<%BANNER%>

Effects of Dislocations on Electronic Properties of III-Nitride Materials

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20110320_AAAADS INGEST_TIME 2011-03-20T21:01:17Z PACKAGE UFE0009281_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 47554 DFID F20110320_AACEEJ ORIGIN DEPOSITOR PATH west_a_Page_113.pro GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
214f380307c51988ab689f2ca5daf38e
SHA-1
057d8ba4db05d177d35c7422112b15af0645ad0d
23620 F20110320_AACDZD west_a_Page_083.jpg
50aa86207c3bfefd806a09762e451a05
4ca1dd0ddc92cb47f07b129f91733df4b35934e6
6482 F20110320_AACEDV west_a_Page_094.QC.jpg
8131f120cbb334f8fbe0d27e1dbc28c4
4c122db165c43f9044aa8b7233824fceeda57850
479882 F20110320_AACDYP west_a_Page_081.jp2
807ddd64af34a8ab9f75e7c6fb8f3520
044a766cabb578eceb8d9d635bef4b57c2b7e641
94302 F20110320_AACEEK west_a_Page_013.jp2
ff29390f77c3c6f0e25417eac31fccf9
512bbdcf00a5211c2a784ac75a24617358fdb07e
102847 F20110320_AACDZE west_a_Page_069.jp2
aa22ac1cb4fef9859f7211cf15db4e18
c08384a40d8c3c6d37028d660cbc80b31fb64f7f
194 F20110320_AACEDW west_a_Page_105.txt
6dea645807a25f5aeda9297357fad618
f0b83b6ad58f8a89a02ea470c92c626bb45799b4
25478 F20110320_AACDYQ west_a_Page_058.jp2
00ea5de9b0a32083b1ceeaf1877ee99b
36d105887f8500002fa24cb73a88b196565169ac
25271604 F20110320_AACEFA west_a_Page_064.tif
f22b93e15070d888d2c8d7d4c398b12d
53f6e94cf6dd11758a70dfec94e477deeda86c7f
F20110320_AACEEL west_a_Page_099.tif
3f223ac35c7ca5ec8db0c35256b233a8
cef4ca56459122c02f1b0d9d21265bca74d8525f
1053954 F20110320_AACDZF west_a_Page_082.tif
0389a526bc3faddc77340cf76c8273f6
a3186128a7bf9189fa97b5789e5944e53f5c5a06
758357 F20110320_AACEDX west_a_Page_034.jp2
dae933707ac88271cc41e70687f3b846
154da2c6617b47488aea4523a7e8b5bb1d089129
30670 F20110320_AACDYR west_a_Page_021.jp2
e0f91d20f3c7556f47e0b7d0ed02e209
6318f6dfa095141b134c4984971a0873cfd66350
34570 F20110320_AACEFB west_a_Page_011.pro
413501afb5f5d1327d33f55d854d0a35
8f15066929a7ced593ca66979ddf0427cccc5de2
6261 F20110320_AACEEM west_a_Page_070thm.jpg
60ef5e1a156033c4be95b6932e5f8a78
dd7ff4085fecd6b3775d9887830efb53de87460a
542 F20110320_AACDZG west_a_Page_064.txt
c3564f8df6aaecca2bca243a47139cfe
a1c866ab8f9eaf55e83fea37010fafcc611d8d2a
940790 F20110320_AACEDY west_a_Page_035.jp2
44410cba3150272d999119e84fa7b19a
5cc1ed90e460ba4f1b3a8c93916a14da17e33362
2257 F20110320_AACDYS west_a_Page_094thm.jpg
5e090a04abf5735b86ec917b369ae2eb
5183e98cffcdbfeb874ef15367d934daee8c3557
23020 F20110320_AACEFC west_a_Page_085.QC.jpg
7d9ec7e5cb8dc32876c2a3d56a2ef64d
f7503304580b8ceeda016494d82c1009438b2eae
68260 F20110320_AACEEN west_a_Page_069.jpg
89129387aeebbfaacb3ec3c877cf9a32
2f0434f14f9347604400a467c86f9694f036de43
23376 F20110320_AACDZH west_a_Page_098.jpg
c87e53e0f0d362bdb3e11dac2c475c0b
bb81f2857768947542e36ae4f52d07b7759526d1
261 F20110320_AACEDZ west_a_Page_102.txt
ce2896f5550e001a5e98424e07a7d364
b2e9917a260eb733754f2007f45705b4a045d8eb
9917 F20110320_AACDYT west_a_Page_097.QC.jpg
9bd444acefc20eaa5f91c653cf395b90
e9afa4b62ad5176093cc0b86979e7612f5dbe3a9
33726 F20110320_AACEFD west_a_Page_004.jpg
2bebf1f2f2b66352616f2d2bc6c45e81
de0b03fd09d7a2f03d7abc656ff4faea8f528e9c
42560 F20110320_AACEEO west_a_Page_086.pro
60401e419517b91f9d9ba309474d6d5c
25e7b5ee1588e709391083695c416f9b35a225a1
12917 F20110320_AACDZI west_a_Page_002.jpg
6ff35fee1aaf0ba30c3582c2adcb81a0
397fcb313bbc6b593ba39a347dde56ce4aca09b0
4036 F20110320_AACDYU west_a_Page_006.txt
76e43322a8454dcc17836cb1b560618c
ced2546d0023f0af681ca9fae9958cc076b7e316
20364 F20110320_AACEFE west_a_Page_081.pro
a5f383c1e80b1ce64c052b276832fe59
ad53925c2205f4d917c004a0aae0578c04b55330
5471 F20110320_AACEEP west_a_Page_106.pro
8a8f4177e64ec303aa39cf2f32205a71
7c5db9e1254b4ae4dfc74504664e9c26b1880c2e
184 F20110320_AACDZJ west_a_Page_002.txt
7c00638fefcede05c5fa097f723e3a88
9b8c12fba27445dbb7d03e4078fe838ad143ffbd
1842 F20110320_AACDYV west_a_Page_045.txt
b2e74ed150506af8090aa8bd147bfb96
64b8c772d0c30b28367b8ca13f9682572042fcab
4588 F20110320_AACEFF west_a_Page_075thm.jpg
16f33d1322a2faab45cde6d643bdd098
3424f30167a99a9cb070253b973b19d95d995c32
8554 F20110320_AACEEQ west_a_Page_062.QC.jpg
bb598b6f693c3605b5bdaedb6243af7b
0feb0177deb2bf94c530b5cfe250e13254c45124
342 F20110320_AACDZK west_a_Page_061.txt
633e0dccb3030f4017e955e453513490
e9afcb69ebb1fd2e44faaf31547e47eb252dce7c
91048 F20110320_AACDYW west_a_Page_031.jp2
e5f600c7c1ee25c0dfe59fa2cde6eea4
21f4a525114d73ba054a7a2595b4771bcac084eb
1733 F20110320_AACEFG west_a_Page_027.txt
3821eb36d5aa2d8fd23b538cde81d80a
df1ad01917d3b0fd7d8615d00c10dd3dcb19717f
200 F20110320_AACEER west_a_Page_052.txt
0e37c5bf293db6895a9d21a047729d66
f2a0e6cc9b9b2302ad68e974c80de8729664f57c
14726 F20110320_AACDZL west_a_Page_075.QC.jpg
a03a9d87de73344434a5fe6b2945fa95
b04496561f03f9c0c593d0f0f702673eede09a88
47057 F20110320_AACDYX west_a_Page_070.pro
e24ef80e4f8a6a4c3799802ee45d7481
6f6c3d1ef6e8abca5ad8401888fce6adabbbe32c
4737 F20110320_AACEES west_a_Page_073thm.jpg
084e79cfaa0b1b263b0ee9466498caf3
f20af542fb7f5a6dd38072ab005d06194bb466f2
189 F20110320_AACDZM west_a_Page_100.txt
d7380b67b194d108c61ce2e6a3fbcbd6
a6979cdc15d0ae29e3082e3a7bbdf751ab55b123
8908 F20110320_AACDYY west_a_Page_096.pro
6ebeecfaf5eac864a9ca039e0153b329
a84ecb4296c939aa18436aa697428fc1526db884
35223 F20110320_AACEFH west_a_Page_029.pro
918b47168ba4ead2a6dd57c72cb495ad
fa6dc342ec7cf44023d76565ef9c7371e8929303
10429 F20110320_AACEET west_a_Page_079.pro
030f7028eddf9ec2f6bad5d9a2ce2b6c
d9b3ca05535c9aca1fd2c372d7952bceb3a1f46e
673 F20110320_AACDZN west_a_Page_056.txt
83785c056d6503553ae7065c95c51c3b
8182e6e9dc92e4fdff9a3064b6d2bbfad8840eae
21296 F20110320_AACDYZ west_a_Page_060.jpg
8d528ede094dbd56598f9b880648661d
0f5d8c9f7f38fc44527d81948da6aa3227de415d
F20110320_AACEFI west_a_Page_019.tif
0d3f76b7ec917c52a35968373ac8eaca
bfa48557aa64b2daa3c78fcb0efc289c927666b1
217323 F20110320_AACEEU west_a_Page_060.jp2
88acf150e814f90f086b9915900c7c8b
2c6c302b41af257601591f890292fa0fbde90efc
51087 F20110320_AACDZO west_a_Page_110.pro
248331d6177d88a1eaec44d0fd8a1bfc
d977b134801fa86607b2f4a0472705166466689a
F20110320_AACEEV west_a_Page_107.tif
5db3d9d5915b066ec7db42776d48242b
732f2fe1d2f1b0b13e8ccb623d2fbe50a5e48fe1
1317 F20110320_AACDZP west_a_Page_057.txt
982b437bee517469f2b83af32851fa61
347264c0c0cad7324fae7387c0dae4f48c176b29
950 F20110320_AACEFJ west_a_Page_050.txt
6ada7c0b667a36e5f19c361815f58ffa
b8f8a7fded3c6af485624030ece6f7af9e645360
F20110320_AACEEW west_a_Page_100.tif
9eae1f96bc7b5a89b47a19f61c55b648
f30a4c0b06fb0c983e90caf5d5dece79b10eaa93
453797 F20110320_AACDZQ west_a_Page_106.jp2
fe9586e059773dcdda3eed52f75d14b3
e702352d9c6a3728dff9adc81081d68e538cafc9
22382 F20110320_AACEFK west_a_Page_046.QC.jpg
f327078f6b9d62334239851a46785ecf
5e0123606071690f497b2ce2d79279d95b39020e
1722 F20110320_AACEEX west_a_Page_086.txt
8af0be800b3274e4b664277535c290c8
66f2bf322fcbf3ac2b60635b2a2663ed867bed81
3117 F20110320_AACDZR west_a_Page_097thm.jpg
a80df244f7b16167abede314acf89db7
aa176e5a268bab433df8c05b57c4a06749be9910
F20110320_AACEGA west_a_Page_010.tif
85937436f9474ffaf5c758a8df53fa4c
658f42a932838e0a44ee2268e831ef8eae3e1821
51467 F20110320_AACEFL west_a_Page_111.pro
e0e8d12f0a5b586d1c5897266a263085
b10de528dc1ddbbb43a0fbc35710a149011c0e7f
72458 F20110320_AACEEY west_a_Page_090.jpg
95d478c3443a315e226642bca7f26e98
3b1f4fc9296013249c68b1535fa2c73631a6586d
242 F20110320_AACDZS west_a_Page_053.txt
d3017257dab902f085d66e6db768d941
a0c04d2e6a747a9639a8603e5d55823bc7296ed9
F20110320_AACEGB west_a_Page_011.tif
ec32a16e884274055d4ee00f1de6a38e
286cc5aeedd3c15745a272c3efd587fb989d1820
69112 F20110320_AACEFM west_a_Page_016.jpg
bf7a0559acc63191accacd3a92babb38
ce0b489932c8c7e363a3e5635fbd3469c1d69aa7
F20110320_AACEEZ west_a_Page_062.tif
52bc16d53bc8d1f22a22018c5dfdfaf8
968a3f39c95c5e8e14855bbcdde8eb55187cc0d9
F20110320_AACDZT west_a_Page_068.tif
f1b7f41a7b417b343796f10e241ab6cd
dfa7aed4c6025802ee83d74aa73c784522b5d06e
F20110320_AACEGC west_a_Page_012.tif
058f26e0df820d454a7868d6ef4e0b30
e977891ad0fd390e55a375d771e323c22402c822
1721 F20110320_AACEFN west_a_Page_003.txt
6df2acc6bcbfad932128501ed4b5a8c4
3932ba13a064f19d5b28522a181cd45ea1d84f4c
7216 F20110320_AACDZU west_a_Page_105.QC.jpg
9b6962c1c1eb29a79ff431c7a40e8a2d
36eda6871af93188f7635336f53dbd2e80f80923
F20110320_AACEGD west_a_Page_014.tif
f898b1f0ca74072e19f2a146b40ddebe
4931e01dfb1a6e34c270a69589daa06baf213cc7
F20110320_AACEFO west_a_Page_031.txt
6c34548813a9cacd284d62d141dec26e
779c8f6b0e089fbbd363b02edc2bffbb37a1b6a1
21672 F20110320_AACDZV west_a_Page_026.QC.jpg
f4fb0e6b690fb7400d6fc0b6950ecc06
c17f115683c888db76b10778ad8cb5eb47591239
F20110320_AACEGE west_a_Page_015.tif
2f739dd59fba83000d8159d756424fe0
a69027b75068a88f1f9e4674829b0c736c4f9b6d
1801 F20110320_AACEFP west_a_Page_068.txt
45a18ad6a6bb8826e507c07d2c5991a2
ec67fc107a05deea8d9d3e0ef6addce263930ce7
F20110320_AACDZW west_a_Page_067.tif
81384007ce41558b4acf70e9983c03a7
cb2ab79625035542e07fa8e4c3a548604b712767
F20110320_AACEGF west_a_Page_017.tif
fa9aeb2905dcae2ab14a1243740fa83e
ada973088cae49ac54911a361012a3824393e822
6362 F20110320_AACEFQ west_a_Page_052.QC.jpg
a0764ec952e717d2bd1bcb220ca994e4
da0ddf586e6e211a69018db319b4a9bcd3740afc
61637 F20110320_AACDZX west_a_Page_019.jpg
c7637e49faab2edeb43a37c674302f53
fc74620bf5985f704db9aed8404335c32e2af05d
F20110320_AACEGG west_a_Page_018.tif
e8161a9b7fe08fdff66b78efe977a2a2
98f430bec471fcd87c78c4bc7db8ff205bc123fd
48403 F20110320_AACEFR west_a_Page_014.pro
c8b78ad7a663b9545ec6fb55cd041fc9
20adcb6cad82dea4b88e4370b89689770818a7b5
68148 F20110320_AACDZY west_a_Page_045.jpg
aa78b0b0884ed8da4fa065c196da66b6
7c348d247d9e148460b14e0942116e7f769e7734
F20110320_AACEGH west_a_Page_021.tif
f4ab553973b40d3883ed6a0c7e110ec3
53057aae235da9a04a5250f51f4b77e808de92e9
94412 F20110320_AACEFS west_a_Page_009.jpg
159ce077c9afc0a63649ad70d2a89389
3a47a2d3470656b8cbb19a4eb7c068115c53b4ee
1051984 F20110320_AACDZZ west_a_Page_009.jp2
eb94ed4ab8e32e8ffb69ab1474673d15
2f0231f2b20eb28482237448e65663d1751db4f4
F20110320_AACEGI west_a_Page_022.tif
7f29958348662d28e88fae7d1fdcb603
6079febefb0111ab257a8154df4dd76bb5c54e0b
187482 F20110320_AACEFT UFE0009281_00001.xml FULL
6e75455c6f727f877ef11b253a2aaf0d
fbebc9ad668ffff25008fdfc95ce8cd7d6f46fd1
F20110320_AACEGJ west_a_Page_023.tif
354945644d74599c0dc3b61f3168598f
ae43c0b39926d0d107f01ac4cca6c92cc9a3b827
F20110320_AACEGK west_a_Page_024.tif
0691a3fffd612910b2a3567b32ee4ec9
e46a0c0475caeacca20045b898d5047dfa8299b3
F20110320_AACEFW west_a_Page_001.tif
976573f7df80c441be9eea71b9fbae92
e87e7cc4168ac067aaf64155347693410d68f1ca
F20110320_AACEHA west_a_Page_048.tif
ffc0aa3f11434067d069538f489735c6
262c3f5c1d602b34980d20e4c93352e85c92402f
F20110320_AACEGL west_a_Page_026.tif
120ca7a0ea52af17a8e0f9eed573ca58
b3f3b526b85890810f852daf380b0d37524121fe
F20110320_AACEFX west_a_Page_005.tif
9e51bd145f502146d1499f52e8bf5bb7
8e36b8ecb11765694bbe19270d753b205f726e38
F20110320_AACEHB west_a_Page_049.tif
9d85dc8abfd9f0e7d9a3537e128785c0
c4ddf2f0ccbf8310302e067e6f5b617d3f5ef907
F20110320_AACEGM west_a_Page_027.tif
cb0b446f82b0e0190b32ae17ff992cb7
37f6978bbb0c32dd75722938886f6d55c7ec8324
F20110320_AACEFY west_a_Page_007.tif
4362d45c7ffc8173eee22e9121ff4aff
94ebd2a31528dbce35e4bc196e76a0d3ccbc9e6e
F20110320_AACEHC west_a_Page_051.tif
82a68b490007662caa14aa58dc64d547
1d367fafb69da318caf38cc6c6f11a5965ce841c
F20110320_AACEGN west_a_Page_028.tif
da2bd8c8c132b54eec4dc976f746a1a0
bc9f8d5a34bdfc541cf241b1b390e3edb52a8fb7
F20110320_AACEFZ west_a_Page_008.tif
de921061a13f75344fa63cb091dd9f29
3d059141b9cf9403d08e13fd2b24490dbee1f0d5
F20110320_AACEHD west_a_Page_052.tif
a9c7bddc26f9ab2b3d9009804752ebaa
095be1a9c75ed07892a00cfa7ee2f5469003a181
F20110320_AACEGO west_a_Page_029.tif
78d75786587f9ff9febe68eb787920ef
1c627ad1e8bf71af85d4b2029ce0fca9f33fd2e7
F20110320_AACEHE west_a_Page_053.tif
8b67899ee49b12d50f599a4d76d27e0d
6ed5f6f025bade0b0af15254f8e31a89444b3605
F20110320_AACEGP west_a_Page_031.tif
3b57e694a2cb97306a1e8b9d63ea246a
1e776409a25f64fde4456db5c3c7bfc0419f3f70
F20110320_AACEHF west_a_Page_056.tif
54cbe83029995a477704ce0099dc6c05
128a9f28d670690da38017ba3387b3d4d3691bff
F20110320_AACEGQ west_a_Page_032.tif
42aa0edda6686f46c24f03dbf0654a6a
d56a43b12e7dfc750163a0e03e35c06342fa9eae
F20110320_AACEHG west_a_Page_057.tif
ac8092a5edc966ac22c59e69c83f6d1a
1a947f4df6f887d1b868625c6114d34b01b3b38b
F20110320_AACEGR west_a_Page_034.tif
bdbb62911e43f0537f4c0e834b16366a
6d8591d13537b1d78db9116736d6f76a1770ed98
F20110320_AACEHH west_a_Page_058.tif
2a4652841a4c1edb77b94f1a9361447a
8126fd9ba1a41d1bc0ad53b65f14bc96c0f5fbf0
F20110320_AACEGS west_a_Page_035.tif
c0b52e27a44e449e93ce0694cb231771
4ea76d85fd3531dc32071a1174e20ef0df4b6cb9
8423998 F20110320_AACEHI west_a_Page_060.tif
0b4527f1abf61b2af2b54762b641d351
b65d45c3120a0d25f58d936000192a8a41d87c33
F20110320_AACEGT west_a_Page_038.tif
e0ddf40128e41c21203f399dfd3e84c7
429eb6eef1ad7c7edfdc5269095fbe3c9545f011
F20110320_AACEHJ west_a_Page_061.tif
4a462f96ceeb7fc78c9bb1b67ed994f7
7c32224422e3841f86cfd762b19c2f0212cc7dfd
F20110320_AACEGU west_a_Page_039.tif
07d18b8d45bcd13cefbc4568577e3bba
35b76f80b4268d34141448036bd932ea37858e2b
F20110320_AACEHK west_a_Page_063.tif
8858b165c32bcab4a1c68b299ac950fe
a2ad7d6e2ea5d0ef65bbfe3948c3198156e50943
F20110320_AACEGV west_a_Page_040.tif
3e92766ea5b18adafad86f7a5b9e76df
1d2e9ddf2f5e539da2f2b39f9ae1b0911c382162
F20110320_AACEGW west_a_Page_041.tif
bc5caa031534b404fdaca52f4b3dd993
e5a0e07d932f59b98ca1b03993d5098d95f57197
F20110320_AACEHL west_a_Page_065.tif
80520808aeb5df824efc1262f42cf712
7fa73cdf4a29627f5b340144d026d975689db38d
F20110320_AACEGX west_a_Page_042.tif
1324f6189ae48148b2f3af79e91f1a3a
25e3e4e6bf7985179173786b1ee3eca36f57b563
F20110320_AACEIA west_a_Page_086.tif
b3434d64cf46ab737d45103d166e67a5
e9bc36e6eb49ac58542e976e9a60dd12239108e0
F20110320_AACEHM west_a_Page_066.tif
dbc3ecf31fe6e740965b091d7bc8f4d1
e5a367ce38f81e5d0db06d9d448305ec24d3c555
F20110320_AACEGY west_a_Page_045.tif
083d23edf146ff25e5773b67d8d85842
987ab344c73a41a2903288b0079c77beae1ac791
F20110320_AACEIB west_a_Page_089.tif
134700628d20d10d90b93c91dd31ef76
50fda98e6d7b1cbcb65a8c7130cad15b2a0a7e4b
F20110320_AACEHN west_a_Page_069.tif
d7fe809a641e8215cf6ab00eb661aae5
27f3de3969b09acd6cf571c932ed55fbf95e5323
F20110320_AACEGZ west_a_Page_047.tif
f91aff0c965c417cfc87da06e9e6854c
f44aa3e6256e2fd41f8c2f7ca06a8d11a8fc0296
F20110320_AACEIC west_a_Page_090.tif
5322cbf472b1a7ae0b84ee56c347f72d
2fd84f774d2eeab2605affa5a7ca860725e44748
F20110320_AACEHO west_a_Page_072.tif
e4d6f0086de778cae193d78125dcf0a2
7df1bd35bf7742624f0d644e4d199500bc5c55cd
F20110320_AACEID west_a_Page_092.tif
16437f3a1fbc76fbea618c2a6d336e32
c963e9a792c23b628be85238e1a50904090524c7
F20110320_AACEHP west_a_Page_073.tif
bb832d3d0c8b8625112f7b0fb33d8c9d
bbe0db893397a8688bce1739e6769f8a5862acde
F20110320_AACEIE west_a_Page_094.tif
aef96250a37b5a10a069f98dcc4d9715
38a9c213f8cc144d053570a211241037ecb0c449
F20110320_AACEHQ west_a_Page_074.tif
25b5a33f023cda8df7f8de6f192af859
ed05ff4922777c46a1ba019d8fa9ef567de0a2a3
F20110320_AACEIF west_a_Page_095.tif
cf7f60e26be64d4d0efa2592824d932a
5250ba34e932c07515e05b85d8aeb2c752a24515
F20110320_AACEHR west_a_Page_075.tif
1c52aec8d7c85a5cf5b8fe5caabc8419
dfef3821a6cf68330d25747a8596657bd9f0dabb
F20110320_AACEIG west_a_Page_096.tif
3cf03f67d61484c80867c9f1b02e5782
9476d9eb3b5654270ff20959fca3f306821d2a1d
F20110320_AACEHS west_a_Page_076.tif
f61ab9f65e9a394f54401550f01389c7
7776ae020b339f7cf46d2acc516b5c0d2aa084cc
F20110320_AACEIH west_a_Page_097.tif
dd6c07eb83cd8479864c49a74af612eb
69f75ef6a1c71a8fbc6ca1cae32f4cac243269e4
F20110320_AACEHT west_a_Page_077.tif
f959756eb33573095a399e0441970cb7
5f2666d2824623775fbdf8c3ad579aecf1d482ed
F20110320_AACEII west_a_Page_098.tif
ee339096ae6e62fa52833b6d7afb0d61
1702d4cbdcefc5548123961b0deea18bcd20771d
F20110320_AACEHU west_a_Page_078.tif
ce7faf2e6a406756a29a138e53447095
c4809d783c3f118cb4f625665de9d4d95b705211
F20110320_AACEIJ west_a_Page_101.tif
89a819b71eada26b1c89cfdcda52114a
5d4c343b3454c3740ffb6b80859289124b798cb1
F20110320_AACEHV west_a_Page_080.tif
ed7755d4c2a33bda037b2d2f16bd51aa
584a6516d07e9f48a91e01937d4aed4fddc39e13
F20110320_AACEIK west_a_Page_102.tif
dcfe598f55dd4998afa87f947641f6e7
7056a0503fd5c95a6b37c15c302c97f25962e2ac
F20110320_AACEHW west_a_Page_081.tif
c4af1009061e313f600a65518c390f35
b595954d11a961866d6bda71955e222e0898e8d9
F20110320_AACEIL west_a_Page_103.tif
8c711ea34e6feab6638ca218b4ebf106
5abc8f6d8dd533bec9d6c8bd91bdbd95418a2d94
F20110320_AACEHX west_a_Page_083.tif
355fa88a0d248082b0abc8a98b875269
58e9673de0a88d3ffb95ba79d1967abecb7ded9f
3045 F20110320_AACEJA west_a_Page_009.txt
d28a8478727e6614ff8bb6d04d304ab4
9ed65ee4d49d9b31e6e0ab51c163bae231bb0780
F20110320_AACEHY west_a_Page_084.tif
18685013c147d2bac72f878b78b5f9f0
a743ebd94e87ce9c35f07678495ac428ff53b64f
1570 F20110320_AACEJB west_a_Page_011.txt
a4de177326fab171a0c96c8fa824ee06
75429efbf91c97f2b80798e2155dd7f9386934fc
F20110320_AACEIM west_a_Page_105.tif
ada96586afc375f8834c8092831af4b5
601392b2fccff2da10b444c27c321e4136326d71
F20110320_AACEHZ west_a_Page_085.tif
d00c52319abb33aef04aa57aa6dd71c9
c5e74dfd15605ce5cc8d27f1404120139e726b71
444 F20110320_AACEJC west_a_Page_012.txt
fc8e62e315ccf41de8b237689d6273ee
e11d683aba0fb8b74e9f5500a5efb597e4c3859d
F20110320_AACEIN west_a_Page_106.tif
a0dc74d72c9ff3132fbacc50b4d0f483
4ab8edbb40fdf818d8c08869b22174f6c49ee043
1873 F20110320_AACEJD west_a_Page_013.txt
7237ed1fcba00666d56b8975b6bd2b8a
dc371ca5f76908bfa992deb0f4c75ba911a21118
F20110320_AACEIO west_a_Page_108.tif
a63b2593bd387ad8137d6f8e1c917897
2d1bcf0de11c08206a3de152ab17efa687525774
1947 F20110320_AACEJE west_a_Page_014.txt
54d2f2d82e527c49cc2f7282079c7384
25bde61b599d934775f10a32b6b12d4a1d83ee69
F20110320_AACEIP west_a_Page_109.tif
1c90191789285532e3593fb9fdf0f107
fb8245eadef6d0e31a7967656724af28b4a948d6
1881 F20110320_AACEJF west_a_Page_015.txt
a994324dadd2538e5fadcfd51f9ac8bc
fdea924d0c58ebd42c38713fe1c740781f32bdd6
F20110320_AACEIQ west_a_Page_112.tif
7ccb05592bcff8fea90776615c64aed0
941bf55803b480a491e310ac97d3801277699dba
1719 F20110320_AACEJG west_a_Page_017.txt
fb84fb372b27528a6ccb2b80c6c2fb23
ac3d7cc76fca9c9104ce3e03870b39fc1881b988
F20110320_AACEIR west_a_Page_113.tif
32196cf81b944afc935c2939fcbb58c4
17acb28324f5df9b5b0c54be96627421967a19bf
1916 F20110320_AACEJH west_a_Page_018.txt
849a9bcde82635dc9e9ee97a9598cf18
41a4574f6f87bed2192836b7c47228232a493f76
F20110320_AACEIS west_a_Page_114.tif
3b60c75c6c1f5c197ec35d14a18a9978
57c7cecbb3397139d6084d7f038f4dd5adc987b7
1682 F20110320_AACEJI west_a_Page_019.txt
3240046e340d010369333531c0442702
8f63f44589afd6a9b17480c5ca08c4e538d1b244
F20110320_AACEIT west_a_Page_115.tif
6a166c961d71ba0788c8cade615750b6
90a4d89846919fb5c937098175fe1a17d31d6833
1073 F20110320_AACEJJ west_a_Page_020.txt
04540051066975655bd19a7d817f2fc8
d8b115a9a84d7ace411fe498b66ce6853b41db0a
F20110320_AACEIU west_a_Page_116.tif
3811636ebfe61dfef8e40b4078ec50ab
94009a4f81854e9ad8c45bf1a8f8d217ecaa2b2f
1628 F20110320_AACEJK west_a_Page_022.txt
a89b0088fdc936585e58b5931065bf05
72a3e520505cdb2389affd69f4e1d4a9038f7e6e
F20110320_AACEIV west_a_Page_117.tif
33c329c7d097bc45c42030e791348d28
226cff0956c65c4d417137134b50c668da171a12
1995 F20110320_AACEJL west_a_Page_023.txt
b75761a95ab631c3463895b807c83c9d
7247702cdc1b24485be461453ddb2b009d385213
463 F20110320_AACEIW west_a_Page_001.txt
9c0046e10a5d906702319ff4e4b1ed01
f5d975a730caf413ca23938846559aa1cd6dca2f
1921 F20110320_AACEKA west_a_Page_049.txt
b2fd16ec3144af2674b89352da3696b1
440b38a7793745721c344809c062677c4f48d87e
1844 F20110320_AACEJM west_a_Page_024.txt
efe6e6ec8b2c23116e47e664c3e9e713
e972867bf96bea944fd4daa1e16d56a5bfb26f34
786 F20110320_AACEIX west_a_Page_004.txt
d0c9817b55812a37d602709e604a3572
17bf1c53859e645f1f0f19ac9c1f7d59340d2b70
500 F20110320_AACEKB west_a_Page_051.txt
2e8ad148af66d003feea31b47d98d193
a86e42942ae885b9f509940be08d476702970aaa
F20110320_AACEIY west_a_Page_007.txt
16071c86ca881afbd9a00382d81a645c
485797d9c4371f728a219deed9711a4a058d0f9a
F20110320_AACEKC west_a_Page_054.txt
845ede63c77507d2dd450bae3acebbea
be13c70a1d2ddd158693d4db812c4329d8462095
1942 F20110320_AACEJN west_a_Page_025.txt
4ff2841670a438bc49525b2bf84e6133
f5794dffe7b2692766214c7fee4618ed2f0685b8
2422 F20110320_AACEIZ west_a_Page_008.txt
ef3139b4a185513cd765fa7d475fa830
f66e84e5f2d28dda7de8aeccb97ce4f3ad9bddaf
287 F20110320_AACEKD west_a_Page_055.txt
4556a7796b6e094db67cc3979e09afe6
08af65cdf0e26798408c78e0fb4f24fab6e78622
1788 F20110320_AACEJO west_a_Page_026.txt
003de9c5e6bb47cc28f9312ffbc3e60b
90290e8aa35183efdb2e4a4a04dc75be9764b3d5
1423 F20110320_AACEJP west_a_Page_029.txt
7dac34c64f3bb4395512243b7733c7fa
72fd866dc843e6b831fefe0660323b512796d28e
609 F20110320_AACEKE west_a_Page_058.txt
6873321fc47ff6f505348c2029ceafb7
2ce4bb97b06699cff92430dce73ed4db137f5a0d
1080 F20110320_AACEJQ west_a_Page_033.txt
bb8e74ceaba1a62a01a0d11d0e0a828e
399ae021fee746c73ebc438ea35743965fb0767e
654 F20110320_AACEKF west_a_Page_059.txt
030200e1456502ff2e398973d9b050b3
f86b6f607714b293d183b4dc14eb3554f19020bc
409 F20110320_AACEJR west_a_Page_034.txt
699c6a60b20cca014fdabe5c78980110
075f7c3fdb4f67d3a3ad1123317dbea607c69413
487 F20110320_AACEKG west_a_Page_060.txt
0e5ae7cb85955dd4debcb18463beb8ce
ed2d36e4cf33a0c48bf033101b85dc2471cf62ea
1835 F20110320_AACEJS west_a_Page_036.txt
866d7b3e2d69b7106075f7b1dc5df5a6
b6ee4f5c68310682ff884bd68d7454a85e731828
420 F20110320_AACEKH west_a_Page_062.txt
789bb371ba23c640763a6aee066b8e72
c3215c8f0785e692c2a19f1d4315d187bfd29719
1834 F20110320_AACEJT west_a_Page_038.txt
598884f4ad148b39a46a3def358cadaf
3bcf715663a83c2451da4bdb02d94a4b8cc07284
458 F20110320_AACEKI west_a_Page_063.txt
ecb95911c95744154997d77e1db2d1a9
d4a9757e07c06a59770f7046bfd37615e3d3ab55
1815 F20110320_AACEJU west_a_Page_039.txt
b6e4d1b6c2b4f81dd21c4934301da830
568d1876c74296b9302965cfbc922b25afdcf796
1623 F20110320_AACEKJ west_a_Page_066.txt
75d9a4efbee85a6f494b71fd59220956
bd71c8edc26739006cfd6ac290794689a9472642
1636 F20110320_AACEJV west_a_Page_040.txt
e042937a0b913556aec14a910eb77cce
9e1ceec926c0a231356bb8bf3920167ff855a13e
1887 F20110320_AACEKK west_a_Page_067.txt
d1ff0dbf762c4756a51cb39d0ca8baf6
5f735d22aa0efa421a61fbf72db93973a4b7a0b6
F20110320_AACEJW west_a_Page_041.txt
7bc4f6d4289224b8e6b3a755c7031f6e
a82202128de138aa822b53b1a41887ba5c5ec1e6
1896 F20110320_AACEKL west_a_Page_070.txt
457d67963b5b7baaadb2aadbfdd11f8b
7308f07b9db0763ad1a46407bc5d09f9e708b94c
1641 F20110320_AACEJX west_a_Page_043.txt
1560786dd429ddf03f2bb14752389430
a8e1186775c3e2b6d406abe8ceadd119dd8cb94a
240 F20110320_AACELA west_a_Page_094.txt
7e54c455322a9dda5b522e410be0691b
829f9f2315a1267d3f8089b2b96c28283c24d364
1327 F20110320_AACEKM west_a_Page_074.txt
0dbf454e616b53e3b9c8355a3af71612
7b0f7a412ad6943d2e5c63d6de10c3bc7c3b1d8b
1907 F20110320_AACEJY west_a_Page_047.txt
f2a73d4f3d3c8fc4f71ec34e162e359e
a59e01ebd63cfc847dd0973c53a69b2b0ea74ebd
425 F20110320_AACELB west_a_Page_096.txt
bfdc413d76d17b62d60333a15d3c1595
7d93197607915777c1eec816ebb126d1ae9dc76c
1530 F20110320_AACEKN west_a_Page_075.txt
9be667af1e75e0c2c578b02d9fe807e2
266c48845cb12b9ddbbb5d561d4d770d4ce06f50
1870 F20110320_AACEJZ west_a_Page_048.txt
bc0e1b5f2b907cbc9bf00ab57d86ee8a
1c40e7ce34869ec5cd5ce9b07a4f7ae2b443e20f
408 F20110320_AACELC west_a_Page_097.txt
0b9b3e3e66155645621d1256afd42890
b5892bccdd905dceb1288281dbdf35d3baf7af7e
177 F20110320_AACELD west_a_Page_098.txt
8b1dbbaf7c105a6ae1252148537bfd43
237d4225cec114298219e5f6246cc5ad6ed654fa
1626 F20110320_AACEKO west_a_Page_076.txt
e9ca7e218d7eb662b274a4b4e72abe10
3db702b260e689b1f2042d41b10bbcb233939d30
327 F20110320_AACELE west_a_Page_099.txt
5dae7203df3f10b803f1c193de073d78
9b9699c6a8f84a4eb319f7a371f05d6ce65fa871
1342 F20110320_AACEKP west_a_Page_077.txt
ad74e77dfe6bfaf662e60ff4ee8db109
e03513ef762e3cd777f1d70940ca46d8106f90ce
270 F20110320_AACELF west_a_Page_101.txt
13f095690b9359e444745f45c18904a4
a577d68d407ad7ac8055727a5449bc5e53f1a842
303 F20110320_AACEKQ west_a_Page_078.txt
606e3d757e7893dd7eb5c123e7690596
afa955f2c267dca24705e4c06a1a9d0ebea87742
F20110320_AACELG west_a_Page_103.txt
a60c2760af4538c5afa96dba74e2481e
3d0ccc130642ba799563522dbae1e61580153a5b
577 F20110320_AACEKR west_a_Page_079.txt
837837a6879a4ab7b8f38d596fe4b7d0
4cc6043a8e780f927d2e9d108cb6578c939f32f8
266 F20110320_AACELH west_a_Page_104.txt
ebad61f5eacac0049f93f50891c65ebb
a850d1a836b8b249973bff8bae84d6d28830e93a
383 F20110320_AACEKS west_a_Page_080.txt
97fc19e010b83f8c977be7a815df2b94
8f398102ace7b140ae27884a9e4d663885d7c399
302 F20110320_AACELI west_a_Page_106.txt
4841b7b594adefd917adcb6db3af477b
dd1e07948f3de92bc2dd369519150903da49ecb9
1961 F20110320_AACEKT west_a_Page_085.txt
78a74e25a4f2eab2fbd39b0faea2eefb
fbdfcd6238ff3c028873bb93921b36583d1254b8
F20110320_AACELJ west_a_Page_107.txt
bf882e9684d8e2556dfbd237cdeae897
79ece83c0e315d7b20ebe83491a41a375e260659
1823 F20110320_AACEKU west_a_Page_088.txt
b687e838436e8a4a9d549298ad81f178
8a117b0e722105ba2c04d73282d5bcf41df4874c
378 F20110320_AACELK west_a_Page_108.txt
d22a531839630b3807511f5019247236
71ec6c659fbbf5b6726e93c5a2890540d2fe18d7
F20110320_AACEKV west_a_Page_089.txt
527dbb154a2ba87e53f92e3f4176da65
50ef7b9e7cde1ac64f1bd8ee55cc2dc6615b248c
F20110320_AACELL west_a_Page_109.txt
ad95786b65a25381b836b9fe8f4f050c
1bbb259a7e7dcd02fd5156a3fca3f8a319edffc5
2003 F20110320_AACEKW west_a_Page_090.txt
74c1a645d0af538e96d584bea4a978ae
72e8d44d5fa3cfc4de25b940c2b2c448bd308536
47604 F20110320_AACEMA west_a_Page_016.pro
843d104b52ac0394f8a2d8c547d3b5ff
d3d8dfe6709cc2c72d369b0deaa6511e5800f51c
2005 F20110320_AACELM west_a_Page_110.txt
aecdcba7c08ecd64469a0e22c59b20d2
9f2f0a94dd877c6b7971035fdc8e8df3687eb9f9
2009 F20110320_AACEKX west_a_Page_091.txt
c95cbccef1caee277d3329b48073ed99
45b685b5308bba7600ed64f2232658714c85b0b0
47154 F20110320_AACEMB west_a_Page_018.pro
d5d44134adf9c8918a42f56bf5cb2e63
9c133a3b1545efb08136eb39ddb9d75b5a3b7a7c
2062 F20110320_AACELN west_a_Page_111.txt
758d9c69e569316e910a3a7ba3dc2b7f
ed89a5acf2cf96aee1e2fc8b02233e7836ac6323
1871 F20110320_AACEKY west_a_Page_092.txt
620d4a68a297c27edc324149c1c537d7
5490568bad9bc27fa908367774d98813659bc375
41242 F20110320_AACEMC west_a_Page_019.pro
a80f72c52fb12a172f76d5b7ee06f48c
a09ead0d8ab7e0d578478859e75e97d112ec31dd
1948 F20110320_AACELO west_a_Page_113.txt
a2ec45f3add6a1707af177165749c1ab
4e04fc94497e5a7a7c2a922d394dee35d486ad12
382 F20110320_AACEKZ west_a_Page_093.txt
897ec38a57e52efecba4b724e4ae57ac
79f09bc495b4367598ebedac2dc1cf90cf460c4a
38190 F20110320_AACEMD west_a_Page_022.pro
41df4cfedfd9832b317909a9ffbd1e77
3ba80a0d0485cedb4f6bc256ffd39908d95a6668
50525 F20110320_AACEME west_a_Page_023.pro
bc7bbc81461b25674ddcd69f67244f70
8f637efca689a1a47b223d5c4beefbf82cbb02b9
2195 F20110320_AACELP west_a_Page_115.txt
880bcc33caffbbd42cb960c77c0773de
e3e7c15b787aa4408a1fb96f6534f451646a3dd8
49167 F20110320_AACEMF west_a_Page_025.pro
520b8c513d9f114e235aaf7875b21b06
8b69ddc83401946218c5bc0a03db2ea549b89cab
1780 F20110320_AACELQ west_a_Page_116.txt
e6bf9ec9a90e54718a004edcc39c8b24
e5a791c0660a61e8c60d5f487fcc766121144956
44866 F20110320_AACEMG west_a_Page_026.pro
3240e67553fd5d310ec48d4396cefcc6
3b0bf76c4aca256ad92ec2643ba97c3510ed25ec
931 F20110320_AACELR west_a_Page_117.txt
78164b65a3b1cb1185c910c0f80e32c3
c60c12b906153193c7b473524fa2d6c126feceef
44939 F20110320_AACEMH west_a_Page_028.pro
5aacfcc662cd97ccf7afa968d24adc55
038e9116a30cfff00794a017bef094a7a0e39c50
8106 F20110320_AACELS west_a_Page_001.pro
e5301cd7b4a7f51a90e42ece1a84bc73
b9993caa75192a9fa45a1f1655ea0a2eaf69ad35
40356 F20110320_AACEMI west_a_Page_030.pro
46e6909c4a072cb1d4a68c6f293e7635
1c6a1c354c1c41e8a81d71199010cd94b0197d3d
3394 F20110320_AACELT west_a_Page_002.pro
6697ff58a106f332f6167c53bee0a76f
af3d8b51f089e8f5a171fa6d6e7dcb94c23cfcb4
41239 F20110320_AACEMJ west_a_Page_031.pro
9a4a91380dbc0fb19ff19c21a7b9ab49
3645356c24f41f34bb019c6fceaca3b64c7a359f
42482 F20110320_AACELU west_a_Page_003.pro
c176cae33bad741fc88f7140c22d900d
631dbeea95ea2c8430fe13453be853972422daae
44768 F20110320_AACEMK west_a_Page_032.pro
7f395d3d58e9979a05cb0fc93507f2a9
21aac977205f2161ed3e6d4939fefd1d0b143ce5
19323 F20110320_AACELV west_a_Page_004.pro
95111c51e858a2c615e618755647f6b7
018ecd8ab0a98fa6628e846d64ffd6c6f6ecde55
26986 F20110320_AACEML west_a_Page_033.pro
1f01ec25ec340fd2e56259fcaa29136a
2a1ecd29564ca17946a430733b3a0c68720a522c
77274 F20110320_AACELW west_a_Page_005.pro
68916178137078918e65fe50a63c0370
96a39bbb31ad820a43399e5ce61459eaf96a31b9
8286 F20110320_AACEMM west_a_Page_034.pro
198cd8f7f09d3fd3e41d7d6d846c4da7
00db81ae110daea4d2ade8c6bee81ea6acba30c0
99752 F20110320_AACELX west_a_Page_006.pro
e1413cba266c2e7256a237ae8eedf5ee
6a4bbc4556f4c827a275f4810b9e58f7f09caf17
4883 F20110320_AACENA west_a_Page_055.pro
74c30b910f48cd8a7cb01710bbe49fba
c614b61e304015e4333905c2cdd1e978d79ff568
45064 F20110320_AACEMN west_a_Page_039.pro
2acb94459284ea8b8a4e2dd0d25a8259
0de6bd255fa9e3a797d6f05ddeb2d34a0a2a9a59
19641 F20110320_AACELY west_a_Page_007.pro
2cc8405ba6439d2f99aac4096d3a6d60
2ac2dd9d806b71de82f509579e9b37acc6ddbff8
22956 F20110320_AACENB west_a_Page_057.pro
a575bf829200297235654b08dfcf979f
b4fe6915e73a55c9d3dcc85330d0add255ea5937
43832 F20110320_AACEMO west_a_Page_041.pro
4f9f4cb7150be64ea15d483a3083a8e2
22027fb6cd4012e1f6c52931486d8387c6f40aeb
60221 F20110320_AACELZ west_a_Page_008.pro
8694af322c208a8558efbdc47fe73923
6b6e757da569f4c93c376faf3653dbe95105733d
10516 F20110320_AACENC west_a_Page_058.pro
9351dee1686e2a34052875ec53da2478
d0f59b14422447d845cd4efb9348bc042b228913
39389 F20110320_AACEMP west_a_Page_043.pro
19b0fb47d43c9c9a156a6e36c6a735d1
17097f0ac2f179209ef719771bb5df1cd63c4aa0
14288 F20110320_AACEND west_a_Page_059.pro
a60a06bc17e7a9057c2cb7359b5e97ef
e36ca6df834e5f6b391901486e1604318c89b899
9457 F20110320_AACENE west_a_Page_060.pro
78cda7aa61be0e22b65a087d8e536870
c07459eda847639a67d30732c031249ff5e84568
50136 F20110320_AACEMQ west_a_Page_044.pro
6a4f8f3370be3600c945a6ae816abeeb
3b5926fbdb412369ca532909fc8d1ff0ac03de94
6267 F20110320_AACENF west_a_Page_061.pro
41707e5e9f2b10de5e7b00b3911cee2f
82d9eaacb9fee660e3bef94e9e27acf99d5631ac
46415 F20110320_AACEMR west_a_Page_045.pro
7a5d18a4d4e08fa455a886f47cd9d452
78a38249af7763ec77194da89d3145ef87f22951
8012 F20110320_AACENG west_a_Page_062.pro
ba716c39ecfa24c33e13d71b0452d4dd
8daee0f887371b9c2438709813567616e6d601f8
47622 F20110320_AACEMS west_a_Page_046.pro
1b31db506e4b329f5ded9c358f14fafa
4f144dec8071c9f61af1ec1fab4df41fce57f614
9989 F20110320_AACENH west_a_Page_064.pro
8b4b5b8a6934e4d408c8a64dfa6661c0
e76f69b2d8cbf00ed721a24e0e56ea949eee94b9
47557 F20110320_AACEMT west_a_Page_047.pro
10b51a208cb7ce5f0a2f9715ae600bc8
b9972e976c3a72b2dd5df70ce49b09177903a24d
45642 F20110320_AACENI west_a_Page_068.pro
551c28308753ab9102756d35c73483e3
f70b5aca66302802fd46601e4b4e0f9a2a80a375
47217 F20110320_AACEMU west_a_Page_048.pro
09469bb8b69e841ec7e97905fefe26f8
37ab95b889d60d502d2c23a9445d69dabc83b3e7
49741 F20110320_AACENJ west_a_Page_072.pro
1584deb9a4a0f01aadb3146cf9179982
876285e0f7687231ec28cc7cf47969af2d3295d7
47969 F20110320_AACEMV west_a_Page_049.pro
857858fad00917d32cc7b69971a7b1c1
3bad74b72304a7aea4bf1b58549bf7814b89aaeb
34704 F20110320_AACENK west_a_Page_073.pro
baeeae84a486ebe31d5180a8027028cb
fa76867fdd87f322b4d9250a1fecfd528e129fd3
23914 F20110320_AACEMW west_a_Page_050.pro
bed9c0ae1279bb72537e862572e8e37a
ad5be681daada63ba71a9b98aab6e32e32011c42
22747 F20110320_AACENL west_a_Page_075.pro
db2630bb10752dea5cafb36dc5287640
5f72cb6f215549f0200bd015867c0336a320bd5d
8578 F20110320_AACEMX west_a_Page_051.pro
c67dad025203de424abdc8fbb5e77dbe
f5c871a16c527e336c505b3ac95a70eaafb1be81
5463 F20110320_AACEOA west_a_Page_102.pro
c9d492e5064849cf522a1efe083af2c7
83991900b52d7d9d6e08dc88b949ad19a0cc2277
24591 F20110320_AACENM west_a_Page_076.pro
0a602140d1b9ecb00204207e14ff7183
3c4bb2897245aaf940ce28177a7007f0f82c9836
3298 F20110320_AACEMY west_a_Page_052.pro
f8084a62304e3ddebabd24fdfa2a1105
c114e5b36e7216015871c62bb68d16e41fcab743
5853 F20110320_AACEOB west_a_Page_104.pro
3785a3e95f9e0d15a5a4eaae76923229
ec30aa5d477f37be80fbd4b9f05e8b0fa671164f
24917 F20110320_AACENN west_a_Page_077.pro
ab3e3861d632ee579c7714622657ae2d
2a15879f8c17a74d5ca3945367a4ce64eaadf8a9
4338 F20110320_AACEMZ west_a_Page_053.pro
41f47eb9cb3107fd4a04d047b5ea548a
6979f6c521881c20c1f537fa34ca437bcfbf11dd
4174 F20110320_AACEOC west_a_Page_105.pro
88a699c0c0feb5a2fe7c8f1374cbecb7
bc0689e1a42dad94b961a0d3ed480441e1481bf3
6035 F20110320_AACENO west_a_Page_078.pro
562fd1788208b0a74bb2957937e3847a
15a53448ec73c23da65da9fa87445e890b30a7f3
4839 F20110320_AACEOD west_a_Page_107.pro
4df8e46067d17c414817c6d944296054
3a767239de42f0d7911dcb88caeaaf631c25742d
15397 F20110320_AACENP west_a_Page_082.pro
7512d89ecff7cdec7b7dc68be8cc78f4
1b5efae59e0f769dde60f01ca224e7037ee88bd2
5988 F20110320_AACEOE west_a_Page_108.pro
5a9bdea12ee999af2664063e83251543
8774d04eba0b35215274e74a5693dec6210d8b28
46927 F20110320_AACENQ west_a_Page_087.pro
986c6e51d97e1c4b11ec9d9cf5a9b2c9
ea496fe9a5df402a1d930f2953e132f1b6654f19
46842 F20110320_AACEOF west_a_Page_109.pro
523d532d2f9ca54e135a6d80e2fd19d7
c3eb3cd473b9d53425c08ad0fa25425161843e1e
48682 F20110320_AACEOG west_a_Page_112.pro
978e8158fc2339de7166ab40030daa37
ec1a4191c76aae733949e5e6dfbe542a8f9fde77
44195 F20110320_AACENR west_a_Page_088.pro
c32bec68facf4d919c1ab209c5b6774b
8a8eef186cbc9d9c9e64e56577d32e68c3fc17cb
55440 F20110320_AACEOH west_a_Page_114.pro
172b0542f1ce40ffac598d3035fec28a
0f53775fdc21480e9182db1cf66e79f64f95e629
46112 F20110320_AACENS west_a_Page_089.pro
997295fbb154f764eb44456daf999d11
ef522cdbddb34b9675eb9355db629b3e04495cc0
43379 F20110320_AACEOI west_a_Page_116.pro
7dedb22619e6f538c0f095872bcece87
cf27338882fc41c5f9ecf31a20b73fc271971257
50977 F20110320_AACENT west_a_Page_091.pro
ff5e27d57af236bf7827541fd5bbc4ca
6d1e56d8e7ac0f923ea51a0c10466fd63be09637
22276 F20110320_AACEOJ west_a_Page_117.pro
9e25485626e2521d67b831a7e1824a42
1292d8e12939af17bb395dd5896451419ecd79a4
46287 F20110320_AACENU west_a_Page_092.pro
a00f817870e353d4d2291647ec1b38c9
ce3734c0f185e20f6f5303db21d3f5d11abb7103
7145 F20110320_AACEOK west_a_Page_001.QC.jpg
000458fa4feb06b0da505093bea2223c
111974d6afcdac2cb8cc0996190603e91cc7f363
8444 F20110320_AACENV west_a_Page_093.pro
20ae51e1c6aec95b823fd0be1b3fc4d8
00d594c5b0f93554465e6275d9ba2359bb8a68c6
20799 F20110320_AACEOL west_a_Page_003.QC.jpg
2b430156952fcbb10e061213556118a7
2354a245af9889ad22d403687e07c4ad75e74a51
4052 F20110320_AACENW west_a_Page_094.pro
c21d369da24a5c79f5776d7d1cf4ea3c
700457727bb0c6568df117235dee1460fda8fa4c
64371 F20110320_AACEPA west_a_Page_013.jpg
62ee9704ec9851523596db1f53d1255d
02f09060b15936e5ba93b18e02fd0c715750df42
72750 F20110320_AACEOM west_a_Page_005.jpg
d12b1de608f434ee086410de06dc0712
b90facbd64f0732c0d69f12e475d968e7f5d9200
7074 F20110320_AACENX west_a_Page_095.pro
00852f0c5bc42985cf66de3fae3d5abe
98b5e6963d8b7be33158dc3f1f1aa51b6d13243e
18168 F20110320_AACEON west_a_Page_005.QC.jpg
dcf1df24faca9be99dfecac193721fb7
2a095f1bca11276a50567c49806a015a9464ce8e
6148 F20110320_AACENY west_a_Page_099.pro
72cce08f9978ad842bafa7b5a0f58cb8
6b7b9af1e41b12248d129db002a2f2fd03b31a81
20577 F20110320_AACEPB west_a_Page_013.QC.jpg
65a7e46fd366fefd9e5a890e23f4c863
38309ba01d34665974b3f3df27d8fe3649281123
84464 F20110320_AACEOO west_a_Page_006.jpg
6ccc633e3c1d926d6c8d65ca5f651786
cdcc69c71d4c92398fba33ebdd01eb5e156f8e02
5229 F20110320_AACENZ west_a_Page_101.pro
dd4bcd1853c7b2c2d4847a1bb046ff7d
aaa2cb04febcf1d30117dcaa234b352ddb568792
69872 F20110320_AACEPC west_a_Page_014.jpg
daa8d03e13fa924334ea06a7ae0ef9ee
df406b3886be1c0a47c99e73a16f60ec63c89460
22005 F20110320_AACEOP west_a_Page_006.QC.jpg
072f70bcf4c78ccc33409fef14064f62
91ae5093f481ff4f854cde9aef67adc663f4c688
22503 F20110320_AACEPD west_a_Page_014.QC.jpg
323f36e67a81f6adacd8f0523436cf19
a8cc8b04f0169cf393952aa2971980945df4776f
6961 F20110320_AACEOQ west_a_Page_007.QC.jpg
f04b1f44fd082a0d5eab45c2ae4da8b7
9e6e6c1ff4928afa5ebaee1e5be3f9a7b4b0a14b
68578 F20110320_AACEPE west_a_Page_015.jpg
33e0bd077f993667a8849901274bc1ef
bedb07c05194c6539ce134ff4efbccd343bde6b3
80461 F20110320_AACEOR west_a_Page_008.jpg
a50af6a4f75d9f81b759ec38d6cd2087
8ed3c6ab074a066d1b1d38fd56a4d2a72de4af0c
22310 F20110320_AACEPF west_a_Page_015.QC.jpg
d841d6717e80d54100b8120cef5d76b4
793824d4fe37b75b3fe023d4c0f605ee4807a281
22446 F20110320_AACEPG west_a_Page_016.QC.jpg
8a3344fbe10a236b0abcf193b8a8e7ff
c7b41f1463b635e4590d978b917cacb77d2fb9f3
22606 F20110320_AACEOS west_a_Page_008.QC.jpg
ae08dafab40508396e84b5360233ee91
53d7caf09cd103dc16728d7ee17b31f15c1f7863
64250 F20110320_AACEPH west_a_Page_017.jpg
0351d3ee73f83f768351dc6240efb8d4
372af509b11aa83edd8ecb047f1cb8894d13c2fb
26553 F20110320_AACEOT west_a_Page_009.QC.jpg
4a83e8611d6edbb768b1653c1dad69eb
5e010315feac2768c487547fdcde94db8ab91a1f
21339 F20110320_AACEPI west_a_Page_017.QC.jpg
b7a309f43cf95415f1a32d5e22e987e4
7ab10bbb19f98e3857f58ef7ab9c8f8087e833ce
20435 F20110320_AACEOU west_a_Page_010.jpg
14b38638541d27c77a0b22c869d92cae
da3582df93ad1e3ea7f02a878c6f54d98dea6502
75206 F20110320_AACEPJ west_a_Page_018.jpg
2dbea305b10e6036733919bc8667c23e
29c8d35d6b196eafb7f20de66241aecbcb5362d3
6047 F20110320_AACEOV west_a_Page_010.QC.jpg
81c2195f3dd825ac813ad88241f2ea28
b2b4bb6197c794ceac21e429a509c50bd9305112
11397 F20110320_AACEPK west_a_Page_020.QC.jpg
60375bf44bdd51db49aa4b67d330f205
ca6ae89201e5cd1aedbe5232b191a56d42dbc930
54868 F20110320_AACEOW west_a_Page_011.jpg
2046c97a9ecf31ce0049d07081233c12
2e651b02f2c945c6f453dc28cfe0785fa8e60989
26307 F20110320_AACEPL west_a_Page_021.jpg
9e0dd42d10924e1c2769fe13c8459b52
7a883f7c4c9364912e2f475570b5c28b83218eec
16889 F20110320_AACEOX west_a_Page_011.QC.jpg
9944e4065601a8c477ab242f6cf3d3b0
d3fa14e90f6066ffb7117216b325594898802dc7
21985 F20110320_AACEQA west_a_Page_035.QC.jpg
730f9d73c896a03a0fe612e10c68c52e
999dbf9e5b070712a1d8c523ad95dfe83bc728cf
8144 F20110320_AACEPM west_a_Page_021.QC.jpg
1d46f1633873817740a1efb40e938008
f1716ba1f6cf61d7236b96a2753f481d037661ad
22647 F20110320_AACEOY west_a_Page_012.jpg
b5512dafcb8bd7c43852c15bb15a34d9
65b6fa8d49d79cc6559c9fc648bbff30c7ef2fb4
73848 F20110320_AACEQB west_a_Page_036.jpg
526483dd243012522e3bebd36bff4579
e03ebcdaed3cbd4731d1d90acc6f116bd1d95249
18972 F20110320_AACEPN west_a_Page_022.QC.jpg
0716b63fa08dcd1454f164589ff10f4b
62a192a0400946e901d2649ee937de6146cb9cc2
7365 F20110320_AACEOZ west_a_Page_012.QC.jpg
1993ecf23f068c5c00a17d39d8700e8c
6217b528ad4eab010a4561e981f216445834d98b
23499 F20110320_AACEQC west_a_Page_036.QC.jpg
092df9028d4d20aaafee04fa1d6c8f57
6d5bbef8b5a97573255501f6800e105e74cdeeb7
73647 F20110320_AACEPO west_a_Page_023.jpg
d0ca693d786af127de76ec99e3b49c48
7db798d1a8fdbd1899aadbd32c6a0c346a859149
23555 F20110320_AACEQD west_a_Page_037.QC.jpg
082998e7f68bf8c121b3e5a3ae89393c
ebc50cf5ebe8684be8cbede03fc3c6d59ede679b
70999 F20110320_AACEPP west_a_Page_025.jpg
b2dafcdf40eaec25208dfb9ab0d5f355
f90bbcfedcf387c1e3ea1969310edb180c2d56b7
74797 F20110320_AACEQE west_a_Page_038.jpg
f968a3f97d2d0524712a1ef3019b8016
79d525b1c447b231026796e99f777c84b356555f
64541 F20110320_AACEPQ west_a_Page_028.jpg
509be5e52f2eb588a5b707475592c6ec
4f8405798ac7d88136aa0026b7ffb13ffc96c53f
23364 F20110320_AACEQF west_a_Page_038.QC.jpg
3333baa2ce67d76ab80f0598fc5bcf3d
fbfc690ae5faa83b9d6b1bffdaa9b02f950161b2
21560 F20110320_AACEPR west_a_Page_028.QC.jpg
c73111f2be7e27d444d76dfbdf7ee8ba
0ca8a2f27fc5ddf0fe9d9bd48ea53ab88ae38307
65676 F20110320_AACEQG west_a_Page_040.jpg
0ac29cd9827f3f9799b18a3a1053a5e8
dfb0d4f5e086bcc4023caaf99e0e177f2cd2f504
59201 F20110320_AACEPS west_a_Page_029.jpg
c633c321d32348efe0aee3e6419cb445
6a05586cf5fd6d2249b2976237227e4613f58cb2
20208 F20110320_AACEQH west_a_Page_040.QC.jpg
b57294e781112058facc75d570c701a7
236e37efdefe78ac77629151bc0b2c1dca9578a7
20739 F20110320_AACEQI west_a_Page_041.QC.jpg
2b79c6e701e8bdf13d0c65d5968fc3f2
380f02fd872c46167e09d9651e7db2d12e58c972
19355 F20110320_AACEPT west_a_Page_029.QC.jpg
a05443eecc04223bf24fa714eb33463b
be2b271d0d36f7327ab10c2540005899d13cd868
71623 F20110320_AACEQJ west_a_Page_042.jpg
ecc93c16fd13dcd1f4d26b4394b0a6c7
7aada1dc5f923cbee8ed9bbc969e603090eae33b
61274 F20110320_AACEPU west_a_Page_031.jpg
e4e8b0ba5fa1e08df59a032f3e2720b8
8e3a010a997e96fd55c2625a018d5115e3655e3e
23391 F20110320_AACEQK west_a_Page_042.QC.jpg
8d6522f51a32e7da9fc9e7e93fd1fd3d
de8fae90436761610c44deeec36e5bdbc207cf00
64003 F20110320_AACEPV west_a_Page_032.jpg
0fce4ca1d923f9e69b1aaf921a24ae08
17820661446153ad5bcebbd3ff4b42ef0255e3d6
64854 F20110320_AACEQL west_a_Page_043.jpg
3f91a6e7b00c7061e0ee1bf7ff658eed
98295b2d340135cb32d7e7b7fd4f95d41eccab89
21367 F20110320_AACEPW west_a_Page_032.QC.jpg
a0bf187502c9de845ac3060703695150
55362d620ca8a6d4627d5fa86b763091280c29cb
29719 F20110320_AACERA west_a_Page_059.jpg
72e45007c9267b7c1ec2a84bf332f7be
8b53a1ede8db2c7239b0ce68b19df66f238933ec
20245 F20110320_AACEQM west_a_Page_043.QC.jpg
f8ce5ec9c42c2a805e7e8b8836a5dc7f
53f470e5510d0f8ed168c8f5a46ab549c90a3dde
14642 F20110320_AACEPX west_a_Page_033.QC.jpg
866b0e8ea6d31f6c988a4d8060765580
11fddb6dd93f936fc3136b21979f26142505054e
16785 F20110320_AACERB west_a_Page_061.jpg
24c488f347d1b94bb48650aff1917cdf
327be707b63196c93767fc7bd38606d3cf3545a3
80181 F20110320_AACEQN west_a_Page_044.jpg
8ade5a7ec2262a820ba72dc9993a2945
99de365b077ad67fecac68b33930d022cf70daf8
45019 F20110320_AACEPY west_a_Page_034.jpg
d9c5869d5bf6d8c77fd9b0a6466fdcd1
bf7051c0a749dbb7ba33f2c4bcda1de4ce1e4503
5626 F20110320_AACERC west_a_Page_061.QC.jpg
0049f24651d3c8135c576d17f2b8e8e7
74152b0f912a64255ca297d23724245609e7b43b
25499 F20110320_AACEQO west_a_Page_044.QC.jpg
448c71bd7a597c5f2b18bba19cb0abcf
bfca97c59e0efa9b48b95662026ed1a90c8ef1ff
13623 F20110320_AACEPZ west_a_Page_034.QC.jpg
8381afa43e617adba6dca92fa52ff956
9cf4e8c29f6bd093d3f52bd7e52a8cb3c98b4c35
25950 F20110320_AACERD west_a_Page_062.jpg
4960bd5f112944fb23a2286f7368cd9c
78072a045e135e2bbbc0d39e1fc1f52def5e5296
21734 F20110320_AACEQP west_a_Page_045.QC.jpg
e70325cb6cb309f8502585c0a8dc965e
aed012529ef8b32aeb59bacca46ce104aeee26f4
25753 F20110320_AACERE west_a_Page_064.jpg
4cf74ef9f7fd72327c497b76a02ce814
6bd6154057df65b69af30f14325f4df5111d6247
22540 F20110320_AACEQQ west_a_Page_048.QC.jpg
736ca5b0b40f93d56cd033ccf498e512
ef5f5780294d2d2da0548c2957b6a1b2d35d23d9
8074 F20110320_AACERF west_a_Page_064.QC.jpg
f1a4efd354ffd4351bdfa176f99cd206
d3255aa7ff877350f0e5e71b8dd4f07c67505df9
24277 F20110320_AACEQR west_a_Page_049.QC.jpg
8676a60ea2f392c1669dc72b197eeee2
cb0a99dd65972bb4a53caf2f7198ef8e5be1aa5f
23894 F20110320_AACERG west_a_Page_065.jpg
279aa2beffcc51330b60a7c32ea8196e
8b23c951d671e37e4c8b0f63e0b6a390809f95f9
7380 F20110320_AACEQS west_a_Page_051.QC.jpg
c295bbbe79de9829cc34f312a072decd
5877848ba96dd6b96cea0df5c1e9f016e9c35be7
59448 F20110320_AACERH west_a_Page_066.jpg
a0f568f97030f9e8fce98534a60a6750
74a1a420fbf0642f92223b15e0151d707a4764a5
19361 F20110320_AACEQT west_a_Page_053.jpg
be05fac34a31daf7feb70bc9d9a41a36
a80d3289b2421d3fb3162d6f4444393a0ea51401
70017 F20110320_AACERI west_a_Page_067.jpg
e3cfdf34ca378c6db5b082ffe387d027
998114fd6c66f4f5216937ccd6088489216c693b
66659 F20110320_AACERJ west_a_Page_068.jpg
c4dcde81c1d2d83c57e428cebc32a11d
39f906f1da8892d1eee4fc9bdccb35fa5e5c3412
16409 F20110320_AACEQU west_a_Page_054.jpg
97e6655f7e3121475ecada2134559a69
192bb7b5ed664fcf7045f1d58948120ab2b850b7
21905 F20110320_AACERK west_a_Page_068.QC.jpg
eefda55fe27ccd78fe824ebcc98180d1
b7198fd0fcf58e0fd48ff80df41a0301887f26e7
17078 F20110320_AACEQV west_a_Page_055.jpg
a837f2a1fa295315c052bf3cc10c1132
ce3befee5c12a8a3a46867f1cbc85d0e430e89b8
69694 F20110320_AACERL west_a_Page_070.jpg
46ef9d753d74666bc61dcd88dc7a24b6
7e99fd3a0361038d5ffc6bc02962da26bc61726b
5079 F20110320_AACEQW west_a_Page_055.QC.jpg
7b1f678eb1c975984345d7c4e676f72b
0f0321d2553b0bbcdefde88e7c0701300e2c0340
22281 F20110320_AACERM west_a_Page_070.QC.jpg
9521957dc845c3e63cabd3479f3333f7
2fa48cff4ff79ef2f2af312bc32a345e5f783a08
11792 F20110320_AACEQX west_a_Page_056.QC.jpg
d6260385f9cb17b0e18d54b9c3dee21b
b88df6a46ac39137c875ab4c3b97811db36dadfb
12320 F20110320_AACESA west_a_Page_079.QC.jpg
d741a4a6bfc33281d46fc05aef61e1eb
93968e747a5fbce231feb390d7297e6829fd43a0
68866 F20110320_AACERN west_a_Page_071.jpg
b1419ec451b2db6bb34b459e50ca1d73
73ba13d2c6ab39c9bbcf9280aa59a2d1fd5f9f84
41500 F20110320_AACEQY west_a_Page_057.jpg
2033cd45264521bf4ecc8e0e2a1222df
c7d1afab32ea96cd480ac525fc00ea715c2dfeef
36508 F20110320_AACESB west_a_Page_080.jpg
80c98e1f7009f02179e92b43de961772
c20fffb52451d433c8bae5c96071677524e3a54b
69638 F20110320_AACERO west_a_Page_072.jpg
f488b8c9b48bbf4fcb5d67ed53278fd3
8accd26ce0491f0e70981dde8aadd63db07a3a94
23883 F20110320_AACEQZ west_a_Page_058.jpg
5ae3a18af94d2750e5524473a364b733
ab30fecabb2626077e4571e52db4e9825831aa89
10908 F20110320_AACESC west_a_Page_080.QC.jpg
51509d42ebbb7b8707ba980124fd2289
283d8dd94ba71392ffbb8cac7e012ed3466ddff4
22883 F20110320_AACERP west_a_Page_072.QC.jpg
b8007fb0e329c095ed3b6117e7bdceaf
c1138f2045973434ae07c447b96c14ce058e67d1
42265 F20110320_AACESD west_a_Page_081.jpg
f8b9da5307c90c19685803a7e198252a
e0f148f7b912a10e5794b8ce561bd98c30106421
52936 F20110320_AACERQ west_a_Page_073.jpg
eb5c1367ba5feb7b6becf9b401eeb39b
53c01e32e062a907fbf52eb86f368ffbec26f75b
12496 F20110320_AACESE west_a_Page_081.QC.jpg
72446c8af29b9dafb4b00f29e418cc2b
70d9eb76dc3593e2e53a187d0645f150af350d50
17114 F20110320_AACERR west_a_Page_073.QC.jpg
8817a72bac8752f2646c492048a981c5
cb2bd0d4b5a04d7628ae8f49d232e93640a0a64d
7930 F20110320_AACESF west_a_Page_082.QC.jpg
b2f71849bb0d4da9432fe77398c5518d
7add57f290ec9bde3a5f90dc9ab4c3eebb27acaf
48186 F20110320_AACERS west_a_Page_074.jpg
399cba935459adc8bcfdbaf41145d049
59ae64f541e327cbdae77cb56651b2307a709ce8
7095 F20110320_AACESG west_a_Page_083.QC.jpg
41d3622902758b0f83456477e67f6005
832c307a566bed789c42a81b00e3a32ba73c6c5d
46149 F20110320_AACERT west_a_Page_075.jpg
c71c0a94213c397f866749a9350e7edd
83b951b3747d19cb728f852154cd06073b6f4d2a
61470 F20110320_AACESH west_a_Page_084.jpg
3b1eea1719429f6db5b74282abce3d12
e450f80947921272e867951b550b78d21a61c72d
45385 F20110320_AACERU west_a_Page_076.jpg
b05bba817801ba8274c3173dfc6e5ed0
10d2bac9a5ed22dd153937f171f9b82e143a5edc
19746 F20110320_AACESI west_a_Page_084.QC.jpg
eb44ff3e6703486d1f86e917308ab25f
8d72e4cf9816c9afe0f37e1714d5ba970a9e5569
20589 F20110320_AACESJ west_a_Page_086.QC.jpg
1e16fb0977e733a00863a02cefec0e85
0583ba0c4cced996b0363a7584b66d71e63e7d28
14070 F20110320_AACERV west_a_Page_076.QC.jpg
0229254aae699a36764ef457211a5374
3528db28979de02a3f32ce6bf49064944baf2e10
21382 F20110320_AACESK west_a_Page_088.QC.jpg
f45f5a9d7dcc66820d810a478a1de25b
dec23885e5dfc0d5bcc685c0b18849abd2c651fe
46120 F20110320_AACERW west_a_Page_077.jpg
b4347a6871825fe058efcfe2a700db89
7d3026adc27badd0a4df9d64548e9532e47de1ef
66892 F20110320_AACESL west_a_Page_089.jpg
bd7daa9e5c79fb5b7dc4fd697fb19c12
f7b18a03f5d36ddf40145af9f413b2f9444dfc0e
38228 F20110320_AACERX west_a_Page_078.jpg
58e3bd167bd2f66438eaf3825eae5ba3
d2c9afc9bdc4209323a83f7f43a3f08728323508
9500 F20110320_AACETA west_a_Page_103.QC.jpg
bcd2de7c1d549ed182f4573ad5c516f9
a62d89dee8dc914a28e9b04661101c3cc8598c05
67413 F20110320_AACESM west_a_Page_092.jpg
1d4982e24cc15f0b9b2faef6ba4f58c8
60012172ff231a18503624c43b7f1c59933df16d
12088 F20110320_AACERY west_a_Page_078.QC.jpg
3f06c6d40c1091c1b23e9e22b9016856
368ac9313aec82424c92121e6bc9d4594f0c243b
11530 F20110320_AACETB west_a_Page_104.QC.jpg
a66459c541c6e3e05ab34788a71bcf61
8721a5785f19ac5e65668f6f6639f0b0ac4e32ce
21645 F20110320_AACESN west_a_Page_092.QC.jpg
08caf33e12f8aa21d1c1cf6dde26f338
7bb0fb603ebe60f18e76141e81af18b1fd2fdc40
24229 F20110320_AACETC west_a_Page_106.jpg
8fc5dfa166fdb7c11c81da60959f7fd6
f5a9acf134c0d15db312d0e5e1814885233faa15
18710 F20110320_AACESO west_a_Page_093.jpg
47ab30b4c87138da05f66250d8e35f71
1caa7af345b0612693338e9da6d08f1ee3cc35c8
37674 F20110320_AACERZ west_a_Page_079.jpg
485127d486bcea77fa2c52b18af7a122
d2e64d1e5dfb5a82dbcef7eddfe708250c85dfc4
37343 F20110320_AACETD west_a_Page_107.jpg
ff335b4ad95635ae170ca8a6367c613c
05dd1c4e86b70bbb62ac9f2f27feb9a6134347bc
6131 F20110320_AACESP west_a_Page_093.QC.jpg
8ec38a4ae6a67921608f5195ba6a6715
41f1d59d6c3fd613d75a35ad2d38f50ce883ff05
11884 F20110320_AACETE west_a_Page_107.QC.jpg
62d0eed858f31fe4bf594a1f8d3bf92d
8a60456c70132a31ddea1650406fe4a0f948c017
23445 F20110320_AACESQ west_a_Page_095.jpg
93105f817a4c1d387c6424e1e2c93a2a
4b3345221800287430d7d9ed6ebe21ba057bd16f
49506 F20110320_AACETF west_a_Page_108.jpg
d07149b454256cfabe907f8c1300a142
9c7e9b59c47bc8398854c3987922d88cb6e49a29
7616 F20110320_AACESR west_a_Page_095.QC.jpg
bb0cd4d7d006857fd6b97e9b9da125f7
51ff108d25b1f58b2f4fe37121618638f17b4fb7
11903 F20110320_AACETG west_a_Page_108.QC.jpg
1753061903fad17c7c5ae9e0a80ae168
04f21d0a72ad9da2161830e6b3d6107c75f8577a
29539 F20110320_AACESS west_a_Page_096.jpg
03f04c65874478680f667f1cf4056ce1
e3d74aa69d663434021a8492e12a306771ccfb6a
22168 F20110320_AACETH west_a_Page_109.QC.jpg
30eb7c7c2d305fb91aa31972787b9856
104ef7e196c6a5b0e303f4ddc05863772ac7eb92
9343 F20110320_AACEST west_a_Page_096.QC.jpg
53ea91cffe4d6ddee3d2d7acf9430f67
da8819ab1a33d9464a46fad4c1ec9892927ccf51
23947 F20110320_AACETI west_a_Page_110.QC.jpg
54bdafc82ab80339aadb97d903410168
ec93b9631148e257d48d48bbb405f02a13df1075
7435 F20110320_AACESU west_a_Page_098.QC.jpg
14bce5ae858bf090545f7482a99137e0
46a250a4f0ba980720a5439c78a2010377acd376
72579 F20110320_AACETJ west_a_Page_111.jpg
597ad8f014a162a4dfe1321f67622e0a
bd064551d53f8508856867c764ece2a761069a6e
30882 F20110320_AACESV west_a_Page_099.jpg
c826bad6b0fe77b1927e4de6c13a914a
e5b81732f9aa1143043c1320d773ca371578665e
70602 F20110320_AACETK west_a_Page_112.jpg
cef70fda5a480e89f807a31f00f7b069
dd426c578a3c3e83231f2659c685faa27e3131e9
22815 F20110320_AACETL west_a_Page_112.QC.jpg
a09f056addb1ad841804258fc223b991
8b8d9f19bb474b19019708d0830b46a8e3507851
24310 F20110320_AACESW west_a_Page_100.jpg
61858b40e44bceafcf26d4be8d018bef
78295cce291dbef12a9eb08813c8be71e4ab96c3
67557 F20110320_AACETM west_a_Page_113.jpg
e924f6126fb879cfc2bac6acbf0881e6
3135b818f14563c61b4e6ffaed4a1b691059885e
9270 F20110320_AACESX west_a_Page_101.QC.jpg
46335f4ba2ee2205067c275239756bce
e87844b50354d0177f3c54bdc05f99630eee2202
102982 F20110320_AACEUA west_a_Page_016.jp2
f9dc238a13e53fd40741767b16db0963
dbddd8e25eb3137da4cae0bec657699725137dbd
21497 F20110320_AACETN west_a_Page_113.QC.jpg
d0a69eb0885397e0d405d3759841e353
1d4b5c8adc6adab77188ff86dcaca2adf37e2f92
35367 F20110320_AACESY west_a_Page_102.jpg
5886c2fa1f36d8654a7809e9e65d3629
c1a34192a2ac1e9b64fb6f9e4c46804aea8a70fe
95275 F20110320_AACEUB west_a_Page_017.jp2
e2d7a61d03a977631be47836582f7c07
27894ac202383217c4ae943a87730a677f7502ef
73534 F20110320_AACETO west_a_Page_115.jpg
c16d4b37f7ffe8a225cc4ba6fd74dd82
fc8f0bfc55309e388126b279e7a6b37507b515f3
10991 F20110320_AACESZ west_a_Page_102.QC.jpg
bd6432d2c72688dd441cb35e62e24058
2b604ff07b4345443787d17c20db0b351999a3f4
85997 F20110320_AACEUC west_a_Page_022.jp2
33192526e10baac0cfba70312fc571e6
7caa864dc9e6b29a6e5b9096f69dc6dbf8ed50ca
23946 F20110320_AACETP west_a_Page_115.QC.jpg
0a50caf739b58762c5e77ce18d348d28
f730f2d357f2cca4960517a56821180fff3623e8
110823 F20110320_AACEUD west_a_Page_023.jp2
37d0b465a943834d83bf5145219224fc
f8027d5f31f405d87680e979faca19013fb4bf3a
62985 F20110320_AACETQ west_a_Page_116.jpg
a01ab96663ebabe22f5c8970d170262c
b117b8de3f7ba1c073671528a27ce5634a3bda5e
98531 F20110320_AACEUE west_a_Page_024.jp2
09588dc97cf6e356c25aecd3c648ad52
d3b7471e6f4aa3172e17be18382fdafc21168490
20039 F20110320_AACETR west_a_Page_116.QC.jpg
b9ca326ead4018ba465d41f299c62514
ff4fcb48ab66e7f59a98e324a7f6484001870f7a
108128 F20110320_AACEUF west_a_Page_025.jp2
21c9eed2d87e01ac851de1f84817e76a
2f1fc2f29e114d10d62c89bca4a13d30b36b29eb
12561 F20110320_AACETS west_a_Page_117.QC.jpg
cee9cb9d0343512cbd55e992a50f0bde
8eca0af6dab43283e3114b5ceb959359a02f3c99
99870 F20110320_AACEUG west_a_Page_026.jp2
856c1ce8a7acae3ac8c82c250b49630a
1185e1e1409b7a76971d5d730fbb7f05adf8c1c5
10112 F20110320_AACETT west_a_Page_002.jp2
adace080dcf47e57a040798d3dfb5347
ffa7c490c3f3a9ba75bae85b526e594fafb7bce6
96512 F20110320_AACEUH west_a_Page_027.jp2
6e774d4292e7e92d6230da2c36e86a72
fea12f3ed3c5a955b50f30a3d029ff69e15028bd
91361 F20110320_AACETU west_a_Page_003.jp2
085cc2b3bb4174edac520eba94ebfc91
58ab1515498e428ee6b0903b35190df78178b7e5
99721 F20110320_AACEUI west_a_Page_028.jp2
1942f40be51c50cc9a6d361c2a0ebdc6
c562b4a20b09885027622a18e6e2afd659976401
1051980 F20110320_AACETV west_a_Page_005.jp2
ea3691b30b3465d73aa3e762eb0aea8e
4793d316748d135392722b0294e3c7c899bcdd74
98735 F20110320_AACEUJ west_a_Page_032.jp2
9898ca63ebc6ee318e59be73d9ac5048
ec4819c1fbe309f7e52de1651ed1af473a51b6f7
421548 F20110320_AACETW west_a_Page_010.jp2
81190df0ccbf3bafa33e27718957f7f2
3623def7c14f088fc7252f59c5491df7e8bc2593
60554 F20110320_AACEUK west_a_Page_033.jp2
09ca16d8a8e528578544d38f8cf7feb7
4fa6cf43b682efefb64cbadc0b4fbac30b985e91
1037246 F20110320_AACEUL west_a_Page_036.jp2
fc7ad8e4802a32cb2d415ecedfb3ca7c
3e89c435a889be61ab79c90e98f8fada12a36611
79000 F20110320_AACETX west_a_Page_011.jp2
f635c2ce897612bb9a6a684ff134b33e
9e497ac0233a3d5d3a9da3f5ff8a2a229795dd7e
34825 F20110320_AACEVA west_a_Page_059.jp2
6f46b5025aab41320f9b9deb366a1b61
46d362b9e6a457c14791302fa7c48d2b2b7e913b
1051969 F20110320_AACEUM west_a_Page_037.jp2
6771fff5e8339b60221c4930d3cf08e9
cfda07b08b1da011115af6ffb46c9bac3764844f
27741 F20110320_AACETY west_a_Page_012.jp2
91b7a21d4f0daafe756cf81856ec4ed0
12c02f4faa04268aba821bddf3611dec093005a5
260531 F20110320_AACEVB west_a_Page_063.jp2
0424df74b0ffdd6177007dca4d5548cb
55f2435998e9864239a8025b7a20d8ffc0925b13
1042022 F20110320_AACEUN west_a_Page_038.jp2
10c56d48396d9434082bff74ddabc7e0
2caa802455421099c30d8c2ff5ca5230994e02ac
103615 F20110320_AACETZ west_a_Page_014.jp2
8bad4f4b3fa353bb47aa84994fb8e64a
dbdb17cdfc666c8c3b5c35abb3df03500a2ad561
22653 F20110320_AACEVC west_a_Page_065.jp2
a94d970cbef891b2fe77eb55449e7614
7538a371a64cb59b2b2cffe358cbba650ba69ada
1028399 F20110320_AACEUO west_a_Page_039.jp2
233997de99b95cb30a64e4aa56920f16
8234053f604856851d879efc677732caac33905d
84234 F20110320_AACEVD west_a_Page_066.jp2
5333d17221d7c1498e6555641b894c87
28de672cbd6b059389e34e116f76cd4131588763
902445 F20110320_AACEUP west_a_Page_040.jp2
b475ec060c82ba35c4c4c1b8d18cc056
f348a4fd12571ec49873daf123b47005be113f23
100114 F20110320_AACEVE west_a_Page_068.jp2
928a544fb8a2a2a2ea61802da4d013ef
38552951ac3ef021d5a2ac52f9a9ed228b72de0c
107543 F20110320_AACEUQ west_a_Page_042.jp2
624474bd46117b1ea4ede6a94d5606db
777d66d657be2b3d2242222b2f826fae440f8fdf
104196 F20110320_AACEVF west_a_Page_070.jp2
b712c0c2de3356e75f5a41c74e13fd4d
75233a3e5b2d90c88445e8dc65f53a3985b14ef8
898630 F20110320_AACEUR west_a_Page_043.jp2
73629c1d408cf472d7924d2ec62bda79
4fd401288fbdb62858c403177b0e36a3e1a9456f
105399 F20110320_AACEVG west_a_Page_071.jp2
f08f52330b39465905589910397decb3
18501d18e34acc01b224a94938003325ed1c84f2
101793 F20110320_AACEUS west_a_Page_046.jp2
d8ad5250f9434961de0339ac8cd37d26
f1b909e9825fd63d1169f261776052aa67063c91
728584 F20110320_AACEVH west_a_Page_074.jp2
5be1a8fd70922d80dc23d232e88758f5
4aa0e36d42f273f5147bf3cac17e72e7bcc05242
F20110320_AACEUT west_a_Page_047.jp2
505279754240e02d503fb4d0a36da496
085af9d8fa40360b17563619aff2741019389f8c
669980 F20110320_AACEVI west_a_Page_075.jp2
e7f69a8acb2f948151b2cc56267b0821
99ff60a610f61939b0b2a575605149fe7fae9b73
104600 F20110320_AACEUU west_a_Page_048.jp2
5a9f95faaf5d87ab192829fc5ef6a6f4
e25b5d9faa8f7a48dcc0b5a869d0ed7a22c86644
647369 F20110320_AACEVJ west_a_Page_076.jp2
7e0b64b7bb05fb1dbd0296a6b6c84e45
7fd3dd26815b52c0924bb5f89338b634a9af2d48
238993 F20110320_AACEUV west_a_Page_052.jp2
7ca650d428ad0531256e9161f3f1b5bb
6050408ab333e4cba8571c2e231e033433f70147
680902 F20110320_AACEVK west_a_Page_077.jp2
1cb5d2a13f5a4feb9cab7d2ae5b1c282
abacc11c346c77796a37eee0ff5e3ceeb3e0ff1f
179450 F20110320_AACEUW west_a_Page_053.jp2
1389b9058a2877068883850bd3f09670
65aeef73b3792042811226949222ad9d355e4a07
526951 F20110320_AACEVL west_a_Page_078.jp2
6ba71fcbb9e0ac0ba7f52d518da2cd92
2379194573b26904d93240b87774cc1268c3b915
224812 F20110320_AACEUX west_a_Page_054.jp2
7f045dc0aec9481021b180078a22497a
9ef5b062a97d1aad700ef0024d538455be91a03a
476124 F20110320_AACEWA west_a_Page_103.jp2
529a6fc324f812ba04e216fb7abc0c53
4747d70f7de8267bf206111df9fd32192c3085ce
520116 F20110320_AACEVM west_a_Page_079.jp2
ce05e788b005d347a919bb44c5c583c1
65b71f93fa5a491bb3293f9e95d26ba71caa5546
648150 F20110320_AACEWB west_a_Page_107.jp2
959134dbafa7a1c9fa1de59bedaf95f2
00f2a117b86addc6f29db9b8538ab047682ceb8c
32590 F20110320_AACEVN west_a_Page_082.jp2
caeb690809775b71060dbc70e824c2c6
f4de46a5d8a67c0682c22142036c636fc0c086c4
240693 F20110320_AACEUY west_a_Page_055.jp2
eaedf6f57e1f1f710bb0429f73962188
4302f51dbe16fefd574f5ea10de55e7db6226949
102741 F20110320_AACEWC west_a_Page_109.jp2
627404e85d4886c4687506ed65090bc0
2e23ded355a2e9a298c4af5fb47acc704acd72f9
88867 F20110320_AACEVO west_a_Page_084.jp2
e13705e35a155d573d94bf3e39650028
8be4b385757f781e5c50412cc3882c5a851e6438
63728 F20110320_AACEUZ west_a_Page_057.jp2
ebf63e5e4986f653fe1d06497b46b3d9
8b04e6a83723e8ec915b4a03e88b14f13d88ae4c
110886 F20110320_AACEWD west_a_Page_111.jp2
a3d9e491ffdb4ec93d76a290239e36c5
37400f2b0952427d9a00d9a433b2bc0c765e5c0a
108492 F20110320_AACEVP west_a_Page_085.jp2
0d56f56275e90f0cac7d0cf0430ae567
d26941c0991991733f46f58f4494813fc90fe298
104711 F20110320_AACEWE west_a_Page_112.jp2
d04e84662eeb2b388dddfbd5b9c66fed
20b06d8f515fe03c99bdbb1ecac9e34ad0a13218
93575 F20110320_AACEVQ west_a_Page_086.jp2
961dbba49ee236f9d4b0809d4e831aa3
984ea78b35cd2d6aff388e4fb4d55760a458b7b1
92949 F20110320_AACEWF west_a_Page_113.jp2
61a9bc53de2f32aee3e75111dd64849f
0162b4bdc4971dbce26dda559e27c894f543b11a
100930 F20110320_AACEVR west_a_Page_087.jp2
295a84f233ab8376f221ff9907baa29e
1c2b7de481a76ce2a789ff8bd2cd763e6a7d73fe
111723 F20110320_AACEWG west_a_Page_114.jp2
b2a2b842cddfffcadf5dffa2045aff10
04fd091e1748dc407a681517234b5ceb4802300a
96949 F20110320_AACEVS west_a_Page_088.jp2
5f24e383eadcd1ae634bba9c0365ccfe
e5b343e306483f99a0879dfd3f70a7ac5f8cc4f6
90952 F20110320_AACEWH west_a_Page_116.jp2
9abb5483d394b4633e6098fd6cbee42c
73389d42268a0f5a4a148c9f8cb3a9de35ff5497
99421 F20110320_AACEVT west_a_Page_089.jp2
c886304f77d30be7f5c7cd8ede25d805
7e975991764d26af23c8dde1fc24220cac8732d4
2410 F20110320_AACEWI west_a_Page_001thm.jpg
c7aeb1b593c6cf8d160ee025bf19e354
aa23f7bf5406558c87437e17f2b9973f6dfa575a
101604 F20110320_AACEVU west_a_Page_092.jp2
d2d9d2bd001ec933b75691a481eb6258
6b049a5a5813339a58fc08733e091336e4e2684f
1591 F20110320_AACEWJ west_a_Page_002thm.jpg
3fa8dd3b73e008478042f328d8b03139
0df07f55d284bcbd05b597514327958f1a16bb71
20842 F20110320_AACEVV west_a_Page_093.jp2
1b2fbe98d2135fe45470e2ac209d6f4b
da109234c0c8e67798a54ef2635e1fb19a0ffdbf
3504 F20110320_AACEWK west_a_Page_004thm.jpg
91540e36114890ee4fedf2a3bc34611a
1f4db7193170a916949d281dec1e795139be23a4
348937 F20110320_AACEVW west_a_Page_094.jp2
4be4efe51f918714704b351105b5d528
be1b4bfb5b3683d5bc271cfa064f780ccaedd6eb
4856 F20110320_AACEWL west_a_Page_005thm.jpg
42e522075b813ad2172f671220ac38b3
2e77f78e825d93d41682f37876dd347d15c0793c
312797 F20110320_AACEVX west_a_Page_096.jp2
75266188661453af942ff40d6c05b13d
3c8fd24b076db78242ea38e5e92f3dc8c1091b61
2405 F20110320_AACEWM west_a_Page_007thm.jpg
8a159f1fa2b15f66758fe1a39235d642
f844e28d844b7f7c96061c71bcad3936df93cea4
415648 F20110320_AACEVY west_a_Page_098.jp2
d9a7e822789a15c28f2e9e5da3e8ba8e
f03c8c280714d2b8719c8542b91a56667c21b889
6050 F20110320_AACEXA west_a_Page_024thm.jpg
28dab823ab336440574ccfe0af2043d3
cd2e0267518d5bf27b5740c9845dad694c140f31
6201 F20110320_AACEWN west_a_Page_008thm.jpg
55f5ee939ba607f84ffe433d85339ea9
00aee815d7cdf62bbcd40d70ca4733ca4d7a9fb4
6357 F20110320_AACEXB west_a_Page_026thm.jpg
75931589f583ffb2178379b0908e782b
ff8b6fff0bc04dda406eb2fe01033b269c4e3014
2175 F20110320_AACEWO west_a_Page_010thm.jpg
78c8c2e4eaef510350783733753eac18
eb7b0daaf17ff368b39045c14e648c279e2420b8
454972 F20110320_AACEVZ west_a_Page_100.jp2
f16e84c25e33714f53212e965910b9cd
6ff277457037812a0b4694536a1f26dbd0fa5e46
4405 F20110320_AACDUA west_a_Page_034thm.jpg
3538430dfb81bcc4c7ad29ee1a31bb04
68e6640efb53f7e8f2c49700fb47d68d1d29d10f
5898 F20110320_AACEXC west_a_Page_027thm.jpg
91efcd8c70f77ffed804f39d2ae282d8
63a2c07935b6cc774924196051bd4eb4e2d6cc46
4870 F20110320_AACEWP west_a_Page_011thm.jpg
ae3748dabb4c86472cf63891ae19d63d
3af0d45d9631b1a0fa380ec09a4f9a5c8a82480a
28317 F20110320_AACDUB west_a_Page_101.jpg
efb539d4db652de4f4ed8457c111f6cb
fa7e9e0d58ca367e586143405a0c9537929e5d62
5849 F20110320_AACEXD west_a_Page_030thm.jpg
f49d42b569a363d9e030daeff20dfbab
1773af3997c81399af07985bb178ddb1e4f208d6
5821 F20110320_AACEWQ west_a_Page_013thm.jpg
9d11684d3084bf672a9d463b7bdde432
8b9f86583b152095d0d354e5611166f65a83232b
3445 F20110320_AACDUC west_a_Page_102thm.jpg
7b14c0076389bed6439c69e616b49cb6
0d9625b35ef7fd4bffe513b284cad3aaf1098a45
5574 F20110320_AACEXE west_a_Page_031thm.jpg
f473c594af4dcd7e99b1a87bab3bdf2b
ab554a65dc8f9ca71db804e2321108fa73443cf0
6354 F20110320_AACEWR west_a_Page_014thm.jpg
c749d23b34a185cb0d3e5d7023c281ac
7f7ff77cdb6f5ac8a90c9c6618739278a2905159
9472 F20110320_AACDUD west_a_Page_099.QC.jpg
670a3a995a0e37971f33bb702fe91cd4
eacff06a5c84de4142d814bff33cef4e825cf987
4045 F20110320_AACEXF west_a_Page_033thm.jpg
ac16b099abdb4b50300310d2b0633a7d
bfa69ad2b7e256274df4c2e0a3602375fe0f4255
6395 F20110320_AACEWS west_a_Page_015thm.jpg
4c1b6ff16ed369ab278bde40c9dd5759
217a7bfb06ab410f4bfeb75e6073eda0268ff0b2
63286 F20110320_AACDUE west_a_Page_086.jpg
52ae5ad250ac021806de47ce78937f26
f208f0d4e887578f74ee4c4b8c431a0bbc183dd8
6091 F20110320_AACEXG west_a_Page_035thm.jpg
0877c6d3c1b50a0ee4ff188f0686ceeb
063709166774ec35185cae8ab41d37f3cb527bc9
6234 F20110320_AACEWT west_a_Page_016thm.jpg
fba58f276a4ec47adcf79cc72e422622
2d523092c05cace0661d9b97f0733ff43db5770e
67078 F20110320_AACEAA west_a_Page_087.jpg
e26d626c40056cb9eff808c87d3cac27
2d30ca655e2715a484bddfb9df9e4202b558f1e3
F20110320_AACDUF west_a_Page_009.tif
eb3108e2752f291e8c844f44d700c98f
2b6ae3193aa9219ad3b2ab6a7378e75f936318e7
6645 F20110320_AACEXH west_a_Page_037thm.jpg
d5adb64dadc83eb94e54408489b07db1
310e7b22fbd5452477d5a74cf1a0868bd5342342
6573 F20110320_AACEWU west_a_Page_018thm.jpg
5d9e9f271df3602b4d4b7663dd6e4ce6
1d739d4c5b601bd1c9e6ebe424906023022aec49
1051942 F20110320_AACEAB west_a_Page_108.jp2
0a78cbe8e9a1e98d8c04f4f839eb3386
5b11ea796df632de25638c3c8aa71f3ce0ce504c
23011 F20110320_AACDUG west_a_Page_091.QC.jpg
9a3de30e8a726a901d421cd7abb2e705
df30098e6dfeda0a6f36a86788d23bf9b67b09b9
6621 F20110320_AACEXI west_a_Page_039thm.jpg
7ae18c0898e3d216c6be626681c7ca8b
7d42f8b65e426539a6f9022abceff6f0c29eb928
21959 F20110320_AACDTS west_a_Page_089.QC.jpg
70e46e1573ce5b13f291ab16b1922371
18deb71a53264963d723cfdb8d03429f215f57a8
5682 F20110320_AACEWV west_a_Page_019thm.jpg
1c2caa8953fe128239e60b60da94dbee
e019c5380ad9796f2d1d1fd96ac04e8c835f4878
46321 F20110320_AACEAC west_a_Page_036.pro
708b350f37489e47a76efc73178c4283
7fee1f8848ebed3f95beddfeb393cb335d4915de
1858 F20110320_AACDUH west_a_Page_061thm.jpg
b20d93a09041c1f14f9a3d32d43b3404
a07eca3154beb85bb770beb8af3fc7288e796bac
5956 F20110320_AACEXJ west_a_Page_040thm.jpg
7557a0372b184cf6bb50014dabf29aed
1b250736760e82dbe357dd6b2c520c168d756072
F20110320_AACDTT west_a_Page_111.tif
98c6a78512a5dd4d6d83b711299cb43f
db625f12eb156372bde0595e04c6168434490adf
3753 F20110320_AACEWW west_a_Page_020thm.jpg
b9a3352675ae9ad8153e2cacea6fc7e8
ff869e8ab44508372c67b0c05b7bab72aeb64942
6226 F20110320_AACEAD west_a_Page_028thm.jpg
a816b89a99317a5b0c8412fe918ed621
b28be326c4d2555adb6d65abe59e0f6e62bd8d7f
1045188 F20110320_AACDUI west_a_Page_018.jp2
848609b57ad829eec4e05a3fac56030c
3ae7798e66b7aa15748e2c38fd0c97999203dd19
5777 F20110320_AACEXK west_a_Page_041thm.jpg
638a7bb1cddc8a5bce385417a120b98f
31215ff882b54242722fe8f511767796911140db
23337 F20110320_AACDTU west_a_Page_051.jpg
ffe448e9ca3d1d58d06519e11932feb8
91981515ceacb8f94063a2e9f745d2970d5f3012
3255 F20110320_AACDUJ west_a_Page_005.txt
8e7fcaa2f6d53a9fb6e1c26598ec0009
71668fa1daf566c4bef239b1816b49b23a7e6445
6698 F20110320_AACEXL west_a_Page_042thm.jpg
cb566b2b506fda35b4e4cf09bd25c9f6
968d8ca2dea94f32e11948dbe1e70c939adf2b88
110838 F20110320_AACDTV west_a_Page_110.jp2
1d011f191d8b657f3849f248de6648e4
9d40722bb1f93cd0a5376f96ccf318915da8fd2f
2577 F20110320_AACEWX west_a_Page_021thm.jpg
4aabae54c25a66e5720c7e7eef71cde2
26664596363a1cc98c2479c418d6a5fa2945e594
2868 F20110320_AACEYA west_a_Page_062thm.jpg
ed4d86d9c781f1c44af8ddff1910851b
08ab8b7f7811cf3983f6c802b85533d8c9bc0f84
F20110320_AACEAE west_a_Page_036.tif
8f388b31eb34114ff31bbe31202f12df
6a52b677fb76efa56c08662463aa2e93a325469f
23638 F20110320_AACDUK west_a_Page_001.jp2
d13b30615fc57b9979193d1a67b30426
90aa2db97dceb0f4fe47b80c27da4a8063cd0b83
6072 F20110320_AACEXM west_a_Page_043thm.jpg
2a2929bc89e4a7e08b8a28a6019ab415
ff64cc9d2178aa818b008fc3601f25224f48e5c1
50690 F20110320_AACDTW west_a_Page_090.pro
749c6dd8fcd6e9af6e54a724e9a9f3e2
07c2043b49fbd0567bd67e74e31197a52ba2c630
5480 F20110320_AACEWY west_a_Page_022thm.jpg
619b8fc09c9bd8de1a2de861a28daea3
09855a9ba5b290a1055528ffce4eec742bc00a62
2825 F20110320_AACEYB west_a_Page_063thm.jpg
c9fff874bb18db7b828e84d8cd1d42ef
df631c976d95a2b52661ade86f664d61ef15b3a2
107770 F20110320_AACEAF west_a_Page_115.jp2
ad78f48676a3271f6c1babd3f03f299b
e6867ace17cb29d8f4470b73954658f35d910730
43390 F20110320_AACDUL west_a_Page_050.jpg
548a567df1807f2d3ea8a6ecc7a14c16
1dd6d1ec78df6cf235859e6f008b9836457f1976
6456 F20110320_AACEXN west_a_Page_045thm.jpg
1cef18c86bf64ddbcaf1a2f210c175cd
1eb866168e63192498951806da88b018d8ea042c
468762 F20110320_AACDTX west_a_Page_101.jp2
2c5a90b3e2210fd31174147205122c05
1ffebbeeb63a3fd7dcfd055a110496a3c5867fa4
6734 F20110320_AACEWZ west_a_Page_023thm.jpg
dcbc7edb12f65365fdd79a1eb4e2e9ba
62ff7c96d01ef7a0c0b7c7e1241143472462bef0
2676 F20110320_AACEYC west_a_Page_064thm.jpg
d2ffb2e3e04ec6bb2586a19cf79ae876
aafa9c5bab5e45c54a2181567d79815803b9cd0a
6009 F20110320_AACEAG west_a_Page_017thm.jpg
9910e16dfc573b86b2112fdccbc75f81
167ecc2f7255a4415a37a12f73f24c273f142e3d
5125 F20110320_AACDVA west_a_Page_103.pro
da0846b1e0e3ad08f4203352f8757ca6
414def3aa7fc3b5d3ac7cb16abbf4504b476cea4
21989 F20110320_AACDUM west_a_Page_069.QC.jpg
665f83ea5c35829711263650b121d4cb
d2af3d1079dce6cc7b103f6f5b9a3b30d13492b9
6237 F20110320_AACEXO west_a_Page_046thm.jpg
52e85144322af77277a2f687beda3c8e
24cec4f46db5d963b1b9b675519c2d6dbe806f15
2898 F20110320_AACEYD west_a_Page_065thm.jpg
1b7cee5d2faffd5cf64a3730ab34b3c7
02bc80ef4d2c14bb57aeba57966e38941b8ede14
25443 F20110320_AACEAH west_a_Page_007.jpg
e16689ba1b1eb239cb931996bbfef3de
7911c11817daf2ee83b720f29392db3f4f1feee6
F20110320_AACDVB west_a_Page_093.tif
3607f707fef475b0bbd41a85a051da8e
2d00717a27b05618ee91f9d94dc7709814d4d4fe
45308 F20110320_AACDUN west_a_Page_024.pro
1cff8fca3617c59e9ca389dfbcbe9b65
33db3970c0eebdac0dd822cc00a1db47c7ab9d6c
6836 F20110320_AACEXP west_a_Page_047thm.jpg
c3b7cf6cb1ca64c3c6f7035d1fcb67a9
c86e96a8efc7435dba757604927666bcd02206eb
7208 F20110320_AACDTY west_a_Page_106.QC.jpg
6d3d673378f9757ead5a62997c298cf9
4099b1e38f44dd016434144ae5a106010022fe6e
5604 F20110320_AACEYE west_a_Page_066thm.jpg
26b641ba46a0a1c41d8986f4be90feac
9e55333a319ad4ef8b20f4ae23b66f18c0541186
F20110320_AACEAI west_a_Page_013.tif
a1bcc8dd4889960ccd51c6717f002dfc
ede7ff1901b9525cc31cb808fe61ad385adfb32f
14765 F20110320_AACDVC west_a_Page_021.pro
d4f056faa519a6b33eaf7ee4a454b949
ae61e69cac0b5af98b918637485112ee900f713c
5683 F20110320_AACDUO west_a_Page_060.QC.jpg
79a053012c1e4a2980ba0459a640d10b
78cf1847d4c9dbd5d106a2985cb88e74ff9391a4
6384 F20110320_AACEXQ west_a_Page_048thm.jpg
28cc9aaf045d3c0207f6977bbb1f7a3a
f6e30b7689a0dcac7baec2f17acc17a299a04e9e
5277 F20110320_AACDTZ west_a_Page_006thm.jpg
18a337a23a838e9ac37a0f5b7723070f
aab0b42fb4a102d9e3acd635f0f69542ca952fde
6196 F20110320_AACEYF west_a_Page_068thm.jpg
2ee66b5ba93511dc676b94e138289fa8
2516b78021d65231edf92f5f0f65c69d83b46a5e
F20110320_AACDVD west_a_Page_079.tif
9922f96ef4bb2b1baff45f510fcd9ecc
0e274ff12afbc20f606ad24c80cd3b1ffc79fccc
67849 F20110320_AACDUP west_a_Page_109.jpg
c7c61460b3e9ade372eea37be603fafe
cc07bdf8f4183e5a321b9fdf44cda39d72bcabd3
6649 F20110320_AACEXR west_a_Page_049thm.jpg
4cc2892b7ff1524bf4f6df1d855d2394
cfad79e2a8ff66a903c46940a326e5c7d11db393
704 F20110320_AACEAJ west_a_Page_021.txt
3b17820db5af889ac881372defeb0a82
16a77b7f9ba7a356e4d1046e7272e11b2f4e6511
6183 F20110320_AACEYG west_a_Page_069thm.jpg
83a3bd93d60d6223b63249b5b4df0c45
1ac09e68a3aecd2ac30796ab7e151334720dbaa7
38223 F20110320_AACDVE west_a_Page_117.jpg
22c6174a48867e747b2989348b5c05aa
560fbdb7b018ed49a1b3435153a014df173dfc38
9260 F20110320_AACDUQ west_a_Page_059.QC.jpg
2c910411bc283fa7f632c908f2c4db3f
a0f244a434c7da697ad2419c0250f0e622de2281
3980 F20110320_AACEXS west_a_Page_050thm.jpg
b68c5b14498c94846e71cd2b3a177c78
89b844fad4a786631d7b1121d9059b3f59c785cb
20381 F20110320_AACEAK west_a_Page_030.QC.jpg
7e22e9df9c0132e49e28855b3b1cc956
3c29dd0cd38c761236dc3441eb4d679ff8ea5ad0
6629 F20110320_AACEYH west_a_Page_072thm.jpg
55b5fa7fa0e2068af1c139b7d08f0cc5
6c496ded5669c4d54552a321af60ed4665a2fdff
92387 F20110320_AACDVF west_a_Page_030.jp2
f94f4a65eb5f65a368c85d9eef059852
d5e3c45f0bb419dd16ef16add4f8fa36b28a5e0e
3750 F20110320_AACDUR west_a_Page_104thm.jpg
aa9645d061ed5de85ce9fdd55c89c1fe
464f0bbce0beda34de5b252ad7404a8344935311
2594 F20110320_AACEXT west_a_Page_051thm.jpg
eb492a668181eab957c5d605cbfe4e0d
295b3ed2581c690cb7de411cad33e953642ceb86
743 F20110320_AACEBA west_a_Page_065.txt
dda5abe442a1e95ef8bcb545c5268d33
bd858246b5c23964f548e3b9d87dc78cf385826a
22980 F20110320_AACEAL west_a_Page_067.QC.jpg
b4f307f4fd1bea8fcc722139fc49b625
0d72a48494b58871dd5786776d6dac543440f32c
4645 F20110320_AACEYI west_a_Page_074thm.jpg
6fa632e17ed5c7997b53733eaa412a3b
20cd52b7b33bd1032f438aee4cd69b16d9a1148c
386 F20110320_AACDVG west_a_Page_095.txt
2e881f39285e65ea2554e33eeb3fa05a
2d5ff57d6dfb929be4464a221c36e7c46bc0581e
254164 F20110320_AACDUS west_a_Page_051.jp2
994ed74fdd2efec5d654d9faae6019f1
f626f22b7addee088fc73a3afd13713af59d3a0b
2166 F20110320_AACEXU west_a_Page_053thm.jpg
812d190940c7d411f8a79ab4b6106b21
5735b29302ecab19e9452bb05a0073f14b5564d8
F20110320_AACEBB west_a_Page_025.tif
d6ea347047f2ba12bd50edbbb1d2b628
efb1b8c4d9b9c8ea67640141130ab9385c8eeebc
7789 F20110320_AACEAM west_a_Page_065.QC.jpg
00dc1c173a312c44b98b253fef875214
9d53e5f832edde545c76770a272cb7944d62b7c3
4444 F20110320_AACEYJ west_a_Page_076thm.jpg
66505da136fbbabb12c5e2733975b379
804d1a352798d8932abd7f1d0b16fb1a8e156c10
39970 F20110320_AACDVH west_a_Page_084.pro
e27f977dec8a78cb5e5f2689c052302a
a639856b30476e258465849a8c20567ebc3059e5
1020 F20110320_AACDUT west_a_Page_081.txt
7748f4743ee22eafa6bed9b2d07c0eba
0acd940ca88ebfd2dbce68212046974c8c17ff6c
1484 F20110320_AACEXV west_a_Page_054thm.jpg
9057219b1daacf702cb741ce693c7d5c
631b93ed37b475b5bd164cd122bd32deb404867d
F20110320_AACEBC west_a_Page_071.tif
51b5460aba8bb1e28ad58762c717a9e8
b24d1fd29b952a84c17aedba7c69f578404592b7
35659 F20110320_AACEAN west_a_Page_020.jpg
64dabce2f64bef02cc084d15d5cea3ce
47368711fba249d8964e1a6beacd4606d58f29ac
3875 F20110320_AACEYK west_a_Page_078thm.jpg
6719c328185b223edb12f33680d03d87
865664bceb83ce5cf1029c491fef8204804b4b5f
6480 F20110320_AACDVI west_a_Page_091thm.jpg
277ef23f9e68ed48a0e6cd7ac544cae3
c704b0b80fdd49caefe591d3d584d4e07548fce6
3610 F20110320_AACDUU west_a_Page_057thm.jpg
2e86838efa854b33a4280b037d703b71
65bef0317efd13f29fb6521403a2b90ea054acd3
1590 F20110320_AACEXW west_a_Page_055thm.jpg
9f771f7bf1b86a835e7a711dede8e0d5
f556a24491c45db0cb7d192905aca53d48ff701d
109438 F20110320_AACEBD west_a_Page_090.jp2
ac599db40c09e1d73ad9b5589a2d1734
3b366ce08be43bac0d04104754210ee0bb068fe4
52473 F20110320_AACEAO west_a_Page_117.jp2
782fcb5f2d2e6700064d36086326b254
07f4ac269d37240c1f7f5be0f376ada9517b1018
3828 F20110320_AACEYL west_a_Page_079thm.jpg
0a844f5695e9a37a807bd788e562f264
e9f169864c34c7e462e481f5a8eedd3e474bb66b
23666 F20110320_AACDVJ west_a_Page_111.QC.jpg
b6515fd08974d08ef4cf7ecacfdfc537
8db12088033a343fff2b1a8dfb51836325e4dd17
6545 F20110320_AACDUV west_a_Page_025thm.jpg
b12fa4457813735bfea6caa5c448c3d9
6c14291636dfe8fa24fe64fe2b1422c3b031ee87
3754 F20110320_AACEXX west_a_Page_056thm.jpg
6a4ccd1836a9631b000b27dfb5f8d3f2
a18ff1634e0ebaa3cabed8be489dca4e26bc7b0b
64971 F20110320_AACEBE west_a_Page_041.jpg
fca6153c04d10be23cd518e12a019930
b13dab2932d19fa58ad3becc6f1a1fa61ae3f974
641104 F20110320_AACEAP west_a_Page_104.jp2
5d0704026c44848b3d873a3cd3e79f5c
a17181040a8e3adcc38d83af659e5c812c672250
F20110320_AACEZA west_a_Page_107thm.jpg
6d5b6027cf2a1f511cd07865e2d0bc7b
a7b455c2f412bf3e041faca547e2f3d95966ae29
3408 F20110320_AACEYM west_a_Page_080thm.jpg
9cee9905708bf01eee1531c0254d394e
317f3444a3ed731cb25933d748ce70d3680eaef0
2947 F20110320_AACEXY west_a_Page_059thm.jpg
4a360e7ba99542685bae28af577141bc
bd78617ef91f328dbc8c5512d0fe3890995b1761
24362 F20110320_AACDVK west_a_Page_090.QC.jpg
7f6e0f6359536a95a8909eeb70085f42
1427a91ebc96cf6a9d5c9c7aea3d676098c14043
77393 F20110320_AACDUW west_a_Page_056.jp2
8a3f3067b690b26c6fe8955682f4e604
c973b33c82e792066be70f1747016844c83b6b33
35976 F20110320_AACEAQ west_a_Page_104.jpg
39bca34c79969dc609e723b037fd785d
36b15bf874df62717b2dea5ddc3800087fecb6c5
F20110320_AACEZB west_a_Page_109thm.jpg
df18c0fabe5ffc810ce9063be076c2d8
40d62d51cf144616609a6714bfd7a5b28835bff3
F20110320_AACEYN west_a_Page_081thm.jpg
628937576e610ca9fd00675bff598867
28a4a0d9594f2659daa2ffafbe49a361c5f32992
F20110320_AACEXZ west_a_Page_060thm.jpg
534be2d398bb790cd1162cde3c7d502f
9a1e3db8711460be0ab2e573e9f5ae9a3af842b6
6273 F20110320_AACDVL west_a_Page_071thm.jpg
ade77c436545cf5dd538dc4002395468
1430c198e43d5ac6334102ff12deb280d6fba72c
309687 F20110320_AACDUX west_a_Page_097.jp2
f655c6c0d3677e72be71548a0d112e5d
b3f02790b4c13b7b115a0ccec833dfee08533179
1891 F20110320_AACEBF west_a_Page_087.txt
1d342cba7bbc99d2a7e2a0043c91ff23
8df87028e8c654d7eed9921d3bb2ee4dce19ee2c
78732 F20110320_AACEAR west_a_Page_073.jp2
4ab14973932dfaed67766869b8b6265b
a0422ed39a39e5cbe7743041cd650f0d7282a7d8
6692 F20110320_AACEZC west_a_Page_110thm.jpg
d508af77c1c682e9b5bb5a1cb6bf6acf
620c7bd38b7003651d5399eae50ac4c8f974a085
2567 F20110320_AACEYO west_a_Page_082thm.jpg
6596d8c77800faed065f6b7d0e29d35b
11439356394fddae75a23992ec15ec6ffc48d074
5746 F20110320_AACDVM west_a_Page_053.QC.jpg
ca5d1f9f9bc4d4d7bbc1c2a041899209
a17af5c736f7720551080cbefac58428b8f5d101
7097 F20110320_AACDUY west_a_Page_009thm.jpg
97ec4732f4c6d7f9684390a677b70d76
95ca88808733f97ab8ceb44dfa88e03fc11d886d
6584 F20110320_AACEBG west_a_Page_114thm.jpg
bf177ca1cdcb45d5668d19a325a8c23a
6466fa1694e13ac254234de4210569bc440ee810
1974 F20110320_AACDWA west_a_Page_071.txt
1d6dd92f86ace6f9a2cb238cb61f9919
91759aa971920f6fa663674df5ef893a358638ea
26189 F20110320_AACEAS west_a_Page_083.jp2
4119106da26ce1c9efc5bfe5a6b9e2bc
bdf9f758e091795c4c2dd3e727e3c83640cbf1bc
6562 F20110320_AACEZD west_a_Page_111thm.jpg
2323f7bddc1385c3b35088836aca9529
600571b729fd61841217c6a88ee542e7d77ad4f1
2171 F20110320_AACEYP west_a_Page_083thm.jpg
79dd346edf59baf7ef2ac38e9d9677e6
4b4c2c69ff0c9863c38e8f5115c64c6cfb1685a0
65764 F20110320_AACDVN west_a_Page_026.jpg
803fc60e9def414ece65711af5e399da
65607350b436bf2652633f104c13bba73cd07e56
67978 F20110320_AACEBH west_a_Page_046.jpg
8c1d37dc7a8bcd942a6979ee16d10f44
d9b106a04c7a1ecf93c5beac39222f3c23210790
23305 F20110320_AACDWB west_a_Page_039.QC.jpg
8e121f6a0c8c82f2cd15bd216f53dd88
eddbb6c63c735d6d915aba7ccb51f97e8f0555dc
813963 F20110320_AACEAT west_a_Page_029.jp2
d1ef8ea0e881d46271080647c5839c97
57fd1f7bc108131b320c94a2a46fea9ebeb462fd
6440 F20110320_AACEZE west_a_Page_112thm.jpg
e43f1d93dac9b949cea2de01749bd6d5
4aea5a492c2ce0355949b07a2499bc78d610511e
6609 F20110320_AACEYQ west_a_Page_085thm.jpg
b31cf9b2c549a7bd24a9155b4e359b5f
fde71045892caf83ec825ea995f1049908d9375c
4452 F20110320_AACDVO west_a_Page_077thm.jpg
2cfc9ea4d8f979564f3d361f27a98a9c
ed15deda381c345d0c8eca15d952dfe1b8cd4d21
3268 F20110320_AACDUZ west_a_Page_108thm.jpg
0aa03f96440dec2502b42b9b79581ca8
6d2685ee4d98920de1c9bc7982cdd2c35812d361
94863 F20110320_AACEBI west_a_Page_041.jp2
745738cda0f776680a83ea31e035e3be
fb4c72d55a4ed64fd6d463a1be610f6a04987a2d
107964 F20110320_AACDWC west_a_Page_091.jp2
9ef7c99e4f754cb9e845f091325ac553
f5ed43f6187a1c65bf975abec8688666605128be
5887 F20110320_AACEAU west_a_Page_086thm.jpg
bf91ed978dc4e3cfa4d23a38619b7266
23e7ed984b1719d37a637e39b2dd00ad49a5af30
5901 F20110320_AACEZF west_a_Page_113thm.jpg
10e7e53699402458346466972a08db6d
2543a97abdb2c874a257c1d289eafa7fc0d38e67
6107 F20110320_AACEYR west_a_Page_088thm.jpg
3576376bd50b84bb1b71c33346885455
e5cea32c61f6af54b7d53281d6c93899fca63fae
14543 F20110320_AACDVP west_a_Page_077.QC.jpg
63af24faacc83a27998625d65f0a11e2
7f6221767cd16daa85cea0b1940b70eaada12e69
F20110320_AACEBJ west_a_Page_003.tif
55c4a1598395d2ab447a47ae9a9e83a4
8c5c5940c15695138a87df1bee39d469be735dcd
1983 F20110320_AACDWD west_a_Page_044.txt
596b2c69af707a4572dab934a3e6bfc3
b50ede1eb1603112ee8f4d2095f5d1974c896dda
1743 F20110320_AACEAV west_a_Page_035.txt
c3a7b3f383fb42c686f10e90641ca70d
ef70e3e68ba7bdf5cdb415cc8f2669ce90377ce2
6608 F20110320_AACEZG west_a_Page_115thm.jpg
b4f24585b4108f3e41b94c481d94815b
7dc0000cf405ca3c294005e771b50e9c71674f95
F20110320_AACEYS west_a_Page_092thm.jpg
0d24cfa1f991bd7ebf4801870a5301ac
c742efd5d9a147853b13b9fb5507e2b48ee1ff7f
39463 F20110320_AACDVQ west_a_Page_040.pro
c11761b15ae46fa362049f2b97d1a4ec
306b40b1fd3d30d04305d04247a13f543988b8df
F20110320_AACEBK west_a_Page_044.tif
32b73d8a152ab67f2120bb013b795d24
e0ee91ab61006f47932888dc93d11b53687b11a4
2244 F20110320_AACDWE west_a_Page_052thm.jpg
03bf3da976863b01c1f051257ca9d5d8
5a3eaca28d162176aa17827c56802f89f5928e82
2364 F20110320_AACEAW west_a_Page_012thm.jpg
fe9609ff1356d52a48f1448b5aae9be2
eff8ad13bd8462fd1bbd0d69b1c94592daee05d4
5692 F20110320_AACEZH west_a_Page_116thm.jpg
22457016764dd5ffbf10c2e5ea5aee65
bfaa1b138c1a698cb1423e58d239f502cfebddb4
2160 F20110320_AACEYT west_a_Page_093thm.jpg
4c27a16715f9388a15e0805fe53ea629
20fff44d45b08e4455c6373d0fce60f055101764
90676 F20110320_AACDVR west_a_Page_019.jp2
fcfbe5ee570982d137a4c00a0c28e64f
e359adfdf7896da77c71cac589142314c868903e
F20110320_AACECA west_a_Page_110.tif
efe42ecbb1901495f95581c613c6bdb6
ad98c1a21189de47859e7957c2e8f5dfe41360c2
46430 F20110320_AACEBL west_a_Page_038.pro
87231f3275c13e92cba3fea1c8fedaaa
bb6d00a70284ff3a8fd2c4c408474be3ccd94b15
1809 F20110320_AACDWF west_a_Page_028.txt
2f73c6e8f1f2688185e59b97b5a0de02
db7e7095e450bcfa60405bdb0c6253fa152bb0a8
61919 F20110320_AACEAX west_a_Page_030.jpg
b389104165e4e77ce72173ee766dffcc
b18962d191912ed79a3a0716bf63f640fcb75e34
6498168 F20110320_AACEZI west_a.pdf
5c0d17f05f40bfe47b549f38b84d5579
cdb65e5ffc0625c14af98b46894d59293b129847
3023 F20110320_AACEYU west_a_Page_096thm.jpg
37710c6944927e765e65cbb0e565b8f4
15a78f9dcd54681180e2ebb549850cc39e196031
45580 F20110320_AACDVS west_a_Page_004.jp2
b0152f39ca7eb22a8702879774314232
fa2ac396cd4002d95473219948dbd60986fcc5ca
F20110320_AACECB west_a_Page_030.tif
9cb48cd863b34b42322f831bbc0b5def
966b6a9664192ef913b2db69bde7b7c90fa7ee65
7316 F20110320_AACEBM west_a_Page_058.QC.jpg
6693ddb59238dbb8c8e6686618abbbe7
a2070217051ebd40e73f0ae2d671921fc561f452
65633 F20110320_AACDWG west_a_Page_024.jpg
777b81a7c0455ed90f595f5d65614a52
40818f77f31e75bd0d24282e8b8d9b0e702bdfa4
22433 F20110320_AACEAY west_a_Page_024.QC.jpg
8fc41c4ce69407c2de6ca52153f701a8
b56ab03dec3a0588120718dbb4553ff675acf957
134325 F20110320_AACEZJ UFE0009281_00001.mets
7ae5f08c85b5c73b1770ddbb23ef5732
60acd6a7e85bb0cbd8ee54b6cd30f23fcf8650af
2715 F20110320_AACEYV west_a_Page_098thm.jpg
e169731ab61676971c3ad896ac05d2f1
9d616900e6468f08be4d900fc559d6e224fbf180
20387 F20110320_AACDVT west_a_Page_027.QC.jpg
9a23aa627e1660df68c5a6ec5907e47b
ab5f481837f9a20ce8aa6400f8b980cc58537aaf
1051961 F20110320_AACECC west_a_Page_006.jp2
778d7d56baab3420699b0ab288c24f57
d734b74e334b932df42362e0767a3b3d26bfb189
23512 F20110320_AACEBN west_a_Page_094.jpg
f344299da264e2aac1b3e1c6d6a6ea0f
2e3dd79dfd2a39005d4687d944f94218a0e1363d
10979 F20110320_AACDWH west_a_Page_012.pro
ab0dbf6199cf71c5fdee2951ae0dfa5e
7ca64db55d6cc4b793a075b45d35603191edad31
12420 F20110320_AACEAZ west_a_Page_083.pro
579c19d82e87884ede2328dfc76fc6ab
03eeb1c799a63aad9c550dc5a9706a120b05f8f0
2993 F20110320_AACEYW west_a_Page_099thm.jpg
85ba1df0f6fd2c36e899a3da5a9eef99
511c5a7d1f297218a9a501ddfd6e9411205e2768
49651 F20110320_AACDVU west_a_Page_085.pro
8daef5e8feccf60f81bcb926f6f2cf07
02cfe69d09f883d5c691dca98f1d3f60329a6ca3
F20110320_AACECD west_a_Page_042.txt
228634f1f81a04fffc41f6b154f5f9cd
a0189b22afab5b904558e4eed04e02f606fd0c25
22459 F20110320_AACEBO west_a_Page_001.jpg
95ccc3a7d22bdd17ca30af4268936e12
39ab230e66c6440146aaffdbf6785d24eac17a27
65943 F20110320_AACDWI west_a_Page_088.jpg
f026f3261fa926005612e7c2de50a3f5
713bc6661306a1dfb43eff81df6a988ed634efb3
3138 F20110320_AACEYX west_a_Page_103thm.jpg
c0baa894f769a63dc8c694fe362f7d8c
7cd4ef56a7f36f95efa0ff2b81ffcfe7e6b4191c
75000 F20110320_AACDVV west_a_Page_037.jpg
a9fb7dbe6e7467ade07a70b1b0e5b146
cdaf1891b7acf7a75ec05ba5e10edddeff4bfa33
50010 F20110320_AACECE west_a_Page_042.pro
61a59015a8aa9e2e1df8cbd0a6c7acb7
a3a00339f8a1ae82b24a996c45a18199c7ea61c1
F20110320_AACEBP west_a_Page_043.tif
c2f7feedd57f1ac4e6ad1e51ceb37259
9636f33659868a55ecc2bcbaeec21b87ee47fdc5
4351 F20110320_AACDWJ west_a_Page_002.QC.jpg
9706d01dda909d42dfacf71a51f5310a
571e30e76b662e80f8ade87512503bf4a137c019
2529 F20110320_AACEYY west_a_Page_105thm.jpg
2542fe26b6f6ba9b8dd01ea80fe74406
9c74700dd3443e4daa295d3ebec11712dd637fc3
F20110320_AACDVW west_a_Page_054.tif
e4870b1cc3554e4ae693a72c21199376
8c3fa969a2d3d1880f6598466f9ce18dd2e2dd76
1772 F20110320_AACECF west_a_Page_032.txt
95a2c4498f6edf5efc8ef36c16f8ce05
8265072aee3c9a0f7e855329b5ee47a5eb8faaa5
71209 F20110320_AACEBQ west_a_Page_091.jpg
e38d08a4082b75c9aca920f93afa72db
21440e5bb7099d28ad10426fef9b47b2da029b66
19597 F20110320_AACDWK west_a_Page_066.QC.jpg
09e516efbd681f60d95e7361d1cb56da
d9b8e1868890e580f58c0f7c601d2edce5a3e565
2561 F20110320_AACEYZ west_a_Page_106thm.jpg
f15d788458dec623da400c39cb946c57
16ec29b3e6e2bcb3d3dd83f638d3eb129f7fd93c
42818 F20110320_AACDVX west_a_Page_017.pro
b937947595e9d8689479a3414a09d0fa
5ab26f35dee77461567a0ff247576cc97530bcf2
1925 F20110320_AACEBR west_a_Page_112.txt
b20acfab8ec6f9b869c3e293a2823d0c
3bf629267eb2e3cc0e954d52e91a32d4fc4be06b
5670 F20110320_AACDWL west_a_Page_084thm.jpg
2a9e3d7ce608da1c8c577d0fa550055d
fc28c70c5b0e6cd10abf40d85a00ae6d47f3975f
11017 F20110320_AACDVY west_a_Page_065.pro
e2a795741143e58f18b45f3930c375b8
f9f7b9ce67144084b2f8252347838290950dce94
19479 F20110320_AACECG west_a_Page_052.jpg
8e51227dcb2097618fff04ad991cf2b6
5f3122b6a5790420b6625252ab4e2e19a073907f
F20110320_AACDXA west_a_Page_033.tif
0ee20f340def97ba540226e456d65e97
f10cbdc510c6f876efb88a8207d00e222309f61a
38539 F20110320_AACEBS west_a_Page_066.pro
8823cdf6fc31eff9f628d9606c40244a
deb2290420d4291454b0bed3bcad3e792ac7f458
F20110320_AACDWM west_a_Page_020.tif
6744824fae92ce7a709312438c50ba5a
6534e78e3b34183c3fef9248a41f78493c1ddc38
19247 F20110320_AACDVZ west_a_Page_020.pro
052d0a92de8fefa584d230e7188c8626
9e44530b23c5b2488d772b0ad871b64247b7e312
76651 F20110320_AACECH west_a_Page_049.jpg
161bca88bf32dcc5f1441a0873e64034
35222e11d5a336aabc082a46cec3f28954352f4a
58376 F20110320_AACDXB west_a_Page_022.jpg
1a5d04b0894332507be4588eb04f8d4a
061e7503301d204228a407e78e94cf2ba67a3af8
108846 F20110320_AACEBT west_a_Page_072.jp2
5599266bdc0fea45fbe9bfa82a29793a
eb922462886f1f81b9e66c594981e8ad7aa8abd2
2913 F20110320_AACDWN west_a_Page_100thm.jpg
9de854960baf2dcfe4a9f05914e66649
2c4cc7ccd7beef6c07341bdd10f4b84fb5e93ff4
8027 F20110320_AACECI west_a_Page_100.QC.jpg
42ebb050126c54bf7fe1841242cd0891
b828c61f05ba35f53e0325edd695b4312f27e008
105802 F20110320_AACDXC west_a_Page_067.jp2
eba850e5abd78520e95526a2a4093e90
2cdeb8733812e09fb4fa40ed0d3af4a646b4d4ad
69386 F20110320_AACEBU west_a_Page_035.jpg
70302031b49c4976b300af548c4f47e0
7469a0345d191afd842de7dbfb71776cd34d86cf
24719 F20110320_AACDWO west_a_Page_023.QC.jpg
f25ee156757d652b7b13c08bb734d6b7
40c46336f056d954f79ee387998d923b0eaf94c2
21438 F20110320_AACECJ west_a_Page_087.QC.jpg
96953c41b27bd420fb572dfba778bec5
64c9a55b1e032acc95dcc487c72b1488614d51d9
53475 F20110320_AACDXD west_a_Page_115.pro
e9393fc70682c1e86244c985611a25bf
05f7358b65116aa58d2a951af695009c91346093
47183 F20110320_AACEBV west_a_Page_015.pro
bc38beeaf99db1e849b9c7bd7b6b7d62
8b8a29c099a0ad122938465d763f045630844cfa
1960 F20110320_AACDWP west_a_Page_072.txt
8b34e2f494f7b2874582773339987f44
2877c7ff9fe6d8237f7e9482c77dda500a92a851
F20110320_AACECK west_a_Page_016.tif
65b140099cdbf6ec933beca31d6f538c
72e9036b50e235f9151a466ecad6d8ef8b367f1d
3180 F20110320_AACDXE west_a_Page_101thm.jpg
db24de52b0b2017155954e79bb1c6727
ddafbe3ab657391656dcf5204ff635aa6ee2915c
43259 F20110320_AACEBW west_a_Page_056.jpg
d33a6cd1a354239c513d80f2868ff586
93cebbebaa7310e334eb71bc29947555fc1c89a3
6738 F20110320_AACDWQ west_a_Page_038thm.jpg
d4ab03cb43c4e8940b15e5457a90ec6a
e438cde6de1d70b8b35b1cc71d4458350c5b93a8
100916 F20110320_AACECL west_a_Page_045.jp2
ca0dad68ae35657a5daa60a4c69212af
bd6529eeeb802c94c7190571176b749e9a86fe62
F20110320_AACDXF west_a_Page_050.tif
3e418c1178255283cfbcc857881aeb6f
8295df516c25d7d16eee4684e58f7a841a707af2
46782 F20110320_AACEBX west_a_Page_069.pro
08acf64be91d567187c7d8c0dc55b3d3
56ff2c366a47e527756447d732b84a91b47ee016
782 F20110320_AACDWR west_a_Page_082.txt
fb747102293ad8310d13af7912fe43c3
cff983d62d7debff38c086a2ae037935fe458d90
6365 F20110320_AACEDA west_a_Page_067thm.jpg
f227fffe9bf0694be0e4eaaa10e97688
e287ffa63d7d626e8be478bd1d8d2ea6b61f5496
1928 F20110320_AACECM west_a_Page_046.txt
a8a833021351510b90d19b9bd8e4b792
1ad82f17aef65776cd9922d32c9d017d391b0f02
47780 F20110320_AACDXG west_a_Page_037.pro
d691be0ff3a642f95a36859a8a3ec588
508c7a54f139fb36bb193df0d2822d6d58a579e5
2684 F20110320_AACEBY west_a_Page_058thm.jpg
6f6e15a231f442dbfb3bc8bea11e1170
2853331f44ac2dd544cdcfa7c28f1aa4b1623ee2
F20110320_AACDWS west_a_Page_002.tif
7a9f331b6f50757c112e5f10caec53fd
e985772c23d2f20285b370170c6688be35d3fbfa
27752 F20110320_AACEDB west_a_Page_105.jpg
6f1d64b0e81d201d3c166699a9bb927a
27f98a08aa2ce53b8981d11158dac03f69ddc884
47684 F20110320_AACECN west_a_Page_067.pro
95acdfff6a27ae72ac9b213d54f7b74e
7419742127a14e2df1f64df05618006e7aede95d
F20110320_AACDXH west_a_Page_037.tif
4c7b3a7ce6b3553d847eb02235c377cd
2e3b5480634fa8df2c16c09d4e4f0570e6b1c628
647 F20110320_AACEBZ west_a_Page_083.txt
3e563b84d1c45238fd3d4677f2da8d71
8c48caba60464f1756907c12256fecca2ddf40a0
F20110320_AACDWT west_a_Page_006.tif
916479a4ab97d19bfe81d3543f8a1b76
d4068711bf436a79e8b642eecf37e2eb60a9f508
10682 F20110320_AACEDC west_a_Page_010.pro
0ba90b9c4e67b448df270ba73ddd2cba
df44c2651791efa418cf9f8864ceb2ed2e957b29
27162 F20110320_AACECO west_a_Page_063.jpg
a0c2e7681245e291f38cc914741f9dde
f1a2e950760e43731bf237ddc77f2419344a2b83
23269 F20110320_AACDXI west_a_Page_074.pro
7aa4ec55ccb69875979a922874c58a0c
339d553e91a47dc648c738633c4543f6f902131f
F20110320_AACDWU west_a_Page_104.tif
797d63888807a0ca2e90effd953636ca
9eb257c3c37cdc43bb54010928116de00e851f0b
22891 F20110320_AACEDD west_a_Page_018.QC.jpg
0af39b80fd15509d9673b6c688ca0917
a31cd1e6258139cf0c6fd24ea2abcc2a08f93f22
535952 F20110320_AACECP west_a_Page_099.jp2
59c235177735b7f447b7aa51a8c102b2
27c697d521b41662eb9a9735369c0a09354e6d3f
F20110320_AACDXJ west_a_Page_004.tif
68610e08f76b8c9170d30bb66027aa8f
dd2e000f36a29844b9b58cefc0502469019bb6ad
102330 F20110320_AACDWV west_a_Page_015.jp2
1d45d6f68d94c73d600742cc38e38034
9c318fa5c4be6b4024382098c9f8e05361692302
20057 F20110320_AACEDE west_a_Page_019.QC.jpg
3b0a0d76b02b73a857ce4bee5119e86e
342113a232efbd0bebfe600b204979f722f19724
20273 F20110320_AACECQ west_a_Page_031.QC.jpg
ef92bf01c51d5358890700b22ee20430
a95a0081578403ef35db040faefdb9cd9ddf7f70
70721 F20110320_AACDXK west_a_Page_085.jpg
647f02a7d58c3a8382f9ed1168ad84f2
5a47ecaaf716226a7b07d50111d4a7f1916d7768
F20110320_AACDWW west_a_Page_055.tif
39327e589c35f1f45e66a35f5838c885
9494033291a570cceedf52620c61f957ef6ebd3f
22455 F20110320_AACEDF west_a_Page_071.QC.jpg
236e816d91407bbcd4639dad02bb6b3e
3d4c31dc3b9fae8f63d2109f8cc1d529ff5dfc19
76809 F20110320_AACECR west_a_Page_009.pro
b98179e99a739a54e0c44bba54b5a8a7
d451f3bf72d856d643d792e4fd5a561c85852af7
F20110320_AACDXL west_a_Page_059.tif
4b8817da5cf77ee7b2016d1d69135404
7d52acc5cc28e4ec1b18e3db0a1108c91c72e96e
2259 F20110320_AACDWX west_a_Page_114.txt
8aea06abb3edd23820930760d1463d01
fec0c847845c5ec95a43599dc1632fe93e8f778b
15664 F20110320_AACEDG west_a_Page_074.QC.jpg
b9bdeae9e9258aa2bdc4a90d6fb6ed7f
f127a54db138a5e6548724292b66b80c1a4c403d
1619 F20110320_AACDYA west_a_Page_030.txt
a574d3d1edd99ace0152e5c5ecf392cc
9322eeabddc3b8f5b1bf5fbd92001bdd16cce42e
1051985 F20110320_AACECS west_a_Page_008.jp2
90e1849d7af08719108f78fa3e36f20f
b25c789d5b3b1f6b8e0e5a5cd1ce57c5958cd1a5
164499 F20110320_AACDXM west_a_Page_061.jp2
95a0436f10a1c0b62338904f63965774
fc8e8a62dc2fa874297c4960fe6686e203186119
4468 F20110320_AACDWY west_a_Page_054.pro
2ba1d68ad7cdaf736905787df7de6645
0ee38ec00d648316699139cc857a931d7487f04e
6253 F20110320_AACDYB west_a_Page_089thm.jpg
c791996a52869e0a23eb21c9aa8c83c5
e5b9a28707492596b88307a3b41732a6ad49ff37
F20110320_AACECT west_a_Page_088.tif
01ce2d90fe190479f7f92bdab352f068
105f26e1a715678004be7ac17d885acc3e795673
44223 F20110320_AACDXN west_a_Page_013.pro
a81b7188281fd86853db417285e410a8
7e241eecd4b8409aa13090ef3ef7152a8155cced
62732 F20110320_AACDWZ west_a_Page_003.jpg
65cb32b23864b168fbf7bd2bd0d5dbf2
94a53f988e97dc91ba25f08fd862bac82f4c84cd
1684 F20110320_AACEDH west_a_Page_084.txt
093754c6054d10474bf23a4109bdf25c
c1741d7426a5e7ed61883bf42cbca1ced3570193
9054 F20110320_AACDYC west_a_Page_063.pro
0ecc44ad89e6d6bc57beeeb42d06d470
007a4496a73a911aa374ebd6371aa0039c1327ca
1051977 F20110320_AACECU west_a_Page_049.jp2
648c4866fe27f60bd67a36ee613aa189
64ac4f81a8c627a325d047d0211c3f259fda4a2e
6045 F20110320_AACDXO west_a_Page_032thm.jpg
4f4d474265ec7ba52e76b852df118c3b
a16ee3c6ba7daf16bdb0814770212ce520772dbc
13714 F20110320_AACEDI west_a_Page_050.QC.jpg
e6c8a3cbcecf6cd6fd5c6432d571a596
5cc7084f82945eec6d8a3f0f07a2740228ebff14
489414 F20110320_AACDYD west_a_Page_080.jp2
574554ed0a9a2d0ab7b92fa98c4ebbdc
51b0f7bcf4e937c5089580e7b1f1cf0ea32e1bbf
10993 F20110320_AACECV west_a_Page_004.QC.jpg
33747b5e2d08bb0347335db222e3e36d
c15a2d37622eccab9e21a14bc2143dfcc054eb1f
506072 F20110320_AACDXP west_a_Page_007.jp2
028f029045da4844bde27adb8178e4e7
20c2619c7244bc93b44e910b0b05a664631a1dc6
42572 F20110320_AACEDJ west_a_Page_033.jpg
8346ed296911ea7ecb6b58fc06d29b35
e2f396be75dc6fe42c6e5a3fb10744e40b9b4596
F20110320_AACDYE west_a_Page_025.QC.jpg
b8fb5e422beb3f28bf87c5ff82e72bd2
182f8b42fcf88cbffad480226e6e9b226cbbfdb8
5620 F20110320_AACECW west_a_Page_029thm.jpg
4fcecd83c3381c35ee3f19e43b898621
940334b176f39c07465550088f1723b829721a7e
1853 F20110320_AACDXQ west_a_Page_069.txt
2f27ae50eb26c328077e93355d7c0acc
bd6cafffdae8d09bc3e8ee73b22b0532657184f2
74336 F20110320_AACEDK west_a_Page_039.jpg
92e7b07c03b5a54cbfc512284544d6e3
c68e67c3d0f9f9bef647a1f49bbc6427f79a0275
649360 F20110320_AACDYF west_a_Page_102.jp2
2d5dd0f118d50d278d924892b0c9b66e
4b449d58126ab53b8a8918a6164cf70b7f3a2043
F20110320_AACECX west_a_Page_070.tif
cc9b2e1bcff6a6f1c9df622626c25a21
2bd4cb7da7a47a9f052f697842f5056211bc678d
49311 F20110320_AACDXR west_a_Page_071.pro
948130729b0e41c5bcb097b57b49d515
d914f58ffbb84b0a3a16ec98d1eea67b0c429939
43419 F20110320_AACEEA west_a_Page_027.pro
3ed6a43b6de87d5de9bc742cde7faee5
6dc49d7c6030366aba7e44457b3da2fd437c16f0
276812 F20110320_AACEDL west_a_Page_064.jp2
00ee36693ef59a135c559938f40e28b5
c2ae86a4d455cb2f36a7a6529c4e71cf3eb43b3b
6168 F20110320_AACDYG west_a_Page_087thm.jpg
9afd1e09dd63d067a043413d8c2137c9
3d75d179152cfd36a6feec0dce6b2b57cdf62686
72723 F20110320_AACECY west_a_Page_110.jpg
c717303f9c884cd7d6b81866f65b634c
bfd48b81a12f93857dd5cbe18bb6cd81b1d7701c
78551 F20110320_AACDXS west_a_Page_114.jpg
8921eb78ff6736e3d8c7658b548ac950
9c0e066e8dab7f22900b055f85524b6625fc035e
7444 F20110320_AACEEB west_a_Page_080.pro
6ab857edc915294e5d69895f0980b530
b693a6eef8a8e0905ff677aa32de9fa085f06e38
63254 F20110320_AACEDM west_a_Page_027.jpg
116f04155444f0e88e07e1eca981e3ea
ed9a9e327d78e7e79e289bfd811566ac397f7fdc
1051866 F20110320_AACDYH west_a_Page_044.jp2
577ab8fd699bb0bc0f167c18fffde9e6
2c62107ab49cd96ef84f7193538eaffb636838ac
2644 F20110320_AACECZ west_a_Page_095thm.jpg
7ecfe4b0c0a931d91a8655874f5cac75
6bb24bd51a912b847e0530e5e4ca62ca39c0b032
3082 F20110320_AACDXT west_a_Page_098.pro
78ad2a779bcdf7d5d3c310d1553d927b
ad15f55429f950ee1a54b328adb03f33e96569fb
67830 F20110320_AACEEC west_a_Page_048.jpg
f4e42b8c2eaff9394511f39ca44bf9b6
7873407d2e6387b962ebb8e20aeea3cdffde155e
251870 F20110320_AACEDN west_a_Page_062.jp2
cff301689956093f61873e8d08bfb3ac
88bfe978d8eaf1eec11b82791cc312083ba9e334
1427 F20110320_AACDYI west_a_Page_073.txt
d3d2e5b59647d084b060af6d8dc1cd4e
03d912c156b1f1e85b1145223bd3430c463f524a
41244 F20110320_AACDXU west_a_Page_035.pro
0fd7bb395773a4ac2d0665cb827f849d
97d91b4e9630d7e4ba596ff3d837442d909fbcfd
4479 F20110320_AACEED west_a_Page_054.QC.jpg
eb8c49565d8019d9354411b49b9360f2
1253116a1670346cb1bdda7062dece2e6108d867
24540 F20110320_AACEDO west_a_Page_047.QC.jpg
38405b8f05402c66ed6466be7b76cc45
d01fd613cb4df66ad91c07a9f6e10da53f98ea21
6503 F20110320_AACDXV west_a_Page_036thm.jpg
f5eb36f91653b4fdbf705acf7a44d004
9a1f55ac724247898b2107269c4b79bf2db641c1
F20110320_AACEEE west_a_Page_091.tif
3796660c892bb1214e52e720e111d079
237f30819cbd6d7cb5713ba46f282d5692326e82
24840 F20110320_AACEDP west_a_Page_114.QC.jpg
9a1186cd251886887f02ae9e2094ec4c
59287b40189fd3c7417258d8b4059a426aaef07d
29612 F20110320_AACDYJ west_a_Page_097.jpg
7a7771a647395d4a405fe8dcdc81db42
95a4d21a037a53b2a2b64f3a0ff9afc310f449da
6803 F20110320_AACDXW west_a_Page_090thm.jpg
8d4d7aec6fe8f53e0cfc479ee6c4fe41
f39f997ceffaaf1a9bc91e85f506e6b16a707721
476783 F20110320_AACEEF west_a_Page_020.jp2
e11e874351e48048e878b8f7e2eb3542
48942b2db5137db3b0e56b45cac770927636c212
29015 F20110320_AACEDQ west_a_Page_103.jpg
30d392d3cc6a0c28b029764297551e8c
b49826f8c2ed17e0c178f3069c82037d7ae651d8
75918 F20110320_AACDYK west_a_Page_047.jpg
abe21b60bf3cfb03edd1df586b7e4617
4473123226a71c5a3dc1d950e6184f7fb0bebb4c
5726 F20110320_AACDXX west_a_Page_003thm.jpg
948f5ca4a2ed3a40c054e66eeb846865
853eb645039f5c10d7f966bc529ed655d3abecb5
546756 F20110320_AACEEG west_a_Page_050.jp2
9fe58af34e8fad7a7e633baec1cd533e
c739ef7b030b96b14016e9d2442733d9abafe7a9
13371 F20110320_AACEDR west_a_Page_056.pro
c50f9f93175b9ca0bc67d2b4e2a77830
302025ff648c3df32af8988486ceb5b7b96f6370
F20110320_AACDYL west_a_Page_087.tif
b5f080a92b1827faee8e3d16408c99ae
49e5fcb2097c0364b8e697c7bbb2ba58759afc01
1885 F20110320_AACDXY west_a_Page_037.txt
7c32e68f904235edc2841cc404bdc086
d4019537018d7f28dd04ba9faab683ac7f52303e
8828 F20110320_AACEEH west_a_Page_097.pro
76362bf3af64449f92834c0fb1ba046f
1c0609d822ee13c175846aa84ac9f4c3bdcb2c1a
F20110320_AACDZA west_a_Page_046.tif
ff4313c24ba1f77f2e4f68f98a074219
ff7bbd9ce8dc629ae46326b7e1de35f93a9e11b1
8798 F20110320_AACEDS west_a_Page_063.QC.jpg
ed2137f7a8fcf175918dd10ad2e06091
1ebab2d4c53d6801322f0630a64cc673979f48d5
426 F20110320_AACDYM west_a_Page_010.txt
0ca7b7c381baaf1bbd1b1cd8a9df68cd
d4d057c2111c99a20b9a1158078df414219fb58f
27792 F20110320_AACDXZ west_a_Page_082.jpg
65bf8d8135154865b15e92e14c6fca18
c9eddf4671887f7c67e34dafda9d11660b520d44
7183 F20110320_AACDZB west_a_Page_044thm.jpg
bae9c3f4a06da1e134536339e1e3e2e7
fad6b07d9cf7f4b0e53fa953c1245f1b54848ee6
222175 F20110320_AACEDT west_a_Page_095.jp2
0a4a0f286e0d61b138bea214d2d07fdd
c82628bf672be51bfb06589a59977c8cb9b9e82d
3817 F20110320_AACDYN west_a_Page_117thm.jpg
ebdbc232f0e0894c33d9f2416f6a9dad
4989d32b455f06cee61dd1f67b17a8def413b1c3
670885 F20110320_AACEEI west_a_Page_105.jp2
23c23f44ac5696f55a7260479f9fb910
f8faa202545a379145fcbe20f54b3fc9676a9cfa
12150 F20110320_AACDZC west_a_Page_057.QC.jpg
03b2f79b6a34061d1c4b52804cf5114d
8c0e689236aafffab5ae56c64c690205e12f34a4
F20110320_AACEDU west_a_Page_100.pro
919f4b7cf65035b9d0321d2e42bdea4d
b0ae385e347f0277079cddbb1012963884dbc518
1971 F20110320_AACDYO west_a_Page_016.txt
840c4e80f343b56d1510bf6a3a9b7a4d
f7de423bcd42f887d48f36b2f3a23a9afea3cf05



PAGE 1

EFFECTS OF DISLOCATIONS ON ELECT RONIC PROPERTIES OF III-NITRIDE MATERIALS By ALLEN M. WEST A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

I hereby dedicate this [dissertation] to . myself. (credit to Rodney Dangerfield as Thornton Melon, Back to School)

PAGE 3

ACKNOWLEDGMENTS I would first like to thank Dr. Cammy Abernathy for all of her guidance and support throughout my tenure in the Materials Science Department. She was the ideal advisor for me, and was crucial to my success. I thank Dr. Stephen Pearton, Dr. Fan Ren, and Dr. David Norton as well for serving on my committee. I especially thank Dr. Crisalle for serving as my external comm ittee member from chemical engineering. I credit Dr. Crisalle for his work as Gradua te Admissions Coordinator for the Chemical Engineering Department. If not for Dr. Crisalle, I would never have ended up at UF. I thank Dr. Karen Waldrip for all of her support and guidance during my time at Sandia. Many of my successes in graduate school were due, in large part, to Karens tireless efforts. I thank Dr. Stephen Lee and Dr. Daniel Koleske for their significant contributions to my research efforts. I further acknowledge Dr. Andrew Allerman for his thorough feedback and instilling strong sense of attention to details in me. I acknowledge Dr. Brent Gila for his contributions to the work in this dissertation, and I thank Dr. Andrea Onstine and Danielle Stodilka for their support and collaboration for various classes and qualifying exams. I especially thank Dr. Jerry Thaler for being a reliable liaison in Florida while trying to complete the bureaucratic requirements for graduation from New Mexico. I gratefully acknowledge Jerrys efforts in this regard. I also thank Jerry, Dr. Omar Bchir, Jeff Shar p, Brandy Colwell, and Sarah Rich for their camaraderie throughout my graduate school experience. iii

PAGE 4

I acknowledge Dr. Sam Kozak for the inspiration and advice he provided during my undergraduate days. I also thank Dr. Ken Van Ness and Dr. Erich Uffelman for serving as my advisors at Washington and Lee. Both were very important to my intellectual development. I also thank Dr. Douglas Szajda, Dr. Marcia France, and Dr. Steven Desjardins. All were superb professors. Dating back to high school, I must acknowledge Mr. James Morrison for piquing my interest in science and tolerating me for three years. I also acknowledge Mr. Robert Bulkeley for doing an outstanding job of advising me throughout high school. Finally, and most importantly, I thank my family for their undying support throughout my graduate school experience. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iii LIST OF FIGURES .........................................................................................................viii ABSTRACT.......................................................................................................................xi CHAPTER 1 INTRODUCTION........................................................................................................1 Background and Motivation.........................................................................................1 Electronic Properties of III-Nitrides.............................................................................1 Chemical, Structural, and Physical Properties of III-Nitrides......................................2 Suitability of III-Nitrides in Optoelectronic Devices...................................................2 Suitability of III-Nitrides in Electronic Devices...........................................................3 Progress in III-Nitride Material Development..............................................................5 Availability of III-Nitride Based Devices.....................................................................6 Statement of the Problem..............................................................................................6 Outline..........................................................................................................................7 2 EXPERIMENTAL METHODS..................................................................................10 Growth Techniques.....................................................................................................10 Metalorganic Chemical Vapor Deposition..........................................................10 Molecular Beam Epitaxy.....................................................................................14 Gallium Nitride Bulk Growth..............................................................................15 Characterization Techniques......................................................................................16 High-Resolution X-Ray Diffraction....................................................................16 Transmission Electron Microscopy.....................................................................17 Secondary Ion Mass Spectroscopy......................................................................18 Superconducting Quantum Interference Device Magnetometry.........................18 Capacitance-Voltage............................................................................................19 Inductively-Coupled Plasma-Mass Spectroscopy...............................................19 Scanning Electron Microscopy............................................................................20 Electron Dispersive X-Ray Spectroscopy...........................................................20 Auger Electron Spectroscopy..............................................................................21 v

PAGE 6

3 DETERMINATION OF DISLOCATION DENSITY OF III-NITRIDE FILMS BY X-RAY DIFFRACTION............................................................................................23 Introduction.................................................................................................................23 Characterization of Dislocations.........................................................................23 Previous Work.....................................................................................................24 Twist....................................................................................................................25 Incoherence..........................................................................................................26 Focus....................................................................................................................26 Experimental Methods................................................................................................27 Method........................................................................................................................29 Sources of Broadening........................................................................................29 Model...................................................................................................................31 Determination of Convolution.............................................................................32 Calculation of Dislocation Density.....................................................................33 Simplified Model.................................................................................................34 Results and Discussion...............................................................................................35 Summary.....................................................................................................................37 4 EFFECT OF DISLOCATIONS ON IMPURITY INCORPORATION IN METALORGANIC CHEMICAL VAPOR DEPOSITION GROWTH.....................54 Introduction.................................................................................................................54 Silicon..................................................................................................................54 Oxygen................................................................................................................55 Carbon.................................................................................................................56 Previous Work.....................................................................................................57 Focus....................................................................................................................58 Experimental Methods................................................................................................58 Results and Discussion...............................................................................................59 Summary.....................................................................................................................61 5 GALLIUM NITRIDE BULK CRYSTAL GROWTH BY DISSOLUTION AND RECRYSTALLIZATION OF GALLIUM NITRIDE POWDER..............................72 Introduction.................................................................................................................72 Previous Work.....................................................................................................72 Proposed Methods......................................................................................................74 Electroplating......................................................................................................74 Electrochemical Reduction of Nitrogen Gas.......................................................75 Dissolution and Recrystallization of Gallium Nitride Crystals...........................75 Focus...........................................................................................................................76 Experimental Methods................................................................................................76 vi

PAGE 7

Results and Discussion...............................................................................................77 Dissolution of Gallium Nitride in Lithium Chloride...........................................77 Gallium Nitride Recrystallization........................................................................78 Summary.....................................................................................................................80 6 SUMMARY AND FUTURE DIRECTIONS.............................................................97 LIST OF REFERENCES.................................................................................................101 BIOGRAPHICAL SKETCH...........................................................................................105 vii

PAGE 8

LIST OF FIGURES Figure page 1-1 Band gap engineering diagram for III-V materials....................................................8 2-1 Schematic of metalorganic chemical vapor deposition operation............................22 3-1 Schematic of reciprocal lattice points probed..........................................................39 3-2 Adjacent domains tilted with respect to one other...................................................40 3-3 Adjacent domains twisted with respect to one another............................................41 3-4 Cross-section transmission electron microscopy image of a low dislocation density sample.......................................................................................................................42 3-5 Cross-section transmission electron microscopy image of a high dislocation density sample.......................................................................................................................43 3-6 Weak beam transmission electron microscopy image and convergent electron beam diffraction image......................................................................................................44 3-7 Broadening effects demonstrated in reciprocal space..............................................45 3-8 Plotted data of full width at half maximum vs. reciprocal lattice vector magnitude46 3-9 Fit of full width at half maximum data vs. angle of inclination...............................47 3-10 Mathematical and geometric derivation of the random distribution method for calculating dislocation density.................................................................................48 3-11 Calculation of dislocation density based on values for coherence length................49 3-12 Comparison of methods of determining dislocation density by XRD.....................50 3-13 Dislocation density determined by transmission electron microscopy vs. dislocation density determined by x-ray diffraction...................................................................51 3-14 Dependence of dislocation density on growth pressure...........................................52 3-15 Effect of dislocation density on room temperature mobility of gallium nitride high electron mobility transistors.....................................................................................53 viii

PAGE 9

4-1 Impurity profiles for low dislocation density at the top...........................................62 4-2 Impurity profiles for mid dislocation density at the top...........................................63 4-3 Impurity profiles for high dislocation density at the top..........................................64 4-4 Impurity profiles for high dislocation density at the flat..........................................65 4-5 Impurity profiles for second series of overgrowths for low dislocation density at the flat.............................................................................................................................66 4-6 Impurity profiles for second series of overgrowths for mid dislocation density at the flat.............................................................................................................................67 4-7 Impurity profiles for second series of overgrowths for high dislocation density at the flat.......................................................................................................................68 4-8 Comparison of impurity profiles for different dislocation density templates..........69 5-1 Gallium nitride crystal with dimensions of 0.9 mm 0.6 mm................................82 5-2 Ideal solubility of gallium nitride.............................................................................83 5-3 Experimentally determined solubility of gallium nitride in lithium chloride compared to the ideal solubility curve.....................................................................84 5-4 Experimentally determined solubility determined after employing a 2 m filter....85 5-5 Scanning electron microscopy image of a bare gallium nitride surface...................86 5-6. Scanning electron microscopy image of a gallium nitride surface exposed to gallium nitride-in-lithium chloride melt...................................................................87 5-7 Close-up scanning electron microscopy image of a gallium nitride crystallite.......88 5-8 Scanning electron microscopy image of a treated gallium nitride surface after exposure to the gallium nitride-in-lithium chloride melt.........................................89 5-9 Scanning electron microscopy image of an untreated gallium nitride surface after exposure to a gallium nitride-in-lithium chloride melt............................................90 5-10 Scanning electron microscopy image focusing on needles on the untreated gallium nitride surface at 750..............................................................................................91 5-11 Scanning electron microscopy image focusing on needles of the untreated gallium nitride surface at 1900............................................................................................92 5-12 Election dispersive spectroscopy images for elemental analysis on needle structures..................................................................................................................93 ix

PAGE 10

5-13 Electron dispersive spectroscopy images of a patch on the gallium nitride surface94 5-14 Scanning electron microscopy image of a deposited gallium nitride crystallite with a thickness of 5 m...................................................................................................95 5-15 Electron dispersive spectroscopy instrument depth calibration...............................96 x

PAGE 11

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECTS OF DISLOCATIONS ON ELECTRONIC PROPERTIES OF III-NITRIDE MATERIALS By Allen M. West May 2005 Chair: Cammy Abernathy Major Department: Materials Science and Engineering By varying metalorganic chemical vapor deposition (MOCVD) growth conditions, III-nitride samples with dislocation densities ranging over two orders of magnitude were grown, and the effects of dislocations were studied. Initially, a novel technique to determine dislocation density by high-resolution x-ray diffraction (HRXRD) was developed. Results were compared to those obtained by transmission electron microscopy (TEM) and agreed within a factor of 1.5. This HRXRD-based technique provided rapid and accurate feedback for use in growth-optimization studies; and became a basis for future studies of the effects of dislocation on specific phenomena associated with III-nitride technology. This included a study of the effect of dislocations on impurity incorporation during MOCVD growth. Dislocations were previously conjectured to provide energetically favorable sites for impurity incorporation, but experimental results did not confirm this. xi

PAGE 12

Finally, a novel growth technique was investigated for gallium nitride (GaN) bulk crystals, based on a simple dissolution/recrystallization process. GaN was successfully dissolved in a molten LiCl medium, and then recrystallized on a nominally homoepitaxial MOCVD-grown GaN template. This provides the basis for a potentially viable method to fabricate high quality GaN bulk crystals for homoepitaxial growth. xii

PAGE 13

CHAPTER 1 INTRODUCTION Background and Motivation Gallium nitride (GaN) and III-nitride alloys (Al-, Ga-, In-N) demonstrate outstanding potential to significantly advance solid-state electronic and optoelectronic technologies. Currently, solid-state electronics comprise a multi-hundred billion-dollar industry, and consist primarily of silicon (Si) and gallium arsenide (GaAs)-based devices. These materials, however, are intrinsically limited for certain applications (namely, high-power electronics and low-wavelength optoelectronics). Such applications have far-reaching implications, with significant promise for commercial, military, and national security purposes. III-nitride-based materials are ideal for these applications because of their favorable intrinsic electronic, structural, and chemical characteristics. Because of these properties, GaN and related alloys have recently been subject to intense research; and scientists have made significant progress in advancing III-nitride technology. Electronic Properties of III-Nitrides Electronically, III-nitrides have a large, direct, tunable band gap, which makes these materials ideally suited for many technologically important electronic and optoelectronic applications. The band gap for III-nitrides ranges from 0.7 eV for indium nitride (InN) [1] to 6.2 eV for aluminum nitride (AlN) [2]. Alloys throughout this entire range are possible to achieve, depending on the composition of group III element (Al, Ga, In). III-nitride materials large band gap gives them distinct capabilities for use in high-power electronic and optoelectronic applications. The tunable nature of the band 1

PAGE 14

2 gap in III-nitrides provides the possibility for large band offsets, which enables greater control over carrier confinement, and better performance in some devices, including HEMTs and multi-quantum well (MQW) LEDs. Chemical, Structural, and Physical Properties of III-Nitrides III-nitrides offer distinct advantages because of their chemical, structural, and physical properties. They are outstandingly hard, corrosion-resistant in most environments, and remarkably stable at high temperatures. III-nitrides exhibit the wurtzite crystal structure, which provides a significant spontaneous polarization field along the c-axis, and a high piezoelectric polarization field [3]. Together, these fields enable greater sheet charge densities in III-nitride heterostructures strained in compression; thus, allowing greater capabilities for III-nitride based electronic devices. III-nitride materials are also significantly more thermally conductive than their III-arsenide counterparts and silicon, allowing for efficient heat removal. This saves substantial resources by reducing the need for external cooling. Silicon-based devices burn out at temperatures exceeding 140C, necessitating the installation of powerful cooling fans for silicon-based devices, for high-power applications. GaN-based devices are operable in environments >300C [3], so the high temperature tolerance and the efficient heat removal of III-nitrides make this material system ideal for high-temperature, high power applications. Suitability of III-Nitrides in Optoelectronic Devices III-nitrides represent one of few known material systems that allow efficient emission of radiation in the green, blue, and deep ultraviolet range. Figure 1-1 shows the band gap dependence on the lattice constant, and the range of achievable wavelengths for

PAGE 15

3 III-nitride emitters. III-nitride LEDs and LDs are exceptionally tolerant of defects, as they show efficient emission of radiation despite high dislocation densities ( = 10 8 11 cm -2 ). Because of the unique capabilities of III-nitride optoelectronic devices, they are ideal for commercial, military, and national-defense applications. Commercially, monochromatic III-nitride LEDs are used for illumination purposes in traffic signals and outdoor displays. One monochromatic traffic light saves 800 kWh/year in power, compared to a typical incandescent white bulb with a monochromatic shutter [6]. III-nitride-based UV VCSELs are also being developed to pump red-green-blue (RGB) phosphors for significantly more efficient and cost-effective white-lighting purposes. This could provide energy savings estimated at $100 billion annually. Furthermore, blue LDs are used commercially low-cost, ultra-high-density optical data storage in digital versatile disks (DVDs) [7] that can store >1 gigabit/cm 2 of information [6]. Militarily, III-nitride photodetectors are currently being researched for use in solar blind detectors for missile warning systems and for malignant biological and chemical agent detection [8,9]. Suitability of III-Nitrides in Electronic Devices Table 1-1 shows critical electronic properties of several semiconductor materials. The intrinsic electronic properties of GaN and AlGaN/GaN heterostructures allow significantly more power output than any other known semiconductor material system. The large, tunable band gap of III-nitrides makes these materials ideal for use in certain electronic devices, such as high electron mobility transistors (HEMTs). Breakdown voltage (V b ) goes as the square of the band-gap energy, giving GaN and AlN-based

PAGE 16

4 electronic devices a distinct advantage over SiC (band gap, E g = 3.2 eV) in terms of power switching, enabling a device to achieve large power density at high frequencies [3,7]. The theoretical maximum power output achievable is directly proportional to the band gap to the fourth power (P max E g 4 ), and the saturation velocity squared (P max v s 2 ). Because of the large band gap and saturation velocity, III-nitride based devices are ideal for high-power applications. For a HEMT device, the output power is shown in Equation 1-1, where I max is the maximum current possible for a particular device. 8maxIVPbout (1-1) Higher saturation currents are achievable with higher carrier concentrations, and the polarization effects of AlGaN/GaN heterostructures provide outstandingly high sheet densities in a two-dimensional electron gas (2DEG). Furthermore, the polarization fields in GaN HEMT structures remove the need for external doping to achieve very high sheet densities. This ability to operate without ionized impurities diminishes the problem of gate leakage in Schottky diodes [5]. Finally, for operable high-power electronic devices, heat dissipation is crucial, since such devices are thermally limited. The high thermal conductivity of GaN enables these HEMT devices to operate more efficiently and more cost-effectively [3]. These characteristics make III-nitrides ideal for use in high-power devices for both commercial and military applications. Commercially, GaN-based HEMTs can be used in cellular communications base stations, telecommunications, hybrid electric cars, and switches for high-power electric grids. For military purposes, GaN-based HEMTs are candidates as amplifiers for synthetic aperture radar (SAR), smart weapons, electronic

PAGE 17

5 warfare, and surveillance. They will eventually replace bulky traveling wave tubes, enabling significantly higher output power, with added benefits of reduced mass, volume, cost, and cooling requirements. Progress in III-Nitride Material Development Performance of III-nitride based devices has improved by leaps and bounds in recent years due to growth and processing of these materials becoming more mature. As growth techniques have progressed, microstructural quality of III-nitride films used in devices has improved; and performance of both electronic [10] and optoelectronic devices has improved accordingly [11]. Optoelectronically, the first nitride-based blue LEDs were fabricated in the 1970s, when the external efficiency hardly exceeded 0.1%. External efficiencies improved with novel growth breakthroughs throughout the 1980s, and reached 1.5% by 1992. As growth of III-nitride films continued to mature, this value reached 20% by 2002 [6]. Further strides with III-nitride optoelectronic devices will be made as microstructural quality of III-nitride films continues to improve. GaN-based electronic devices have experienced similar improvements as III-nitride technology has become more advanced. The earliest reported GaN HEMT state-of-the-art was 1.1 W/mm in 1995, and exceeded 10 W/mm by 2001 [10]. Since 2001, GaN-based technology has progressed rapidly, and the current state-of-the-art GaN HEMT power density is an astounding 32.2 W/mm [12]. Conversely, GaAs-based HEMTs have attained merely 1 W/mm for power density, despite the fact that GaAs-based technology is significantly better studied.

PAGE 18

6 Availability of III-Nitride Based Devices Considering these substantial advantages and the potential widespread use of GaN-based devices, commercial devices based on GaN are currently in the research stage. Comprehensive studies of IIInitride materials are underway, in order to better exploit these materials advantages, where considerable growth and processing issues are being addressed. The most significant of these issues involves the effects of dislocations on III-nitride thin films. Because there are no commercially available substrates that allow for homoepitaxial growth of IIInitride thin films, these devices must be grown heteroepitaxially on sapphire or silicon carbide, at high lattice mismatch. This causes extremely high dislocation densities (=10 8 11 cm -2 ). Statement of the Problem Currently, there is much disagreement concerning the exact effects of dislocations on III-nitride electrical properties and device performance. To advance the science of IIInitrides, and consequently IIInitride based device performance, the precise effects of dislocations on electronic properties must be determined. In other semiconductor material systems (such as Si and GaAs), a dislocation density of ~10 4 cm -2 renders a device inoperable. IIInitride films with dislocation densities of ~10 11 cm -2 have been used in operable devices with surprisingly high efficiency. Because of this high tolerance for defects, some researchers have postulated that dislocations are electrically inactive in III-nitride materials [13]. Others, however, have indicated that dislocations have a deleterious effect on device performance [14], and will significantly affect reliability and scalability of these devices. Determination of the effects of dislocation density on

PAGE 19

7 electronic properties would greatly aid optimization of IIInitride device performance, especially as long as these devices are grown heteroepitaxially. Outline The remainder of this dissertation is structured as follows: Chapter 2 describes the growth and characterization techniques used in this work. Chapter 3 illustrates one technique in depth, based on x-ray diffraction (XRD) to determine dislocation density in IIInitride thin films. This method develops a theoretical model to determine dislocation density from XRD results; and allows for rapid, accurate feedback for dislocation density. This is extremely helpful in optimization studies for metalorganic chemical vapor deposition (MOCVD) IIInitride growth studies. Chapter 4 describes a study on the effect of dislocation density on impurity incorporation in GaN films during MOCVD growth. Impurities can dramatically affect electronic properties of GaN, by forming deep trap states and acting as scattering centers. Understanding how dislocations affect impurity incorporation is critical in optimizing MOCVD growth parameters for device structures. Chapter 5 describes an investigation into a novel method to fabricate IIInitride bulk crystals. Bulk IIInitride substrates would allow homoepitaxial growth of IIInitride thin films, which would minimize dislocation density for device structures. This would greatly enhance performance for devices for which dislocations have a deleterious effect. Finally, Chapter 6 summarizes significant results and recommends future work for these projects.

PAGE 20

8 0.2 0.4 0.3 5.2 4.8 5.0 4.6 6.6 6.4 Energy Gap (eV) Lattice Constant (Angstroms) 6.2 6.0 5.8 5.6 5.4 Wavelength (m ) 0.5 5.0 2.0 1.0 0.0 4.2 4.4 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 Arsenides (near-IR) Antimonides (far/mid-IR) Phosphides (yellow/red) InSb GaSb InAs InP AlAs GaAs GaP AlSb AlN GaN InN Nitrides (blue/UV) Figure 1-1. Band gap engineering diagram for III-V materials. III-nitrides are ideal candidates for use as green, blue, and deep-UV light emitters.

PAGE 21

9 Table 1-1. Electronic properties of semiconductor materials used in electronic device applications. Semiconductor Material Si GaAs SiC GaN Band Gap (eV) 1.1 1.4 3.3 3.4 Saturated Electron Velocity (10 7 cm/s) 1 2.1 2 2.7 Breakdown Field (MV/cm) 0.3 0.4 2 3.3 Thermal Conductivity (W/cm-K) 1.5 0.5 4.5 >1.7 Bulk Electron Mobility (cm 2 /V-s) 1500 8500 700 1200 Heterostructure None AlGaAs/InGaAs None AlGaN/GaN 2DEG Sheet Density (10 12 cm -2 ) NA <4 NA 20

PAGE 22

CHAPTER 2 EXPERIMENTAL METHODS Chapter 1 showed the technological importance of III-nitride-based technology and identified the fundamental problem of high dislocation density. Chapter 2 describes the experimental procedures involved in growth and characterization of III-nitride materials used in these studies. This includes a description of the two different film-growth techniques, specifically metalorganic chemical vapor deposition (MOCVD) and molecular beam epitaxy (MBE), followed by a description of a novel bulk-growth method for GaN. The characterization techniques used are also explained. Growth Techniques Both MOCVD and MBE were used to accomplish different goals for growing films in this work. Growth of III-nitride materials is relatively new, because of the lack of III-nitride lattice-matched substrates and the difficulty associated with incorporation of nitrogen. For MOCVD, ammonia (NH 3 ) is the most prevalent precursor; but its stability makes it difficult to incorporate nitrogen, and requires high growth temperatures (usually ~1050C). For MBE growth of III-nitrides, the nitrogen source is a reactive nitrogen plasma, which consists of nitrogen gas and an electron cyclotron resonance (ECR) plasma source. Metalorganic Chemical Vapor Deposition MOCVD offers the following advantages over MBE: high growth rates, high throughput, ability to grow at high temperatures, external feeds for precursors, high 10

PAGE 23

11 growth pressures, and lower costs associated with this technique over MBE. Higher growth rates decrease the time required to perform material optimization studies, and higher throughput is desirable when fabricating commercial devices, once the material-optimization growth studies are complete. Growth at higher temperatures enables more effective cracking of nitrogen precursor to incorporate nitrogen into the film; which, in turn, results in higher growth rates. The external feed lines that provide the precursors are advantageous, since they allow the growth chamber to be isolated when replacing the sources on depletion. This is a significant advantage over MBE, since replacing the solid sources requires venting the growth chamber to atmosphere. Baking out the impurities from this exposure can take up to 2 weeks, whereas changing group III gas bubblers or ammonia sources for MOCVD usually takes less than 1 hour. The growth pressures associated with MOCVD are relatively high (usually 40 to 500 Torr for III-nitride growth) compared to those of MBE (<10 -6 Torr). This removes the need for ultra high vacuum (UHV) conditions, which require diffusion and cryogenic pumps. Diffusion pumps are problematic because they are susceptible to back-streaming oil into the growth chamber, which may incorporate into the film. Cryogenic pumps require a constant flow of liquid nitrogen (L-N 2 ), which can be expensive, depending on the time and frequency of growth experiments. Finally, for all of the advantages that MBE provides for devices in other material systems, MOCVD-grown III-nitride films for electronic devices have demonstrated superior electronic properties and device performance over their MBE grown counterparts [18]. The MOCVD growths were performed in an EMCORE D-125 short jar reactor with the wafers placed on a silicon carbide-coated graphite rotating platen, containing up

PAGE 24

12 to three wafers. There are several advantages to this type of reactor. First, it allows for a high degree of aluminum incorporation into films, since the short jar design significantly reduces the parasitic pre-reactions of the aluminum precursor with ammonia due to upstream heating. Second, the multi-wafer EMCORE D-125 allows a greater throughput of device-quality films once the growth conditions are optimized. Third, this reactor allows for excellent thermal control and uniformity across the platen, and therefore across the individual wafers. This is especially useful when growing a film that is temperature sensitive. Fourth, the rotation rate of the platen can reach 2100 revolutions per minute (rpm). This allows a high degree of control over the boundary layer thickness, which could improve growth uniformity and growth rate. Growth uniformity can be monitored in-situ, since the EMCORE D-125 contains three view ports where it is possible to monitor the film growth at different points across the wafers on the platen. Finally, this reactor provides a load-lock, directly connected to a turbo-molecular pump (TMP) backed by a dry-scroll rough pump. This allows the reaction chamber to remain under vacuum during sample exchange, which reduces the amount of contamination during film growth. Figure 2-1 is a schematic of the MOCVD operation. The precursors used for the films are trimethyl gallium (TMG) and trimethyl aluminum (TMA) for gallium and aluminum respectively, and ammonia (NH 3 ) for nitrogen, with hydrogen (H 2 ) and nitrogen (N 2 ) serving as carrier gases. The precursors undergo the following reactions: (CH 3 ) 3 Ga (g) + NH 3(g) GaN (s) + 3CH 4(g) (CH 3 ) 3 Al (g) + NH 3(g) AlN (s) + 3CH 4(g)

PAGE 25

13 The solid III-nitride film deposits on the substrate, while the gaseous byproducts are pulled through an EBARA rough pump, and out through the exhaust. In all likelihood, the reactions rarely occur as shown above, as the reactions shown above are oversimplified. Many side reactions and parasitic pre-reactions complicate the MOCVD film deposition process; but overall, the reactions shown above are a useful simplification. The parameters that affect the growth of the film are temperature and pressure, as well as flows of metalorganics, ammonia, and carrier gases. One challenge posed by the MOCVD growth technique is determining which parameter(s) to vary, in order to grow the film with the desired qualities. Since our goal was to determine the effects of dislocations on electronic properties and device performance for III-nitride materials, it was necessary to grow films over a wide range of dislocation densities by MOCVD. The MOCVD growth parameters that were varied include the type of nucleation layer (AlN vs. GaN), growth temperature, and growth pressure, in order to grow films of varying dislocation density. To further expand this study, two different substrates were used: silicon carbide (SiC) and sapphire (Al 2 O 3 ). SiC has a less severe lattice mismatch than sapphire with III-nitrides (3.5% vs. 16% for GaN and 13.29% vs. 1% for AlN) [19]; and therefore, enables experimenters to grow lower-dislocation-density-films. Sapphire is more widely used, because it is significantly less expensive; insulating sapphire wafers are $60 each, while device-grade SiC wafers are $5000 each. For growths on SiC, a 1,400 high-temperature (1125C) AlN nucleation layer was first deposited, followed by a GaN buffer layer at 1050C. The pressure for these growths varied from 100 Torr. For growths on sapphire, more variables were

PAGE 26

14 explored. For some films, a low-temperature (600C) AlN nucleation layer was first deposited; while others had a low temperature (530C) GaN nucleation layer, depending on the desired dislocation density. The pressure of the overlying GaN layers varied from 20 Torr, and the temperature of this layer varied from 1040C. Another factor that affected the film microstructure was recovery time. A fast recovery provided a film with a higher dislocation density, while a longer delay between the growth of the nucleation layer and the GaN over-layer provided a film with a higher microstructural quality [22]. By altering the growth conditions as needed, it was possible to grow GaN films with a very broad range of dislocation densities. The material grown for our study ranged from 4.2110 8 cm -2 1.0710 11 cm -2 This allowed for thorough studies of the effects of microstructure on electronic and magnetic properties, as well as device performance and impurity incorporation. Molecular Beam Epitaxy Molecular beam epitaxy (MBE) is advantageous when abrupt interfaces and precise control of layer thickness are required. MBE is also an inherently safer process when growing III-nitride films, since all precursors used in MOCVD are either toxic or pyrophilic; whereas none of the precursors used in MBE present such hazards. This removes the need for strict engineering safety controls and gas sensors that are required for MOCVD. For our study, MBE offered a distinct advantage when growing GaN films with transition metal dopants and when growing oxide films. MBE is better suited for transition-metal dopant growths, because of the difficulty associated with obtaining an

PAGE 27

15 MOCVD transition-metal source that can incorporate efficiently into a III-nitride film and that has a high enough vapor pressure to provide a substantial molar flux in the carrier gas. MBE is also better for oxide growths, because of the lack of oxygen sources for MOCVD. For the MBE growths in our study, two series of films were grown: chromium (Cr)-doped aluminum nitride AlN and scandium oxide (Sc 2 O 3 ). For the Cr-doped AlN films, the group III precursor was solid 99.99999% pure (7N) aluminum metal, the dopant source was 5N solid Cr, and the nitrogen precursor was 6N nitrogen gas. To create reactive nitrogen radicals, the nitrogen gas was ionized by a radiofrequency (RF) plasma unit with a frequency of 13.56 MHz. For the scandium oxide growths, the precursors were solid 4N scandium metal and 6N oxygen gas, which was ionized at 13.56 MHz. All of the solid sources were evaporated by resistively heating poly-boronitride (PBN) containers, and the atomic fluxes were controlled by adjusting the shutter apertures. Gallium Nitride Bulk Growth The GaN bulk growth studies were performed in a home-built stainless-steel container, fitted for a 1 diameter quartz tube. The method to grow GaN bulk crystals was simply a dissolution/recrystallization process of polycrystalline GaN powder in a molten lithium chloride (LiCl) solvent. The chamber was resistively heated up to ~1000C to effectively dissolve GaN powder and maximize deposition of GaN crystallites. Procedurally, the solubility was tested with respect to temperature. Once it was determined that GaN dissolved sufficiently well at achievable temperatures, the

PAGE 28

16 experiment was repeated to test the recrystallization of GaN on a nominally homoepitaxial template of MOCVD-grown GaN on sapphire substrate. Other effects such as convective fluid dynamics and surface treatment were tested to maximize growth on the surface. Characterization Techniques After the MOCVD, MBE, and bulk growths, samples were removed from their respective growth chambers; and their properties were assessed using several techniques. Specifically, the microstructural, magnetic, and electronic characteristics were determined. All of the techniques used are described in the subsections to follow. High-Resolution X-Ray Diffraction High-Resolution X-Ray Diffraction (HRXRD) is a powerful technique that provides information about a materials microstructure and lattice constant. Our study used only rocking curves (-scans) to ascertain the microstructure of the GaN thin films. The microstructural quality of a thin film is closely related to the width of the diffracted x-ray intensity, with the full-width at half maximum (FWHM) used as the figure of merit. Since HRXRD is a diffraction phenomenon, its measurements must be taken with respect to reciprocal space, where reciprocal lattice points (RLPs) correspond to specific lattice planes. If it were possible to perform HRXRD analysis on a defect-free crystal of infinite dimensions, the results would be a series of delta functions at every RLP. Once finite dimensions and defects are imposed on the crystal, broadening can be observed and measured by HRXRD. The extent to which broadening is observed is indicative of the microstructural quality; the higher the measured FWHM, the higher the dislocation density.

PAGE 29

17 All of the HRXRD measurements were made using a Philips (Almelo, Holland) XPert MRD high-resolution triple-axis x-ray diffractometer with a sealed Cu anode collimated by a four-crystal Ge(220) monochromator. The Cu anode produced an x-ray wavelength of Cu k )=1.5404 Transmission Electron Microscopy Transmission Electron Microscopy (TEM) is a universally accepted method to determine dislocation density in materials. TEM operates by emitting a focused electron beam through a sample, into a series of magnetic lenses, and into a fluorescent screen detector. It is a robust technique for determining microstructural characteristics of materials, since it provides high spacial resolution (with a magnification of 10 6 ). Using this technique, dislocation density is simply computed as the number of dislocations counted over a known measured area. TEM is a destructive technique, since sample preparation requires a thickness that is transparent to the electron beam (<3000 ). The microscopy was performed with a Philips CM20 instrument at 200kV. Some specimens were prepared in plan-view while others were prepared in cross section. The samples observed in cross-section were prepared by a focused ion beam; while those observed in plan-view were prepared by a traditional mechanical polishing, dimple, and ion mill procedure.

PAGE 30

18 Secondary Ion Mass Spectroscopy Secondary Ion Mass Spectroscopy (SIMS) is a superior method to determine impurity concentrations with respect to film depth. It is a destructive technique, but is known to provide accurate results for impurity profiles. SIMS results are obtained by accelerating heavy metal ions at a target sample, sputtering off constituent atoms. These atoms from the target sample become ionized, and are attracted to a charged detector. The SIMS work in our study was performed in a PHI Quadrupole SIMS spectrometer (Charles Evans and Associates, Cambridge, MA). The sputtering ion was cesium (Cs + ), with a primary ion energy of 4keV. The cesium angle of incidence was 60, and the instrument was optimized for sensitivity for carbon, oxygen, and silicon. Detection limits were 110 16 cm -3 for carbon, 510 15 cm -3 for oxygen, and 610 15 cm -3 for silicon. Superconducting Quantum Interference Device Magnetometry A Superconducting Quantum Interference Device Magnetometry (SQUID) Quantum Design Magnetic Properties Measurement System was used to ascertain magnetic properties of DMS materials. Hysteresis loops from magnetization vs. field measurements verify that a sample is ferromagnetic. The second technique to determine magnetic properties was Field-Cooled/Zero Field-Cooled (FC/ZFC) measurement. This technique involves cooling a sample to 10K under an applied field while the samples magnetization is measured (FC), and again while the samples temperature increases to room temperature (ZFC).

PAGE 31

19 Capacitance-Voltage Capacitance-voltage (CV) is a non-destructive technique that provides information about a films electrical properties. In our study, a mercury (Hg) contact CV was used to probe for a buried conductive layer in a nominally insulating GaN film, and to confirm the existence of a conductive 2DEG in a AlGaN/GaN HEMT structure. The depletion region of a HEMT structure is modeled as a parallel plate capacitor, where the mercury droplet acts as a metal Schottky contact, and the sheet density of a 2DEG is given as 123)(rVCAqCxN (2-1) where C is the capacitance, q is the carrier charge, is the permittivity, A is the area of the mercury drop, and V r is the reverse bias voltage. The apparatus used for experimentation was an MSI Electronic Model 412-2L mercury probe, with a droplet diameter of 2.0310 -3 cm -2 Electronic measurements were obtained by a Hewlett Packard 4284A Precision Inductance-Capacitance-Resistance meter, operating at 1 MHz. Inductively Coupled Plasma-Mass Spectroscopy Inductively coupled plasma-mass spectroscopy (ICP-MS) is a powerful technique to determine the existence and concentrations of trace elements. This technique operates by injecting an aerosolized liquid sample into an argon plasma at a temperature of 7000 K. When the atoms from the sample collide with the energetic argon ions, the become charged. The ions are then drawn into a high vacuum chamber with a quadrupole mass analyzer. The quadrupole filters out atoms of undesired masses by alternating

PAGE 32

20 radiofrequency (RF) and direct current (DC) fields, so that only atoms of the desired mass can pass into the detector. The detector is calibrated by correlating the intensity of the signal created by these collisions with the detector to a set of standards with known concentrations. In this study, a Perkin Elmer (Boston, MA) Elan 6100 instrument was used to determine the concentration of gallium and lithium in samples of GaN dissolved in LiCl. Scanning Electron Microscopy Scanning electron microscopy (SEM) is a useful technique to determine morphological information on a surface at extremely high magnifications (10,000). SEM operates by emitting an electron beam at the sample surface in a vacuum. When this electron beam hits the surface, secondary electrons are emitted and collected in a detector, which forms an image, giving precise topographical information. The instrument used was a JEOL (Peabody, MA) JSM5800 tungsten filament scanning electron microscope. The electron beam energies ranged from 5 keV. Electron Dispersive X-Ray Spectroscopy Electron dispersive x-ray spectroscopy (EDS) is used in conjunction with SEM to provide information about elemental analysis on surface features of a sample. Other than emitting secondary electrons when exposed to an incident electron beam, a sample will also emit x-rays. This is because, when inner shell electrons are ejected from an atom, electrons at high energy shells will fill the inner shell vacancy, which emits a photon. The energy of the photon emitted is related to the element. EDS measures the energy and intensity of the emitted radiation, which provides qualitative and quantitative analysis of

PAGE 33

21 surface impurities. The EDS data in this study was obtained using an EDAX Phoenix system, fitted with a 3,000 window, allowing for analysis of low-Z elements. Auger Electron Spectroscopy Auger Electron Spectroscopy (AES) was employed for surface elemental analysis. AES operates by emitting an incident electron beam at a sample, which causes surface atoms to emit electrons of an energy that is characteristic of the atom. This technique operates in the following manner: an incident electron knocks off an inner shell electron on the sample surface, which causes an electron of higher energy to fall into this inner shell. This causes emission of a photon, which is reabsorbed by the surface atom, and causes an outer shell electron to be emitted. It is this emitted electron that is detected, and its characteristic energy is indicative of its atom. The Auger instrument used in this research was a Perkin Elmer 6600, that is capable of providing compositional data for elemental analysis for samples of 1 atomic %.

PAGE 34

22 Cold Source Gases Reactants Reaction Z one Gaseous Products Substrate Heat Flux Figure 2-1. Schematic of metalorganic chemical vapor deposition operation. Cold precursors and carrier gases flow over a heated substrate. Precursor gases undergo a heterogeneous reaction at the substrate surface, depositing a solid film, while gaseous byproducts are pumped out through the system exhaust.

PAGE 35

CHAPTER 3 DETERMINATION OF DISLOCATION DENSITY OF III-NITRIDE FILMS BY XRAY DIFFRACTION Introduction There is currently no universally accepted method that allows for rapid and accurate determination of threading dislocation density in III-nitride thin films. X-ray diffraction is a promising technique to determine dislocation density because it provides results quickly and nondestructively, and when modeled correctly, allows for accurate determination of microstructural parameters of crystalline materials. The goal of this chapter is to develop a method to determine dislocation density based on x-ray diffraction that provides accurate results as quickly as possible. Such a method would facilitate the establishment of a correlation between dislocation density and film properties. This information, in turn, would allow experimenters to determine how dislocations influence electrical and optical properties of device layers. Characterization of Dislocations Several characterization techniques have been employed to ascertain dislocation density in III-nitride films, including transmission electron microscopy (TEM), atomic force microscopy (AFM), cathodoluminescence (CL), and high-resolution x-ray diffraction (HRXRD). TEM is generally accepted as the most accurate and best understood method, but the sample preparation is time consuming and destructive, and some care must be used in how the dislocations are imaged and counted [23]. AFM has also been used, with the dislocations identified as slight depressions near step 23

PAGE 36

24 terminations, and these depressions have been correlated to dislocations with a screw component [24]. Dislocations with a pure edge component are occasionally observed on terraces; however, the large tip radius usually precludes direct observation. Using CL imaging, dark spots are observed on the GaN surface and have been correlated to the dislocation density in TEM samples [25]. In CL, the dislocations that act as non-radiative recombination are imaged as dark spots while dislocations that radiate may have the same intensity as the background GaN. X-ray diffraction provides quick results that are correlated to the microstructure of thin films, and the microstructure is directly related to dislocation density. Because of the limitations associated with TEM, AFM, and CL, HRXRD is an attractive alternative to ascertain dislocation density for III-nitride thin films. Previous Work Researchers began using XRD to study microstructures in metals as early as the 1950s [26]. By comparing the angular distribution of x-ray intensity for a given reflection of cold-worked vs. annealed specimens (i.e., those with high and low dislocation densities, respectively), a correlation emerged between the XRD line width and the dislocation density. Several models were proposed to calculate dislocation density based on the degree of diffracted x-ray scattering. These studies provided a valuable theoretical and experimental foundation for determination of dislocation density by XRD. In the 1990s, these classic works were applied to heteroepitaxial GaN-based thin films [30]. These modern works accounted for line width broadening due to strain, coherence length, rotations of crystallographic domains, and interdependence of these

PAGE 37

25 rotations; however, there were disagreements concerning the relative importance of each of these broadening effects when estimating threading dislocation density. As a result, there is currently no unified analytical model from which to determine dislocation density based on XRD line width results. This work resolves these differences by providing insight into each of the significant broadening mechanisms, concentrating heavily on twist and incoherence. Twist When ascertaining the crystalline quality of III-nitride films by XRD, it is crucial to determine the microstructural twist. Threading edge dislocations (b = 11-20) are often dominant in heteroepitaxial III-nitride films, and these are directly related to crystallographic twist. Recognizing this, Heying et al. [35] explicitly demonstrated the need for asymmetric scans in skew geometry in order to better characterize the dislocation character of GaN films. Several models were subsequently proposed on how to best determine twist in GaN films. Metzger et al.[30] and Kang et al.[31] proposed similar procedures that employ a -scan in order to determine twist. However, it is not possible to obtain accurate results by this method with most XRD systems because the instrumental out-of-plane broadening of a scan is ~1.5, which corresponds to a dislocation density on the order of 10 11 cm -2 thereby grossly overestimating the edge component of dislocation density based on twist. Srikant et al. [32] proposed a model that employs asymmetric rocking curves (scans), and accounts for tilt and twist dependence on both inclination angle () and measured full width at half maximum (FWHM). By treating twist as a free parameter, several scans at different angles of inclination are required for this method, with twist

PAGE 38

26 being determined by optimizing the fit to the FWHM data. This model was also the basis of the work of Chierchia et al. [33] and Sun et al. [34] to determine twist. One significant disadvantage with this method is the necessity of performing so many scans at asymmetric reflections, lengthening the measurement and analysis time. Incoherence Another phenomenon often neglected when determining dislocation density by XRD is the effect of incoherence broadening on the measured FWHM. The model proposed by Srikant et al. only considers tilt and twist contributions to broadening [32]. Chierchia et al. discussed broadening due to incoherence for symmetric scans, but did not explicitly include it when calculating dislocation density [33]. Both Chierchia et al. and Sun et al. [34] stated that their calculations of twist were overestimations because they neglected coherence length broadening upon determining twist [33]. The degree to which incoherence broadening affects the calculation of dislocation density, however, was never described in detail. Focus In this chapter, a theoretical and experimental approach is presented for using HRXRD to measure threading dislocation densities of III-nitride thin films in a rapid, non-destructive manner. This method relates HRXRD line widths to dislocation density, allowing resolution of screw and edge dislocation densities from the total dislocation content. The model developed in this work unifies and simplifies other proposed models and accounts for all significant microstructural causes for broadening. This model was verified by comparing the dislocation density results obtained by HRXRD to those obtained by TEM on III-nitride films grown by metalorganic chemical vapor deposition

PAGE 39

27 (MOCVD) on both SiC and sapphire substrates. The model was further simplified by carefully considering the magnitudes of the contributions of each broadening term, leading to accurate dislocation density determinations from as few as two HRXRD reflections. Experimental Methods To obtain the dislocation density from the HRXRD line widths, the individual broadening effects from tilt, twist, and lateral coherence length were deconvolved from a series of symmetric ((0002), (0004), (0006)) and skew-symmetric ((10-11), (20-22)) rocking curves. These reciprocal lattice points are shown in Figure 3-1. Crystallographic tilt is defined as the out-of-plane rotation of adjacent domains in the crystal, as demonstrated in Figure 3-2, and could be determined directly from the symmetric series of scans. The asymmetric reciprocal lattice points were chosen based on the significant twist component and the relatively high measured intensity associated with them. Twist is defined as the in-plane rotation of adjacent domains (Figure 3-3). Measurements were made using a Philips XPert MRD high-resolution triple-axis x-ray diffractometer with a sealed Cu anode collimated by a four-crystal Ge(220) monochromator. The Cu anode produced an x-ray wavelength of Cu k )=1.5404 Of the ten samples studied by HRXRD, seven were selected for comparison with TEM over the widest range of dislocation density. The microscopy was performed with a Philips CM20 instrument at 200kV. One specimen was prepared in plan view while the other six were prepared in cross section. Three were prepared by focused ion beam, while the balance was prepared by a traditional mechanical polishing, dimple, and ion mill

PAGE 40

28 procedure. Cross-section TEM (XTEM) images of low and high dislocation density samples are shown in Figure 3-4 and Figure 3-5, respectively. To determine dislocation densities of the samples in cross section, dislocations were imaged and counted as usual, under two-beam bright field, dark field, and weak beam (Figure 3-6a) conditions and counted as usual. Specimen thicknesses were determined by convergent beam electron diffraction (CBED) using a (11-20) two-beam condition (Figure 3-6b). The number of parallel diffraction intensity oscillations (fringes) observed in the CBED disks is related to the extinction distance g and the specimen thickness t. Cross sectional thicknesses were assigned based on an average of the first two linear fits of the fringe spacings according to the formula si2nk2+ 1 g2nk2=1 t2 where s i is the fringe spacing of the i th fringe, and n k is an integer. For more detailed information on this technique, see reference 15. The samples were grown by MOCVD in an EMCORE D-125 reactor on both silicon carbide and (0001) sapphire substrates under a variety of growth conditions in order to vary the dislocation densities; the growth temperature for all films was 1050C, and the precursors were trimethylgallium (TMG), trimethylaluminum (TMA), and ammonia (NH 3 ) with hydrogen (H 2 ) and nitrogen (N 2 ) carrier gases. The pressure was varied from 140 300 Torr. All samples analyzed were either GaN or Al x Ga (1-x) N, with aluminum content as high as 45%.

PAGE 41

29 Method There are several challenges posed by employing HRXRD to determine threading dislocation density. These involve the following: Identifying the significant sources of line width broadening. Formulating a model that accounts for these broadening effects. Establishing a procedure using XRD scans to probe the film layers, using the proper geometry. Determining the extent to which microstructural defects influence measured HRXRD line width. calculating dislocation density from this information. Computation of dislocation density is important for comparison studies between materials. Microstructural quality of III-nitride thin films is often reported by quoting the FWHM values of select reflections. However, a simple comparison of line widths for particular XRD scans may be insufficient when comparing crystalline quality of thin films. There are two reasons for this: first, dissimilar XRD systems may exhibit significantly different intrinsic broadening, which must be accounted for. Second, research groups may employ any of a variety of reflections to ascertain crystalline quality of thin films, and these reflections are affected differently by microstructural imperfections. Since there is no one universally employed XRD reflection, calculating dislocation density with a microstructural model is the only way to compare different samples conclusively. Sources of Broadening When performing HRXRD rocking curves on single crystal materials, the potential sources of broadening are due to: Crystallographic rotations. Inhomogeneous strain fields. Curvature of the film. Intrinsic rocking curve width.

PAGE 42

30 Broadening due to the monochromator. For the purposes of this study, broadening effects due to film curvature, intrinsic rocking curve width and the monochromator are neglected, since all of these effects are less than 12 arc sec. [36] Figure 3-7 demonstrates specific broadening effects in reciprocal space and how the diffractometer elucidates each of these broadening effects. Figure 3-7(a) is a view of the film in cross-section with the relevant broadening effects at an asymmetric reciprocal lattice point (RLP) in skew geometry. K hkl is the reciprocal lattice vector between the origin of reciprocal space and the RLP being probed. Its direction is related to the angle between the surface normal of the crystal and the normal of the set of diffracting planes, while its length is related to the d-spacing of the set of planes associated with the RLP. Rotations of crystallographic domains broaden the RLP transverse to K hkl while inhomogeneous strain broadens in the direction parallel to K hkl Broadening due to incoherence occurs both parallel as well as transverse to K hkl Specific broadening effects due to crystallographic rotations and incoherence are demonstrated in Figure 3-7(a). Pure domain tilt and twist are observed about the K y and K z axis, respectively. The effects of tilt and twist measured by the diffractometer are the projections of the rotations about K y and K z respectively, on the RLP. The extent to which each of these rotational broadening contributions affects the measured line width of a rocking curve (-scan) is geometrically related by hkl the angle between K z and the RLP being probed. Figure 3-7(a) also demonstrates broadening due to incoherence, as domain thickness (h) and lateral coherence length (L) are shown explicitly. Because threading dislocations propagate along the growth plane normal during III-nitride growth,

PAGE 43

31 the domain thickness is assumed to be equal to the film thickness. The lateral coherence length is defined as the average distance between threading dislocations. The tilt, twist, and lateral coherence length all affect broadening of an asymmetric RLP when a rocking curve is performed in skew geometry. Since the film thicknesses in this study were 8000 vertical coherence length had an insignificant effect on measured broadening. Furthermore, inhomogeneous strain broadening was neglected because all scans employed in this study were triple axis rocking curves, which diminish the observed broadening effects in the direction parallel to K hkl Therefore, the microstructural imperfections that significantly contributed to our measured values for FWHM were projections of rotations about both K y and K z (tilt and twist) on the probed RLP and lateral incoherence (2/L). Figure 3-7(b) shows broadening effects of the skew-symmetric RLP in plan-view. The measured broadening ( hkl ), is the measured FWHM and broadening due to tilt, twist and incoherence are indicated explicitly. These microstructural broadening effects were included in the proposed model that was used to calculate dislocation density. Model The following reciprocal space model describes the (hkl) dependence of the convolved peak-width contributions from the combined effects of tilt ( tilt ), twist ( twist ) and coherence length (L) broadening: nhklnhkltwistnhkltiltnhklLK||2sincos (3-1)

PAGE 44

32 The geometric dependence of tilt and twist on measured line width is apparent, and the observed broadening due to incoherence is inversely proportional to the magnitude of the reciprocal lattice vector. The exponent, n, denotes the type of convolution of the model. Determination of Convolution One major discrepancy among previous works is the type of convolution, i.e., the value of n is used in the model (Equation 3-1). The choice of n is derived from the bestfit function to the intensity distribution [32]. For example, for a Lorentzian convolution, n=1 and for a Gaussian convolution, n=2. For symmetric scans, both Metzger et al. [30] and Chierchia et al. [33] assumed a Lorentzian convolution, where tilt and symmetric coherence length were determined from a Hall-Williamson plot. The results obtained in this study indicated that the lateral coherence length value was unrealistically high when assuming a Lorentzian convolution, and a more plausible value for coherence length was obtained when a Gaussian convolution was used. For example, a film with a screw dislocation density of 7.710 8 cm -2 measured from TEM gave the following results: assuming a Lorentzian convolution, the lateral coherence length was ~31,000 indicating a dislocation density, ~110 7 cm -2 For the same sample, assuming a Gaussian convolution, the lateral coherence length was 5,000 indicating a dislocation density, ~410 8 cm -2 Similar results were reported by Chierchia et al.[33]. Therefore, a Gaussian convolution (n=2) was used in our analyses with Eq. 1, which is similar to the models of Hordon and Averbach [26] and Ayers [36]. Figure 3-8 shows FWHM data vs. the magnitude of the reciprocal lattice vector (|K| hkl ) for a typical sample assuming a Gaussian convolution. Figure 3-9 shows the dependence of the FWHM on angle of inclination () assuming a Gaussian convolution

PAGE 45

33 for one sample of GaN on sapphire. The Gaussian assumption of the model fits the FWHM vs. data well, and explicitly demonstrates that setting the exponent, n=2, provides accurate results. Calculation of Dislocation Density With the proposed model, two series of scans at different angles were required to deconvolve each individual broadening effect on the experimentally determined FWHM. This deconvolution procedure involves fitting the FWHM data to the model, as shown in Figure 3-8, and is described in detail in the Results and Discussion section. From the experimentally determined values for tilt, twist, and coherence length (both symmetric and asymmetric), two independent methods were used to calculate dislocation density. The first method uses the tilt and twist values with the classic formulation of Dunn and Koch [27], where the density of dislocations, is given by 2 /4.36b 2 Figure 3-9 shows the mathematical and geometrical derivation upon which this method is based. Here, is the calculated rotational broadening effect (tilt or twist) and b is the applicable Burgers vector. With this formula, it is possible to calculate both the screw and edge dislocation density with the deconvolved values for tilt and twist, respectively, using the appropriate Burgers vector. It is important to note that while the coherence length contribution was not used explicitly in calculating dislocation density by this method, its broadening effect was accounted for in Equation 3-1. The second method employed to compute dislocation density was that proposed by Hordon and Averbach [26]. This method assumes that the measured value for coherence length equals the root mean square spacing of randomly spaced threads, where

PAGE 46

34 the dislocation density, =1/L 2 as shown in Figure 3-10. In order to compute the total dislocation density, the value of coherence length was L (l0-ll) the coherence length due to all dislocations (i.e. the coherence length as determined by asymmetric scans in skew geometry, since these are sensitive to both tilt and twist). In order to calculate screw dislocation density, the symmetric coherence length, L (000l) was input for the value of L. The edge dislocation density could be resolved by taking the difference between the total threading dislocation density and screw dislocation density. Simplified Model It is possible to simplify Eq. 1 by neglecting lateral coherence length broadening in order to determine dislocation density more rapidly: 222sincoshkltwisthkltilthkl (3-2) This modification makes it possible to determine the relative tilt and twist of the crystal using only one symmetric scan (e.g., (0004)) and one asymmetric scan (e.g., (10-11)) in skew geometry. Eq. 2 provides an upper bound for rotational broadening effects and a concomitant overestimation of dislocation density. However, Eq. 2 is useful for providing more rapid results to ascertain relative crystalline quality of III-nitride thin films. For example, when characterizing an 8000 thick GaN film with a dislocation density ~10 9 cm -2 this two-scan procedure can provide results within an hour, as opposed to >3 hours for the full five-scan procedure required when incoherence broadening is included in the analysis. This is a considerable improvement when rapid results are desired for growth optimization studies. Comparing dislocation density results determined from the full five-scan procedure to results obtained using this condensed model, the abridged procedure

PAGE 47

35 provided results to within 15%. In all samples studied, the spacing of dislocations was assumed to be random rather than piled-up. A piled-up distribution of dislocations may affect the incoherence broadening term more significantly and must be accounted for, as the value for coherence length may decrease significantly for such a distribution. Furthermore, when the dislocations are piled up, the calculation of dislocation density must be performed using a different method, the details of which are described in Reference 8. Results and Discussion For each sample, the individual broadening effects were deconvolved using Equation 3-1 with the measured FWHM data. Figure 3-8 demonstrates a fit of the model to the FWHM data. From the fit of the model to the data, the magnitude of each broadening contribution was determined. The three symmetric scans were employed to determine tilt and symmetric coherence length, as twist does not affect these data. Once this value for tilt was established, twist and asymmetric coherence length were next determined by fitting the model to the FWHM values of the skew-symmetric reflections. It is important to note that coherence length for a symmetric family of planes may be significantly different than that of an asymmetric family of planes. Two independent methods were used to calculate dislocation density using these values for tilt, twist, and coherence length (both symmetric and asymmetric). These included the formulation of Dunn and Koch [27] ( 2 /4.36b 2 ) and that proposed by Hordon and Averbach [26] (=1/L 2 ). Figure 3-12 shows the agreement between the two methods within a factor of 3. However, the formulation of Dunn and Koch was more reliable when compared to TEM results. This was expected since the coherence length

PAGE 48

36 broadening effect was slight compared to those of crystallographic rotations, and therefore, had a greater intrinsic error. Figure 3-13 shows dislocation densities determined by HRXRD, using the formulation of Dunn and Koch compared to those determined by TEM. The values for dislocation density ranged from 5.410 8 cm -2 to 7.910 9 cm -2 All of the values for dislocation density determined by HRXRD were obtained using Eq. 1 and the full fivescan procedure, for which incoherence broadening was accounted. The values for dislocation density determined by HRXRD agreed with those determined by TEM within a factor of 1.5 for all seven samples. With confidence that the XRD-based method described in this chapter is accurate with respect to TEM, it served as a basis for all subsequent growth optimization studies described throughout this dissertation with respect to dislocation density. The first such optimization study observed the effect of growth pressure on dislocation density. Figure 3-14 demonstrates these results, and exhibits that increasing the growth pressure causes a decrease in dislocation density. Performing this study on so many samples using TEM would be impractical because of the time and cost associated with TEM sample preparation. Such studies can only be performed by a non-destructive, rapid technique such as XRD. Ultimately, III-nitride researchers seek to conclusively determine the effect of dislocations on device performance. Using the XRD-based method described above to determine dislocation density, the effect of dislocations on HEMT performance was ascertained with respect to electronic properties. For the purposes of this experiment, the HEMT performance assumed to be proportional to the sheet densitymobility product (n s

PAGE 49

37 ), widely used as the figure of merit for GaN HEMT device performance [38]. In a series of experiments, electronic properties obtained by Hall measurements were compared to dislocation density. These results are shown in Figure 3-15. While dislocations appear to inhibit room temperature mobility for a given range of sheet density, no trend could be concluded. The reason is that the three samples of highest dislocation density were grown early in the HEMT growth optimization process, and other room temperature scattering mechanisms are believed to have had a stronger effect on these samples compared to those that exhibit higher mobility. These scattering mechanisms include interface roughness and alloy disorder. As the HEMT growths were optimized, higher quality interfaces with fewer variations in alloy disorder from the AlGaN barrier layer were most likely the cause for the improved mobility. Further study is necessary to conclusively determine the precise effects of dislocations on room temperature mobility. Summary This chapter describes a method to determine dislocation density from measured HRXRD line widths. A geometrically derived reciprocal space-based model was developed that allows for the determination of individual microstructural broadening effects in III-nitride thin films. This model unifies previous works that employ x-ray diffraction to calculate dislocation density, but demonstrates distinct advantages in terms of simplicity, accuracy, and time efficiency. Realizing the microstructural tilt, twist, and coherence length, dislocation density was calculated by two separate, independent methods, which agreed within a factor of 3. The model was simplified to neglect broadening due to incoherence, which provided an upper bound for dislocation density

PAGE 50

38 from only two scans. Finally, the validity of this model was tested by comparing dislocation density results to those obtained by TEM, and agreement within a factor of 1.5 was found. These results show that the HRXRD-based method to determine dislocation density presented here is accurate, as demonstrated by the agreement of our XRD-based values for dislocation density with TEM, and time-efficient, as results were obtained within an hour. Thus, this HRXRD-based method to determine dislocation density may provide researchers with a valuable tool for optimizing growth conditions for heteroepitaxial III-nitride films. The method demonstrated in this chapter serves as a basis for the following chapters of this dissertation, since rapid and accurate feedback are required to perform intensive surveys on the effects of dislocations on electronic properties and device performance of GaN thin films.

PAGE 51

39 K (l0l) (0002) (0004) (0006) (10-11) (20-22) K II Figure 3-1. Schematic of reciprocal lattice points probed. Procedurally, the reciprocal lattice points probed were symmetric (000l) and asymmetric (l0-ll). Together, these data could be used to determine the tilt, twist, and coherence length in the crystal.

PAGE 52

40 Figure 3-2. Adjacent domains tilted with respect to one other. The axis of rotation is orthogonal to the c-axis.

PAGE 53

41 Figure 3-3. Adjacent domains twisted with respect to one another. The axis of rotation is orthogonal to the plane of the domain, parallel to the c-axis.

PAGE 54

42 g 11-2-2 0.5 m Figure 3-4. Cross-section transmission electron microscopy image of a low dislocation density sample. The dislocation density for this sample is 2.110 9 cm -2

PAGE 55

43 g 11-2-2 0.5 m Figure 3-5. Cross-section transmission electron microscopy image of a high dislocation density sample. The calculated dislocation density for this sample is 6.810 9 cm -2

PAGE 56

44 A B Figure 3-6. Weak beam transmission electron microscopy image and convergent electron beam diffraction image. A) A sample weak beam TEM image of a film with a dislocation density of 2.5910 9 cm -2 B) Convergent beam electron diffraction image showing the oscillations in diffracted intensity. Disks are the (000) and (11-20) reflections.

PAGE 57

45 Figure 3-7. Broadening effects demonstrated in reciprocal space. In order to accurately determine twist, asymmetric rocking curves in skew geometry are necessary. In (a), the film is shown in cross section. The crystallographic rotations affect measured broadening proportional to their projections of their respective axes onto the reciprocal lattice point being probed. Broadening due to incoherence is proportional to 2/L. In (b), the measured broadening effects are shown in plan view. These include tilt, twist, and incoherence.

PAGE 58

46 0.1700.2100.2500.2902468|K| (1/)FWHM (degrees) Symmetric data Skew-symmetric data Figure 3-8. Plotted data of full width at half maximum vs. reciprocal lattice vector magnitude. As |K| hkl increases, incoherence broadening diminishes. The curvature of the plot as |K| hkl approaches 0 indicates the coherence length contribution on the measured FWHM.

PAGE 59

47 0.050.070.090.110.130.150102030405060708090Chi (degrees)FWHM (degrees) Figure 3-9. Fit of full width at half maximum data vs. angle of inclination. The data shown are for one sample of GaN on sapphire at every observable reciprocal lattice point. The fit of the data assumes a Gaussian convolution without accounting for broadening due to incoherence. The data fit well, and demonstrate that a Gaussian convolution is acceptable. Most of the data lie above the fit due to the neglecting of incoherence broadening.

PAGE 60

48 L ||2ln2b Figure 3-10. Mathematical and geometric derivation of the random distribution method for calculating dislocation density. The coherence length, L, and the rotational broadening effect, are indicated explicitly. The opposite side of this right triangle was derived to be a constant multiplied by the appropriate Burgers vector.

PAGE 61

49 L L L L Figure 3-11. Calculation of dislocation density based on values for coherence length. Since coherence length is defined as the average distance between dislocations, dislocation density is simply calculated as =1/L 2

PAGE 62

50 8.79.19.59.98.79.19.59.9Log Dislocation Density (Dunn/Koch)Log Dislocation Density (Hordon/Averbach) Figure 3-12. Comparison of methods of determining dislocation density by XRD. The agreement shown here demonstrates internal consistency from the XRD-based methods to determine dislocation density.

PAGE 63

51 0.0E+003.0E+096.0E+099.0E+090.0E+003.0E+096.0E+099.0E+09XRD Dislocation Density (cm-2)TEM Dislocation Density (cm-2) Figure 3-13. Dislocation density determined by transmission electron microscopy vs. dislocation density determined by x-ray diffraction. All values determined by XRD agreed with TEM within a factor of 1.5.

PAGE 64

52 1.0E+091.5E+092.0E+092.5E+093.0E+0950100150200Pressure (Torr)Dislocation Density (cm-2) GaN 150torr GaN 100torr Figure 3-14. Dependence of dislocation density on growth pressure. Higher growth pressure causes a lower dislocation density. Dislocation density results were obtained by x-ray diffraction.

PAGE 65

53 04008001200160020000.00E+002.00E+094.00E+096.00E+09Dislocation Density (cm-2)Mobility (cm2/V-s) ns = 5.2 E+12 ns = 9 E+12 ns = 2.2 E+13 ns = 1.3 E+13 Figure 3-15. Effect of dislocation density on room temperature mobility of gallium nitride high electron mobility transistors. No clear effect is evident.

PAGE 66

CHAPTER 4 EFFECT OF DISLOCATIONS ON IMPURITY INCORPORATION IN METAL ORGANIC CHEMICAL VAPOR DEPOSITION GROWTH Introduction Aside from high dislocation density, MOCVD grown GaN and III-nitride films contain extremely high concentrations of impurities. Due to the harsh MOCVD conditions required for growth of III-nitride materials, typical GaN films have impurity concentrations 10 16 cm -3 and sometimes as high as 10 20 cm -3 [39]. These impurities have a profound effect on electronic properties of III-nitride thin films and make it difficult to optimize MOCVD growth conditions. Impurities have also been implicated in providing mid-gap states and inhibiting device performance. The most prominent of these impurities are silicon, oxygen, and carbon. Silicon Silicon is a common impurity in GaN and III-nitride thin films. It is used as an ntype dopant, as it acts as a shallow donor in a III-nitride lattice when it is a substitutional impurity on a gallium lattice site (Si Ga ). The energy differential of the silicon donor state to the conduction band minimum (CBM) has been widely reported, although there is some disagreement. Researchers have reported values of 22 meV [40], 30.18 meV [41], 30.8 meV [42], 31.7 meV [43], and 42 meV [44]. Despite the variation in these reported results, all of these values are indicative of a shallow donor in GaN and AlGaN films, and is the dopant of choice when n-type conductivity is desired. The 54

PAGE 67

55 source of silicon in unintentionally doped (UID) III-nitride thin films is widely believed to be from decomposition of silicon carbide (SiC) substrates [45] and susceptors. Oxygen Oxygen is another impurity that has been implicated as an n-type dopant when it incorporates substitutionally on a nitrogen lattice site (O N ). Many researchers have implicated oxygen as being responsible for the significant background n-type conductivity of UID GaN films. Several research groups have reported that O N donor level states are between 32-34 meV from the CBM [40-42] while other groups have reported lower values. Joshkin et al. reported a value of 23.5 meV [46], while others have reported values between 2-10 meV [44,47]. Despite the variance in these values, oxygen is widely considered a shallow donor. Many researchers believe that oxygen impurities act as n-type dopants, simply because of its extra valence electron relative to nitrogen, while others have proposed a more complicated mechanism for oxygen providing the background n-type conductivity of UID GaN. For example, Oila et al. have proposed that oxygen impurities promote the existence of gallium vacancies (V Ga ), as oxygen-gallium vacancy complexes (O N -V Ga ) are energetically stable in the GaN lattice. According to Oila et al., it is the gallium vacancies that provide the background n-type conductivity [48]. Overall, however, many researchers attribute the background conductivity of GaN to the presence of oxygen in the lattice, despite the disagreements in the proposed mechanisms. The sources of oxygen include residual oxygen gas in the ammonia source [47] and atmospheric contamination of the growth chamber. Aside from these sources of contamination, the sapphire (Al 2 O 3 ) substrate [45] provides a significant source of

PAGE 68

56 oxygen. When sapphire is subjected to the high temperatures employed during MOCVD growth of in a hydrogen ambient, the sapphire surface is etched, and the oxygen atoms incorporate into the GaN film. Performing electrical measurements, researchers have found that the GaN film is significantly n-type within ~0.4 m of the sapphire interface, due to a buried conductive layer (BCL) [49]. This BCL is undesirable when attempting to grow semi-insulating GaN films, and researchers have attempted to compensate these residual carriers by intentionally incorporating deep acceptors, like carbon and iron [49]. Carbon Carbon is another common impurity in MOCVD grown GaN films. Carbon is an amphoteric dopant, as its electronic properties depend on whether it incorporates substitutionally on a gallium site or on the nitrogen site [54,55]. If carbon incorporates on a gallium site, it acts as a shallow donor, with an ionization energy of ~0.2 eV. Conversely, if carbon incorporates on a nitrogen site, it acts as a shallow acceptor with values of ionization energy reported as ~0.3 eV [55] and 34 meV [42]. If carbon incorporates interstitially, its energy level is mid-gap, and its donor/acceptor properties depend on the Fermi level. If carbon incorporates interstitially into n-type GaN, it will act as a deep acceptor, but if carbon incorporates into p-type GaN, its energy state will be indicative of a deep donor [55]. Carbon is generally considered a deep acceptor, and when GaN films must be insulating, MOCVD researchers often intentionally adjust growth conditions to favor carbon incorporation in order to compensate the background n-type conductivity of UID GaN. However, growers try to minimize carbon

PAGE 69

57 incorporation near active regions, as it acts as a scattering center, and has been implicated in gate leakage and dispersion in GaN-based field effect transistors (FETs). The sources of carbon in MOCVD growth of III-nitrides are the metalorganic group III precursors, and, to a lesser extent, SiC substrates and susceptors. Carbon is a common impurity in any MOCVD growth process, and its incorporation can dramatically affect the films properties, but growth conditions can be adjusted to control the extent to which carbon incorporates. Previous Work Growth experimenters have demonstrated that MOCVD growth conditions have a significant effect on impurity incorporation in GaN films. Specifically, the effects of temperature, pressure, and flows of TMG, ammonia, and hydrogen on silicon and carbon incorporation were studied in-depth [39]. A correlation between resistivity, dislocation density, and carbon concentration in GaN films has also been established, indicating that threading edge dislocations and/or carbon act as compensating centers [56]. Furthermore, threading edge dislocations have been shown to provide energetically favorable sites for impurity incorporation, due to the miscoordinated atoms along extended line defects [57,58] in other material systems. Studies on growth of III-nitrides have only investigated the effects of growth conditions on impurity incorporation on templates with nominally identical microstructures. What is lacking to date is the effect of dislocations on impurity incorporation in MOCVD-grown III-nitride films [39,56]. Both Koleske et al. and Wickenden et al. have indicated that threading edge dislocations facilitate carbon incorporation, but conclusive studies to confirm this have not been performed.

PAGE 70

58 Focus The experimentation in this chapter aims to determine the effects of dislocation density on impurity incorporation during MOCVD growth. Specifically, silicon, oxygen, and carbon incorporation are observed with respect to growth conditions on templates of variable dislocation densities. By subjecting templates with vastly different dislocation densities to identical growth conditions, the effect of dislocation density on impurity incorporation could be determined. Experimental Methods Templates were grown by MOCVD on sapphire substrates, and dislocation density was intentionally altered to obtain two series of GaN films with the largest range of dislocation density possible. Specifically, the growth conditions that were altered were growth pressure, nucleation layer, and recovery time. Template dislocation densities ranged from 4.5910 8 cm -2 8.7610 10 cm -2 for the first growth series, as shown in Table 4-1, and 4.4610 8 cm -2 7.610 10 cm -2 for the second growth series, as shown in Table 4-2. As expected, conditions that minimized dislocation density were higher growth pressures, GaN nucleation layers, and delayed recoveries. Higher growth pressures cause larger grain sizes during the initial stages of high temperature growth, which cause fewer dislocations upon coalescence [59]. GaN nucleation layers provide lattice-matched templates upon which to grow high temperature GaN. Delayed recovery is believed to cause dislocations to propagate laterally rather than thread along the c-axis to the surface. The characterization performed in this experimentation included XRD to determine dislocation density, using the method described in Chapter 3, and SIMS to determine impurity incorporation. Two series of growth runs were performed on

PAGE 71

59 templates of dislocation density varying over two orders of magnitude, while the only varied growth parameter was pressure in the overgrowth. For both overgrowth runs, variable dislocation density templates were grown at 1050C, while pressure was varied from 20 Torr, and silicon was incorporated as a marker when changing pressure. Impurity incorporation was measured for all three samples from both runs. Results and Discussion The first series of SIMS data is shown in Figures 4-1 to 4-4. In all samples, oxygen and silicon were at the detection limit of the instrument. The carbon profiles changed in all cases, presumably due to changes in growth pressure, but this could not be confirmed due to the fact that the silane did not incorporate. Another observation was that the radial position of the wafer had a distinct effect on impurity incorporation. The profile shown in Figure 4-3 was taken from the same wafer as that of Figure 4-4, but the SIMS results from Figure 4-3 were obtained near the top of the wafer, while those from Figure 4-4 were obtained near the flat. This demonstrates that impurity profiles are more pronounced near the flat part of the wafer. From the results of the first series of runs, two adjustments were made: 1) all SIMS analysis was performed near the flat part of the wafer to obtain more distinct profiles, and 2) the silane source was confirmed to be online by Hall measurements of previous samples to ensure intentionally doped n-type conductivity. The SIMS results for this second series of runs are shown in Figures 4-5 to 4-7. Once again, oxygen impurities were below detection limit, which demonstrates that oxygen does not incorporate significantly into GaN films grown in this reactor. Silicon was also at the detection limit, except when it was intentionally incorporated to indicate changes in

PAGE 72

60 pressure. Carbon profiles, however, changed significantly with changes in pressure for all samples. As expected, carbon incorporated inversely with pressure; the maximum carbon incorporation occurred when the growth pressure was 20 Torr, and the minimum incorporation occurred at 500 Torr for all samples. In order to determine the effect of dislocation density on carbon incorporation, average values of carbon concentration were determined for each sample at each pressure and compared. These results are shown in Figure 4-8, and show that there is no clear trend confirming that dislocation density promotes carbon incorporation. For all growth pressures, the template with a dislocation density of 5.5910 9 cm -2 had the highest carbon concentration. Although dislocation density did not have an observable effect on impurity incorporation, pressure had a significant effect. Figure 4-8 indicates that carbon incorporation is, to some extent, a pressure-dependent process. At higher pressures, there are greater molecular interactions between carbon and hydride gases, which promote carbon removal from the film during MOCVD growth. Threading dislocations, however, do not appear to provide significantly energetically favorable sites for impurity incorporation. In order to confirm the finding that dislocations do not promote incorporation of impurities (specifically carbon) during MOCVD growth, two series of experiments are proposed. First, a similar series of experiments could be performed varying parameter of growth temperature. It is possible that the carbon removal mechanism at high pressures is significantly different than that of higher temperature, so dislocations may have an observable effect on carbon incorporation by altering temperature. Second, a similar series of experiments could be performed with broader range of dislocation density

PAGE 73

61 templates. This could be achieved by growing templates employing unconventional methods to minimize dislocation density, including epitaxial lateral overgrowth (ELO) or cantilever epitaxy (CE). These techniques essentially filter dislocations and have been used to grow GaN films with dislocation densities below 10 7 cm -2 This would provide a broader range of dislocation densities for the templates for the purpose of this experimentation. Summary The effect of threading dislocation density on impurity incorporation was investigated in this chapter. Oxygen and silicon were at the SIMS detection limit in all cases (except when silane was intentionally flowed into the growth chamber to indicate pressure changes), indicating that an insignificant amount of oxygen and silicon incorporated into the GaN films. Carbon incorporation depended heavily on pressure, and higher growth pressures caused significantly less carbon to incorporate into the films. Dislocation density, however, had no observable effect on carbon incorporation. This result indicates that threading dislocations do not provide significantly energetically favorable sites for impurity incorporation in MOCVD grown GaN films. The carbon incorporation in this experiment was pressure-driven by a mechanism dependent on molecular interactions.

PAGE 74

62 1E+141E+151E+161E+171E+181E+191E+201E+210246810DEPTH (microns)CONCENTRATION (atoms/cc)1E+001E+011E+021E+031E+041E+051E+061E+071E+08Counts Per Second Ga-> CO Si Figure 4-1. Impurity profiles for low dislocation density at the top. The impurities under study were silicon, oxygen, and carbon for the first series of overgrowth studies. The dislocation density of the template is 4.5910 8 cm -2 The SIMS sample was taken from the top part of the wafer.

PAGE 75

63 1E+141E+151E+161E+171E+181E+191E+201E+210246810DEPTH (microns)CONCENTRATION (atoms/cc)1E+001E+011E+021E+031E+041E+051E+061E+071E+08Counts Per Second OGa-> C Si Figure 4-2. Impurity profiles for mid dislocation density at the top. The impurities under study were silicon, oxygen, and carbon for the first series of overgrowth studies. The dislocation density of the template is 6.6610 9 cm -2 The SIMS sample was taken from the top part of the wafer.

PAGE 76

64 1E+141E+151E+161E+171E+181E+191E+201E+210246810DEPTH (microns)CONCENTRATION (atoms/cc)1E+001E+011E+021E+031E+041E+051E+061E+071E+08Counts Per Second OGa-> C Si Figure 4-3. Impurity profiles for high dislocation density at the top. The impurities under study were silicon, oxygen, and carbon for the first series of overgrowth studies. The dislocation density of the template is 8.7610 10 cm -2 The SIMS sample was taken from the top part of the wafer.

PAGE 77

65 1E+141E+151E+161E+171E+181E+191E+201E+210246810DEPTH (microns)CONCENTRATION (atoms/cc)1E+001E+011E+021E+031E+041E+051E+061E+071E+08Counts Per Second OGa-> C Si Figure 4-4. Impurity profiles for high dislocation density at the flat. The impurities under study were silicon, oxygen, and carbon for the first series of overgrowth studies. The dislocation density of the template is 8.7610 10 cm -2 The SIMS sample was taken from the flat part of the wafer.

PAGE 78

66 Impurity Depth Profile, = 4.5x108 cm-21.00E+151.00E+161.00E+171.00E+181.00E+1902468Depth (m)Concentration (cm-3) Carbon Silicon Oxygen Figure 4-5. Impurity profiles for second series of overgrowths for low dislocation density at the flat. The impurities under study were silicon, oxygen, and carbon. The dislocation density of the template is 4.4610 8 cm -2

PAGE 79

67 Impurity Depth Profile, = 5.6x109 cm-2 1.00E+151.00E+161.00E+171.00E+181.00E+19012345Depth (m)Concentration (cm-3) Carbon Silicon Oxygen Figure 4-6. Impurity profiles for second series of overgrowths for mid dislocation density at the flat. The impurities under study were silicon, oxygen, and carbon. The dislocation density of the template is 5.5910 9 cm -2

PAGE 80

68 Impurity Depth Profile, = 7.6x1010 cm-21.00E+151.00E+161.00E+171.00E+181.00E+19012345Depth (m)Concentration (cm-3) 6 Carbon Silicon Oxygen Figure 4-7. Impurity profiles for second series of overgrowths for high dislocation density at the flat. The impurities under study were silicon, oxygen, and carbon. The dislocation density of the template is 7.610 10 cm -2

PAGE 81

69 Carbon Incorporation vs. Pressure1.00E+161.00E+171.00E+181.00E+190100200300400500600Pressure (Torr)Carbon Concentration (cm-3) High dislocation density template Low dislocation density template Mid dislocation density template Figure 4-8. Comparison of impurity profiles for different dislocation density templates. Impurity profiles for silicon, oxygen, and carbon for variable dislocation density templates during MOCVD growth with variable pressure are shown. The low dislocation density template is 4.4610 8 cm -2 the mid dislocation density template is 5.5910 9 cm -2 and the high dislocation density template is 7.610 10 cm -2 The SIMS samples were all taken from the flat part of the wafer.

PAGE 82

70 Table 4-1. Growth conditions for the templates for the first series of SIMS results. Conditions that promote lower dislocation density are higher pressure, delayed recovery, and GaN nucleation layers. Dislocation densities were intentionally altered to provide templates with as large of a range for dislocation density possible. Sample Dislocation density (cm -2 ) Growth pressure (Torr) Nucleation Layer Recovery Time 1 4.5910 8 500 GaN Delayed 2 6.6610 9 50 GaN Rapid 3 8.7610 10 70 AlN Rapid

PAGE 83

71 Table 4-2. Growth conditions for the templates for the second series of SIMS results. Dislocation densities were intentionally altered to provide templates with as large of a range for dislocation density possible. Sample Dislocation density (cm -2 ) Growth pressure (Torr) Nucleation Layer Recovery Time 1 4.4610 8 500 GaN Delayed 2 5.5910 9 50 GaN Rapid 3 7.610 10 70 AlN Rapid

PAGE 84

CHAPTER 5 GALLIUM NITRIDE BULK CRYSTAL GROWTH BY DISSOLUTION AND RECRYSTALLIZATION OF GALLIUM NITRIDE POWDER Introduction Poor material quality is widely considered to be the most significant factor limiting GaN-based device performance, and this is primarily due to a lack of availability of lattice-matched substrates. Ideally, GaN film growth would be performed on single crystal GaN bulk substrates, but the intrinsic properties of this material makes fabrication of bulk crystals difficult. In other semiconductor material systems, such as silicon and gallium arsenide, bulk crystals have been fabricated from melts, where boules of high crystalline quality material have been achieved with dimensions as large as 12-inches 12 feet. Similar methods are not feasible for GaN because the GaN dissociates prior to melting at atmospheric pressure. Because of this, no commercially viable bulk growth method for GaN has been developed to date, despite the potential benefits of low defect density GaN bulk crystals. Previous Work Researchers have demonstrated several methods to successfully fabricate GaN bulk crystals. The two most prominent of these are an ultra high nitrogen pressure method, and an ammonothermal technique. Both approaches offer a distinct set of advantages and disadvantages relative to the other. The ultra high nitrogen pressure technique entails subjecting a pool of liquid gallium metal to extremely harsh conditions in a nitrogen gas (N 2 ) environment [60], with 72

PAGE 85

73 temperatures of ~1600C and nitrogen overpressures of ~45,000 atm. GaN crystals are formed by dissolving the nitrogen gas into the gallium liquid, where N 2 dissociates into atomic nitrogen, and bonds with the gallium. Using this method, researchers have successfully fabricated crystals of GaN that are 1 cm 2 100 m, with extremely low dislocation densities (~100 cm -2 ). The drawbacks to using this approach include slow growth kinetics, high impurity concentrations, the inability to grow boules of material, and the high pressure and high temperature requirements. The GaN crystals are grown as a crust on the liquid gallium surface, and a 1 cm 2 100 m crystal is formed after ~1 month under the extreme conditions required. Furthermore, these crystals contain ~10 18 cm -3 oxygen impurities, which may have an undesirable effect on the electronic properties of the GaN crystal. Finally, the temperature and pressure requirements add significant cost to this method. The ammonthermal technique has been demonstrated to successfully grow GaN bulk crystals by dissolving GaN feedstock into liquid ammonia, and precipitating single crystal GaN upon supersaturation [61]. This method demonstrates significant improvement over the high nitrogen pressure-driven process described above with respect to growth rate (~0.5 mm/week). Furthermore, large area boules are possible to extract from a seed crystal. The disadvantages of this method include higher dislocation densities (~10 6 cm -2 ) and high impurity incorporation (~10 17-18 cm -3 ). Moreover, this process also requires harsh conditions, with temperatures of ~550C and pressures of ~4000 atm in order to dissolve GaN into a liquid ammonia medium. Although both methods of GaN bulk growth described above were significant breakthroughs, neither are commercially viable processes to fabricate GaN crystals. This

PAGE 86

74 is primarily due to the high-pressure requirement, which prevents both processes from being scalable and manufacturable. If a method to fabricate GaN bulk crystals at atmospheric pressure were possible, this would provide a distinct advantage over other methods currently employed. Proposed Methods In order to develop an atmospheric pressure process to form GaN crystals, three series of experiments have been proposed. The first involves electroplating of GaN using a gallium metal electrode and a reactive nitride ion (N 3) from a lithium nitride (Li 3 N) precursor. A second possible technique comprises an electrochemical reduction of nitrogen gas (N 2 ) to form nitride ions (N 3) that react with gallium to form GaN. A third potential method is a simple dissolution/recrystallization process of GaN powder in a liquid medium, where the formation of GaN single crystal would be driven by a thermal gradient. Electroplating Electroplating of GaN from gallium and nitride precursors could prove to be a promising method to form high quality GaN crystals. The difficulty, however, was determining a suitable host environment in which a nitride ion could exist without reacting quickly and explosively. In an unrelated series of experiments, Goto et al. determined that a molten alkali-halide host environment could successfully dissolve stable nitride ions [62]. Using this, the possibility of forming GaN crystals from gallium and nitride precursors was investigated. Figure 5-1 shows a GaN crystal of macroscopic dimensions (0.9 mm 0.6 mm) that was successfully synthesized using this method.

PAGE 87

75 This result was encouraging, but the electroplating method suffered from four significant drawbacks: Electroplating does not provide material with optimal crystalline quality. The electroplating process would be very difficult to scale, in order to develop a controllable, manufacturable process. The Li 3 N precursor is extremely expensive, making this process less cost-effective. The buildup of Li + ions with GaN deposition complicates the electrochemistry [63]. In order to make this process more cost effective and controllable, the electrochemical reduction of nitrogen gas to nitride ions was investigated (1/2 N 2 + 3 e N 3). Electrochemical Reduction of Nitrogen Gas Goto and Ito investigated the possibility of electrochemically reducing nitrogen gas into nitride ions in a molten alkali-halide liquid medium [63]. Using this result, the possibility of fabricating GaN crystals from electrochemically oxidized gallium with electrochemically reduced nitrogen gas was investigated, and GaN crystals were successfully formed. Using this method, bulk GaN crystals could possibly deposit on a seed crystal, and be extracted as a boule. While this process offers several advantages, the major disadvantage is the difficulty of controlling the fluid dynamics to form the crystals. Because nitride ions are so reactive, they must be separated from gallium ions until they reach the desired growth surface, which would be difficult to model. Dissolution and Recrystallization of Gallium Nitride Crystals The ideal method to grow bulk GaN crystals would be based on a simple dissolution/recrystallization process of GaN in a liquid medium. Using such a method, it would be possible to control the rate of crystal growth by controlling the temperature

PAGE 88

76 gradient and the fluid mechanics in the growth vessel. This process could eventually be used to form large GaN crystals, that could be extracted as a boule from the melt. Focus In this chapter, a novel bulk growth method of GaN crystals is introduced, based on a simple dissolution/recrystallization process in a molten alkali-halide salt. If GaN can dissolve sufficiently in such a medium, then researchers could develop a temperaturedependent, atmospheric pressure process to fabricate GaN bulk crystals. Such a method could prove to be a scalable and manufacturable process that could lead to production of commercially available GaN bulk crystals and wafers for homoepitaxial GaN growth. Specifically, this chapter examines whether a molten halide medium is capable of dissolving GaN, and if so, whether it is possible to recrystallize GaN on a nominally homoepitaxial GaN template. Experimental Methods All of the experimentation in this chapter was performed in a home-built stainless steel heater vessel fitted for a 1 quartz tube in a glove box with a nitrogen gas ambient. To determine the temperature-dependent solubility of GaN, aliquots of GaN in molten lithium chloride (LiCl) were taken at various temperatures with a quartz rod. The solubility was determined by inductive coupled plasma-mass spectrometry (ICP-MS), where the ratio of the gallium to lithium concentration was calculated. Recrystallization studies were performed in quartz tubes in the same stainless steel chamber, where the recrystallization surface was MOCVD-grown GaN on sapphire. Deposited GaN crystallites were observed by scanning electron microscopy (SEM) and elemental

PAGE 89

77 analysis on these crystallites was performed by energy dispersive spectroscopy (EDS) in order to determine the extent to which GaN recrystallized on the wafer surface. Results and Discussion In order to grow GaN crystals at a fast enough rate to be commercially viable, it is necessary to dissolve a relatively high concentration of GaN into solution. If the solubility is found to be sufficiently high, it would be possible to grow GaN bulk crystals at a reasonable rate upon supersaturation. The basis for the experimentation demonstrated in this chapter is the ideal solubility curve for GaN, shown in Figure 5-2. This curve was calculated by the following equation [64]: 1lnTTRSxmFnn (5-1) Under diffusion-limited conditions, a solubility of ~0.1% for GaN could provide a crystal growth rate of ~0.5 mm/h. This corresponds to a temperature of ~1230C on the ideal solubility curve, and since this process occurs at atmospheric pressure, these conditions are mild enough for a commercially viable method of GaN crystal growth. If the selected solvent has favorable interactions with GaN, then experimentally, the data should lie above this curve, corresponding to higher solubilities, and consequently, potentially higher growth rates. The first series of experiments determined the temperature dependent solubility of GaN powder in LiCl. Dissolution of Gallium Nitride in Lithium Chloride The temperature dependent solubility of GaN in LiCl was experimentally determined by performing ICP-MS analysis of aliquots of solution at various temperatures. These temperatures ranged from 750C 1000C, as shown in Figure 5-3. The experimental data indicated that, in most cases, more GaN dissolved than was

PAGE 90

78 predicted by the ideal solubility curve. The scatter in the data was most likely due to the fact that the ICP-MS is an inorganic, water-based technique, and since GaN is insoluble in water, some of the crystallites may have precipitated out, and segregated. In order to address this, a 2 m filter was used to filter samples for analysis prior to injection into the ICP-MS. These results, shown in Figure 5-4, represent a lower bound for solubility of GaN in LiCl, but they were still significantly above the ideal solubility curve. These data indicate that 10-50 ppm of GaN were soluble in LiCl at temperatures ranging from 800C 950C, corresponding to a growth rate of ~50-100 m/h. These values were sufficient to commence experimentation of recrystallization of GaN on a GaN template. Gallium Nitride Recrystallization In order to determine whether GaN could recrystallize out of solution, a series of experiments was performed, where GaN powder in LiCl was heated to ~950C, and then, became supersaturated upon cool-down. This supersaturated GaN would presumably deposit on a lattice-matched template. SEM images of a surface not exposed to the melt (Figure 5-5) were obtained and compared to one that was (Figure 5-6). Note the distinct differences in surface morphology. Upon further investigation, individual features were observed at higher magnifications, as shown in Figure 5-7. This feature appears to exhibit a hexagonal shape, indicative of the wurtzite crystal structure. To further promote GaN recrystallization, another experiment was run with stirring at the maximum temperature (950C), prior to cool-down. The result of this experiment was that more surface features appeared. From this finding, a third experiment was performed, with 1) half of the wafer surface treated with nitric acid (HNO 3 ) at 50C, and 2) constant agitation during cool-down. The surface treatment was

PAGE 91

79 performed in order to etch the native oxide off the growth surface, and the agitation was performed in order to provide forced convection, and greater recrystallization. This provided two interesting results. First, more features were formed on the wafer surface. Second, the features that formed on the treated surface tended to be different from those on the untreated surface. On the treated surface, there was a high density of patchy features, shown in Figure 5-8, whereas, the untreated surface had a high density of needle-like features, shown in Figures 5-9 to 5-11. In order to determine the chemical composition of the two observed features, elemental analysis was performed by EDS. The EDS results for a needle structure is shown in Figure 5-12, and strongly indicate that this structure is comprised of etched quartz, as evidenced by the heavy silicon and oxygen concentrations as well as the gallium void indicated. Elemental analysis of a patch structure is shown in Figure 5-13, and is indicative of GaN; the gallium and nitrogen concentrations are high, and silicon and oxygen are practically nonexistent. Another interesting finding is that silicon oxide (SiO x ) preferentially deposited on the untreated surface, which presumably contained the native oxide, whereas GaN preferentially deposited on the treated surface. A cross-sectional view of the patch demonstrated in the SEM image in Figure 5-13 is shown in Figure 5-14. From the scale shown, this GaN crystallite is ~5 m in thickness, which indicates that the growth rate in this experiment was ~2.5 m/h; this is significantly higher than any other bulk growth technique. Furthermore, the depth of field of the EDS instrument is shown in Figure 5-15 for the highest power used in this study. Since the feature probed is ~5 m, and the depth of the instrument is ~1600 nm, this further supports that the crystallite is that of pure GaN.

PAGE 92

80 While the experimentation described in this chapter is relatively unrefined, and far from optimized, it demonstrates that a simple dissolution/recrystallization process is a potentially viability one for the fabrication of GaN bulk crystals from a molten alkalihalide medium. The results were encouraging, and could provide a first step in the development of a novel method for growing GaN crystals. Significant improvements could be made to the process, including solvent optimization, greater purity of reagents, greater control over the fluid dynamics, and using a true seed crystal on which to crystallize GaN from the molten solvent. Eventually, using this method, researchers could potentially develop a commercially feasible, rapid growth rate process to fabricate high crystalline quality GaN boules for homoepitaxial GaN growth. Summary An ideal solubility curve was calculated for GaN with respect to temperature, based on intrinsic thermodynamic parameters. The solubility of GaN in LiCl was experimentally determined, and demonstrated that GaN was more soluble than the ideal curve indicated. From these results, a GaN template was placed in a LiCl medium containing GaN powder, which was heated to ~1000C, and then cooled. Deposition was observed on the GaN template surface by SEM. Further experimentation demonstrated that GaN crystallites deposited at a higher rate after pretreatment of the GaN template surface in an acid solution, and with constant fluid agitation during cool-down. Elemental analysis determined that patchy structures as thick as 5 m were, in fact, GaN. These crystallites formed after merely 2 hours, indicating an extraordinarily high growth rate, relative to other GaN crystal growth techniques.

PAGE 93

81 Potential improvements to this experimentation include More sophisticated equipment to allow greater control over thermal gradients and fluid dynamics. Use of higher purity reagents. More comprehensive studies optimizing GaN solubility. Use of a true seed crystal upon which to deposit dissolved GaN.

PAGE 94

82 Figure 5-1. Gallium nitride crystal with dimensions of 0.9 mm 0.6 mm. The bulk growth technique to form this crystal was electrodeposition.

PAGE 95

83 Ideal Solubility of GaN00.050.10.150.20.2510001100120013001400Temperature (oC)Solubility (%) Figure 5-2. Ideal solubility of gallium nitride. The data for this curve were determined by Equation 5-1.

PAGE 96

84 Solubility of GaN in LiCl02040608010012075080085090095010001050Temperature (oC)GaN Solubility (ppm) measured ideal Figure 5-3. Experimentally determined solubility of gallium nitride in lithium chloride compared to the ideal solubility curv e. These data points were obtained without the use of a filter.

PAGE 97

85 Solubility of GaN in LiCl05101520253035406507508509501050Temperature (oC)Solubility (ppm) Ideal Solubility Experimental Data Figure 5-4. Experimentally determined solubility determined after employing a 2 m filter. These data represent a lower bound for solubility of gallium nitride in lithium chloride.

PAGE 98

86 Figure 5-5. Scanning electron microscopy image of a bare gallium nitride surface. The morphology is smooth.

PAGE 99

87 Figure 5-6. Scanning electron microscopy image of a gallium nitride surface exposed to gallium nitride-in-lithium chloride melt. The morphology is noticeably rougher than the bare, untreated surface.

PAGE 100

88 Figure 5-7. Close-up scanning electron microscopy image of a gallium nitride crystallite. Note the hexagonal structure.

PAGE 101

89 Figure 5-8. Scanning electron microscopy image of a treated gallium nitride surface after exposure to the gallium nitr ide-in-lithium chloride melt. Note the prevalence of patchy structures.

PAGE 102

90 Figure 5-9. Scanning electron microscopy image of an untreated gallium nitride surface after exposure to a gallium nitride-in-lithium chloride melt. Note the prevalence of needle-like structures.

PAGE 103

91 Figure 5-10. Scanning electron microscopy image focusing on needles on the untreated gallium nitride surface at 750. Note th e prevalence of needle structures on this part of the wafer.

PAGE 104

92 Figure 5-11. Scanning electron microscopy image focusing on needles of the untreated gallium nitride surface at 1900. The stru ctures appear to have a hexagonal structure and exhibits columnar growth.

PAGE 105

93 Needle Figure 5-12. Election dispersive spectroscopy images for elemental analysis on needle structures. Note the high con centrations of silicon and oxygen.

PAGE 106

94 Figure 5-13. Electron dispersive spectroscopy images of a patch on the gallium nitride surface. Note the high concentrations of gallium and nitrogen, and low concentrations of silicon and oxygen.

PAGE 107

95 Figure 5-14. Scanning electron microscopy image of a deposited gallium nitride crystallite with a thickness of 5 m. This crystallite was formed after approximately 2 hours.

PAGE 108

96 Figure 5-15. Electron dispersive spectroscopy instrument depth calibration. This represents a Monte Carlo simulation for the highest voltage used (15keV).

PAGE 109

CHAPTER 6 SUMMARY AND FUTURE DIRECTIONS In this dissertation, specific effects of dislocations in III-nitride materials are investigated. Samples of a wide range of dislocation density were grown by MOCVD to determine their specific effects. In order to enable run-to-run feedback for growth optimization studies, a method to determine dislocation density based on HRXRD results was developed. This method allows researchers to obtain results for dislocation density within one hour, and was shown to be accurate by its agreement with TEM results within a factor of 1.5 for all seven samples analyzed. This XRD-based method provided the foundation for subsequent research investigating the effects of dislocations on electronic properties and impurity incorporation of GaN films during MOCVD growth. Further study of this technique should be performed to investigate its accuracy with respect to dislocation type. Screw dislocations are considered to have a different effect on electronic properties than edge dislocations, and determining the density of dislocations with a screw component could be essential for further growth optimization studies. The method shown in Chapter 3 is accurate for dislocation density with an edge component, but has not been shown to be accurate with respect to dislocations with a screw component. In the MOCVD films grown, the edge-type dislocation density was greater than the screw-type dislocation density by approximately an order of magnitude in all cases. Inaccuracies for screw dislocation density would be suppressed, since edge dislocation density is significantly higher. For more conclusive results, a similar series of experiments should be performed studying the accuracy of the technique compared to 97

PAGE 110

98 TEM results for dislocations with a screw component. Furthermore, the technique should be investigated for samples of relatively low dislocation densities ( 10 7 cm -2 ). As growth optimization and III-nitride bulk crystal development continue to progress, samples of lower dislocation density will be grown, and this techniques accuracy should be confirmed for higher quality III-nitride crystals. Furthermore, in samples grown by cantilever epitaxy (CE) and epitaxial lateral overgrowth (ELO), the distribution of dislocations may be piled-up rather than random, so the method to calculated dislocation density proposed in Chapter 3 may need to be modified to account for this. Finally, the effect of film thickness on XRD line width must be ascertained. Currently, the model does not account for broadening due to strain. For samples of low thickness (<4000 ), strain broadening has a noticeable effect on measured XRD line width, providing anomalously high results for dislocation density when using the model in Chapter 3. Determining the thickness where strain broadening must be accounted for would be useful, and including strain broadening into the model for thin samples could provide benefits for growth optimization studies. Therefore, further studies are required to include broadening due to strain for thinner films. For the growth studies investigating impurity incorporation into GaN films, more experimentation is necessary to conclusively determine whether threading edge dislocations provide energetically favorable sites for carbon incorporation during MOCVD growth. One possible series of experiments involves the measurement of carbon incorporation on variable dislocation density templates as temperature is varied. It has been shown that subjecting templates to lower growth temperatures causes more carbon to incorporate. Repeating the experimentation described in Chapter 4 with respect

PAGE 111

99 to temperature could provide more insight on the effect of threading edge dislocations on impurity incorporation. Impurities may incorporate by a different mechanism with respect to growth temperature, as opposed to growth pressure. The proposed mechanism with respect to pressure is one that involves molecular interactions between carbon on the growth surface and the hydride gases in the ambient. The carbon removal mechanism was proposed to be one that involves carbon reacting with hydrides to form methyl radicals, which are subsequently removed from the growth surface and pumped out through the exhaust. Varying temperature while pressure is held constant could provide insight on whether threading edge dislocations provide energetically favorable sites for carbon incorporation. Furthermore, samples over a broader range of dislocation density would be desired for this study. It is possible that the effect of threading edge dislocations (if any) is different for lower dislocation density samples. Growing on CE samples, where dislocation density, ~10 6 cm -2 could provide further insight on potential mechanisms for impurity incorporation. Where dislocations demonstrate a deleterious effect on IIInitride films and IIInitride based devices, III-nitride bulk crystals are desired. In order to further investigate the possibility of depositing GaN bulk crystals, the research from Chapter 5 should be continued to optimize the dissolution/recrystallization process demonstrated. This includes performing solubility studies with a wide variety of alkali-halide salts in order to determine which best dissolves GaN. Once this is determined, deposition optimization studies should be performed. This includes gaining better control over the fluid dynamics and thermal gradients, in order to maximize the rate of GaN crystal formation. More sophisticated equipment will be required for this, such as a multistage

PAGE 112

100 heater and higher purity reagents. Moreover, similar experimentation could be performed to ascertain the possibility of fabricating AlN, InN, bulk crystals as well as alloys of bulk III-nitride crystals. If no solvent is deemed to be suitable for a commercially viable process to extract bulk GaN (or bulk AlN, InN, or III-nitride alloy) crystals, the solubility results could still prove to be beneficial, even if this dissolution/ recrystallization method is not ultimately used. The method based on electrochemical generation of precursors could be employed, where knowing the solubility would be beneficial since it would indicate the extent to which liquid phase pre-reactions of Ga 3+ and N 3ions must be minimized. Using such a method, the growth rate of a GaN bulk crystal would be controlled by fluid dynamics, but developing a process where pre-reactions are minimized could be difficult, due to the highly reactive nature of N 3ions. Finally, if this electrochemical generation of precursors method provides growth rates that are too low to be a commercially viable process, then electro-deposition of GaN films could be further optimized. While this technique for fabricating high quality bulk crystals demonstrates several difficulties, a process could be developed where single crystal GaN wafers could be formed and extracted from solution by employing fluid dynamics to ensure uniform N 3coverage over a liquid gallium electrode. The reactive N 3ions would react with Ga 3+ ions at the gallium liquid interface, forming a single crystal crust of GaN that could be extracted. Using such a method would remove the need to cleave a boule of bulk GaN, as such a process would produce individual GaN wafers. Optimization of the electro-deposition process could then be scaled to produce wafers of desired dimensions.

PAGE 113

LIST OF REFERENCES 1. W. Walukiewicz, S.X. Li, J. Wu, K.M. Yu, J.W.Ager III, E.E. Haller, H. Lu, W.J. Schaff, J. Cryst. Growth, 269, p. 119, (2004). 2. K.B. Nam, J. Li, K.H. Kim, J.Y. Lin, H.X. Jiang, Appl. Phys. Lttr., 78, (23), p. 3690, (2001). 3. L.F. Eastman, U.K. Mishra, IEEE Spectrum, p. 28, (2001). 4. O. Ambacher, B. Foutz, J. Smart, J.R. Shealy, N.G. Weimann, K. Chu, M. Murphy, A.J. Sierakowski, W.J. Schaff, L.F. Eastman, R. Dimitrov, A. Mitchell, M. Stutzmann, J. Appl. Phys., 87, (1), p. 334, (2000). 5. L.F. Eastman, V. Tilak, V. Kaper, J. Smart, R. Thomson, B. Green, J.R. Shealy, T. Prunty, Phys. Stat. Sol., 194, (2), p. 433, (2002). 6. I. Akasaki, J. Cryst. Growth, 237-239, p. 905, (2002). 7. S.J. Pearton, C.R. Abernathy, M.E. Overberg, G.T. Thaler, A.H. Onstine, B.P. Gila, F. Ren, B. Lou, J. Kim, Mat. Today, p. 24, (2002). 8. M. Razeghi, Proc. IEEE, 90, (6), p. 1006, (2002). 9. M.M. Wong, U. Chowdhury, C.J. Collins, B. Yang, J.C. Denyszyn, K.S. Kim, J.C. Campbell, R.D. Dupuis, Phys. Stat. Sol., 188, (1) p. 333, (2001). 10. S. Keller, Y. F. Wu, G. Parish, N. Ziang, J.J. Xu, B.P. Keller, S.P. DenBaars, U.K. Mishra, IEEE Trans. Elec. Dev., 48, (3), p. 552, (2001). 11. I. Akasaki, J. Cryst. Growth, 221, p. 231, (2000). 12. Y.F. Wu, A. Saxler, M. Moore, R.P. Smith, S. Sheppard, P.M. Chavarkar, T. Wisleder, U.K. Mishra, P. Parikh, IEEE Elec. Dev. Lett, 25, (3), p. 117, (2004). 13. J. Elsner, R. Jones, P. Sitch, V. Porezag, M. Elstner, T. Frauenheim, M. Heggie, S. Oborg, P. Briddon, Phys. Rev. Lett. 79, p. 3672, (1997). 14. V. Kirchner, M. Fehrer, S. Figge, H. Heinke, S. Einfeldt, D. Hommel, H. Selke, P. Ryder, Phys, Stat. Sol. 216, p. 659, (1999). 15. D. Look, J. Sizelove, Phys. Rev. Lett. 82, p. 1237, (1999). 101

PAGE 114

102 16. P. Hansen, Y. Strausser, A. Erickson, E. Tarsa, P. Kozodoy, E. Brazel, J. Ibbetson, U. Mishra, V. Narayanamurti, S. DenBaars, J. Speck, Appl. Phys. Lett. 72, p. 2247, (1998). 17. H. Amano, S. Kamiyama, I. Akasaki, Proc. IEEE, 90, (6), p. 1015, (2002). 18. R. Dimitrov, M. Murphy, J. Smart, W. Schaff, J.R. Shealy, L.F. Eastman, O. Ambacher, M. Stutzmann, J. Appl. Phys., 87, (7), p. 3375, (2000). 19. J. Kozlowski, R. Paszkiewicz, R. Korbutowicz, M. Tlaczala, Physica B, 308-310, p. 114, (2001). 20. C.J. Sun, P. Kung, A. Sexler, H. Ohsato, K. Haritos, M. Razeghi, J. Appl. Phys., 75, (8), p. 3964, (1994). 21. N. Onojima, J. Suda, H. Matsunami, J. Cryst, Growth, 237-239, p. 1012, (2002). 22. D.D. Koleske, A.J. Fischer, A.A. Allerman, C.C. Mitchell, K.C. Cross, S.R. Kurtz, J.J. Figiel, K.W. Fullmer, W.G. Breiland, Appl. Phys. Lett., 81, (11), p. 1940, (2002). 23. D. Follstaedt, N. Missert, D. Koleske, C. Mitchell, K. Cross, Appl. Phys. Lett. 83, p. 4797, (2003). 24. D. Kapolnek, X. H. Wu, B. Heying, S. Keller, B. P. Keller, U. K. Mishra, S. P. DenBaars, and J. S. Speck, Appl. Phys. Lett. 67, (11) p.1541, (1995). 25. T. Sugahara, H. Sato, M. Hao, Y. Naoi, S. Kurai, S. Tottori, K. Yamashita, K. Nishino, L. T. Romano, S. Sakai, Jpn. J. Appl. Phys. 37, (4A) p. L398, (1998). 26. M. Hordon, B. Averbach, Acta Metall. 9, (3), p. 237, (1961). 27. C. Dunn, E. Koch, Acta. Metall. 5, (10), p. 548, (1957). 28. G. Williamson, W. Hall, Acta Metall. 1, (1), p. 22, (1953). 29. P. Gay, P. Hirsch, A. Kelly, Acta Metall. 1, (3), p. 315, (1953). 30. T. Metzger, R. Hopler, E. Born, O. Ambacher, M. Stutzmann, R. Stommer, M. Schuster, H. Gobel, S. Christiansen, M. Albrecht, H. Strunk, Phil. Mag. 77, (4), p. 1013, (1998). 31. H. Kang, N. Spencer, D. Nicol, Z. Feng, I. Ferguson, S. Guo, M. Pophristic, B. Peres, Mat. Res. Soc. Symp. Proc., 743, p. 405, (2003). 32. V. Srikant, J. Speck, D. Clarke, J. Appl. Phys. 82, (9), p.4286, (1997). 33. R. Chierchia, T. Bottcher, H. Heinke, S. Einfeldt, S. Figge, D. Hormel, J. Appl. Phys. 93, (11), p. 8918, (2003).

PAGE 115

103 34. Y. Sun, O. Brandt, T. Liu, A. Trampert, K. Ploog, J. Blasing, A. Krost, Appl. Phys. Lett. 81, (26), p. 4928, (2002). 35. B. Heying, X. Wu, S. Keller, Y. Li, D. Kapolnek, B. Keller, S. DenBaars, J. Speck, Appl. Phys. Lett. 68, (5), p. 643, (1996). 36. J. Ayers, J. Cryst. Growth 135, (135), p. 71, (1994). 37. B. Williams, C. Carter, Transmission Electron Microscopy, Plenum Press, New York, 1996, pp. 301-323. 38. N. Zhang, V. Mehrota, S. Chandrasekaran, B. Moran, L. Shen, U. Mishra, E. Etzkorn, D. Clarke, IEEE Pwr. Elec. Spec. Conf., p. 233, (2003). 39. D.D. Koleske, A.E. Wickenden, R.L. Henry, M.E. Twigg, J. Cryst. Growth, 242, p. 55, (2002). 40. W. Gotz, N.M. Johnson, C. Chen, H. Liu, C. Kuo, W. Imler, Appl. Phys. Lett., 68, (22), p. 3144, (1996). 41. W.J. Moore, J.A. Freitas Jr., G.C.B. Braga, R.J. Molnar, S.K. Lee, K.Y. Lee, I.J. Song, Appl. Phys. Lett., 79, (16), p. 2570, (2001). 42. H. Wang, A.B. Chen, J. Appl. Phys., 87, (11), p. 7859, (2000). 43. D. Volm, K. Oettinger, T. Streibl, D. Kovalev, M. Ben-Chorin, J. Diener, B.K. Meyer, J. Majewski, L. Eckey, A. Hoffman, H. Amano, I. Akasaki, K. Hiramatsu, T. Detchprohm, Phys. Rev. B, 53, (24), p. 16543, (1996). 44. X. Xu, H. Liu, C. Shi, Y. Zhao, S. Fung, C.D. Beling, J. Appl. Phys., 90, (12), p. 6130, (2001). 45. J.E. Van Nostrand, J. Solomon, A. Saxler, Q.H. Xie, D.C. Reynolds, D.C. Look, J. Appl. Phys., 87, (12), p. 8766 (2000). 46. V.A. Joshkin, C.A. Parker, S.M. Bedair, J.F. Muth, I.K. Shmagin, R.M. Kolbas, E.L. Piner, R.J. Molnar, J. Appl. Phys., 86, (1), p. 291, (1999). 47. M.A. di Forte-Poisson, F. Huet, A. Romann, M. Tordjman, D. Lancefield, E. Pereira, J. Di Persio, B. Pecz, J. Cryst. Growth, 195, p. 314, (1998). 48. J. Oila, V. Ranki, J. Kivioja, K. Saarinen, P. Hautojarvi, J. Likonen, J.M. Baranowski, K. Pakula, T. Suski, M. Leszczynski, I. Grzegory, Phys. Rev. B, 63, p. 045205-1, (2001). 49. S. Heikman, S. Keller, S.P. DenBaars, U.K. Mishra, Appl. Phys. Lett., 81, (3), p. 439, (2002).

PAGE 116

104 50. J.W.P. Hsu, D.V. Lang, S. Richter, R.N. Kleiman, A.M. Sergent, R.J. Molnar, Appl. Phys. Lett., 77, (18), p. 2873, (2000). 51. M.E. Twigg, D.D. Koleske, A.E. Wickenden, R.L. Henry, S.C. Binari, Appl. Phys. Lett., 79, (26), p. 4322, (2001). 52. D.C. Look, R.J. Molnar, Appl. Phys. Lett., 70, (25), p. 3377, (1997). 53. M.G. Cheong, K.S. Kim, C.S. Oh, N.W. Namgung, G.M. Yang, C.H. Hong, K.Y. Lim, E.K. Suh, K.S. Nahm, H.J. Lee, D.H. Lim, A. Yoshikawa, Appl. Phys. Lett., 77, (16), p. 2557 (2000). 54. C.H. Seager, A.F. Wright, J. Yu, W. Gotz, J. Appl. Phys., 92, (11) p. 6553, (2002). 55. A.F. Wright, J. Appl. Phys., 92, (5), p. 2575, (2002). 56. A.E. Wickenden, D.D. Koleske, R.L. Henry, M.E. Twigg, M. Fatemi, J. Cryst. Growth, 260, p. 54, (2004). 57. R.J. Graham, K.V. Ravi, Appl. Phys. Lett., 60, (11), p. 1310, (1992). 58. R.A. Brown, O. Kononchuk, G.A. Rozgonyi, S. Koveshnikov, A.P. Knights, P.J. Simpson, F. Gonzalez, J. Appl. Phys., 84, (5), p. 2459, (1998). 59. A.E. Wickenden, D.D. Koleske, R.L. Henry, R.J. Gorman, M.E. Twigg, M. Fatemi, J.A. Freitas, W.J. Moore, J. Elec. Mat., 29, (1), p.21, (1999). 60. M. Leszczynski, I. Grzegory, H. Teisseyre, T. Suski, M. Bockowski, J. Jun, J.M. Baranowski, S. Porowski, J. Domagala, J. Cryst. Growth, 169, p.235, (1996). 61. M.J. Callahan, B.G. Wang, L.O. Bouthillette, S.Q. Wang, J.W. Kolis, D.F. Bliss, Mat. Res. Soc. Symp. Proc., 798, p. 263, (2003). 62. T. Goto, M. Tada, Y. Ito, Electrochim. Acta, 39, (8-9), p. 1107, (1994). 63. T. Goto, Y. Ito, Electrochim. Acta, 43, (21-22), p. 3379, (1998). 64. M.J. Gentilcore, Chem. Eng. Proc. Mag. 100, (3), p. 38, (2004).

PAGE 117

BIOGRAPHICAL SKETCH Allen M. West was born on January 26, 1975 in Baltimore, MD. After graduating from the Gilman School in 1994, he attended college at Washington & Lee University in Lexington, VA, where he earned a Bachelor of Science in chemistry-engineering. After graduating, he commenced graduate studies in chemical engineering at Florida State University in Tallahassee, FL. After one year, he transferred to the University of Florida, where he studied under Dr. Oscar Crisalle, and earned a Master of Engineering in chemical engineering. Upon graduation, he joined the Materials Science and Engineering Department, under Dr. Cammy Abernathy. Dr. Abernathy arranged an opportunity for him at Sandia National Laboratories in the spring of 2001, where he completed his graduate research in III-nitride semiconductor materials in 2004. 105


Permanent Link: http://ufdc.ufl.edu/UFE0009281/00001

Material Information

Title: Effects of Dislocations on Electronic Properties of III-Nitride Materials
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0009281:00001

Permanent Link: http://ufdc.ufl.edu/UFE0009281/00001

Material Information

Title: Effects of Dislocations on Electronic Properties of III-Nitride Materials
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0009281:00001


This item has the following downloads:


Full Text












EFFECTS OF DISLOCATIONS ON ELECTRONIC PROPERTIES OF III-NITRIDE
MATERIALS















By

ALLEN M. WEST


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2005




























"I hereby dedicate this [dissertation] to ... myself."

(credit to Rodney Dangerfield as Thornton Melon, Back to School)















ACKNOWLEDGMENTS

I would first like to thank Dr. Cammy Abernathy for all of her guidance and

support throughout my tenure in the Materials Science Department. She was the ideal

advisor for me, and was crucial to my success. I thank Dr. Stephen Pearton, Dr. Fan Ren,

and Dr. David Norton as well for serving on my committee. I especially thank

Dr. Crisalle for serving as my external committee member from chemical engineering.

I credit Dr. Crisalle for his work as Graduate Admissions Coordinator for the Chemical

Engineering Department. If not for Dr. Crisalle, I would never have ended up at UF.

I thank Dr. Karen Waldrip for all of her support and guidance during my time at

Sandia. Many of my successes in graduate school were due, in large part, to Karen's

tireless efforts. I thank Dr. Stephen Lee and Dr. Daniel Koleske for their significant

contributions to my research efforts. I further acknowledge Dr. Andrew Allerman for his

thorough feedback and instilling strong sense of attention to details in me.

I acknowledge Dr. Brent Gila for his contributions to the work in this dissertation,

and I thank Dr. Andrea Onstine and Danielle Stodilka for their support and collaboration

for various classes and qualifying exams. I especially thank Dr. Jerry Thaler for being a

reliable liaison in Florida while trying to complete the bureaucratic requirements for

graduation from New Mexico. I gratefully acknowledge Jerry's efforts in this regard.

I also thank Jerry, Dr. Omar Bchir, Jeff Sharp, Brandy Colwell, and Sarah Rich for their

camaraderie throughout my graduate school experience.









I acknowledge Dr. Sam Kozak for the inspiration and advice he provided during

my undergraduate days. I also thank Dr. Ken Van Ness and Dr. Erich Uffelman for

serving as my advisors at Washington and Lee. Both were very important to my

intellectual development. I also thank Dr. Douglas Szajda, Dr. Marcia France, and

Dr. Steven Desjardins. All were superb professors.

Dating back to high school, I must acknowledge Mr. James Morrison for piquing

my interest in science and tolerating me for three years. I also acknowledge Mr. Robert

Bulkeley for doing an outstanding job of advising me throughout high school.

Finally, and most importantly, I thank my family for their undying support

throughout my graduate school experience.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ......... .................................................................................... iii

LIST OF FIGURES ............. ................... ............ .......... ............... .. viii

ABSTRACT .............. .......................................... xi

CHAPTER

1 IN TRODU CTION .................................... .......... .... ............. .

Background and M otivation ...................................................... .......................
Electronic Properties of III-N itrides ............. ............................................................
Chemical, Structural, and Physical Properties of III-Nitrides......................................2
Suitability of III-Nitrides in Optoelectronic Devices ............................................
Suitability of III-Nitrides in Electronic Devices........................................................3
Progress in III-Nitride Material Development.................. ...........................5
Availability of III-N itride Based D evices................................. ...................... 6
State ent of the P problem ............................................................................. ........ 6
O u tlin e ............................................................................ . 7

2 EXPERIM ENTAL M ETHOD S........................................... ........................... 10

G row th Techniques................. ... ...................... .. ............ .... .............. 10
Metalorganic Chemical Vapor Deposition.................. ......................................10
M olecular B eam Epitaxy ......................................................... ............. 14
G allium N itride Bulk G row th ................................... ........................... .. ........ 15
C haracterization T techniques ........................................................... .....................16
High-Resolution X-Ray Diffraction.....................................................16
Transmission Electron Microscopy ........................... ........ ...............17
Secondary Ion Mass Spectroscopy ........... ................................. ...............18
Superconducting Quantum Interference Device Magnetometry .........................18
Capacitance-Voltage .................... ........................................ .. ............... 19
Inductively-Coupled Plasma-Mass Spectroscopy ............................................19
Scanning Electron M icroscopy....................................... ......................... 20
Electron Dispersive X-Ray Spectroscopy ................................. ............... 20
A uger Electron Spectroscopy ................................... .......................... .. ......... 21




v









3 DETERMINATION OF DISLOCATION DENSITY OF III-NITRIDE FILMS BY
X -R A Y D IF FR A C T IO N ................................................................... ....................23

In tro du ctio n ......................................................... ............. ................ 2 3
Characterization of D islocations ........................................ ....................... 23
Previous W ork ............. .... ............ .................................. ....... 24
Tw ist ............... ......... ....... ..... .... ... ............ ........... 25
Incoherence ............... .......... ..... ........... .......... .......... 26
F o c u s ............................................................................................................... 2 6
E xperim ental M ethods......................................................................... ..................27
M eth od ............................................. ........................................... ..........................2 9
S ou rces of B roadening ............................................................... ....................2 9
M o d el .............. .......... .. ...................................................... ... .3 1
Determination of Convolution......................... ............................. 32
Calculation of Dislocation Density ............... ....... .. ..................................33
Simplified Model .............. .................. .. ... ................ 34
R results and D discussion ........................... ...... ... .. ...... .............. 35
S u m m a ry ........................................................................................................3 7

4 EFFECT OF DISLOCATIONS ON IMPURITY INCORPORATION IN
METALORGANIC CHEMICAL VAPOR DEPOSITION GROWTH...................54

Introduction ...................................................................................... .. ... .... ..54
S ilic o n .......................................................................................5 4
O x y g e n ................................................................5 5
C a rb o n ............................................................................................................ 5 6
Previous W ork ..................................... ....................... 57
F o c u s ............................................................................................................... 5 8
Experim ental M ethods................................................. 58
R esu lts an d D iscu ssion .......................................................................................... 59
S u m m a ry ......................................................................................................6 1

5 GALLIUM NITRIDE BULK CRYSTAL GROWTH BY DISSOLUTION AND
RECRYSTALLIZATION OF GALLIUM NITRIDE POWDER ..............72

Introdu action .............. ................. .................................................................. 72
Previous Work ................ ......... ..................72
P proposed M methods ...............................................................74
Electroplating ................................................. 74
Electrochemical Reduction of Nitrogen Gas .....................................................75
Dissolution and Recrystallization of Gallium Nitride Crystals ...........................75
F o c u s ...................................................................................................................... 7 6
Experim ental M ethods................................................. 76









R results and D discussion ............................... ........................................ .. ..........77
Dissolution of Gallium Nitride in Lithium Chloride ...................................77
Gallium Nitride Recrystallization................... ....... ........................... 78
Summary ......... ........ ............ ............. ............... 80

6 SUMMARY AND FUTURE DIRECTIONS............... .................................97

L IST O F R E F E R E N C E S ............. .............................................................................. 10 1

B IO G R A PH ICA L SK ETCH ........... ..................................................... .....................105
















LIST OF FIGURES


Figure p

1-1 Band gap engineering diagram for III-V materials ..........................................8

2-1 Schematic of metalorganic chemical vapor deposition operation..........................22

3-1 Schem atic of reciprocal lattice points probed ............ ................... ..................39

3-2 Adjacent domains tilted with respect to one other ...........................................40

3-3 Adjacent domains twisted with respect to one another..........................................41

3-4 Cross-section transmission electron microscopy image of a low dislocation density
sam p le ......................................................... .................. 4 2

3-5 Cross-section transmission electron microscopy image of a high dislocation density
sam p le ......................................................... .................. 4 3

3-6 Weak beam transmission electron microscopy image and convergent electron beam
diffraction im age ...................... .................... .. .. ........... .... ....... 44

3-7 Broadening effects demonstrated in reciprocal space.................. ................45

3-8 Plotted data of full width at half maximum vs. reciprocal lattice vector magnitude46

3-9 Fit of full width at half maximum data vs. angle of inclination .............................47

3-10 Mathematical and geometric derivation of the random distribution method for
calculating dislocation density ........................................... .......................... 48

3-11 Calculation of dislocation density based on values for coherence length ..............49

3-12 Comparison of methods of determining dislocation density by XRD ..................50

3-13 Dislocation density determined by transmission electron microscopy vs. dislocation
density determined by x-ray diffraction.... ........... ....................................... 51

3-14 Dependence of dislocation density on growth pressure...........................................52

3-15 Effect of dislocation density on room temperature mobility of gallium nitride high
electron m obility transistors ......................................................... ............. 53









4-1 Impurity profiles for low dislocation density at the top ...........................................62

4-2 Impurity profiles for mid dislocation density at the top ...........................................63

4-3 Impurity profiles for high dislocation density at the top........................................64

4-4 Impurity profiles for high dislocation density at the flat.......................................65

4-5 Impurity profiles for second series of overgrowths for low dislocation density at the
flat.............. .................... ..................................... ......... ...... 6 6

4-6 Impurity profiles for second series of overgrowths for mid dislocation density at the
flat.............. .................... ..................................... ......... ...... 6 7

4-7 Impurity profiles for second series of overgrowths for high dislocation density at
th e fla t .......................................................................... 6 8

4-8 Comparison of impurity profiles for different dislocation density templates ..........69

5-1 Gallium nitride crystal with dimensions of 0.9 mm x 0.6 mm .............................82

5-2 Ideal solubility of gallium nitride ......................................................... ..................83

5-3 Experimentally determined solubility of gallium nitride in lithium chloride
com pared to the ideal solubility curve ........................................ ............... 84

5-4 Experimentally determined solubility determined after employing a 2 |tm filter....85

5-5 Scanning electron microscopy image of a bare gallium nitride surface...................86

5-6. Scanning electron microscopy image of a gallium nitride surface exposed to
gallium nitride-in-lithium chloride melt........... ............................ .............87

5-7 Close-up scanning electron microscopy image of a gallium nitride crystallite .......88

5-8 Scanning electron microscopy image of a treated gallium nitride surface after
exposure to the gallium nitride-in-lithium chloride melt ......................................89

5-9 Scanning electron microscopy image of an untreated gallium nitride surface after
exposure to a gallium nitride-in-lithium chloride melt ........................................90

5-10 Scanning electron microscopy image focusing on needles on the untreated gallium
nitride surface at 750x ........................ .............. .............................91

5-11 Scanning electron microscopy image focusing on needles of the untreated gallium
nitride surface at 1900 x ............................................................................. .... 92

5-12 Election dispersive spectroscopy images for elemental analysis on needle
stru ctu res ............................................................................ 9 3









5-13 Electron dispersive spectroscopy images of a patch on the gallium nitride surface 94

5-14 Scanning electron microscopy image of a deposited gallium nitride crystallite with
a thickness of 5 m ............. .... ...................................................... ......... ...... 95

5-15 Electron dispersive spectroscopy instrument depth calibration ............................96















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EFFECTS OF DISLOCATIONS ON ELECTRONIC PROPERTIES OF III-NITRIDE
MATERIALS

By

Allen M. West

May 2005

Chair: Cammy Abernathy
Major Department: Materials Science and Engineering

By varying metalorganic chemical vapor deposition (MOCVD) growth conditions,

III-nitride samples with dislocation densities ranging over two orders of magnitude were

grown, and the effects of dislocations were studied. Initially, a novel technique to

determine dislocation density by high-resolution x-ray diffraction (HRXRD) was

developed. Results were compared to those obtained by transmission electron

microscopy (TEM) and agreed within a factor of 1.5. This HRXRD-based technique

provided rapid and accurate feedback for use in growth-optimization studies; and became

a basis for future studies of the effects of dislocation on specific phenomena associated

with III-nitride technology. This included a study of the effect of dislocations on

impurity incorporation during MOCVD growth. Dislocations were previously

conjectured to provide energetically favorable sites for impurity incorporation, but

experimental results did not confirm this.









Finally, a novel growth technique was investigated for gallium nitride (GaN) bulk

crystals, based on a simple dissolution/recrystallization process. GaN was successfully

dissolved in a molten LiCl medium, and then recrystallized on a nominally homoepitaxial

MOCVD-grown GaN template. This provides the basis for a potentially viable method to

fabricate high quality GaN bulk crystals for homoepitaxial growth.














CHAPTER 1
INTRODUCTION

Background and Motivation

Gallium nitride (GaN) and III-nitride alloys (Al-, Ga-, In-N) demonstrate

outstanding potential to significantly advance solid-state electronic and optoelectronic

technologies. Currently, solid-state electronics comprise a multi-hundred billion-dollar

industry, and consist primarily of silicon (Si) and gallium arsenide (GaAs)-based devices.

These materials, however, are intrinsically limited for certain applications (namely,

high-power electronics and low-wavelength optoelectronics). Such applications have

far-reaching implications, with significant promise for commercial, military, and national

security purposes. III-nitride-based materials are ideal for these applications because of

their favorable intrinsic electronic, structural, and chemical characteristics. Because of

these properties, GaN and related alloys have recently been subject to intense research;

and scientists have made significant progress in advancing III-nitride technology.

Electronic Properties of III-Nitrides

Electronically, III-nitrides have a large, direct, tunable band gap, which makes

these materials ideally suited for many technologically important electronic and

optoelectronic applications. The band gap for III-nitrides ranges from 0.7 eV for indium

nitride (InN) [1] to 6.2 eV for aluminum nitride (A1N) [2]. Alloys throughout this entire

range are possible to achieve, depending on the composition of group III element (Al, Ga,

In). III-nitride materials' large band gap gives them distinct capabilities for use in

high-power electronic and optoelectronic applications. The tunable nature of the band









gap in III-nitrides provides the possibility for large band offsets, which enables greater

control over carrier confinement, and better performance in some devices, including

HEMTs and multi-quantum well (MQW) LEDs.

Chemical, Structural, and Physical Properties of III-Nitrides

III-nitrides offer distinct advantages because of their chemical, structural, and

physical properties. They are outstandingly hard, corrosion-resistant in most

environments, and remarkably stable at high temperatures. III-nitrides exhibit the

wurtzite crystal structure, which provides a significant spontaneous polarization field

along the c-axis, and a high piezoelectric polarization field [3-5]. Together, these fields

enable greater sheet charge densities in III-nitride heterostructures strained in

compression; thus, allowing greater capabilities for III-nitride based electronic devices.

III-nitride materials are also significantly more thermally conductive than their

III-arsenide counterparts and silicon, allowing for efficient heat removal. This saves

substantial resources by reducing the need for external cooling. Silicon-based devices

burn out at temperatures exceeding 1400C, necessitating the installation of powerful

cooling fans for silicon-based devices, for high-power applications. GaN-based devices

are operable in environments >3000C [3], so the high temperature tolerance and the

efficient heat removal of III-nitrides make this material system ideal for

high-temperature, high power applications.

Suitability of III-Nitrides in Optoelectronic Devices

III-nitrides represent one of few known material systems that allow efficient

emission of radiation in the green, blue, and deep ultraviolet range. Figure 1-1 shows the

band gap dependence on the lattice constant, and the range of achievable wavelengths for









III-nitride emitters. III-nitride LEDs and LDs are exceptionally tolerant of defects, as

they show efficient emission of radiation despite high dislocation densities (p = 10-1011

cm-2).

Because of the unique capabilities of III-nitride optoelectronic devices, they are

ideal for commercial, military, and national-defense applications. Commercially,

monochromatic III-nitride LEDs are used for illumination purposes in traffic signals and

outdoor displays. One monochromatic traffic light saves 800 kWh/year in power,

compared to a typical incandescent white bulb with a monochromatic shutter [6].

III-nitride-based UV VCSELs are also being developed to pump red-green-blue (RGB)

phosphors for significantly more efficient and cost-effective white-lighting purposes.

This could provide energy savings estimated at $100 billion annually. Furthermore, blue

LDs are used commercially low-cost, ultra-high-density optical data storage in digital

versatile disks (DVDs) [7] that can store >1 gigabit/cm2 of information [6]. Militarily,

III-nitride photodetectors are currently being researched for use in solar blind detectors

for missile warning systems and for malignant biological and chemical agent detection

[8,9].

Suitability of III-Nitrides in Electronic Devices

Table 1-1 shows critical electronic properties of several semiconductor materials.

The intrinsic electronic properties of GaN and AlGaN/GaN heterostructures allow

significantly more power output than any other known semiconductor material system.

The large, tunable band gap of III-nitrides makes these materials ideal for use in certain

electronic devices, such as high electron mobility transistors (HEMTs). Breakdown

voltage (Vb) goes as the square of the band-gap energy, giving GaN and AIN-based









electronic devices a distinct advantage over SiC (band gap, Eg = 3.2 eV) in terms of

power switching, enabling a device to achieve large power density at high frequencies

[3,7]. The theoretical maximum power output achievable is directly proportional to the

band gap to the fourth power (Pmax oc Eg4), and the saturation velocity squared (Pmax oc

Vs2). Because of the large band gap and saturation velocity, III-nitride based devices are

ideal for high-power applications.

For a HEMT device, the output power is shown in Equation 1-1, where Imax is the

maximum current possible for a particular device.


out b max (1-1)
8

Higher saturation currents are achievable with higher carrier concentrations, and the

polarization effects of AlGaN/GaN heterostructures provide outstandingly high sheet

densities in a two-dimensional electron gas (2DEG). Furthermore, the polarization fields

in GaN HEMT structures remove the need for external doping to achieve very high sheet

densities. This ability to operate without ionized impurities diminishes the problem of

gate leakage in Schottky diodes [5]. Finally, for operable high-power electronic devices,

heat dissipation is crucial, since such devices are thermally limited. The high thermal

conductivity of GaN enables these HEMT devices to operate more efficiently and more

cost-effectively [3].

These characteristics make III-nitrides ideal for use in high-power devices for

both commercial and military applications. Commercially, GaN-based HEMTs can be

used in cellular communications base stations, telecommunications, hybrid electric cars,

and switches for high-power electric grids. For military purposes, GaN-based HEMTs

are candidates as amplifiers for synthetic aperture radar (SAR), smart weapons, electronic









warfare, and surveillance. They will eventually replace bulky traveling wave tubes,

enabling significantly higher output power, with added benefits of reduced mass, volume,

cost, and cooling requirements.

Progress in III-Nitride Material Development

Performance of III-nitride based devices has improved by leaps and bounds in

recent years due to growth and processing of these materials becoming more mature. As

growth techniques have progressed, microstructural quality of III-nitride films used in

devices has improved; and performance of both electronic [10] and optoelectronic

devices has improved accordingly [11]. Optoelectronically, the first nitride-based blue

LEDs were fabricated in the 1970s, when the external efficiency hardly exceeded 0.1%.

External efficiencies improved with novel growth breakthroughs throughout the 1980s,

and reached 1.5% by 1992. As growth of III-nitride films continued to mature, this value

reached 20% by 2002 [6]. Further strides with III-nitride optoelectronic devices will be

made as microstructural quality of III-nitride films continues to improve. GaN-based

electronic devices have experienced similar improvements as III-nitride technology has

become more advanced. The earliest reported GaN HEMT state-of-the-art was 1.1

W/mm in 1995, and exceeded 10 W/mm by 2001 [10]. Since 2001, GaN-based

technology has progressed rapidly, and the current state-of-the-art GaN HEMT power

density is an astounding 32.2 W/mm [12]. Conversely, GaAs-based HEMTs have

attained merely 1 W/mm for power density, despite the fact that GaAs-based technology

is significantly better studied.









Availability of III-Nitride Based Devices

Considering these substantial advantages and the potential widespread use of

GaN-based devices, commercial devices based on GaN are currently in the research

stage. Comprehensive studies of III-nitride materials are underway, in order to better

exploit these materials' advantages, where considerable growth and processing issues are

being addressed. The most significant of these issues involves the effects of dislocations

on III-nitride thin films. Because there are no commercially available substrates that

allow for homoepitaxial growth of III-nitride thin films, these devices must be grown

heteroepitaxially on sapphire or silicon carbide, at high lattice mismatch. This causes

extremely high dislocation densities (p=10-1011 cm-2).

Statement of the Problem

Currently, there is much disagreement concerning the exact effects of dislocations

on III-nitride electrical properties and device performance. To advance the science of

III-nitrides, and consequently III-nitride based device performance, the precise effects of

dislocations on electronic properties must be determined. In other semiconductor

material systems (such as Si and GaAs), a dislocation density of 104 cm-2 renders a

device inoperable. III-nitride films with dislocation densities of-1011 cm2 have been

used in operable devices with surprisingly high efficiency. Because of this high tolerance

for defects, some researchers have postulated that dislocations are electrically inactive in

III-nitride materials [13]. Others, however, have indicated that dislocations have a

deleterious effect on device performance [14-17], and will significantly affect reliability

and scalability of these devices. Determination of the effects of dislocation density on









electronic properties would greatly aid optimization of III-nitride device performance,

especially as long as these devices are grown heteroepitaxially.

Outline

The remainder of this dissertation is structured as follows: Chapter 2 describes the

growth and characterization techniques used in this work. Chapter 3 illustrates one

technique in depth, based on x-ray diffraction (XRD) to determine dislocation density in

III-nitride thin films. This method develops a theoretical model to determine dislocation

density from XRD results; and allows for rapid, accurate feedback for dislocation density.

This is extremely helpful in optimization studies for metalorganic chemical vapor

deposition (MOCVD) III-nitride growth studies. Chapter 4 describes a study on the

effect of dislocation density on impurity incorporation in GaN films during MOCVD

growth. Impurities can dramatically affect electronic properties of GaN, by forming deep

trap states and acting as scattering centers. Understanding how dislocations affect

impurity incorporation is critical in optimizing MOCVD growth parameters for device

structures. Chapter 5 describes an investigation into a novel method to fabricate III-

nitride bulk crystals. Bulk III-nitride substrates would allow homoepitaxial growth of

III-nitride thin films, which would minimize dislocation density for device structures.

This would greatly enhance performance for devices for which dislocations have a

deleterious effect. Finally, Chapter 6 summarizes significant results and recommends

future work for these projects.











b.U 0.2
AIN

5.0 e ANitrides
4.5 (blue/UV)
4.0 0.3
4.0 Phosphides 0
3 Ga (yellow/red)
M' 3.0 0.4 j
S2.5 p0.5
0,
S1. Arsenides
e 1.5fr (near-IR) b 3
1.0 GaAs 1.0
Antimonides GaSb
(far/mid-IR) InAs
0.0 1 I. I. ..L .. .L L L . .. 5.0
4.2 4.4 4. 4.8 5.0 5.2 5.4 5.6 5.8 6. 6. 6.4 6.
Lattice Constant (Angstroms)



Figure 1-1. Band gap engineering diagram for III-V materials. III-nitrides are ideal
candidates for use as green, blue, and deep-UV light emitters.









Table 1-1. Electronic properties of semiconductor materials used in electronic device
applications.
Semiconductor Material Si GaAs SiC GaN
Band Gap (eV) 1.1 1.4 3.3 3.4
Saturated Electron Velocity (x107 cm/s) 1 2.1 2 2.7
Breakdown Field (MV/cm) 0.3 0.4 2 3.3
Thermal Conductivity (W/cm-K) 1.5 0.5 4.5 >1.7
Bulk Electron Mobility (cm2/V-s) 1500 8500 700 1200
Heterostructure None AlGaAs/InGaAs None AlGaN/GaN
2DEG Sheet Density (x1012 cm-2) NA <4 NA 20














CHAPTER 2
EXPERIMENTAL METHODS

Chapter 1 showed the technological importance of III-nitride-based technology

and identified the fundamental problem of high dislocation density. Chapter 2 describes

the experimental procedures involved in growth and characterization of III-nitride

materials used in these studies. This includes a description of the two different

film-growth techniques, specifically metalorganic chemical vapor deposition (MOCVD)

and molecular beam epitaxy (MBE), followed by a description of a novel bulk-growth

method for GaN. The characterization techniques used are also explained.

Growth Techniques

Both MOCVD and MBE were used to accomplish different goals for growing

films in this work. Growth of III-nitride materials is relatively new, because of the lack

of III-nitride lattice-matched substrates and the difficulty associated with incorporation of

nitrogen. For MOCVD, ammonia (NH3) is the most prevalent precursor; but its stability

makes it difficult to incorporate nitrogen, and requires high growth temperatures (usually

-10500C). For MBE growth of III-nitrides, the nitrogen source is a reactive nitrogen

plasma, which consists of nitrogen gas and an electron cyclotron resonance (ECR)

plasma source.

Metalorganic Chemical Vapor Deposition

MOCVD offers the following advantages over MBE: high growth rates, high

throughput, ability to grow at high temperatures, external feeds for precursors, high









growth pressures, and lower costs associated with this technique over MBE. Higher

growth rates decrease the time required to perform material optimization studies, and

higher throughput is desirable when fabricating commercial devices, once the

material-optimization growth studies are complete. Growth at higher temperatures

enables more effective cracking of nitrogen precursor to incorporate nitrogen into the

film; which, in turn, results in higher growth rates. The external feed lines that provide

the precursors are advantageous, since they allow the growth chamber to be isolated

when replacing the sources on depletion. This is a significant advantage over MBE, since

replacing the solid sources requires venting the growth chamber to atmosphere. Baking

out the impurities from this exposure can take up to 2 weeks, whereas changing group III

gas bubblers or ammonia sources for MOCVD usually takes less than 1 hour.

The growth pressures associated with MOCVD are relatively high (usually 40 to

500 Torr for III-nitride growth) compared to those of MBE (<10-6 Torr). This removes

the need for ultra high vacuum (UHV) conditions, which require diffusion and cryogenic

pumps. Diffusion pumps are problematic because they are susceptible to back-streaming

oil into the growth chamber, which may incorporate into the film. Cryogenic pumps

require a constant flow of liquid nitrogen (L-N2), which can be expensive, depending on

the time and frequency of growth experiments. Finally, for all of the advantages that

MBE provides for devices in other material systems, MOCVD-grown III-nitride films for

electronic devices have demonstrated superior electronic properties and device

performance over their MBE grown counterparts [18].

The MOCVD growths were performed in an EMCORETM D-125 short jar reactor

with the wafers placed on a silicon carbide-coated graphite rotating platen, containing up









to three wafers. There are several advantages to this type of reactor. First, it allows for a

high degree of aluminum incorporation into films, since the short jar design significantly

reduces the parasitic pre-reactions of the aluminum precursor with ammonia due to

upstream heating. Second, the multi-wafer EMCORETM D-125 allows a greater throughput

of device-quality films once the growth conditions are optimized. Third, this reactor

allows for excellent thermal control and uniformity across the platen, and therefore across

the individual wafers. This is especially useful when growing a film that is temperature

sensitive. Fourth, the rotation rate of the platen can reach 2100 revolutions per minute

(rpm). This allows a high degree of control over the boundary layer thickness, which

could improve growth uniformity and growth rate. Growth uniformity can be monitored

in-situ, since the EMCORETM D-125 contains three view ports where it is possible to

monitor the film growth at different points across the wafers on the platen. Finally, this

reactor provides a load-lock, directly connected to a turbo-molecular pump (TMP)

backed by a dry-scroll rough pump. This allows the reaction chamber to remain under

vacuum during sample exchange, which reduces the amount of contamination during film

growth. Figure 2-1 is a schematic of the MOCVD operation.

The precursors used for the films are trimethyl gallium (TMG) and trimethyl

aluminum (TMA) for gallium and aluminum respectively, and ammonia (NH3) for

nitrogen, with hydrogen (H2) and nitrogen (N2) serving as carrier gases. The precursors

undergo the following reactions:

(CH3)3Ga(g) + NH3(g)4 GaN(s) + 3CH4(g)

(CH3)3Al(g) + NH3(g) + AIN(S) + 3CH4(g)









The solid III-nitride film deposits on the substrate, while the gaseous byproducts are

pulled through an EBARA rough pump, and out through the exhaust. In all likelihood,

the reactions rarely occur as shown above, as the reactions shown above are

oversimplified. Many side reactions and parasitic pre-reactions complicate the MOCVD

film deposition process; but overall, the reactions shown above are a useful

simplification.

The parameters that affect the growth of the film are temperature and pressure, as

well as flows of metalorganics, ammonia, and carrier gases. One challenge posed by the

MOCVD growth technique is determining which parameters) to vary, in order to grow

the film with the desired qualities. Since our goal was to determine the effects of

dislocations on electronic properties and device performance for III-nitride materials, it

was necessary to grow films over a wide range of dislocation densities by MOCVD. The

MOCVD growth parameters that were varied include the type of nucleation layer (A1N

vs. GaN), growth temperature, and growth pressure, in order to grow films of varying

dislocation density. To further expand this study, two different substrates were used:

silicon carbide (SiC) and sapphire (A1203). SiC has a less severe lattice mismatch than

sapphire with III-nitrides (3.5% vs. 16% for GaN and 13.29% vs. 1% for A1N) [19-21];

and therefore, enables experimenters to grow lower-dislocation-density-films. Sapphire

is more widely used, because it is significantly less expensive; insulating sapphire wafers

are $60 each, while device-grade SiC wafers are $5000 each.

For growths on SiC, a 1,400 A high-temperature (11250C) A1N nucleation layer

was first deposited, followed by a GaN buffer layer at 10500C. The pressure for these

growths varied from 100-200 Torr. For growths on sapphire, more variables were









explored. For some films, a low-temperature (6000C) A1N nucleation layer was first

deposited; while others had a low temperature (530C) GaN nucleation layer, depending

on the desired dislocation density. The pressure of the overlying GaN layers varied from

20-500 Torr, and the temperature of this layer varied from 1040-10600C. Another factor

that affected the film microstructure was recovery time. A fast recovery provided a film

with a higher dislocation density, while a longer delay between the growth of the

nucleation layer and the GaN over-layer provided a film with a higher microstructural

quality [22].

By altering the growth conditions as needed, it was possible to grow GaN films

with a very broad range of dislocation densities. The material grown for our study ranged

from 4.21x10 cm-2 1.07x1011 cm-2. This allowed for thorough studies of the effects of

microstructure on electronic and magnetic properties, as well as device performance and

impurity incorporation.

Molecular Beam Epitaxy

Molecular beam epitaxy (MBE) is advantageous when abrupt interfaces and

precise control of layer thickness are required. MBE is also an inherently safer process

when growing III-nitride films, since all precursors used in MOCVD are either toxic or

pyrophilic; whereas none of the precursors used in MBE present such hazards. This

removes the need for strict engineering safety controls and gas sensors that are required

for MOCVD.

For our study, MBE offered a distinct advantage when growing GaN films with

transition metal dopants and when growing oxide films. MBE is better suited for

transition-metal dopant growths, because of the difficulty associated with obtaining an









MOCVD transition-metal source that can incorporate efficiently into a III-nitride film

and that has a high enough vapor pressure to provide a substantial molar flux in the

carrier gas. MBE is also better for oxide growths, because of the lack of oxygen sources

for MOCVD.

For the MBE growths in our study, two series of films were grown: chromium

(Cr)-doped aluminum nitride A1N and scandium oxide (Sc203). For the Cr-doped A1N

films, the group III precursor was solid 99.99999% pure (7N) aluminum metal, the

dopant source was 5N solid Cr, and the nitrogen precursor was 6N nitrogen gas. To

create reactive nitrogen radicals, the nitrogen gas was ionized by a radiofrequency (RF)

plasma unit with a frequency of 13.56 MHz. For the scandium oxide growths, the

precursors were solid 4N scandium metal and 6N oxygen gas, which was ionized at 13.56

MHz. All of the solid sources were evaporated by resistively heating poly-boronitride

(PBN) containers, and the atomic fluxes were controlled by adjusting the shutter

apertures.

Gallium Nitride Bulk Growth

The GaN bulk growth studies were performed in a home-built stainless-steel

container, fitted for a 1" diameter quartz tube. The method to grow GaN bulk crystals

was simply a dissolution/recrystallization process of polycrystalline GaN powder in a

molten lithium chloride (LiC1) solvent. The chamber was resistively heated up to

-1000C to effectively dissolve GaN powder and maximize deposition of GaN

crystallites.

Procedurally, the solubility was tested with respect to temperature. Once it was

determined that GaN dissolved sufficiently well at achievable temperatures, the









experiment was repeated to test the recrystallization of GaN on a nominally

homoepitaxial template of MOCVD-grown GaN on sapphire substrate. Other effects

such as convective fluid dynamics and surface treatment were tested to maximize growth

on the surface.

Characterization Techniques

After the MOCVD, MBE, and bulk growths, samples were removed from their

respective growth chambers; and their properties were assessed using several techniques.

Specifically, the microstructural, magnetic, and electronic characteristics were

determined. All of the techniques used are described in the subsections to follow.

High-Resolution X-Ray Diffraction

High-Resolution X-Ray Diffraction (HRXRD) is a powerful technique that

provides information about a material's microstructure and lattice constant. Our study

used only rocking curves (co-scans) to ascertain the microstructure of the GaN thin films.

The microstructural quality of a thin film is closely related to the width of the diffracted

x-ray intensity, with the full-width at half maximum (FWHM) used as the figure of merit.

Since HRXRD is a diffraction phenomenon, its measurements must be taken with respect

to reciprocal space, where reciprocal lattice points (RLPs) correspond to specific lattice

planes. If it were possible to perform HRXRD analysis on a defect-free crystal of infinite

dimensions, the results would be a series of delta functions at every RLP. Once finite

dimensions and defects are imposed on the crystal, broadening can be observed and

measured by HRXRD. The extent to which broadening is observed is indicative of the

microstructural quality; the higher the measured FWHM, the higher the dislocation

density.









All of the HRXRD measurements were made using a Philips (Almelo, Holland)

X'Pert MRD high-resolution triple-axis x-ray diffractometer with a sealed Cu anode

collimated by a four-crystal Ge(220) monochromator. The Cu anode produced an x-ray

wavelength of k(Cuk~1)=1.5404 A.

Transmission Electron Microscopy

Transmission Electron Microscopy (TEM) is a universally accepted method to

determine dislocation density in materials. TEM operates by emitting a focused electron

beam through a sample, into a series of magnetic lenses, and into a fluorescent screen

detector. It is a robust technique for determining microstructural characteristics of

materials, since it provides high spacial resolution (with a magnification of 106x). Using

this technique, dislocation density is simply computed as the number of dislocations

counted over a known measured area. TEM is a destructive technique, since sample

preparation requires a thickness that is transparent to the electron beam (<3000 A).

The microscopy was performed with a Philips CM20 instrument at 200kV. Some

specimens were prepared in plan-view while others were prepared in cross section. The

samples observed in cross-section were prepared by a focused ion beam; while those

observed in plan-view were prepared by a traditional mechanical polishing, dimple, and

ion mill procedure.









Secondary Ion Mass Spectroscopy

Secondary Ion Mass Spectroscopy (SIMS) is a superior method to determine

impurity concentrations with respect to film depth. It is a destructive technique, but is

known to provide accurate results for impurity profiles. SIMS results are obtained by

accelerating heavy metal ions at a target sample, sputtering off constituent atoms. These

atoms from the target sample become ionized, and are attracted to a charged detector.

The SIMS work in our study was performed in a PHI Quadrupole SIMS

spectrometer (Charles Evans and Associates, Cambridge, MA). The sputtering ion was

cesium (Cs ), with a primary ion energy of 4keV. The cesium angle of incidence was

600, and the instrument was optimized for sensitivity for carbon, oxygen, and silicon.

Detection limits were 1x1016 cm-3 for carbon, 5x1015 cm-3 for oxygen, and 6x1015 cm-3

for silicon.

Superconducting Quantum Interference Device Magnetometry

A Superconducting Quantum Interference Device Magnetometry (SQUID)

Quantum Design Magnetic Properties Measurement System was used to ascertain

magnetic properties of DMS materials. Hysteresis loops from magnetization vs. field

measurements verify that a sample is ferromagnetic. The second technique to determine

magnetic properties was Field-Cooled/Zero Field-Cooled (FC/ZFC) measurement. This

technique involves cooling a sample to 10K under an applied field while the sample's

magnetization is measured (FC), and again while the sample's temperature increases to

room temperature (ZFC).












Capacitance-Voltage

Capacitance-voltage (CV) is a non-destructive technique that provides

information about a film's electrical properties. In our study, a mercury (Hg) contact CV

was used to probe for a buried conductive layer in a nominally insulating GaN film, and

to confirm the existence of a conductive 2DEG in a AlGaN/GaN HEMT structure. The

depletion region of a HEMT structure is modeled as a parallel plate capacitor, where the

mercury droplet acts as a metal Schottky contact, and the sheet density of a 2DEG is

given as


N(x)-C3 (2-1)
qsA' AVJ


where C is the capacitance, q is the carrier charge, ; is the permittivity, A is the area of

the mercury drop, and Vr is the reverse bias voltage. The apparatus used for

experimentation was an MSI Electronic Model 412-2L mercury probe, with a droplet

diameter of 2.03x103 cm-2. Electronic measurements were obtained by a Hewlett

Packard 4284A Precision Inductance-Capacitance-Resistance meter, operating at 1 MHz.

Inductively Coupled Plasma-Mass Spectroscopy

Inductively coupled plasma-mass spectroscopy (ICP-MS) is a powerful technique

to determine the existence and concentrations of trace elements. This technique operates

by injecting an aerosolized liquid sample into an argon plasma at a temperature of 7000

K. When the atoms from the sample collide with the energetic argon ions, the become

charged. The ions are then drawn into a high vacuum chamber with a quadrupole mass

analyzer. The quadrupole filters out atoms of undesired masses by alternating









radiofrequency (RF) and direct current (DC) fields, so that only atoms of the desired

mass can pass into the detector. The detector is calibrated by correlating the intensity of

the signal created by these collisions with the detector to a set of standards with known

concentrations. In this study, a Perkin Elmer (Boston, MA) Elan 6100 instrument was

used to determine the concentration of gallium and lithium in samples of GaN dissolved

in LiCl.

Scanning Electron Microscopy

Scanning electron microscopy (SEM) is a useful technique to determine

morphological information on a surface at extremely high magnifications (10-300,000x).

SEM operates by emitting an electron beam at the sample surface in a vacuum. When

this electron beam hits the surface, secondary electrons are emitted and collected in a

detector, which forms an image, giving precise topographical information. The

instrument used was a JEOL (Peabody, MA) JSM5800 tungsten filament scanning

electron microscope. The electron beam energies ranged from 5-20 keV.

Electron Dispersive X-Ray Spectroscopy

Electron dispersive x-ray spectroscopy (EDS) is used in conjunction with SEM to

provide information about elemental analysis on surface features of a sample. Other than

emitting secondary electrons when exposed to an incident electron beam, a sample will

also emit x-rays. This is because, when inner shell electrons are ejected from an atom,

electrons at high energy shells will fill the inner shell vacancy, which emits a photon.

The energy of the photon emitted is related to the element. EDS measures the energy and

intensity of the emitted radiation, which provides qualitative and quantitative analysis of









surface impurities. The EDS data in this study was obtained using an EDAX Phoenix

system, fitted with a 3,000 A window, allowing for analysis of low-Z elements.

Auger Electron Spectroscopy

Auger Electron Spectroscopy (AES) was employed for surface elemental analysis.

AES operates by emitting an incident electron beam at a sample, which causes surface

atoms to emit electrons of an energy that is characteristic of the atom. This technique

operates in the following manner: an incident electron knocks off an inner shell electron

on the sample surface, which causes an electron of higher energy to fall into this inner

shell. This causes emission of a photon, which is reabsorbed by the surface atom, and

causes an outer shell electron to be emitted. It is this emitted electron that is detected,

and its characteristic energy is indicative of its atom. The Auger instrument used in this

research was a Perkin Elmer 6600, that is capable of providing compositional data for

elemental analysis for samples of 1 atomic %.


































Figure 2-1. Schematic of metalorganic chemical vapor deposition operation. Cold
precursors and carrier gases flow over a heated substrate. Precursor gases
undergo a heterogeneous reaction at the substrate surface, depositing a solid
film, while gaseous byproducts are pumped out through the system exhaust.














CHAPTER 3
DETERMINATION OF DISLOCATION DENSITY OF III-NITRIDE FILMS BY
X-RAY DIFFRACTION

Introduction

There is currently no universally accepted method that allows for rapid and

accurate determination of threading dislocation density in III-nitride thin films. X-ray

diffraction is a promising technique to determine dislocation density because it provides

results quickly and nondestructively, and when modeled correctly, allows for accurate

determination of microstructural parameters of crystalline materials. The goal of this

chapter is to develop a method to determine dislocation density based on x-ray diffraction

that provides accurate results as quickly as possible. Such a method would facilitate the

establishment of a correlation between dislocation density and film properties. This

information, in turn, would allow experimenters to determine how dislocations influence

electrical and optical properties of device layers.

Characterization of Dislocations

Several characterization techniques have been employed to ascertain dislocation

density in III-nitride films, including transmission electron microscopy (TEM), atomic

force microscopy (AFM), cathodoluminescence (CL), and high-resolution x-ray

diffraction (HRXRD). TEM is generally accepted as the most accurate and best

understood method, but the sample preparation is time consuming and destructive, and

some care must be used in how the dislocations are imaged and counted [23]. AFM has

also been used, with the dislocations identified as slight depressions near step









terminations, and these depressions have been correlated to dislocations with a screw

component [24]. Dislocations with a pure edge component are occasionally observed on

terraces; however, the large tip radius usually precludes direct observation. Using CL

imaging, dark spots are observed on the GaN surface and have been correlated to the

dislocation density in TEM samples [25]. In CL, the dislocations that act as non-radiative

recombination are imaged as dark spots while dislocations that radiate may have the same

intensity as the background GaN. X-ray diffraction provides quick results that are

correlated to the microstructure of thin films, and the microstructure is directly related to

dislocation density. Because of the limitations associated with TEM, AFM, and CL,

HRXRD is an attractive alternative to ascertain dislocation density for III-nitride thin

films.

Previous Work

Researchers began using XRD to study microstructures in metals as early as the

1950s [26-29]. By comparing the angular distribution of x-ray intensity for a given

reflection of cold-worked vs. annealed specimens (i.e., those with high and low

dislocation densities, respectively), a correlation emerged between the XRD line width

and the dislocation density. Several models were proposed to calculate dislocation

density based on the degree of diffracted x-ray scattering. These studies provided a

valuable theoretical and experimental foundation for determination of dislocation density

by XRD.

In the 1990s, these classic works were applied to heteroepitaxial GaN-based thin

films [30-34]. These modern works accounted for line width broadening due to strain,

coherence length, rotations of crystallographic domains, and interdependence of these









rotations; however, there were disagreements concerning the relative importance of each

of these broadening effects when estimating threading dislocation density. As a result,

there is currently no unified analytical model from which to determine dislocation density

based on XRD line width results. This work resolves these differences by providing

insight into each of the significant broadening mechanisms, concentrating heavily on

twist and incoherence.

Twist

When ascertaining the crystalline quality of III-nitride films by XRD, it is crucial

to determine the microstructural twist. Threading edge dislocations (b = (11-20)) are

often dominant in heteroepitaxial III-nitride films, and these are directly related to

crystallographic twist. Recognizing this, Heying et al. [35] explicitly demonstrated the

need for asymmetric scans in skew geometry in order to better characterize the

dislocation character of GaN films. Several models were subsequently proposed on how

to best determine twist in GaN films. Metzger et al.[30] and Kang et al.[31] proposed

similar procedures that employ a 4-scan in order to determine twist. However, it is not

possible to obtain accurate results by this method with most XRD systems because the

instrumental out-of-plane broadening of a (-scan is -1.50, which corresponds to a

dislocation density on the order of 1011 cm2, thereby grossly overestimating the edge

component of dislocation density based on twist.

Srikant et al. [32] proposed a model that employs asymmetric rocking curves

(co-scans), and accounts for tilt and twist dependence on both inclination angle (2) and

measured full width at half maximum (FWHM). By treating twist as a free parameter,

several scans at different angles of inclination are required for this method, with twist









being determined by optimizing the fit to the FWHM data. This model was also the basis

of the work of Chierchia et al. [33] and Sun et al. [34] to determine twist. One significant

disadvantage with this method is the necessity of performing so many scans at

asymmetric reflections, lengthening the measurement and analysis time.

Incoherence

Another phenomenon often neglected when determining dislocation density by

XRD is the effect of incoherence broadening on the measured FWHM. The model

proposed by Srikant et al. only considers tilt and twist contributions to broadening [32].

Chierchia et al. discussed broadening due to incoherence for symmetric scans, but did not

explicitly include it when calculating dislocation density [33]. Both Chierchia et al. and

Sun et al. [34] stated that their calculations of twist were overestimations because they

neglected coherence length broadening upon determining twist [33]. The degree to which

incoherence broadening affects the calculation of dislocation density, however, was never

described in detail.

Focus

In this chapter, a theoretical and experimental approach is presented for using

HRXRD to measure threading dislocation densities of III-nitride thin films in a rapid,

non-destructive manner. This method relates HRXRD line widths to dislocation density,

allowing resolution of screw and edge dislocation densities from the total dislocation

content. The model developed in this work unifies and simplifies other proposed models

and accounts for all significant microstructural causes for broadening. This model was

verified by comparing the dislocation density results obtained by HRXRD to those

obtained by TEM on III-nitride films grown by metalorganic chemical vapor deposition









(MOCVD) on both SiC and sapphire substrates. The model was further simplified by

carefully considering the magnitudes of the contributions of each broadening term,

leading to accurate dislocation density determinations from as few as two HRXRD

reflections.

Experimental Methods

To obtain the dislocation density from the HRXRD line widths, the individual

broadening effects from tilt, twist, and lateral coherence length were deconvolved from a

series of symmetric ((0002), (0004), (0006)) and skew-symmetric ((10-11), (20-22))

rocking curves. These reciprocal lattice points are shown in Figure 3-1. Crystallographic

tilt is defined as the out-of-plane rotation of adjacent domains in the crystal, as

demonstrated in Figure 3-2, and could be determined directly from the symmetric series

of scans. The asymmetric reciprocal lattice points were chosen based on the significant

twist component and the relatively high measured intensity associated with them. Twist

is defined as the in-plane rotation of adjacent domains (Figure 3-3).

Measurements were made using a Philips X'Pert MRD high-resolution triple-axis

x-ray diffractometer with a sealed Cu anode collimated by a four-crystal Ge(220)

monochromator. The Cu anode produced an x-ray wavelength of k(Cuk l)=1.5404 A. Of

the ten samples studied by HRXRD, seven were selected for comparison with TEM over

the widest range of dislocation density. The microscopy was performed with a Philips

CM20 instrument at 200kV. One specimen was prepared in plan view while the other six

were prepared in cross section. Three were prepared by focused ion beam, while the

balance was prepared by a traditional mechanical polishing, dimple, and ion mill









procedure. Cross-section TEM (XTEM) images of low and high dislocation density

samples are shown in Figure 3-4 and Figure 3-5, respectively.

To determine dislocation densities of the samples in cross section, dislocations

were imaged and counted as usual, under two-beam bright field, dark field, and weak

beam (Figure 3-6a) conditions and counted as usual. Specimen thicknesses were

determined by convergent beam electron diffraction (CBED) using a (11-20) two-beam

condition (Figure 3-6b). The number of parallel diffraction intensity oscillations

("fringes") observed in the CBED disks is related to the extinction distance g and the

specimen thickness t. Cross sectional thicknesses were assigned based on an average of

the first two linear fits of the fringe spacings according to the formula



S,2 + 1
11 2 7 211 t2

where s, is the fringe spacing of the ith fringe, and n is an integer. For more detailed

information on this technique, see reference 15.

The samples were grown by MOCVD in an EMCORETM D-125 reactor on both

silicon carbide and (0001) sapphire substrates under a variety of growth conditions in

order to vary the dislocation densities; the growth temperature for all films was 10500C,

and the precursors were trimethylgallium (TMG), trimethylaluminum (TMA), and

ammonia (NH3) with hydrogen (H2) and nitrogen (N2) carrier gases. The pressure was

varied from 140 300 Torr. All samples analyzed were either GaN or AlxGa(l-x)N, with

aluminum content as high as 45%.









Method

There are several challenges posed by employing HRXRD to determine threading

dislocation density. These involve the following:

Identifying the significant sources of line width broadening.
Formulating a model that accounts for these broadening effects.
Establishing a procedure using XRD scans to probe the film layers, using the
proper geometry.
Determining the extent to which microstructural defects influence measured
HRXRD line width.
calculating dislocation density from this information.

Computation of dislocation density is important for comparison studies between

materials. Microstructural quality of III-nitride thin films is often reported by quoting

the FWHM values of select reflections. However, a simple comparison of line widths for

particular XRD scans may be insufficient when comparing crystalline quality of thin

films. There are two reasons for this: first, dissimilar XRD systems may exhibit

significantly different intrinsic broadening, which must be accounted for. Second,

research groups may employ any of a variety of reflections to ascertain crystalline quality

of thin films, and these reflections are affected differently by microstructural

imperfections. Since there is no one universally employed XRD reflection, calculating

dislocation density with a microstructural model is the only way to compare different

samples conclusively.

Sources of Broadening

When performing HRXRD rocking curves on single crystal materials, the

potential sources of broadening are due to:

Crystallographic rotations.
Inhomogeneous strain fields.
Curvature of the film.
Intrinsic rocking curve width.









Broadening due to the monochromator.

For the purposes of this study, broadening effects due to film curvature, intrinsic rocking

curve width and the monochromator are neglected, since all of these effects are less than

12 arc sec. [36]

Figure 3-7 demonstrates specific broadening effects in reciprocal space and how

the diffractometer elucidates each of these broadening effects. Figure 3-7(a) is a view of

the film in cross-section with the relevant broadening effects at an asymmetric reciprocal

lattice point (RLP) in skew geometry. Khkl is the reciprocal lattice vector between the

origin of reciprocal space and the RLP being probed. Its direction is related to the angle

between the surface normal of the crystal and the normal of the set of diffracting planes,

while its length is related to the d-spacing of the set of planes associated with the RLP.

Rotations of crystallographic domains broaden the RLP transverse to Khkl, while

inhomogeneous strain broadens in the direction parallel to Khkl. Broadening due to

incoherence occurs both parallel as well as transverse to Khkl.

Specific broadening effects due to crystallographic rotations and incoherence are

demonstrated in Figure 3-7(a). Pure domain tilt and twist are observed about the Ky and

Kz axis, respectively. The effects of tilt and twist measured by the diffractometer are the

projections of the rotations about Ky and Kz, respectively, on the RLP. The extent to

which each of these rotational broadening contributions affects the measured line width

of a rocking curve (co-scan) is geometrically related by Xhkl, the angle between Kz and the

RLP being probed. Figure 3-7(a) also demonstrates broadening due to incoherence, as

domain thickness (h) and lateral coherence length (L) are shown explicitly. Because

threading dislocations propagate along the growth plane normal during III-nitride growth,









the domain thickness is assumed to be equal to the film thickness. The lateral coherence

length is defined as the average distance between threading dislocations.

The tilt, twist, and lateral coherence length all affect broadening of an asymmetric

RLP when a rocking curve is performed in skew geometry. Since the film thicknesses in

this study were >8000 A, vertical coherence length had an insignificant effect on

measured broadening. Furthermore, inhomogeneous strain broadening was neglected

because all scans employed in this study were triple axis rocking curves, which diminish

the observed broadening effects in the direction parallel to Khkl. Therefore, the

microstructural imperfections that significantly contributed to our measured values for

FWHM were projections of rotations about both Ky and Kz (tilt and twist) on the probed

RLP and lateral incoherence (27T/L).

Figure 3-7(b) shows broadening effects of the skew-symmetric RLP in plan-view.

The measured broadening (Fhkl), is the measured FWHM and broadening due to tilt, twist

and incoherence are indicated explicitly. These microstructural broadening effects were

included in the proposed model that was used to calculate dislocation density.

Model

The following reciprocal space model describes the (hkl) dependence of the

convolved peak-width contributions from the combined effects of tilt (Frlt), twist (Ftwst)

and coherence length (L) broadening:


Frh (Fr tt COS Xhki ) + (rtF sin hkI) + 2; (3-1)
| K hki L









The geometric dependence of tilt and twist on measured line width is apparent, and the

observed broadening due to incoherence is inversely proportional to the magnitude of the

reciprocal lattice vector. The exponent, n, denotes the type of convolution of the model.

Determination of Convolution

One major discrepancy among previous works is the type of convolution, i.e., the

value of n is used in the model (Equation 3-1). The choice of n is derived from the

best-fit function to the intensity distribution [32]. For example, for a Lorentzian

convolution, n=1 and for a Gaussian convolution, n=2. For symmetric scans, both

Metzger et al. [30] and Chierchia et al. [33] assumed a Lorentzian convolution, where tilt

and symmetric coherence length were determined from a Hall-Williamson plot.

The results obtained in this study indicated that the lateral coherence length value

was unrealistically high when assuming a Lorentzian convolution, and a more plausible

value for coherence length was obtained when a Gaussian convolution was used. For

example, a film with a screw dislocation density of 7.7x108 cm-2 measured from TEM

gave the following results: assuming a Lorentzian convolution, the lateral coherence

length was -31,000 A, indicating a dislocation density, p -1x107 cm-2. For the same

sample, assuming a Gaussian convolution, the lateral coherence length was 5,000 A,

indicating a dislocation density, p -4x10 cm-2. Similar results were reported by

Chierchia et al.[33]. Therefore, a Gaussian convolution (n=2) was used in our analyses

with Eq. 1, which is similar to the models of Hordon and Averbach [26] and Ayers [36].

Figure 3-8 shows FWHM data vs. the magnitude of the reciprocal lattice vector

(I/lhkl) for a typical sample assuming a Gaussian convolution. Figure 3-9 shows the

dependence of the FWIHM on angle of inclination (X) assuming a Gaussian convolution









for one sample of GaN on sapphire. The Gaussian assumption of the model fits the

FWHM vs. data well, and explicitly demonstrates that setting the exponent, n=2,

provides accurate results.

Calculation of Dislocation Density

With the proposed model, two series of scans at different angles were required

to deconvolve each individual broadening effect on the experimentally determined

FWHM. This deconvolution procedure involves fitting the FWHM data to the model, as

shown in Figure 3-8, and is described in detail in the Results and Discussion section.

From the experimentally determined values for tilt, twist, and coherence length (both

symmetric and asymmetric), two independent methods were used to calculate dislocation

density.

The first method uses the tilt and twist values with the classic formulation of

Dunn and Koch [27], where the density of dislocations, p, is given by p=F2/4.36b2

Figure 3-9 shows the mathematical and geometrical derivation upon which this method is

based. Here, F is the calculated rotational broadening effect (tilt or twist) and b is the

applicable Burger's vector. With this formula, it is possible to calculate both the screw

and edge dislocation density with the deconvolved values for tilt and twist, respectively,

using the appropriate Burger's vector. It is important to note that while the coherence

length contribution was not used explicitly in calculating dislocation density by this

method, its broadening effect was accounted for in Equation 3-1.

The second method employed to compute dislocation density was that proposed

by Hordon and Averbach [26]. This method assumes that the measured value for

coherence length equals the root mean square spacing of randomly spaced threads, where









the dislocation density, p=1/L2, as shown in Figure 3-10. In order to compute the total

dislocation density, the value of coherence length was L(lo-11), the coherence length due to

all dislocations (i.e. the coherence length as determined by asymmetric scans in skew

geometry, since these are sensitive to both tilt and twist). In order to calculate screw

dislocation density, the symmetric coherence length, L(ooo0) was input for the value of L.

The edge dislocation density could be resolved by taking the difference between the total

threading dislocation density and screw dislocation density.

Simplified Model

It is possible to simplify Eq. 1 by neglecting lateral coherence length broadening

in order to determine dislocation density more rapidly:

2l = (tt COS zhk )2 + (Fr t sin hl )2 (3-2)

This modification makes it possible to determine the relative tilt and twist of the crystal

using only one symmetric scan (e.g., (0004)) and one asymmetric scan (e.g., (10-11)) in

skew geometry. Eq. 2 provides an upper bound for rotational broadening effects and a

concomitant overestimation of dislocation density. However, Eq. 2 is useful for

providing more rapid results to ascertain relative crystalline quality of III-nitride thin

films. For example, when characterizing an 8000 A thick GaN film with a dislocation
9 -2
density -109 cm-2, this two-scan procedure can provide results within an hour, as opposed

to >3 hours for the full five-scan procedure required when incoherence broadening is

included in the analysis. This is a considerable improvement when rapid results are

desired for growth optimization studies.

Comparing dislocation density results determined from the full five-scan

procedure to results obtained using this condensed model, the abridged procedure









provided results to within 15%. In all samples studied, the spacing of dislocations was

assumed to be random rather than piled-up. A piled-up distribution of dislocations may

affect the incoherence broadening term more significantly and must be accounted for, as

the value for coherence length may decrease significantly for such a distribution.

Furthermore, when the dislocations are piled up, the calculation of dislocation density

must be performed using a different method, the details of which are described in

Reference 8.

Results and Discussion

For each sample, the individual broadening effects were deconvolved using

Equation 3-1 with the measured FWHM data. Figure 3-8 demonstrates a fit of the model

to the FWHM data. From the fit of the model to the data, the magnitude of each

broadening contribution was determined. The three symmetric scans were employed to

determine tilt and symmetric coherence length, as twist does not affect these data. Once

this value for tilt was established, twist and asymmetric coherence length were next

determined by fitting the model to the FWHM values of the skew-symmetric reflections.

It is important to note that coherence length for a symmetric family of planes may be

significantly different than that of an asymmetric family of planes.

Two independent methods were used to calculate dislocation density using these

values for tilt, twist, and coherence length (both symmetric and asymmetric). These

included the formulation of Dunn and Koch [27] (p=F2/4.36b2) and that proposed by

Hordon and Averbach [26] (p=1/L2). Figure 3-12 shows the agreement between the two

methods within a factor of 3. However, the formulation of Dunn and Koch was more

reliable when compared to TEM results. This was expected since the coherence length









broadening effect was slight compared to those of crystallographic rotations, and

therefore, had a greater intrinsic error.

Figure 3-13 shows dislocation densities determined by HRXRD, using the

formulation of Dunn and Koch compared to those determined by TEM. The values for

dislocation density ranged from 5.4x10 cm-2 to 7.9x109 cm-2. All of the values for

dislocation density determined by HRXRD were obtained using Eq. 1 and the full

five-scan procedure, for which incoherence broadening was accounted. The values for

dislocation density determined by HRXRD agreed with those determined by TEM within

a factor of 1.5 for all seven samples.

With confidence that the XRD-based method described in this chapter is accurate

with respect to TEM, it served as a basis for all subsequent growth optimization studies

described throughout this dissertation with respect to dislocation density. The first such

optimization study observed the effect of growth pressure on dislocation density. Figure

3-14 demonstrates these results, and exhibits that increasing the growth pressure causes a

decrease in dislocation density. Performing this study on so many samples using TEM

would be impractical because of the time and cost associated with TEM sample

preparation. Such studies can only be performed by a non-destructive, rapid technique

such as XRD.

Ultimately, III-nitride researchers seek to conclusively determine the effect of

dislocations on device performance. Using the XRD-based method described above to

determine dislocation density, the effect of dislocations on HEMT performance was

ascertained with respect to electronic properties. For the purposes of this experiment, the

HEMT performance assumed to be proportional to the sheet density-mobility product (ns









x [t), widely used as the figure of merit for GaN HEMT device performance [38]. In a

series of experiments, electronic properties obtained by Hall measurements were

compared to dislocation density. These results are shown in Figure 3-15. While

dislocations appear to inhibit room temperature mobility for a given range of sheet

density, no trend could be concluded. The reason is that the three samples of highest

dislocation density were grown early in the HEMT growth optimization process, and

other room temperature scattering mechanisms are believed to have had a stronger effect

on these samples compared to those that exhibit higher mobility. These scattering

mechanisms include interface roughness and alloy disorder. As the HEMT growths were

optimized, higher quality interfaces with fewer variations in alloy disorder from the

AlGaN barrier layer were most likely the cause for the improved mobility. Further study

is necessary to conclusively determine the precise effects of dislocations on room

temperature mobility.

Summary

This chapter describes a method to determine dislocation density from measured

HRXRD line widths. A geometrically derived reciprocal space-based model was

developed that allows for the determination of individual microstructural broadening

effects in III-nitride thin films. This model unifies previous works that employ x-ray

diffraction to calculate dislocation density, but demonstrates distinct advantages in terms

of simplicity, accuracy, and time efficiency. Realizing the microstructural tilt, twist, and

coherence length, dislocation density was calculated by two separate, independent

methods, which agreed within a factor of 3. The model was simplified to neglect

broadening due to incoherence, which provided an upper bound for dislocation density









from only two scans. Finally, the validity of this model was tested by comparing

dislocation density results to those obtained by TEM, and agreement within a factor of

1.5 was found. These results show that the HRXRD-based method to determine

dislocation density presented here is accurate, as demonstrated by the agreement of our

XRD-based values for dislocation density with TEM, and time-efficient, as results were

obtained within an hour. Thus, this HRXRD-based method to determine dislocation

density may provide researchers with a valuable tool for optimizing growth conditions for

heteroepitaxial III-nitride films. The method demonstrated in this chapter serves as a

basis for the following chapters of this dissertation, since rapid and accurate feedback are

required to perform intensive surveys on the effects of dislocations on electronic

properties and device performance of GaN thin films.














(0006)




(20-22)

(0002) (22
X(o01)
(10-11)


KII

Figure 3-1. Schematic of reciprocal lattice points probed. Procedurally, the reciprocal
lattice points probed were symmetric (0001) and asymmetric (10-11). Together,
these data could be used to determine the tilt, twist, and coherence length in
the crystal.






40










Figure 3-2. Adjacent domains tilted with respect to one other. The axis of rotation is
orthogonal to the c-axis.






41












Figure 3-3. Adjacent domains twisted with respect to one another. The axis of rotation is
orthogonal to the plane of the domain, parallel to the c-axis.






42








Figure 3-4. Cross-section transmission electron microscopy image of a low dislocation
density sample. The dislocation density for this sample is 2.1x109 cm-2







43










Figure 3-5. Cross-section transmission electron microscopy image of a high dislocation
density sample. The calculated dislocation density for this sample is 6.8x109
c -2
cm



















































Figure 3-6. Weak beam transmission electron microscopy image and convergent electron
beam diffraction image. A) A sample weak beam TEM image of a film with a
dislocation density of 2.59x109 cm-2. B) Convergent beam electron
diffraction image showing the oscillations in diffracted intensity. Disks are
the (000) and (11-20) reflections.


: : Y: : 777 7:: 7 V P I, Ac:::::: =


" ':I'"~~~~'~:':~.'_'':'~'"' 7 "T: :;


..... .....
.. ..
.........











Ewald
sphere \ s


\ I Ir I
/ ~'d K etector-
aperture
/ integration

2u r 1

S / bbroadening due to
(b) K rotations about K, and K,
(D) K \K

Domain
S twist broadening due to
rotations about K,
4 inmanin

'-- .....--
J Li





((hkl) planes/ domain tilt


Figure 3-7. Broadening effects demonstrated in reciprocal space. In order to accurately
determine twist, asymmetric rocking curves in skew geometry are necessary.
In (a), the film is shown in cross section. The crystallographic rotations affect
measured broadening proportional to their projections of their respective axes
onto the reciprocal lattice point being probed. Broadening due to incoherence
is proportional to 27T/L. In (b), the measured broadening effects are shown in
plan view. These include tilt, twist, and incoherence.







46



-Symmetric data
3 0.290
0.290 -m -Skew-symmetric data

S0.250

0.210

0.170
2 4 6 8
IKI (1/A)


Figure 3-8. Plotted data of full width at half maximum vs. reciprocal lattice vector
magnitude. As |Khkl increases, incoherence broadening diminishes. The
curvature of the plot as KIhkl approaches 0 indicates the coherence length
contribution on the measured FWHM.











0.15

0.13


' 0.11

M 0.09

LL
0.07

0.05


0 10 20 30 40 50 60 70 80 90
Chi (degrees)


Figure 3-9. Fit of full width at half maximum data vs. angle of inclination. The data
shown are for one sample of GaN on sapphire at every observable reciprocal
lattice point. The fit of the data assumes a Gaussian convolution without
accounting for broadening due to incoherence. The data fit well, and
demonstrate that a Gaussian convolution is acceptable. Most of the data lie
above the fit due to the neglecting of incoherence broadening.






48


-2-T In 2 b l

L

Figure 3-10. Mathematical and geometric derivation of the random distribution method
for calculating dislocation density. The coherence length, L, and the
rotational broadening effect, F, are indicated explicitly. The opposite side of
this right triangle was derived to be a constant multiplied by the appropriate
Burger's vector.







49





LL
LL
L


Figure 3-11. Calculation of dislocation density based on values for coherence length.
Since coherence length is defined as the average distance between
dislocations, dislocation density is simply calculated as p=/L2.













: 9.9





0 0 9.1

0
-j
8.7
8.7 9.1 9.5 9.9
Log Dislocation Density (Dunn/Koch)


Figure 3-12. Comparison of methods of determining dislocation density by XRD. The
agreement shown here demonstrates internal consistency from the XRD-based
methods to determine dislocation density.













c 9.0E+09
E



o 6.0E+09

o
o
3.0E+09


I-
0.0E+0
O.OE+00 3.0E+09 6.0E+09 9.0E+09
XRD Dislocation Density (cm"2)


Figure 3-13. Dislocation density determined by transmission electron microscopy vs.
dislocation density determined by x-ray diffraction. All values determined by
XRD agreed with TEM within a factor of 1.5.











3.0E+09
OGaN 150torr
E GaN 10Otorr
2.5E+09
: C

o 2.0E+09 -
o 1

8 1.5E+09 -


1.0E+09
50 100 150 200
Pressure (Torr)


Figure 3-14. Dependence of dislocation density on growth pressure. Higher growth
pressure causes a lower dislocation density. Dislocation density results were
obtained by x-ray diffraction.











2000
S*ns = 5.2 E+12
1600 ns = 9 E+12
Ans = 2.2 E+13
S1200 ns = 1.3 E+13
E 1200 -
0

800
O

400
A
0
0.00E+00 2.00E+09 4.00E+09 6.00E+09
Dislocation Density (cm-2)

Figure 3-15. Effect of dislocation density on room temperature mobility of gallium
nitride high electron mobility transistors. No clear effect is evident.














CHAPTER 4
EFFECT OF DISLOCATIONS ON IMPURITY INCORPORATION IN METAL
ORGANIC CHEMICAL VAPOR DEPOSITION GROWTH

Introduction

Aside from high dislocation density, MOCVD grown GaN and III-nitride films

contain extremely high concentrations of impurities. Due to the harsh MOCVD

conditions required for growth of III-nitride materials, typical GaN films have impurity

concentrations >1016 cm-3, and sometimes as high as 1020 cm-3 [39]. These impurities

have a profound effect on electronic properties of III-nitride thin films and make it

difficult to optimize MOCVD growth conditions. Impurities have also been implicated in

providing mid-gap states and inhibiting device performance. The most prominent of

these impurities are silicon, oxygen, and carbon.

Silicon

Silicon is a common impurity in GaN and III-nitride thin films. It is used as an

n-type dopant, as it acts as a shallow donor in a III-nitride lattice when it is a

substitutional impurity on a gallium lattice site (SiGa). The energy differential of the

silicon donor state to the conduction band minimum (CBM) has been widely reported,

although there is some disagreement. Researchers have reported values of 22 meV [40],

30.18 meV [41], 30.8 meV [42], 31.7 meV [43], and 42 meV [44]. Despite the variation

in these reported results, all of these values are indicative of a shallow donor in GaN and

AlGaN films, and is the dopant of choice when n-type conductivity is desired. The









source of silicon in unintentionally doped (UID) III-nitride thin films is widely believed

to be from decomposition of silicon carbide (SiC) substrates [45] and susceptors.

Oxygen

Oxygen is another impurity that has been implicated as an n-type dopant when it

incorporates substitutionally on a nitrogen lattice site (ON). Many researchers have

implicated oxygen as being responsible for the significant background n-type

conductivity of UID GaN films. Several research groups have reported that ON donor

level states are between 32-34 meV from the CBM [40-42] while other groups have

reported lower values. Joshkin et al. reported a value of 23.5 meV [46], while others

have reported values between 2-10 meV [44,47]. Despite the variance in these values,

oxygen is widely considered a shallow donor. Many researchers believe that oxygen

impurities act as n-type dopants, simply because of its extra valence electron relative to

nitrogen, while others have proposed a more complicated mechanism for oxygen

providing the background n-type conductivity of UID GaN. For example, Oila et al. have

proposed that oxygen impurities promote the existence of gallium vacancies (VGa), as

oxygen-gallium vacancy complexes (ON-VGa) are energetically stable in the GaN lattice.

According to Oila et al., it is the gallium vacancies that provide the background n-type

conductivity [48]. Overall, however, many researchers attribute the background

conductivity of GaN to the presence of oxygen in the lattice, despite the disagreements in

the proposed mechanisms.

The sources of oxygen include residual oxygen gas in the ammonia source [47]

and atmospheric contamination of the growth chamber. Aside from these sources of

contamination, the sapphire (A1203) substrate [45] provides a significant source of









oxygen. When sapphire is subjected to the high temperatures employed during MOCVD

growth of in a hydrogen ambient, the sapphire surface is etched, and the oxygen atoms

incorporate into the GaN film. Performing electrical measurements, researchers have

found that the GaN film is significantly n-type within -0.4 |jm of the sapphire interface,

due to a buried conductive layer (BCL) [49-53]. This BCL is undesirable when

attempting to grow semi-insulating GaN films, and researchers have attempted to

compensate these residual carriers by intentionally incorporating deep acceptors, like

carbon and iron [49].

Carbon

Carbon is another common impurity in MOCVD grown GaN films. Carbon is an

amphoteric dopant, as its electronic properties depend on whether it incorporates

substitutionally on a gallium site or on the nitrogen site [54,55]. If carbon incorporates

on a gallium site, it acts as a shallow donor, with an ionization energy of -0.2 eV.

Conversely, if carbon incorporates on a nitrogen site, it acts as a shallow acceptor with

values of ionization energy reported as -0.3 eV [55] and 34 meV [42]. If carbon

incorporates interstitially, its energy level is mid-gap, and its donor/acceptor properties

depend on the Fermi level. If carbon incorporates interstitially into n-type GaN, it will

act as a deep acceptor, but if carbon incorporates into p-type GaN, its energy state will be

indicative of a deep donor [55]. Carbon is generally considered a deep acceptor, and

when GaN films must be insulating, MOCVD researchers often intentionally adjust

growth conditions to favor carbon incorporation in order to compensate the background

n-type conductivity of UID GaN. However, growers try to minimize carbon









incorporation near active regions, as it acts as a scattering center, and has been implicated

in gate leakage and dispersion in GaN-based field effect transistors (FETs).

The sources of carbon in MOCVD growth of III-nitrides are the metalorganic

group III precursors, and, to a lesser extent, SiC substrates and susceptors. Carbon is a

common impurity in any MOCVD growth process, and its incorporation can dramatically

affect the film's properties, but growth conditions can be adjusted to control the extent to

which carbon incorporates.

Previous Work

Growth experimenters have demonstrated that MOCVD growth conditions have a

significant effect on impurity incorporation in GaN films. Specifically, the effects of

temperature, pressure, and flows of TMG, ammonia, and hydrogen on silicon and carbon

incorporation were studied in-depth [39]. A correlation between resistivity, dislocation

density, and carbon concentration in GaN films has also been established, indicating that

threading edge dislocations and/or carbon act as compensating centers [56].

Furthermore, threading edge dislocations have been shown to provide energetically

favorable sites for impurity incorporation, due to the miscoordinated atoms along

extended line defects [57,58] in other material systems. Studies on growth of III-nitrides

have only investigated the effects of growth conditions on impurity incorporation on

templates with nominally identical microstructures. What is lacking to date is the effect

of dislocations on impurity incorporation in MOCVD-grown III-nitride films [39,56].

Both Koleske et al. and Wickenden et al. have indicated that threading edge dislocations

facilitate carbon incorporation, but conclusive studies to confirm this have not been

performed.









Focus

The experimentation in this chapter aims to determine the effects of dislocation

density on impurity incorporation during MOCVD growth. Specifically, silicon, oxygen,

and carbon incorporation are observed with respect to growth conditions on templates of

variable dislocation densities. By subjecting templates with vastly different dislocation

densities to identical growth conditions, the effect of dislocation density on impurity

incorporation could be determined.

Experimental Methods

Templates were grown by MOCVD on sapphire substrates, and dislocation

density was intentionally altered to obtain two series of GaN films with the largest range

of dislocation density possible. Specifically, the growth conditions that were altered were

growth pressure, nucleation layer, and recovery time. Template dislocation densities

ranged from 4.59x10 cm-2 8.76x1010 cm-2 for the first growth series, as shown in Table

4-1, and 4.46x108 cm-2 -7.6x1010 cm-2 for the second growth series, as shown in Table

4-2. As expected, conditions that minimized dislocation density were higher growth

pressures, GaN nucleation layers, and delayed recoveries. Higher growth pressures cause

larger grain sizes during the initial stages of high temperature growth, which cause fewer

dislocations upon coalescence [59]. GaN nucleation layers provide lattice-matched

templates upon which to grow high temperature GaN. Delayed recovery is believed to

cause dislocations to propagate laterally rather than thread along the c-axis to the surface.

The characterization performed in this experimentation included XRD to

determine dislocation density, using the method described in Chapter 3, and SIMS to

determine impurity incorporation. Two series of growth runs were performed on









templates of dislocation density varying over two orders of magnitude, while the only

varied growth parameter was pressure in the overgrowth. For both overgrowth runs,

variable dislocation density templates were grown at 10500C, while pressure was varied

from 20-500 Torr, and silicon was incorporated as a marker when changing pressure.

Impurity incorporation was measured for all three samples from both runs.

Results and Discussion

The first series of SIMS data is shown in Figures 4-1 to 4-4. In all samples,

oxygen and silicon were at the detection limit of the instrument. The carbon profiles

changed in all cases, presumably due to changes in growth pressure, but this could not be

confirmed due to the fact that the silane did not incorporate. Another observation was

that the radial position of the wafer had a distinct effect on impurity incorporation. The

profile shown in Figure 4-3 was taken from the same wafer as that of Figure 4-4, but the

SIMS results from Figure 4-3 were obtained near the top of the wafer, while those from

Figure 4-4 were obtained near the flat. This demonstrates that impurity profiles are more

pronounced near the flat part of the wafer.

From the results of the first series of runs, two adjustments were made: 1) all

SIMS analysis was performed near the flat part of the wafer to obtain more distinct

profiles, and 2) the silane source was confirmed to be online by Hall measurements of

previous samples to ensure intentionally doped n-type conductivity. The SIMS results

for this second series of runs are shown in Figures 4-5 to 4-7. Once again, oxygen

impurities were below detection limit, which demonstrates that oxygen does not

incorporate significantly into GaN films grown in this reactor. Silicon was also at the

detection limit, except when it was intentionally incorporated to indicate changes in









pressure. Carbon profiles, however, changed significantly with changes in pressure for

all samples. As expected, carbon incorporated inversely with pressure; the maximum

carbon incorporation occurred when the growth pressure was 20 Torr, and the minimum

incorporation occurred at 500 Torr for all samples. In order to determine the effect of

dislocation density on carbon incorporation, average values of carbon concentration were

determined for each sample at each pressure and compared. These results are shown in

Figure 4-8, and show that there is no clear trend confirming that dislocation density

promotes carbon incorporation. For all growth pressures, the template with a dislocation

density of 5.59x109 cm-2 had the highest carbon concentration.

Although dislocation density did not have an observable effect on impurity

incorporation, pressure had a significant effect. Figure 4-8 indicates that carbon

incorporation is, to some extent, a pressure-dependent process. At higher pressures, there

are greater molecular interactions between carbon and hydride gases, which promote

carbon removal from the film during MOCVD growth. Threading dislocations, however,

do not appear to provide significantly energetically favorable sites for impurity

incorporation.

In order to confirm the finding that dislocations do not promote incorporation of

impurities (specifically carbon) during MOCVD growth, two series of experiments are

proposed. First, a similar series of experiments could be performed varying parameter of

growth temperature. It is possible that the carbon removal mechanism at high pressures

is significantly different than that of higher temperature, so dislocations may have an

observable effect on carbon incorporation by altering temperature. Second, a similar

series of experiments could be performed with broader range of dislocation density









templates. This could be achieved by growing templates employing unconventional

methods to minimize dislocation density, including epitaxial lateral overgrowth (ELO) or

cantilever epitaxy (CE). These techniques essentially filter dislocations and have been

used to grow GaN films with dislocation densities below 107 cm-2. This would provide a

broader range of dislocation densities for the templates for the purpose of this

experimentation.

Summary

The effect of threading dislocation density on impurity incorporation was

investigated in this chapter. Oxygen and silicon were at the SIMS detection limit in all

cases (except when silane was intentionally flowed into the growth chamber to indicate

pressure changes), indicating that an insignificant amount of oxygen and silicon

incorporated into the GaN films. Carbon incorporation depended heavily on pressure,

and higher growth pressures caused significantly less carbon to incorporate into the films.

Dislocation density, however, had no observable effect on carbon incorporation. This

result indicates that threading dislocations do not provide significantly energetically

favorable sites for impurity incorporation in MOCVD grown GaN films. The carbon

incorporation in this experiment was pressure-driven by a mechanism dependent on

molecular interactions.











1E+21 1E+08


1E+20 Ga-> 1E+07

1 1E+06
u 1E+19
E 0-
O C
S1E+05 O
Z 1E+18
O C
D1 E+04 (
< n0
1E+17 -
Z 1E+03 3
U O
w O O
Z 1E+16
S1E+02

1E+15
1E+01


1E+14 1E+00
0 2 4 6 8 10
DEPTH (microns)


Figure 4-1. Impurity profiles for low dislocation density at the top. The impurities under
study were silicon, oxygen, and carbon for the first series of overgrowth
studies. The dislocation density of the template is 4.59x 10 cm-2. The SIMS
sample was taken from the top part of the wafer.











1E+21 1E+08



1E+20 Ga-> 1 1E+07


1E+06
S1E+19
E -0
o 1E+05 c
1E+18 C o
Z (D
O
-1 E+04 D

1E+17
Z
LU 1E+03 o
o 00
01E+16
U.' 1E+02


1E+15
1E+15 E+01


1 E+14 1E+00
0 2 4 6 8 10
DEPTH (microns)


Figure 4-2. Impurity profiles for mid dislocation density at the top. The impurities under
study were silicon, oxygen, and carbon for the first series of overgrowth
studies. The dislocation density of the template is 6.66x109 cm-2. The SIMS
sample was taken from the top part of the wafer.










1E+21 1E+08


1E+20 Ga-> 1E+07

1+ 1E+06
E 1E+19 "0
o0 C
1E+05 8
Z 1E+18 00
O
F- 1E+04 (
1E+17
S1E+03
U O 0
O 1E+16 O
Z 1E+02
O
1E+15 1E+01


1E+14 1E+00
0 2 4 6 8 10
DEPTH (microns)


Figure 4-3. Impurity profiles for high dislocation density at the top. The impurities under
study were silicon, oxygen, and carbon for the first series of overgrowth
studies. The dislocation density of the template is 8.76x1010 cm-2. The SIMS
sample was taken from the top part of the wafer.










1E+21 1E+08


1E+20 Ga-> 1E+07

S1E+06
E 1E+19 1E
0 2 -
o 1E+05 0
Z 1E+18
0
DP 1E+04 (
S1E+17 0U)
1E+03 0
uJ o
U 1E+16 C)
Z 1E+02
O
0
C) 1E+15 1E+01


1E+14 1E+00
0 2 4 6 8 10
DEPTH (microns)


Figure 4-4. Impurity profiles for high dislocation density at the flat. The impurities
under study were silicon, oxygen, and carbon for the first series of overgrowth
studies. The dislocation density of the template is 8.76x1010 cm-2. The SIMS
sample was taken from the flat part of the wafer.






























Figure 4-5. Impurity profiles for second series of overgrowths for low dislocation density
at the flat. The impurities under study were silicon, oxygen, and carbon. The
dislocation density of the template is 4.46x10 cm-2












1.00E+19 --r ..... .... ...

1 n: ,, ,n II "

E 1.00E+18
C

1.00E+17



S1.00E+16



1.00E+15
0 1 2 3 4 5
Depth (|tm)


Figure 4-6. Impurity profiles for second series of overgrowths for mid dislocation
density at the flat. The impurities under study were silicon, oxygen, and
carbon. The dislocation density of the template is 5.59x109 cm2.























0 1 2 3 4 5 6
Depth (u)m)


Figure 4-7. Impurity profiles for second series of overgrowths for high dislocation
density at the flat. The impurities under study were silicon, oxygen, and
carbon. The dislocation density of the template is 7.6x1010 cm2.











Carbon Incorporation vs. Pressure
1.00E+19
High dislocation density template
E Low dislocation density template
Si ,lid dislocation density template

c
.0 1.00E+18 -


o -

C
O
S 1.OOE+17 -

0



1.00E+16
0 100 200 300 400 500 600
Pressure (Torr)


Figure 4-8. Comparison of impurity profiles for different dislocation density templates.
Impurity profiles for silicon, oxygen, and carbon for variable dislocation
density templates during MOCVD growth with variable pressure are shown.
The low dislocation density template is 4.46x10 cm-2, the mid dislocation
density template is 5.59x109 cm-2, and the high dislocation density template is
7.6x1010 cm-2. The SIMS samples were all taken from the flat part of the
wafer.






70


Table 4-1. Growth conditions for the templates for the first series of SIMS results.
Conditions that promote lower dislocation density are higher pressure, delayed
recovery, and GaN nucleation layers. Dislocation densities were intentionally
altered to provide templates with as large of a range for dislocation density
possible.
Dislocation density Growth pressure Nucleation Recovery
Sample (cm-2) (Torr) Layer Time
1 4.59x108 500 GaN Delayed
2 6.66x109 50 GaN Rapid
3 8.76x1010 70 A1N Rapid






71


Table 4-2. Growth conditions for the templates for the second series of SIMS results.
Dislocation densities were intentionally altered to provide templates with as
large of a range for dislocation density possible.
Dislocation density Growth pressure Nucleation Recovery
Sample (cm-2) (Torr) Layer Time
1 4.46x108 500 GaN Delayed
2 5.59x109 50 GaN Rapid
3 7.6x1010 70 A1N Rapid














CHAPTER 5
GALLIUM NITRIDE BULK CRYSTAL GROWTH BY DISSOLUTION AND
RECRYSTALLIZATION OF GALLIUM NITRIDE POWDER

Introduction

Poor material quality is widely considered to be the most significant factor

limiting GaN-based device performance, and this is primarily due to a lack of availability

of lattice-matched substrates. Ideally, GaN film growth would be performed on single

crystal GaN bulk substrates, but the intrinsic properties of this material makes fabrication

of bulk crystals difficult. In other semiconductor material systems, such as silicon and

gallium arsenide, bulk crystals have been fabricated from melts, where boules of high

crystalline quality material have been achieved with dimensions as large as 12-inches x

12 feet. Similar methods are not feasible for GaN because the GaN dissociates prior to

melting at atmospheric pressure. Because of this, no commercially viable bulk growth

method for GaN has been developed to date, despite the potential benefits of low defect

density GaN bulk crystals.

Previous Work

Researchers have demonstrated several methods to successfully fabricate GaN

bulk crystals. The two most prominent of these are an ultra high nitrogen pressure

method, and an ammonothermal technique. Both approaches offer a distinct set of

advantages and disadvantages relative to the other.

The ultra high nitrogen pressure technique entails subjecting a pool of liquid

gallium metal to extremely harsh conditions in a nitrogen gas (N2) environment [60], with









temperatures of 16000C and nitrogen overpressures of -45,000 atm. GaN crystals are

formed by dissolving the nitrogen gas into the gallium liquid, where N2 dissociates into

atomic nitrogen, and bonds with the gallium. Using this method, researchers have

successfully fabricated crystals of GaN that are 1 cm2 x 100 |tm, with extremely low

dislocation densities (-100 cm-2). The drawbacks to using this approach include slow

growth kinetics, high impurity concentrations, the inability to grow boules of material,

and the high pressure and high temperature requirements. The GaN crystals are grown as

a crust on the liquid gallium surface, and a 1 cm2 x 100 |tm crystal is formed after -1

month under the extreme conditions required. Furthermore, these crystals contain -1018

cm-3 oxygen impurities, which may have an undesirable effect on the electronic

properties of the GaN crystal. Finally, the temperature and pressure requirements add

significant cost to this method.

The ammonthermal technique has been demonstrated to successfully grow GaN

bulk crystals by dissolving GaN feedstock into liquid ammonia, and precipitating single

crystal GaN upon supersaturation [61]. This method demonstrates significant

improvement over the high nitrogen pressure-driven process described above with respect

to growth rate (-0.5 mm/week). Furthermore, large area boules are possible to extract

from a seed crystal. The disadvantages of this method include higher dislocation

densities (-106 cm-2) and high impurity incorporation (-1017-18 cm-3). Moreover, this

process also requires harsh conditions, with temperatures of 550C and pressures of

-4000 atm in order to dissolve GaN into a liquid ammonia medium.

Although both methods of GaN bulk growth described above were significant

breakthroughs, neither are commercially viable processes to fabricate GaN crystals. This









is primarily due to the high-pressure requirement, which prevents both processes from

being scalable and manufacturable. If a method to fabricate GaN bulk crystals at

atmospheric pressure were possible, this would provide a distinct advantage over other

methods currently employed.

Proposed Methods

In order to develop an atmospheric pressure process to form GaN crystals, three

series of experiments have been proposed. The first involves electroplating of GaN using

a gallium metal electrode and a reactive nitride ion (N3-) from a lithium nitride (Li3N)

precursor. A second possible technique comprises an electrochemical reduction of

nitrogen gas (N2) to form nitride ions (N3-) that react with gallium to form GaN. A third

potential method is a simple dissolution/recrystallization process of GaN powder in a

liquid medium, where the formation of GaN single crystal would be driven by a thermal

gradient.

Electroplating

Electroplating of GaN from gallium and nitride precursors could prove to be a

promising method to form high quality GaN crystals. The difficulty, however, was

determining a suitable host environment in which a nitride ion could exist without

reacting quickly and explosively. In an unrelated series of experiments, Goto et al.

determined that a molten alkali-halide host environment could successfully dissolve

stable nitride ions [62]. Using this, the possibility of forming GaN crystals from gallium

and nitride precursors was investigated. Figure 5-1 shows a GaN crystal of macroscopic

dimensions (0.9 mm x 0.6 mm) that was successfully synthesized using this method.









This result was encouraging, but the electroplating method suffered from four significant

drawbacks:

Electroplating does not provide material with optimal crystalline quality.
The electroplating process would be very difficult to scale, in order to develop a
controllable, manufacturable process.
The Li3N precursor is extremely expensive, making this process less cost-
effective.
The buildup of Li+ ions with GaN deposition complicates the electrochemistry
[63].

In order to make this process more cost effective and controllable, the electrochemical

reduction of nitrogen gas to nitride ions was investigated (1/2 N2 + 3 e- N3-).

Electrochemical Reduction of Nitrogen Gas

Goto and Ito investigated the possibility of electrochemically reducing nitrogen

gas into nitride ions in a molten alkali-halide liquid medium [63]. Using this result, the

possibility of fabricating GaN crystals from electrochemically oxidized gallium with

electrochemically reduced nitrogen gas was investigated, and GaN crystals were

successfully formed. Using this method, bulk GaN crystals could possibly deposit on a

seed crystal, and be extracted as a boule. While this process offers several advantages,

the major disadvantage is the difficulty of controlling the fluid dynamics to form the

crystals. Because nitride ions are so reactive, they must be separated from gallium ions

until they reach the desired growth surface, which would be difficult to model.

Dissolution and Recrystallization of Gallium Nitride Crystals

The ideal method to grow bulk GaN crystals would be based on a simple

dissolution/recrystallization process of GaN in a liquid medium. Using such a method, it

would be possible to control the rate of crystal growth by controlling the temperature









gradient and the fluid mechanics in the growth vessel. This process could eventually be

used to form large GaN crystals, that could be extracted as a boule from the melt.

Focus

In this chapter, a novel bulk growth method of GaN crystals is introduced, based

on a simple dissolution/recrystallization process in a molten alkali-halide salt. If GaN

can dissolve sufficiently in such a medium, then researchers could develop a

temperature-dependent, atmospheric pressure process to fabricate GaN bulk crystals.

Such a method could prove to be a scalable and manufacturable process that could lead to

production of commercially available GaN bulk crystals and wafers for homoepitaxial

GaN growth. Specifically, this chapter examines whether a molten halide medium is

capable of dissolving GaN, and if so, whether it is possible to recrystallize GaN on a

nominally homoepitaxial GaN template.

Experimental Methods

All of the experimentation in this chapter was performed in a home-built stainless

steel heater vessel fitted for a 1" quartz tube in a glove box with a nitrogen gas ambient.

To determine the temperature-dependent solubility of GaN, aliquots of GaN in molten

lithium chloride (LiC1) were taken at various temperatures with a quartz rod. The

solubility was determined by inductive coupled plasma-mass spectrometry (ICP-MS),

where the ratio of the gallium to lithium concentration was calculated. Recrystallization

studies were performed in quartz tubes in the same stainless steel chamber, where the

recrystallization surface was MOCVD-grown GaN on sapphire. Deposited GaN

crystallites were observed by scanning electron microscopy (SEM) and elemental









analysis on these crystallites was performed by energy dispersive spectroscopy (EDS) in

order to determine the extent to which GaN recrystallized on the wafer surface.

Results and Discussion

In order to grow GaN crystals at a fast enough rate to be commercially viable, it is

necessary to dissolve a relatively high concentration of GaN into solution. If the

solubility is found to be sufficiently high, it would be possible to grow GaN bulk crystals

at a reasonable rate upon supersaturation. The basis for the experimentation

demonstrated in this chapter is the ideal solubility curve for GaN, shown in Figure 5-2.

This curve was calculated by the following equation [64]:


Inx = T -1 (5-1)


Under diffusion-limited conditions, a solubility of -0.1% for GaN could provide a crystal

growth rate of -0.5 mm/h. This corresponds to a temperature of 1230C on the ideal

solubility curve, and since this process occurs at atmospheric pressure, these conditions

are mild enough for a commercially viable method of GaN crystal growth. If the selected

solvent has favorable interactions with GaN, then experimentally, the data should lie

above this curve, corresponding to higher solubilities, and consequently, potentially

higher growth rates. The first series of experiments determined the temperature

dependent solubility of GaN powder in LiC1.

Dissolution of Gallium Nitride in Lithium Chloride

The temperature dependent solubility of GaN in LiCl was experimentally

determined by performing ICP-MS analysis of aliquots of solution at various

temperatures. These temperatures ranged from 750C 10000C, as shown in Figure 5-3.

The experimental data indicated that, in most cases, more GaN dissolved than was









predicted by the ideal solubility curve. The scatter in the data was most likely due to the

fact that the ICP-MS is an inorganic, water-based technique, and since GaN is insoluble

in water, some of the crystallites may have precipitated out, and segregated. In order to

address this, a 2 |tm filter was used to filter samples for analysis prior to injection into the

ICP-MS. These results, shown in Figure 5-4, represent a lower bound for solubility of

GaN in LiC1, but they were still significantly above the ideal solubility curve. These data

indicate that 10-50 ppm of GaN were soluble in LiCl at temperatures ranging from 800C

- 950C, corresponding to a growth rate of -50-100 [tm/h. These values were sufficient

to commence experimentation of recrystallization of GaN on a GaN template.

Gallium Nitride Recrystallization

In order to determine whether GaN could recrystallize out of solution, a series of

experiments was performed, where GaN powder in LiCl was heated to -9500C, and then,

became supersaturated upon cool-down. This supersaturated GaN would presumably

deposit on a lattice-matched template. SEM images of a surface not exposed to the melt

(Figure 5-5) were obtained and compared to one that was (Figure 5-6). Note the distinct

differences in surface morphology. Upon further investigation, individual features were

observed at higher magnifications, as shown in Figure 5-7. This feature appears to

exhibit a hexagonal shape, indicative of the wurtzite crystal structure.

To further promote GaN recrystallization, another experiment was run with

stirring at the maximum temperature (950C), prior to cool-down. The result of this

experiment was that more surface features appeared. From this finding, a third

experiment was performed, with 1) half of the wafer surface treated with nitric acid

(HNO3) at 500C, and 2) constant agitation during cool-down. The surface treatment was









performed in order to etch the native oxide off the growth surface, and the agitation was

performed in order to provide forced convection, and greater recrystallization. This

provided two interesting results. First, more features were formed on the wafer surface.

Second, the features that formed on the treated surface tended to be different from those

on the untreated surface. On the treated surface, there was a high density of patchy

features, shown in Figure 5-8, whereas, the untreated surface had a high density of

needle-like features, shown in Figures 5-9 to 5-11.

In order to determine the chemical composition of the two observed features,

elemental analysis was performed by EDS. The EDS results for a needle structure is

shown in Figure 5-12, and strongly indicate that this structure is comprised of etched

quartz, as evidenced by the heavy silicon and oxygen concentrations as well as the

gallium void indicated. Elemental analysis of a patch structure is shown in Figure 5-13,

and is indicative of GaN; the gallium and nitrogen concentrations are high, and silicon

and oxygen are practically nonexistent. Another interesting finding is that silicon oxide

(SiOx) preferentially deposited on the untreated surface, which presumably contained the

native oxide, whereas GaN preferentially deposited on the treated surface.

A cross-sectional view of the patch demonstrated in the SEM image in Figure

5-13 is shown in Figure 5-14. From the scale shown, this GaN crystallite is -5 |tm in

thickness, which indicates that the growth rate in this experiment was -2.5 [tm/h; this is

significantly higher than any other bulk growth technique. Furthermore, the depth of

field of the EDS instrument is shown in Figure 5-15 for the highest power used in this

study. Since the feature probed is -5 |tm, and the depth of the instrument is -1600 nm,

this further supports that the crystallite is that of pure GaN.









While the experimentation described in this chapter is relatively unrefined, and far

from optimized, it demonstrates that a simple dissolution/recrystallization process is a

potentially viability one for the fabrication of GaN bulk crystals from a molten

alkali-halide medium. The results were encouraging, and could provide a first step in the

development of a novel method for growing GaN crystals. Significant improvements

could be made to the process, including solvent optimization, greater purity of reagents,

greater control over the fluid dynamics, and using a true seed crystal on which to

crystallize GaN from the molten solvent. Eventually, using this method, researchers

could potentially develop a commercially feasible, rapid growth rate process to fabricate

high crystalline quality GaN boules for homoepitaxial GaN growth.

Summary

An ideal solubility curve was calculated for GaN with respect to temperature,

based on intrinsic thermodynamic parameters. The solubility of GaN in LiCl was

experimentally determined, and demonstrated that GaN was more soluble than the ideal

curve indicated. From these results, a GaN template was placed in a LiCl medium

containing GaN powder, which was heated to -10000C, and then cooled. Deposition was

observed on the GaN template surface by SEM. Further experimentation demonstrated

that GaN crystallites deposited at a higher rate after pretreatment of the GaN template

surface in an acid solution, and with constant fluid agitation during cool-down.

Elemental analysis determined that patchy structures as thick as 5 |tm were, in fact, GaN.

These crystallites formed after merely 2 hours, indicating an extraordinarily high growth

rate, relative to other GaN crystal growth techniques.






81


Potential improvements to this experimentation include

* More sophisticated equipment to allow greater control over thermal gradients and
fluid dynamics.
* Use of higher purity reagents.
* More comprehensive studies optimizing GaN solubility.
Use of a true seed crystal upon which to deposit dissolved GaN.






82















Figure 5-1. Gallium nitride crystal with dimensions of 0.9 mm x 0.6 mm. The bulk
growth technique to form this crystal was electrodeposition.












Ideal Solubility of GaN


0.25

0.2

$ 0.15

0.1
o
S0.05

0


1000 1100 1200 1300 1400

Temperature (C)


Figure 5-2. Ideal solubility of gallium nitride. The data for this curve were determined
by Equation 5-1.


.


r;


I


I












-120
E
0100
> 80

S60
o
S40 -
z
S20
0
750


Solubility of GaN in LiCI


800 850 900 950 1000
Temperature (oC)


1050


Figure 5-3. Experimentally determined solubility of gallium nitride in lithium chloride
compared to the ideal solubility curve. These data points were obtained
without the use of a filter.











Solubility of GaN in LiCI


40

35

- 30 -
E
. 25

S20

| 15
O
0) 10

5

0-
650


Experimentally determined solubility determined after employing a 2 |tm
filter. These data represent a lower bound for solubility of gallium nitride in
lithium chloride.


750 850 950 1050
Temperature (oC)


Figure 5-4.





























Figure 5-5. Scanning electron microscopy image of a bare gallium nitride surface. The
morphology is smooth.























S,- .











Figure 5-6. Scanning electron microscopy image of a gallium nitride surface exposed to
gallium nitride-in-lithium chloride melt. The morphology is noticeably
rougher than the bare, untreated surface.





























Figure 5-7. Close-up scanning electron microscopy image of a gallium nitride crystallite.
Note the hexagonal structure.