Chua, P. s., SAMPE Q. 18 (1987) 10. Couchman, P. R and Karasz, F. E., Macromolecyfl, 11 (1978) 117. Crossman, F. W., Mauri R. E., and Warren, W. J., Advanced Composite Materifls — Environmental Effects, J. R. Vinson ed., p. 205, ASTM STP 658, American Society for Testing and Materials (1978). d’Almeida, J. R M., Composites- 22 (1991) 448. DeIasi, R and Whiteside, J. B., Advanced Composito Materials — Environmentg Effects, J. R Vinson ed., p. 2, ASTM STP 658, American Society for Testing and Materials (1978) DeNeve, D. and Shanahan, M. E. R., flow 34 (1993) 5099. Dewimille, B. and Bunsell A. R., Composites, January (1983) 35. Dynes. P. J. and Kaelble. D. H..Cnmp_csite_Materia15_Iesiirigaad_Desiaa S. W. Tsai ed., p 566, ASTM, Philandalphia, 1980. Ellis, T. S. and Karasz, F. E., flow 25 (1984) 664. Etxeberria, 1.. Franco, J -, and Valea, A. Wigwam 14 (1995) 988. Farra, N. R. and Ashbee, K. H. G.,.l. flu. Q:Ap,pL Elm... 11 (1978) 1009. Farra, N. R. and Ashbee, K. H. G., W ACS Symposium Series, Vol. 132, p 435, C. May ed, American Chemical Society, Washington, DC, (1980) Fedors, T. R., Pobmers, 21 (1980) 713. Frassine, R. and Pavan, A., Comps. Sci. and T ech., 51 (1994) 495. Garg, A. C. and Ishai, Fracture Mechanics 22 (1985) 423. Gazit, S., J. Am]. Pom. Sci, 22 (1978) 3547. Ghorbel, I. and Valentin, 1)., Poo. Como, 14 (1993) 324. Gordon, J. M., Bouse, J. B., Gibbs, J. H., and Risen, W. M., ,1. Chem. Phys” 66 (1977) 4971. 170 Grayson, M. A. and Wolf, C. J., Proc. of the 5th International Conference on Commsite Materials ICCM/S, p. 1463, ed. by Harrigan, Jr., W. C., Strife, J., and Dhingra, A. K., San Diego, 1985. Grayson, M. A and Wolf, C ., J. ofPolymer Sci. : PaLt B Polymer Physics, 25 (1987) 31. Gupta, V. P., & Drzal, L. T., J. 0f Applied Polymer Science. 30 (1985) 4467. Hahn, H. T., ,1. (lamp. Mater., 10 (1976) 266. Hahn, H. T., J. Eng. Mater. Tech, 109 (1987) 3. Halpin, J. C., The Role of the Polymeric Matrix in the PrMssing sod Stmcture Properties of Composite Materials, J. C. Seferis and L. Nicolais ed, p 3. Plenmun Press, New York, 1983. Helander, R D. and Tolly, W. B., Nat. SAMPE Smposium, 29 (1983) 1373. Illinger, J. L. and Schneider, N. S., V_V_ater in Polymers, ACS Symposium Series, Vol. 127, p 571, S. P. Rowland ed, American Chemical Society, Washington, DC, (1980). Imaz, J. J., Rodriguez, J. L., Rubio, A., and Mondragon, 1., MW Letters, 10 (1991) 662. Jelinski, L. W., Dumais, J. J., Cholli, A. L., Ellis, T. S., and Karasz, F. E., Weeks. 18 (1935) 1091- Johncock, P. and Tudgey, G. T., W 18 (1986) 292. Jordan, W. M. and Bradley, W. 1..., Ionghepedfiomposites, N. S. Johnson ed., p 95, ASTM STP 937, 1987. Jost, M., Diffusion in Solids, Liquids, and Gases, Acadamic Press. 1960. Karasek, M. L., Strait, L. H., Amateau, M. F., and Runt, J. P., J. of Camp. Tech. & Research 17 (1995) 3. Karasek, M. L., Strait, L. H., Amateau, M. P., and Runt, J. P., J. at Camp. Tech. & Research 17(1995) 11. Kelley F. N. and Bueche, F., J. QfPolymer Science. L (1961) 549. Kong, E. S. W. and Adarnson M. J., Polym. Commo 24 (1983) 171. 171 Kuzenko, M. V., Browning C. E., and Fowler, C. F., Resins for Aerospace, ACS Symposium Series Vol. 132, C. May ed., p 275, ACS, Washington, 1980. Kwei, T. K., J. Appl. POM. Sci, 10 (1966) 1647. Law, G. E. and Mlkins, D. J ., Delamination Fracture Qritiria for Composite Strumres, NAV-GD-0053, Naval Air Systems Command, Washington, DC, 1984 Lee, C. Y., Pfeifer, M., Theompson, B. S., and Gandhi, M. V., J. Comp. Mate; 23 (1989) 819. Lee, M. C. & Peppas, N. A., Prog. Fob/Ln. Sci, 18 (1993) 947. Lee, M. C. & Peppas, N. A., .1. Qngplied Polymer Science. 47 (1993) 1349. Lee, S. B., Rockett, T. J., and Hoflinan, R D., Polyper, 33 (1992) 3691. Leung, C. L. and Kaelble, D. H., Resins for the Aerospace, ACS Symposium Series Vol. 132, C. May ed., p 419, ACS, Washington, 1980. Levy, D. J. and Arnold, C. R., Not. SAMPE Symposium, 29 (1984) 1326. Loos, A. C. and Springer, G. S., J. Comosite Materials. 13 (1979) 131 loos, A C and Springer. G. 8.. in W G. S. Springer ed., p. 34, Technomic Publishing Co. Westport (1981). Lubin, 6., lisndbook of CompositesNan Norstrand Reinhold Co. Inc., 1982. Lucas, J. P. and Odegard, B. 0, Advanced in Thermalplastics Composites, G. Newaz ed., ASTM STP 1044, p 231, 1989. Lucas, J. P. & Zhou, J., Proc. of the 9th International Conference on Composite Materials, Vol. 5, p 633, Woodhead Publishing Ltd., Zaragoza, Spain, 1993. Lucas, J. P. & Zhou, J., JOM 45-12 (1993) 37. Lucas, J. P. & Zhou, J ., Proc. of the 10th Internetiofinal Conference on Composite Materials, Vol. 6, p 247, Woodhead Publishing Ltd, Canada, 1995. Marsh, L L., Lasky, R., Seraphirn, D. P., and Springer, G. S., W onQomposijeMmerials, Vol.3, G. S. Springer ed., p. 51, Technomic Publishing Co. Westport, (1988). McKague, Jr., E. L., Reynolds, J ..D., and Halkias, J. J. Qngpl. Polymer Sci. 22 (1978) 1643. 172 Mehring, M., Prinoiples of Hoigh-Resolution NMR in Solids, Springer-Verlag, New York, 1983. Mijovic, J. and Weinstein, S. A., Polym. Commu_n_., 26 (1985) 237. Mikols, W. J. and Seferis, J. C., We; ACS Symposium Series, Vol. 132, p 293, C. May ed., American Chemical Society, Washington, DC, (1980). Mikols, W. J ., Seferis, J. C., Apicella, A., and Nicolais, L., Polym. Cpomosites. 3 (1982) 1 18. Morgan, R. J. and Mones, E. T., Resins in Aerospace, ACS Symposium Series Vol. 132, C. May ed., p 233, ACS, Washington (1980). Morgan, R. J. and O’Neal, J. E., J. 01 Materials Science, 15 (1980) 751. Morri, T., T animoto, T., ComaSciind Tech., 49 (1993) 209. Monis, G. E., ASTM STP 696. B. Pipes ed., p 274,1977. Moy, P. & Karasz, F. E., Polymer Eng’neering and Science, 20 (1980) 315. Rao, R M. V. G. K., Chanda, M., and Balasubramanian, N., Environmental Effeots on Composite Materials, Vol.2, G. S. Springer ed., p. 179, Technomic Publishing Co. Westport, (1984). Rao, R M. V. G. K., Kumari, H. V. S., and Raju, K. S., J. of Reinforced Plastics and Cowsites. “(1995) 513. Rehage, G. and Brorchard, W., The Physics of Glassy Polypaers, p 54, R. N. Haward ed., Wiley, New York, 197 3. Sahlin, J. J. and Peppas, N. A., 1nd Eng. Chem. Res, 9 (1991) 211. Schutte, C. L., Materials Sci. and Eng, R13 (1994) 265. Shen, C. H. and Springer, G. S., 1. Composite Materials. 10 (1976) 2. Shell. C. H. and Springer. G. S..EaummnamaLEff2915_Qa_CQmp_Q5nc_Maienala G. S. Springer ed., p. 15, Technomic Publishing Co. Westport (1981). Shirrell, C. D., Adv, Comp. Mat.— Environmental Effeots, J. R. Vrnson ed, p. 21, ASTM STP 658, American Society for Testing and Materials (1978). 173 Shirrell C. D. and Leisler, W. H., ondestru iv Eval ation an Flaw Criti i f r Composite Materials, R. P. Pipes ed., p 209, ASTM, Philadelphia, 1979. Spathis, G., Kefalas, B., and Theocaris, P. S., Interrelations Between Processingfitosotm and Promrties of Polmeric Materials, J. C. Seferis and P. s. Theocaris ed., p 391, Elsevier, Amsterdam, 1984. Springer, G. S., Enviropmental Effects on Composite Materials, G. S. Springer ed, Vol. 1, p 126, Technomic Publishing Co. Westport (1981). Springer, G. 8., Environmental Effects on Composite Materials G. S. Springer ed., Vol. 1, Technomic Publishing Co. Westport (1981). Springer, G, S., Environonental Effects on Composite Materials, G. S. Springer ed, Vol. 2, Technomic Publishing Co. Westport (1984). Springer, G. S., Environmental Effects on Composite Materials, G. S. Springer ed., Vol. 3, Technomic Publishing Co. Westport, 1988. Waterbury, M. C. and Drzal, L. T., J. ofReigforced Plastics and Composites. 8 (1989) 627. Weitsman, Y. J ., Effects of Fluids on Polymeric Composites - a Review, Reported to Office of Naval Research N00014-90-J-1556, 1995. Whitney, J. M. and Browning, C. E., Adv. Comp. Mat. —Environmentng_fl‘gs, J. R. Vinson ed., p. 43, ASTM STP 658, American Society for Testing and Materials, 1978. “filliams, M. L., Landel, R. F., and Ferry, J. D., J. Am. Chem. Soc. 77 (1955) 3701. Wolff, E., SAMPE Journal, 29-3 (1993) 11. Wong, T. C. and Broutman, L. J., Palm. Eng. Sci, 25 (1985) 529. W00. M- and Piggott. M., W. 9 (1987) 101- Wright, W. W., Commsites, p. 201 (1980). Xiang.Z D andJones F R MWWM MaterialsICCM;2, Vol.5, p. 601, Woodhead Publishing Ltd. Zaragoza, Spain 1993. Zhou, J. and Lucas, J. P., hoe, of the 2:11; Conf, of the ASC, p 1213, Technomic Publishing, 1994. 174 Zhou, J.and Lucas, J. P., Cmsites Sci. and Tech. 53 (1995) 57. HICHIGRN STRTE UNIV. LIBRQRIES 11 l IHI Llllllllllfll III 312 30 42 5457