<%BANNER%>

Model for Suspended Gate Field Effect Transistors Used in Laboratory Animal CagexxMonitoring

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 E20101122_AAAACB INGEST_TIME 2010-11-23T04:06:07Z PACKAGE UFE0011981_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 21734 DFID F20101122_AABBDA ORIGIN DEPOSITOR PATH supan_k_Page_035.QC.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
cdab6186cff4291b171ddb827dd50d40
SHA-1
592ee8e55d038922587e937b73b903e15e04c8c4
6113 F20101122_AABBCL supan_k_Page_025thm.jpg
0c66bd2dce707bf0c77ac36d0ed1f345
d9d7c078fa33731f8839133bdb99a71883e99f5d
22486 F20101122_AABBBX supan_k_Page_017.QC.jpg
bc94517ca42b512866cbf91efc276190
7dac881c23d731bea77eddbfccdc5322e6e3a0bf
1862 F20101122_AABAXF supan_k_Page_035.txt
76c757e45c75d31fcf3d802591966564
a20e4887181f424b6203360bea59e4a86cae1976
1891 F20101122_AABAWR supan_k_Page_019.txt
3a4ad6938753bb4751e7ee87a0b14931
5511844a9c5afa1366b7b96216ae232c17f9e786
6082 F20101122_AABBDB supan_k_Page_035thm.jpg
bb8145fdd83788d78e81551679a54cda
db4faefc9d4f5047ba7943cb287c7d09f325d48a
23516 F20101122_AABBCM supan_k_Page_026.QC.jpg
2a71c8ecbf9ff5c93faba224dad6ef92
8bb6a7c9608b07e40acd85cd4e1d3347ea2129d6
6270 F20101122_AABBBY supan_k_Page_017thm.jpg
030a531d4dc6cddb9e0bd0605b70e0b4
7bf7c16f10ed54b2a5fdcd899854cbb14c145ca0
1905 F20101122_AABAXG supan_k_Page_036.txt
33fd65595471f9f54d5d796167db66a2
47bedc9cba737f9674c7393074883eee7451f65e
1307 F20101122_AABAWS supan_k_Page_020.txt
c2ad21c9f4906d3f1cdc916f79f3d614
a10fb3bf2770bd7c5fc49085fbdb6b389a2877fe
21189 F20101122_AABBDC supan_k_Page_036.QC.jpg
d495b3bc968063f15a26310f35335e20
13f4dc11f4f87c035ea9efe1a264b2d82fb7c82b
6443 F20101122_AABBCN supan_k_Page_026thm.jpg
d2dc39d5af28c71948bee1b4291559dd
047de6ae954604ae98e0ef3bc9d2137877f1faa3
6810 F20101122_AABBBZ supan_k_Page_018thm.jpg
71b1de3aa1f98baa1ca59e43785313ab
8a5ccc8ad72b1ef650edff30e9bf2658b78e842e
1921 F20101122_AABAXH supan_k_Page_037.txt
1ba971af71b635b158247906bcd33715
00a7b2add15aef9930d53be75c4be54db35dd266
1966 F20101122_AABAWT supan_k_Page_022.txt
57135a4914f2d0dfb755ea611ef69379
e85c468dc359ae1706bf7bd694d9e442f71f01dd
61413 F20101122_AABAAA supan_k_Page_144.jpg
16f8dcb8cf81c7505544bf1f371942dd
fca88b1f1c70d7e6d91107ca5456842648f85e94
6127 F20101122_AABBDD supan_k_Page_036thm.jpg
dfe9bf968ed35b6e38aa334850bccfd4
1e99b8aab7843d6a51a8d4313e2b497005f0b1d9
20176 F20101122_AABBCO supan_k_Page_027.QC.jpg
e3c983e1748fbead9b7cfedbcd7f60cc
77b8267c49475f17cb79444f0d78060ff95f6f3c
1879 F20101122_AABAXI supan_k_Page_038.txt
041362c1635b11b57802fa59be81a5ed
08047d8b22472568b83634e92a55a8ea539e96a0
F20101122_AABAWU supan_k_Page_023.txt
644f14a7ce41d41fd7d7a0d4f2f864dd
2b023cb6d2f5bb86111d48d1285c1c4f743f2fe7
25083 F20101122_AABAAB supan_k_Page_143.QC.jpg
ff6512d1d866554c239ec2809e952c53
31f77365acafbbb0e5d055bbdd12902454c8a820
6196 F20101122_AABBDE supan_k_Page_037thm.jpg
c83c26e0b4fc96f73b485951ee3f439c
764857c10f3ca20ed067328ff64bec35e0e8b358
5755 F20101122_AABBCP supan_k_Page_027thm.jpg
1e9614e7383417c200f5283a8af11191
326d6cf6bcb71a7ed32510c2bfd4a6ce29dde12e
1845 F20101122_AABAXJ supan_k_Page_039.txt
17d8875cd1643f83e9c459e40ee8d9ae
00f67634ecf34a4d860d5d06a0178bbca01589b9
1893 F20101122_AABAWV supan_k_Page_024.txt
9765940d5a784c5b8091dd3194bf9eac
66144d912a1b0d043fe4142f69c8340941ce5784
9747 F20101122_AABAAC supan_k_Page_145.QC.jpg
8b53dc82267af477d5827fe833927450
1f0f1b02d4e0f10242ce3d7174af2eefaa1e1c37
21740 F20101122_AABBDF supan_k_Page_038.QC.jpg
972146d9bc7ac5aef1ca865f029a86be
b138cb6b3943a139cec6a113ba6c3809fd548c3e
23287 F20101122_AABBCQ supan_k_Page_028.QC.jpg
1d8475449f1f9634c198924079b0d3f2
820d68068b9e5b1f298b85991c07decd24089e71
1831 F20101122_AABAXK supan_k_Page_040.txt
97331859fac752186f475d59ec9d37e1
1ccbb73b0a18cf5f4b535947fb5e390e7c6fb0f0
1718 F20101122_AABAWW supan_k_Page_025.txt
f9f710f616a5c1995c8aa1217514fda6
b93bb4bbf4eb2a860f2a024e6c0334b14422db07
25271604 F20101122_AABAAD supan_k_Page_081.tif
b42ea9683424e6bf9088dc8e2b6e637b
6dcc6ab0b9fc343cde8d2b2f8242b607323ca076
5985 F20101122_AABBDG supan_k_Page_038thm.jpg
1a38d9151e8eb84efce8e4298f9c4739
2a01341c5156998d38919fff6492540182fadabc
6425 F20101122_AABBCR supan_k_Page_028thm.jpg
f0ae7854cc77e1b927b85063e0002d65
e05076bd7d29b6f897beb985485ba2fb4a28c644
1960 F20101122_AABAXL supan_k_Page_041.txt
8d8073cb248173dc1037115c76741956
8f7f2713ecc50d75d9e1ec4cfad2273a57ce1a9c
2016 F20101122_AABAWX supan_k_Page_026.txt
6f7a89f6b470d815900f2a4607e0fb52
6b90be3d02f1d3dc3b7c8413e110e4da299dc53c
1054428 F20101122_AABAAE supan_k_Page_129.tif
94c24cb0b7fea3ba7a5096f8edc8c2b5
8f952be5098325f952a5c647059afe5598f745f2
22373 F20101122_AABBCS supan_k_Page_030.QC.jpg
02f045c6fe19b138b3c4c8c7078ac0d1
13b4b62b9baee856b27aa333fbfcdd69034ee9de
2209 F20101122_AABAYA supan_k_Page_057.txt
467b8870c53c7032c8f84fb601a1880a
c54cd060324a97fd4e511674cb2843eea61feac1
1887 F20101122_AABAXM supan_k_Page_042.txt
2ab9c1a074ae348fabfd106f287dee76
6ad08d0ef654a98e689d2ee3c30b9afca09630af
1674 F20101122_AABAWY supan_k_Page_027.txt
1a313279a522ccdfe806c9b023772554
59f2caa3ec80d90e4e006a1f8584535c3a80d498
22064 F20101122_AABBDH supan_k_Page_039.QC.jpg
7735d372a6fddb76a04a9b451668107a
8d29b8b865e81db750d48d42d145855b3a9130ed
22490 F20101122_AABBCT supan_k_Page_031.QC.jpg
f1e86f3356e7ff86d2b638f585fb567a
a8c70c4b9fdb5bfe3d401220842f748d5fc301a7
1485 F20101122_AABAYB supan_k_Page_059.txt
5265bab4c0c4d1204c9768f1e65b1e73
9d239b2688a18fdf32827de74ef904b66cb063d3
2004 F20101122_AABAXN supan_k_Page_043.txt
50db11ec5a40f56913f0b7f983b8ddb7
c99bf9796208f29fc4bd698c9de6a9586b1f7d81
1806 F20101122_AABAWZ supan_k_Page_029.txt
5d4437f7f48d318d5b11c18a235046ee
f5fbc0f5c9eca8a18fbeb48ba8cea4c1d56a2652
46330 F20101122_AABAAF supan_k_Page_039.pro
413dab8d59f848332d21862142f9a0e8
fd073d9a7fef5632d75301082b9864a99ab794a6
6235 F20101122_AABBDI supan_k_Page_039thm.jpg
35099c07b02606fafa16de396c1f99f3
d84c86116f42690bb4df8d8c5b32c9bf97b96ed6
6356 F20101122_AABBCU supan_k_Page_031thm.jpg
341bb73078a98d034b802b7790c7d937
3e0e5b6e53a5724a65d81d8f3c8f6dc64ae99526
1468 F20101122_AABAYC supan_k_Page_060.txt
515e5e7fbaa817fbafe425f9a26f58fa
56b338a78956ffc496ab583a81d3f1e6e9cfad0b
1849 F20101122_AABAXO supan_k_Page_044.txt
6b937bc981ab5b45ba28d6eb0f8a4506
b5a041bca49fa420d55d875f7a283483ac7d8012
44278 F20101122_AABAAG supan_k_Page_044.pro
c0a95289051f6d10dff2beb63f9e7028
f23286ca3327176f5bf6951a4d701e4b528f65d1
6420 F20101122_AABBDJ supan_k_Page_040thm.jpg
a782ae3ef8d27fb34e8753696584e7c5
8e4eae2fcb9db86e4d158dec8eb6a4e37b53d884
21923 F20101122_AABBCV supan_k_Page_032.QC.jpg
846bbcaebd6f0f3682b703e553cce71c
1ebeafcbec54063b4ff0a31e1cf7aa271829c7cc
1978 F20101122_AABAYD supan_k_Page_061.txt
e032d8e24714dee022d5d1f13c77794c
bf7fc06a5e4fcf91e248b0d10e9e4db70caf9771
1801 F20101122_AABAXP supan_k_Page_045.txt
46d045780b871d043332c9fbdcbc054b
5491ffc9b11d220ddade169d78272d6f77d3da36
1053954 F20101122_AABAAH supan_k_Page_026.tif
8d9156cefd68f75188c18c1d3f3717ec
aeb09061ac1c7e3d35446224b4205d993ac0fdf7
23813 F20101122_AABBDK supan_k_Page_041.QC.jpg
937225607cc77cd57cd5fc4cbc1893c4
68c3b1aa8024a914381baf7e1a0a80a28304027c
6090 F20101122_AABBCW supan_k_Page_032thm.jpg
f41555ee2ad2776eb9bc46e810261831
879cdee5f484ca48ff3120c61019a0604249ccc4
1672 F20101122_AABAYE supan_k_Page_062.txt
24ac212cc1ab901c08e96fa050172c03
4bea469223658e2f927963e36d3a92f55474faa7
1529 F20101122_AABAXQ supan_k_Page_046.txt
90cac512aa8e84acd76ea597cfcb191a
3fe547595c616f53f95dcf98764b5c9e634887e3
3811 F20101122_AABAAI supan_k_Page_007.txt
556afe8e39138b0cb50bae1b22fba52e
3cbfb5690e9e8e64ad155ab4a8cc1fe6a4f1693f
5947 F20101122_AABBEA supan_k_Page_050thm.jpg
705cbcf84658a1f1d3c5706b93494b9f
e3b7cf059adb771d6766141f160674ea06282feb
6709 F20101122_AABBDL supan_k_Page_041thm.jpg
7ab457f9f6c4906ffcd84873636ad542
8169736808f1e73abe94059918f22ff741a10847
19366 F20101122_AABBCX supan_k_Page_033.QC.jpg
42e739690f928f950b3d21135f5fe12b
0bd4967a09c6598c4fa3a004841bd2ba2a55ce1f
2048 F20101122_AABAYF supan_k_Page_063.txt
be77cd2d863942fe8f019961df5bd826
3c7a549fe48283a1704aebecc438beeab03bc34d
1765 F20101122_AABAXR supan_k_Page_047.txt
3835f3416fd9d4afc59058a93ac0e70a
a1cb91fe2cdf67c12954d8a97a2545e839d1b599
5363 F20101122_AABAAJ supan_k_Page_033thm.jpg
2f7e25ad4e9f21608b94f5655c57f000
98619050d447d2e2217722e708d13b85fee805db
6173 F20101122_AABBEB supan_k_Page_052thm.jpg
96633658b477c4f41342c85d5ed828c1
f2928200cc1cb2628bc39ec9ce2fb0fa6e888884
6355 F20101122_AABBDM supan_k_Page_042thm.jpg
613001635f3eea446a770b351501338e
cac6792a3571fa90b8e59932763df6e8c28c6d56
19646 F20101122_AABBCY supan_k_Page_034.QC.jpg
f2e81f78dc45db46a11ff92d88a7d521
9d003869057d247bc275387ef2cfc64542017821
2641 F20101122_AABAYG supan_k_Page_064.txt
57d0ec7ed4e50e0a9e1c8d76388ac493
cd0a6d31a9ef5dc700e9b945c36a2b7eab629ba3
1526 F20101122_AABAXS supan_k_Page_048.txt
ca42ff4382a1ca1404c5094334328fd0
6d41c8fa12b9b43b0b0d74263f4a303a599443b4
F20101122_AABAAK supan_k_Page_098.txt
f6e04de37f8af2f30874431fbefe7686
0568e539c6e0e0b7e76f2c5054aa896118c28f47
21118 F20101122_AABBEC supan_k_Page_053.QC.jpg
78da5189f808e8882b181f824e916916
d1a38ed54cebe1f4ab481cf04b3440d9d4957288
23960 F20101122_AABBDN supan_k_Page_043.QC.jpg
3d23b577a2c95e264ebd9095a58a3a23
dcbbdc2b5193d7774ddb972481c114059420f014
5807 F20101122_AABBCZ supan_k_Page_034thm.jpg
1bc3f92048fdc826979c0a859cd5ffed
d4d4bc242a52b11bc6a97fd6a2d7adca2650b196
1847 F20101122_AABAYH supan_k_Page_066.txt
3616b56b84463e9c0fdf5dba230aa554
62d4975a1faffe08b81aee167b44c332e913d0ba
1469 F20101122_AABAXT supan_k_Page_049.txt
90e58af4a1a0a147939d79fa849c6b16
21f1966114de3fa29ddcaf8aaf8905b4310c8c70
5769 F20101122_AABABA supan_k_Page_044thm.jpg
611cd0660f39d04a0b494b116ee73f5c
0b91b955f24588a0b7629c0fd0c48c8d3936b563
F20101122_AABAAL supan_k_Page_143.tif
1b225d121a90e4c16d5109df7c060913
a2ef8cfd5b2e0f067c3aca61ff1aa85f31111f29
6077 F20101122_AABBED supan_k_Page_053thm.jpg
154e6b7b6b1a5bb35c83c1681c0b2c1d
c2bcf146de3501f06fe7ae511ecca7f0107fe237
6610 F20101122_AABBDO supan_k_Page_043thm.jpg
f669a04013da7b131e1941b890f92f1b
6ebf4c4df437bbb55f123a0b4732b23b589eaaf3
1773 F20101122_AABAYI supan_k_Page_068.txt
d85cbe89f5b8f7c24b390eff28d85441
42a947e752e38c55f24c1de5da9782534c22298c
1575 F20101122_AABAXU supan_k_Page_050.txt
e96c3538e27bddc815e31b7a4fc44084
e2405ea4de96c15fa7dc6b2ee51fe8b555781718
F20101122_AABABB supan_k_Page_028.tif
38d2dc8ff56f5d575f38762e364e85cc
c532d142c1ef150d4e2c70e6799226a0d9aecce7
26799 F20101122_AABAAM supan_k_Page_125.jp2
e4701f37eaad4e1dedd2cd50bdd4f76c
7fedd566e9607834b233ccec5754c378404e4a38
19120 F20101122_AABBEE supan_k_Page_054.QC.jpg
e680301d6477b29fce4df27b117366db
a477e40808cc9d7bbea229cf6765fe9f231a311d
20685 F20101122_AABBDP supan_k_Page_044.QC.jpg
d4a3357b4eb323bc5406607b0114921b
8ad0c6ea32425aed3d3553e8b6f6669a279f9313
1781 F20101122_AABAYJ supan_k_Page_069.txt
1b2cb00c7aa7aa39e9ea8c32f5a524af
06ac347339c904717bb1fed3039d2da5df30fc14
1843 F20101122_AABAXV supan_k_Page_051.txt
a5b5d4d0c7be95d7dce8a442bdd89540
0cf4d427a5d8726c4ae246378ea9e691985814e1
7467 F20101122_AABABC supan_k_Page_001.QC.jpg
4c84ab3065e211f5534e85e9328fea59
fb35136fc4e7fb5c4df137734e583690a72a19ce
7280 F20101122_AABAAN supan_k_Page_003.jp2
6713139d4147fed7f45ca5177f1b8def
fc5cabe6d0e2c8e51d307b0ded6d86e2986b1d52
5624 F20101122_AABBEF supan_k_Page_054thm.jpg
3eb4150e6e792036244d3d971aaedc51
ebe43d025a65be6b4ce8438bb3e7ca62ee1dda5a
21252 F20101122_AABBDQ supan_k_Page_045.QC.jpg
9e90b545bc31c6dba98716be620794fa
108633f2593286a91355e29682186bab98285abf
1410 F20101122_AABAYK supan_k_Page_070.txt
c496f80396211433f0f51d3ee56e1cd8
9bbb582613e3322b5eff581bf597eabccffff0d8
1819 F20101122_AABAXW supan_k_Page_052.txt
09bf9a2874aefbfb506c28366e2e7df0
9861f8f6c14204e6fb229d72bab10dd1fcde4cb2
693 F20101122_AABABD supan_k_Page_145.txt
20921ab18ad0a3819df9d79f5fc8c199
a9a250ed17d619b00a847869b9dea55ad541b4be
1147 F20101122_AABAAO supan_k_Page_058.txt
67c0ca8b98915924841e2da5e12655ba
0b3f9b43c1869a323ea2ba0c92d14a1302153625
3635 F20101122_AABBEG supan_k_Page_055.QC.jpg
450bb05cba82bbf527d421afdf24f54c
88a8f99c4c0b0a05f33e0c5de7712f8c52a923ee
6222 F20101122_AABBDR supan_k_Page_045thm.jpg
ddd3582e53c45cc79d8ba4ce08da94f9
d42c1de90fd9986c85a00d4cacb8eb6ce338fcc9
F20101122_AABAYL supan_k_Page_072.txt
129d85e269ff5aa7ae1934f61972347b
22170edbe9ca3db0590e0de4828f68562c01ebc7
1740 F20101122_AABAXX supan_k_Page_054.txt
6eb599daa82bcdd0cb883e265fa7a77c
fc992030632c6d3856adb464144eb2664180cd3d
F20101122_AABABE supan_k_Page_080.tif
b9ce5cdcf60992414844b413ca6463af
39fef73c12ca92ef197d989494b261a6917ac30b
F20101122_AABAAP supan_k_Page_085.tif
02e522d0ff29a99b06f50674d084ab27
e1c0406afb6fdf3a84321402e39197a0a7b2eb2d
F20101122_AABBEH supan_k_Page_055thm.jpg
e54eb1e56ac7b3bb0bccbf53e3ff35cd
d6550a5e1e79dba69f822a0c962d558fd57d6d05
18795 F20101122_AABBDS supan_k_Page_046.QC.jpg
7b68e753f408187d33f0388e2c480891
dd13fa6c38c233afabf90a40fc86859f19766f5f
1770 F20101122_AABAYM supan_k_Page_073.txt
f0662eecf76d87dd1af249496e4f265e
5f7e2f3569771bacf50c7a9165c004fee6356d6b
118 F20101122_AABAXY supan_k_Page_055.txt
2197a17cd726a073259b8697016bdf2e
eb21b49a5c5f3b8bc0c6ffb03dff09a59d56c2c1
99891 F20101122_AABABF supan_k_Page_037.jp2
aaf342354ed19860838e6c4551aec647
1edee46955def69ca9de94f17e75f128c8a27876
21678 F20101122_AABAAQ supan_k_Page_099.pro
10eb770c4d1325a543b8aeb15f6f45c8
91773269b605ea57bc226b5084f0b5768542fc90
1705 F20101122_AABAZA supan_k_Page_089.txt
9b4d083cd1f228a718ee072cef41eb6b
b7280e8a5d23c24048f2324bf9c8c2226f488d46
5507 F20101122_AABBDT supan_k_Page_046thm.jpg
0ee46c0a5e390d53a5d51072470b6afa
a94eaaeecfe503b9e89545201618baa529b4288c
1712 F20101122_AABAYN supan_k_Page_075.txt
298d8bc414d0171768f93e839858c487
9dc1ae09f21f68cd40a1b5817069c10a440e0b94
2043 F20101122_AABAXZ supan_k_Page_056.txt
4cde9092509dfdd52f9c138dbbb08d0a
a2088d9348ee75616cdb2d842745ea0f3e548a6e
23254 F20101122_AABAAR supan_k_Page_022.QC.jpg
f3fe91e2a2afad1717de832303442ef5
dbc62534da2f574dbb9b933979192a90d4745d09
1791 F20101122_AABAZB supan_k_Page_090.txt
f2459b775ca797d707e0d6f0299f5531
437d0cb59a6934e8249367874a99a02126c0085f
16038 F20101122_AABBEI supan_k_Page_056.QC.jpg
ee3fc07209fe74c7b4928b5e1965bbcb
33fd079bf3828147e782578a54922a60d8617bb0
19850 F20101122_AABBDU supan_k_Page_047.QC.jpg
a54190af23a95fae5e8fb983cf1c758e
12b5de00e289462e8d7d598546987b1e05d51f56
1647 F20101122_AABAYO supan_k_Page_076.txt
df8743e524ebaf618ee365bcfa8f0b01
243012224e3e039140512724407201f80d61e07c
20665 F20101122_AABABG supan_k_Page_117.QC.jpg
4477a981b0a3888bf4e1633ee8648c7a
0bb751661c28c1a2ad1ccfb3c0cd7b2d7ae91463
6396 F20101122_AABAAS supan_k_Page_030thm.jpg
51e8075c1e15f84ab03820374f320d69
4e90c4bbf0e23cad59576873d21c12b55a8f0bb6
1370 F20101122_AABAZC supan_k_Page_091.txt
1655db75e0e932d889b2a0582050e13b
88126953eacf6d03cf5ad6ee70169035c60a3cad
4675 F20101122_AABBEJ supan_k_Page_056thm.jpg
6b11cb8cde05616a445d7d8cfce467a8
b8ce53557c5517f4fe074abff79f6be1b77a65fe
5680 F20101122_AABBDV supan_k_Page_047thm.jpg
6cda4e15522648a10be13362491c220e
009e9ce25260e1556e06fb9b8149edada2d35572
1640 F20101122_AABAYP supan_k_Page_077.txt
2a0da3a6d8d9154d1375c8944df387a2
d79be4bf548a009a55443a7da57addc7cace82dd
F20101122_AABABH supan_k_Page_048.tif
bc11553de610741eb4a489f9ac2b48bd
50a35f33ab62fd81c3868275928a1616c8727de1
65437 F20101122_AABAAT supan_k_Page_091.jp2
7a04fb4de4306e4e55eeaea5dc88da93
3bf7a7a3ec5debf3b805e07bd5bd11c6ab755a07
1766 F20101122_AABAZD supan_k_Page_092.txt
0d4a92a8bfd05c6098160f00686a9d96
58a5f9e31e5018ba8577db0c58d03c452db16b10
21595 F20101122_AABBEK supan_k_Page_057.QC.jpg
fbd63c26cd02c9f95f230fa7b30723ba
51f19cce4a1644cf6799f953668b5d388c6bb3ad
6287 F20101122_AABBDW supan_k_Page_048thm.jpg
0f21a569f76f921d9a3a07bc09b254fe
d5d9e87aeb316f042949ffdcb08f31225458213f
1743 F20101122_AABAYQ supan_k_Page_078.txt
8f59d482de9eb059b42e6c01ba9163ce
5641e563b03a4c4e1347aa6c4c884cb2d89a20b4
4175 F20101122_AABABI supan_k_Page_061thm.jpg
5a02d055f90366c8bcc261c0a74840c7
0d95cc8bba581c0fa2fdd1f01521ab16701b29c1
F20101122_AABAAU supan_k_Page_025.tif
8f511199dfc2bb097c421be5724eea27
d2bb527144b2a27176b0b23cbca7304f86c60473
2331 F20101122_AABAZE supan_k_Page_093.txt
5f258b08f2b6936a4db529b52c2f4e6b
6db3c47da3d413e403d3a161aeca3c256b650bb5
F20101122_AABBFA supan_k_Page_067thm.jpg
1ff3a1fd17ef74086fc8027ce2f6b33a
98929988fe786caec8855f14b0d01dac1aab9ab7
21665 F20101122_AABBEL supan_k_Page_058.QC.jpg
50f959b1824544777590d3cfc781d2e7
1dbb66c72f87b9d3e8869167dd3f9e4429ea3310
20823 F20101122_AABBDX supan_k_Page_049.QC.jpg
2cb9121ff28196cf2f64460679af939b
49d8107cff41874292baa21c49cb0df14c69676a
1549 F20101122_AABAYR supan_k_Page_079.txt
7848f7765b9e1eee4559c37db4197066
903a8ce8aa5f474c51eadf8854542a05923b6ab1
43249 F20101122_AABABJ supan_k_Page_076.jpg
479fa6998133afc5c7f76f43073a0246
938ed0b34682bef1759502b998cfe9ab3c40ba5a
F20101122_AABAAV supan_k_Page_105.tif
6cb6f1b63b2befe6f1fa7d4d322d8940
760e62133b682a78feb59583a57f1d14f9d2cc63
F20101122_AABAZF supan_k_Page_094.txt
e767dc0953849a19c9ffbc0a60393e0f
903899f9f0d1b4aac90a3a528055625a970b9259
20090 F20101122_AABBFB supan_k_Page_068.QC.jpg
2fe1ff598ade307781939d3fdd30c662
55be9dcaebd293ec38be1e30c04bdee2d38c2dc6
6251 F20101122_AABBEM supan_k_Page_059thm.jpg
5763039d9c68adc7b4559ac1318d0566
ef2270d85228225a400afe83fbd8502962f3e86d
6216 F20101122_AABBDY supan_k_Page_049thm.jpg
f15e362463d5e74af2d5680e379517d5
a516534e19174ce01ad24aafd90067889fb31ab0
2176 F20101122_AABAYS supan_k_Page_080.txt
fa842059e2f43ae53fddb2b5697abb1f
963a70f9456fe9cf33b69bb9e316a223a6bd3c68
5279 F20101122_AABABK supan_k_Page_096thm.jpg
a589d3263717644378c49d2a06a483b8
264404cb76d21ea01fcf6927ccb91c025569555e
F20101122_AABAAW supan_k_Page_139.txt
0645dd6eef826c0726b36ac1abc22efb
85d631df39f3a953cbce543d28089948e658eb5c
F20101122_AABAZG supan_k_Page_096.txt
990a49421e5c9fb5319bc6997bd88ec8
9e6a30c2a13cf92bfcf6f3d67ed5f50690b9727e
5661 F20101122_AABBFC supan_k_Page_068thm.jpg
3a7835c2f1117329bfc7f9b558535a27
c6a16ab94791a2ea5e51c81f46e8a046f9feef76
21298 F20101122_AABBEN supan_k_Page_060.QC.jpg
8c849787dc5ad6a6f45e30be3eebe6e5
5bedaf2f35f7f109659e59985335067aed5a60b2
20690 F20101122_AABBDZ supan_k_Page_050.QC.jpg
afe1a745ce03118c99417d462a14be9f
d19673c4ddd72b4f02b2b50085e0a09a6456a55d
1903 F20101122_AABAYT supan_k_Page_081.txt
7a0b753b31df481db870ba47d8ac9243
60bead238d4f64b0fe6aec13e798c97a2ec130b4
931530 F20101122_AABACA supan_k_Page_050.jp2
c2999daa49b996338ab223177e6d8a17
b46818e58e0fcab6463ba4fd4d6bcdaa03fcd36c
19498 F20101122_AABABL supan_k_Page_084.QC.jpg
b3f50700f0c0aab6629a978f500311f4
42efcd2bc378bcd5e5fae6f75b6c89d5ad113ae0
22127 F20101122_AABAAX supan_k_Page_113.QC.jpg
b2adcd36947c0f658c0fa772435f5eab
bf3066e63ec2700125af1d79cdb0d761578a15f0
1700 F20101122_AABAZH supan_k_Page_099.txt
c8a9ca8095abb93ad3dd89665eb33567
fe7183ac0474e40dea5c531214716d8d20ffe542
15353 F20101122_AABBFD supan_k_Page_069.QC.jpg
a0e3cb17914242c917cde542783e62dc
ec06ecb34eafc2860c1ca0014b3171d7b1db1f50
6317 F20101122_AABBEO supan_k_Page_060thm.jpg
abf4970a97ebfa92542c2390c96622a3
328e5a16da9987ef165c0fa65c79bef0b31da4d7
1733 F20101122_AABAYU supan_k_Page_082.txt
8746c92f952ac41ed8624116fa25bd3d
29b5d13c68d5cc90f53ef236be5586aadb70efc2
6027 F20101122_AABACB supan_k_Page_073thm.jpg
50afe6c7e1aecbcab7123e8d729752d7
4008a8488d650d007c0c90a0b379ab3020734472
89808 F20101122_AABABM supan_k_Page_054.jp2
861f917d2803c5ef0b3d76f26dd120b2
32d83a2ce41dfe77b58afa238933938edf5318d5
21849 F20101122_AABAAY supan_k_Page_121.QC.jpg
b5178a40735b2571b1076a9bf639a26f
82b1e05d81e624d15b448578f66e25f12b0fa1bd
1946 F20101122_AABAZI supan_k_Page_100.txt
6532cff72235d9452b3d0e78cd2289f9
0a8c7e5bd050adedb26e25dd2adddbc6f6866e39
4877 F20101122_AABBFE supan_k_Page_069thm.jpg
0d5177235c7638a93b08d6826836228d
3aec46eb9f5b96a3596fd906c38d95c060958277
14689 F20101122_AABBEP supan_k_Page_061.QC.jpg
64404846676bbeee8ca7e203b96700a1
cb075ca5090f535cf5ebbf12725ae001cf83c095
2115 F20101122_AABAYV supan_k_Page_083.txt
f6c2eecbc77ced0f43dac24a47f31bba
3754c8f87e186e6602f2a444efe6a6247fb35741
16448 F20101122_AABACC supan_k_Page_006.QC.jpg
587dba8520cd99d548bd5a2b1158d2d6
374fe4b876073470e8f25fd4b6f944a4ee4753e9
999985 F20101122_AABABN supan_k_Page_102.jp2
345f6dc914b7dcdabbab2afb03c53b76
5feb5d8837e1561dab3f71275c35189035c5f7f2
20914 F20101122_AABAAZ supan_k_Page_052.QC.jpg
2fcb07fc8644c0d69817ae3035f9e1eb
a78c0a3db227e9ed65066c84163aa6d5c8861be9
1381 F20101122_AABAZJ supan_k_Page_101.txt
9bfbab15f2a725c3daed5927bd86f8c6
5e93c7c075ba8d1621af12d979c5cdb477457e95
12408 F20101122_AABBFF supan_k_Page_070.QC.jpg
55fd0e35273d61bf0bf48df3ef63be50
926dd8afb1db60236d3be7d91c8980390cf879ca
18469 F20101122_AABBEQ supan_k_Page_062.QC.jpg
9023e38788717176a40bf2c773401e4b
68ef1698dbed41d9725f0270ed70455392f47a8f
1714 F20101122_AABAYW supan_k_Page_085.txt
d55aeddea6e7a320149df2662ad4cf42
0f2678cf2aef32d7ce3977c9b1e434dcee421d97
55705 F20101122_AABACD supan_k_Page_075.jpg
e62c0184960b18a7515f770e25d4397d
7dedd0538a6d359de36187afaa43f76cbe0320e9
12370 F20101122_AABABO supan_k_Page_005.jpg
1f5aa80c31d666e86ef0691b29fdfd66
3feb041004e3c4e74ae4b55633d34d4a99c22899
1395 F20101122_AABAZK supan_k_Page_102.txt
f9266783b436a13d868d70e220f51e93
9f87eeaffc6d9742ab981169258a05ebff9fb164
4044 F20101122_AABBFG supan_k_Page_070thm.jpg
fc6d6e2f057b9baceb54b96af17fc82e
45d62932085904e918b9f31a836f5903f660c1ec
5505 F20101122_AABBER supan_k_Page_062thm.jpg
7c814d3172dadf0dc3424d3573d74531
b702eaf9c454e2dd6e3b4e6c09d93d21c5537583
1757 F20101122_AABAYX supan_k_Page_086.txt
04df4777d15463b898e1bf81cfe47b44
5d6a7e4ed77998ce0ab561cb2f83b31848427bef
F20101122_AABACE supan_k_Page_056.tif
78623e2ee8f0b232b05caa98104970a7
024955d14a6cddc44bca0dfaca8e6e3d57125063
3769 F20101122_AABABP supan_k_Page_134thm.jpg
c18f864090541a19ef151dda1864ed31
bbc08a2f3b7c2052e60ad22934b469f2aa5f62a7
1039 F20101122_AABAZL supan_k_Page_103.txt
2c383bd0c3c9b014bc840de771ec0958
79072679680cc6fbff1a779283d078e4832ae3ce
15478 F20101122_AABBFH supan_k_Page_071.QC.jpg
403bda5a7ef769c5556fae891cd87942
ff3534b85f4da1ac2ec8c70ef872bbf9ea295c72
19816 F20101122_AABBES supan_k_Page_063.QC.jpg
9384694dd19572fbff7ee69d75486114
f4de430aba9ae0731dccb55bb5a8c33bbb99e199
1965 F20101122_AABAYY supan_k_Page_087.txt
a591856a9c0fc9083d20fcfe34c1affd
8e85f44e62fe066a000def7faf46512ee4b86487
3287 F20101122_AABACF supan_k_Page_002.QC.jpg
8448c31903f8e0f12a5193ff41d021d8
e8a4c96aca0aae03ed460ebd689d5d1b5ff1c9e5
16633 F20101122_AABABQ supan_k_Page_107.QC.jpg
9c6d0d231f6f460fa0662cfbb6c7ea75
504650688937da8e75def5df01a89d7a21b13b56
2055 F20101122_AABAZM supan_k_Page_104.txt
ddc51405093cbdaa943c59cebcab24c9
b295545011fd2d5c81190859b37ed5013c080fda
4903 F20101122_AABBFI supan_k_Page_071thm.jpg
8eeee175ab3d5381c068d635794ec797
eb1bc690627dc9750d37a94b7afdaa0ced4a228e
5785 F20101122_AABBET supan_k_Page_063thm.jpg
7c5349ce8735d4590891660ad6b36826
adb5fb6de63c943a5ff31f23f9c53c44ff395111
F20101122_AABAYZ supan_k_Page_088.txt
4aa151ce55244361654bfdcd035ad7bf
2a136c2eef5b245757cbeab14f83cbfd868b4a39
47042 F20101122_AABACG supan_k_Page_118.pro
30a0a11da78edda86ecd16df04f18bfb
02b40e3cf5b2eaea2eb081cd117917cc2517b3ba
110326 F20101122_AABABR supan_k_Page_026.jp2
c46014621845c985b92ec2aee3320902
f7458861a08d9f547bf087163d99f115a10875a5
1797 F20101122_AABAZN supan_k_Page_105.txt
70bd98bcdf9bec62ce0d71013a4ae935
3558c31c24e65a2e715068f9a96d45e8a86e9f73
16355 F20101122_AABBEU supan_k_Page_064.QC.jpg
eaff9d6c142c4f1c122304cf1b67db67
f85d5d1ccb80dae2790330f11c2b397100567535
43250 F20101122_AABABS supan_k_Page_089.jpg
64184ad015ee9b7e880a05123c77a51e
bbc5d85c61e61b86833d091f80d78552d97f6fa1
1973 F20101122_AABAZO supan_k_Page_106.txt
09ef4d9421d884d4d3c85a681b1fe41e
2db4a89e38f3a0cd3ed560a90e6c9d30eba94e28
17546 F20101122_AABBFJ supan_k_Page_072.QC.jpg
af4fea2e560266b3080d47e0809cf125
04cfd1b10e8b71d97842d6ca9ed76eb8d3fbb1b9
4932 F20101122_AABBEV supan_k_Page_064thm.jpg
8a7e9214d1ecdc19e8ad73d034498988
cde54885f6ebe8138cd6a96ad81ab24208ce1476
18489 F20101122_AABACH supan_k_Page_066.QC.jpg
e9f77ed89b075e7aafd9abcecc4120af
89e00d71a53e3526479f67d3f1a9efb99397c6cb
18683 F20101122_AABABT supan_k_Page_083.QC.jpg
3c898d79f94ce29cf7e660eb6b1c614e
316a6509eb619d2816a16ab9d3e081041b5a93b7
1484 F20101122_AABAZP supan_k_Page_107.txt
f6df7f273fdc3a37307b43545d8d36ee
1af6b4d864e609759ca219a06983a71b3dccd66c
5077 F20101122_AABBFK supan_k_Page_072thm.jpg
6f24626e9b79244846592cad6a7c20ab
f0747e2fdb91c235bf311bd029026f51bf88ba50
20303 F20101122_AABBEW supan_k_Page_065.QC.jpg
9e519edee46c3fab9c44c2d544253e81
23d7a6440e495e2be4d208818fced4e6ab247245
1051978 F20101122_AABACI supan_k_Page_010.jp2
ba696d983775edcdc3726a89e14bc5dd
b3308d35e3dd1d53f64d9684af67c22630cb88b8
91176 F20101122_AABABU supan_k_Page_011.jpg
cde0434dbc2e44232bf4bb5249a52be8
c58bbf01d38b725b7c89c4187d6f1d2c80ddae19
2012 F20101122_AABAZQ supan_k_Page_108.txt
b9736fb3b78e88f9c6290b1809aabd50
6044813d2ce1ecda28c3d6211d531795630dd389
22757 F20101122_AABBGA supan_k_Page_082.QC.jpg
f7dfa854b4b9adf0a6bc8725972e553a
f2f74e64cd6d91b84b026b0ab0ba6900a424dbe2
21433 F20101122_AABBFL supan_k_Page_073.QC.jpg
70403f8b53f79c9b4c7d59e49e9dc272
633b1b38f0a20382491a80e0c6c1c646e904b267
5652 F20101122_AABBEX supan_k_Page_065thm.jpg
3c9532f45dd6573bd015eaf3f7f4a344
6f5b339338bd8f3979137e250ab43c7255e9acff
89994 F20101122_AABACJ supan_k_Page_120.jp2
e6d2d17e8191e3ca068b213b63102cce
8b6f19824c6c44ce8b973bbecc9b9cbd826af386
48054 F20101122_AABABV supan_k_Page_017.pro
bb7434c3f15bed07d1fdcab7933dedc2
09609ca3d706313c2a1586822436d09010d00632
2178 F20101122_AABAZR supan_k_Page_109.txt
5232ccf462ddd818004de3e544ac7ac0
54b4925e89f13a61ebb0922087c11ba0ecc8ab77
6318 F20101122_AABBGB supan_k_Page_082thm.jpg
bdd7bf801f6fe259797761c85ce2429d
17d9c8dfffc06bdeef1724ae7e4063391da18f13
17745 F20101122_AABBFM supan_k_Page_074.QC.jpg
5a535bf17f6fe91b3a29fa26eb6e3888
2c539e6726771085c1d8df816fce1c9eabfd7d14
5502 F20101122_AABBEY supan_k_Page_066thm.jpg
30627d605d0403a5011dc0d1e74dc16c
10ea1654905c17d33cd39685029e55782c44b432
1937 F20101122_AABACK supan_k_Page_028.txt
d92652bfa00891bf6af8e587b6db9111
a4babee40afb78cf0eef3fe79566008d2bf7f98a
6104 F20101122_AABABW supan_k_Page_029thm.jpg
c068744c876fed0eb90d6f9bacaa0668
0c9bf64972edfbb75629277d630cc927f2cd7998
2105 F20101122_AABAZS supan_k_Page_110.txt
dbe01c1582329b3122485b67c535c6be
e1980e03ae5f011b8828572a3cc944339a7e4820
5239 F20101122_AABBGC supan_k_Page_083thm.jpg
ca58e306ec3c7fce915a7aeee39dc19f
fc44e41087d4f6cd3d38cbbe59e84bd053185dce
17862 F20101122_AABBFN supan_k_Page_075.QC.jpg
7212f8ac37d93b5e1a6199bfb4319f86
613792ef5fa772ed44e58ab912982401e7a61fe0
3870 F20101122_AABBEZ supan_k_Page_067.QC.jpg
2dfd931aae6c8d8bae5f4c00b268eceb
f578486560d3481c9be6e38d87125e16a738c5cb
F20101122_AABACL supan_k_Page_036.tif
a8add65ff0361820a7e17f01cbc76e24
5fe1ef8210d8a314e196557ae8dbd756fa57ade1
5404 F20101122_AABABX supan_k_Page_139thm.jpg
42970fc0d0315826de5fd31898d3264c
d39162bab52e326b66b08ce5917640bb9f0bc5f6
2477 F20101122_AABAZT supan_k_Page_111.txt
b6571e7f4d190d310ee35b25f51914a4
f2695ea30ffc9ce268079bcc950c75900618de88
13126 F20101122_AABADA supan_k_Page_008.jpg
b9d41cccce52375acf013daa32d22715
b1cca982335133237ebf6d3307c22fd5225a8bba
5591 F20101122_AABBGD supan_k_Page_084thm.jpg
b427102421f78d0a2852b4a953ceac96
4c92f1a051f4923e5f2b7367083dcee1f1f02739
5002 F20101122_AABBFO supan_k_Page_075thm.jpg
22c77f0aba7fb2bfa54cb07511a18285
c6f363e544f884a99568ecb03a1ecc2a08f3ad7d
F20101122_AABACM supan_k_Page_140.tif
e6ba783adbfafadf104d83fedbc00143
1001ea7323a89bbe78e493cb5170fb33fe435da0
36780 F20101122_AABABY supan_k_Page_060.pro
a6c6273716d8a7a5de48442584938915
a0b6c9c7fcaf458118f82a476bff9fea1939b63d
2123 F20101122_AABAZU supan_k_Page_112.txt
bfbd86494340b733addeb02047fb770b
1b062b13969ed270361c684fef9d7598b09c2390
68513 F20101122_AABADB supan_k_Page_009.jpg
84792a72130f2b2c47bfd4b2a2b4f396
90dda6177b31ca859b276dd5e12702e19965533e
20317 F20101122_AABBGE supan_k_Page_085.QC.jpg
09977bc99df07e9157d9a279b0b0dea2
6ba93a5e1c39be056ee8c89d31e664d93a25f358
14088 F20101122_AABBFP supan_k_Page_076.QC.jpg
0e5b507228b8ded776432893c48e546c
c78673941411bd1cb32dbbbea4eafd860ab1f4ce
6333 F20101122_AABACN supan_k_Page_058thm.jpg
f16a70c935c346f9228d17b87a90bfe5
12d539175c99fdd90c51e94f333fc6a88e554030
F20101122_AABABZ supan_k_Page_077.tif
2d079d73951d3feeb5e36d008bb0baa5
d32a6f16ec35cfff0c243f5b18c7c1d8cec8cd12
1892 F20101122_AABAZV supan_k_Page_113.txt
a53b7e109297f9921793878241d81aff
c4f0402d5f137e86897bff63cfc510c8024eb98c
67032 F20101122_AABADC supan_k_Page_010.jpg
1142d53f488900188e8093666091d6fd
a16e7fcfae47f03679c743513635683b7c62acbd
6037 F20101122_AABBGF supan_k_Page_085thm.jpg
d86b443039f1b356686b9b372a8af80f
42391ea0a605c74a06896ce20eca1aed48f50b56
4554 F20101122_AABBFQ supan_k_Page_076thm.jpg
27697cca190054343da69ad02afc28be
9b9ee7303d304bb7324ed21050985b6eb44d8f6d
21638 F20101122_AABACO supan_k_Page_051.QC.jpg
13a6bf71b627a0060f3cafe08a240e74
6de95067ab6e311953f20acf16bba7e382e5d6a1
540 F20101122_AABAZW supan_k_Page_114.txt
00fc7cadffa4a26b217ed0dc2ff82dd5
98e55568145e5249cb606df3da549e8be9d4e41e
62895 F20101122_AABADD supan_k_Page_012.jpg
8ea14bb8ba946dddee736493ede17370
7c75a20cfeb63662839acca28990ab7314be10fd
21193 F20101122_AABBGG supan_k_Page_086.QC.jpg
e241c48933102a37388cfccf476570a8
09b22b0e9d7bfb673815f6cabc515d62d68def6b
14569 F20101122_AABBFR supan_k_Page_077.QC.jpg
747b4b32f38dd5d18ea5e94dffbacf58
e6e30eeeef4bb3b86c6aa75fd3cbef8050bdd676
22775 F20101122_AABACP supan_k_Page_016.QC.jpg
628dfa365d4b691fa4462dac776ef56d
8cd3b49aba736b0195cfa685cdc987718d2d3593
1684 F20101122_AABAZX supan_k_Page_115.txt
cd24d5541525f0e597b49c801e3e0290
e6787c05c9953bcc9587c8a0b915318be24eb11f
66836 F20101122_AABADE supan_k_Page_013.jpg
72ff0eb7097b2b6db46f60c43aac5cd4
6d0d4c15eca7b12f92ab4c6929b5b1ec89ad6a77
23721 F20101122_AABBGH supan_k_Page_087.QC.jpg
b65e01304a16ee2a05437bb808d5fe15
c57be94bf491ce60fd8493fa2ae04a7ca23abeb9
4408 F20101122_AABBFS supan_k_Page_077thm.jpg
0cbf4be095d77f1cb167de41c800c87b
9b565b1a29f29a7dcf01588cd5952b84ff555f41
8423998 F20101122_AABACQ supan_k_Page_117.tif
0ef5c0bf1c40c0fb8516a08057884795
03fd55aec04e401092a69450eadd4b577774c212
1916 F20101122_AABAZY supan_k_Page_116.txt
a2357c3e333d4866209cdd1e67f5f489
96ceec6915faab49fc4d0612d9744447864bbc3a
63184 F20101122_AABADF supan_k_Page_015.jpg
981552ef30fe4e039ca12d4fbdf1193d
dc566f2260f4a4e5dde4c78f05b41107a5f07bcb
6571 F20101122_AABBGI supan_k_Page_087thm.jpg
df06dc79195eff2df3f98d8916e5dd34
403ab7e5e2fc2c418e83580cc682291a11cf4448
21096 F20101122_AABBFT supan_k_Page_078.QC.jpg
dced28ee525889905d043bb842c1d09b
45bc326ddab40c7bf6b9c0d7856c5afaef9dd8a1
2555 F20101122_AABAZZ supan_k_Page_117.txt
bd43f9a93d3cecfbe6beb2d9782fdba8
75e2f15558883cf089557cee04f0dc29820ec36e
70451 F20101122_AABADG supan_k_Page_016.jpg
00153a90dbdd53eadaa1fdd2d0326b96
e7ca22f7a13af26d139801aff9a85bb4afbf9940
22976 F20101122_AABACR supan_k_Page_040.QC.jpg
d4bc53142c1435fc538287b8e89de66f
9e93c0b2fd1bcac9c1b46c60617d5f932e3a3e8f
19325 F20101122_AABBGJ supan_k_Page_088.QC.jpg
ae2b62b53f5bf10fe2b7b7f2a3d6e024
dd6651226135077dc8d7cb26068ab54061800b84
5786 F20101122_AABBFU supan_k_Page_078thm.jpg
35a5c37ad8442912ec5a48d3d3b90d53
ff21f57d5973d323c7995a363603de3dd8c9b6dd
69979 F20101122_AABADH supan_k_Page_017.jpg
368ad327c40ee8a5efe859f0afde5270
a43204cafe6958dceefb6b058263324f4d635659
215813 F20101122_AABACS UFE0011981_00001.xml FULL
13ab2af0a09d328aa2a7c2e991b5a926
c017be351605e7a897b7fcaaeffa5bf5ed86332b
19848 F20101122_AABBFV supan_k_Page_079.QC.jpg
80b0368cf89b4b816841d600ac8983da
7ad5248156f18cc5d7e69e246bc65c8ffd0f5161
5440 F20101122_AABBGK supan_k_Page_088thm.jpg
c2519c1daac43d868c175115bbb01323
29bcd203ba3219db7f561412c077d14012131d7c
5815 F20101122_AABBFW supan_k_Page_079thm.jpg
8088263952b1db604b613002d7525e1d
38b3d9c2af865384af0033cad30d3b4e56ae0bad
69343 F20101122_AABADI supan_k_Page_018.jpg
eeb58a2cedac5ac6173fdf897c5a221d
9e1c8f59d534dee938054e87c90d921dbd5b1d77
19160 F20101122_AABBHA supan_k_Page_098.QC.jpg
4acbda7ab1045c989a9de68ff783f5a4
54a63ed1e18f6d0835c862a833ee366fb82aab7c
14622 F20101122_AABBGL supan_k_Page_089.QC.jpg
95e03706ff2456d6c851d2d4bf88d876
849982ff687abdb3d1e73e52d52953ac8c7cc1a1
22845 F20101122_AABBFX supan_k_Page_080.QC.jpg
953f2f1174190120d81d8549fb3032ac
38003e00d753e15c2b21c9135acede3e6370aa2f
69046 F20101122_AABADJ supan_k_Page_019.jpg
f3abe026d136a91df85656d4bf4c0501
861abd76ae75819951001d3b632f2c8e48cbc5d0
23798 F20101122_AABACV supan_k_Page_001.jpg
70d4a87b0a07883ccb678473e6fb989e
58be0cef56c6d0998fa95667e22e02fa096f5373
5558 F20101122_AABBHB supan_k_Page_098thm.jpg
4434fdc92fdc4beee0e4e09febdd2b02
14ca8496998609243a40374b158af517d7c3583e
5171 F20101122_AABBGM supan_k_Page_090thm.jpg
b6e72b6e8c32298d7b9392a39035b1a5
51fe44418ba12162057bc32c24ddac6b7bb1eabf
6204 F20101122_AABBFY supan_k_Page_080thm.jpg
614aaae1385877b02505c677670a8549
4731e2a2cf22fc6e2503928c208a3cc80264f560
75778 F20101122_AABADK supan_k_Page_020.jpg
0f000ef23b106886d314cf2c8ecb33f5
61759af44e777686905c61f36b438abd06108ff5
10277 F20101122_AABACW supan_k_Page_002.jpg
b5f92e60abf2e1036c4dc8a76961e898
4e1fe05429dbfdbb6e20eb91388f27f04a5d9e99
12808 F20101122_AABBHC supan_k_Page_099.QC.jpg
8e390bb1de29bc5758ed492fd6bb3320
1f7bf3f0dfef04d8e6eee4920f989037b2f05047
14388 F20101122_AABBGN supan_k_Page_091.QC.jpg
c4a1dd605adfe675d49106c3bbbf5da3
9866b910ce8fe3a73a3833e8e565fb95afc04d7a
6376 F20101122_AABBFZ supan_k_Page_081thm.jpg
b40bd1148a6cfbb55f4c98a69a8fa38e
b1876808a1a82a523b3aef2d9e13f1cce0c0f39b
65983 F20101122_AABAEA supan_k_Page_037.jpg
ee606281cabd1c3429e6366cb76b9e87
cd948833ef76439c8d1af417c343cfc71900d83f
71218 F20101122_AABADL supan_k_Page_021.jpg
d11511bc391edecfd6ad1638f8daa4fe
4141d94dd6f0ab6d37fe13ab46862f35bb5caf69
11431 F20101122_AABACX supan_k_Page_003.jpg
668a8d2a9904f50e8e79018c70a25896
dce861a9fc21fa2430b3692c455a2ebbed7b5885
3850 F20101122_AABBHD supan_k_Page_099thm.jpg
bf6b6442fbd3870d44bc831bd883331c
60cff01b2490b0a4e25b33290728696974cc6ab6
4440 F20101122_AABBGO supan_k_Page_091thm.jpg
b62357602012ef6e06d4bd9954063870
0b81636c3fbb31a0a7e63aee50efcf1b8dedfc9c
66222 F20101122_AABAEB supan_k_Page_038.jpg
a7465b5861e7d272770bb76112e9526c
d6d0fd45f996dd1641648d554b2e36f9e24b8178
66860 F20101122_AABADM supan_k_Page_023.jpg
b8efa8e38a22e71a1793e6db79208a9e
ba11185f3e84116196a6a60c061fb2d346babcc7
61986 F20101122_AABACY supan_k_Page_004.jpg
3d26f5febb3a9a9c81635b34a1cc170f
d7a72bbe913e36018319ce3d18744825b0041b0a
19212 F20101122_AABBHE supan_k_Page_100.QC.jpg
5f9a7aba21975159359f14aa2b8f03d6
db9de8f68a5de8516bfb69b5a8788af2b829b70c
19438 F20101122_AABBGP supan_k_Page_092.QC.jpg
6a832daec165874fe20b9c3bb4b45e6a
b4c79bb8d4415c7ce6257301089bdcb222413c81
67866 F20101122_AABAEC supan_k_Page_039.jpg
f28536e52e84444595617ca53056b7c5
58302be3fe7413c0eff2c7b22fad55962e0340a1
67129 F20101122_AABADN supan_k_Page_024.jpg
06083c345456c20c1a88148a7740b40f
8e9a4b5bf8246852e5e635fef1e2105135a4a55a
62705 F20101122_AABACZ supan_k_Page_006.jpg
39ab8865fc63bebb3b0c11de409b64b1
2a5a6e92581bed146e9ad729823699d659152063
5944 F20101122_AABBHF supan_k_Page_100thm.jpg
5ea06835624d6a59bb604e6fb95d4fb9
6515f5439db6263478161ae286f6d9f661deb73e
5629 F20101122_AABBGQ supan_k_Page_092thm.jpg
60407eb9e395e6cef80d271e81e5f6db
03e24d8ba854e0aad00b86621f6bbe9744f06059
67034 F20101122_AABAED supan_k_Page_040.jpg
9907f117606d88a5a8c6b7aa4f29cd58
96953e7b9855657313c88e817e7dfbf06a399190
61889 F20101122_AABADO supan_k_Page_025.jpg
d6747fcfa272e226b1c00d0145d7b924
c4ebf050b1d981bae15d6139e81d1b4c45b787a5
15984 F20101122_AABBHG supan_k_Page_101.QC.jpg
b1b117a086b29cd9ac906e57358840cc
04256d5c89faef7f49278a66438a855d6d27546b
20347 F20101122_AABBGR supan_k_Page_093.QC.jpg
02a159b0efe6f251564e2011f9e28a47
a17c8dc97c64dc240cf3f0d749cffdba0323261e
72365 F20101122_AABAEE supan_k_Page_041.jpg
638fc2485e7f478725f173f0bd0e5be1
34ea9bd0c6dc74c9ca6d6cf6d2c49a3c84d4fd42
74005 F20101122_AABADP supan_k_Page_026.jpg
4003c50725f84b0ce25945028d2098a0
5d4665eb15b578de6b602f396dec2bb53acc03f4
4918 F20101122_AABBHH supan_k_Page_101thm.jpg
09bb6871d3fdfad084c95635370cb5b2
572e2aabdb8be72011792d15f8d5ee507eced65c
5585 F20101122_AABBGS supan_k_Page_093thm.jpg
2217c030e35e536bb7adb9e9efafab9b
52453e81f40b6424e8d451e3f03df55496213591
68650 F20101122_AABAEF supan_k_Page_042.jpg
21b5ea959073dbeeae9f822ca6d11ac9
83a34854bf0792d603d64b7498a9bab23cde63b7
61277 F20101122_AABADQ supan_k_Page_027.jpg
95096408e8024f41009c5bd57693af1a
4b066136d19d5304d84be5b6865edf92c5c7ffd9
20533 F20101122_AABBHI supan_k_Page_102.QC.jpg
278e555d7433a4720582a8a65a59b182
9a944e04d698228a02483ed57fc573baefece256
20087 F20101122_AABBGT supan_k_Page_094.QC.jpg
bfb4403b61753c751a14feca9e8b4e6a
759c3f1661162c0f445cf045c67e383243dcc165
71901 F20101122_AABAEG supan_k_Page_043.jpg
19efb5c0be45c842d84304be6ea4c787
c4689400b55fed2e01ac89fbaf696d2b2755f095
70768 F20101122_AABADR supan_k_Page_028.jpg
63ed0bdeb66ec03d8be2356f561cbb8c
6a3f12112b9b3b5b919b0eb151c4e3846d48c817
15770 F20101122_AABBHJ supan_k_Page_103.QC.jpg
59db190af9a017c4a97ba753dacfca5f
577c7d70934eb669a3cc860c3766b6416edcd5f3
5488 F20101122_AABBGU supan_k_Page_094thm.jpg
bb67eb8303908d2b8ca5259996b3a0ac
7985e4e9b5d19412df1aa24786f63e1fa2747e8b
64202 F20101122_AABAEH supan_k_Page_044.jpg
ca959f5400c55bec97bd3dd98bfae110
5e7f4fe0764006968569b2a07a4ba54bf79282ff
67234 F20101122_AABADS supan_k_Page_029.jpg
05a815c6852a634663913e521a58fc91
2ef44238b58db0192337f1a437ac1eeaba409618
4459 F20101122_AABBHK supan_k_Page_103thm.jpg
98b21dd7dd1248c84c419d418a6e85bf
bb74067b6d8c6e7e85c2e246de62448c38c1622e
18152 F20101122_AABBGV supan_k_Page_095.QC.jpg
cbdb9dc543cfb3c798451dc0d2ee2a24
9efc0e88fbf05c8359138cc1f96bfde5b2f2b73d
64776 F20101122_AABAEI supan_k_Page_045.jpg
5b1d638518e40f2f89b48ac55829a465
32920057225cda30b4178007f8a1fad7c6434b29
69881 F20101122_AABADT supan_k_Page_030.jpg
037c284d934fa5f04c7ea38ab1557dbf
3343cccd8eafe31a2d69952ecba2e0caca925c5a
5387 F20101122_AABBGW supan_k_Page_095thm.jpg
efabc5f312bcacb31779098d94aed7b8
cf803f8fc7a4bc33b6e15dab1ef8f4662d93de7c
67836 F20101122_AABADU supan_k_Page_031.jpg
e72d924c064f42ffb17a645ea82f0fa8
7abefec00db662017418e693461741ffb8a93740
10344 F20101122_AABBIA supan_k_Page_114.QC.jpg
2945bc6026876522ea445fd4c50a1bbc
5c8f677fd2cdbb731c37781bbca5176a6f458715
21637 F20101122_AABBHL supan_k_Page_104.QC.jpg
821cd4cde0a27a12ab895b85e9d323ba
421b0368f7587d6112b1c3decfdb8844bafe94a1
17677 F20101122_AABBGX supan_k_Page_096.QC.jpg
8aee29b0fe2fcc26e282466e8bb3b25a
0ec711b69dd1d8043ac687a3f8d7f7fa555e2bf9
58102 F20101122_AABAEJ supan_k_Page_046.jpg
39cb9ad636a35e048f19a60ef6f4f3d8
a2173c4c1b0f705f3f668b67b332ffd21ba31aba
66400 F20101122_AABADV supan_k_Page_032.jpg
8d3c5ba3cbbf461706c1fee527f2b8e7
5a2d28e4bfb601024d8dcc21306b9de93d145c0d
3112 F20101122_AABBIB supan_k_Page_114thm.jpg
265423505ac2ee1882ddd04b71574f1d
40787a23249326da928fde5d3d357d870681a7ad
6403 F20101122_AABBHM supan_k_Page_104thm.jpg
79d388f5b9121c52a9445914287e6234
cccb54d7fe1951c064e1303b4dad03e80ba49f23
16516 F20101122_AABBGY supan_k_Page_097.QC.jpg
9f457bd75ef53e3b2b49293790ad22af
cd0a37cd93b4683273ad7bf8d2d7a086ad8af923
60327 F20101122_AABAEK supan_k_Page_047.jpg
5584f89f7dd62478c9a8bfaf921b2dc8
980117c002be5137b15847729638bac6c59fe22b
57739 F20101122_AABADW supan_k_Page_033.jpg
03831d1f3d1de307ccd7b20dbcfa66c2
79b2a1756ea35c08a512ef3e655f79fe58b4f379
17131 F20101122_AABBIC supan_k_Page_115.QC.jpg
3cfdbbbf47d255d5426f7106aca69e0c
94e678c8359d1e8a6c364c99a5727ba9b808b6d8
20844 F20101122_AABBHN supan_k_Page_105.QC.jpg
eea379141ec605db5b3b746c004b1c02
a100c49ddf2d9ff8581e094d102c0125582289d8
4617 F20101122_AABBGZ supan_k_Page_097thm.jpg
e6806afb3e2549812016f92ed2e9a0b1
27999bac8b40d46cc2ffb8ea3dd4e606bac86092
63327 F20101122_AABAEL supan_k_Page_048.jpg
5fa31af45f63d6ab1f1222644d012d05
d5ece62a8a556ff9890ba95698192a3395ad7211
60298 F20101122_AABADX supan_k_Page_034.jpg
cc993a5fd807d3e6910246c4da90a48b
224446388ee5ea71ca0fac2243f6c95945ee30db
53326 F20101122_AABAFA supan_k_Page_064.jpg
a4c9ca2a98776474c72787ffb1230bb0
146692cedcf69f276c655ab1927ebac0114a054c
4793 F20101122_AABBID supan_k_Page_115thm.jpg
6f9b673d59ae4e68522ec0e2f5f05266
976514e138eaf01ccff1fcc67a466f1c2200084a
6156 F20101122_AABBHO supan_k_Page_105thm.jpg
faf39752b633a02426021351fc9f52e4
a60e1f963fe3f70681e58ed4ff48ce0e97dbb289
62831 F20101122_AABAEM supan_k_Page_049.jpg
f54c9cea6aed0da3f47267e75a9a716f
ac2765c83999317796d76549885efdb67eb76e80
67111 F20101122_AABADY supan_k_Page_035.jpg
84f895b54f8ba03faf2c3172413ba36c
d45fb6d2fb6e9bcf48b7c3d3dd16f62518dbf3cf
64240 F20101122_AABAFB supan_k_Page_065.jpg
213193944316d419970ff0509b66bdd0
a357a22c9118e15eb1dcebaf0ca8b16ee7217ee2
22178 F20101122_AABBIE supan_k_Page_116.QC.jpg
7da5d6d63ebc6402f47e11177b2b6697
033d2c8bebc59ec327dad88e52d4091306c1acd4
23023 F20101122_AABBHP supan_k_Page_106.QC.jpg
b8f1e7f002025bec328da1b755731173
183c3072ad6b78d5b2186d6faf3fb6eeefe2cc4c
65130 F20101122_AABAEN supan_k_Page_050.jpg
427a0ae1af65b182979aad1dee1a27b9
0fb3fae68551def41bb3b3d35779701f7471a9c6
66306 F20101122_AABADZ supan_k_Page_036.jpg
36a67bad6510ca498efe07ccb91e72c1
4d463c3b3caf5df37a4ba70acda45a2352e1228e
56186 F20101122_AABAFC supan_k_Page_066.jpg
5c96d2cc2cff6c6c166fbbfa73c4f464
99be936735cf0c0fe0e5725653c6fa2c5b155cef
6458 F20101122_AABBIF supan_k_Page_116thm.jpg
270d5e9dae463baacc487f37242ad9c2
d4829ffd180aff3743dd495a7e87a8299f4528ae
4664 F20101122_AABBHQ supan_k_Page_107thm.jpg
2da667a36d06532ed3a29f941747f525
d48a426e4d07e5ff20f29efa23ba02824fbc76ae
66025 F20101122_AABAEO supan_k_Page_051.jpg
39e1fb6bbfe0bd03cce1b54c27a10d4a
fc20ce2631a21d8f14f1f3bcc868d397ba8a668e
11669 F20101122_AABAFD supan_k_Page_067.jpg
1945f2fb4041278d2237a32449917d8f
714472189e10c077103a03ff8cbdec6df179c3e0
5609 F20101122_AABBIG supan_k_Page_117thm.jpg
8efcf9df4527d17a0d1df1a2f6597bf1
11e125b49a94203eb25572f82fe1475368388829
23701 F20101122_AABBHR supan_k_Page_108.QC.jpg
124fca8333f7b72271b48852601f7a24
927c7e1341cdeae34660b1ca92cab142e0491c5c
64124 F20101122_AABAEP supan_k_Page_052.jpg
e08cc17d8e4930fd92b087c973613830
03638c5f74ef0d0355bb3b6b9e625860b6abbe87
61756 F20101122_AABAFE supan_k_Page_068.jpg
98ffccdc99056f7cf88aed0f4f7ca51b
cde6640bb0e72f37d2dd9c50e37d4c8ffd855f4a
20540 F20101122_AABBIH supan_k_Page_118.QC.jpg
694d02ed8812988157d28084dac213cc
dca0b5fefe3e923a4aa9f9d91062a71659a36560
6631 F20101122_AABBHS supan_k_Page_108thm.jpg
abe794d14105c797dd68f8b35d454375
e067276efd9f8f98e761625eababe85da22ffbe1
63430 F20101122_AABAEQ supan_k_Page_053.jpg
9263744ae58d75094e718e508b647cbe
c1c5ef60c36cc52a7587b42341c0747f1700fe15
37368 F20101122_AABAFF supan_k_Page_070.jpg
a18f482508c850f785d64e8fc97b02b5
e87f544654f03c202808465a25a5f2e7402d5892
6093 F20101122_AABBII supan_k_Page_118thm.jpg
b3c33fd1c04a6e1515bf678ede60a6b5
68a5c77033d771142f3b5be173a9c609fe24572b
20798 F20101122_AABBHT supan_k_Page_109.QC.jpg
3a2c3c6a248a92b968844aed87008e19
87e61c7cc0ba26a76c5c411385426066ee955fcf
58446 F20101122_AABAER supan_k_Page_054.jpg
1f25080b77156797c0f30ef192d91ac9
77e68c6597f9c30f754a0c654b265b3decb29a7a
47622 F20101122_AABAFG supan_k_Page_071.jpg
81dc63f19195696945d7c83ced3920bb
1a5b82a9a1b6e1278a3a79f457fe57b30a540704
11294 F20101122_AABBIJ supan_k_Page_119.QC.jpg
667a1fa946b615568b1857c389d97d9f
9feb418ed8125d7c027c036443e15a2ea487ca17
21000 F20101122_AABBHU supan_k_Page_110.QC.jpg
cc3e885c1f146d24dbc20bf920320585
f56b3e4fba3c08dc5ed7862da65217b221c03017
11024 F20101122_AABAES supan_k_Page_055.jpg
b7f19e0d838dd19b49abc81bc81624a7
afc0afb086c03c699e4a8a66a2e335b5ba344be8
50255 F20101122_AABAFH supan_k_Page_072.jpg
0a5673a03dfb7257da9ba4e1606ea8e4
0f0524c0ec2e3544716db439d53af84c1f8fee9b
3174 F20101122_AABBIK supan_k_Page_119thm.jpg
4e2bd3e2a2453ce8cdeba579c0a60ebc
017788afacaa5f37a889c74d6ce0f59138c5b714
5857 F20101122_AABBHV supan_k_Page_110thm.jpg
30f94cb02357bf19a96600d30ee86729
6fc360f07935bcf58a857f605f2f7a7c6cbb3218
50429 F20101122_AABAET supan_k_Page_056.jpg
382583581af0d9e0f83297926475dae5
cf4b170037c4c0dd46a2542b276dacaa5513eca8
65053 F20101122_AABAFI supan_k_Page_073.jpg
7d1523f1cdbe2582e0098af4b5d4e057
66d62302c64948d3941ee449d9de7daa5e720b0d
19626 F20101122_AABBIL supan_k_Page_120.QC.jpg
ecd1884e93163812b317e51be003e4ad
1e22b2a1dc33383d3278eb81be02945d293e45dc
24370 F20101122_AABBHW supan_k_Page_111.QC.jpg
46d4cce61beb07c12f73e47b6f8aa110
8c0badfa712fa07432682b33ef54d70d4d0b7fe3
71239 F20101122_AABAEU supan_k_Page_057.jpg
95362462419d800a5816e9264c03a0ce
e742e2a8c7cd96040573b071e074f49bed272702
53269 F20101122_AABAFJ supan_k_Page_074.jpg
5919d398d96f7c9cd62045c76ca1bd15
ef66e0871a24a967c3382a5a86256257e4f49c80
3665 F20101122_AABBJA supan_k_Page_128thm.jpg
bbef23e52f4943adccaf7888f4fa8bdf
ae4706fc991ba41af8157adf12c0e486b682dae3
6608 F20101122_AABBHX supan_k_Page_111thm.jpg
ddece582bf127858b928ca820325f399
28167c8392a0c9fa5ca57c0f8203d4a31c5e3e62
68124 F20101122_AABAEV supan_k_Page_058.jpg
219d562579787bf226218de12f99f9e2
b78bd27f2d315fae3ae3da1ca74c1e2490e86317
12766 F20101122_AABBJB supan_k_Page_129.QC.jpg
8772383d163d6179920ee10b24ddac3a
3583cbc9a2df1cdc6e2437c420312595e10e4cc4
5608 F20101122_AABBIM supan_k_Page_120thm.jpg
83ca94091221f9d390039e7256d1778b
afa3f9eaa5af3fcb8aea1f010062047bb5444482
5780 F20101122_AABBHY supan_k_Page_112thm.jpg
6d54f02cdd519481370a29dbdd9b0980
9a6ecfda8ad33d3d69dcc30769d2496db912311a
67804 F20101122_AABAEW supan_k_Page_059.jpg
307a129b6338cd0f4345d549587b2fab
25f64490098eef394306ddbe53e453d6e1a6950b
64541 F20101122_AABAFK supan_k_Page_078.jpg
1e3b6a8dae65fe0d4921a003538221fd
c6c3ef7574a8dd3de273b5a7b23e604235ec5e89
3594 F20101122_AABBJC supan_k_Page_129thm.jpg
5549b0007f2ca04bbf530b0166e19b98
06b65c4940578bf05dee8a6dad7c05fc659d09cd
6246 F20101122_AABBIN supan_k_Page_121thm.jpg
adb26f4d29e6a8f8819f337f51325f97
773c3915a766bab6f573bd107b4a0e21b1c55f1a
6400 F20101122_AABBHZ supan_k_Page_113thm.jpg
6603b5cfd1472ff7188475a6217df616
17dfa62da24cfd3d38e8c7ea905cb3aa34e0f4a7
67770 F20101122_AABAEX supan_k_Page_060.jpg
3c376e4fc8ec2eb38108b91ca80d5394
f05bad1c09d93ff6d792f5ceaa8228d03746e1eb
50962 F20101122_AABAGA supan_k_Page_097.jpg
cd438abdcbd8b77c611869a04d959384
31210fc674d0fe88fb13202986dc44c6b316dda6
63655 F20101122_AABAFL supan_k_Page_079.jpg
eddab8ea98eeecfb94da80f9c21f5b37
f9d15329214528615ce5f99dff6f00b944a6444c
14455 F20101122_AABBJD supan_k_Page_130.QC.jpg
170a2b7e82f5f9319cad47bdc6271fde
97b3f6e1d70aa9e331bc693031f57d6d247a7a55
22531 F20101122_AABBIO supan_k_Page_122.QC.jpg
acfb16a04dc53c05d4ad7dfc8029c6b1
354230cb8f6e1c4f6757014345d2809d1bb32ce9
59324 F20101122_AABAEY supan_k_Page_062.jpg
d951e136505d4d15317a979d135f1636
f73cef4d0e458a33b5612ea3b4a15b230e0814ca
60137 F20101122_AABAGB supan_k_Page_098.jpg
142716765341e05d6284c9217f15b7fc
8b923105ed8d98eb6db32e425eb31a8808660c39
71557 F20101122_AABAFM supan_k_Page_080.jpg
19bb9a24b4979399dab16872d5dfba0c
882f1b0c89821fef468eac95ee1022b567cb789c
4616 F20101122_AABBJE supan_k_Page_130thm.jpg
f8a364a3d38cb94f954322ab28b76f0e
4f80f186c14c0b43cdcad40695a3f5898a59e734
6141 F20101122_AABBIP supan_k_Page_122thm.jpg
875d666e5c7bd068ec9cb68665545337
bbbeef86d9589593c9ec93e83cac1c081758b8ca
63511 F20101122_AABAEZ supan_k_Page_063.jpg
aafe080a87a3913cc8ac5734c3292570
eb353dd8b00b4ba3f787986b8228973cf4593b15
41423 F20101122_AABAGC supan_k_Page_099.jpg
82d91120b1a7c15648a6b620261cc063
6b86d929e2595b91ad64a417638d4b6d04d3e4a5
69720 F20101122_AABAFN supan_k_Page_081.jpg
166ed378852b0c36842c71b20648b6ed
05d49b976e1e692ea05bca54cda7dd8c3886ba62
14204 F20101122_AABBJF supan_k_Page_131.QC.jpg
282b01913d9537b50288cccdffb4b833
836ede662c30d72f518e8ddd8a404caa26f70289
6623 F20101122_AABBIQ supan_k_Page_123thm.jpg
0f7f75a5c23916129613921b297afbc9
426bace011c933e32b8f9537323108787219983a
58750 F20101122_AABAGD supan_k_Page_100.jpg
a79b77bbd79009f659516c282d3bcbe7
3fcfe0f2b848ccdb9241aa787b61bd6fa0511b23
70529 F20101122_AABAFO supan_k_Page_082.jpg
4dc1a3e4b594d247ebc1c78326240749
3510cc95789d1d685c4b2a4cf16a93d0dad87548
4821 F20101122_AABBJG supan_k_Page_131thm.jpg
0968fb1f54e94ca903bd7e8d39cf4e9e
89d1b1541aabe076591b6d4e6a348c18e950cd2f
26257 F20101122_AABBIR supan_k_Page_124.QC.jpg
c8f7202a795e8052ae0e8560a6ece943
7103ead2f10acd74af5392d6cf13e11d8eb6aa75
48319 F20101122_AABAGE supan_k_Page_101.jpg
980286191e6bee52f450bc21eb51ce46
1a592e6bfecc6505e02a4f9ff9a07050e0f73ba8
59580 F20101122_AABAFP supan_k_Page_083.jpg
b0cd11f7213b0b8adce0bc6c9eb665ec
804092454754c1f41888b6436ae8331d76e9c1fb
13923 F20101122_AABBJH supan_k_Page_132.QC.jpg
00309c1c22bffc0b8a488bc671ee3d04
3dabc17f0f690dd067e37c00a8d5ae331d352ec1
6757 F20101122_AABBIS supan_k_Page_124thm.jpg
e3a1b39c4f483f1331a1e7e13725831c
a38f01d60c5ba71ed683c3a9426a4d5e19a33c7e
69255 F20101122_AABAGF supan_k_Page_102.jpg
e2bb31688dc537eba79410ee20ff09f1
0523e9dd9980407c8aaf539225ae209a4c90b132
65294 F20101122_AABAFQ supan_k_Page_085.jpg
7857ebbafebed1db31886977041c2292
67ee658617c2e1fca100b6c7fec2a527f643a9a8
4241 F20101122_AABBJI supan_k_Page_132thm.jpg
9384ba19f53b367fa1b0ae5d0457ad2e
d799f7f4038f0ecffe84625943b0374248c63b8c
6822 F20101122_AABBIT supan_k_Page_125.QC.jpg
8e17343f7061e04431943c7397f73bc1
6f554aa2c2e1412120bb92b3310fd52662c0ca4c
52064 F20101122_AABAGG supan_k_Page_103.jpg
149bce431839ff59ebe9d07929f66ca6
cc7af74fbab0b29eeda2fb7ee229022e37128762
71397 F20101122_AABAFR supan_k_Page_087.jpg
59b4002ab8be56affd2b8f4dd5c0ef77
af1a90b0567930baa85818fe2e928d6ddc765293
11334 F20101122_AABBJJ supan_k_Page_133.QC.jpg
7d7ea708f2388b32c48fd8879746ad28
eac8ae9899675355a4901640bbb327da4b89d5da
2180 F20101122_AABBIU supan_k_Page_125thm.jpg
d7241b9e091ed4a4bf6e924807fe2887
19e57297a7f0155497c87f774a7f30b44d9ef5ae
67703 F20101122_AABAGH supan_k_Page_104.jpg
0e4fa3b01ceb8f98b6c597e3fb473c4c
b67fa1436be4f8aebc8278ba7b6face93e048ff2
56243 F20101122_AABAFS supan_k_Page_088.jpg
280e9ffdcb4ec53b472da0c6cfafad3a
85d57f487bf4d06cca329464b7cfc7f6b016351b
3228 F20101122_AABBJK supan_k_Page_133thm.jpg
2148e881588a2849807d3e830e679971
99f51b703df4f42afa86bfe74f76862efa5389fe
3341 F20101122_AABBIV supan_k_Page_126.QC.jpg
37492af3444ff3ee2d85103da9afd65b
3d2330de69a610b301ace9305e09d9fca0fea404
69011 F20101122_AABAGI supan_k_Page_105.jpg
275536105494aab2fe6dcd938083effe
39ee68dd90b9a45b5a6db3cd647cb018c60a0480
51721 F20101122_AABAFT supan_k_Page_090.jpg
b1ac0d4c5a2ad3173711b06032ac8d64
60e0267dc2f4295a544ed3d8aa5e72af4f409c36
13536 F20101122_AABBJL supan_k_Page_134.QC.jpg
2b836d7d50ec68f838e7e895145c3b90
2999b86845099bc00c685317db5373c8ce4bf95c
1351 F20101122_AABBIW supan_k_Page_126thm.jpg
349f8efccca2f5135e9374a91257849b
647634aee73bce85fa172cd5341903b342861600
70611 F20101122_AABAGJ supan_k_Page_106.jpg
1817658a88e02d8bfb7d264d25fdc77a
900415d08f1ecd93bb55e3f4dc3b189fc21aaade
45029 F20101122_AABAFU supan_k_Page_091.jpg
49c14a7f3e26037a436f5fb517cfb1b2
3cf964f4529c1dfe86ef411bbe5d63946281c797
6617 F20101122_AABBKA supan_k_Page_143thm.jpg
d7074aa1905320ef0c540fcbeab4f4d0
8cca865c158d92d4cab7068f9cfdd54a0f0d6a01
7631 F20101122_AABBJM supan_k_Page_135.QC.jpg
06d8306bc104a4ad783b8647d5726653
5f4115b797faa705206f274ded6e793a615858be
13268 F20101122_AABBIX supan_k_Page_127.QC.jpg
57ad7a3e6e275b899005579dace062ea
b8d47c8b71db27c3080ae1e4727f4ba4dc0f63df
54534 F20101122_AABAGK supan_k_Page_107.jpg
6fdbb37f3ca174b5d866a972fe04cae2
5caf0fe9f828407aa1198232111a609b18f5d113
60242 F20101122_AABAFV supan_k_Page_092.jpg
2d847650e7f1799f5d340c127eeab437
55adc08f3888e076cacbfb583345bce4b05b113d
17894 F20101122_AABBKB supan_k_Page_144.QC.jpg
925199e840388043b864720766a7a9c7
398fbecb5414910fe9c87a8ac1bce28e1deb5194
3720 F20101122_AABBIY supan_k_Page_127thm.jpg
e391dbdcb31d5c42a75a1a63510a16ff
db11c3435ff9668777a5f350a9575ac601a93863
65305 F20101122_AABAFW supan_k_Page_093.jpg
b311287b371bef4c20ae04e2cdf49cca
a817f03691787a665382e715c3ea37247aa8b568
F20101122_AABBKC supan_k_Page_144thm.jpg
027d9a1437e677a259a70dddb4b57995
e222d4c8307a058b01f1a9f0308628bbebce90eb
2688 F20101122_AABBJN supan_k_Page_135thm.jpg
72abd4fa4374bb0989f9b0d6826ca988
8c30f3797458e5340717dbc4189ed5eece45b8be
13444 F20101122_AABBIZ supan_k_Page_128.QC.jpg
098d349a8ea1653f5882fd7fcb03f913
bbc3fab49f33946526663028215033125b89ce5d
94516 F20101122_AABAHA supan_k_Page_124.jpg
16ba4635e0ed541b0221b21beaab8db9
18ad4a3ad98a61a7c8110f0616b4cda6bbf72bc0
68451 F20101122_AABAGL supan_k_Page_109.jpg
895452ffd8960d6aac602a16184bc84d
6c6f4e735bd11743b197d1b461c644871f75bd4c
60991 F20101122_AABAFX supan_k_Page_094.jpg
8dc27498e2fac8638e10a51720d8e3f6
4e9150c735bcd60729f87ec7c4e0a3a0b5673188
3023 F20101122_AABBKD supan_k_Page_145thm.jpg
4abeee54049e5c0a9da8dbeeb0ad375c
ba6bf8b14ba928a5486feb24e208abbe5d2a2a47
16946 F20101122_AABBJO supan_k_Page_136.QC.jpg
2f1ffde6c1f35f32fd3b91c17b932b9a
6eee1cd6ab1aa1e2381f226da741cfa731138e9d
22754 F20101122_AABAHB supan_k_Page_125.jpg
98dccaecc7ca5656158b9a4309601307
73d2967767099358b7ca1a5514a9fda17e8a615f
68313 F20101122_AABAGM supan_k_Page_110.jpg
919ea8bb58eb1d4accf02219667069e2
8e65f63e592db72aaff03555566e44b17f4119f4
55063 F20101122_AABAFY supan_k_Page_095.jpg
dfb7ce15ad415ea03e991e9c0ce27bfe
6e24adfb78929bd92ef21106c9d2b6b422bc7920
166715 F20101122_AABBKE UFE0011981_00001.mets
0c8477c7406480b4fe070a885a6fe1c2
39fce9fde094a40b48471cc73147259f4df0f68f
F20101122_AABBJP supan_k_Page_137.QC.jpg
0c2a6dbbc7e4a4dc5e29e4afebe17462
b0408078a66d5541b68a2408b5016ea76a806982
10425 F20101122_AABAHC supan_k_Page_126.jpg
8e7f5e00c3f118e4ed9c51b663997855
b8005acee1c050b0d400523085b217191571e9b0
79589 F20101122_AABAGN supan_k_Page_111.jpg
92cfca0755c8f57ebdd1b65bd4210858
a2c510d6ffe5d6969c97bdbc00be212f9a9e97d9
55747 F20101122_AABAFZ supan_k_Page_096.jpg
cf66921037ce4272a6aeb91cb42e5bd2
18e8df827b7cb60048af649aa0b63ad1b97d0641
4470 F20101122_AABBJQ supan_k_Page_137thm.jpg
5a982e76334cc7275112accb0c1c2ecf
02da0bd3b4dcd943421ac4d47def39077c7b8b10
43745 F20101122_AABAHD supan_k_Page_127.jpg
822b1d603208f442667d1102253e205a
30dbf860ceea92af77b07b9ae0ad8790871f6e0b
64212 F20101122_AABAGO supan_k_Page_112.jpg
860f310b9b7e6dfbecf49f058221d7d2
db665562834043426c2f5457e20fa9a6e784e525
15338 F20101122_AABBJR supan_k_Page_138.QC.jpg
a8cec39659e1026ad366404f270ef798
4fa3db1edca953e759925ad12e3adc5d1be49ed6
44934 F20101122_AABAHE supan_k_Page_128.jpg
3eeeea2dcf8462ea70221190339be9f9
191d8c17ee2c63c6e1d563b04033add3e182eef2
67168 F20101122_AABAGP supan_k_Page_113.jpg
1df950b6d9dbc7350ce5134e0c35ca3b
79c19a8a655ec914f27dc69ab02217e59846571e
4713 F20101122_AABBJS supan_k_Page_138thm.jpg
780cf081e5717cbdf39887ef0c66e8bc
cf29cc03c45d2da1728ea0c7284af15fe10f3862
43406 F20101122_AABAHF supan_k_Page_129.jpg
c6d1a141ff341420fce70ebb88d83d4d
094823d5ac67796e41e400c7aaf7990563764626
31796 F20101122_AABAGQ supan_k_Page_114.jpg
4e172905dedcee9517410ba6e1a13e3d
2ef1f8b8eebb070e41e06bdbb009976011e2f965
18783 F20101122_AABBJT supan_k_Page_139.QC.jpg
933687f0a0d5a5a5f225628f3f2d5637
49e97f83e1e7aaf26cbc7c6cbddbcc7906acba0d
44530 F20101122_AABAHG supan_k_Page_130.jpg
cf56398bdc750d3b049a13bc5638671c
94ee5f9f414325bb2758be58420b1ffab36796f8
54391 F20101122_AABAGR supan_k_Page_115.jpg
454bed602b50e54b13f86a94d050d9bb
5611285700bf4ea155295f2f34a45834f7dbc1de
24708 F20101122_AABBJU supan_k_Page_140.QC.jpg
597e24e4d736d82fabe3ede59bda2d94
fa5b4aa2b706e3397706aceac1f802d70e0561c2
45448 F20101122_AABAHH supan_k_Page_131.jpg
252b0847f3263c9f9413bfbf2b7ba911
43a9504013c27161f9fd755c4a82c1a192ae101b
69664 F20101122_AABAGS supan_k_Page_116.jpg
7693bcbf85f3198c1de41f8d9d58ceaa
c8310ede2b911cafb6b38b2f7d7a932746fd4621
6603 F20101122_AABBJV supan_k_Page_140thm.jpg
d8776ad34e223b8148e79efc69110523
fe57849f30ce91f6056e05226704cdc2a6b1f0eb
44876 F20101122_AABAHI supan_k_Page_132.jpg
a0cccabf4cdc54cad6db397e7b170e04
2fb81ea428372a054feadd1146a116934032fc4c
67580 F20101122_AABAGT supan_k_Page_117.jpg
09625131eb3f5676c757c57e9b1893df
a8a5eda2b8047a5753fe3c68ca11f55e3f577a05
24347 F20101122_AABBJW supan_k_Page_141.QC.jpg
c692e2004f626749427935bae967dde2
4e501d1130303435ae247ac7fff5ae4b63e89203
38145 F20101122_AABAHJ supan_k_Page_133.jpg
2fc1874189d72fac1d90d325b0d66147
8500b2d9483a5af53cbb6565a52223f7396dbba8
65172 F20101122_AABAGU supan_k_Page_118.jpg
5dcc45b290a0be04303e6300b9431e6d
2d98a12464d04a148588cb0b78e3d6ea43f2a930
6722 F20101122_AABBJX supan_k_Page_141thm.jpg
a0f1a4fa1a0edcf2097dd960c4ff5769
893f963967e44fcaa46eb2098df742770b48f759
52075 F20101122_AABAHK supan_k_Page_134.jpg
8da33b92d94a5a554979180582dbeedb
d689dd42293ba45085dab61275509deef4862bf1
34472 F20101122_AABAGV supan_k_Page_119.jpg
7ef05a773f9b49c5cd0ccd9fe577cf84
43d061c4d7c24665a5fc4157bf999f8340f88fbb
26744 F20101122_AABBJY supan_k_Page_142.QC.jpg
bc7a36bbb7586422cacfc2f51c1f3406
cf5efcd126fc013c0fff0a9cdf4df5736ea93f24
23230 F20101122_AABAHL supan_k_Page_135.jpg
7c0954043b8df3c275184172f444a790
c432e060b832a8acc37da29e2694e63c09e14924
60379 F20101122_AABAGW supan_k_Page_120.jpg
ba506eca0bdf8b67138d2cf6ed3555ac
f268143221407b8fccf355994f73a542b5c0e0ce
6991 F20101122_AABBJZ supan_k_Page_142thm.jpg
ec69ee025cf0b31fd5c5daee207df9e6
7f8db6a78b654f22d28b04fbf46f6928c63c47cb
65511 F20101122_AABAGX supan_k_Page_121.jpg
94042a4f7bd7fc27b6e6816accbaddff
bc00a33f77898f48fe566bd7e25017d1b8d0383b
154884 F20101122_AABAIA supan_k_Page_008.jp2
df7c2235d92d2faa6abd40625c23b95f
71353172d83ed281e7e4d0850871d85c463a5d43
51001 F20101122_AABAHM supan_k_Page_136.jpg
e26a777b9c66fd7bd7b45493f8731ad0
36205480313372e861ea0961e3d4646aad381ca2
68841 F20101122_AABAGY supan_k_Page_122.jpg
e73b14313bd7047c36f3d848aa0298df
0baee60f3c91e6b48f20153ccfda776de0b518fc
1051982 F20101122_AABAIB supan_k_Page_009.jp2
56d82fd0541097ce5482aaa14c24439e
ed9551bf3541d096047ffa2da63795a063f14b52
46136 F20101122_AABAHN supan_k_Page_137.jpg
d24f3806890abf9e115d55c318b410ba
52958c90e01eaa4d2f98f2d04d0653db36977d9b
73128 F20101122_AABAGZ supan_k_Page_123.jpg
14a739b23721a74b9da25ec0c032308b
e4fe89b1cd80fd99cc3229180b6bf6c40327a388
1051974 F20101122_AABAIC supan_k_Page_011.jp2
5154ce86bf47c2d7f80bb5450fd54594
90ee1932bd2e9458909fd8d886b59989022c00a5
1051983 F20101122_AABAID supan_k_Page_012.jp2
a0cb0672bca3c0e0a43803089324f24e
64d24bb57f94d2cd81de783c9557349fc5ba4ea2
58592 F20101122_AABAHO supan_k_Page_139.jpg
1960fabb9a5a2ad9070514fb1a2f5adb
c064f01d3b1bef24b41b706caccc6e277d4a9c65
868635 F20101122_AABAIE supan_k_Page_013.jp2
ea81aa39f8a0b34d57629ff33b0e0bca
4180349a5f7f599a8baa210c9e6e45467a867970
83195 F20101122_AABAHP supan_k_Page_140.jpg
34e0564a0b0e09df209079ac39db8727
c867f6183182c1c0fffd45d7616981f90be765c5
88473 F20101122_AABAIF supan_k_Page_014.jp2
e7faf875a3f5e399810de9d45ce0bf22
cab6869b3f8548ff19675144573fc02696a6855a
85580 F20101122_AABAHQ supan_k_Page_141.jpg
06f4098998a8f506b9950bb1dcdfc0f9
f7a070a8d33ce59dae4429f803e716fb0bba9e1c
93614 F20101122_AABAIG supan_k_Page_015.jp2
646426f7870db59b71fd2e44fac42539
afa92240bcad1a0b7a6fec6e63ac6c07af33daed
90531 F20101122_AABAHR supan_k_Page_142.jpg
8e6fe9bc7de56003cb736d37cb468eb0
19dcff9ff99b7604a15e3dd7103fe040328554a6
107979 F20101122_AABAIH supan_k_Page_016.jp2
8285290ca46856604c7322e6d65a0b47
0cf5b916503284fdcd300ee4c5ec7cd7ba0ef92c
86059 F20101122_AABAHS supan_k_Page_143.jpg
4723119426aa7c3bb6088672b2fec9f2
587971a89eacaa9ca17d9bedce51811ac9c70cde
1051931 F20101122_AABAII supan_k_Page_018.jp2
d3fe3bb8a4ec415ea3206fe7a88368a0
389581191702db055e8ae94638f0666647748021
30162 F20101122_AABAHT supan_k_Page_145.jpg
d940387714829ae80fa7a64359e6fc7c
50a55e7e88f8b1ff6f7015d3df1dfc46ccc2c471
104661 F20101122_AABAIJ supan_k_Page_019.jp2
e7f6da82f5d01c1da0e04fa934a23aff
c6e63cdf8b0ac1205e7d48cffdfc7cff97aa1960
24637 F20101122_AABAHU supan_k_Page_001.jp2
8747af65e3ce0c125bec0ffde9527353
7c4015b8b29b0e00a8749fd1862b99eaa58af365
1051925 F20101122_AABAIK supan_k_Page_020.jp2
c5fe1a1131e7d7aaf71d2bb6863218e8
8c2b0eab7bdd453f25baf4125be63d227c03f6f5
5528 F20101122_AABAHV supan_k_Page_002.jp2
07c28793901145cc978013138ed479a9
d31b79d4f5bdefd425a751541699d997b651eda0
108438 F20101122_AABAIL supan_k_Page_021.jp2
c2911f06a1116aa6525f014bec9eb8ba
1bb97ed5908fed22ab1a27df4f05904dffadf47f
91913 F20101122_AABAHW supan_k_Page_004.jp2
c2c87bf91a1e4e2bc1d42cf32fd8a810
4a546c5baec312387204e4899a78a5c7649ea4f3
101197 F20101122_AABAJA supan_k_Page_040.jp2
8282ae50e0fc115ae5cbad2d7c845882
5f0a22979f74dcf6b761c3054b06eacc9c1e4ede
109990 F20101122_AABAIM supan_k_Page_022.jp2
8a48aef600bf77f39b286103bceb3ad1
9fedbd3c5ad92e70e5feb049541b695f0470004d
64103 F20101122_AAAZXA supan_k_Page_089.jp2
876a7414e22253cb6692b5bd73ff79b3
44f955a6b94d1719f9dff36b9b6676e66eea088f
10167 F20101122_AABAHX supan_k_Page_005.jp2
dd600e03e4dcf49b94f997d0bf88336f
4bb0bdd4d5b42a3dccee15958130c3600fea6dae
109501 F20101122_AABAJB supan_k_Page_041.jp2
de4cffe194f7859dd048bb2f8683e88e
4a987d4eb215f00caf8466293bafac0205078861
22323 F20101122_AAAZXB supan_k_Page_042.QC.jpg
27221c3df4ef4f6be1667686d4684a28
99df21763792f76aacf50dc2a8839b64cc7e3809
1051977 F20101122_AABAHY supan_k_Page_006.jp2
1c0d0583ea13140d9e73c61dd0699cfd
abf299aeb2ed8841e73d325fe939d09e5df3c1bc
111973 F20101122_AABAJC supan_k_Page_043.jp2
aad223c693441413a42817c2af656154
e6ce12f79f1ba53e8b684f8a9e4acb9f8f7ecdf0
101390 F20101122_AABAIN supan_k_Page_023.jp2
22de1d9e32dd4565279ed4e152e2fa88
f3088b510098cb671af345894a89e341612f730d
1051976 F20101122_AABAHZ supan_k_Page_007.jp2
d56bc24470795464e7261bae11e2090d
81d3d0a125d6c50c418925ef261bcdceeb181f8c
96751 F20101122_AABAJD supan_k_Page_044.jp2
5e57fe644d04eee69fa11b512270a27d
368e4ddfb36f836ba9fd4b0eda1542c13abc64e3
101070 F20101122_AABAIO supan_k_Page_024.jp2
10a4b310f8725d38d7f515bee4c25b1b
05533899cc79fc6ec773ef6e8d510064a89ddfc7
F20101122_AAAZXC supan_k_Page_083.tif
042d06d8f53c36db3f32fdc2c6876936
6383054ad58da95ff8506a20c60fcd9c3e6f0812
99057 F20101122_AABAJE supan_k_Page_045.jp2
21dd08304961eed2e21f6f35f1862873
1db5aa13d3e58be859b5e73f3d37b75518dd28bd
94334 F20101122_AABAIP supan_k_Page_025.jp2
352a75401c923775df74e04cfeaa8bc5
1e4074a3ee172af4ffba72fdb5d1c054cf5010a0
18188 F20101122_AAAZXD supan_k_Page_090.QC.jpg
e398c122e66597a2c56bf955cfaeb65e
846c9b00e271c0e1c07bb42c0ba960e6b9c353e4
813987 F20101122_AABAJF supan_k_Page_046.jp2
07fd48de141ceacdd7a7a88d1cc7f7e7
afc734e02b3732e648aaa4320658e7aae2be3fff
91428 F20101122_AABAIQ supan_k_Page_027.jp2
9553511c03fb0167d2860d0ed6d89e7b
c7484ff05cc99f6df4942bfc72e5afab01337165
19197 F20101122_AAAZXE supan_k_Page_014.QC.jpg
066669da085ae4f7608ec2e41fc60e4e
79a0c20cd4dd8d73439efe941a0de68d7f1d1861
90247 F20101122_AABAJG supan_k_Page_047.jp2
763bbf0ff88ea7eb87c366724cdc0ff5
8cdcef64f5717516649162ec707b8db1b2e13457
106878 F20101122_AABAIR supan_k_Page_028.jp2
d8ebf54bfccec25390bffc41b2f92123
ec911f0202d53741a2c9309dfd4bf428135df53b
58778 F20101122_AAAZXF supan_k_Page_014.jpg
cb7dc76cd846a55579317724658beb1e
0dc8fa1b6fb48ed9939c6c77da7dc5962152a860
873683 F20101122_AABAJH supan_k_Page_048.jp2
147844175ce90fcd1f162f3e4ba4b228
1b3a72ee216c114446466dc5020a571022e02989
102208 F20101122_AABAIS supan_k_Page_029.jp2
20e5b851f85f058c67a3a6fb6a2f1277
7ab06608739348d27e2604641dc48d26aae6f89e
6262 F20101122_AAAZXG supan_k_Page_086thm.jpg
a164009e42c89784e3608bf37191ce58
9ada115e9b68260553e978a90e1c6211cc4e0ae4
858259 F20101122_AABAJI supan_k_Page_049.jp2
148f12179ac119a0b0fc7e44af2f5dcb
b7af14ccc55c9a531d66cc4a50e2e3d0eacfd5f7
105874 F20101122_AABAIT supan_k_Page_030.jp2
9275b6019f422e17b6190d2cfe5d9b18
fa4dcb83bba298a0c5de9d14367b12a7d245b53c
F20101122_AAAZXH supan_k_Page_116.tif
48f338613ea121d830f61c8d3712f39f
2548c41ced7274874f4f3eba47d7d8ee101f9609
98499 F20101122_AABAJJ supan_k_Page_051.jp2
fb4b77867b9693f2b7773b2f7268c45b
9040692434c0f5c3fc3c1d658bf7e6dc95e35164
99616 F20101122_AABAIU supan_k_Page_032.jp2
e290e8f1f081068d248b7c7b7e31e1f6
8f1aeb03d63532e3d567a1bf52bab7bacb65d448
45622 F20101122_AAAZXI supan_k_Page_069.jpg
cf3197b62be88fab4e259670a5739dec
b195e90311a7bf1c0f749b84041649b90904cd7c
98083 F20101122_AABAJK supan_k_Page_052.jp2
8dd86fdfc53a18af03f2aad9cb509af8
935d37b6cbc1a674ec86d25d55790e5804741a8e
84941 F20101122_AABAIV supan_k_Page_033.jp2
c9744aaeb7e3e0dddf83f57b61b85490
59732c0fb91e99ef1b4800bfad8c5cb87f0b171a
104636 F20101122_AAAZXJ supan_k_Page_042.jp2
c091bf5564b22c1076548bca13ab6309
be7e8c7a43a1e6d018c6809e8fd9a9a92e467f8a
95988 F20101122_AABAJL supan_k_Page_053.jp2
c6017558149852a7f6a688909e6ce79f
83043855c060abd2ff7f90be5b5260e07baf6a29
90525 F20101122_AABAIW supan_k_Page_034.jp2
213dd50a8452f9171456c6018b1f6875
5d256a2a2bb3fc6e2c59f9bfe5ba50b28af40c6a
1786 F20101122_AAAZXK supan_k_Page_015.txt
d8d143f7d36463619796cdf43142f483
22efed2f68810a51bf451f3a7fa217474024f5ab
7182 F20101122_AABAJM supan_k_Page_055.jp2
7f8e2c02df96a107b74c1ccfe1f9470b
5f51ca7a5fb7c3b5c4e424360a62050bb252a5a5
34053 F20101122_AAAZYA supan_k_Page_079.pro
a7735db5a29a619a6fd31dad8ff0d36c
99989e88ba73941af909af2cbf07c7b7324706be
101208 F20101122_AABAIX supan_k_Page_035.jp2
1ad7170e61feecd5b27927a02ed1b140
60d52eaeca5f8a0b65d0c82e59504991f3a2ca81
1768 F20101122_AAAZXL supan_k_Page_095.txt
edd553279cb633fd6d83d69d74aeb1db
430a65c1637f0481f80d6dcf6cdd93343b5427c9
52551 F20101122_AABAKA supan_k_Page_070.jp2
d05f98f2dc653b596de16e9400d08c30
2655d30fe37ae79f8b853044d253827c6fcacc18
102422 F20101122_AABAJN supan_k_Page_057.jp2
dc3620bdc65eb0be06c55791c71b65a8
b250da5a10879737ca012a176f599e4f72d78f08
30243 F20101122_AAAZYB supan_k_Page_059.pro
b37c23309a1e5463a1c828efc85286cf
f7d1e598fe489a81cc696494f013b7fcf9b2e391
99973 F20101122_AABAIY supan_k_Page_038.jp2
e318123545a68159f25da0ad27cd5f19
d40420aeebe4356815b89c11655c7fa70fde3189
1776 F20101122_AAAZXM supan_k_Page_071.txt
65e542ce3e551aae614e61019c4a1efc
d3b729d40df0f75f54ec71fc33b057b9abfd9250
69458 F20101122_AABAKB supan_k_Page_071.jp2
3b6e57716dfb9306047f8724938eb0a4
754642c28bcccf3efffaf2fab17771de2bc0edaa
6269 F20101122_AAAZYC supan_k_Page_016thm.jpg
b798047fb7d00d4932eaf56ec54cbf6d
cc49a4ad9a3e101327154709086737b71245d0a4
101816 F20101122_AABAIZ supan_k_Page_039.jp2
eef2fcc24bb3e1a9aef88b3f2cc8ec40
b42cb9aff1dd4b878410a0a64daa6e5d37d026f7
F20101122_AAAZXN supan_k_Page_120.tif
53b1b9473e233bc3dab6f6b55d658167
e6a6b04e40b1cdebc33693a5cb3b21c668650ff5
75738 F20101122_AABAKC supan_k_Page_072.jp2
a5995d9a840dcc6a801f4b9ad03cebe2
8fac974309aec697a5d223c8a247a9d1911bc722
35656 F20101122_AAAZWZ supan_k_Page_084.pro
9dd83761a57900571508ae291052d179
3465c7a2865e5a4629ad45d23c465593d300f9eb
1051955 F20101122_AABAJO supan_k_Page_058.jp2
8cc79f70c091b73543c5240a14fd6265
32bf6fcb1c6288effc6e7ef95a665efa2a3f142f
20921 F20101122_AAAZXO supan_k_Page_059.QC.jpg
25c5aeb91cd1b6f50ee77189a0e9b37e
ed7c3277c1c6ebeb8d8601c686a252d8e8dfd55e
F20101122_AABAKD supan_k_Page_073.jp2
1a1d373745660171cb821c8c068a7a3e
5c7124a08f44a5a147283f65831e3f702997f073
1051920 F20101122_AABAJP supan_k_Page_059.jp2
665abac14521e209d7c2b95660e8244a
130d5ff2295deef3a3324721474677532595e9eb
73519 F20101122_AAAZYD supan_k_Page_108.jpg
02232d0f12f6403aa4d2024cc282e996
0a6a6ee5a1c0f96c2ec35318a739d0c97a98f96b
2013 F20101122_AAAZXP supan_k_Page_021.txt
92015979eb64e474ad8c4d0dffe2f9c4
9f3795d2729423c776f0fcbc2da34458457bf7be
79400 F20101122_AABAKE supan_k_Page_074.jp2
2e0d616f8cef0cc7ad86adc7ec2b30f7
1ef16e5e5c5ee5d0a2adaaa0cada5da6bf97ccca
1041255 F20101122_AABAJQ supan_k_Page_060.jp2
4da09112719ea8a0b14cfacccabd03da
f3e7c729e60aba4eb2e80caaa1ca7309b343a7c8
74316 F20101122_AAAZYE supan_k_Page_056.jp2
5a153ae544683c7817607e04dae5df1b
df6be185d90d11bf8478aa7556347352263b0c8b
F20101122_AAAZXQ supan_k_Page_106thm.jpg
d03b6773318e4e1300934030270a6b28
2895debdf62a7c5114347b10338894af66fb5e48
81577 F20101122_AABAKF supan_k_Page_075.jp2
c05c5093579634d307033c5847e45eb2
5b03f63b29891364ca26766e27eae1df26b73538
59536 F20101122_AABAJR supan_k_Page_061.jp2
8e08974657d76c26e78f228b0c27876a
f498603e6ebf99f39f4596ae2db5a2e00131083a
19804 F20101122_AAAZYF supan_k_Page_048.QC.jpg
c1a4b2e8d3f0755b7dd44ff824ca4499
e40485cd476a3d62242fba81d8252d447194b570
1803 F20101122_AAAZXR supan_k_Page_065.txt
5ba0b59718c02e3decee0d3cd383ae54
f58c99bec67bfc22ba6f56e5438a561a8f91ddaf
63631 F20101122_AABAKG supan_k_Page_076.jp2
d7afa0d3736c99a27b21be4e60aea02a
fa184e4b5646f4516519aec03d42207a80355857
823119 F20101122_AABAJS supan_k_Page_062.jp2
d070fdf5a77b5f9991d27869630715ae
e634a56a2efe1fe4fd231b483a0c0937e5513360
49435 F20101122_AAAZYG supan_k_Page_016.pro
b179638c7869030bd78cd9319547ea2d
753b1cc072754a4fbcef041dc9538bc4020962de
F20101122_AAAZXS supan_k_Page_089.tif
96bde0eddadb795b39e4be91f32daae6
e34fa577b0a2e1d7b3e30d0fa34ac10f60898b3b
63728 F20101122_AABAKH supan_k_Page_077.jp2
54c466a1184b840c7a2fb4d14f707a2e
3d41d48a12a596573c1f0aba62532f0e3db75856
883360 F20101122_AABAJT supan_k_Page_063.jp2
9979650d597bbeaa3bc9e2e5bc768024
e71bd4c72520d7b93a100ff0b18f7e7b18ae8195
104314 F20101122_AAAZYH supan_k_Page_017.jp2
6493c9460657609eb6c7852b2248cffa
ed8c7aab5691c2172abdeec76d9e1c31c3925ffa
43772 F20101122_AAAZXT supan_k_Page_061.jpg
9cb36dd8eb3977c3c4693f1dc1d6416e
cd6cd4e8443a48691dc0d6542466e10934fea182
1044885 F20101122_AABAKI supan_k_Page_079.jp2
45bcf98ed546ddc9ea3722f8555f9e31
6daa6107fa67dc9150d6ed472ab15b0b850c3ce2
729353 F20101122_AABAJU supan_k_Page_064.jp2
b6fb608b6ef97d01f7e3bdfd78db0c29
80da1717cce7057c2e76607f32ba6f3e9d198bea
F20101122_AAAZYI supan_k_Page_070.tif
3b674dddcb830de03302d0b05b0e8643
3f10f4385dbc1c84079e45b4ed842339347dffdd
F20101122_AAAZXU supan_k_Page_053.txt
f6ddee9ab92df133089163828dc72864
95e7958d17b96e819731e3d5ae2f96381895cd5f
102407 F20101122_AABAKJ supan_k_Page_080.jp2
1ab1e30c8c1cf2a55b98b853c6afc988
86c1aae15f62682823505cec771ba4dcf3fd589c
914830 F20101122_AABAJV supan_k_Page_065.jp2
10c69b190b9a17b557fd7196730cfb98
4b983da0e9f3e34ffa0b1c7581a95a9afb323612
4963 F20101122_AAAZYJ supan_k_Page_136thm.jpg
7c63a1dd91d4589ed232c799c04030a3
a44964f8bf4d9c2b2eaa353f66f853ffe04798f7
45599 F20101122_AAAZXV supan_k_Page_077.jpg
04539e4d7146aa70052a27e000cb2048
72d3213855bb84852a56ae7689686e5b4c113957
1051969 F20101122_AABAKK supan_k_Page_081.jp2
5c47f28bb479ba9853c01c5d7daf4a49
6dd829a003e3b3cabd519ee16b2366f381606347
800515 F20101122_AABAJW supan_k_Page_066.jp2
f9328dbd26dd5a0455bb53abdaf65c1c
a2f49d32a73e1edf80392d572f8ec13702226ca5
97108 F20101122_AAAZYK supan_k_Page_036.jp2
bcbb598a2079d46510aa3efc7ecec425
a16c5da68bc48c8744456061989ceeff0cd778a7
20437 F20101122_AAAZXW supan_k_Page_112.QC.jpg
19f610aba9d38a03a5c679c2af7807e3
15994824d76ee20d005b65c957111de92c7c7d17
F20101122_AABAKL supan_k_Page_082.jp2
f818263416957227f1134eaaed760e80
6847e9c3b77f90802c97306a36d02fc9dc6674ec
9532 F20101122_AABAJX supan_k_Page_067.jp2
6c238065bf57e89207f6c4622f0e6abe
1b4b5c6989a6dc98c2994ed95629dde60b729f3d
1544 F20101122_AAAZYL supan_k_Page_018.txt
c37fc5d20e965553243686ccddeea2bc
37a2ba2a07cefe5c2c19e32a3b12c10dcd77a112
522211 F20101122_AABALA supan_k_Page_099.jp2
4651e14117ab7f1d05fe18cf616f0b26
1e2ce6afab7589aa2f0758d7585346af36a9e8b4
22537 F20101122_AAAZXX supan_k_Page_018.QC.jpg
3bbe9c7155be4de031932a17632fb2c8
648cd99c46468273d06e3c3a49b38f427dd34899
812001 F20101122_AABAKM supan_k_Page_083.jp2
46cadee8862b1362a0fb61cef5b489fe
d524d5f62dc95ce63cab9f28f38c1e42c63fa3b0
103773 F20101122_AAAZZA supan_k_Page_031.jp2
5e00b15d710ff9f1010456c83f19c65b
a648878000ceb58c7de4208f6821c0802153ca3c
90686 F20101122_AABAJY supan_k_Page_068.jp2
da0aacf3cccb299b4fb302917d9c9123
124b250eb2be9032b86e7afc99548c3b94809304
F20101122_AAAZYM supan_k_Page_104.tif
49d9ff56718afa65b2dd9fd4712966c0
1c9c43813bc01d16fa4937059dee227bfdd095b2
88954 F20101122_AABALB supan_k_Page_100.jp2
ef5692d94eb534b4246ce8a22252b094
3f6b8f1b23e606fecf1c6660fc630f99b83f3a4a
6531 F20101122_AAAZXY supan_k_Page_020thm.jpg
cef9c6e42843155d32a3d1ddb1176f10
a69721041383844ef8843d3c9925bc1b421c2a08
889285 F20101122_AABAKN supan_k_Page_084.jp2
1c106dbfab24ba1fa11ac53d62ff8517
1a4dab71937d7667447720607067a0bfc60a9b11
F20101122_AAAZZB supan_k_Page_012.tif
8f8db527b96e84c93bf1694ff4910465
50dae7ad3bdb6ff9cc2c63bbd6e856ccf99dbede
68964 F20101122_AABAJZ supan_k_Page_069.jp2
9c77d1f8f9b01c91bce32834385088bf
b967b69951094dac3b5f8f3f47dfb4a4cc42ac10
537117 F20101122_AAAZYN supan_k_Page_130.jp2
f4565c20fe83d06810b62a40eed777c0
b3fbd576a713da62123e8b8c6ca3ac3460c2a709
71752 F20101122_AABALC supan_k_Page_101.jp2
0bea138342d1796b388e464879a3acef
b9a14ad20da827ddc2e4929bd0e99715fe8e59b5
63057 F20101122_AAAZXZ supan_k_Page_084.jpg
20748f8fb356ab75ae8345519bdce9c6
31c1a6045425d85e89adbaa1d06dd946cf1757c3
939954 F20101122_AABAKO supan_k_Page_085.jp2
e0b15e7ef076d8e0149da61f4c389b8d
e199c2b79d31f2cca87475d67afb6cd4beb5fdff
129 F20101122_AAAZZC supan_k_Page_003.txt
a209bb49e53fa62f245eba2270f61a6a
40e18ab4767b32e42ef92bb5ff15f3b9dcd916ef
F20101122_AAAZYO supan_k_Page_022.tif
026f1e779b1cdaac7d1d6e8d94320bb0
c72b28120de1ae222f0cd5ff434e43c96ecf329a
674864 F20101122_AABALD supan_k_Page_103.jp2
cc34e669b4e45b1acd2f5e06121c93a5
2cff24fe771191ea49219a6fffe7cea5b6742e76
5368 F20101122_AAAZZD supan_k_Page_009thm.jpg
789f01eb67cade8f1c119b862bac7222
39d29954ffa98312d0974d3c4723f294820c8d0b
4504 F20101122_AAAZYP supan_k_Page_089thm.jpg
5717bf348842bc6b24212631efe24d3b
0c5ec6b7276bbd7a4260e8e8db70a6928d890770
101782 F20101122_AABALE supan_k_Page_104.jp2
0b674cf66a80a9bb8c684a08954e1469
5c13d49a24fc6f42f68fe7ca552f97841a1b1f1e
97943 F20101122_AABAKP supan_k_Page_086.jp2
b2e95d058634314d056cc0532e15e9f0
008d6be08b107260ee3ed67926244717981d78e1
45704 F20101122_AAAZYQ supan_k_Page_036.pro
84fdaee4fadc75cef3faad3dc91b6974
d2c88cc2bea99b998205b999351c130432775929
908790 F20101122_AABALF supan_k_Page_105.jp2
d5e1554e1a7396efc3154c8f5819eb89
1e6343c057312bdaf96382977e555fdaef76f600
109323 F20101122_AABAKQ supan_k_Page_087.jp2
29d9670f116159abfae5fe64182f6c94
fb48e1c1202dc570f8d0ed8f00644798424d85ae
94556 F20101122_AAAZZE supan_k_Page_078.jp2
e577a6712cc911dee448aecfac223951
da47519b915d8cbf6a6475979d76bd23f57ecd66
852 F20101122_AAAZYR supan_k_Page_133.txt
4c1011436e60c248913b86c22b854d15
fe49374198d0e7a8ea10195a010f98698f55763b
106570 F20101122_AABALG supan_k_Page_106.jp2
8c962ce26d8a3b94850a41c9ee85eb45
939883fe820661664268648ad8f1cf489f9eb2e0
84503 F20101122_AABAKR supan_k_Page_088.jp2
af9700004539bf127293904404f4aa14
2a3cf98fd2898a23797929ce501fafeed557a958
F20101122_AAAZZF supan_k_Page_113.tif
9644237ca06acc5b1a04552eac8e0a5a
0040202ac373e97b86aaf7720c907fbe2ec34b1c
F20101122_AAAZYS supan_k_Page_030.tif
3b262f91444d26b81f59c72ee568c7ff
a32af592fb56a2c1d79b67eacade9eeaf3b34eec
657758 F20101122_AABALH supan_k_Page_107.jp2
e27b77900bf06dc710d29c8a57464ebd
1872874c2e9c326080c5f853a27eff1df3febf87
76637 F20101122_AABAKS supan_k_Page_090.jp2
38b788f80197d18f00bdca84322b746d
1086708d3f488fb8f62c987319854a8c7aa86707
F20101122_AAAZZG supan_k_Page_074.txt
42c84b94854471c6067aa2d5bec859e8
578cab9320e63bf70df9b8f5eef7e4ef362c44b5
6100 F20101122_AAAZYT supan_k_Page_057thm.jpg
7dcf9f75ced8b6c53b52a3c57aa422bb
bc667ddea9653b2eac042a83cf8ce084bf73556f
110561 F20101122_AABALI supan_k_Page_108.jp2
0a729db752cc250af17309fe29779412
8695e051824d30b2d11d50ccc34090758b32e929
89750 F20101122_AABAKT supan_k_Page_092.jp2
9d05a758f0a04384876fefd2d1fe5e25
bd7ea9f08cd2bfe4d011bc17dde90665bb6f0568
21695 F20101122_AAAZZH supan_k_Page_081.QC.jpg
874f6430509443a39146cd4cb9637705
9acc0497f05531b8b138e9feee7c9ee2c920c851
63880 F20101122_AAAZYU supan_k_Page_086.jpg
c265e220e6561818d734999a5a41ea59
e246b145e585728802716c76725d75bd1b8e6595
971462 F20101122_AABALJ supan_k_Page_109.jp2
f3eb46e280f96e5cadceb9f1b4174965
2c614687c64a423441475848481875061133c461
880902 F20101122_AABAKU supan_k_Page_093.jp2
66515ecb8299a73079152fd4a32a4346
df1e2019a9d145e4d5e8f783873dbea68f70e3ed
1596 F20101122_AAAZZI supan_k_Page_084.txt
aaa6fede9ace3c3b930fd685fd6273f2
5d73b991537129454921bddbc07a3d5856b24ef9
14028 F20101122_AAAZYV supan_k_Page_131.pro
8d0b5401d5dc6c0eab622810bbd9fdc5
a254d00e8d9e3de618b832938b903f4c1b3323eb
997975 F20101122_AABALK supan_k_Page_110.jp2
6d3f39a9fe692359d5567e7476cb6db4
dade9024dceff8701458a02965dcf81016325677
87401 F20101122_AABAKV supan_k_Page_094.jp2
7943d825e94cae8384dbae5c3f91f068
2d220dd336b41eedc88a74b51fe47154e0c55694
48623 F20101122_AAAZZJ supan_k_Page_138.jpg
415eddbca3f9aeb3702130c809234412
b362f0f36a5770f654f59652b8dce728b5fe08db
23242 F20101122_AAAZYW supan_k_Page_123.QC.jpg
33e1d86a8d35df62abde3fb7114b1c27
13c38d999ea14682edcd274a09e8195c99f1e8f2
1051971 F20101122_AABALL supan_k_Page_111.jp2
32cbf4d4d55bdb7266b11667e990bfa0
fbdf622e8e06703811b05e23d833f8e18cde5253
80860 F20101122_AABAKW supan_k_Page_095.jp2
03079612adfd71c8f348f608c0479082
6eeb0baced3bdc4b22a8280dd7f1618fc251721f
1724 F20101122_AAAZZK supan_k_Page_097.txt
a7a18686e21ff61f2957e702b32ec600
d4df84d196ec1a5f61cf1c6cc567088b67e13cf3
54903 F20101122_AABAMA supan_k_Page_129.jp2
e63e717d3fe5af91b5a418395c0f59e1
cdfe95abc5fc72ebde8bcacd7d660cd690dbe609
5372 F20101122_AAAZYX supan_k_Page_074thm.jpg
b9c8ebe074da1c767a7e3eb3a2a694fa
03e4f4584c7ccf14edfe9bb12cbfd249412b7ace
98050 F20101122_AABALM supan_k_Page_112.jp2
7d766471f5f2fe2682c9d2e6a1cf2e2b
97f18569f632c462fbd03f868752e7fea3a33a5e
771949 F20101122_AABAKX supan_k_Page_096.jp2
788bbe09a84f59e162902ba436d98868
eb61bf23c29244b65f175c04e10491620430c3da
F20101122_AAAZZL supan_k_Page_029.tif
6a1d85adc03c79e31d93458c5a828dfc
12e421f65bb3ffea5bdd393624456b1f49f39cce
636067 F20101122_AABAMB supan_k_Page_131.jp2
69b18d92172d9c1f2d7142910cd54773
e5ef53634befd1c151a1bdaa9ef00ab665ea797c
21471 F20101122_AAAZYY supan_k_Page_037.QC.jpg
45b116490b01ce0fdeb47353e82c73b3
f4001b173b59d93e5b4254b58a1dfd304625056a
103843 F20101122_AABALN supan_k_Page_113.jp2
6d00dc5fc666d6b119881a43682ac007
61c1774dc9a1f45bfe671e29baf13fdb06f825c9
710116 F20101122_AABAKY supan_k_Page_097.jp2
4fcd797941455cb5b30a4297f42450d7
cde79d4c62290f9d384578e4e717376eb1bcb157
29797 F20101122_AAAZZM supan_k_Page_101.pro
f139b3e1c77a8bb61b6d24e48403ea8d
f1a3f33a5c0ca25448ceb686d0c1f0b5b0e3ba1f
1051835 F20101122_AABAMC supan_k_Page_132.jp2
dac910248255bb1873b391ed91a7cb33
466e369bfc4d8ba562659eeb52bb80b85dcbebd0
5720 F20101122_AAAZYZ supan_k_Page_102thm.jpg
f6f4939ffe9c6bf06ac7b90318d351f8
c19890d569f507974d6c182f11aa5a8aa2667cb1
578587 F20101122_AABALO supan_k_Page_114.jp2
8e1159ff0333eb749a0718620f60a95d
08b52f7d4a83895f610ce4054e7e9b8f78af9187
865031 F20101122_AABAKZ supan_k_Page_098.jp2
022cb01c718344968cfe943b66d59fcc
c4ee31794d842abe13fc01afc00b3d20bcccf688
2337 F20101122_AAAZZN supan_k_Page_118.txt
3532a61538ab963b53b530d21c2cbd8b
867a5f0ce833ffceaf32093092ede5178f70b253
1051923 F20101122_AABAMD supan_k_Page_133.jp2
763ea1f8525fcb060398d48e4b92b165
5dddbbe3b09f016c951cb11e1f98c4d4c1f98283
668846 F20101122_AABALP supan_k_Page_115.jp2
b1551da59c647d4b0866671f89de4017
4b1166aa2b6eff3572aea81943cde567291e3964
121545 F20101122_AAAZZO supan_k_Page_141.jp2
793eb7383aa94c9dc893e7f976ebf7ff
e38834ed2782e0f6976b8d67961576c7baaece04
1051984 F20101122_AABAME supan_k_Page_134.jp2
e6474bc35e29611bfc7234aaa8a65f05
a99456deeca415ce2b3a033c9f5554f0c3f64b09
6372 F20101122_AAAZZP supan_k_Page_051thm.jpg
158e7edd31b643b451291c3b05de4e9f
212f1bdbe148302e05455149254ce1bddf680f5b
659694 F20101122_AABAMF supan_k_Page_135.jp2
2a292aac0b84a19b1de9619a36f83b13
dd80310b8ad63ade209195972156930ee343197f
103208 F20101122_AABALQ supan_k_Page_116.jp2
f9a911506dbdc6498c85ed3e1558e440
7d33fb888ce60ae4c752004c63eb811c27f05b24
22043 F20101122_AAAZZQ supan_k_Page_029.QC.jpg
3ff3c84f6f4e7737459c50f46470b535
2169181025d4f4e19585bd4d8f54313201ae695c
73387 F20101122_AABAMG supan_k_Page_136.jp2
8b29f4688c5f29785334db5729334a85
36265673858278c97f942b182023511e76b38805
889374 F20101122_AABALR supan_k_Page_117.jp2
52d11daa35d9729c6db65d0f84e475c5
880932ae3fbbc07bbe7b2833232a89562739ebfc
156 F20101122_AAAZZR supan_k_Page_067.txt
23e8e0e8060bfa8fca1b6f3df2ecd64b
f87c9a117e18ec43aff5d1cca5a8e29dcc0466e9
64233 F20101122_AABAMH supan_k_Page_137.jp2
65025363a836b2bfab77f6b857107f2c
e78dad45329529b08c5f2aadbb4b4b753c3c36df
442129 F20101122_AABALS supan_k_Page_119.jp2
7bd81a5563034d33664f562e0634ddf5
918d95ea490f5a72ef1d9a06cb5f3d027b00896f
81558 F20101122_AAAZZS supan_k_Page_007.jpg
e5b2ce14ba3d650ae192338bba9e0671
1624225042db46f73e1eeade537d881081a7feb9
69510 F20101122_AABAMI supan_k_Page_138.jp2
ccd085554042c5b5d555d891f870d303
12ce8412d09b887305b59cdc6f64ef4cea72b68d
100666 F20101122_AABALT supan_k_Page_121.jp2
ea2f318cd1c0100040793d463f442fd1
f71117b253836dcd8c8735cfe8088a7b3e56f8ff
15564 F20101122_AAAZZT supan_k_Page_133.pro
b7ed09f9e16fd21a4ae52d6d6aa245c5
c16852d40a9b51915feba6b1416f8e4e82baf93c
95076 F20101122_AABAMJ supan_k_Page_139.jp2
6cf442abb4fcb0a8c68ec9e209062368
1eac3360efb03621b330010ce9bdf22969e8e3fa
104812 F20101122_AABALU supan_k_Page_122.jp2
45fce9d038a1f7871b15e8c569519dfc
f7ea7fb4b4b4d8ea61c4b1edbafb62a062237759
72205 F20101122_AAAZZU supan_k_Page_022.jpg
e3640781977ea57138ca3a604bfd485c
6705de68d38c0dc69f944caa533f3133c8e411fc
125391 F20101122_AABAMK supan_k_Page_140.jp2
4076c8aa45a875670f6fc7586fbff087
d5ba49036dad2956cf618e8aeb94f6dbd4cc6c4f
112189 F20101122_AABALV supan_k_Page_123.jp2
66b5a0345a4b30fa8c395456f996879c
eea56c4af6d942d4f5699c670882e63075e4dff1
F20101122_AAAZZV supan_k_Page_109thm.jpg
02ba623a9cee872df2eaa95346c156f7
4042f4195f24dde211b936db4ac834ce5edc4dd6
131451 F20101122_AABAML supan_k_Page_142.jp2
b481c62deaae320a6eee3d6cb9edaf21
e372701ddd8d9ec966a7cacc0553a690bb7ded59
139155 F20101122_AABALW supan_k_Page_124.jp2
237bd8b2cd014c5e9b69fd1aee0745db
0eb675b5f65d2474566fdd3c68ac60452cfb1ac6
47649 F20101122_AAAZZW supan_k_Page_104.pro
fa38995c0d5ce14906a90e606a15251e
fd39480ec569af6cff2c8304370e468a58448ccb
5872 F20101122_AABALX supan_k_Page_126.jp2
77d4012b56d2959a3c39786e17b7b865
d70012d03b29e90d4c95842d271dd0b502ea46ae
F20101122_AABANA supan_k_Page_013.tif
b8c86338aff399ad10f92d0b97bc45df
ea3ea1ccdbe1c0cc9f65c3da25145752646ed62b
96019 F20101122_AAAZZX supan_k_Page_118.jp2
a79ce924255b055dd58f4b32335c417f
fcb275f176666cfd1ee375730c2355c91cfb2b24
124132 F20101122_AABAMM supan_k_Page_143.jp2
256aee21abd2762401fd143cc71afda9
dc9f658d59e1116b30b23bdf9cd3350e09e66b63
58301 F20101122_AABALY supan_k_Page_127.jp2
ecd65eaa5650d09fdc52e4a97b74e8d9
2981b090953d3489a319c9e659a4c0626813a1d3
F20101122_AABANB supan_k_Page_014.tif
2a4152b577d452f96d7fa9beb4827f83
30b878ca4be3395b6ce6ed8fdcfad7142a4983c1
2433 F20101122_AAAZZY supan_k_Page_143.txt
1fd952206d8c29b3ac59c48d9a307020
9761d4b6621479d981427cbea887f019a2a58741
83194 F20101122_AABAMN supan_k_Page_144.jp2
09f6d77eb249a0e436f14083c871e225
79d980b66fbd0157df2996560eac86c654032ac5
57993 F20101122_AABALZ supan_k_Page_128.jp2
e362c690a79a4a75b5b4e34c00e5a35d
64bd3ece3bce10ce9768198c1bc7dafb69b1eef1
F20101122_AABANC supan_k_Page_015.tif
e8a907017f496848e4ab49445b30dfaf
32f87d404c7774c401d7a1e19ff7ea29e3ff52b5
40620 F20101122_AAAZZZ supan_k_Page_094.pro
a77fdc4144287de01aa5373faea794b5
d48fdc92d08429aa29960f3d0a26ee3746b45bab
39160 F20101122_AABAMO supan_k_Page_145.jp2
d891b374f908cbd38f8003dd826186df
92f7c85be21035fc837f5da53bb70bf89c5b344b
F20101122_AABAND supan_k_Page_016.tif
69543faa84df6aaf71102be3a6457ea8
07ddb17045c03704da57b6a433fb0b2a84262039
F20101122_AABAMP supan_k_Page_001.tif
9b12432ba3e60adbbe8c4964668aeb83
c6c5ca8969ba5a5c88d4950c32a24cb6dc74cf84
F20101122_AABANE supan_k_Page_017.tif
76c3d19191b819ad6a8d5b5a9eba1cec
3d9a4848e49f0bc96532bed10090b190706621b1
F20101122_AABAMQ supan_k_Page_002.tif
201d335037e176284ed2f7da2744bb04
701e97d417516fd5bde3e8c1fa58dbebd7d03f4a
F20101122_AABANF supan_k_Page_018.tif
17bf0130f64e710ecc12e731c55fab9b
001a827c5b58bdad9f874116811f1c6079af41e0
F20101122_AABANG supan_k_Page_019.tif
cbaef2593ea10414f2e641cc3e3dc496
f4e34012753b95a0918f9ff1c49e345a94cd1f95
F20101122_AABAMR supan_k_Page_003.tif
d025992ee33e9e90bf5faec95c3ae5e2
a863de7f9cf33539b47152ff21f6e14c0ebaf094
F20101122_AABANH supan_k_Page_020.tif
7b306c07d95525de7918460982045235
3554447b94c66e90cffca02b595d48b36939fb8c
F20101122_AABAMS supan_k_Page_004.tif
b45b0526f81c2d995a31dff550fee092
7a5ba017d1b15192009c63272c8e0be0702ecdd7
F20101122_AABANI supan_k_Page_021.tif
a2329b441dabcb02850c1a262b81f0b4
d9f710a10e1475b42e38bbbb389c9c09a1cf8cdd
F20101122_AABAMT supan_k_Page_005.tif
d99ff955a7a5b78fa8dd25054177fae2
488da82bd44510fd8f0e2b9195c41d05437b36c1
F20101122_AABANJ supan_k_Page_023.tif
59698cce4051d174aa782467c7245455
4224d1f0cd9af66bb88ff1529035c3f465da453e
F20101122_AABAMU supan_k_Page_006.tif
90efbcca4eab643084a313eeb28d95db
86259e92af27c8a4a622c12c320114497764c491
F20101122_AABANK supan_k_Page_024.tif
e024f8e9a12dd0f2a385c9bac3fe8bf9
efeafb5c579234fd77865378cf682c332a1530c6
F20101122_AABAMV supan_k_Page_007.tif
af261da14cea938769087347688e5900
a4b1c815355f9e91ef54dd86c5d9099619edb95d
F20101122_AABANL supan_k_Page_027.tif
326d09bdc1fb02e7ac563c8f9ed9b1c0
7372542a64989a77bf0ca2f984213a43ff6a5519
F20101122_AABAMW supan_k_Page_008.tif
1266b15916d4159a2469f38cf0a88b72
4cd37b3b1bbc446a15e0d6d9fc3358c0c5bf207e
F20101122_AABAOA supan_k_Page_046.tif
597fcefb35802746d88b6bb39379cee4
f1eacd67a92d6dda3073f32189bf38fc8907a2b7
F20101122_AABANM supan_k_Page_031.tif
c47af2bc5330a25fe551e09421b2fd8c
edeb6b8a9b795a7c7a8d4dfec8019392fe71084f
F20101122_AABAMX supan_k_Page_009.tif
92b5b04d1270a43700130f4e309d2267
e8f10913922a92f15edabf3dbd92fd76e99197f6
F20101122_AABAOB supan_k_Page_047.tif
874e105ff0a1674a16fe812bdb1e16d4
7211da2c6c1a029c182e42345ceb0b84d7f44c1b
F20101122_AABANN supan_k_Page_032.tif
eb9c213203ee7d198dbe6ccd98c0589f
6002d345eb971b41efef8cf706af66e1de283555
F20101122_AABAMY supan_k_Page_010.tif
626e582ae0e01d74115a5f54cc732959
536f402cdf2a944c3e5433ef04d1f07a5071f888
F20101122_AABAOC supan_k_Page_049.tif
48f762822b674e45a361a5630773a626
d377b23cb35bf44e6fc8eaf64dcab1c907564df0
F20101122_AABANO supan_k_Page_033.tif
f95ab3b49b28f9ba2e03653424276ca9
15a06425f4c600adfcb9726125691a4d072e9b7d
F20101122_AABAMZ supan_k_Page_011.tif
0b91df7e16e8e1ef3091ea4c838420f1
7bd90cdef0564de91f843a7fe2495bdc6c776e29
F20101122_AABAOD supan_k_Page_050.tif
c943061328da2b803db78245a4e0f417
b12204556ce042cc2c05b4de511dad2d158bdb7e
F20101122_AABANP supan_k_Page_034.tif
9a06cef0c6363cceff57afb513666c24
72e2a32f40af44ef1f5b18cdd8316d1437abd0e3
F20101122_AABAOE supan_k_Page_051.tif
c9e693e401fb22aab073d0ffaf06e67e
e63b626705d65e5c705f6cff731c1a35a3b35562
F20101122_AABANQ supan_k_Page_035.tif
ef093f2ec1e54f3a110e70aa9a4cf83c
06468542b69979b57829c132fc8007f8f222f940
F20101122_AABAOF supan_k_Page_052.tif
3a6984df1b74d8d9f8cf7efd11934946
14cc20a1e0f611181541472da3e0277399517e33
F20101122_AABANR supan_k_Page_037.tif
52ab1fd5afb763d36d24e5f4bd857555
0e5738926044ff3d66c86577878e55278060c013
F20101122_AABAOG supan_k_Page_053.tif
d97da6389001816586a7889543206d8a
076f8c7a6907f2bf4e4ab493196e9d5e6ccca9b4
F20101122_AABAOH supan_k_Page_054.tif
3e02a7ad7c98b95c6972046746508d53
9daa6b0a1abaa15ceefe3ada2f6980e797c1cce1
F20101122_AABANS supan_k_Page_038.tif
d6f2e3148fd3bacb9da9c04af6d3ea4d
bec822066a4120ae1c59942dc4691ca1c74aa273
F20101122_AABAOI supan_k_Page_055.tif
a4a42ad2784c8b0972bd2780a78a8cab
b422cac508b65947f9a79e2918a3772f4dadbeaa
F20101122_AABANT supan_k_Page_039.tif
1f718f8cd9ec219e8c9edcd2c880a7c7
deac31a5ff0e5b928010e5cf06adca2c1902d2ad
F20101122_AABAOJ supan_k_Page_057.tif
743ab976564b4b0da7d05fc12d46e146
d45ba1987e2cbd1d1d71f5e4da009122051bf45d
F20101122_AABANU supan_k_Page_040.tif
a15cb61ede98e72d813b26302f4024b6
ba0c0a96b20be5402e7f14a2d0111d6aa98d01df
F20101122_AABAOK supan_k_Page_058.tif
a9bc857d16b5f1932fdc8bd683483cc6
ec32c82d670cf18fad68ce9a4edd75c3a1f4b57f
F20101122_AABANV supan_k_Page_041.tif
e881bff848f2e9ea5952c12992599037
40fb4092fa53cec6f7417859211e2878648b1648
F20101122_AABAOL supan_k_Page_059.tif
94ed4dd57b8621e6178806bece8e4d74
f3f35e2d53a5d2f800c2739a074e0eaec1e3b122
F20101122_AABANW supan_k_Page_042.tif
9d15bfa6f33071a59b4c5f89ef25e75b
99751a3205a33c7379e906d7f52524d350bfa6e3
F20101122_AABAOM supan_k_Page_060.tif
610a36d6f897bd39e5c52fac5cd2ad6f
9c53848e01e133b5d2761ef97234a43498f94ad4
F20101122_AABANX supan_k_Page_043.tif
3e29dc6f904f61148833436a872f6133
387ce0f831e09f61f247a29ba818bf4a8f45a65f
F20101122_AABAPA supan_k_Page_075.tif
de603a5f8d86a0057fb756837333386e
e829b29e27b3288ed2938d8fb8d4e0bcfcbcf70c
F20101122_AABAON supan_k_Page_061.tif
afebc877d72d3f99e1715c76eaddcf13
c9ece60a3b9a370161b0253d95ea6a000eb9bc91
F20101122_AABANY supan_k_Page_044.tif
1c16a13d9ae9ae859f5beb341d2ca59e
34e5bfe21f838c64bc0e85885b3723be0bdf0849
F20101122_AABAPB supan_k_Page_076.tif
2792b25ff70127154d436ad6c92c1900
6cd36499109d8a79b24ea93a606c7df5912de0b9
F20101122_AABAOO supan_k_Page_062.tif
5c662beaf556fe05e29aee90df236614
8b3d3c2c718bcf46c4eb5072b15c6f85bdd8579d
F20101122_AABANZ supan_k_Page_045.tif
504a889239e9bc81ead1aaaeda213083
f1f260f24a728a61cbfaad1f0474fdbbfef12484
F20101122_AABAPC supan_k_Page_078.tif
602fa084f7cae6ff57294e33ff2ab229
36226234f6669d32d590f4ae766582196c313e6e
F20101122_AABAOP supan_k_Page_063.tif
7de9224f8f21681094efeac6aefbe604
83107f19c73bfe5a16f81abd7a6372a265078329
F20101122_AABAPD supan_k_Page_079.tif
f277ef28de5fa97e94983a07b70f3116
2fff64bcc1f9cf557b6be07e35fe0263d386e09e
F20101122_AABAOQ supan_k_Page_064.tif
b19dbb02d69815832870162f64c687ac
fea1e5da51f9949b17fa1ff489f5774872f689f1
F20101122_AABAPE supan_k_Page_082.tif
8c78e520507d2c47e99fad6e9666edcc
25728101bae08bfa938df90221e0ae56492d246c
F20101122_AABAOR supan_k_Page_065.tif
2832f39550fcbfb2f59af1c77629487f
d75b6602b78db653b8eac11218444be7f8a1140c
F20101122_AABAPF supan_k_Page_084.tif
fbdf9ba536ebd8a93294ae6143f7ecce
427be9a7c6e8a9429f908203276a01723d573ff9
F20101122_AABAOS supan_k_Page_066.tif
175698182a9dfbb739f0566414194316
6db99b0ce55593f2cbfcb057be60bd2d24d68465
F20101122_AABAPG supan_k_Page_086.tif
f13ac5aa9ad2e9e46a0dbd933f640509
8d2fa9d8a710cb703073de02f4a76f2025bb3a0a
F20101122_AABAPH supan_k_Page_087.tif
881590a9de413b8faa89c66ede3fb32a
272d0462f37c335281f138c17bdca883796598af
F20101122_AABAOT supan_k_Page_067.tif
4c56ca34c99438287cb20ec0011fa125
af82f4461f3a32a142f628d9309606d5dd1cebdb
F20101122_AABAPI supan_k_Page_088.tif
757ae5ce79ce5870675541c391427d80
9fed8ad2c66d02c54b44b78038da6cfb4c0e53c3
F20101122_AABAOU supan_k_Page_068.tif
14e66e32707cc988f2734f113d24e7d7
21e368a66849baec2759858af2193fabf1d405b1
F20101122_AABAPJ supan_k_Page_090.tif
33e38b0755dd95af8cba6df5ea0f2fa9
488f4c76bba86be92abd7c6c7696453f6feadeb0
F20101122_AABAOV supan_k_Page_069.tif
1a02726805fb82baa57c6fd283791a43
d5428da4ba98043e4a14e7a08d2cac559eb256d1
F20101122_AABAPK supan_k_Page_091.tif
c58f0ac1a0a0b5dd9b7549f5466efb6b
826adcfe2d562369f5ea615350b14fb90f53040c
F20101122_AABAOW supan_k_Page_071.tif
9b483047e86510edd6c5592d892143ae
5f26be9845beee923b5173a7a201cf3a0088991d
F20101122_AABAPL supan_k_Page_092.tif
f282f048fed454c896234231d27a2106
9d752d53831512e0757b2ae43acb7ff015ee2964
F20101122_AABAOX supan_k_Page_072.tif
fad2e13554a55af5adb9af385ef83793
b970a358bdbaa4f3149e89e4db4e6de36edef391
F20101122_AABAQA supan_k_Page_109.tif
1a1b69ed932f7cf8ee35848ce7659e33
2c513cf6e5a408f64ab94af1ede76dc84d180c97
F20101122_AABAPM supan_k_Page_093.tif
db0072e116abcbef74ef0adca4121083
0a2e9b463bd3bfacde6566b725b27d613f477fa0
F20101122_AABAOY supan_k_Page_073.tif
88e91c65afc067b6a636f90846dde18a
4654846ac57fdde4f5854f5887a55173bf1d6f27
F20101122_AABAQB supan_k_Page_110.tif
ec6dff45286c6f9dd7966a5001dd882a
d159d0472c3278a29b10f804ee2717448ee83ea3
F20101122_AABAPN supan_k_Page_094.tif
a0695d07f3f877889b259f55f476e9c9
6107da2e26c658bf95b8cc108f938a97a26c0cc6
F20101122_AABAOZ supan_k_Page_074.tif
2b8e66c7fffcaa003ab9542d4539f919
25376b5459cb877cc819f73fd95c530f865d9c17
F20101122_AABAQC supan_k_Page_111.tif
0c61dcab76bb7ce2106d56f2591feb6c
448a711e8008099f8af61485c64f62f7174011e1
F20101122_AABAPO supan_k_Page_095.tif
c9243bbffe72223c4462af6d10b5cec1
41a52f76eb7eadc9a13a48624e47b93125ca7cc6
F20101122_AABAQD supan_k_Page_112.tif
11cc1f4df80ec8ba5f65a29d360b3344
f1507960d0c65c81d17922e3f493fa823900bfc3
F20101122_AABAPP supan_k_Page_096.tif
c94b55c7b9b7b3e02a915a2130273264
103fc384fd5ec01f1f991a571f273705d6cf2ea9
8425398 F20101122_AABAQE supan_k_Page_114.tif
fd64e637c9abb7e1ec23fb1938362777
9a4a0b181b35c43f63e7b7acfa1cb69813ba3da5
F20101122_AABAPQ supan_k_Page_097.tif
a21a2735d7f69df8a5a680575e9dfac0
14f83eb6919027001dc9452fcc462b7d126596c5
F20101122_AABAQF supan_k_Page_115.tif
5a5cd693a620cf2464678387944a38bc
751ee83a36725eca29172e1bfb4e8c178c4b52ca
F20101122_AABAPR supan_k_Page_098.tif
cb2b0aabdde23e2dde687b9a5d21e76a
ad7900cb9eca9cb083d568c51981a701afd110a0
F20101122_AABAQG supan_k_Page_118.tif
153629417949d460e092604b20c701de
f93e8739b1f75384118d636c45515cbea2511b1b
F20101122_AABAPS supan_k_Page_099.tif
bd32fbc34a6d014fbee6d10bf288f2ef
2587e549aa9bddbc4b24d6f8eb75652470dd153f
F20101122_AABAQH supan_k_Page_119.tif
5331860e2bca9ccc5de918f0939854f1
7669b70cb267713788b882a4ac1d9ea7e2377f13
F20101122_AABAPT supan_k_Page_100.tif
db46b8ecf3f5234562b62abb8edfbb42
b2bce02f5c74cb8e765aeb7628f94a0184a242d7
F20101122_AABAQI supan_k_Page_121.tif
6050602114318e6a8398c3897845f183
a3d5945141bd0cee7a580d9a5e7391e7171722c9
F20101122_AABAQJ supan_k_Page_122.tif
26791ac98aea2431cd468f875c3713c8
cb59191c930238c42d3af5d4763bd22dd5086098
F20101122_AABAPU supan_k_Page_101.tif
07030ae5df4c682b046ee804cad1a419
239647d4f09a9cdcb4a061969d57205a25d7970c
F20101122_AABAQK supan_k_Page_123.tif
787d623a8d7a345b957fb3045ec471fd
35b27eefa5365853142130e4574dacd5a2e4d85d
F20101122_AABAPV supan_k_Page_102.tif
30f17c35fe54921f58f9677d3972bf61
efa2a3ca78e3aebe944144a51a30ff8f1ae00398
F20101122_AABAQL supan_k_Page_124.tif
3cc96919e810fb13c90db64bb8111ae5
67a74bf61b85e67bf376f3934df92ac7a5bb1261
F20101122_AABAPW supan_k_Page_103.tif
80f564df53b7ec577d4634ee3759d91f
533fdd04605947db281d4f317a098786dab2784e
F20101122_AABARA supan_k_Page_141.tif
41c4742b058213dae0b2b79d38578196
9a9bbf90a806a3fa15c0b72cdb1f537196774824
F20101122_AABAQM supan_k_Page_125.tif
c868030372c83048f1841f6a2c81519c
657944cd1b12d078d47fbbe1354a0f7da0252236
F20101122_AABAPX supan_k_Page_106.tif
bfa05c0b314bdb2b2a25a0f53cc29e3b
1245415ec45cc3c17a31a74595249e6eab63f9e5
F20101122_AABARB supan_k_Page_142.tif
abc6733b00a5d0f17b1c1469affb89db
63b4514b9209923843118531d730ef547228217b
F20101122_AABAQN supan_k_Page_126.tif
4c73f08835f30f9a08a7848e2c30438c
350863d2ebe43622ee59a56ec38b05655b394580
F20101122_AABAPY supan_k_Page_107.tif
1b9585fa288d188d30e7e6c5fd36b518
b69ed94a5de22e4af447a789aacf1a9b7fa78fc0
F20101122_AABARC supan_k_Page_144.tif
3b38067b6e74c6dd82e5040756fd7869
2b009520e5665b0ebeaba527d6af9dffc513aae7
F20101122_AABAQO supan_k_Page_127.tif
3d996e2706b2847038ae0a8fb0f061ed
dffdf7e4542281ee32cf269d900ac6e89474b260
F20101122_AABAPZ supan_k_Page_108.tif
3bdca77c8892b5f22e05db15e85136f8
504eff91d8b201d910a64981f09df2ef47e45fe4
F20101122_AABARD supan_k_Page_145.tif
a560847d114e8da740a53f41395290ec
6c93f4d2f036ed14299314fe580a07eb8430d90f
F20101122_AABAQP supan_k_Page_128.tif
95491060b23e8804e15f8be11c019af3
26c7f50858d91ed13ac5acdd41d331869cd448b7
8354 F20101122_AABARE supan_k_Page_001.pro
31304d4faedd35817db606ae06c5ae38
071517befe9c5e94dd62293dcc5127361fc1c524
F20101122_AABAQQ supan_k_Page_130.tif
b9eb764530646abc311ed59f4d5e9c36
c1d7e788df7157ce670f47718b587e2f12fbf7c2
F20101122_AABARF supan_k_Page_002.pro
8b6a5abdaf1eac8a42fb3b793da72883
4188880280ec81cc2fc4b85e19730bf4bb018987
F20101122_AABAQR supan_k_Page_131.tif
b42e93b0bbb8e6d63ad33b353a1a00da
d8ff2a16beb541e7916d8fa01072adeee6493521
1949 F20101122_AABARG supan_k_Page_003.pro
348d532fe3099fa6e35076ecbe2b1814
20a5c4b2097a907ca591182012d8d62840bc1b5f
25265604 F20101122_AABAQS supan_k_Page_132.tif
73c6e1469194e8bab1b3644411511467
db3548a74e0f9103490c36146899a2b1593d4dc7
42044 F20101122_AABARH supan_k_Page_004.pro
7d54812f945de77504adee0fd87c1e6c
9186b0c2a986df7260d2ef5b498bf24a0594dd04
F20101122_AABAQT supan_k_Page_133.tif
c048e088e71c50e6e634a3bd5e70d14b
6f17438c0a7b49add46511c0c4c2ebf2d57f4e64
3212 F20101122_AABARI supan_k_Page_005.pro
0f1bb7e33a205fa25e65462e4059bfb5
bcd65f2dbb493fc69be101b53cc7f422c1da135f
F20101122_AABAQU supan_k_Page_134.tif
08076a8c0e5ffa58b26d90b8f1a06ae8
c392cce4b116375c1146787cda5b7cb1c22e481c
75466 F20101122_AABARJ supan_k_Page_006.pro
205fba705e4e85e3d4f994b1acebb7b9
4338a69312940984babfa251a2894252a95a9089
F20101122_AABAQV supan_k_Page_135.tif
e70865ced46307d9e618381cc631007b
144df467d84541ed835e7f35e5670d034c2a5cb3
94322 F20101122_AABARK supan_k_Page_007.pro
b6fd43159cc87f8efe28add086c32e71
93fe2f5c797b2f964157ff66c242f99f184aaf17
F20101122_AABAQW supan_k_Page_136.tif
2ba72b2d14b1e8dee8ca97f58162e18c
5c0ed487fa4c7992d75060e9968591fa9ca42733
6374 F20101122_AABARL supan_k_Page_008.pro
60feb20fc2375d6476a39b51ffd26a43
4424ef4435d7939a368f95275227c9a54c86fa9e
F20101122_AABAQX supan_k_Page_137.tif
ce70889288585b99406de574a596fffb
f9df4069fe192d53fa6d3ce9c67ed95427522add
42367 F20101122_AABASA supan_k_Page_025.pro
daac06c3ff2b9870c99c7b2e2a0bfb56
becd7cbe723c83fac5a8766a6fae8518046c44c2
56634 F20101122_AABARM supan_k_Page_009.pro
61474e06b4273e02ee6efef5c4ed7ac1
f4f0c7d4a17b0448976b521e9f2beb1cced69c62
F20101122_AABAQY supan_k_Page_138.tif
219dcefa8d6ba96468329f871baa7aa9
d51a7fa1aab29abea75794f126de52086b459e24
50941 F20101122_AABASB supan_k_Page_026.pro
b5524475f44538a2544cc56b7635f379
75749aef6aeef5db476d75161a567ae44dc0d394
57464 F20101122_AABARN supan_k_Page_010.pro
159b2d5df10448cf5e982646933c6ced
ad0e0dfdc9f7d17416680f4618e6d8caabbe6119
F20101122_AABAQZ supan_k_Page_139.tif
3164ebcad19cd2196cb37b69e56ed281
a6d0ab31343ceb888bb90c4aed2dcdccc24ecb87
41510 F20101122_AABASC supan_k_Page_027.pro
856bf5869022eb4f0f958204dfa77c7c
d23e53474a48ebdc96a0fce1f0f1930b1e8f509f
76313 F20101122_AABARO supan_k_Page_011.pro
8deef4aa7914bc2f02377138ce7cd4bc
228098ac50f4fb1f9df0eb58ddf4c1a78f2cf590
48801 F20101122_AABASD supan_k_Page_028.pro
fd155b98b74f4acebdae0d9498b6acce
c245463b7a3ff949d3ade9d4272ea0fa76de8cfd
49316 F20101122_AABARP supan_k_Page_012.pro
49325ac45680e09d112d831b0a7a0609
7bcc75edd96cbaf686ae9bcf3e778eb31ffb22d4
45648 F20101122_AABASE supan_k_Page_029.pro
4879de2f593b54d2b35de3529182328e
00e8f550f98a0de48386640f0cf401b0a5cb08c1
38298 F20101122_AABARQ supan_k_Page_013.pro
10d06f0b1ee92e2e38dcfda9cc5affb3
e93030fe0bb41c4097fdf12b6e8782881b408d33
48435 F20101122_AABASF supan_k_Page_030.pro
e817773a1ff957ef756b55fbc0116c50
15d92922a1aa8fac00b6266300cddf131bb68a67
40543 F20101122_AABARR supan_k_Page_014.pro
e51f54e5a64f8de03d651a4c046f6931
be528fbc85a9a65e0446ca365c939d8b1b5727a2
47076 F20101122_AABASG supan_k_Page_031.pro
36636b07321530219b01f1c4197c9a86
cd0da73da256ec1d36e00199ae5e27cd4e211e87
43385 F20101122_AABARS supan_k_Page_015.pro
4685b4b7dc41f200a8a5c2d267c9438a
24c9e7bb7d8f1ad7986b9a22ed80a68476d8220d
45504 F20101122_AABASH supan_k_Page_032.pro
22982c5da84ba5944b9aefb16a30ec5d
1eda79febfa542439dbbf342f57b87b95f8d5f00
36177 F20101122_AABART supan_k_Page_018.pro
c509d75e129e420ca2562172426a2c4b
0e34e9ec591a6df5e42e388d4829ba387993743e
37545 F20101122_AABASI supan_k_Page_033.pro
b2171cb33d11e5a2b66819082434d6a4
05d72e8569df869d536a554a5cde6df6ac40cf70
47834 F20101122_AABARU supan_k_Page_019.pro
21cd5a056ea01229abcd2cbea22c4763
118ec14becb485e29ce5ae50c8c2e579f1a818f0
41395 F20101122_AABASJ supan_k_Page_034.pro
72079c626c9737626cfa56c357c60bf1
e911d62ec1497e66c9388655f727bf9a81431a37
31318 F20101122_AABARV supan_k_Page_020.pro
7a1aec298376d0bdced6c5ebc20461dd
bf04dc7df9af68ac8d63c1d1967429821419d6f2
46863 F20101122_AABASK supan_k_Page_035.pro
4cb18fdb120e5ff29f535527da1f0e3b
606489f423ee2c3023042b652a722d351e7ecb41
46730 F20101122_AABASL supan_k_Page_037.pro
fbb54cbc96d276f7ec2c874501123803
d1e0d1a3dd23a67e18a85c84a12dcd8a1ddece44
50463 F20101122_AABARW supan_k_Page_021.pro
1c4f1113ab0827d263948084751b6dd2
66af7f0dd358e021a4b30809a9c1c56b6cdeb98c
40707 F20101122_AABATA supan_k_Page_054.pro
67f386c64c57b7e777fb9d42b348aec9
8c2cf8bf1f503563bbc319dee7db7172a2264baa
44355 F20101122_AABASM supan_k_Page_038.pro
842f97a9eed0c8650174d13df9748e46
84931fe83bb71b0d24f7815a3cd2f986da9b8ea3
49729 F20101122_AABARX supan_k_Page_022.pro
44f6fbfc9327962519403818317ae160
3e8d817717e7b3e3bf93b0cb2077d6a7f8dbd957
2076 F20101122_AABATB supan_k_Page_055.pro
937cba3b9e82bfd67fb2efb59a036c3c
7e43a2ec341800e2dc449060df12f148f4b87872
46049 F20101122_AABASN supan_k_Page_040.pro
eecd970694abd589e1e4e9a4884c98c3
90484234bd358056eec877dd2bbddcedeae3a0bd
46753 F20101122_AABARY supan_k_Page_023.pro
dd109d7d5c4b58d5955aa00766055409
c943eb975dea7dea7b90f68d41637744a6b405f6
36343 F20101122_AABATC supan_k_Page_056.pro
d9f1c8c35ba7fd47a6c5bc12cf88ec53
3cddd2a9d2ed928121f28f25788775bfb4724f82
49246 F20101122_AABASO supan_k_Page_041.pro
6d775f9f1d83cbf50f8901aa4341d3ad
62401a02efe8f8478c21cf86af12ecddfd7b8a4b
47289 F20101122_AABARZ supan_k_Page_024.pro
d02842283f34e7fc95acd71beb9c9164
615e7a315663352e0fe3648e9db033448debab13
51062 F20101122_AABATD supan_k_Page_057.pro
6e976c883d80159e612eddacc96ac211
1917899f9d96af6178d8c68c8ec71e0af85aef1e
47358 F20101122_AABASP supan_k_Page_042.pro
aa4140250dcd3db6c578fe48aa842928
655d86a3e0fd232d57719f700f1460dd6b621555
24195 F20101122_AABATE supan_k_Page_058.pro
ebea29f7a6d00576b7838e5f427ee854
1e09728b71f5d573cfea7d43bbb242f43488521b
50423 F20101122_AABASQ supan_k_Page_043.pro
e4c66fadece26009e9b88ce3ef7fc8ae
671a43578b87f1401fdd0c88fbfe844c69a909ab
32964 F20101122_AABATF supan_k_Page_061.pro
3e7acfeaef5dc6feb693d7d74f50ce51
aca91cbdac6e50d118184b63714cce3221c4fee3
45163 F20101122_AABASR supan_k_Page_045.pro
d2e2d98fe562c402ad21053b6c21974f
eb6e4b211e1aa2caacde297f6c2114adb0c1c915
36077 F20101122_AABATG supan_k_Page_062.pro
c42272c93d72e3ce61dfc59deb0939b5
ab1a52462dd7036aadab2590f8b2691a5c2c0b2c
33102 F20101122_AABASS supan_k_Page_046.pro
6721579f3ed08c736df6c3a62fcb632c
35b53f54435b3583e34c61c5be1b378fa3790f53
40749 F20101122_AABATH supan_k_Page_063.pro
6854bc14ef8b2ef047fbd8e8730460e8
8c5a00ffd220d46040c0912d470afa80c0f19048
41941 F20101122_AABAST supan_k_Page_047.pro
e679cd8ed4eb3ee4303f4700b993639a
6d006ce06006a6304cacf41cef0b7abc1f4d2ce2
44281 F20101122_AABATI supan_k_Page_064.pro
b9c2f37215cdf20d095a3462d0131307
3e6341e4e88c89444ca145b05751025c44150945
34136 F20101122_AABASU supan_k_Page_048.pro
11b31922ba26b892a51eb87081441f78
9a4587403cf6c377807832bbada172daf3947538
39586 F20101122_AABATJ supan_k_Page_065.pro
14ccedfeac87d58ab1522017f7581089
0339cd62551c4db6ef65e5138327708a8adbf9d8
29505 F20101122_AABASV supan_k_Page_049.pro
5a2f09e3db7e619dfdbedf3c05235390
e025faf6a22a479fa8e46c9741fb68b754ea541a
35330 F20101122_AABATK supan_k_Page_066.pro
f4505e67bb7ea8787967934203ac4a1c
37768348d752357442f2c68ee4874bfdce69d878
37253 F20101122_AABASW supan_k_Page_050.pro
cea701da2c3d8fa8f76375b221fb4db3
7d0b70f4d68cb032d2882f9ff93ffdfe1fbf72f4
2804 F20101122_AABATL supan_k_Page_067.pro
668cbfbe7373582120e3f4b243e511f6
db8b565ff48ebc4774b967fed64832329ea8c11b
42061 F20101122_AABATM supan_k_Page_068.pro
246ee0e087d42aea9c40af8600fc6a78
95604c65331a8b3269b972776177ea5afeebd8b1
44801 F20101122_AABASX supan_k_Page_051.pro
c94701dfbd930709d8df8c120dca02f6
551920543e9217cc8b8b374662dba014f4902047
34471 F20101122_AABAUA supan_k_Page_083.pro
772fed70800ad19b0c86b54c9cb8c016
31e3882ba0d4014e564d67959e6c01e5e3365516
32798 F20101122_AABATN supan_k_Page_069.pro
c12745356b75552b385dd003e82f8a1f
46445045e4785fced9bd074135de124fc88463b4
44232 F20101122_AABASY supan_k_Page_052.pro
b44dc12b04b605beb01b4d35bfa6fc36
e1f195c056b8db27bb81ec8f07237e3f3e3bf12c
37729 F20101122_AABAUB supan_k_Page_085.pro
7cc1be8b2befe17d14b7e113a891f14e
0ae2885b4d31e382233c419c84a14c0d4bdd713e
22905 F20101122_AABATO supan_k_Page_070.pro
fc0a2008bf6f2de000c42ec0ba481106
13aa2ab78993fb3cea54b83e7394b637ecb6c0e8
43656 F20101122_AABASZ supan_k_Page_053.pro
8af7039ad1c714bf1974401477feea6c
764b9d5d0eb9d08f82f1fbca9a5177fb5134e000
43807 F20101122_AABAUC supan_k_Page_086.pro
a10cfc9d1003d713d4d7d8cb03ebd97a
863b89b1664fdfeb7b392fd889edcaf273be638b
33025 F20101122_AABATP supan_k_Page_071.pro
1ab8955849e0e6e1f75b5a4825aed192
953d51f55c4b1c82769aa006e56961d53878e788
49861 F20101122_AABAUD supan_k_Page_087.pro
c1b6047e150c1f754ec3438ce81be8f8
0d137ee0626659431fee53067244266364b271ec
34003 F20101122_AABATQ supan_k_Page_072.pro
394bb4b4fa05a3df19ed9fcd59c85de0
fc9b6af8f3a725c8c2c2b44e568ead78c85c70f1
39225 F20101122_AABAUE supan_k_Page_088.pro
3bc31e099b29b78d4204e43ecb5598cb
41d32cf5015bed3d3d26b0398778801e45602be6
44633 F20101122_AABATR supan_k_Page_073.pro
57e6722ccb9920612abd81f12475d21d
a4aeb3e47acddaae059c7246572038db89e676ad
841 F20101122_AABBAA supan_k_Page_119.txt
84205c255cb535a71d25f046b2f931b0
1d5261236cf5e0e288cd591096d82d9706383c4d
28745 F20101122_AABAUF supan_k_Page_089.pro
698ec76546705423709e8915f9b31d78
bdfbd63f77cc3594c364517604375bda643fc7c2
35470 F20101122_AABATS supan_k_Page_074.pro
2e86939a5a1f0804418077646db9cd28
fcee209a484cb61b3c0992d9f46d8ce7ee86c7df
1693 F20101122_AABBAB supan_k_Page_120.txt
7fd5c81941e717c8809418597738f58a
3e0f9de3b70a3db135e184f5bbeefcd9ddfb09a5
34484 F20101122_AABAUG supan_k_Page_090.pro
4281674bee5493eb505d276f2dee4478
0f35c430fca1ddef7635ccaec69360f923a0146c
37114 F20101122_AABATT supan_k_Page_075.pro
927b295eb33fea2d165990fdf9ef4fe1
c4f3f6472f716c692a6eda0d0784efa66c90fa8a
1822 F20101122_AABBAC supan_k_Page_121.txt
b271d76c9268469005ac767c99600e11
33b317da881780143857a6f35e5cf4ac165aef3c
29055 F20101122_AABAUH supan_k_Page_091.pro
ebe56effdff1485a79fd7f7440c5727d
0d3ccf1034a881e12dfbf0ec220bf82f38a56737
29057 F20101122_AABATU supan_k_Page_076.pro
536d435001304a536dca9a6b1d953a44
a4f078dbb2402db66fe7dbc4243ec81457356c38
1917 F20101122_AABBAD supan_k_Page_122.txt
d6202ef6c332f40ec01451366e28fa92
1bdb0f8b1e3176b21e59476c9ca221fe1ee262f0
42045 F20101122_AABAUI supan_k_Page_092.pro
c85b56e2aaf4c060e8ce4d5059ddedbf
70626c0f252579ae115276d18147876ffdedc080
27289 F20101122_AABATV supan_k_Page_077.pro
68481c64266dccb4f689a6d298176ef1
bec8f7673045a927d5f6e720d60620a5003a3a50
43323 F20101122_AABAUJ supan_k_Page_093.pro
8b315ac848ecf8c582aadbc17a0123b6
3820b79a96993acfdb8e5306f60f7f61ad80e116
42588 F20101122_AABATW supan_k_Page_078.pro
4dc2f925e12d75f01b4641a078884a02
810a6d8c2ab326c23922d4f6066944a856a42746
2112 F20101122_AABBAE supan_k_Page_123.txt
526aa5fbc2164194d82fa24d64aa5705
b3daacbd89b96d7695bde0a4dff35a3f3b408ccc
36422 F20101122_AABAUK supan_k_Page_095.pro
0ba01131ca7cfa6d1edd5b4d527a631c
8a780971b3bae174a40352a36dcffb7327fb501c
50708 F20101122_AABATX supan_k_Page_080.pro
be1d82c0bc54225a9987305ff7ad3f23
545f5af1ce9438c6e4add61ba118d55bd3940b85
2877 F20101122_AABBAF supan_k_Page_124.txt
43852d29bba9150f5c24717d5b327d6f
a90d7214d08bf307a59d53a63ecbc66269ec1cae
33293 F20101122_AABAUL supan_k_Page_096.pro
6ebe52d5aa6190d2d725fed0407e5ba6
70777cad466dc06833ba95965c087186f4bb7733
516 F20101122_AABBAG supan_k_Page_125.txt
2e5fc306c109796817c4767a995c657d
5da18965a6caaa64c7251c8a72027dec53ebe841
9975 F20101122_AABAVA supan_k_Page_114.pro
11e902b5ed14f56322892bc39760aed3
9eab87d0a71650fa8d5e742df1f642c49fb71ab6
30871 F20101122_AABAUM supan_k_Page_097.pro
44a9c1965cde1ef7fc04db57d677cb53
635333cadbf8dff43382cab274f960fef6f50479
42592 F20101122_AABATY supan_k_Page_081.pro
d165a1563c6ec15bc83d48d0ae54434b
3973baee986c5a575497524e13e2f7647cd0b946
68 F20101122_AABBAH supan_k_Page_126.txt
8408b110641d37e3cd2fc5318fd81791
9328ba787cb26baa92c9a9c1063500046d1456de
27067 F20101122_AABAVB supan_k_Page_115.pro
359dd4d6a6d40bd68fb5d549299d2174
a2c6cb50fcf67f9f064379626606115fa689980b
37324 F20101122_AABAUN supan_k_Page_098.pro
e8e8022e9b8a1f049d49044a2edaa6cf
89bfb8cb9d9fdac41f7afc484475315e60706026
F20101122_AABATZ supan_k_Page_082.pro
daf1fc8096c6814105e4b4ab698b4e90
bd9b0514eb1bf63178ed32b9583dbbc15eed4cc5
3093 F20101122_AABBAI supan_k_Page_127.txt
460543c070a6ec1362c3a14498e9f0e7
a33139b2a700d5e6174e12ec06fbb62de6b98688
48516 F20101122_AABAVC supan_k_Page_116.pro
31cf6afa2567bd23a5bfb911fa946b46
94a32a1b98d7e5ae1e1efd9f41871be4a368137e
41589 F20101122_AABAUO supan_k_Page_100.pro
f9367c22eed9c7592f3b3c474a67dd25
558736020c19014ac5de30f0650c3a14cc67327d
3090 F20101122_AABBAJ supan_k_Page_128.txt
2b5fffa53c203e6acc386b2cf41367f4
39cdd42ab8b37fd1e001d9b3f8ba74415765d9d6
47012 F20101122_AABAVD supan_k_Page_117.pro
2a401d0580ac299e99123515826d8d39
964dcef9d6d9986b541d7c3a0bd7a2ef7f7071bd
32829 F20101122_AABAUP supan_k_Page_102.pro
a74d7ccfff361124bfa3648a480edd25
bd54d5cd960d4e9a2bc4508caab2ac0cd966b042
2950 F20101122_AABBAK supan_k_Page_129.txt
e1607ce1ba6dfcf428fadf5d8fc7ad6a
9abca9bf94c8edc54e0ded1c26fcffa1242a0b54
16926 F20101122_AABAVE supan_k_Page_119.pro
6b9fb1124dd28b77dc7e56212d9b245a
f7cbeb07702ee8c588fded5f4689dd8fcd85566a
23073 F20101122_AABAUQ supan_k_Page_103.pro
b0c53d7dd3122f4391d6cf3ef785f6de
8e89d6eb813aa7a55e8a3fc0dde0208a055eb534
1099 F20101122_AABBAL supan_k_Page_130.txt
403fc704aac684c5e3e8c007409c1036
a6b3fae207300abe3fd222e377ddf7a52593f3db
40324 F20101122_AABAVF supan_k_Page_120.pro
3616ba74e26df62bfe1e4100d17f551c
bd5be24e5efd010a80c064d08bb3e963f9d3b036
36662 F20101122_AABAUR supan_k_Page_105.pro
cc3da1f93e0f669d5b420ffd66b09619
8136c14152d2f39ce815cc27250da048f44576ae
3403 F20101122_AABBBA supan_k_Page_003.QC.jpg
c3a287dcf60e617fc87c0112fff3ebee
4beff7fbb459d7d87c6ff2167fd318dd67d6a45f
849 F20101122_AABBAM supan_k_Page_131.txt
2cb45c957e36aaf2afc909421eef623b
86fb78c6a2b4208aa9bb89f1471ea74e05b801e2
46038 F20101122_AABAVG supan_k_Page_121.pro
7928f8e3c06df275be57609b552a9385
e3d53726e29064e95fe77b5a8b48aa3f34e6d516
50075 F20101122_AABAUS supan_k_Page_106.pro
f7b6e3a1fa3df339b535bdc20b5eb457
17bfbd1ab326b08e4b738aecbd8c26336b164c8a
1423 F20101122_AABBBB supan_k_Page_003thm.jpg
686f11048fde0bb9546dc765c4a5f969
845c64e025f727f641dc940c431733e3f338256d
206 F20101122_AABBAN supan_k_Page_132.txt
a7947899dcd97765c26e12be0fb9723a
42c72c9d013ca7ea18a7368f721611d04d11cd72
48500 F20101122_AABAVH supan_k_Page_122.pro
3b136854a462ce1ef1f943443d026235
5ce747c19f57571a969c592e318b206ce968db33
29227 F20101122_AABAUT supan_k_Page_107.pro
627283fdc8b68f083d1373c6f2351db7
a65f651dde1433fa9167d93777abb2687b9e3d0f
19974 F20101122_AABBBC supan_k_Page_004.QC.jpg
cfd45804303237576a3e3b7e2b46316f
bb324bf5c694ec54b9df6ba8f12ce5e4ef5e0833
169 F20101122_AABBAO supan_k_Page_134.txt
e80dad182606faeeb2ec2d7b92c0650c
8a8e5c949fa6734064d7dcb01817aa6af98bbc8e
52523 F20101122_AABAVI supan_k_Page_123.pro
c07f3bbf13cf62efdd065a84deb51947
0c16eee28fc34a7a9180a56693d2d6fb0a2bd14c
51270 F20101122_AABAUU supan_k_Page_108.pro
ea1c2a6965b722d7e0cd34d476b6af63
ad8f4ca4a3963c28ff5ffc967106f4cd9923fcbb
5763 F20101122_AABBBD supan_k_Page_004thm.jpg
31678c94f11d3cf9cd8a58252e62f4e1
82aff2352c7f0bc2efb7c4318afdbf997d5a6096
521 F20101122_AABBAP supan_k_Page_135.txt
58ba2230bcca2720c3474ae261535575
d3f57c119c45e0817c6c7710223f1242dfa21c30
69501 F20101122_AABAVJ supan_k_Page_124.pro
e060ea46f30c5c3a94053f04783c208d
6cc74234b325c4d83a3269e3f7827f27d39377a8
46381 F20101122_AABAUV supan_k_Page_109.pro
f006d8780f8bf0bdbfec8e8d9bf47c69
37a3856acaa71ad6067b21eee7cc0fd230cd2229
4194 F20101122_AABBBE supan_k_Page_005.QC.jpg
0a47e9f8ac97b4156d9b19eb854326e3
ad77b8ddf77b91e7e90f3f88b7bb71f6234e0386
1587 F20101122_AABBAQ supan_k_Page_136.txt
54ad3b21d27f5cde877d0b417a551ddc
094c00ab481f2a6051c8c5dc8c4ea8f84d4e5a4b
11806 F20101122_AABAVK supan_k_Page_125.pro
267d082e7771900941ff42e162940edc
f3e093588b382f042956a731c811dd68e25efeb9
44624 F20101122_AABAUW supan_k_Page_110.pro
efad7b4494dc0a1f2dff1eb13c5f2c71
3535ab0d97e6118cbb9014a421c92bd1c47e3113
1603 F20101122_AABBBF supan_k_Page_005thm.jpg
0acfd394eeecd4a31cad31e49faf94b2
cdfdbc2ebc304580642e61b99ec6b71579bbe8ef
1574 F20101122_AABBAR supan_k_Page_137.txt
934537f453eca65f6c308e9a6aa770a5
f62a0153085165573593f7979c11604901f0e3e0
1132 F20101122_AABAVL supan_k_Page_126.pro
c2fb7d25c0ebc1b1d5dfea20873efe82
3d23e6c99feee7fb2f9d81a0fec8de6e8b568e5d
55112 F20101122_AABAUX supan_k_Page_111.pro
a5e9723d7a235910885182127ada572a
b28f56c99b076020b1eb33812cfe1e42d9778cd2
4542 F20101122_AABBBG supan_k_Page_006thm.jpg
adcc45f9342bb24afea9ddc823935eb2
e18c96db184e21ed930ae7f1cbcf87e72e3124d7
58637 F20101122_AABAWA supan_k_Page_143.pro
741eeaee32a0f309428b9c2b0b98235f
0d30ed4abb65b96414aba6ce6ea88e92f3497759
1513 F20101122_AABBAS supan_k_Page_138.txt
6e001db3da41c6b4d5c3a85ec249d6e4
9e72e7d8f20832e1fac99d8c326822568e64aa05
48746 F20101122_AABAVM supan_k_Page_127.pro
c6f1f217ca528ac67554cff0b3a9a529
106fb5a8fc5f3eee64bc8fe7d58e0336b2e4075b
47682 F20101122_AABAUY supan_k_Page_112.pro
0df21a1f49ef3db68afbcccd2b361fcb
04568b13071fa68af0226ea850a836734f53b918
21108 F20101122_AABBBH supan_k_Page_007.QC.jpg
4d318d3732e8bfb211e55b6f8eb52407
f46bbda9501d1a888d04deac7c8ee11469f6e88a
38761 F20101122_AABAWB supan_k_Page_144.pro
dbf8b32160ccf99b4c3b50575d23f5e0
3e1aff82abe6a44ff2c9e96cb5c571abdd2a792d
2478 F20101122_AABBAT supan_k_Page_140.txt
427b71ad1e6b0e706e6b6a97a0e21e78
cdc5a4d8815411c4d4eb8e269e0b553b73da8bd5
52263 F20101122_AABAVN supan_k_Page_128.pro
54037b7e0af65a21791da1079b047f3c
72a23ee4979630c23b5fa1ce05b86ddc4680e0da
5359 F20101122_AABBBI supan_k_Page_007thm.jpg
e10c17e58912d6ab6e7bfde4f49a18bd
17f4bbdd4b272ab475bd72f3e4130fa34d50f4a1
16187 F20101122_AABAWC supan_k_Page_145.pro
ca511e701cb52732c295828c4f3e61cd
fcddea388969822c735ae37c845f7a736639ef1b
2363 F20101122_AABBAU supan_k_Page_141.txt
d898069ad357e95ca5d270c30ab43700
4c711f9057e7fbc6bf96c432b731e3952a2e0a1e
50349 F20101122_AABAVO supan_k_Page_129.pro
5e305f6656a73d093845d21e2b5a13ce
9622145bc014c3c9ea8f5590c9b48f7971c1a201
47027 F20101122_AABAUZ supan_k_Page_113.pro
a91aed85cbe2f9dbc89bb0ffe846cc53
efa9b3c84c361816402cb465dca43ccb4c5a5800
3954 F20101122_AABBBJ supan_k_Page_008.QC.jpg
074c647911238270788977cdbdbd7671
3cab0a090e49560feba9b555a03af4fad7770044
464 F20101122_AABAWD supan_k_Page_001.txt
a5cc7b49b96bb8f3437fe1b49a31c3da
fdb71aac792989e9b6aa540080dab8039790b0b3
2640 F20101122_AABBAV supan_k_Page_142.txt
413ea6610db4b43528eacd59d8926a2a
8cb1de7d392931b55c64feda7ab203795f3cd524
21340 F20101122_AABAVP supan_k_Page_130.pro
ee0fd9407a388782ba9d58540429ab38
6e2b00a178019af560b2b1d5fc193cb9158e3ee8
1555 F20101122_AABBBK supan_k_Page_008thm.jpg
8e1656e7d441e7f520584e26eb347f18
7d500d9207b50b529afc6e4bbe48642d37ab365e
110 F20101122_AABAWE supan_k_Page_002.txt
0dbaf9518d806af3bd700fd393d4f227
85725696af48466a7618801908a642777c6c5e5f
1630 F20101122_AABBAW supan_k_Page_144.txt
0d8a64d2034065b347de238fe7acea9f
f2f382144b4bb0d1343ce1065ad555f53affe302
4655 F20101122_AABAVQ supan_k_Page_132.pro
f0a7e2a78fd8d0a310a80c7f8a6175e0
53bdd1b70c60f5c21791d1c399272735ed3acb2a
22173 F20101122_AABBCA supan_k_Page_019.QC.jpg
f937999fbe11bc8efc1912a57fd4feb4
a9d2f130bd90666f6d02ae2da94a04d6a0fc1efe
20009 F20101122_AABBBL supan_k_Page_009.QC.jpg
6f2d0a6905eb78b6d1cd7ce28740e151
720b38ca8642664be5f6668c726732b22b431a90
1709 F20101122_AABAWF supan_k_Page_004.txt
a4d6dacdd9dfa3e19623c884a4fe0708
5da2aa72715ffc01e4dcdff7b01bd5be6127848c
2431 F20101122_AABBAX supan_k_Page_001thm.jpg
90a222e9a85728a80e9aad44a5087d57
bc9eb86113d956aba6329e8b76e646e007017945
2573 F20101122_AABAVR supan_k_Page_134.pro
50d7c6bed0ca6776c5b50d02216d513e
8bfc8dbca065db1f7dc6a210d6ec6c2f61c1e87d
6214 F20101122_AABBCB supan_k_Page_019thm.jpg
d460842b447d1749425f27085b23482a
7e59e64c8bedbff1dc4bfc896146476cbdf73d98
19621 F20101122_AABBBM supan_k_Page_010.QC.jpg
b71ce61a7e4d8b14848f630043e41cc8
d7c10a0f730b0f56f17f3b92568cdd66f4748f90
2144867 F20101122_AABBAY supan_k.pdf
200f5af183297d32ed0529ce0b11f067
bd4601fe5fa91854a1be392cc3827914aad8f206
10499 F20101122_AABAVS supan_k_Page_135.pro
07fd78f7b3fdbd2ea73d972bff7afadd
45588774fee97b1e6b6774e5039cfe21e0769814
132 F20101122_AABAWG supan_k_Page_005.txt
b7d33ec7c79d1e9711566366ef4f4d65
7871edb1fee27b6dd49c6f1656f1474bc947d4eb
23991 F20101122_AABBCC supan_k_Page_020.QC.jpg
a330c567b02ffbdec98c1cc87cbe03c6
18eb738216786aed4f4f80b71d0f88f5a422ff0c
4942 F20101122_AABBBN supan_k_Page_010thm.jpg
d71e1f334b0598f78df97142d64db628
77e301add733cd1d3cf76153107780eb909bb949
1378 F20101122_AABBAZ supan_k_Page_002thm.jpg
c84cdd109f70734cc843c67a75e633b2
eb3bb315e8422740001c0a24ed1296c3efbe1fef
32133 F20101122_AABAVT supan_k_Page_136.pro
95248babec99ad5e0013326795860b37
f91fd84b4d29aee296b723a2745ee470e067c665
3153 F20101122_AABAWH supan_k_Page_006.txt
f045f91be34b6c213fa1c9017f1cbf65
4807e83443fab76c10230cb12030d8cbe491323a
23004 F20101122_AABBCD supan_k_Page_021.QC.jpg
6b369cd66e79af7227ed16265457fd6b
b3fa7e92431af529d586fc32a024e2b0035adfcc
26167 F20101122_AABBBO supan_k_Page_011.QC.jpg
f1449798316412189226f46ccaddb657
aae84e6e0f6deffb06594c0c2ad62f4042b54dad
28060 F20101122_AABAVU supan_k_Page_137.pro
037b6753a723396e0f685845fe119063
148065b815a14d7347e359e8d21e01c36f1991b4
255 F20101122_AABAWI supan_k_Page_008.txt
3365efc704aecfe4ece59f98b84e7ade
a0fd37e51fbccfb3e1661cbc68b155b0da6f851d
6370 F20101122_AABBCE supan_k_Page_021thm.jpg
3e4aa749b7edc4756705b577a0702460
8e83f4a2c7e346ee8b565cc82f01206fc2328b22
7015 F20101122_AABBBP supan_k_Page_011thm.jpg
ea50fe995a24402c12d75745b7b58083
26d479c0faaba10b0a918d0fd1d104c01756c9be
30066 F20101122_AABAVV supan_k_Page_138.pro
44e13959f5a0de01a740ecc39eacedf4
e4eb0ee71afa39ed16e12612d8bef75f61ff758e
2302 F20101122_AABAWJ supan_k_Page_009.txt
06327eb2d57de600d537e644137f163a
aa0805f117eb6895949ec22e5f6dc79586a85588
6491 F20101122_AABBCF supan_k_Page_022thm.jpg
c3d73952e754110c46262da2f2d07063
4cb5aaf3bc23a263d0f88b3ceb519044b7951ef0
18210 F20101122_AABBBQ supan_k_Page_012.QC.jpg
b66734ea0b70b314e743f65554be6932
bac1b8be9c0befcbf3d956c746e062975d459b42
43383 F20101122_AABAVW supan_k_Page_139.pro
c0c842e6c4e0c1364ef9fdc0970875f9
4b86bfea833317c688c10b22cd2e6aa28092da5d
2329 F20101122_AABAWK supan_k_Page_010.txt
107eed71f5d72791ad8b85d04fb385b2
4e12002c1b6082b5d990f4e9ecd444b441ce6345
4904 F20101122_AABBBR supan_k_Page_012thm.jpg
7f4b7cdfaa8af1405bf854def03fd1d7
fba21c15109d34c387e5085fe5e65500c463d22f
59651 F20101122_AABAVX supan_k_Page_140.pro
95636ba51f99488ee9b3b4dd61d18617
b8ac14ba579224341775dbff605601eff8736a71
3018 F20101122_AABAWL supan_k_Page_011.txt
deceeff2550ced62dec5d46a6ebcd4cd
64cac1fbe40d21e8b4c18d74754e5405fb25c86a
21985 F20101122_AABBCG supan_k_Page_023.QC.jpg
29c1da4754e5c57a623e3cb52b917030
fc06f1f7aab1f4d1edca4a9238c10ae0c4139830
20044 F20101122_AABBBS supan_k_Page_013.QC.jpg
451ed9f8b71e4fe9a1594830edb17948
f8a5af1c927e420cf30c2f16701b90882d606c1f
56430 F20101122_AABAVY supan_k_Page_141.pro
99ca51bd7a4af6c420e9a305223d5a42
41bd52c9220346752f99036cc374b9832a581eb9
1918 F20101122_AABAXA supan_k_Page_030.txt
3c403efa08f47d6c8b60a2b719a26335
618245323b8920f9f6427258a92e3988f09793c5
F20101122_AABAWM supan_k_Page_012.txt
b42d3197a7e3203a24dd5d74298a3925
cf04896320910fdf2258e22974d196ed0e5705bc
6220 F20101122_AABBCH supan_k_Page_023thm.jpg
cf64d3bbd94730f8c95a9bf87e30ea31
83960d41a4953e0913273291ef4cfb4a527cb7c5
5418 F20101122_AABBBT supan_k_Page_013thm.jpg
201c47f8f7aaa14fdbc2934c4319902f
879daf55b69d69ceb896104573ab465b165878b1
63421 F20101122_AABAVZ supan_k_Page_142.pro
a7e7c4baf15456d2b322ed2f077eb7e9
88c1cf2978faa068c1eda2e440e67937714ad35b
1860 F20101122_AABAXB supan_k_Page_031.txt
f323052b9fd109cc52a968e3d1a94530
9dc1dd33439fbee59f88ef64b4e187e12a134f49
1689 F20101122_AABAWN supan_k_Page_013.txt
9ba3672621e63fd1aba543db78476335
df85df5a318cad1fc33ae6b436245a7966e85ee2
21622 F20101122_AABBCI supan_k_Page_024.QC.jpg
4f73fa904002145f5c635cf736019342
bd1d46ff8155923d1a34e9ed2be66bcd348c8eb3
5510 F20101122_AABBBU supan_k_Page_014thm.jpg
3f1bff4b89ef919f64f53956c310651c
bc2f35e110a461987ef258efcf86cf37e10e45f6
1826 F20101122_AABAXC supan_k_Page_032.txt
731bd2ad32d80bdb3ace0f27909f7dec
f896d2c40a360b05cf5e952440bed2b4720ccf76
1621 F20101122_AABAWO supan_k_Page_014.txt
82b10551f9e61470c319e993478eb371
f6985284417ee3acca2b15d18eaade8322d2ee2b
5862 F20101122_AABBCJ supan_k_Page_024thm.jpg
a22a0da17766d3e917b8b81b848b3f03
7b8277c1eb63a43b02c1d5e7432d468f176f8916
20414 F20101122_AABBBV supan_k_Page_015.QC.jpg
64a9ad80e966629fd4c13e8460c52939
b3805929795e76f99f119a60803322b1de03070b
1548 F20101122_AABAXD supan_k_Page_033.txt
103f442b8a7e3c39653fbbbfef27b9d0
dfef7fece7343f50f435b0ef20ce2c9a6fb93e4e
1959 F20101122_AABAWP supan_k_Page_016.txt
5d9d2ea2ea73111675684803e3a2b8fd
8b33e44b7f8a0bbeb8da9342d0063bed43fdff65
20199 F20101122_AABBCK supan_k_Page_025.QC.jpg
4b1b3e9a84b80efb65043656493fd6f7
a0e5d1c03c2330218ed1bce915c2897c6250b998
F20101122_AABBBW supan_k_Page_015thm.jpg
1535cd2813679a71211ac9d46cf6514a
a8b93e0ad013d62181c0a81a1874f41f52eae048
1668 F20101122_AABAXE supan_k_Page_034.txt
d0d91eb62750dd9c15f03bfa6e3dd786
14d37940f7b414fe0570894bf1577b3e8b39fe5c
1899 F20101122_AABAWQ supan_k_Page_017.txt
0ee040f6649c3ac799b08a2ec5207f67
b8d7cd48bdc54840267216913fa43a6b618b9379



PAGE 1

MODEL FOR SUSPENDED GATE FIELD EFFECT TRANSISTORS USED IN LABORATORY ANIMAL CAGE MONITORING By KAREN E. SUPAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORI DA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Karen E. Supan

PAGE 3

This document is dedicated to my parents, Mary Jo and Fred Timm.

PAGE 4

iv ACKNOWLEDGMENTS First, I would like to thank my advisor, Dr. Herbert Ingley, for his constant support and encouragement of this work. From our fi rst meeting, he has been a great mentor to me in engineering, teaching, and life in genera l. I thank him for the countless words of wisdom he imparted to me. I acknowledge my appreciation to Dr. David Hahn for the generous amount of time he devoted to this project by providing t echnical assistance on the experimental portion of this project, offe ring his expertise in a new area to me and for serving on my committee. I also express gratit ude to Dr. Sherif Sherif, Dr. Bill Lear, and Dr. Jason Weaver for being committee members and providing guidance in their area of proficiency. I am grateful to Osman Ahmed of Si emens Building Technology for the inception and financial support for this project. Danke schn to Roland Pohle, Peter Gulden, and many others who made the trek across the Atlantic from Siemens Core Technology in Munich, Germany, to collaborate on this proj ect. Special thanks go to Dr. Gus Battles and Mike Cormier from Animal Care Services for opening their facilities for this project and answering many questions along the way. It is impossible to forget all the friends and colleagues I have met at UF who have made the time here so enjoyable. Above all, I express thanks to my parent s and family for creating an environment where education was a high priority and paving the way for me to follow in their tracks. My deepest gratitude goes to my husband, Br ian, for providing encouragement to pursue

PAGE 5

v this degree. I will be forever indebted to him for the distances he went to help me achieve this accomplishment.

PAGE 6

vi TABLE OF CONTENTS Page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES.............................................................................................................x ABSTRACT.....................................................................................................................xiii CHAPTER 1 BACKGROUND..........................................................................................................1 Comparison of Static Isolator Ca ges to Ventilated Caging Systems............................4 The Macroenvironment................................................................................................7 The Microenvironment.................................................................................................9 Ammonia.............................................................................................................10 Carbon Dioxide...................................................................................................14 Other Contaminants.............................................................................................16 Contact Bedding..................................................................................................17 Relative Humidity...............................................................................................19 Ventilation...........................................................................................................20 Additional Environmental Factors......................................................................21 Previous Environmental Studies.................................................................................22 Cost Analysis for Current Husbandry Practices.........................................................29 Air Sampling Techniques...........................................................................................30 Semiconductors...................................................................................................31 Field Effect Transistors.......................................................................................33 Gas Sensing.........................................................................................................35 Summary.....................................................................................................................40 2 EXPERIMENTAL FACILITIES AND METHODS.................................................42 Experimental Setup.....................................................................................................42 Experimental Procedures............................................................................................48 Carbon Dioxide Sensor........................................................................................48 Ammonia Sensor.................................................................................................49

PAGE 7

vii 3 THEORETICAL MODELING..................................................................................54 Gibbs Free Energy......................................................................................................54 Adsorption..................................................................................................................58 Surface Reaction Rate Expressions............................................................................60 Langmuir Adsorption Isotherm...........................................................................60 Dissociative Adsorption......................................................................................62 Competitive Adsorption......................................................................................63 Proposed Mechanisms................................................................................................64 Ammonia and Hydroxide....................................................................................64 Ammonia Dissociation........................................................................................66 Molecular Adsorption..........................................................................................72 Reaction Kinetics........................................................................................................74 Summary.....................................................................................................................77 4 RESULTS AND DISCUSSION.................................................................................78 Drift Tests...................................................................................................................78 Ammonia Sensor Results............................................................................................79 Sensor Response and Mechanism.......................................................................80 Diffusion..............................................................................................................87 Temperature Effects............................................................................................89 Sensor Performance.............................................................................................91 Gradual ramping tests...................................................................................94 Cross-sensitivity to humidity and carbon dioxide........................................96 Carbon Dioxide Sensor Results..................................................................................99 Summary...................................................................................................................104 5 SUMMARY AND CONCLUSIONS.......................................................................106 Summary of Results..................................................................................................106 Ammonia Sensor...............................................................................................106 Sensor response and mechanism................................................................106 Performance and feasibility........................................................................107 Carbon Dioxide Sensor......................................................................................108 Recommendations.....................................................................................................109 Future Work..............................................................................................................110 APPENDIX A ENVIRONMENTAL STUDIES..............................................................................112 B DETAILS OF FIELD EFFECT TRANSISTOR......................................................116 C ANALYSIS OF EXPERIMENTAL UNCERTAINTY...........................................122

PAGE 8

viii LIST OF REFERENCES.................................................................................................125 BIOGRAPHICAL SKETCH...........................................................................................131

PAGE 9

ix LIST OF TABLES Table page 1-1. Recommended space for lab mice.............................................................................22 1-2. Ammonia concentration levels after seven days in static isolator cages...................23 1-3. Gas sampling pumps used in environmental studies.................................................30 2-1. Sensor testing equipment...........................................................................................43 2-2. Ammonia and carbon dioxide concentrations...........................................................47 2-3. Ammonia concentrations used for ramping tests......................................................50 2-4. Experimental parameters for time response tests......................................................52 3-1. Infrared (IR) and XPS measurements and assignments for adsorbed species of ammonia on surface catalysts...................................................................................66 4-1. Rates of baseline signal drift for the carbon dioxide and ammonia sensor...............79 4-2. Curve fit coefficients, R-squared valu es, and rate constants for desorption and adsorption of 100 ppm ammonia on sensor.............................................................83 4-3. Adsorption and desorption time constants for the ammonia sensor using forced and diffusion flow regimes.............................................................................................90 4-4. Curve fit coefficients, R-squared valu es, and rate constants for desorption and adsorption of ammonia on sensor at varying surface temperatures.........................91 4-5. Slopes for ramp down and ramp up curves completed in diffusion box...................96 4-6. Magnitude of signal response for cha nges in ammonia concentration and relative humidity...................................................................................................................98 4-7. Adsorption time constants.......................................................................................104 4-8. Desorption time constants.......................................................................................104

PAGE 10

x LIST OF FIGURES Figure page 1-1. A static microisolator cage...........................................................................................4 1-2. Equipment used to ensure a healthy mi croenvironment in an animal laboratory........6 1-3. When excited, electrons move from th e valence band to the conduction band across the energy gap..........................................................................................................32 1-4. Schematic of field effect transistor.............................................................................34 1-5. Classical FET configuration.......................................................................................35 1-6. Suspended Gate FET configuration............................................................................35 1-7. Photograph of a hybrid flip chip FET sensor device..................................................36 2-1. Schematic of experimental facilities...........................................................................42 2-2. Humidification section of experimental facilities......................................................44 2-3. Mixing section of e xperimental facilities...................................................................44 2-4. Two flow regimes used in sensor testing....................................................................45 2-5. Carbon dioxide sensor................................................................................................46 2-6. Signal from DC power supply....................................................................................48 2-7. Graphical depiction of experimental pa rameters used to test the carbon dioxide sensor for cross-sensitivity to humidity...................................................................49 2-8. Graphical depiction of ammonia con centrations used in ramping tests.....................50 2-9. Graphical depiction of expe rimental parameters used to test the ammonia sensor for cross-sensitivity to humidity while ammonia was present.......................................51 2-10. Graphical depiction of expe rimental parameters used to test the ammonia sensor for cross-sensitivity to humidity and carbon dioxide.....................................................52

PAGE 11

xi 3-1. Diffuse Reflectance Infrared Fourier Tr ansform Spectra (DRIFT-spectra) for a TiN screen-printed film...................................................................................................65 3-2. Mechanism for the reduction of NO by NH3 over a V2O5 sensing layer in the presence of oxygen...................................................................................................65 3-3. The XPS N(1 s ) core-level spectra for ammonia.........................................................67 3-4. The XPS N(1 s ) core-level spectra for ammonia on (a) Si(100)-(2x1) and (b) Si(111)(7x7).........................................................................................................................68 3-5. Surface species and desorption products from ammonia on Si(100).........................69 3-6. Ball and stick models for the adsorption geometry for -NH2 and H on two different surfaces.....................................................................................................................70 3-7. The XPS N 1s spectra of Ti/Si (100) surface.............................................................71 4-1. Long-term drift test results.........................................................................................79 4-2. Average desorption data and curv e fit for 100 ppm desorption tests.........................82 4-3. Average adsorption data and curv e fit for 100 ppm adsorption tests.........................83 4-4. Actual 50 ppm average ( N =5) desorption curve plotted along with the predicted curve.........................................................................................................................84 4-5. Actual 50 ppm average ( N= 6) adsorption curve plotte d along with the predicted curve.........................................................................................................................85 4-6. The slope of the Gibbs energy changes as the reaction proceeds...............................87 4-7. Diffusion of ammonia in a semi-infinite region.........................................................88 4-8. Average ( N= 6) desorption curve for the diffusi on case compared with the forced curve fit for 100 ppm data........................................................................................89 4-9. Average ( N= 6) adsorption curve for the diffusion case compared with the forced curve fit for 100 ppm data........................................................................................89 4-10. Average desorption and adsorption curv es at 50 ppm ammonia for heater voltages of (a) (b) 2 V ( N= 10) and (c) (d) 3 V ( N= 7 and N= 8).......................................91 4-11. Single analyte test results ( N= 1) for the ammonia sensor........................................93 4-12. Calibration curve and linear fit for ammonia sensor................................................93 4-13. Ammonia concentration ramped up and down with ammonia sensor in diffusion box............................................................................................................................95

PAGE 12

xii 4-14. Ammonia sensor tested for cross-sensitivity to humidity........................................96 4-15. Ammonia sensor tested for cross-sens itivity to humidity w ith ammonia in the system.......................................................................................................................97 4-16. Average (N = 6) ammonia sensor re sponse to humidity, carbon dioxide, and ammonia concurrently............................................................................................100 4-17. Raw data ( N= 1) from carbon dioxide single analyte test.......................................101 4-18. Calibration curve for carbon dioxide sensor...........................................................101 4-19. The average ( N =4) response time and respective cu rve fits of the carbon dioxide sensor......................................................................................................................103 4-20. Carbon dioxide sensor tested for cross-sensitivity to humidity..............................105 B-1. Schematic section through a hybrid flip chip field effect transistor........................116 B-2. Schematic cross-section of a susp ended gate field effect transistor........................117 B-3. Scanning electron microscope cross-sect ion of a hybrid flip chip field effect transistor.................................................................................................................117 B-4. Printed circuit board for the suspe nded gate field eff ect transistor.........................118 B-5. Schematic for one channel of the suspended gate field effect tran sistor control board, where the drain to source vo ltage is indicated by UDS.........................................119 B-6. Detailed drawing of the suspended gate field effect transist or control board..........120 B-7. Simple drawing of the suspended gate field effect transi stor control board............121

PAGE 13

xiii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MODEL FOR SUSPENDED GATE FIELD EFFECT TRANSISTORS USED IN LABORATORY ANIMAL CAGE MONITORING By Karen E. Supan December 2005 Chair: H.A. Ingley III Major Department: Mechanic al and Aerospace Engineering Over the past century, great advances in medicine have been achieved through the use of laboratory animals, specifically rodents. The quality of the animal environment is important to the rodents hea lth and welfare, and their we ll-being directly affects the quality of research involving th eir use. There can be signifi cant variability in air quality between cages depending on a number of factors such as population si ze and air flow. A way to accommodate for the variability betw een cages is to monitor environmental quality indicators within th e cage, such as ammonia, ca rbon dioxide, temperature, and relative humidity. Since rodent cages ar e approximately the size of a shoebox, commercially available sensors would be too la rge for this application. Therefore, microsensors, or field effect transistors, were in vestigated for application in a rodent cage. Since these sensors were on the forefront of technology, a theoretical model was developed for the ammonia sensor to furt her understand the chemical reaction taking place on its surface.

PAGE 14

xiv The sensors were tested in a controlled environment, where the air quality was known. The magnitude and time of the response to different levels of contaminants (e.g., ammonia and carbon dioxide) were determine d. The study showed th at the sensors can detect changes in air quality in a sufficiently short amount of time (5 minutes) so that corrective action could be taken to preven t the rodents from overexposure to harmful levels of air contaminants. At the present development stage, the sensors used for this investigation will require further improveme nts before implementation in a laboratory animal cage. These improvements include but are not limited to eliminating drift of baseline signal, increasing sensitivity of sensor, amplifying signal output, and coupling each gas sensor with a humidity sensor. The reaction mechanism selected for th e model which was best supported by the literature and the experiments was molecular adsorption of ammonia on a titanium nitride surface. The experimental results were fitt ed to the model to obtain the adsorption and desorption rate constants, the equilibrium c oncentration constant, equilibrium constant, and Gibbs free energy, which respectively were 6.28 L/mol*s 6.43 x 10-3 s-1, 976.7 L/mol, 39.04, and -9.25 kJ/mol. Based on th ese values, it was determined that the forward reaction, or adsorption, occurs spontaneously. There was good correlation between the theoretical model a nd the experimental results, i ndicating that the theoretical model was sufficient for this application.

PAGE 15

1 CHAPTER 1 BACKGROUND Nearly every medical breakthrough in the last century has come as a result of research with animals (1). In the nineteenth century, animals contributed to the treatments for rabies, smallpox, and anthra x. The early 1900s saw breakthroughs in cardiac catheterization t echniques, treatment for rickets, a nd the discovery of insulin. In the 1930s research with dogs contributed to th e development of modern anesthesia, while horses aided in the prevention of tetanus. The 1940s saw treatment of rheumatoid arthritis and whooping cough, prev ention of diphtheria, and de velopment of antibiotics. In the 1950s a cure for polio was found through the use of rabbits, monkeys, and rodents. That decade also saw the development of ope n-heart surgery, cancer chemotherapy, and tranquilizers. In the past 50 years, animals have helped find treatments for diseases such as rubella, measles, and Hansens disease. They have also furthe r advanced research on organ transplants, breast cancer, cystic fibrosis, multiple sclerosis, and Lou Gehrigs disease. While much of this research was done with large animals, such as dogs, monkeys, or sheep, todays medical rese arch mainly utilizes rodents. Mice have been utilized in a wide array of medical research. For example, mice have been used in cancer research since 1894 (2). In 1921 inbred strains, which were susceptible to tumors, were created. More strains were developed in 1929 with the founding of one of the first an imal laboratories, Jackson La boratories. In 1962, a mutant mouse with low immunity was discovered which led to human tumor transplantations.

PAGE 16

2 The late 1980s saw a boom in mice research with the development of a transgenic mouse whose genes were altered to produce a desire d characteristic. Fr om this research, oncogenes, a gene that can cause a normal cell to become cancerous, could then be studied. Genetically engineered mice are also used today to determine Vitamin Cs role in health and illness (3). Today, research is ongoing in such diverse areas as diabetes, hearing loss, ovarian cancer, and glaucoma (4). Mice are valuable for medical research b ecause of their genetic similarity to humans (1). Laboratory mice, Mus domesticus belong to the family Muridae and are a domesticated variant of the house mouse, Mus musculus (5) Adult mice are adapted to live in groups and generally li ve no longer than two years. They typically weigh 20-40 grams (0.7-1.4 ounces), have a length of 12-15 centimeters (5-6 inches), and when standing on hind legs achi eve a height of 10-12 centimeters (4-5 inches). The quality of the environment is imperative to the rodents health and welfare, and their well-being directly a ffects the quality of research collected from them (5). Rodent cages have evolved over the years to better suit the needs of the animals. Lisbeth Kraft, in the late 1950s, was the firs t to separate mice into isolator cages to prevent the spread of communicable diseas es through direct contact, specifically epidemic diarrhea of infant mice (6). Filter-top covers were added to Krafts isolator system to prevent the transmission of airborne diseases as well as to elimin ate the exchange of feces, soiled bedding, and hair between adjacent cages (7). Robert Se dlacek developed the modern filter top while working with a large gnotobiotic mouse col ony used in radiation biology research at Massachusetts General Hospital. The filter top was a polycarbonate frame fitted with a

PAGE 17

3 polyester filter medium held in place by a perf orated aluminum plate. The rim at the bottom of the filter top, where it fit over the underlying cage, formed a lip. One of the first static isolator cages consisted of a cy lindrical cage with a solid galvanized bottom and a tight-fitting lid with metal mesh sides wrapped with fiberglass insulation that filtered the incoming air. The contemporary static isolator caging system has given institutions the ability to keep rodents clean while hous ing contaminated and clean rodents simultaneously (6). Further improvements were made to the st atic isolator cages with the onset of ventilated caging systems (VCS). These sy stems were developed to improve intracage ventilation and to increase housing capacity. The first ventilated cages were developed under the direction of Dr. Ed Les at Jack son Laboratories in 1960, around the same time as the filter-top cages. The earliest version deposited mouse odors and allergens out into the room. Jackson Laboratories and Thor en Caging Systems collaborated to further improve upon this design in the late 1970s by filtering the incoming and exhaust air. The first systems were commercially available in the early 1980s. By the early 1990s VCS had gained widespread popularity (6). Meanwhile, other companies such as Hazelton Systems were developing a ventilated rack system to help reduce allergies of staff members (8). In Hazeltons system, air was blown into each shelf row of cages or each individual cage at a low velocity and removed by a main exhaust system. In a modern VCS, high efficiency particulate arresting (HEPA) filtered air is blown into cages through a manifold under positive-pre ssure. Air is either exhausted directly from the cage or filtered before sending it to the room or venting outside (9).

PAGE 18

4 In addition to the VCS, HEPA filtered flow work areas were developed for use when changing the bedding or restocking food a nd water in a cage. The work area helps maintain the microbial barrier created by the stat ic isolator cage and VCS. Ambient air is drawn into the station, moved through a pre-fi lter and HEPA filter, and is then blown horizontally or vertically across the workstation. Comparison of Static Isolator Cages to Ventilated Caging Systems Static microisolator cages are still used in many animal laboratories. Isolator cages are cost effective and allow containment at the cage level without expensive ventilation (10). They are a proven technology for protecting valu able mice from microbial contamination (5). Isolator caging systems provide a separate microenvironment and aid in the development and upkeep of disease-free rodents for use in rese arch. Static cages are useful for studies where containment at the cage level is desirable, for example, in vivo administration of hazardous agents (6). Photographs of static microisolator cages are depicted in Figure 1-1. (a) (b) Figure 1-1 A static microisolator cage (a) side view and (b) top view. While static cages may be cost effectiv e, there are drawbacks that must be considered. The use of these cages can be labor intensive depending on how often the

PAGE 19

5 bedding is changed. Changing and handling of the cages is physically intensive potentially leading to back or hand injuri es. Due to frequent cage changing a large supply of bedding is required. The cages mu st be washed frequently which uses a significant amount of water and electricity a nd can accelerate the degr adation of the cage (9). The static cages in use today are ad vantageous because they are durable, transparent, and have a replaceable filter t op and tight-fitting lid, which is not easy to dislodge (11). When the filter top was firs t introduced, the advantage of containment outweighed the effects on the ai r quality inside the cage. Res earch soon revealed that the filter top was a barrier to air and moisture exchange increasing the intracage relative humidity in one study by 38% compared to the macroenvironmental humidity (6). The filter top impedes air exchange between the micro and macro environments. The only ventilation in static cages comes from the rodents breathing patterns (8). Individually ventilated caging systems (IVCS) have addressed the problem of little to no air exchange between the micro and macro environments. The IVCS combines the static cages with individu al ventilation. The whole p ackage, as depicted in Figure 1-2, should include a microisolation cage, ventilat ed cage rack, and a Class 100 or Class II change cabinet, which helps rodents remain disease-free in a healthy microenvironment (5). Microenvironmental air quality is better and the variabil ity in air quality between cages is reduced when using ventilated cages Intracage ammonia, carbon dioxide, and relative humidity are also lower. In additi on, the day on which ammonia is first detected can be delayed (6, 12, 13). One study revealed that di rect ventilation, 23 air changes per

PAGE 20

6 hour (ACPH) to each cage, improved the mi croenvironmental conditions in comparison with a static isolator cage. The relativ e humidity, ammonia, and carbon dioxide in the ventilated cages were 8% lower, 240 parts per million (ppm) lower, and 2900 ppm lower, respectively (11). Ventilated systems also help prevent mouse urinary protein (MUP) from spreading. A study showed that less than 0.05 ng/m3 (5x10-7 ppb) was detected in a room with IVCS, whereas a high level, 4.6 ng/m3 (4.6x10-6 ppb), of MUP was measured when mice were housed in open cages (14). (a) (b) Figure 1-2 Equipment used to ensure a healt hy microenvironment in an animal laboratory should include (a) a cage changing stati on and (b) a ventilated cage rack. Operational savings can also be achiev ed with IVCS. With the improved air quality, cage changing can be delayed to weekly or longer, which translates into labor savings. The time spent sanitizing cages and the quantity of bedding used are reduced as well. Due to the decreased cage changing in tervals the service life of the cages and caging systems is increased. The stocking de nsity per room can be increased, thereby

PAGE 21

7 allowing more efficient use of space in a laboratory animal facility (6). Overall, an improved microenvironment lead s to lower operational costs. Despite the improvements in micro and macro environmenta l air quality and savings in operational costs, IVCS are not used in every laboratory animal facility. To begin with, IVCS are expensive to acquire. Despite the advantages of IVCS, a new system is decidedly capital intensive (6). Other costs accrue from electricity for operating the system, maintenance, and replaceme nt of filters. The IVCS are also more difficult to sanitize than a standard rack hol ding isolator cages. Bl owers, shelves, and access panels must be removed before placi ng the IVCS in a rack washer. Extensive washing by hand is required and access to a ll plenums and ducts may not be feasible. The heat gain in the housing room may be increased due to the supply and exhaust blowers. If stocking density is increased the heat gain due to the mice will rise causing an increase in the total cooling load. Nois e generation from the IVCS exhaust and supply blowers may be an issue for employees working in the room and mice in the microenvironment. Finally, excess intrac age ventilation can cause chilling and dehydration, especially with neonates and hairless mutants. Whichever system a laboratory animal facility chooses to use, th e most important issue is the health and welfare of the animals and laboratory personnel. The Macroenvironment To better meet the needs of the animals and personnel, the environment in which they work or reside must be examined. The overall facility or bu ilding is known as the megaenvironment, the items in an animals room are considered the macroenvironment, and the items in an animals cage or im mediate surroundings are the microenvironment (15). Each environment can be treated sepa rately, but the couplings between each system

PAGE 22

8 must also be considered. For example, a high-level exhaust system can improve thermal ventilation efficiency for the mega and m acro environments. This high-level exhaust system may not provide enough circulation to properly vent ilate the individual static isolator cages and thereby does nothing to decrease the ammonia and carbon dioxide concentrations in the microenvironment (6). Therefore improvements for the macroenvironment will not necessarily improve the microenvironment. One of the main reasons laboratory anim al facilities are concerned with the condition of the macroenvironment is for labor atory animal personnel. According to the National Institute for Occupational Safety a nd Health (NIOSH), 33% of animal handlers have allergic symptoms and 10% have symptoms of animal-induced asthma (5). Laboratory workers can have laboratory an imal allergies (LAA) to prealbumin and albumin, which are derived from mouse urine and skin (16). Aeroallergens from mice are highest during cage changing, while handli ng male mice on an unventilated table, and while dumping the bedding from the cage without a dumping station. A study by Sagakuchi et al. (16) revealed that using female mice, filter-top covers, and corncob bedding could reduce LAA. Preal bumin and albumin levels were reduced by 90% and 40%, respectively, when using female mice versus male mice. Using a filtertop cover rather than no cover reduced pr ealbumin by 90% and albumin by 60%. When wood shavings were replaced by corncob bedding prealbumin dropped 57% and albumin by 77%. Other studies have shown that usi ng IVCS and carrying out animal husbandry and research procedures in ventilated cabinets can reduce exposure to aeroallergens (5). A healthy macroenvironment is impera tive for the overall facility. The environmental conditions in the cage and r oom directly affect how an animal will

PAGE 23

9 respond to laboratory procedures (17). Apposite housing and ma nagement are critical to animal welfare, the quality of research data and teaching or testing programs where animals are used, and the health and safety of employees (18). The Microenvironment The microenvironment should meet the various needs of the mice. First, the primary enclosure should allow for normal phys iological and behavioral needs of the animals. These include allowing for mainte nance of proper body temperature, urination, defecation, normal movement, and postural ad justments. The cage should be large enough so that the mouse can turn around a nd make typical movements. Second, the primary enclosure should allow for social inte raction and hierarchical development. In addition, the cage should provide a clean, dry, safe area with adequate food, water, and ventilation. Lastly, the pers onnel should be able to view the animal with minimal disturbance (15). Taking into account all of thes e measures can help ensure a healthy microenvironment. The microenvironments condition has b een observed scrupulously due to the advent of the filter-top. Befo re the filter-top, observations of the macroenvironment were sufficient to maintain the health and well-being of the animals. In todays animal facility it is inadequate to supply the macroenvir onment with 15 ACPH, keep the temperature and relative humidity at the recommended le vels, and then completely disregard the conditions in the microenvironment (11). This is because air exchange occurs at the junction of the lid with the cage and not th rough the filter medium. A study by Keller et al. showed that the lid redu ced the air exchange rates wi thin the cages to one ACPH while the room was ventilated at 12 ACPH (19).

PAGE 24

10 With such an airtight environment the cag e can easily become contaminated. There are four major environmental quality indicators within the cage. The first is ammonia, which is produced from the urea found in the animals excrement. Second is carbon dioxide, which is generated as a metabolic wa ste product. Third is moisture, indicated by relative humidity, from respiration, excremen t, and the drinking water for the mice. Additionally, thermal loads from metabolic activity can contribute to a rise in cage temperature (11). In general, ammonia, ca rbon dioxide, relative humidity, and temperature are used to assess th e microenvironmental conditions. The level of air quality is dependent on a number of factors, which include but are not limited to population size, strain and stock of animal, location of a cage on the rack, type of filter, airflow within the room, and relative humidity (12). Lipman identifies four ways to address poor microenvironmental air quality (6): 1. Change cages at sufficient frequencies. 2. Use contact bedding with desirabl e performance characteristics. 3. Reduce the macroenvironmental relative humidity. 4. Increase the macroenvironmental temper ature (dry-bulb) without altering the moisture content in the air. Choi et al. determined that restricting the number of animals per cage, regularly changing soiled bedding, and increasing ventil ation were techniques to prevent relative humidity and concentrations of carbon dioxi de and ammonia from increasing in static cages (12). Ammonia The build-up of intracage ammonia concen trations is the primary reason for ventilating a microisolation cage (5). Above a certain moisture threshold the ureasepositive bacteria grows, lead ing to ammonia production (15). The moisture threshold is

PAGE 25

11 dependent on the type of bedding used, for example, pine shavings, recycled pulp, or corncob bedding. Enteric bacteria produce ammonia through two possible mechanisms of enzymatic activity. In the first mechanism, bacterial ur ease acts as a catalyst for the hydrolysis of urea to ammonia and carbamate: Carbamate Ammonia UreaUrease (1.1) For the second mechanism, Dand L-am ino acid oxidases remove the amino group from amino acids to form keto acid and ammonia (12). Ammonia levels. The human ammonia threshold limit value (TLV) used by NIOSH of 25 ppm was established through th e work of Gamble and Clough (20). A TLV is the concentration to which humans can be exposed to for 8 hours a day 5 days a week without any harmful effects (21). The Occupational Health and Safety Administration (OSHA) sta ndards for ammonia are 50-ppm time-weighted average (TWA) and 35-ppm short-term exposure. The human TLV is accepted for animals, although for mice the concentration capable of reducing respirat ory rate by 50% (RD50) is approximately 300 ppm (22). Effects of ammonia. Ammonia acts as an irritant and can alter or destroy the tracheal epithelium. More specifically, when the epithelium becomes irritated, the cilia are paralyzed, mucus flow is altered, and th e surface layers of the epithelial lining are destroyed (20). The epithelial thickness also changes, which increases the airflow and deposition of airborne pa rticles to that area (21). The abnormal increase in the tracheal epithelium is dependent on the amount of ammonia and how long the animals are exposed to it.

PAGE 26

12 Gamble and Clough documented the effect s of ammonia on the rat tracheal epithelium (20). After four days of exposure at 200 50 ppm, there was a transitionalstratified appearance to the epithelium and irregularities were noticeable. Some gross changes were noted after eight days; the cilia disappeared, stratification increased, folds formed on the surface of the lumen, and the am ount of mucus increased. At twelve days of exposure, the epithelial thickness increa sed and there was an acute inflammatory reaction with increased cellularity and alteration of cell types. Few studies showing long-term effects on mice exposed to ammonia exist and the amount, which causes harmful effects, seems to vary for different studies done on rats. For example, rats exposed to ammonia at 180 ppm for 90 days did not show any problems, while in other studies a level greater than 25 ppm pr omoted growth of infective agents in the respiratory tract (5). Coon et al. completed rat inhalation st udies on ammonia for both repeated and continuous exposure (23). Repeated expos ure (30 repeated exposures, 8 hours/day, 5 days/week) to 0.155 ppm of ammonia produced no adverse effects. At 0.770 ppm of repeated exposure there were nonspecific infl ammatory changes in the lungs of rats. Under continuous 90-day exposure, 0.040 ppm led again to nonspecific inflammatory changes in the lungs. At 0.127 and 0.262 ppm the same changes were seen in the kidneys and lungs. At 0.455 ppm, 32 of 51 rats died by day 25 of exposure and 50 by day 65, when the experiment was terminated. The ra ts showed mild signs of dyspnea and nasal irritation. At 0.479 ppm, 13 of 15 rats died and the following c onditions were found: Focal or diffuse interstitial pneumonitis Calcification of renal tubul ar and bronchial epithelia Proliferation of renal tubular epithelium Myocardial fibrosis and fatty changes of liver plate cells

PAGE 27

13 Ammonia can also be harmful for rats that already have a weakened immune system. For example, Broderson et al. found that ammonia plays an important role in increasing natural murine respiratory mycoplasmosis (MRM) in rats infected with mycoplasma pulmonis (24). Despite the limited number of studies of chronic exposure on mice, ammonia has been monitored in experiments where other para meters were the independent variable. In an early experiment, ammonia concentrations were 400% higher in a cage with a punched lid than one with a wire mesh lid (21). White and Mans (25) found that at low ammonia concentrations in their experiments, ther e was little to no systemic accumulation of environmental ammonia in the animals. Afte r four days of exposure with an ammonia concentration greater than 200 ppm, histological changes in the respiratory tract were visible. Ammonia levels great er than 500 ppm were consider ed lethal. Choi et al. (12) investigated the effect of population size on intracage ammoni a levels and did not detect ammonia in the ventilated cages for the duration of the study, 32 days. According to NIOSH, mice exposed to ammonia for 6 hours/day for 5 days showed signs of nasal lesions including hypertrophy, hyperplasia, epithelial erosion, ulceration, and necrosis (22). Ammonia production. The production of ammonia is affected by a number of factors (21, 25): Number of animals in a cage Frequency of bedding changes Ambient temperature Relative humidity Time of day Type of caging Ventilation rate and air flow

PAGE 28

14 Improving cage-washing procedures and animal room cleanliness can reduce the concentrations of ammonia pr oducing bacteria. Choosing a di fferent strain of mice can also help reduce ammonia concentration, as the formation of ammonia is strain dependent. Using female mice instead of ma les can also lower ammonia levels. One study showed that males produce noticeably more ammonia than females when housed on vermiculite, pulp, and pine shaving beddings (21). Carbon Dioxide Carbon dioxide is a metabolic byproduct of respiration and is generally used as a metric to determine whether there is enough fresh air in a conditioned space. The carbon dioxide concentration in atmo spheric air is roughly 300 to 350 ppm, so comparably low levels are not harmful (26). The activity level of animals, population density, and air exchange rate with the macroenvironment influence formation and accumulation of carbon dioxide (12). Carbon dioxide generation. A 25 g (0.88 oz) resting mouse consumes 1.65 ml (0.06 fluid oz.) of oxygen per gram of bodyw eight per hour and converts 1 ml (0.06 in3) of oxygen to 1 ml of carbon dioxide. Five 30 g (1.05 oz.) mice housed in a filter-topped Type II cage, 350 cm2 (54 in2) and 19 cm (7.5 in) high, generated 250 ml (15.26 in3), approximately 37,000 ppm, if unventilated or undiluted, in one hour (27). In a study by Krohn and Hansen (26), mice housed in a static filter t op cage with a stocking density of 20 g/L stabilized at a carbon dioxide level of 5000 ppm after two hours. They also measured carbon dioxide concentrations in IVCS cages without ventilation. The level reached values between 20,000 and 80,000 ppm within two hours. Carbon dioxide recommended levels. There are currently no official limits for acceptable exposure of rodent s to carbon dioxide. The guideline for humans of 5000

PAGE 29

15 ppm is applied most often to rodents (26). However, since rodent s are adapted to live in tunnels where carbon dioxide levels can re ach 14,000 ppm, higher values have been investigated (5). It is recommended that intrac age carbon dioxide for IVCS and static cages not exceed 5,000 and 30,000 ppm, respectively (5). Lipman states that carbon dioxide levels can be up to 4,000 ppm higher than those observed in the macroenvironment when housing the maximum number of mice (6). It is also advised that if animals are exposed to a level a bove 15,000 ppm, which is significantly higher than the atmosphere, they should be used for experimental purpose with caution and allowed a few days of recovery after exposure (27). Levels less than 30,000 are acceptable for studies involving physiological or biochemical parameters, while when between 30,000-50,000 ppm animals should be give n ample time to recover. Exposures greater than 50,000 ppm should not be accepte d because the impact on the animals may be harmful and irreversible (26). Carbon dioxide effects. The reaction of animals exposed to carbon dioxide mimics a stress reaction, with elevated serum corticosterone levels, increased respiration, reduced numbers of eosinophils and lymphocyt es, and a fight-or-f light reaction along with the release of adrenaline. Humans exposed to less than 10,000 ppm showed only minor effects, which then normalized after 10-15 days. Animals and humans exposed to higher concentrations, 10,000-15,000 ppm, expe rienced the stress reaction described above. Rodents exposed to levels greater th an 30,000 ppm had elevated respiration rates and high levels of circulating corticoste rone, indicating physiological and hormonal changes (27).

PAGE 30

16 Other Contaminants Ammonia and carbon dioxide are only two of many contaminants in laboratory animal cages. Other contaminants found in cages include acetic acid, sulfur dioxide, formaldehyde, dimethylamine, ethanol, ethylene glycol, methane, and hydrogen sulfide. Experiments have found uncharacterized ai r contaminants in isolator cages (6). Acetic acid. Perkins and Lipman in their comparison study of bedding materials, detected acetic acid (mean = 0.86 ppm) in st atic isolator cages with and without mice containing corncob bedding (10). Acetic acid was off-gassed, presumably from the decay of vegetative material relate d with the corncob bedding, ra ther than from bacteria associated with the mice (6). Acetic acid, while not always present in static cages, has a low threshold limit value. The permissible exposure limit was se t by OSHA for acetic acid at 10 ppm for an 8-hour TWA. A 10 ppm 8-hr TWA was also set by NIOSH, as well as a 15 ppm shortterm exposure limit (15 minutes). Exposure can occur through inha lation, ingestion, eye or skin contact, and absorption through the sk in. The vapors cause eye, skin, mucous membrane, and upper respiratory tract irri tation. Mice exposed to 1,000 ppm of acetic acid vapor had eye and upper resp iratory irritation (28). Decreased lung mechanics were observed in guinea pigs exposed to 5 to 500 ppm of airborne acetic acid for one hour (10). Sulfur dioxide. Perkins and Lipman found sulfur dioxide (mean = 0.42 ppm) in static isolator cages, but only in the presence of mice a nd corncob bedding (9). The OSHA standard for sulfur dioxide is 5 ppm averaged over an 8-hour work shift, while NIOSH recommends that the limit be reduced to 0.5 ppm for the TWA for up to a 10hour work shift for a 40-hour week (29). Su lfur dioxide can affect the body if it is

PAGE 31

17 inhaled or comes in contact with the eyes or skin. As a gas it is a severe irritant of eyes, mucous membranes, and skin. It rapidly forms sulfurous acid on contact with moist membranes (16). Other contaminants. Mild inflammatory changes, primarily in the lungs, were noted when rats were exposed continuously to formaldehyde (0.0046 ppm), dimethylamine (0.009 ppm), and ethanol (0.086 ppm). When exposed repeatedly to ethylene glycol at 0.010 and 0.057 ppm no cha nges were seen. After 8 days of continuous exposure to 0.012 pp m of ethylene glycol, two out of fifteen rats suffered corneal damage with apparent blindness (23). Methane and hydrogen sulfide concentrations were evaluated in static isolat or cages. Methane leve ls were greater than 500 ppm after seven days, while no increase in hydrogen sulfide was detected. The physiological relevance and e ffects of these two gases on mice are still unclear (6). The aforementioned contaminants are a sa mpling of what could be present in a laboratory animal housing environment. The presence of any one of these contaminants is dependent on the combination of factors su ch as type of bedding, cage type, and strain of mice. A well-monitored cage is a key step to ensure a healthy and safe environment for all involved. Contact Bedding Bedding in animal cages is a controllable environmental parameter, which can influence animal welfare and research data (18). Beddings should be chemically and biologically inert, contaminant free, highly absorptive, nontoxic, dust-free, compatible with the research study, easily disposable, and inexpensive (10,15). Additionally, the ideal contact bedding should enhance the phys ical and psychological well-being of the animal, while not influencing it biologically (21).

PAGE 32

18 Bedding materials produce environmental pollu tants, such as ammonia. Ammonia production is influenced by the following proper ties of bedding: partic le size, absorption properties, and the presence of urease or a urease activator. The particle size plays an important role in desiccating fecal pelle ts and thus reducing ammonia production. Potgieter and Wilke state that because large particles have a larger exposure area they may dry faster (21). Smaller particles, however, have greater surface to volume ratio and it seems they would dry faster. Urease, an en zyme that catalyzes the hydrolysis of urea, is widely distributed in plants, which ar e the source for most bedding materials (21). Ammonia build-up in cages can be controll ed by the frequency of bedding changes, as the two are inversely related (21). The Guide for the Care and Use of Laboratory Animals henceforth referred to as The Guide, recommends that soiled bedding should be removed as often as is necessary to keep animals clean and dry. The frequency and intensity of cleaning and di sinfection should depend on what is needed to provide a healthy environment for an animal (18). The kind of contact bedding chosen can a ffect air contaminants such as ammonia (18). Common contact beddings include recy cled paper, ground co rncob, cellulose, and wood chips. Perkins and Lipman evaluated se veral contact beddings in static isolator caging, with 15 room ACPH, and four mice per cage (10). The beddings in the study were ranked from lowest to highest m ean ammonia concentration as follows: Corncob Virgin cellulose pelleted Recycled paper Hardwood chip Virgin pulp loose Reclaimed wood pulp Pine shavings Aspen shavings

PAGE 33

19 Corncob proved to be the best bedding under these conditions as no ammonia was detected after seven days of exposure. Corncob bedding is among the most popular beddings currently used. Relative Humidity Relative humidity is the ratio of th e partial pressure of water vapor, pv, in a given moist air sample to the partial pressu re in a saturated moist air sample, pg, at the same temperature and total pressure, p T g vp p,. (1.2) At room temperature, 25C (77F), the pre ssure of saturated wate r vapor is 3.2 kPa (0.5 psi). Monitoring relative humidity is another way to help control the microenvironment. Higher macroenvironmental relative humidity leads to higher intracage relative humidity, which increases ammonia pr oduction within the cage (10, 30). The Guide recommends a relative humidity range from 30-70% (18). A relative humidity threshold has been found where above this level ammonia production is independent of contact bedding. The threshold, however, varies with contact bedding (30). For example, when the macroenvironmental relative humidity was greater than 70%, the ammonia generation curve for a static isolator cage was similar between pine shavings and corncob bedding. When the relative humidity was reduced to 60% the slope of the corncob reduced to zer o for seven days while there was no change in the pine shavings curve (6). Potgieter and Wilke discovered in their experiments that

PAGE 34

20 when relative humidity and te mperature reached 50% and 21 C (70 F) ammonia production increased (21). Ventilation An average mouse (25-30 g) inhales approxi mately 35 liters (45 g) of air in a 24hour period, which is more than the total we ight of its food and water. The quality, quantity, and distribution of air are more directly associated with the animals health, comfort, and overall well-being th an other environmental factors (7). In addition, air associated pollutants can negatively affect the animals quality of life and general welfare (21). The Guide states that the purpose of ventilation is to (18): Supply adequate oxygen Remove thermal loads caused by animal respiration, lights, and equipment Dilute gas and particulate contaminants Adjust moisture conten t of room and cage air Create static-pressure differential between adjoining spaces Ten to fifteen fresh ACPH are recommende d for secondary enclosures and have been the standard for many years. This guide line does not take into account possible heat loads, species, size, number of animals, t ype of bedding or frequency of cage changing, room dimensions, or effici ency of air distribution (18). For individually ventilated cages, cage air change rates should be adjustable from 30 to 100 ACPH depending on the number of mice and changing frequency. The air velocity at the inlet to the cage should be less than 15 m/min. (50 ft/min.), which is consistent with still air. This reduces the risk of high-velocity air-cooling and dehydration of cage occupants (5).

PAGE 35

21 Reeb et al. (31) studied the impact of room ventilation rates on microenvironmental parameters for static isolator cages. The study found that the microenvironment maintained adequate levels of ammonia, ca rbon dioxide, and relative humidity at low (5 ACPH) room ventilation rates. They also di scovered that increasing the room ventilation rate had minimal effect on intracage ventilat ion except for cages on the highest row just below the fresh air supply. Increased room ventilation did, however, decrease the humidity in the room and cages. For exampl e, with 5 room ACPH the relative humidity was at 50%, but dropped to 22% as the room ACPH increased to 20. Ventilation rate can be an important fact or for controlling environmental ammonia concentration. However, studies indicate th at changes in ammonia concentration and ventilation rate are not li near. White and Mans (25) found that the mean ammonia concentration in unventilated cages did not va ry in direct proportion to the room air exchange rate. In a study by Besch (32), doub ling and tripling the r oom ventilation rate did not produce proportional decreases in am monia concentration. Serrano also found that increasing the ventilation rate di d not proportionally decrease the ammonia concentration (7). The lack of linearity is most li kely correlated to physical limitations placed on air motion patterns within a room, which is a function of the type of air diffusion system and the face velocity of ai r from the diffuser at varying room air exchange rates (25). Additional Environmental Factors In addition to the environmen tal factors previously descri bed there are a few others, which create a suitable living environment for mice. These include temperature, cage space, noise, and light levels.

PAGE 36

22 Temperature. Since rodents are warm-blooded an imals, they must maintain their body temperature within normal varia tion for their overall well-being (18). Recommended dry-bulb temper atures for mice are 18-26 C (64-79 F). The cage temperature may be higher than the macroe nvironment due to animal heat load, heat transferred from fan motors, and in efficient cage ventilation rates (5). Cage space. The recommended cage space for lab mice as defined by the Guide is listed in Table 1-1. Solid-bottom caging, with beddi ng is suggested as it is preferred by rodents (18). It is important to follow the re commended space guidelines as the number of animals in a cage positively influences ammonia production, carbon dioxide levels, temperature, and relative humidity (12, 21). Table 1-1 Recommended space for lab mice Weight, g Floor Area/Animal, in2 Height, in. <10 6 5 Up to 15 8 5 Up to 25 12 5 >25 >15 5 Noise and light levels. The control of noise and light levels is primarily for the comfort of the mice. Mice can hear frequencies ranging from 80 to 100 kHz, but are most sensitive to 15 to 20 kHz and 50 kHz. They hear high frequency and ultrasound, which is why intracage ultrasound should be mi nimized. Mice have adapted to low light levels of approximately 40-60 lux (3.7-5.6 candles). In comparison, ordinary office lighting is less than 500 lux (46 candles) (5). Previous Environmental Studies There are many factors that contribute to the animals environment as depicted in the previous section; and different combinati ons offer varying results. For example, one combination of contact bedding, bedding cha nge frequency, and ventilation rate may

PAGE 37

23 provide a suitable environment, while alteri ng one of those factors may allow ammonia production to occur earlier. The following is a review of previous environmental studies illustrating the numerous combinations that ha ve been tried to improve the animals environment. Ammonia concentrations in filter-top cages. Serrano (7) was influential in the evolution of the modern cage with his st udy on static isolator cage types in 1971. Through his study he determined the effect of different types of covers on the distribution of gases in cages with varying population sizes Four types of filter tops were used: fiberglass, molded laminated polyester, and two types of steel-wire mesh (40 by 40 and 20 by 20). In one experiment, eight mice were housed per cage with corncob bedding and contaminant gas levels were measur ed after seven days. Carbon dioxide concentration levels were less than 4000 ppm fo r all types of filter tops. For the polyester type filter, which is similar to present da y filter tops, the mean concentration on the seventh day for cages with 4, 8, and 16 mice were <2, 21 18, and 90 28 ppm, respectively. Ammonia concentrations varied significantly from cage to cage as seen in Table 1-2 with the 40 by 40 mesh having the highest concentration. Table 1-2 Ammonia concentration levels afte r seven days in static isolator cages Filter Top [NH3] (ppm) Fiberglass 63 33 Molded Laminated Polyester 21 18 40 by 40 Mesh 177 64 20 by 20 Mesh 35 35 Comparison between macroand microenvironment. Murakami (17) in 1971, compared the environment within the cage to the ambient air. He found that changes in temperature and relative humidity within the cage paralleled changes in the ambient air,

PAGE 38

24 with negligible difference between the inte rnal and external temperature. Relative humidity and ammonia concentrations, however, were notably higher than in the room. Decreases in room relative humidity. Lipman (11) conducted an environmental study on filter-top cages, specifically Sedl acek-type. First, microenvironmental parameters were measured while the room wa s held constant at 50% relative humidity. Then, the room relative humidity was decrea sed to 20% and the same parameters were measured. Below 50% relative humidity, the cage relative humidity was 20% higher, ammonia concentrations were 150 ppm higher, and carbon dioxide concentrations were 2300 ppm higher than in a cage without a lid. Ammonia was detected in the cages on day 4. When the room relative humidity was d ecreased to 20%, there was a 15% decrease in cage humidity, the mean weekly relative hum idity was 58%, and ammonia was less than 20 ppm and not detected until day 7. Static isolator cages and strains of mice. Hasenau et al. (33) compared four different static isolator cages for microenvi ronmental temperature, relative humidity, and ammonia concentrations. Three cages had polycarbonate bases and lids with Reemay 2024 filter material, while the fourth was used as a control without a filter. Comparisons were made of BALB/c and CD-1 same sex mi ce at four and two per cage under varying microenvironmental conditions. The followi ng parameters were used in the study: Room ACPH:20 Room Temp: 22.8 1.7 C (standard RH) 24.1 0.7 C (<40% RH) Relative Humidity: 51.5 8.2% (standard) 22.7 7.7% (below normal) Bedding:Autoclaved hardwood bedding Bedding Change Frequency:Every 7-9 days

PAGE 39

25 Under standard relative humidity (4070%) conditions in the macroenvironment, the relative humidity in th e filter-top cages, each housi ng 4 mice, ranged from 17-28% higher than the room levels. At below nor mal relative humidity (<40%) levels, the cage humidity ranged from 25-38% higher than the room. Ammonia concentrations varied signifi cantly between strains of mice and cage types. In the first study, where 4 BALB /c mice per cage were utilized, all cages accumulated less than 5 ppm by day 9. With two mice per cage the concentrations dropped to less than 3 ppm on day 9. When th e CD-1 mice were housed 4 per cage under standard relative humidity levels, ammonia c oncentrations after 8 days ranged from 1.9 ppm (control) to 117.1 ppm. The ammonia le vels dropped significantly for all cages when the macroenvironmental relative humid ity dropped to below 40%; ammonia levels varied from 0.1 to 8.7 ppm. At a stocking dens ity of two mice per cage, all cages had an ammonia concentration of less than 5 ppm at day 8. In conclusion, Hasenau et al. (33) determined a number of factors that could be altered to improve the envir onment. They found that reducing the macroenvironmental relative humidity reduced the ammonia produc tion in the cages. Also, reducing the stocking density from 4 to 2 mice decreased ammonia levels more than decreasing the room relative humidity. There was also a significant difference in ammonia levels depending on the strain of mice. Population size and cage type. Choi et al. (12) studied the effect of population size on the buildup of ammonia and relative hum idity in static and ventilated cages over time. The pressurized individually ventil ated cages received 50-60 ACPH, while the room received 15 ACPH.

PAGE 40

26 Environmental parameters varied between the static and individually ventilated cages. Ammonia levels were less than 1 ppm throughout the 32-day study for the individually ventilated cage s housing both 2 and 4 mice. Relative humidity increased slightly with the number of mi ce. In the static cages, no ammonia was detected in cages with 1 or 2 mice after 8 days. The relative humidity increased with the number of mice, which in turn increased the ammonia levels Levels with 3 and 4 mice were 5.5 ppm after 8 days. For all the cages, the cornc ob bedding appeared dry throughout the study and there were no wet areas in one spot, indica ting that the mice did not have one spot to urinate. Comparison of contact beddings. Potgieter and Wilke (21) investigated three different contact beddings for dust conten t, dust generation, mo isture absorption properties, and ammonia production. The cont act beddings: vermiculite, pine shavings, and unbleached eucalyptus pulp were chosen b ecause they were readily available in South Africa, the location of the st udy. The room was kept at 24.7 1.1 C (76.5 F) and 51.3 5.3% RH and received eight ACPH. On e hundred forty-four adult inbred conventional BALB/c mice were divided am ong the three bedding types and housed in static isolator breeding cage s that were changed weekly. Ammonia concentrations were surpri singly low throughout the study and never exceeded 3.5 ppm. The lowest ammonia level ( 1 ppm) measured on day 7 was from the eucalyptus pulp. Potgieter and Wilke (21) do not recommend using the vermiculite as contact bedding due to the quant ity of dust it produced. They advise using the eucalyptus pulp due to its moisture absorption propert ies and low levels of ammonia and dust.

PAGE 41

27 IVCS and absorbent bedding. Huerkamp and Lehner (9) characterized and compared microenvironments of three IVCS and a static is olator cage with ammoniainhibiting contact bedding to a standard static isolator cage contai ning corncob bedding. The cage ACPH for the three IVCS were 74, 106, and 112, respectively. Each IVCS was changed every 14 days, while the static cages were changed weekly. The room was held constant at 22.4 0.3 C (72 F) and 42 6% relative humidity with 15 ACPH. With the use of five and ten percent of th e absorbent bedding in the static isolator cages, the ammonia detection was delayed by on e day. However, over the course of six days the ammonia production was not altered. Ammonia was not detected in the three IVCS after seven days and after 14 days the levels were still low. Conversely, ammonia was detected after four days in the static cages and exceeded 100 ppm after seven days. Carbon dioxide was reduced in the IVCS ( 1050 1650 ppm) compared to an average of 2050 ppm in the static cages. Methane was dete cted in all cages at an excess of 500 ppm, while hydrogen sulfide was not detected. Comparison of individually ventilated cages. Hoglund and Renstrm (34) evaluated two different IVCS (BioZone Ven tiRack and Tecniplast Sealsafe) for ammonia concentrations after two weeks, carbon di oxide build-up during a one-hour simulated power failure, and the ability to maintain a po sitive or negative pre ssure differential for long periods of time. Male mice, 10 weeks ol d, were used in the study, housed three per cage. Aspen wood shavings were used for th e contact bedding. The room was held at 22 1 C (71.6 F) and 55 5% RH and received 17 ACPH. The VentiRack from BioZone provided a more uniform and balanced differential pressure, but the systems exhibited similar behavi or in all other areas. Under either the

PAGE 42

28 negative or positive pressure differential the ammonia content in the cages was less than 10 ppm after 10 days when the bedding was not soaked. If the bedding was soaked, the ammonia concentration remained high re gardless of the ventilation rate. Carbon dioxide did not build up to harmfu l concentration levels in the one-hour simulated power failure due to the filter-top cages that were used with the IVCS. Ventilation and frequency of bedding changes. Reeb et al. (13) evaluated the microenvironment in pressurized individually ventilated (PIV) cages under two different conditions: varying cage air change rates a nd reduced frequency of bedding changes. Cage ventilation rates were held constant for 1 week at 30, 40, 60, 80, and 100 ACPH. Bedding was not changed for 26 days. Two gr oups of mice were evaluated: 9-11 week old males and trio groups for mating with pups less than 14 days old. The bedding was autoclaved white pine shavings. The micr oenvironmental parameters measured were temperature, relative humidity, ammonia, and carbon dioxide. The results from increased cage ventilation show that the environment improves with more circulation. Ammonia and car bon dioxide decreased significantly with increased ventilation rates. For all ventil ation rates the ammonia level was less than 3 ppm. Relative humidity was significantly higher at 30 and 40 ACPH, while it was not significantly different for 60, 80, and 100 ACPH. For less than 60 ACPH the temperature was 25.0.02C and dropped significantl y at 80 ACPH to 23.3C. Higher ventilation rates could reduce fre quency of bedding changes to once every two weeks. Mean ammonia concentrations st ayed low for 21 days and increased to 12 ppm between days 21 and 26. The relative humidity was highest at day 21 (45%) and

PAGE 43

29 decreased by day 26. Carbon dioxide and temp erature fluctuated, but did not increase in relation to the number of da ys with soiled bedding. In a similar study, Reeb-Whitaker et al (35) compared three different cagechanging frequencies (7, 14, and 21 days) at three different cage ACPH (30, 60, and 100). Twelve breeding pairs and twelve breeding tr ios were evaluated for general health over seven months. Pressurized individually ventil ated cages with white pine shavings were used to house the mice. Ammonia was gr eater than 25 ppm at 30 ACPH at all frequencies of bedding changes and at 60 AC PH after 21 days. The pup mortality rate was higher when cages were changed every seven days. Reduced frequency of cage changes had no effect on the following health areas: weanling wei ght, animal growth, plasma corticosterone concentrations (important for carbohydrate and protein metabolism), immune function, breede r mortality, and breeder productivity. It is evident that many environmental studies have been done which combine a wide range of environmen tal parameters. Appendix A includes a comprehensive matrix classified by the dependent variables in each study. Cost Analysis for Current Husbandry Practices Besides maintaining a healthy environmen t, many of the previous studies were driven by economics. One of the major cost savings areas is th e frequency of bedding changes. For example, at the University of Florida Animal Care Services static microisolator cages are changed twice a week and not autoclaved, which translates into $116 per cage per year. Agrawal (36) recomm ends decreasing the frequency from 3.5 days to 7 days and autoclaving the bedding, which will reduce the cost to $76, a savings of $40 per cage per year. Likewise for i ndividually ventilated cages, if the bedding change frequency can be extended from 14 to 21 days a cost savings of $16 per cage per

PAGE 44

30 year could be realized. Considering the An imal Care Services current housing needs, $100,300 could be saved per year. With real-time monitoring of cages, th e bedding change frequency could be reduced and savings realized without compromi sing the health and welfare of the animals and laboratory personnel. The goal of this work was to explore the possibilities of continuous monitoring of laboratory animal cages through the use of field effect transistors. Air Sampling Techniques Presently, there is no standa rdized sampling method for rodent aeroallergens nor is there a uniform procedure for measuring a nd quantifying rodent allergen exposure in rooms with IVCS systems (5). Current husbandry practices rely on environmental measurements taken in studies and regulated by the Guide (18) to control aeroallergens. Environmental measurements in studies were typically taken with a gas-sampling pump as seen in Table 1-3. While these sampling pumps have low measurement error, .5% (35), sampling is often done on an infrequent ba sis due to cost and time restraints. For example, in one study measurements were ta ken three times between 1:00 pm and 5:00 pm on days 6, 13, and 20 of a 21-day cage ch anging cycle. Both experiments and husbandry practice would benefit through con tinuous or more frequent monitoring of cages. Table 1-3 Gas sampling pumps us ed in environmental studies Type Manufacturer Reference Gas Analyzer, Model 1302 and 1303 Brel and Kjr (35) Aspirating Pump #8014-400A Matheson Gas Products (12) Multi Gas Detector Drger (21) Toxic Gas Monitor, Model SC-9 Riken Keiki (34)

PAGE 45

31 One way to provide continuous monitoring in an animal cage is through low-power miniaturized gas sensors offered as field effect transistors (FET). For background purposes, the following section includes general information on semiconductor properties, field effect transistors, and surface reactions betwee n sensing films and ammonia or carbon dioxide. Semiconductors In general FETs operate on the principle of electrical manipulation of fields. The field manipulation is controlled by a gate, which acts on the conduction of carriers in a semiconductor channel (37). Semiconductors are materials consisting of elements from group IV of the periodic table with electrical proper ties lying somewhere betw een insulating and conducting materials. Conducting material is character ized by a large number of conduction band electrons that have a weak bond with the basi c structure of the material. Therefore an electric field easily transmits energy to the oute r electrons and allows the flow of electric current (38). Semiconductors act as conductors when the electrons possess enough energy to exceed the energy gap, Eg, between the valence band, the en ergy level filled by electrons in their lowest energy state, and conducti on band, the unfilled energy level into which electrons can be excited to pr ovide conductivity as seen in Figure 1-3 (39). Semiconductors have a lattice structure, which is characterized by covalent bonding. Whenever a free electron leaves the lat tice structure, it crea tes a positive charge or hole. Electrons move to fill the holes, consequently creating more holes. When voltage is applied electrons move towards th e positive band, while ho les shift towards the

PAGE 46

32 negative band. The movement of electrons and holes conducts current. The number of electron-hole pairs determines the conductivity according to the following relationship, h h e eq n q n (1.3) where ne and nh are the number of electrons in the conduction and valence bands, respectively, e and h are the mobility of electrons and holes, respectively, and q is the charge (39). Figure 1-3 When excited, electrons move fr om the valence band to the conduction band across the energy gap. The number of charge carriers in semiconducto rs is controlled by temperature. At absolute zero, all the electr ons are in the valence band, while the conduction band is empty. As the temperature increases it is more likely that an energy level in the conduction band will be occupied. The number of electrons in the conduction band is equal to the number of holes in the vale nce band and is related to temperature, T by kT E n n n ng o h e2 exp (1.4) where k is Boltzmanns constant and no is a constant (39). If the voltage source or exciting energy is removed, the holes and electrons will recombine over a period of time, designated as,

PAGE 47

33 t n noexp (1.5) where t is the time after the field is removed and is a constant known as the recombination time (39). The behavior of an intrinsic semiconductor cannot be accurately controlled due to its sensitivity to slight variations in temp erature. Therefore a dding impurities or dopants that determine the number of charge carriers can create an extrinsic semiconductor. An n-type has an extra electron that lowers the energy level, whereas a p-type does not have enough electrons and a hole is created (39). Field Effect Transistors A transistor is a three-terminal se miconductor device, which performs two functions: amplification and switching (38). Transistors have three connections, where the voltage on (current into/ out of) switch has the effect of controlling the ease with which current can flow between the other two terminals (40). The effect is to make a resistance whose value can be altered by the input signal. The patterns of signal fluctuation can be transferred from a sma ll input signal to a larger output signal. Specifically for a metal oxide semiconductor-F ET (MOSFET), there are three terminals: gate, drain, and source (Figure 1-4). The gate is a metal film layer that is separated from the bulk by a thin oxide layer. When a voltage is applied to the gate an electric field is created which repels positive charge carriers away from the surface of the bulk in which the negative charge carriers dominate and are available for conduction. By increasing the gate voltage, the depth of the channe l can subsequently be increased.

PAGE 48

34 Figure 1-4 Schematic of field effect transistor (FET) (40). Upon gas exposure, the surface of the se nsing film registers a work function changea, which is seen electronically as a gas se nsitive potential. This potential is then added to the gate voltage and ope rates the transistor (41). There are two basic types of FETs used in gas sensing: a classical FET and a suspended gate FET (SGFET) or hybrid flip-c hip FET (HFCFET). In the classical FET the sensing layer lies directly on the gate su rface, whereas in the SGFET the sensing film is separated from the gate surface by an air gap (Figure 1-5 and Figure 1-6). In order to ensure adequate capacitive coupling the air gap must be no larger than a few micrometers, but at least one micrometer to allow for sufficient gas diffusion through the channel. There must be no contact between the sensitive material and channel-insulating layer (41). An FET consists of two parts: an alumin a substrate that contains the conducting structures for the flip-chip contacts and a sensitive laye r on a separate electrode (Figure a The smallest amount of energy, measured in electron volts, required to remove an electron from the boundary of an element.

PAGE 49

35 1-7). The sensor film is applied to a de signated area using a freel y selectable deposition process. If a heater element is needed, it is on the backside of the alumina substrate. For an HFCFET the etching steps in thin-film technology ensure th at a defined air gap for gas diffusion is formed when the FET is mounted. An advantage of using an air gap is that the same substrate design can be used for all types of transducers because the geometry does not change (41, 42). Additional schematics and de tailed electrical drawings can be found in Appendix B. Figure 1-5 Classical FET configuration where th e sensing film lies in the same plane as the gate. Figure 1-6 Suspended Gate FET configuration where the sensing film is separated from the gate by an air gap. Gas Sensing Field effect transistors can be used to detect a wide variety of gases by choosing the sensing layer that reacts with or catalyzes the specific gas. The reaction mechanism,

PAGE 50

36 adsorption, and desorption rates will vary fo r each combination and creates a challenge to the designer to find the best combination. Figure 1-7 Photograph of a hybrid flip chip FET sensor device (left) mounted on a printed circuit board (right). (Data of Simon et al. (46)) Hydrogen detection is the most basic r eaction and aids in understanding more complex reactions involved with ammonia and carbon dioxide sensing. For sensing of hydrogen the reaction occurs in three steps. First, the hydrogen mo lecules dissociate on the catalytic metal surface of the device. Secondly, the hydrogen atoms are transported through the metal film. Lastly adsorption of hydr ogen occurs at the interfacial layer between the metal and insula tor where a dipole layer form s. The adsorbed hydrogen disrupts the electric fields acr oss the device structure and is detected through changes of the electrical characteristics of the device (43). Ammonia reactions. For ammonia detection, one surf ace that is used is titanium nitride (TiN). Ostrick et al. (44) claim th at water and ammonia adsorb on the TiN film and change the work function of the film. The TiN surface is covered with hydroxide (OH) and water and upon exposure to a mmonia the OH species is reduced and compounds related to ammonia are formed. Two reaction mechanisms are possible for

PAGE 51

37 ammonia adsorption. In the first, the a mmonia removes OH from its binding site. The second mechanism proposes that the OH groups become binding sites for the ammonia, as seen here surface surface gasNH O OH NH4 3 (1.6) The change of the work function, is due to the difference of dipoles on the surface of OH and ONH4, given by, OH ONHd d4~ (1.7) where is the coverage of ONH4 and d is the dipole moment (44). Ostrick et al. (44) found that the reacti on of ammonia on TiN occu rred reversibly at room temperature and was not hindered by preadsorption. No cross sensitivities to the following gases were found: carbon monoxide (30 ppm), carbon dioxide (3000 ppm), nitrogen dioxide (1 ppm), hydrogen (10 ppm), methanol (10 ppm), and acetone (10 ppm). Sensitivity to ammonia was found to be inde pendent of relative humidity (5-80%). At room temperature ammonia solves almost completely in water under formation of NH4 + and OHions. Since the ammonia concentra tion is low (<100 ppm) compared with the surface water concentration (>5000 ppm), al l of the ammonia react s with the water. In the next reaction step, the ammonia ions may react directly with the surface or with adsorbed hydroxide ions. Due to the ex cess of water molecules, the found ammonia sensitivity may be independent of the wate r concentration. At higher temperatures, however, the sensitivity to am monia decreases. The sol ubility of ammonia in water decreases at higher temperatur es as well as the amount of water on the surface. The amount of ammonia on the TiN surface is also reduced, therefore the ammonia is less likely to react with adsorption sites and the sensitivity decreases.

PAGE 52

38 A reaction mechanism proposed for amm onia detection in metal-insulatorsemiconductor (MIS) field effect devices is that adsorbed NH4 + is detected on the oxide surface in the holes and cracks of the metal f ilm. Another proposed mechanism is that ammonia dissociates on the catalytic surface and reacts similarly to hydrogen. The SGFET and MIS devices diffe r in that, for an SGFET the response results from adsorption of the detectable species on th e surface, whereas in an MIS it occurs by adsorption in the interface between metal and oxide. Results from Abom et al. (43) imply that when a porous TiN film is used for an SGFET, no NH3 response is seen. Even when the surface is covered with NH4 +, the molecules cannot reach the reactant surface SiO2 and produce a response. However, it was found that NH3 is dissociated if Pt is present and the resulting atom ic hydrogen can be detected. Carbon dioxide reactions. Ostrick et al. (45) outlin ed the temperature dependent reaction mechanisms for barium carbonate (BaCO3) as the sensing film used to detect CO2. At 50C the reaction of CO2 is dependent on water and occurs only if water is present, yet it is independent of the partia l pressure of oxygen. When the temperature increases, above 200C, the reaction is mo re complicated and is dependent on oxygen and humidity. The reaction mechanisms at low temperatures are predicted as follows, surface gas surfaceHCO CO OH 3 2 (1.8) bulk surface gas surface surfaceCO O H OH HCO2 3 / 2 3 (1.9) At high temperatures, for CO2 in O2 as the dry carrier gas, th e reaction of formation of the carbonate appears as, bulk metal gas gasCO e O CO2 3 2 2 1 22 (1.10)

PAGE 53

39 which results in an electromotive force (emf) change of ) ln( 4 ) ln( 22 2O B CO B op e T k p e T k E E (1.11) where k is the reaction rate constant, e is the elementary charge, and pi is partial pressure. In terms of work function ch ange, the reaction appears as, ) log( ) log(2 2 2 2O O CO CO op S p S (1.12) where S is the sensitivity. When humidity is present, the sensitivity to oxygen is reduced and another reaction dominates, possibly th e formation of hydrogen carbonate from carbonate, carbon dioxide, and water, surface gas gas bulkHCO O H CO CO 3 2 2 2 32. (1.13) The work function equation for this reaction is ) log( ) log(2 2 2 2O H O H CO CO op S p S (1.14) Another surface layer used to detect CO2 is barium titanate (BaTiO3) (46). This compound can exist with a low excess quantity of barium. The excess of barium is not compensated by Tior Ovacancies or by fo rming low amounts of a new barium titanate phase with a higher stoichiometry of barium. A mixture of BaTiO3 with CuO has been reported as a highly sensitive material for CO2 sensing using the capacitance change in a temperature range of 200-1000C. Kelv in probe measurements of BaTiO3 indicate that it has a fast response time, less than one mi nute, and a sufficient sensitivity to CO2 (20 mV/decade). Similar to the BaCO3 surface, the sensitivity of CO2 is dependent on the presence of humidity. The sensor sh owed significant cross-sensitivity to NO2 and drift effects tended to increase. Ostrick et al. (47) also investigated the different reactions occurring in a multilayer system, Pt/NiO/BaCO3, used to detect CO2. It was found that besides the CO2

PAGE 54

40 reaction, a separate reaction to NO2 could occur at the NiO in terface. Inserting inert metal oxide layers stopped this reaction. Summary As developed in this chapter, there is an evident need for co ntinuous environmental monitoring in laboratory animal cages. By continuously monitoring the cages, laboratory animal personnel can determine when a cage ne eds to be changed so that the mice are disrupted only when necessary. Prolonged cag e changing intervals al so translate into labor savings for the animal care facility. Through real-time mon itoring, individually ventilated cages can be prope rly ventilated; increased vent ilation during active times and reduced ventilation if the mice are inactive a nd air contaminant levels are low. Field effect transistors provide a way to achiev e continuous monitoring of ammonia, carbon dioxide, temperature, and relative humidity in a laboratory animal cage. Most imperatively, continuous monitoring can provi de a better environment for the animals and laboratory personnel. As a prelude to the following chapters, the goals of this project were to: Assess feasibility of applying field eff ect transistors for monitoring laboratory animal cages through the following tests o Single analyte in air o Time-response o Cross-contamination Theoretically model the chemical kinetics and catalysis between the sensing film and contaminant gases and more specifically, o Define reaction mechanisms o Determine adsorption and desorption rates o Find position of chem ical equilibrium

PAGE 55

41 o Determine equilibrium constants o Explore the role of diffusion

PAGE 56

42 CHAPTER 2 EXPERIMENTAL FACILITIES AND METHODS The purpose of the experimental portion of this study was to assess the feasibility of applying field effect tran sistors for monitoring micr oenvironments in laboratory animal cages used in animal research f acilities; specifically the work focused on ammonia and carbon dioxide sensors. Experimental Setup The experimental facilities included hu midification, mixing, and sensor-testing regions as seen in the schematic in Figure 2-1. Figure 2-1 Schematic of experimental faciliti es, which included a mixing, humidification, and sensor-testing region. In general, the desired gas mixture (e.g ., air, ammonia, carb on dioxide, and water vapor) was prepared by mixing the desired gases (see Table 2-1 for stock gas) and then passing through the test chamber. Table 2-1 provides a comprehensive list of the equipment for the experiments. If humidi fication was necessary, compressed synthetic Humidification section Mixing Section

PAGE 57

43 air (Praxair) was preheated to 100C, and th en water was added through a variable flow pump (Fisher Scientific, Mode l No. 13-876-4), pictured in Figure 2-2. Once humidified, ammonia or carbon dioxide (P raxair) was added to the air system through flow controllers (Alicat Scientific, MC Series) as pictured in Figure 2-3. Table 2-1 Sensor testing equipment Sensor Testing Equipment Manufacturer Part No. Vacuum Piping ----------Flow meters Alicat Scientific -----Fiberglass Cloth Tape Insulati on Fisher Scientific 01-472A Variable Flow Peristaltic Pu mp Fisher Scientific 13-876-4 Laptop Gateway -----Data Acquisition System Nati onal InstrumentsNIDAQPad-6020E Labview Software National InstrumentsLabview 7.1 Controller Omega CN1A-TC-24V Handheld Thermometer Omega HH-26K Heating Tape, 1/2"x4' Omega SRT-051-040 Heating Tape, 1/2"x6' Omega SRT-051-060 Multimeter Omega HHM-11 T/C TO ANALOG CONVERTER, "K" Omega SMCJ-K Type K Connector Omega OST-K-MF Type-K ex. Wire, 100' Omega EXPP-K-20-100 Type-K Probe Omega KQIN-116G-12 2% Ammonia in Nitrogen Praxair -----75% Nitrogen, 25% Carbon Dioxide Praxair -----Synthetic Air Praxair -----DC Power Supply Protek 303 The combined gas stream flowed either directly over the sensor, Figure 2-4a, or into a Plexiglas box, Figure 2-4b. The Plexiglas box was used to simulate cage conditions where diffusion would be the main m ode of mass transfer to the sensor. When the gas stream flowed directly over the sensor the flow rate was limited to between 1 to 2 L/min as specified by the sensor developers Higher flow rates could induce baseline drift and add noise to the si gnal. At a flow rate of 2 L/min the Reynolds number was 460, which indicates the flow was laminar.

PAGE 58

44 (a) (b) Figure 2-2 Humidification sect ion of experimental facilities: (a) PID temperature controller heats the air befo re and after adding water, (b) Variable flow pump adds water to the air stream. Air N H 3 CO 2 Sensor Wate Prehea Humidified air Air N H 3 CO 2 Sensor Water Preheat (a) (b) Figure 2-3 Mixing section of e xperimental facilities: (a) Sc hematic of mixing section (b) Flow meters used to control flow rate of air, ammonia, and carbon dioxide In the diffusion case, the relevant fl ow rates for the calibration gases were determined based on the air exchange rates a nd velocity of air leaving an individually ventilated mouse cage. In a typical cage, air enters the cage through a small tube and exits the cage by way of a small gap between the lid and cage bottom. The flow rate calculations were based on th e following cage parameters: Cage air changes per hour 60 Volume of cage 7.067 L (431.25 in3)

PAGE 59

45 Cage width 12.7 cm (5 in.) Gap opening 1.27 cm (0.5 in.) Area of gap 16.1 cm2 (2.5 in2) min / 38 4 129 16 min 067 72m cm L A Q Vcage (2.1) where Q is the volumetric flow rate, A is the cross-sectional area, and Vcage is the velocity of air leaving the cage. (a) (b) Figure 2-4 Two flow regimes used in sensor te sting: (a) Gas stream flowed directly over the sensor or (b) Gas stream diffused onto the sensor mounted in a Plexiglas box used to simulate the cage environment. The volumetric flow rate in the test section was calculat ed based on the velocity of air leaving the cage. With a cross-sectional area of 9.62 cm2 (1.5 in2) the volumetric flow rate was calculated as follows, min / 21 4 min 38 4 62 92L m cm Qgas (2.2) In keeping with this value, the air flow ra te was set to 5 L/min in the diffusion test section. The stock ammonia and carbon di oxide were suspended in nitrogen with concentrations of 2% and 25% by volume, respectively (see Table 2-1). The ammonia and carbon dioxide streams were th en diluted to the desired c oncentrations by the pure air

PAGE 60

46 stream. The desired analyte concentration wa s entered into a Labview program (National Instruments (2004) Labview 7.1), which subseque ntly calculated the necessary flow rates and used analog output channels to control the flow controllers (Alicat Scientific). The accuracy specification for the flow controller s was % of the full scale, where the full scales for the air, ammonia, and carbon dioxide controllers were 5 L/min, 50 cc/min, and 100 cc/min, respectively. The accuracy for the compressed gas cylinders was % and .25% for the ammonia and carbon dioxide cy linders, respectively. The uncertainty associated with the ammonia and carbon dioxi de concentrations for each experimental condition is listed in Table 2-2. Detailed uncertainty calculations can be found in Appendix C. The actual sensor was on a micro-scale as depicted in Figure 2-5. For ease of use in the laboratory, the sensor was moun ted on a larger box housing the electrical connections, as shown in Figure 2-4b. In actual field operation, the sensor would be mounted in a cage with wireless feedback to th e electrical board. To prevent the signal from drifting, the surface of the sensor was heated by means of an internal heater. The internal heater was powered by voltage from an external regulated DC power supply (Protek, Model No. 303). Figure 2-5 Carbon dioxide sensor. (S cale bar = 1 inch (2.54 cm)).

PAGE 61

47 Table 2-2 Ammonia and carbon dioxide concentr ations and their associated uncertainties for each experimental condition. Flow regime Air flow rate (L/min) NH3 conc. (ppm) NH3 uncertainty (%) CO2 conc. (ppm) CO2 uncertainty (%) 5 5 40.0 5 10 20.1 5 15 13.5 5 20 10.2 5 25 8.3 5 30 7.0 5 40 5.5 5 50 4.6 5 75 3.5 5 100 3.0 5 150 2.6 Diffusion 5 200 2.4 2 50 10.5 2 100 5.9 2 300 41.9 2 1000 13.6 2 2500 7.4 2 3000 6.9 2 5000 6.1 Forced 2 7000 5.8 5 50 4.5 3000 5.3 5 50 4.5 4800 5.2 Diffusion 5 100 3.0 3000 5.3 Signals from the sensors were collected through a data acquisition board (National Instruments, NIDAQPad-6020E) and processe d using Labview software (National Instruments (2004) Labview 7.1). Inputting a two-volt signal from a regulated DC power supply (Protek, Model No. 303) directly into the board for a six hour time period tested the accuracy and precision of the board and softwa re. The results of this test can be seen in Figure 2-6.

PAGE 62

48 0123456 1.96 1.98 2.00 2.02 2.04 Signal (V)time (hr) Figure 2-6 Signal from DC power supply as co llected in Labview, which shows precision and accuracy of the data acquisition system over six hours. Experimental Procedures Before starting any of the other tests, th e carbon dioxide and a mmonia sensors were tested for drift of the baseline signal. Out put signals were collected every 30 minutes for over 40 hours while air flowed over the sensor at a rate of 2.00 L/min. The sensors were not exposed to either ammonia or carbon dioxide during these tests. Carbon Dioxide Sensor Single analyte tests, where the sensor wa s only exposed to mixtures of air and carbon dioxide, were completed to establish a calibration curve. Since carbon dioxide levels should not exceed 5000 ppm; the meas urable range was set to 300 to 7000 ppm. The sensor was exposed to concentrati ons between 300 and 7000 ppm carbon dioxide for 10 to 30 minutes at a time with a 10 minute purging cycle with air between exposures. These tests were all conducted using the forced flow regime. The other experiment completed with the carbon dioxide sensor tested for crosssensitivities to humidity. First, air flowed over the sensor for 30 minutes to establish a baseline. Then the sensor was exposed to 3000 ppm carbon dioxide for 30 minutes. The

PAGE 63

49 relative humidity was then increased from dry (2%) to humid (50 60%). After humidifying the air, the sensor was ag ain exposed to 3000 ppm carbon dioxide. A graphical depiction of this test can be seen in Figure 2-7. 020406080100120140 0 500 1000 1500 2000 2500 3000 3500 Gas concentration (ppm)time (min) CO2 conc.0 25 50 75 100 rh Rel. Humidity (%) Figure 2-7 Graphical depiction of experimental parameters us ed to test the carbon dioxide sensor for cross-sensitivity to humidity. Ammonia Sensor A number of single analyte te sts were conducted on the ammonia sensor in order to establish a calibration curve. Since the recommended threshold limit value (TLV) for ammonia is 25 ppm, the desired measuring range was 25 to 100 ppm. The sensor was exposed to concentrations of 25, 50, 75, and 100 ppm ammonia for 10 to 20 minutes at a time with a 10 minute purging cycle with air between exposures. These tests were all conducted using the forced flow regime. In actual operation the sensor will not have a purging cycle after it has been exposed to air contaminants. Therefore, tests were conducted where the ammonia concentration was ramped up or down to simula te gradual changes that may be seen in the environment. The parameters used for the ramping tests are listed in Table 2-3, while

PAGE 64

50 a graphical depiction is given in Figure 2-8. The ramping tests were completed using the diffusion box. Table 2-3 Ammonia concentrations used for ramping tests. Ramp up Ramp down NH3 Concentration (ppm) Duration (min) NH3 Concentration (ppm) Duration (min) 0 10 0 10 5 10 200 10 10 5 150 5 15 5 100 5 20 5 75 5 25 5 50 5 30 5 40 5 40 5 30 5 50 5 25 5 75 5 20 5 100 5 15 5 150 5 10 5 200 5 5 5 Increment back to 0 Increment back to 200 020406080100120140 0 50 100 150 200 Gas concentration (ppm)time ( min ) 020406080100120140 0 50 100 150 200 Gas concentration (ppm)time ( min ) (a) (b) Figure 2-8 Graphical depicti on of ammonia concentrations used in ramping tests. Since the laboratory animal cage is not a homogeneous environment, the effects of other air quality factors were tested. First, the ammonia sensor was tested for crosssensitivity to humidity. The sensor was e xposed to air under dry conditions (2% rh) for 20 minutes, the relative humid ity was increased to 40% rh for 20 minutes, and then this cycle was repeated twice. Six repetitions of this cycle were completed for the forced and

PAGE 65

51 diffusion flow regimes. Second, the sensor was tested for cross-sensitivity to humidity while ammonia was present. The sensor wa s first exposed to ammonia for 30 minutes, followed by 30 minutes of air, then the relativ e humidity was increased, and lastly the sensor was exposed again to ammonia. A graphical depiction of the experimental parameters used for this test can be seen in Figure 2-9. 0306090120150 0 25 50 75 100 Gas concentration (ppm) Relative humidity (%)time (min) NH3 conc. RH Figure 2-9 Graphical depiction of experimental parameters used to test the ammonia sensor for cross-sensitivity to humidity while ammonia was present. Next the ammonia sensor was tested for cross-sensitivities to humidity and carbon dioxide with ammonia in the system. A gr aphical depiction of the combination of parameters used is shown in Figure 2-10. The introduction of each new parameter was at least 10 minutes after the last ch ange in experimental conditions to ensure that the effect on the sensor was from the intended parameter. Six repetitions of this experiment were conducted. Time response tests were completed to determine the amount of time required by the sensor to display 95% of a step change in gas concentration. This test also aided in determining adsorption and desorption rates of the analyte on the sensor. To complete this test, air flowed across the sensor fo r 10 minutes reaching steady state, and then

PAGE 66

52 ammonia was added to the system for 10 minut es. The analyte stream was shut off and air again flowed over the sensor for 10 minutes This cycle was repeated six times for each of the eight experime ntal conditions listed in Table 2-4. 0306090120150 10 100 1000 Gas conc. (ppm)time (min) CO2 NH30 25 50 75 100 rh Rel. Hum. (%) Figure 2-10 Graphical depiction of experimental parameters used to test the ammonia sensor for cross-sensitivity to humidity and carbon dioxide with ammonia in the system. Table 2-4 Experimental parameters for time re sponse tests (X indicates tests completed). Flow Humidity (%) 50 ppm NH3 100 ppm NH3 Forced 2 X X Diffusion 50 X X Forced 2 X X Diffusion 50 X X In the last set of experime nts, the surface temperature of the sensor was varied to test for temperature effects on the response tim e of the sensor. The sensor was exposed to 50 ppm ammonia for 10 minutes at 31C (1.0 V), 39C (2.0 V) and 52C (3.0 V). The temperature of the surface was controlled by an internal heater, which was powered by an external DC power supply (Protek Model No. 303). The output voltage was calibrated to

PAGE 67

53 the above-mentioned temperatures At each temperature, the experiment was repeated six times.

PAGE 68

54 CHAPTER 3 THEORETICAL MODELING The theoretical modeling of the reactions between the sensin g film and analyte gases in the field effect transi stors allowed for a number of qu estions to be investigated. The first topic of research dealt with th e reaction mechanism fo r the ammonia gas on a titanium nitride surface. Th e second inquiry focused on the adsorption and desorption rates for each reaction. Th e goal was to model the time response for adsorption and desorption and then determine the position of ch emical equilibrium. In the process of conducting these queries, the equilibrium cons tants for each reaction were explored. Diffusion was also investigated as it played an important role in transporting the molecules to and from the surface so that ad sorption or desorption may take place. The experimental portion of the study was used to confirm the adsorption and desorption rates found through the theoretical modeling. To determine the adsorption and desorption rates for each reaction, the position of chemical equilibrium, and the equilibrium constants, the following heterogeneous chemistry theory was reviewed and applied to the analysis. Gibbs Free Energy Gibbs free energy, G, is essential in determining the driving force or spontaneity of chemical reactions, the equilibrium constant, and the position of chemical equilibrium, where the equilibrium position of a reaction is said to lie far to the right if almost all reactants are used up and far to the left if scarcely any product is formed. Gibbs energy can be described by the following equation,

PAGE 69

55 TS H G (3.1) where H is enthalpy, T is the absolute temper ature, and S is entropy. For an isothermal process, the change in free energy as the process proceeds can be written as, S T H G (3.2) A chemical process will continue in the dire ction that decreases the free energy. For example, if 0 Gthen the forward reaction will contin ue spontaneously. Likewise, if 0 Gthen the reverse reaction w ill occur spontaneously. The process continues in the direction to minimize the free energy of the system until0 G, at which point equilibrium is achieved. The free energy change can be related to th e reaction equilibrium constant. First, consider the elementary reaction yY xX bB aA (3.3) If the reaction proceeds by a differential amount, d then the number of moles of each chemical species changes according to ad dnA (3.4) bd dnB (3.5) xd dnX (3.6) yd dnY (3.7) The total differential of mixture free energy is thus K k k kdn Vdp SdT dG1, (3.8) where is the chemical potential. At fixed temperature and pressure, this equation converts to Y Y X X B B A Adn dn dn dn dG (3.9)

PAGE 70

56 d y d x d b d a dGY X B A (3.10) d b a y x dGB A Y X) ( (3.11) ) (,B A Y X T pb a y x d dG (3.12) Equilibrium is reached when 0, T pd dG. (3.13) Therefore at equilibrium, 0 r B A Y Xb a y x (3.14) Considering the chemical potential at standard state, Eq. 3.12 can be written as, 0 ln ln ln ln B o B A o A Y o Y X o Xp RT b p RT a p RT y p RT x (3.15) With this result, a constraint is put on the pressures that the four gases can have at equilibrium. Rearranging Eq. 3.15 leads to o B o A o Y o X eq b B a A y Y x Xb a y x p p p p RT ln, (3.16) o r eq b B a A y Y x XG p p p p RT ln. (3.17) Because the right hand side of Eq. 3.17 is constant for a given temperature, the logarithmic term must also equal a constant, eq b B a A y Y x X pp p p p K (3.18)

PAGE 71

57 known as the equilibrium constant. If Kp is greater than one, th e products of reaction are favored over the reactants and the forward reaction proceeds. Based on Eq. 3.17 and Kp, at equilibrium, p o rK RT Gln (3.19) where it can be rewritten with the help of Eq. 3.2 as p o r o rK RT S T H ln (3.20) or p o r o rK R S T Hln (3.21) When the reaction is exothermic, T Ho r relates to a positive change of entropy of the surroundings and favors the formation of the products. As the temperature increases, T Ho r decreases and the increasing entropy of the surroundings has a less powerful effect. Resultantly, the equilibrium lies less to the right. If the reaction if endothermic, the primary factor is the increasing entropy of the reaction system The significance of the unfavorable change of the entropy of th e surroundings is lessene d as the temperature increases and the reaction can shift towards the products. Looking at the forward and reverse rates of progress at equilibrium develops the equilibrium concentration constant, Kc. If the reaction is at equilibrium, then, ki kiK k k i r K k k i fX k X k 1 1 (3.22) or kie K k k i r i f i cX k k K 1 , ,, (3.23)

PAGE 72

58 where e indicates equilibrium, and ki is the net stoichiometric coefficient for species k in reaction i. Concentration is related to pr essure through the ideal gas law, RT p V n Xk k k (3.24) Equation 3.23 now becomes, kie K k k i cRT p K 1 (3.25) The equilibrium pressure constant fo r a general reacti on is written as, K k e o k i pkip p K1 (3.26) where po is the standard-state pressure. The two equilibrium constants can be combined into one equation as follows, kiK k o i p i cRT p K K 1 ,. (3.27) Gibbs free energy and associated thermodynamic terms were an essential part of examining the reaction of ammoni a on a titanium nitride surface. Adsorption The reactions between the titanium nitride surface on the field effect transistor and ammonia occurred through adsorption, attach ment of particles to the surface. The substance that adsorbs is called the adsorbate while th e underlying material is the adsorbent. The reverse of adsorption is de sorption. Adsorption can occur in two ways: physical adsorption, physisorption, or chemical adsorption, chemisorption. Physical adsorption is due to van der Waals inte ractions between the adsorbate and the adsorbent. Van de r Waals interactions have a long range, but are weak. The energy

PAGE 73

59 released is on the same order of magnitude as the enthalpy of condensation, roughly 20 kJ/mol. These small energies are absorbed as vibrations of the lattice structure and dissipated as thermal motion. A molecule wi ll bounce around and fi nally adsorb to the surface in a process called accommodation. The bonds do not break; therefore a physisorbed molecule retains its identity (48). In chemical adsorption the molecules adhe re to the surface by forming a chemical bond, typically covalent. In comparison with physisorption, the enthalpy of chemisorption is ten times greater at appr oximately 200 kJ/mol. The distance between the surface and closest adsorbate atom is shorter for chemisorption than for physisorption. A chemisorbed molecule may be torn apart at the demand of unsatisfied valences of surface atoms. The existence of molecular fragments on the surface is one reason why solid surface s catalyze reactions. Chemisorption is most often exothermic which can be proven by examining the Gibbs equation, S T H G For chemisorption to be a spontaneous process, 0 G and because the translational freedom of the adsorbate is reduced when it is adsorbed, 0 S. Therefore, the enthalpy of adsorption, H must be negative, which indicates an exothermic nature. The enthalpy of adsorption, however, is dependent on the extent of surface coverage because the adsorbat e particles interact with each other. For example, if the particles repel each other like CO on palladium, adsorption becomes less exothermic as the coverage increases. If the particles attract each other, such as O2 on tungsten, the process becomes more exothe rmic as the particles cluster together.

PAGE 74

60 The rate and extent to which a surface is covered are important when considering heterogeneous reactions. Molecules will quick ly cover a surface exposed to a gas. The collision flux, wZ, can be expressed as, 2 12mkT p Zw (3.28) As an example, air with molecular weight of 29 g/mol at 1 atm and 25C, will have a collision flux of 3 x 1027 m-2s-1. So for a one meter square metal surface containing 1019 atoms, each atom is struck approximately 108 times each second. The fractional coverage, is given by, V V available sites adsorption of number occupied sites adsorption of number (3.29) where V is the volume of adsorbate correspondi ng to complete monolayer coverage. The rate of adsorption, dt d is determined by observing the change of fractional coverage over time. Surface Reaction Rate Expressions A number of classic rate expressions can be used to typify hete rogeneous reactions. These include adsorption isotherms, compe titive adsorption, and dissociative adsorption. Langmuir Adsorption Isotherm The free gas and adsorbed gas are in dynamic equilibrium where the fractional coverage of the surface depends on the pressure of the overlying gas. The variation of the fractional coverage with pressure at a specified temperature is known as the adsorption isotherm.

PAGE 75

61 The Langmuir adsorption isotherm desc ribes the equilibrium between a singlecomponent gas, A, and adsorbed species, A(s) at a surface. It is based on three assumptions (48): Adsorption cannot proceed beyond monolayer coverage. All sites are equivalent a nd the surface is uniform. Ability of a molecule to adsorb at a given site is independent of the occupation of neighboring sites. The isotherm expression rela tes the fraction of surface, A, covered by the adsorbed species as a function of partial pressure, pA, exposed to the surface and is given as follows, A A AKp Kp 1 (3.30) At low partial pressures, the coverage of adsorbed species increas es linearly with the partial pressure. As the partial pressure of A increases, the amount of adsorbed A(s) begins to saturate, and the coverage A approaches unity. The monolayer has thus been completed and further adsorption cannot take place. An equivalent expression for the isot herm can be developed using mass-action kinetics. For a gas molecule reacting with th e surface, the adsorption process proceeds as follows, ) ( ) (1 1s A s O Ak k (3.31) where O(s) represents an open site on the surface and k1 and k-1 are the equilibrium constants for the forward and reverse proces ses, respectively. At equilibrium the concentration of the surface adsorbed species is constant and can be represented by

PAGE 76

62 ) ( ) ( 0 ) (1 1s A k s O A k dt s A d (3.32) If the open site is related to the site density, by ) ( ) (s A s O (3.33) then at steady-state, ) (1 1 1s A A k k A k (3.34) The coverage A then becomes, A K A K A k k A k s Ac c A 1 ) (1 1 1 (3.35) Using Eq. 3.2 the coverage can be expre ssed in terms of pressure as o A p o A p Ap p K p p K 1 ) (. (3.36) Dissociative Adsorption Some molecular species undergo dissociati on upon adsorption, especially on metal surfaces. For example, H2 dissociates on a metal surface in to two surface adsorbed H(s) atoms. Likewise, methane dissociates into CH3(s) and H(s). Dissociative adsorption is assumed to require two open sites on the surface. The process of adsorption and dissociation are thought to o ccur in a single step. The site fraction is given by 2 1 2 11 Kp KpA (3.37) Mass-action kinetics provides a surface reaction of s A s O Ak k2 21 12 (3.38) and leads to the surface site frac tion on a concentration basis:

PAGE 77

63 2 1 2 1 2 1 2 12 1 2 1 ,1A K A Kc c A (3.39) Competitive Adsorption In competitive adsorption, two gases A and B are considered when modeling. Both gases are present above a surface and comp ete for available s ites on the surface for adsorption. The coverage of A and B are represented as follows B B A A B B B B B A A A A ABp K Ap K p K Bp K Ap K p K 1 1 (3.40) From mass-action kinetics the reaction rate expressions for A and B are s B s O B s A s O Ak k k k2 2 1 1 (3.41) The steady state analysis, similar to the deve lopment of the Langmuir isotherm, leads to an expression for the surface coverage for both A and B: B K A K B K B K A K A Kc c c B c c c A 2 1 2 2 1 1 ,1 1 (3.42) Adsorption of ammonia on the surface was assumed to occur by chemisorption. The Langmuir adsorption isotherm was used as the basis for the reaction mechanism. Competitive adsorption was important when c onsidering ammonia in humid air, where both ammonia and water were compet ing for available surface sites.

PAGE 78

64Proposed Mechanisms Due to the limited amount of testing th at could be done with the carbon dioxide sensor, the model focused only on reactions oc curring on the ammonia sensor. To detect ammonia a titanium nitride (TiN) film was used as the sensing layer. The change in ammonia concentration was detected by a chan ge in the work function of the sensing layer. Several hypotheses for the reaction m echanism, which causes the change in the work function, were considered. Ammonia and Hydroxide The first reaction mechanism considered involved ammonia and hydroxide (OH). Previous work by Ostrick et al. (44) indicated that ammoni a may bond to OH groups or OH-precovered sites on the sensing layer. Two mechanisms are possible, either the ammonia removes the OH from its binding sites or the ammonia binds directly to the OH groups already on the surface. The mechan ism proposed was based on the following: 1. Peaks seen in Diffuse Reflectance Infrar ed Fourier Transfor m Spectra (DRIFTspectra), Figure 3-1, 2. Sensitivity of the TiN sensing layer at room temperature versus at higher temperatures (120C), and 3. Experiments by Takagi-Kawai et al. (49) where a similar mechanism was proposed. Considering first the spectra of Figure 3-1, the peaks at 1450 cm-1 and between 3600 3800 cm-1 were attributed to OH groups, the deformation vibration and free or Hbridge-bonded valence vibration, respectively. Increases in signal were attributed to ammonia as follows: 1600 cm-1, asymmetric deformation band of NH3, 2700 2800 cm-1, (N-H) of ammonia ions, and 3000 3300 cm-1, NH stretching region.

PAGE 79

65 Figure 3-1 Diffuse Reflectance Infrared Fourier Transform Sp ectra (DRIFT-spectra) for a TiN screen-printed film, where absorbance was used to distinguish between the species. (Data of Ostrick et al. (44)). Next Ostrick et al. (44) indicated that because of the high affinity between ammonia and water, the small concentrati on of ammonia molecules (<100 ppm) will react entirely with the high concentration of surface water (>5000 ppm). Also, in the second reaction step, the ammonia ions may react directly with the sensing layer or with adsorbed OH ions. When the temperatur e rises, the amount of water on the surface decreases along with the solubility of amm onia in water. Resultantly, the amount of ammonia on the TiN surface is reduced and the sensitivity decreases. From Takagi-Kawai et al. (49) evidence of a reacti on between ammonia and hydroxide on a surface is depicted in Figure 3-2. Figure 3-2 Mechanism for the reduction of NO by NH3 over a V2O5 sensing layer in the presence of oxygen. (Data of Takagi Kawai et al. (49).)

PAGE 80

66 This mechanism was derived from in frared (IR) and x-ray photoelectron spectroscopy (XPS). The spectra from several surfaces were compared as seen in Table 3-1. Table 3-1 Infrared (IR) and X PS measurements and assignments for adsorbed species of ammonia on surface catalysts. Degene rate deformation and symmetric deformation are indicated by d and s, respectively. (Data of Takagi Kawai et al. (49).) IR (cm-1) XPS (ev) Surface sNH4 + dNH3 sNH3 sNH4 + dNH3 sNH3 V2O5 1413 ---------------400.9 -----------------V2O5/Al2O3 1410 1610 1275 401.0 400.2 400.2 Al2O3 ---------1610 1275 ---------400.0 400.0 V2O5/SiO2 1435 1620 --------401.0 ------------------V2O5/TiO2 1424 1605 1238 401.0 399.4 399.4 TiO2 ---------1605 1177 ---------399.6 399.6 NH4VO3 1410 ---------------400.9 -----------------By comparison, only one peak from Figure 3-1, at 1600 cm-1, matches up with those in Table 3-1. Shimanouchi (50) has al so reported this mode, degenerate deformation, with great certainty as 1627 cm-1. Ostrick et al. (44) indicated there was a peak between 2700 2800 cm-1, which was evidence of NH4 +. Since this peak does not match up with the values for sNH4 + in Table 3-1, another l ook suggests that the absorbance in this region is not large enough to warrant it as a peak. For example, the absorbance seen in the region between 2700 2800 cm-1 is no greater than at wave number 2000 cm-1, which was not considered by Ostrick et al. (44). With that in mind, NH4 + may not even be a product and therefore an other mechanism should be considered. Ammonia Dissociation The second mechanism considered was dissociation of ammonia on the TiN surface. Several studies examining ammoni a reactions on various surfaces have been done. The earliest studies evaluated the r eaction of ammonia on s ilicon surfaces. Later

PAGE 81

67 investigations examined ammonia on other tran sition metal surfaces such as titanium and silicon, titanium nitride a nd platinum, and nickel. Hlil et al. (51) used x-ray photoemissi on spectroscopy (XPS) and ultra-violet photoemission spectroscopy (UPS) to study chem isorption of ammonia on a Si(100) surface at substrate temperatures from 100 to 700 K. When the surface was exposed to ammonia at low temperatures ( 100 K), a line appeared at 400.0 eV binding energy, curve (a) in Figure 3-3, which was attributed to mole cular condensation on the substrate. At room temperature another line emerged at 398.5 eV, curve (b) in Figure 3-3, which correlates to NHX (X = 1,2). Additionally, the pr esence of weak silicon to hydrogen bonds evidences partial dissociat ion and co-adsorption of NHX radicals and atomic H. Figure 3-3 The XPS N(1 s ) core-level spectra for ammoni a over the Si(100) surface as a function of substrate temperatur e. (Data of Hlil et al. (51)). Bozso and Avouris (52) also studied reacti ons of two different silicon substrates with NH3 and atomic nitrogen using XPS and UPS. On a Si(100)-(2x1) surface at 100 K a peak at 400.1 eV, physisorbed molecular a mmonia, was visible. Another peak was evident at 398.5 eV, although not as clearly defined as the NH3 peak. As the temperature increased to 300 K, this peak became more de fined. Again this peak was most probably due to NH2 or NH. Upon further annealing to hi gher temperatures, the peak broadened to

PAGE 82

68 a lower binding energy indicating conversion of this species. When annealed to 950 K, a peak appeared at 397.7 eV, atomic nitrogen on the silicon surface, which resulted from complete dissociation of the NHX intermediate at 398.5 eV. Both the molecular ammonia (400.1 eV) and N bonded to silicon (397.7 eV) species were also seen on the Si(111)(7x7) surface. A surface species produced by dissociation was also apparent on this surface, but at a slightly hi gher binding energy of 398.8 eV. Figure 3-4 illustrates the transition from low to high temperature on both surfaces. (a) (b) Figure 3-4 The XPS N(1 s ) core-level spectra for amm onia on (a) Si(100)-(2x1) and (b) Si(111)-(7x7) surfaces as a function of substrate temperature. (Data of Bozso and Avouris (52).) Bischoff et al. (53) used ultra high va cuum (UHV) multilayer preparations, which combined the different species, to identify by XPS the nitrogen chemical environments in the Si/NH3 system. The assignment of the binding energies was found as follows: Nitride ) (3N Si 397.4 eV Imide ) (2H N Si 398.0 eV Amide ) (2H N Si 398.6 eV NH3 physisorbed 400.1 eV

PAGE 83

69 Zhou et al. (54) studied the decomposition of NH3 on Si(100) using static secondary ion mass spectroscopy (SSIMS). Th is procedure probed the surface directly and followed reaction intermediates in real time. At low temperatures, 100 K, adsorption of NH3 on Si(100) was dissociative and produced NH2(a) and H(a). Some thermal decomposition of NH2(a) to N(a) and H(a) occurred at 320 K, but most of the NH2(a) was stable up to 630 K. NH2(a) decomposed rapidly between 630-730 K, with no evidence of NH(a) at this temperature. Some NH2(a) recombined with H(a) at 685 K to liberate NH3(g). Temperature programmed desorption (TPD) was also completed and revealed that NH3(g) was desorbed at room temperature. Figure 3-5 illustrate s the surface species and desorption products found as the te mperature increased from 100 to 1000 K. Figure 3-5 Surface species and desorption produ cts from ammonia on Si(100). (Data of Zhou et al. (54).) Adsorption 100 K NH3(a, 2) NH3(a, 1) NH2(a), H(a) NH3(a, 1) NH2(a), H(a) NH2(a) H ( a ) NH2(a) H(a), N(a) H ( a ), N ( a ) Silicon nitride NH3(g) NH3(g) Negligible desor p tion NH3(g) ( small ) H2(g) 190 K 190 400 K 400 630 K 630 730 K 730 1000 K Surface Species Temperature Programmed Desorption (TPD) Products NH3(g) Heat

PAGE 84

70 Chen et al. (55) examined ammonia su rface chemistry on two different surfaces, Si(111) (7x7) and Si(100) (2x1) using a high re solution electron energy loss spectrometer (HREELS) and TPD. On th e Si(111) surface dissociative adsorption produced NH2(a) and H(a), and between 300 and 600 K further dissociation occurred to produce NH(a). On the Si(100) surface, the NH2(a) species remained thermally stable until approximately 600 K. The differing geomet ries of the two surfaces contributed to the species that were stable between 300 and 600 K. Figure 3-6 shows a ball and stick model of the two different surfaces, as well as the adsorption geometry of ammonia on the surface. (a) (b) Figure 3-6 Ball and stick models for the adsorption geometry for -NH2 and H on two different surfaces: (a) Si(111) (7x7) a nd (b) Si(100) (2x1). (Data of Chen et al. (55).) Siew et al. (56) used XPS to observe the adsorption and reaction of ammonia on a titanium/silicon surface (100). At low temp eratures, 120 K, three N1s peaks emerged: 397.8 398.1, 400.5 400.8, and 402.2 402.6 eV, attributed to NHX (X = 1 or 2), molecular NH3, and NH4 +, respectively. At room temperature more NH3 molecules desorbed and NH4 + disappeared indicating the species was not stable at room

PAGE 85

71 temperature. Some conversion of the NHX species occurred and the N diffused into the film. Figure 3-7 shows the spectra at 120 and 300 K. Figure 3-7 The XPS N 1s spectra of Ti/Si (100 ) surface at 120 and 300 K. (Data of Siew et al. (56)). Abom et al. (43) used titanium nitride as part of a sensing layer of a field effect metal-insulator-semiconductor device. The th ree different layers investigated for responses to ammonia and hydr ogen were: TiN, a double layer with platinum on top of TiN, and two-phase Pt TiN films formed by co-sputtering. The sensor response to ammonia, hydrogen, propene, and acetalde hyde was measured and it was found that devices containing platinum were responsive to all gases. Devices without platinum did not respond to any of the test gases. Abom et al. (43) indicated that the change in signal of the sensor was due to an interaction with hydrogen. It was assumed that only atomic hydrogen diffused through the film and that atomic hydrogen was only created when Pt was present. From the results it was apparent that the response to ammonia was due to dissociation of ammonia molecules and detection of atomic hydrogen. Laksono et al. (57) studied the adsorption of NH3 on clean and oxygen pre-treated nickel (111) surfaces at room temperature using XPS. Without oxygen on the nickel

PAGE 86

72 surface, no adsorption of ammonia was observed. Similarly, the surface reactivity was strongly linked to the presence of adsorbed oxygen; it increased with increasing adsorbed oxygen coverage. Two N adspecies were detected from N1s core level peak: 399.8 (molecular) and 397.8 (dissociated). Looking at the O1s core le vel peak revealed that at low ammonia exposures, the hydroxyl component increased, while the main feature at 529.9 eV decreased. Laksono et al. (57) found that the concurre nt transformation of both peaks indicates that hydrogen is removed from ammonia by the adsorbed oxygen to produce OH and NHX. Quantitative treatment of the XPS offered the following stoichiometry for the reaction: ) ( ) ( ) ( ) (2 3a OH a NH a O a NH Additionally, the experiments indi cated that the kinetics of desorption were faster than the kinetics of dissociation. Molecular Adsorption This last piece of evidence leads into the next mechanism, molecular adsorption of ammonia. While dissociation of amm onia was seen on many surfaces, it is possible that in the present study the rate of ammoni a desorption is faster than the rate of dissociation. Diebold and Madey (58) investigated adsorption and electron stimulated desorption of NH3 on TiO2(110) by XPS and low-energy io n scattering (LEIS). Three different surfaces were studied: (1) a stoich iometric surface, (2) a thermally treated, slightly oxygen deficient surface, and (3) a sputtered, highly oxygen deficient surface. Ammonia was seen to adsorb molecularly on all three surfaces and the saturation coverage was governed by repulsive interac tions between the molecules. Diebold and

PAGE 87

73 Madey (58) proposed a model in which ammonia adsorbs at titanium cation sites and where lateral repulsion between the ammoni a molecules along one-dimension limits the saturation coverage. For the stoichiometric surface, ammonia appeared to desorb in molecular form, whereas for the highly oxygen deficient surface electron stimulated dissociation of ammonia was seen. The fina l product of this dissociation process was atomic nitrogen. The previous experiment s were done at 160 K. On the highly oxygendeficient surface, heating the surface to 395 K desorbed all nitrogen-containing species from the surface, with no evidence of thermall y induced dissociation. Therefore, it was concluded that ammonia was adsorbed as an intact molecule. Karthigeyan et al. ( 59) studied an iridium oxide thin film integrated HSGFET which was selectively sensitive to ammonia at room temperature. An increase in sensor signal for ammonia at higher temperatures wa s seen, but the nature of the response to ammonia was unchanged from room temper ature to 100C. A possible reaction mechanism for ammonia on the sensing laye r was based on the sharing of lone donor bond to positively charged vacancies by chemisor ption. The sensor showed no response to concentrations of hydrogen up to 10,000 pp m. Additionally, the signal response was negligible to CO, SO2, Cl2, and NO2. Karthigeyan et al. (59) indicated that in the classical model, a charge transfer between adsorbed molecules and the surface of the sensitive layer and/or dipole moments of adsorbed molecules on the surface create a surface dipole layer which causes a work function change. A decrease in the work function upon ammonia adsorption indicates that there is a net electron transfer from ammonia to cation sites on the surface. The strong nature of the ammonia reaction is furthe r indicated by the eff ect of temperature on

PAGE 88

74 the reaction and recovery prope rties seen in the transient response of the sensor. The increase of the work function signal with temperature came from diffusion of the ammonia molecule into the film. Diffused molecules attached to cation sites lying underneath the surface. The desorption wa s complicated by residual gases in the measurement chamber and changes in bonding nature of adsorbed molecules before adsorption, i.e. surface diffusion to alternativ e bonding sites, lateral interactions between ammonia molecules, or imme diate compound formation due to ammonia dissociation. Reaction Kinetics Based on the preceding evidence, molecular adsorption of ammonia on the titanium nitride surface was chosen as the reaction mechanism for the model. Therefore the reaction mechanism was described as follows, ) ( ) (3 3s NH s O NHr fk k (3.43) where O(s) is an open surface site and ) (3s NH represents ammonia adsorbed on the surface. For the rate of the forward react ion or adsorption the forward and reverse reactions were considered together as, ) ( ) ( ) ( ) ( ) ( ) (3 3 . 3 3 33s NH k s NH S g NH k s NH k s O p k dt s NH dr s s f r NH f (3.44) Upon integration, ) ( 1 1 ) (3 ) ( . 33g NH k k e S s NHf r t k g NH k s sr f (3.45) where [ S ]s.s. is the total con centration of sites on the surface where the ammonia can adsorb at steady-state. This is different than total site density, because [ S ]s.s. varies

PAGE 89

75 with the free stream gas concentration. For example, the number of sites the analyte occupied was less at a lower analyte free str eam concentration than at a higher analyte concentration. For the reverse reaction the analyte was removed by forced flow rather than by diffusion thereby quickly forcing the ammoni a out of the free stream, which caused the first term on the right hand side of Eq. 3.44 to be negligible compared with the reverse reaction. Therefore, the rate of th e reverse reaction or desorption is ) ( ) (3 3s NH k dt s NH dr (3.46) Integration leads to t k S s NHr s s exp ) (. 3. (3.47) At equilibrium, 3 31 ) (. 3 NH NH s sKp Kp S s NH (3.48) where K=kf /kr. Also at equilibrium, RT H R S RT G K / exp / exp / exp (3.49) The rate equations were made dimensionles s to allow them to be used in other applications where the physical parameters, such as the area of the sensing film, may be different. The dimensionless variables were as follows, . 3) (s sS s NH (3.50) and L W Dt (3.51)

PAGE 90

76 where D is the diffusion coefficient and W and L are the width and length of the sensitive film. Equation 3.44 in dimensionless form now becomes, D L W k g NH k d dr f 1 ) (3. (3.52) As noted above, molecular adsorption was the focus of this study, however since dissociation was not ruled out as a possible reaction mechanism, the rate equations for dissociation were formulated and integrat ed. The reaction mechanism for ammonia dissociation on a titanium nitr ide surface was as follows, ) ( ) ( ) ( 22 3s H s NH s O NHr fk k (3.53) The dissociative adsorption was assumed to require two open surface sites and the adsorption and breaking-apart of the molecule were taken to occur in a single step. Assuming that ) (2ad NH and ) ( ad H adsorb and desorb at the same rate on the surface, Eq. 3.53 was simplified as follows ) ( 2 ) ( 22s A s O Ar fk k (3.54) where A2 represents 3NH and 2 A signifies the two surface sites that ) (2s NH and ) ( s H occupied. The rate of the forward reaction was given as 2 2 . 2 2 2 2) ( ) ( ) ( ) ( ) ( s A k s A S A k s A k s O A k dt s A dr s s f r f (3.55) Integrating the rate equation a nd applying the initial condition,0 0 s A t, provides, 2 . 2 1 2 . 2 1 2tanh ) ( A k k t S A k k S A k k s Af r s s r f s s r f (3.56)

PAGE 91

77 The rate of the reverse reaction or deso rption was considered without the forward reaction because the analyte was removed by fo rced flow rather than by diffusion thereby preventing the particles from adsorbing to the surface again. Hence the gas phase concentration2A was assumed to be negligible, yielding Eq. 3.57 from Eq. 3.55. The rate of the reverse r eaction or desorption was 2s A k dt s A dr (3.57) Integrating using th e initial condition, ., 0s sS s A t led to .1 1 ) (s s rS t k s A (3.58) Summary The foundation for the theoretical modeling of the ammonia sensor was laid in this chapter by first presenting the relationship between thermodynamic properties and reaction kinetics. Then adsorption theory and surface reaction rate expressions were discussed, followed by a literature review of mechanisms that led to the choosing of molecular adsorption as the mechanism for the model. Finally, the reaction kinetics for molecular adsorption were mathematically form ulated to be used with the experimental results presented in the next chapter.

PAGE 92

78 CHAPTER 4 RESULTS AND DISCUSSION The results from the experimental porti on of this study are presented in this chapter. Throughout the course of this proj ect, the sensor develope rs were continuously upgrading the performance of the field effect transistors; resu ltantly, three generations of sensors were used in testing. The latest versi ons were always utilized in the experiments. For the ammonia sensor, first presented are the response times fitted to the model to obtain rate information. Then, the performan ce and feasibility of the sensor for use in a laboratory animal environment are discussed. Detector reliability and sensitivity limited the useful data obtained from the carbon dioxi de sensor. Consequently, for the carbon dioxide sensor, only a calibration curve, response times of the sensor, and crosssensitivity to humidity are reported. Experiments for modeling purposes were not completed with the carbon dioxide sensor. Drift Tests Prior to conducting any tests with the analyte gases, the carbon dioxide and ammonia sensors were tested for drift of th e baseline signal. The sensors were exposed to air at a flow rate of 2 L/min for over 40 hour s. The results of these tests can be viewed in Figure 4-1. Each sensor had two channels of output offering the same signal with different amplification factors. Table 4-1 lists the rates of the signal drift. The carbon dioxide sensor had a negative drift rate, while the ammonia sensor drift rate was positive. The magnitude of the drift rate for the carbon di oxide sensor was slightly larger than for

PAGE 93

79 the ammonia sensor. For both sensors the drif t rate was much higher in the first five hours than for the rest of the test, which show s that the stability of the baseline increased with time. Therefore, for the following tests the sensor was turned on at least one hour before each experiment to help stabil ize the baseline signal. 010203040506070 1.25 1.50 1.75 2.00 Ch A Ch BSignal (V)time (hr) 01020304050 3.4 3.6 3.8 4.0 4.2 4.4 4.6 4.8 Ch A Ch BSignal (V)time (hr) (a) (b) Figure 4-1 Long-term drift te st results (sample size, N = 1) for the (a) carbon dioxide and (b) ammonia sensors. Signal ou tput recorded every 30 minutes. Table 4-1 Rates of baseline signal drift for the carbon dioxide and ammonia sensor exposed to air at 2 L/min for over 40 hours. Slope (mV/hr) Sensor Channel A Channel B Carbon dioxide -7.3 -5.5 Ammonia 1.8 2.6 Ammonia Sensor Results For the ammonia sensor, first presented are the time response tests fitted to the model to obtain rate constants, the equilibri um constant, and Gibbs free energy. Second, the role of diffusion is discussed along with the presentation of tim e response results in the diffusion box. Third, results from test s for temperature depende ncy and the resulting Arrhenius parameters are presented. Fina lly, the performance and feasibility of the sensor for use in a laboratory animal environment are discussed.

PAGE 94

80Sensor Response and Mechanism The reaction mechanism selected for th e ammonia sensor model was molecular adsorption of ammonia on the titanium nitride surface, ) ( ) (3 3s NH s O NHr fk k (4.1) where O(s) is an open surface site and ) (3s NH represents ammonia adsorbed on the surface. The Langmuir adsorption isotherm wa s used when developing the mechanism. In conjunction with the Langmuir m odel, the assumptions made were Adsorption does not proceed beyond single layer, All sites are equivalent a nd the surface is uniform, Ability of a molecule to adsorb at a given site is independent of the occupation of neighboring sites. The forward and reverse reactions were considered together to determine the rate of the forward reaction, ) ( ) ( ) ( ) ( ) ( ) (3 3 . 3 3 33s NH k s NH S g NH k s NH k s O p k dt s NH dr s s f r NH f (4.2) where kf and kr are the forward and reverse rate constants, respectively, ) (3s NH and ) (3g NH are the concentrations of ammonia on the surface and in the gas phase, respectively and [ S ]s.s. is the total concentration of sites on the surface where the ammonia can adsorb at steady-state. This is different than total site density, because [ S ]s.s. varies with the free stream gas concentrat ion. For example, the number of sites the analyte occupied was less at a lower analyte free stream concentration than at a higher analyte concentration. Solving Eq. 4.2 by integration using the initial condition: t = 0, ) (3s NH =0, led to the following,

PAGE 95

81 ) ( 1 1 ) (3 ) ( . 33g NH k k e S s NHf r t k g NH k s sr f (4.3) The rate of the reverse reaction or deso rption was considered without the forward reaction because the analyte was removed by forced flow rather than by diffusion thereby, preventing the particles from adso rbing to the surface again. Hence the gas phase concentration,) (3g NH was assumed to be negligible, yielding Eq. 4.4 from Eq. 4.2. The rate of the revers e reaction or de sorption is ) ( ) (3 3s NH k dt s NH dr (4.4) Integrating using the in itial condition: t = 0, . 3) (s sS ad NH provided the following t k S s NHr s s exp ) (. 3. (4.5) To solve for the rate constants, the dime nsionless concentration, or the fractional surface coverage, was used, . 3 s sS s NH (4.6) To experimentally determine the sensor was exposed to air for 10 minutes, then 100 ppm ammonia for 10 minutes, and then air ag ain for 10 minutes. The surface coverage for the forward reaction was cal culated by dividing all data points from when gas was introduced into the system until it was removed by the maximum voltage signal. Hence the signal was normalized, yielding a scale fr om 0 to 1, with 1 being maximum surface coverage. Similarly for the reverse reaction, was determined by dividing all data points from when gas was removed from the system by the maximum voltage signal.

PAGE 96

82 The rate constants for the forward and reverse reactions were solved by curve fitting the experimental data. The reverse reaction rate constant was analyzed first because the forward reaction is a function of the reverse rate constant. For the reverse reaction the surface coverage, was plotted versus time and the following equation was used for the curve fit, btae (4.7) where the fitting parameter b yields the rate constant, rk b For the normalized data, ideally a = 1. This was typically the nominal value, (see Table 4-2) altho ugh the curve fit was not forced to a pre-exponential of unity. The desorption data w ith the curve fit can be seen in Figure 4-2. Based on the curve fit, the reverse reaction rate constant, kr, is 6.43 x 10-3 s-1. The coefficients, R-squared values, and rate constants for both adsorption and desorption are listed in Table 4-2. 0100200300400500600 0.0 0.2 0.4 0.6 0.8 1.0 Surface coverage, time ( s ) Figure 4-2 Average desorption data and curve fit for 100 ppm desorption tests. The error bars represent standard deviation at selected data points (N=5). For the forward reaction, was also plotted versus time with the following equation used for the curve fit,

PAGE 97

83 bte a 1. (4.8) where again a is a pre-exponential constant and r fk g NH k b 3. The curve fit for the adsorption data can be seen in Figure 4-3. The constant b was used to solve for kf as follows, s mol L gNH mol mgNH gNH ppm L mgNH ppm s g NH k b kr f / 28 6 17 1 1000 1 1 1 100 00643 0 04335 03 3 3 3 1 3 (4.9) 020406080100120140 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Surface coverage, time (s) Figure 4-3 Average adsorption data and curve fit for 100 ppm adsorption tests. The error bars represent standard deviation for 6 runs at selected data points. Table 4-2 Curve fit coefficients, R-squared va lues, and rate consta nts for desorption and adsorption of 100 ppm ammonia on sensor. a b R2 k Desorption 0.915 -0.00643 0.96 6.43 x 10-3 s-1 Adsorption 1.01 0.04335 0.99 6.28 L/mol*s Time response experiments were also c onducted at 50 ppm, where the sensor was exposed to ammonia for 10 minut es with 10 minutes of air be fore and after the exposure. The rate constants calculated us ing 100 ppm data were then us ed to predict the adsorption

PAGE 98

84 and desorption curves at 50 ppm ammonia. The actual and predicted curves for desorption and adsorption are depicted in Figure 4-4 and Figure 4-5, respectively. Percent error between the actual and predic ted curves was calculated as follows, % 100 % redicted P redicted P Actual Error (4.10) For the desorption curve, the average percent error was 27%, where the error significantly increased as the predicted curve approached zero, which was the limit of the exponential curve fit. Neglecti ng the error at the limit, the av erage percent error from 0 450 seconds was 10%. The percent error for the adsorption curve was 38%, which is expected when comparing the two curves in Figure 4-5. The actual curve for 50 ppm adsorption has large standard deviations at the representative points, which may contribute to the large error seen between the ac tual and predicted curves It is noted that the detection limit for the ammonia sensor is roughly 100 ppm (see be low); therefore the 50 ppm data must be considered in this context. 0100200300400500600 0.0 0.2 0.4 0.6 0.8 1.0 Actual PredictedSurface coverage, time (s) Figure 4-4 Actual 50 ppm average ( N =5) desorption curve pl otted along with the predicted curve based on th e rate constant determined from 100 ppm curves. Error bars represent standard de viation for selected data points.

PAGE 99

85 020406080100120140160180 0.0 0.2 0.4 0.6 0.8 1.0 Actual PredictedSurface coverage, time (s) Figure 4-5 Actual 50 ppm average ( N= 6) adsorption curve pl otted along with the predicted curve based on the rate cons tants determined from 100 ppm curves. Error bars represent standard de viation for selected data points. Once kr and kf were determined, the equilibrium concentration constant, Kc, was determined from the following, mol L s s mol L k k Kr f c/ 7 976 10 43 6 / 28 61 3 (4.11) The equilibrium constant, K can be calculated from the equilibrium concentration constant by, 04 39 1 1 304 1 315 8 1 10 01 1 1000 1 976 12 5 3 m N J K J K mol bar m N bars L m mol L RT p K p RT K p K Ko c o c o p (4.12)

PAGE 100

86 where po is standard pressure and = -1 for molecular adsorption. Also at equilibrium, RT G Kr/ exp (4.13) where Gr is the standard Gibbs energy of reaction, R is the ideal gas constant, and T is the surface temperature of the sensor. Therefore, Grwas calculated as follows, mol kJ K K mol J K RT Gr26 9 04 39 ln 304 31 8 ln (4.14) In comparison, the standard Gibbs ener gy of reaction for ethylene hydrogenation on platinum is -148.2 kJ/mol and for ammoni a synthesis on iron it is 62.0 kJ/mol (60). The Gibbs energy of reaction predicts the direction of spontaneous change of a reaction at constant temperatur e and pressure. The reaction Gi bbs energy is also defined as the slope of the graph of the Gibbs ener gy plotted against the extent of the reaction, : T p rG G, (4.15) and it is further derived as the diffe rence between the chemical potentials, of the reactants and products at the com position of the reaction mixture, ants react products rG (4.16) Due to the fact that the chem ical potentials vary with th e composition, the slope of the plot of Gibbs energy against extent of r eaction varies as the reaction continues (Figure 46). Further, because the reaction pr oceeds in the direc tion of decreasing G it is apparent that the forward reacti on is spontaneous when products ants react or Gr<0. The reverse reaction is spontaneous when ants react products Gr >0, whereas the system is at equilibrium when ants react products Gr =0. Since Gr < 0 for the reaction of

PAGE 101

87 ammonia on the titanium nitride surface, th e forward reaction, a mmonia adsorption, is spontaneous. rG<0rG>0rG=0 Gibbs energy, GExtent of reaction, Figure 4-6 The slope of the Gibbs energy cha nges as the reaction proceeds. The arrows indicate the direction in which the reaction spontaneously proceeds. The reaction is at equilibrium when the curve is at its minimum. Diffusion Time response tests were also completed in the diffusion box to aid in understanding the role diffusion played in the sensors response to ammonia. In the forced flow case, the ammonia in the gas pha se was a constant because the distance from the surface to the free stream was not large enough to create a space or time gradient. However, the diffusion case can be modeled as a semi-infinite medium, as illustrated in Figure 4-7, which allows the gas phase ammonia to be a function of x and t The mathematical formulation of th e gas phase ammonia is given as 0 0 ) ( ) (3 t x in t x f g NH (4.17) with the following boundary conditions: 0 ) (3 3 t x as NH g NHin, (4.18) 0 0 ) ( ) ( ] [ ) (3 3 3 t x at ad NH k x g NH t S k t ad NHr f. (4.19)

PAGE 102

88 The second boundary condition, at0 x, describes the overall rate of the surface reaction described in Eq. 4.1, for the diffusion case. The rate of the overall reaction now becomes a partial derivative due to the presence of the space and time gradients with respect to the ammonia gas. The solution cannot be obtained analytically due to th e coupled nature of the equation; numerical methods such as finite difference must be employed to solve this problem. The solution of Eq. 4.1 was not within the scope of this investigation, but was considered because the main mode of mass tr ansfer in a laboratory animal cage will be diffusion. 0 x ) ( ) ( ] [ ) ( 3 3 3ad NH k x g NH t S k t ad NH r f in NH g NH 3 3) ( Figure 4-7 Diffusion of ammoni a in a semi-infinite region. To determine the time response for the diffusion case, the sensor was exposed to 100 ppm ammonia for 10 minutes, with 10 minutes of air before and after exposure. The desorption and adsorption results are plot ted along with the forced curves in Figure 4-8 and Figure 4-9, respectively. The time constants for the diffusion case are listed in Table 4-3. As is evident from Figure 4-8 and Figure 4-9, the diffusion time constants were much greater than for the

PAGE 103

89 forced flow regime. For desorption, the per cent difference in time constants relative to the forced condition was 27%. In the case of adsorption, the percent difference was 67%. 0100200300400500600 0.0 0.2 0.4 0.6 0.8 1.0 Diffusion ForcedSurface coverage, time ( s ) Figure 4-8 Average ( N= 6) desorption curve for the diffusion case compared with the forced curve fit for 100 ppm data. Error bars represent standard deviation for select data points. 050100150200250300 0.0 0.2 0.4 0.6 0.8 1.0 Diffusion ForcedSurface coverage, time ( s ) Figure 4-9 Average ( N= 6) adsorption curve for the diffusion case compared with the forced curve fit for 100 ppm data. Error bars represent standard deviation for select data points. Temperature Effects The sensor was exposed to ammonia for 10 minutes at 31C (1.0 V), 39C (2.0 V) and 52C (3.0 V), where the voltage corresponde d to a calibrated temperature voltage

PAGE 104

90 curve and was altered by a DC power s upply. The adsorption and desorption time response plots for 31C (1.0 V) were shown previously in Figure 4-4 and Figure 4-5, respectively. The rate cons tants calculated from 100 ppm da ta were established as the accepted rate constants at 1.0 V. Table 4-3 Adsorption and desorption time cons tants for the ammonia sensor using forced and diffusion flow regimes. Reaction Flow a b R2 Time Constant, = 1/b (s) Adsorption Forced 1.01 0.043 0.9923.3 Diffusion 1.02 0.014 0.9571.4 Desorption Forced 0.9150.006430.96156 Diffusion 1.00 0.004650.98215 The results and respective curve fits for 39C (2.0 V) and 52C (3.0 V) data are plotted in Figure 4-10. Rate constants for 39C (2.0 V) and 52C (3.0 V) data were determined from curve fits, where the equations used for curve fitting desorption and adsorption plots were btae and bte a 1, respectively. The curve fit coefficients, R-squared values, rate constants, and equilibriu m concentration constants for all three temperatures are compiled in Table 4-4. For both the forward and reverse reactions the rate constants decreased with increasing temperature. The equilibrium concentration constant increased with in creasing temperature. From the Langmuir adsorption model this result is counterintuitiv e. The desorption rate is expected to increase with increasing temperature. In the experiments, was defined as . 3 s sS s NH when to compare temperature eff ects the fractional surface coverage should really be max 3S s NH where maxSis the absolute number of sites where ammonia can adsorb for a particular isotherm. If the latter fractional surface coverage is used, a result which follows th e Langmuir model may be found.

PAGE 105

91 0100200300400500600 0.0 0.2 0.4 0.6 0.8 1.0 Surface coverage, time (s) (a) 050100150200250300 0.0 0.2 0.4 0.6 0.8 1.0 Surface coverage, time ( s ) (b) 0100200300400500600 0.0 0.2 0.4 0.6 0.8 1.0 Surface coverage, time (s)(c) 050100150200250300350400 0.0 0.2 0.4 0.6 0.8 1.0 Surface coverage, time ( s ) (d) Figure 4-10 Average desorption and adsorpti on curves at 50 ppm ammonia for heater voltages of (a) (b) 2 V ( N= 10) and (c) (d) 3 V ( N= 7 and N= 8). Error bars represent standard deviat ion for select data points. Table 4-4 Curve fit coefficients, R-squared va lues, and rate consta nts for desorption and adsorption of ammonia on sensor at varying surface temperatures. Reaction Temp. (K) Heater Voltage (V) a b R2 kf (L/mol*s) kr (s-1) Kc (L/mol) 304 1.0 0.915 -0.00643 0.96 6.43 x 10-3 976 312 2.0 0.925 -0.00521 0.97 5.21 x 10-3 1032 Desorption 325 3.0 0.974 -0.00378 0.97 3.78 x 10-3 1132 304 1.0 1.01 0.04335 0.99 6.28 312 2.0 0.984 0.02104 0.99 5.38 Adsorption 325 3.0 0.993 0.01638 0.98 4.28 Sensor Performance Single analyte tests, where air and ammonia were the only gases in the flow stream, were conducted on the ammonia se nsor to establish a calibra tion curve. Seeing as the ammonia threshold limit value for a laborator y animal cage is 25 ppm, the sensors were

PAGE 106

92 tested at concentrations of 25, 50, 75, and 100 ppm. The sensor was exposed to the set concentration for 10 to 20 minutes with a mi nimum of 10 minutes of air between analyte exposures. For the forced flow case, 10 mi nutes of purging with air was a sufficient amount of time to remove ammonia from the system. Figure 4-11 shows a sample of single analyt e tests with six exposures to 100 ppm and three exposures to 50 ppm a mmonia. From the raw signal, Figure 4-11(a), it was difficult to determine the magnitude of the re sponse due to the significant drift of the baseline signal. To correct for the baseline drift, the average of 30 data points before gas introduction into the system and 30 data point s after gas removal were used to determine a slope for baseline drift. Then the slope was subtracted from ev ery data point along the curve to give the baseline corrected data in Figure 4-11(b). From Figure 4-11(a), it is evident that the sensor responds to ammonia, but Figure 4-11(b) illustrates that the magnitude of the signal was not consistent. For the 100-ppm pulses the response varied from 20 to 50 mV. Additionally, four of th e 100-ppm pulses were of the same magnitude as the 50-ppm pulses. The calibration curve for all of the si ngle analyte tests can be seen in Figure 4-12. The average response and respective standard deviation to exposures of 25, 50, 75, and 100 ppm along with a linear fit is plotted. Th e linear fit has an R-squared value of 0.85. Although the R-squared value is statistically high, the error bars for the 75 and 100-ppm data are greater than the slope of the linear fit. For this concentration range, the sensor cannot accurately predict the ammonia concentration. From Figure 4-12 a level of detection (LOD) was calculated in order to determine if the testing was completed in the sensors operating range. The L OD was calculated as follows,

PAGE 107

93 m LOD 2 (4.20) where is the standard deviation and m is the slope of the linear curve fit. With a standard deviation of 0.005 and a slope of 7.78 x 10-5, the level of detection for the ammonia sensor was found to be approximately 100 ppm. 020406080100120140160180200 3.56 3.60 3.64 3.68 Signal (V)time (min)0 25 50 75 100 NH3 [gas] (ppm) 020406080100120140160180200 0.00 0.02 0.04 0.06 Signal (V)time (min)0 25 50 75 100 NH3 [gas] (ppm) (a) (b) Figure 4-11 Single anal yte test results ( N= 1) for the ammonia sensor: (a) raw data and (b) baseline corrected. 0255075100 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 S ignal ( V ) [NH3] (ppm) Mean Linear Fit Figure 4-12 Calibration curve and linear fit for ammonia sensor. The error bars represent one standard deviation for two to six runs.

PAGE 108

94 Gradual ramping tests While the single analyte tests in forced flow were useful to obtain response data, in actual operation the sensor will not have s udden increases and decreases in ammonia concentration. Therefore, to imitate the gr adual changes seen in the cage environment tests were conducted where the ammonia conc entration was ramped up or down at fiveminute intervals. The gradual ramping tests completed in the diffusion box are depicted in Figure 413. Figure 4-13(a) illustrates the response to 200 ppm ammo nia, followed by a gradual decrease to zero concentration, and then back to 200 ppm. Figure 4-13(b) depicts a gradual ramp to 200 ppm and then a gradual decrease back to 0 ppm. Local maximums and minimums are labeled on the curves in Figure 4-13 and the slopes for the signal change over concentration change for each segment are tabulated in Table 4-5. Segments A B and E F, where the concentrations we re increasing, have comparable slopes, 0.25 and 0.23 mV/ppm, respectively. The slope for se gment C D was expected to be similar to the aforementioned slopes, however its sl ope of 0.15 mV/ppm was closer to the slopes for the decreasing ramps. The decreasing ramps, segments B C and F G, had slopes of 0.11 and 0.13 mV/ppm, respectively. These two smaller slopes indicate that the ammonia may not have completely diffuse d out of the box to the desired input concentration. Therefore when more ammoni a was added it increased the concentration that was already in the box. To better estima te the concentration at point C, the voltage difference from B C was divided by the slope from A B. This resulted in a difference of 80 ppm from B C or a tota l concentration of 120 ppm at point C. The same theory was applied to better estimate the concentrati on at point D. The change in concentration from C D was calculated as 116 ppm and addi ng it to 120 ppm from point C resulted in

PAGE 109

95 a concentration of 236 ppm at point D. Since the segment C D directly followed a decrease in concentration the slope may have been affected by residual ammonia left in the system. 020406080100120140 0.00 0.01 0.02 0.03 0.04 0.05 0.06 D C B A Signal (V)time (min)0 50 100 150 200 NH3 [gas] (ppm) 020406080100120140 -0.01 0.00 0.01 0.02 0.03 0.04 G F E Signal (V)time (min)0 50 100 150 200 NH3 [gas] (ppm) (a) (b) Figure 4-13 Ammonia concentration ramped up and down with ammonia sensor in diffusion box (a) average ( N= 6) ramp down and (b) average ( N= 6) ramp up. The local maximums and minimums are labeled on each curve as (A) (G). Further proof of the slow diffusion process is that there was a five-minute lag time at lower concentrations in both Figure 4-13(a) and Figure 4-13(b). The five-minute lag is more apparent in Figure 4-13(b) where the sensor does not respond to ammonia until 20 minutes of testing when the concentrati on is increased from 5 to 10 ppm. Two conclusions can be drawn from these result s; either the sensor does not respond to concentrations less than 10 ppm or the e xposure time of 10 minutes was not long enough for the sensor to respond to the low concentra tion. Given the LOD, it is concluded that 20 ppm is not a valid level statistically. Ba sed on earlier tests, 10 minutes should be sufficient time. Baseline drift was significant for both tests. If an observer did not know what the concentration of ammonia was, the changes in voltage could be from 1) changes in

PAGE 110

96 ammonia concentration, 2) baseline drift, or 3) changes in relative humidity, as will be shown next. Table 4-5 Slopes for ramp down and ramp up curves completed in diffusion box. Signal (mV) [NH 3 ] (ppm) Graph Segment 1 2 1 2 Slope (mV/ppm) A B 0 47.8 0 195 0.25 B C 47.8 27.5 195 5 0.11 C D 27.5 56.6 5 195 0.15 E F -8.8 35.4 5 195 0.23 F G 35.4 10.3 195 0 0.13 Cross-sensitivity to humi dity and carbon dioxide The ammonia sensor was tested for cr oss-sensitivity to humidity through two different experiments. In the first set of experiments the sensor was exposed to humidified air in the absence of ammonia. The humidity was increased to 40 60% rh three times for 20 minutes each with 20 minutes of dry air between each cycle. For the second set of experiments, the ammonia was increased to 50 or 100 ppm ammonia in dry air, followed by the same cycle in humidified air. The results of both experiments are illustrated in Figure 4-14 and Figure 4-15, respectively. 0102030405060708090 -0.04 -0.02 0.00 0.02 0.04 signal (V)time ( min ) 0 25 50 75 100 RH (%) (a) (b) Figure 4-14 Ammonia sensor tested for crosssensitivity to humidity (a) six individual tests and (b) average of 6 runs. Test s were completed in the absence of ammonia.

PAGE 111

97Figure 4-14(a) shows the results of all six tests completed in the absence of ammonia. The average response is shown in Figure 4-14(b). From th ese six tests it is evident that there was no correlation between signal response and changes in relative humidity. Any changes in signal were from se nsor instability and drift of the baseline. 0306090120150 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 signal (V)time (min)0 25 50 75 100 [NH3] rh [gas] (ppm) rh (%) 0306090120150 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08 signal (V)time (min)0 25 50 75 100 [NH3] rh [gas] (ppm) rh(%) Figure 4-15 Ammonia sensor tested for cross-sensitivity to humidity with ammonia in the system at two different concentration levels: (a) 50 ppm and (b) 100 ppm. Figure 4-15(a) and Figure 4-15(b) depict the cross-sensitivity to humidity at 50 and 100 ppm ammonia, respectively. The magnitudes of the sensor response to changes seen in Figure 4-15 are listed in Table 4-6. A response to changes in ammonia for both dry and humid air was seen. The magnitudes for adsorption varied from 12 to 60 mV. For desorption, the magnitudes varied from -13 to -55 mV. The magnitude of the response was not dependent on the relative humidity of the air. The sensor also responded to an increase in relative humidity with a decrea se in signal of to mV. This response was similar in magnitude and directio n to the ammonia desorption response. Since the sensor responded to relative humidity only when ammonia was recently in the system, the change in signal could be attributed to the water molecules removing residual ammonia from the surface. This competition between water and ammonia for available surface sites provides further evidence to rule out the ammonia and hydroxide

PAGE 112

98 mechanism where the water acted as the ve hicle for the sensor to respond to the ammonia. Notably, above 10% the sensor doe s not respond with the same magnitude as when the humidity increases from 2 to 50%. The relative humidity operating range in a laboratory animal cage is between 30 to 70% (18). Therefore the work function change seen when going from dry to humid air s hould not hinder the sensor in detecting ammonia. Table 4-6 Magnitude of signa l response for changes in ammonia concentration and relative humidity. Description Signal Change (mV) 0 to 50 ppm, dry air (2% rh) +12 0 to 50 ppm, humid air (50% rh) +32 0 to 100 ppm, dry air +60 0 to 100 ppm, humid air +36 100 to 0 ppm, dry air -43 100 to 0 ppm, humid air -37 50 to 0 ppm, dry air -13 50 to 0 ppm, humid air -55 0 to 60% rh with previous exposure to 50 ppm NH3 -30 0 to 80% rh with previo us exposure to 100 ppm NH3-40 Next the ammonia sensor was tested for cross-sensitivities to humidity and carbon dioxide with ammonia in the system. The introduction of each new parameter was at least 10 minutes after the last ch ange in experimental conditions to ensure that the effect on the sensor was from the intended parameter. Figure 4-16 illustrates the average results of this test. The sensor re sponded to ammonia similarly to previous tests in dry air. Upon the introduction of carbon dioxide, at 35 minutes, the signal experienced a sudden spike. At 90 minutes, when the car bon dioxide increased fr om 3000 to 5000 ppm, the signal again spiked. No change in re sponse was apparent when the carbon dioxide concentration decreased at 150 minutes. The sensor responded to humidity as previously

PAGE 113

99 discussed, where the signal decreased when the humidity went from dry to humid. Additionally, the signal increased when the air went from humid to dry, which further supports the theory that water and ammonia ar e competing for sites on the sensor surface. Carbon Dioxide Sensor Results Single analyte (i.e. carbon di oxide in dry air), time re sponse, and cross-sensitivity to humidity tests were completed for the ca rbon dioxide sensor. For the single analyte tests, at each concentration level the sensor was exposed to the specified carbon dioxide concentration for 10 to 30 minutes and the si gnal was collected every 10 seconds. Raw data for one of the single analyte tests, where the sensor was exposed to 300, 1000, 2500, and 5000 ppm of carbon dioxide, can be seen in Figure 4-17. When creating the calibration curve, the baseline of each signal was set to 1.00 thereby normalizing the curve to ensure that the signal responses were comparable. To eliminate the transient response from the calib ration, the first two minutes of data were not used when calculating the averages shown in Figure 4-18. The carbon dioxide sensor responded to ch anges in gas concentration with a decrease in voltage. For concentrations of 300 and 1000 ppm, the signal changed by 60 and 74 millivolts (mV), respectively. There wa s a 130 mV change in signal for a change in gas concentration at 2500 ppm. The ch ange in signal at 30 00 ppm was 90 mV, which was less than the response at 2500 ppm. Thes e data were taken from a different channel than the others and consequently the amplifica tion of the signal was slightly less than the others. At 5000 and 7100 ppm, the signal changed by 150 and 200 mV, respectively. A linear fit of the data was conducte d with an R-squared value of 0.88.

PAGE 114

100 020406080100120140160 0.00 0.01 0.02 0.03 0.04 Average signal (V)time (min)10 100 1000 [CO2] [NH3][gas] (ppm) 0 25 50 75 100 rh (%) rh Figure 4-16 Average (N = 6) ammonia sensor response to humidity, carbon dioxide, and a mmonia concurrently. Error bars represen t standard deviation fo r select data points.

PAGE 115

101 0306090120150180 0.4 0.5 0.6 0.7 0.8 0.9 signal (V)time (min)100 1000 10000 [CO2][gas] (ppm) Figure 4-17 Raw data ( N= 1) from carbon dioxide single anal yte test, where baseline drift has not been eliminated. Signal was collected every 10 seconds. 010002000300040005000600070008000 0.72 0.76 0.80 0.84 0.88 0.92 0.96 1.00 Normalized signal (V)[CO2] (ppm) Mean Linear Fit Figure 4-18 Calibration curve for carbon dioxide sensor. The error bars represent one standard deviation. Figure 4-19 depicts the transi ent response of the sensor to introduction and removal of gas from the flow stream. Each curve is an average of four te sts with the standard deviation plotted at several points along th e curve. For each test the baseline was established for 30 minutes by running dry ai r through the system, then the sensor was exposed to 3000 ppm of carbon dioxide for 30 minutes, and finally the system was

PAGE 116

102 flushed out with dry air for 15 minutes. The curves have been normalized on a scale from zero to 1.00, where zero is the baseline and 1.00 is the maximum change in signal for the sensor. For the introduction of gas, or adsorption process, the sensor reaches 95% of the maximum in 16.4 minutes. When th e carbon dioxide is re moved, or desorption process, the sensor returns to 5% of the baseline signal in 8 minutes. A similar time response test was completed in a humidified air stream and the results are shown in Figure 4-19(c)-(d). The average response of the sensor to the introduction of gas into the humidified ai r stream reaches a maximum of 90% in 15 minutes. Each individual curve reaches 1.00 at some point, but even after 30 minutes, the average does not stabilize to the normalized maximum. For the desorption process, the sensor reaches 5% of the baseline signal in 9.5 minutes. A curve fit was conducted for each of the plots in Figure 4-19. The equation used to curve fit the adsorption plots was bte a 1, where is the normalized signal and 1/ b is the time constant, For the desorption plots, the equation btae was used. The time constants for the adsorption and desorption graphs are listed in Table 4-7 and Table 4-8, respectively. Also listed in Table 4-7 and Table 4-8 for comparison are time constants for the ammonia sensor. The adsorption time constant in 50% relative humidity, 83.3 s, was smaller th an for dry air at 200 s. Ho wever, with a lower R-squared value of 0.79 the uncertainty of the humid ad sorption time constant is higher. For desorption, the humid time constant is greater than in dry air, but the two values are closer to each other than were the adsorpti on time constants. The time constants for the carbon dioxide and ammonia sensors ar e on the same order of magnitude.

PAGE 117

103 051015202530 0.0 0.2 0.4 0.6 0.8 1.0 Normalized signaltime (min) (a) 02468101214 0.0 0.2 0.4 0.6 0.8 1.0 Normalized signal time (min) (b) 051015202530 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Normalized signaltime (min) (c) 02468101214 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Normalized signaltime (min) (d) Figure 4-19 The average ( N =4) response time and respective curve fits of the carbon dioxide sensor to (a) adsorption and (b ) desorption in dry air, < 2% relative humidity, (c) adsorption a nd (d) desorption in humid air, 50% rh. The error bars represent standard devi ation for selected data points. Figure 4-20 illustrates the test for cross-sensitivity to humidity. Three tests were used to establish the mean curve. For each test the procedure was as follows: 30 minutes dry air, 30 minutes of 3000 ppm carbon dioxide, 15 minutes dry air, 15 minutes humid air, 30 minutes of 3000 ppm carbon dioxide in a humid air stream, and 30 minutes of humid air. Under dry conditions the ma ximum change in signal to 3000 ppm carbon dioxide was 140 mV, while in humid air the signal only changed by 80 mV. Another important result is that when the relative humidity changed from less than 10% (dry) to 60% (humid), the sensor registered a 230 mV increase; it responded as if the carbon

PAGE 118

104 dioxide concentration was decreasing. Th e same phenomenon was seen with the ammonia sensor and can be attributed to residual analyte on the surface, where when the humidity is increased the water molecules co mpete with the analyte for adsorption sites. The change in signal is due to the water mol ecules removing the residual analyte from the surface. Table 4-7 Adsorption time constants. Equation used for curve fitting was bte a 1. Sensor Flow Relative Humidity (%) Gas concentration (ppm) a b R2 Time Constant, = 1/b (s) CO2 Forced 2 3000 0.890.005 0.95 200 CO2 Forced 50 3000 0.760.012 0.79 83.3 NH3 Forced 2 100 1.010.043 0.99 23.3 NH3 Diffusion 2 100 1.020.014 0.95 71.4 Table 4-8 Desorption time constants. Equation used for curve fitting was btae Sensor Flow Relative Humidity (%) Gas concentration (ppm) a b R2 Time Constant, = 1/b (s) CO2 Forced 2 3000 1.04 0.005430.98 184 CO2 Forced 50 3000 0.91 0.003500.93 286 NH3 Forced 2 100 0.9150.006430.96 156 NH3 Diffusion 2 100 1.00 0.004650.98 215 Fortunately above 10%, for minor fluctuati ons in humidity, the sensor does not respond with the same magnitude as when changing from dry to humid air. Since the guideline for laboratory animal cage relative humidity levels is between 30 to 70%, the significant work function change seen when going from dry to humid air should not hinder the sensor in de tecting carbon dioxide. Summary Experimental results for the carbon dioxide and ammonia sensors were presented and discussed in this chapter. The theore tical model was used with the experimental

PAGE 119

105 results to determine rate constants, the e quilibrium constant, a nd Gibbs free energy for the molecular adsorption of ammonia on a ti tanium nitride surface. A mathematical formulation and experimental results for the diffusion flow regime were presented and found to be comparatively slower than for the forced flow. 0306090120150 0.85 0.90 0.95 1.00 1.05 1.10 1.15 Mean signal (V)time (min)10 100 1000 CO2[gas] (ppm) 0 25 50 75 100 RH %RH Figure 4-20 Carbon dioxide sensor tested for cr oss-sensitivity to humidity. The error bars represent standard deviation for 3 runs at selected data points.

PAGE 120

106 CHAPTER 5 SUMMARY AND CONCLUSIONS Summary of Results There were two main goals for this project. The first goal was to assess the feasibility of applying ammonia and carbon di oxide sensors for monitoring laboratory animal cages through single analyte, time-response, and cross-contamination experiments. The second goal was to theo retically model the chemical kinetics and heterogeneous chemistry of the reaction on the ammonia sensor by defining the reaction mechanism, determining adsorption and desorpti on rates and rate constants, determining the equilibrium constant, and e xploring the role of diffusion. Ammonia Sensor Sensor response and mechanism The mechanism that was best supported by the literature and the experiments was molecular adsorption of ammonia on a titanium nitride surface. Resulting adsorption and desorption reaction rate equations were formulated and time response tests were completed and compared with the model. Curve fits were completed for response time tests using 100 ppm ammonia in dry air (2% rh). From the expe rimental data, the forward and reverse rate consta nts were evaluated as 6.28 L mol-1s-1 and 6.43x10-3 s-1, respectively. The rate constants from the 100 ppm data were used to predict adsorption and desorption curves for 50 ppm data. Time response experiments were also conducted at 50 ppm, where the error between the actual and predicted curves for adsorption and desorption was 38% and 10%, respectively. Th e equilibrium concentration constant was

PAGE 121

107 determined to be 976.7 L/mol. The equilibrium constant was calculated as 39.04, which was used to determine the standard Gibbs energy of reaction, -9.25 kJ/mol. From the equilibrium constants and the Gibbs energy of reaction it was determined that the forward reaction would occur spontaneously. Time response tests were also completed using diffusion dominated delivery to aid in understanding the role diffusion played in the sensors response to ammonia. The diffusion time constants were markedly more than for the forced flow regime. For desorption, the diffusion time constant was 67% more than the forced time constant, while for adsorption it was 27% more. Th e diffusion case was also mathematically formulated as a semi-infinite medium wher e the gas phase ammonia concentration was a function of space and time. To examine temperature effects, time response experiments were completed at three different temperatures. For both th e forward and reverse reactions the rate constants decreased with increasing temperat ure. The equilibrium constant increased with increasing temperature. Performance and feasibility The ammonia sensor responded to changes in ammonia concentration. However, the calibration curve for the ammonia sensor showed that standard deviation for the response to 100 ppm was comparably higher than the total slope of the calibration curve, essentially yielding a detecti on limit of approximately 100 ppm. In future applications, where the concentration is not pre-determine d, it would be difficult to accurately predict the ammonia concentration at levels below 100 pp m. Further, when the sensor is used in a laboratory animal cage, where the air qua lity is unknown, changes in signal could be

PAGE 122

108 from changes in ammonia concentration, change s in relative humidity, or drift of the baseline signal. When the sensor was exposed to dramatic changes in relative humidity, from 2 to 60%, the sensor responded as though more ammonia was being desorbed from the system. This phenomenon only occurred when a mmonia was present in the gas stream or if it had recently been in the system, approximately 10 minutes before humidity increased. If ammonia was not present, then the signal from the sensor did not change when the humidity increased. The sensors cross-sensitivity to humidity not only played into feasibility of using the sensor for this application, but also o ffered insight into the mechanism for the model, where the ammonia and water compete for open surface sites. As for cross-sensitivity to carbon dioxide, it caused a transient spike in the signal when it was introduced into the system. Carbon Dioxide Sensor The carbon dioxide sensor responded to ch anges in carbon dioxide concentration with a decrease in voltage of 60 to 200 mV for concentration changes of 300 to 7000 ppm, respectively. A calibration curve of si gnal versus carbon dioxide concentration was established with a linear curv e fit (R-squared = 0 .88). If the concentration is unknown, as will be the case in actual operation in a laboratory animal cage, it would be difficult to accurately determine the absolute value of carbon dioxide concentration based on the established calibration curve. The transient response of the sensor to addition and removal of analyte was evaluated in dry and humid air. Fifteen to sixteen minutes was the average response time to the introduction of the analyte, while 8 9 minutes was the average response time to removal of the analyte under forced flow conditions. For adsorption the signal was at

PAGE 123

109 50% of the maximum at 2.8 minutes, while the half-life for desorption was 2 minutes. In a laboratory animal cage, these would be suffi cient response times for two reasons. First, the cages are not expected to experience rapi d increases in carbon dioxide concentration. If the concentration does increase it is expe cted to do so gradually over hours or even days. Therefore the sensors response time of 15 16 minutes woul d give sufficient time to monitor gradual changes. Second, if the carbon dioxide co ncentration were to rapidly increase, such as in the case where a ventilat ed cage would lose its air supply due to a power failure, a 50% change would be seen in 2 3 minutes, which would provide ample time to react to the situation and prev ent the animals from overexposure to carbon dioxide. The carbon dioxide sensor was sensitive to changes in relative humidity, especially when the air went from dry (2%) to humid (50%). Additionally, the magnitude of the response to carbon dioxide was reduced when the air was humid with a 40% reduction in signal response between dry and humid air. According to the Guide (18), the cage must be held between 30 70% relative humidity, so the cage will not cha nge rapidly from 2 50%. The cross-sensitivity to water vapor a nd the smaller response in humid air leads to the same conclusion as for the ammonia se nsor, as noted above, the water vapor competes with the analyte for open surface sites. Recommendations At the current development stage, thes e sensors require further modifications before implementation into an animal cage. Adjustments recommended include: Eliminate drift of the baseline signal. In a laboratory animal cage, the ammonia and carbon dioxide concentrations are expe cted to gradually change over hours or days. Since the drift of the baseline a nd response to the gases were on the same order of magnitude, the signal drift may be mistaken for changes in ammonia or carbon dioxide.

PAGE 124

110 Increase sensitivity of the sensors. For example, the ammonia sensors accuracy was not reliable until 200 ppm, which is not desirable for this application where the maximum ammonia concentration stated by the Guide (18) is 25 ppm. Amplify the signal, which would ensure that the magnitude of the signal is consistent and at a different level for each concentration. Couple each sensor with a humidity sensor. This coupling would allow for the gas sensors response to adjust to changes in humidity. For example, if the sensor had a decrease in the work function signal, there should be a mechanism that would crosscheck with the humidity sensor to see if a change was regist ered there as well. If there was a response on the humidity se nsor, then the sensor would output that no change was seen in the gas concentration. Future Work Further work with the field effect transistors includes: Implement the above sensor recommendati ons into the next generation ammonia and carbon dioxide sensors. Develop a theoretical model, similar to th e ammonia model, for the carbon dioxide reaction on a surface that is selective to carbon dioxide such as barium carbonate (BaCO3) or barium titanate (BaTiO3). Additionally, more thorough single analyte, time-response, and cross-sensitivity expe riments could be completed with the carbon dioxide sensor. Then the model could be compared with the limited experimental data completed in this pr oject as well as with the more thorough experiments. Conduct single analyte experiments on th e ammonia sensor at concentrations starting at 5 ppm and increasing until the sensing surface is saturated (the signal does not change with increasing ammonia c oncentration). From these experiments, two plots can be generated (1) 1 vs. 31NHp and (2) 1 vs. 2 131NHp. If plot (1) is linear, then the reaction mechanism be tween ammonia and the titanium nitride surface can be ruled as molecular adsorp tion. If plot (2) is linear, then the mechanism could be ruled as ammonia dissociation. Apply numerical methods to solve the part ial differential equations formulated for the diffusion flow regime. Examine the role of water on the sensitive layer of the sensor. Since the sensors will be used at room temperature with 40-70% relative humidity, it will be important to understand the water surface ch emistry and investigate such items as whether the water adsorption follows a Langmuir model. Over a period of 2 days for static cages and 14 days ventilated cages analyze cage air samples for particulates. The part iculate should be sized, counted, and

PAGE 125

111 characterized (e.g., fibers and spores). With this information filters could be developed for protecting the sensors from contamination in a cage environment. Conduct a field study with the sensors in the laboratory rodent cages. The study would look at the optimal position in the cag e for the sensors, evaluate the rodents response and interaction with the sensors, and determine the overall performance of the sensor in the cage.

PAGE 126

APPENDIX A ENVIRONMENTAL STUDIES

PAGE 127

113 Independent Variables Dependent Variable NH3 CO2 RH Temp.Cage ACPH Room ACPH Cage Air Velocity Cage Type Cage Population Size Bedding Change Frequency Bedding type Mice (M), Rats (R) Ref. ACH xx M (61) ACH xx M (31) Airborne Allergens xx xx M (16) Brain NH3 xx xx xx xx R (25) Breeding Performance xx xx M (35) Physiological Changes xx R (20) CO2 xx xx M (7) CO2 xx M (34) CO2 xx M (10) CO2 xx xx M (35) CO2 xx M (31) CO2 xx xx xx M (13) CO2 xx M (62) CO2 xx xx M (63) Differential pressure xx M (34) Dust content and Generation xx M (17) Dust content and Generation xx M (21) Histological Effects xx R (24) Histological Effects xx xx M (35) Immune function xx xx M (35) Light Intensity xx M (61) Microenv. Contaminants xx M (10)

PAGE 128

114 Independent Variables Dependent Variable NH3 CO2 RH Temp.Cage ACPH Room ACPH Cage Air Velocity Cage Type Cage Population Size Bedding Change Frequency Bedding type Mice (M), Rats (R) Ref. Moisture Absorption xx M (21) Moisture in Bedding xx xx M,R(64) NH3 xx xx M (12) NH3 xx xx M (7) NH3 xx M (17) NH3 xx M (34) NH3 xx xx xx xx R (25) NH3 xx xx xx R (20) NH3 xx M (21) NH3 xx M (10) NH3 xx xx M (35) NH3 xx M (31) NH3 xx xx xx M (13) NH3 xx xx R (65) NH3 xx M (62) NH3 xx xx M (63) NH3 xx M (66) NH3 xx xx M,R(64) NH3 xx M,R(67) NH3 in different cages xx xx xx M (68) NH3, CO2, Temp xx (63) NH3, CO2, Temp xx(flow over cage) xx M (69) NH3, Temp, RH xx xx R (70) Plasma cort. xx xx M (35)

PAGE 129

115 Independent Variables Dependent Variable NH3 CO2 RH Temp.Cage ACPH Room ACPH Cage Air Velocity Cage Type Cage Population Size Bedding Change Frequency Bedding type Mice (M), Rats (R) Ref. Levels Preference xx xx R (71) Rel. Hum. xx xx M (12) Rel. Hum. xx M (17) Rel. Hum. xx M (61) Rel. Hum. xx M (10) Rel. Hum. xx xx M (35) Rel. Hum. M (30) Rel. Hum. xx xx M (13) RH xx M (62) RH xx xx M (63) Room ACH xx M (61) Sound levels xx M (61) Telemetry(heart rate) xx xx R (71) Telemetry(heart rate) xx R (72) Temp xx M (17) Temp xx M (61) Temp xx M (10) Temp xx xx M (35) Temp M (13) Temp xx xx M (13) Temp xx xx M (63) Tracer Gas xx M (73) Weanling wt. And Growth xx xx M (35)

PAGE 130

116 APPENDIX B DETAILS OF FIELD EFFECT TRANSISTOR While the primary focus of this study was the reaction between the analyte gas (ammonia or carbon dioxide) and the sensitive la yer of the field effect transistor, it is important to provide additional details in schematic form (Figure B-1 and Figure B-2), photographs (Figure B-3 and Figure B-4), and electrical drawings (Figure B-5, Figure B6, and Figure B-7). Figure B-1 Schematic section through a hybrid flip chip field effect transistor in a grossly enlarged vertical representation, where PFC indicates polymer flip chip and Si represents silicon. (Data of Pohle et al. (42).)

PAGE 131

117 Figure B-2 Schematic cross-section of a suspe nded gate field effect transistor. (Data of Pohle et al. (42).) Figure B-3 Scanning electron micr oscope cross-section of a hybrid flip chip field effect transistor, where PFC indicates polymer flip chip. The air gap height is approximately 4 microns. (Data of Pohle et al. (42).)

PAGE 132

118 Figure B-4 Printed circuit board for the su spended gate field effect transistor.

PAGE 133

119 Figure B-5 Schematic for one channel of the suspended gate field effect transistor c ontrol board, where the drain to source vol tage is indicated by UDS.

PAGE 134

120 Figure B-6 Detailed drawing of the suspended gate field effect tran sistor control board.

PAGE 135

121 Figure B-7 Simple drawing of the suspended gate field effect transistor control boar d, where OS designates the potentiometers u sed to adjust the offset for the four channels and UDS indi cates the drain source current, which can not be changed.

PAGE 136

122 APPENDIX C ANALYSIS OF EXPERIMENTAL UNCERTAINTY The ammonia and carbon dioxide concentra tions were determined by iteratively solving the following two equations, 3 2 33 3NH CO air NH in outQ Q Q Q NH NH (C.1) and 3 2 22 2 NH CO air CO in outQ Q Q Q CO CO (C.2) where 3NHQ 2COQ and airQ were the volumetric flow rate s for ammonia, carbon dioxide, and air, respectively. inCO2 and inNH3 were the concentrations of the carbon dioxide and ammonia compressed gas cylinders, respec tively. As an example, experimental uncertainties were analyzed for the expe rimental condition of 50 ppm ammonia, 3000 ppm carbon dioxide, and an air fl ow rate of 5 L/min. The fi rst step was to estimate the uncertainty interval for each measured quant ity. The accuracy specification for the flow controllers was % of the fu ll scale. The full scales fo r the ammonia, carbon dioxide, and air flow meters were 50 cc/min, 100 cc/mi n, and 5 L/min, respectively. The accuracy for the compressed gas cylinders was % and .25% for the ammonia and carbon dioxide cylinders, respectively. Therefore the error estimates for the measured flow rates were min / 1 729 60 min / 5 0 684 122 3cc Q cc QCO NH

PAGE 137

123 ppm CO ppm NH L Qin in air500 12 000 250 400 000 20 min / 05 0 000 52 3 The relative uncertainties in measured quantities were 05 0 000 250 500 12 02 0 000 20 400 01 0 min / 000 5 min / 05 0 0165 0 min / 729 60 min / 1 0394 0 min / 684 12 min / 5 02 32 3 ppm ppm CO ppm ppm NH L L u cc cc u cc cc uin in Q Q Qair CO NH. The uncertainty interval for the calculated ammonia concentration was determined from the following equation, 2 1 2 3 3 3 3 2 3 3 2 3 3 2 3 33 2 2 2 3 3 3 3 in air CO NH outNH in out out in Q air out out air Q CO out out CO Q NH out out NH NHu NH NH NH NH u Q NH NH Q u Q NH NH Q u Q NH NH Q u (C.3) The partial derivative terms were 1 1 01 0 13 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 33 2 3 2 2 2 2 3 2 2 3 3 out out in out out in NH CO air air out out air out out air NH CO air CO out out CO out out CO NH CO air CO air out out NH out out NHNH NH NH NH NH NH Q Q Q Q NH NH Q NH NH Q Q Q Q Q NH NH Q NH NH Q Q Q Q Q Q NH NH Q NH NH Q (C.4) Substituting into Eq. C.3 gave

PAGE 138

124 % 52 4 0452 0 02 0 1 01 0 1 0165 0 01 0 0394 0 13 32 1 2 2 2 2or u uNH NH (C.5) The uncertainty interval for the carbon dioxide concentration was determined in a similar fashion as the ammonia concentration. Th e uncertainty interval was calculated from 2 1 2 2 2 2 2 2 2 2 2 2 2 2 2 22 2 2 2 3 3 3 2 in air CO NH outCO in out out in Q air out out air Q CO out out CO Q NH out out NH COu CO CO CO CO u Q CO CO Q u Q CO CO Q u Q CO CO Q u (C.6) where the partial derivative terms were 1 1 1 003 02 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 23 2 3 2 3 2 2 3 2 3 3 3 out out in out out in NH CO air air out out air out out air NH CO air NH air out out CO out out CO NH CO air NH out out NH out out NHCO CO CO CO CO CO Q Q Q Q CO CO Q CO CO Q Q Q Q Q Q CO CO Q CO CO Q Q Q Q Q CO CO Q CO CO Q (C.7) Substituting into Eq. C.6 gave % 35 5 0535 0 05 0 1 01 0 1 0165 0 1 0394 0 003 02 22 1 2 2 2 2or u uCO CO (C.8) For this example, the ammonia and carbon di oxide concentrations fell between the interval 47.75 52.25 ppm and 2840 3160 ppm, respectively.

PAGE 139

125 LIST OF REFERENCES 1. Americans for Medical Progress. Medical Milestones. Animal Research. www.amprogress.org/Issues/Issue sList.cfm?c=10. (July 2005). 2. University of California Center for Anim al Alternatives. The mouse in science. Cancer Research. www.vetmed.ucdavis.edu/Animal_Alternatives/cancer.htm. (May 2004). 3. Lam, M. Mice research useful for vitamin C update. News and Views. www.drlam.com/news_and_views/MiceResea rchUsefulforVitaminCUpdate.cfm (July 2005). 4. The Jackson Laboratory. Research areas. Research. www.jax.org/research/researc h_areas.html. (July 2005). 5. Novak, G. & Sharpless, L. Whats be st for the mouses house: selecting an individually ventilated ca ging system caging system. Lab Anim. (NY) 32(7), 41-47 (2003). 6. Lipman, N. Isolator rodent cag ing systems: a critical view. Contemp. Top. Lab. Anim. Sci. 38(5), 9-17 (1999). 7. Serrano, L. Carbon dioxide and ammonia in mouse cages. Lab. Anim. Sci. 21(1), 75-85 (1971). 8. Schulhof, J. Keeping clean with f ilter top and ventilated cages. Lab Anim. (NY) 19(1), 31-35 (1990). 9. Huerkamp, M. & Lehner, N. Comparative effects of forced-air, individual cage ventilation or an absorbent bedding additive on mouse isolator cage environment. Contemp. Top. Lab. Anim. Sci. 33(2), 58-61 (1994). 10. Perkins, S. & Lipman, N. Char acterization and qua ntification of microenvironmental contaminants in isol ator cages with a variety of contact bedding. Contemp. Top. Lab. Anim. Sci. 34(3), 93-97 (1995). 11. Lipman, N. Microenvironmental conditi ons in isolator cages: an important research variable. Lab Anim. (NY) 21(6), 23-27 (1992).

PAGE 140

126 12. Choi, G., McQuinn, J., Jennings, B., Hasse tt, D. & Michaels, S. Effect of population size on humidity and ammonia le vels in individually ventilated microisolation rodent caging. Contemp. Top. Lab. Anim. Sci. 33(6), 77-81 (1994). 13. Reeb, C., Jones, R., Bearg, D., Bedigian, H., Myers, D., & Paigen, B. Microenvironment in ventilated animal cag es with differing ventilation rates, mice populations, and frequency of bedding changes. Contemp. Top. Lab. Anim. Sci. 37(2), 43-49 (1998). 14. Gordon, S., Fisher, S., & Ronald, R. Elimination of mouse allergens in the working environment: assessment of indivi dually ventilated cage systems and ventilated cabinets in the cont ainment of mouse allergens. J. Allergy Clin. Immunol. 108(2), 288-294 (2001). 15. Sharp, P. & LaRegina, M. The Lab Rat. (CRC Press, Boca Raton, 1998). 16. Sakaguchi, M., Inouye, S., Miyazawa, H., Kamimura, H., Kimura, M., & Yamazaki, S. Evaluation of counterm easures for reduction of mouse airborne allergens. Lab. Anim. Sci. 40(6), 613-615 (1990). 17. Murakami, H. Differences between intern al and external environments of the mouse cage. Lab. Anim. Sci. 21(5), 680-684 (1971). 18. National Research Council. Guide for the Care and Use of Laboratory Animals. (National Academy Press, Washington, D.C., 1996). 19. Keller, L., White, W., Snider, M., & La ng, C. An evaluation of intracage ventilation in three animal caging systems. Lab. Anim. Sci. 39(3), 237-242 (1989). 20. Gamble, M. & Clough, G. Ammonia buildup in animal boxes and its effect on rat tracheal epithelium. Lab. Anim. 10, 93-104, (1976). 21. Potgieter, F. & Wilke, P. The dust c ontent, dust generation, ammonia production, and absorption properties of three different rodent bedding types. Lab. Anim. 30, 79-87, (1996). 22. National Institute of Occupational Health and Safety. Occupational Health Guideline for Ammonia. 92-110, (1992). 23. Coon, R., Jones, R., Jenkins, L., & Sieg el, J. Animal inhalation studies on ammonia, ethylene glycol, formalde hyde, dimethylamine, and ethanol. Toxicol. Appl. Pharmacol. 16(3), 646-655 (1970). 24. Broderson, J., Lindsey,J. & Crawford, J. The role of environmental ammonia in respiratory mycoplasmosis of rats. Am. J. Pathol. 85(1), 115-127 (1976).

PAGE 141

127 25. White, W. & Mans, A. Effect of beddi ng changes and room ventilation rates on blood and brain ammonia levels in normal rats and rats with portacaval shunts. Lab. Anim. Sci. 34(1), 49-52, (1984). 26. Krohn, T. & Hansen, A. Carbon dioxide con centrations in unventilated IVC cages. Lab. Anim. 36, 209-212, (2002). Krohn, T. & Hansen, A. Carbon dioxide concentrations in unventilated IVC cages. Lab. Anim. 36, 209-212, (2002). 27. Krohn, T. & Hansen, A. The effects of and tolerances for carbon dioxide in relation to recent developments in laboratory animal housing. Scand. J. Lab. Anim. Sci. 27, 173-181, (2000). 28. National Institute of Occupational Health and Safety. Occupational Health Guideline for Acetic Acid. 92-110, (1992). 29. National Institute of Occupational Health and Safety. Occupational Health Guideline for Sulfur Dioxide 92-110, (1992). 30. Lipman, N. & Perkins, S. Macroenvironm ental relative humidity and bedding in isolator cages. Contemp. Top. Lab. Anim. Sci. 39(6), 7 (2000). 31. Reeb, C., Jones, R., Bearg, D., Bedigi an, H., & Paigen, B. Impact of room ventilation rates on mouse cage ve ntilation and microenvironment. Contemp. Top. Lab. Anim. Sci. 36(1), 74-79 (1997). 32. Besch, E. Environmental quality within animal facilities. Lab. Anim. Sci. 30, 385398, (1980). 33. Hasenau, J., Baggs, R. & Kraus, A. Mi croenvironments in microisolation cages using BALB/c and CD-1 mice. Contemp. Top. Lab. Anim. Sci. 32(1), 11-16 (1993). 34. Hoglund, A. & Renstrm, A. Evaluation of individually ventilated cages for lab rodents. Lab. Anim. 35, 51-57, (2001). 35. Reeb-Whitaker, C., Paigen, B., Beam er, W., Bronson, R., Churchill, G., Schweitzer, I., & Myers, D. Impact of reduced frequency of cage changes on the health of mice housed in ventilated cages. Lab. Anim. 35, 58-73, (2001). 36. Agrawal, R. A model for minimizing cost for housing laboratory mice. Masters Thesis University of Florida. 2003. 37. Oxner, E. FET Technology and Application. (Marcel Dekker Inc., New York, 1989). 38. Askeland, D. The Science and Engineering of Materials. (PWS Publishing Company, Boston, 1994).

PAGE 142

128 39. Rizzoni, G. Principles and Applications of Electrical Engineering. (McGraw Hill, Boston, 2000). 40. Lesurf, J. Field effect transistor. The First Eleven. University of St. Andrews. 22 April 2004, www.standrews.ac.uk/~www_pa/Scots_Gui de/first11/part7/page1.html. 41. Fleischer, M., Ostrick, B., Pohle, R., Sim on, E., Meixner, H., Bilger, C., & Daeche, F. Low-power gas sensors based on work-f unction measurement in low-cost hybrid flip chip technology. Sens. Actuators B Chem. 80, 169-173, (2001). 42. Pohle, R., Simon, E., Schneider, R., Fl eischer, M. & Meixner, H. Improved CCFET and SGFET type gas sensors in hybrid flip chip technology: realization and characterization of a new sensor concept. Not yet published. 43. Abom, E., Hultman, L., Eriksson, M., & Twes ton, R. Properties of combined TiN and Pt thin films applied to gas sensing. J. Vac. Sci. Technol. A 20(3), 667-673, (2002). 44. Ostrick, B., Pohle, R., Fleischer, M., & Meixner, H. TiN in work function type sensors: a stable ammonia sensitive mate rial for room temperature operation with low humidity cross sensitivity. Sens. Actuators B Chem. 68, 234-239, (2000). 45. Ostrick, B., Fleischer, M., Meixner, H., & Kohl, C. Investigation of the reaction mechanisms in work function type sensor s at room temperature by studies of the cross-sensitivity to oxygen and water: the carbon-carbon cioxide system. Eurosensors XIII 81-84, (1999). 46. Simon, E., Lampe, U., Pohle, R., Fleisc her, M., Meixner, H., Frerichs, H.-P., & Lehmann, M. Novel carbon dioxide gas se nsors based on field effect transistors. Not yet published. 47. Ostrick, B., Fleischer, M., & Meixner, H. The influence of interfaces and interlayers on the gas sensitivity in work function type sensors. Sens. Actuators B Chem. 95, 271-274, (2003). 48. Atkins, P. Physical Chemistry. (W.H. Freeman and Company, New York, 1995). 49. Takagi-Kawai, M., Soma, M., Onishi, T., & Tamaru, K. The adsorption and the reaction of NH3 and NOx on supported V2O5 catalysts: effect of supporting materials. Can. J. Chem. 58(20), 2132-2137 (1980). 50. Shimanouchi, T. Tables of Molecular Vi brational Frequencies. (National Bureau of Standards, Washington, D.C., 1972). 51. Hlil, E., Kubler, L., Bischoff, J., & Bolmont, D. Photoemission study of ammonia dissociation on Si(100) below 700 K. Phys. Rev. B., Condens. Matter. 35(11), 5913-5916, (1987).

PAGE 143

129 52. Bozso, F. & Avouris, P. Photoemission studies of the reactions of ammonia and N atoms with Si(100)-(2x1) and Si(111)-(7x7) surfaces. Phys. Rev. B., Condens. Matter. 38(6), 3937-3942, (1988). 53. Bischoff, J., Lutz, F., Bolmont, D. & Kubl er, L. Use of multilayer techniques for XPS identification of various nitr ogen environments in the Si/NH3 system. Surface Science. 251/252, 170-174, (1991). 54. Zhou, X., Flores, C., & White, J. Decomposition of NH3 on Si(100): a SSIMS study. Surface Science Letters. 268, L267-273, (1992). 55. Chen, P., Colaianni, M., & Yates, J. Silicon backbond strain effects on NH3 surface chemistry: Si(111)-(7x7) compared to Si(100)-(2x1). Surface Science Letters. 274, L605-610, (1992). 56. Siew, H., Qiao, M., Chew, C., Mok, K., Chan, L., & Xu, G. Adsorption and reaction of NH3 on Ti/Si(100). Applied Surface Science. 173, 95-102, (2001). 57. Laksono, E., Galtayries, A., Argile, C., & Marcus, P. Adsorption of NH3 on oxygen pre-treated Ni(111). Surface Science. 530, 37-54, (2003). 58. Diebold, U. & Madey, T. Adsorption and electron stimulated desorption of NH3/TiO2(110). J. Vac. Sci. Technol. A 10(4), 2327-2335, (1992). 59. Karthigeyan, A., Gupta, R., Scharnagl, K ., Burgmair, M., Sharma, S., & Eisele, I. A room temperature HSGFET ammonia sensor based on iridium oxide thin film. Sens. Actuators B Chem. 85, 145-153, (2002). 60. Mhadeshwar, A., Wang, H., & Vlachos, D. Thermodynamic consistency in microkinetic development of su rface reaction mechanisms. J. Phys. Chem. B. 107, 12721 12733, (2003). 61. Clough, G., Wallace, J., Gamble, M., Merryweather, E., & Bailey, E. A positive, individually ventilated cagi ng system: a local barrier syst em to protect both animals and personnel. Lab. Anim. 29, 139-151, (1995). 62. Lipman, N, Corning, B., & Coiro, M. The effects of intracage ventilation on microenvironmental conditions in filter-top cages. Lab. Anim. 26, 206-210, (1992). 63. Corning, B. & Lipman, N. A compar ison of rodent caging systems based on microenvironmental parameters. Lab. Anim. Sci. 41, 498-503, (1991). 64. Wu, D, Joiner, G., & McFarland, A. A fo rced-air ventilation system for rodent cages. Lab. Anim. Sci. 35, 499-504, (1985).

PAGE 144

130 65. Ishii, T., Yoshida, K., Hasegawa, M., Mizuno, S., Okamoto, M., Tajima, M., & Kurosawa, T. Invention of a forced-air -ventilated micro-isolation cage and rack system. Environment within cages: temp erature and ammonia concentration. Appl. Anim. Behav. Sci. 59, 115-123 (1998). 66. Keller, G., Mattingly, S., & Knapke, F. A forced-air individually ventilated caging system for rodents. Lab. Anim. Sci. 33, 580-582, (1983). 67. Weissman, S., Beethe, R., & Redman, H. Ammonia concentrations in an animal inhalation exposure chamber. Lab. Anim. Sci. 30, 974-980, (1980). 68. Eveleigh, J. Murine cage density: ca ge ammonia levels during the reproductive performance of an inbred strain and tw o outbred stocks of monogamous breeding pairs of mice. Lab. Anim. 27, 156-160, (1993). 69. Corning, B., & Lipman, N. The effects of a mass air displacement unit on the microenvironmental parameters within isolator cages. Lab. Anim. Sci. 42, 91-93, (1992). 70. Hirsjarvi, P. & Valiaho, T. Microc limate in two types of rat cages. Lab. Anim. 21, 95-98, (1987). 71. Krohn, T., Hansen, A., & Dragsted, N. Th e impact of cage ventilation on rats housed in IVC systems. Lab. Anim. 37, 85-93, (2003). 72. Krohn, T., Hansen, A., & Dragsted, N. The impact of low levels of carbon dioxide on rats. Lab. Anim. 37, 94-99, (2003). 73. Keller, L., White, W., Snider, M., & Lang, C. An evaluation of intra-cage ventilation in 3 animal caging systems. Lab. Anim. Sci. 39, 237-241, (1989).

PAGE 145

131 BIOGRAPHICAL SKETCH Karen (Timm) Supan grew up in Stoughton, Wisconsin. She received her bachelors degree in mechanical engineering from Minnesota State University, Mankato, in May 2000. A summer internship with th e Institute of Paper Science and Technology in Atlanta, Georgia, while still an undergra duate led her to pursue a masters degree in pulp and paper science. She completed he r masters degree in 2002 and returned to mechanical engineering by entering the doctorate program at the University of Florida. Karen and her husband, Brian, were married in 2002 and currently reside in Tampa, Florida.


Permanent Link: http://ufdc.ufl.edu/UFE0011981/00001

Material Information

Title: Model for Suspended Gate Field Effect Transistors Used in Laboratory Animal CagexxMonitoring
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: UFE0011981:00001

Permanent Link: http://ufdc.ufl.edu/UFE0011981/00001

Material Information

Title: Model for Suspended Gate Field Effect Transistors Used in Laboratory Animal CagexxMonitoring
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: UFE0011981:00001


This item has the following downloads:


Full Text












MODEL FOR SUSPENDED GATE FIELD EFFECT TRANSISTORS USED IN
LABORATORY ANIMAL CAGE MONITORING















By

KAREN E. SUPAN


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


































Copyright 2005

by

Karen E. Supan

































This document is dedicated to my parents, Mary Jo and Fred Timm.















ACKNOWLEDGMENTS

First, I would like to thank my advisor, Dr. Herbert Ingley, for his constant support

and encouragement of this work. From our first meeting, he has been a great mentor to

me in engineering, teaching, and life in general. I thank him for the countless words of

wisdom he imparted to me. I acknowledge my appreciation to Dr. David Hahn for the

generous amount of time he devoted to this project by providing technical assistance on

the experimental portion of this project, offering his expertise in a new area to me and for

serving on my committee. I also express gratitude to Dr. Sherif Sherif, Dr. Bill Lear, and

Dr. Jason Weaver for being committee members and providing guidance in their area of

proficiency.

I am grateful to Osman Ahmed of Siemens Building Technology for the inception

and financial support for this project. Danke \, heil to Roland Pohle, Peter Gulden, and

many others who made the trek across the Atlantic from Siemens Core Technology in

Munich, Germany, to collaborate on this project. Special thanks go to Dr. Gus Battles

and Mike Cormier from Animal Care Services for opening their facilities for this project

and answering many questions along the way.

It is impossible to forget all the friends and colleagues I have met at UF who have

made the time here so enjoyable.

Above all, I express thanks to my parents and family for creating an environment

where education was a high priority and paving the way for me to follow in their tracks.

My deepest gratitude goes to my husband, Brian, for providing encouragement to pursue









this degree. I will be forever indebted to him for the distances he went to help me

achieve this accomplishment.

















TABLE OF CONTENTS
Page


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

LIST OF TABLES ........................ ... ....... ............. ix

LIST OF FIGURES ................................. ...... ... ................. .x

A B S T R A C T .......................................... ..................................................x iii

CHAPTER

1 B A C K G R O U N D ................................................................ ....................... .

Comparison of Static Isolator Cages to Ventilated Caging Systems............................4
T he M acroenvironm ent ............................................................... ......................... 7
The M icroenvironm ent .............................................. ........ ........ ........ ..
Am m onia ................ ......... ..... ................ ......... ............. 10
C arb on D iox id e .................................................. ................ 14
Other Contaminants ........... ...... ..... .. ... ................ ....... ....... 16
Contact Bedding ........... .... ...... .. .... .................... .... .. 17
R elative H um idity .................................. .. .. ...... ............ 19
Ventilation ...................................................................................... ................. 20
A additional E nvironm mental Factors ............................................. ....................21
Previous Environm ental Studies........................................... .......................... 22
Cost Analysis for Current Husbandry Practices........................................ ...............29
A ir Sam pling T echniqu es ........................................ ............................................30
S em ic o n d u cto rs ............................................................................................. 3 1
Field E effect T ransistors ............................................... ............................ 33
G a s S e n sin g ................................................................................................... 3 5
Summary ....................... ......... ............................40

2 EXPERIMENTAL FACILITIES AND METHODS ............. ............... 42

E x p erim en tal S etu p ........................................................................... .... ..............4 2
E xperim mental P procedures ......... ................. ............................................................4 8
Carbon Dioxide Sensor........... .................. ................. ..... ............... 48
Amm onia Sensor .......... .. ...... .......... .. ............ .... .... ... .. 49









3 THEORETICAL MODELING ...........................................................................54

Gibbs Free Energy ................................. ........ ... ........ ......... 54
A d so rp tio n .............................................................5 8
Surface R action Rate Expressions ........................................ ........................ 60
Langm uir A dsorption Isotherm ........................................ ....................... 60
D issociativ e A dsorption ........................................................... .....................62
C om petitiv e A dsorption ........................................................... .....................63
P proposed M mechanism s ........................................................................ .................. 64
Am m onia and Hydroxide ............................................................................64
Ammonia Dissociation ............... ........... ...... ............... 66
M olecu lar A d sorption ............................................................... .....................72
R action K inetics.......... ...................................................................... ........ .. .. ..74
S u m m a ry .....................................................................................................7 7

4 RESULTS AND DISCU SSION ........................................... .......................... 78

D rift T e sts .............. .............................................................................................. 7 8
Ammonia Sensor Results.......... ......................... ..................... 79
Sensor Response and M echanism ............................................ ............... 80
D iffu sio n ........................................................................................................ 8 7
T em perature E effects ............................................ .. ........ .......... .....89
Sensor Performance...... ......... .. ................. ............................... 91
G radual ram ping tests.................... ....................... ............... ... 94
Cross-sensitivity to humidity and carbon dioxide.......................................96
C arbon D ioxide Sensor R results ......... .............................................. ... ............ 99
S u m m ary ................................104.............................

5 SUMMARY AND CONCLUSIONS .................................................................106

Sum m ary of R esults......... ............................................................... .... .... .... ... 106
A m m onia Sensor ............. ......... .... ..... ........... .... ... .... .. ............ 106
Sensor response and mechanism .........................................................106
Performance and feasibility...................... .............. ..................107
Carbon Dioxide Sensor............... ........................ .. ............. ............... 108
Recommendations........ ........ ........ .. ................. ....... 109
F utu re W ork ........... ................................... ............... ................... 110

APPENDIX

A ENVIRONMENTAL STUDIES ........................... ......... ..................... 112

B DETAILS OF FIELD EFFECT TRANSISTOR.................................................116

C ANALYSIS OF EXPERIMENTAL UNCERTAINTY ................ .................122









L IST O F R E FE R E N C E S ......................................................................... ................... 125

BIOGRAPHICAL SKETCH ............................................................. ..................131
















LIST OF TABLES

Table pge

1-1. Recom m ended space for lab mice ................................................................. 22

1-2. Ammonia concentration levels after seven days in static isolator cages .................23

1-3. Gas sampling pumps used in environmental studies..............................................30

2-1. Sensor testing equipm ent ......................................................... ............... 43

2-2. Ammonia and carbon dioxide concentrations ................................. ............... 47

2-3. Ammonia concentrations used for ramping tests ....................... .....................50

2-4. Experimental parameters for time response tests ............................................... 52

3-1. Infrared (IR) and XPS measurements and assignments for adsorbed species of
am m onia on surface catalysts........................................................ ............... 66

4-1. Rates of baseline signal drift for the carbon dioxide and ammonia sensor ..............79

4-2. Curve fit coefficients, R-squared values, and rate constants for desorption and
adsorption of 100 ppm ammonia on sensor. ................................. .................83

4-3. Adsorption and desorption time constants for the ammonia sensor using forced and
diffusion flow regime es. ................................................ ................................ 90

4-4. Curve fit coefficients, R-squared values, and rate constants for desorption and
adsorption of ammonia on sensor at varying surface temperatures .......................91

4-5. Slopes for ramp down and ramp up curves completed in diffusion box .................96

4-6. Magnitude of signal response for changes in ammonia concentration and relative
hum idity ............................................................................98

4-7. A dsorption tim e constants ............................................... ............................ 104

4-8. D esorption tim e constants ............................................... ............................ 104
















LIST OF FIGURES

Figure pge

1-1. A static m icroisolator cage ........................................ ................................. 4

1-2. Equipment used to ensure a healthy microenvironment in an animal laboratory ........6

1-3. When excited, electrons move from the valence band to the conduction band across
the energy gap. ................................................................32

1-4. Schem atic of field effect transistor............... .... ................................. ........ ... 34

1-5. Classical FET configuration ........... ................. ........ .............................. 35

1-6. Suspended Gate FET configuration..................... ..... ........................... 35

1-7. Photograph of a hybrid flip chip FET sensor device...............................................36

2-1. Schem atic of experim ental facilities...................................... ......................... 42

2-2. Humidification section of experimental facilities ............................................... 44

2-3. Mixing section of experimental facilities.................. ........................... 44

2-4. Two flow regimes used in sensor testing.......... ......... ................... ...............45

2-5. C arb on dioxide sen sor ............................ ........................................ ........................46

2-6. Signal from DC power supply ....... ......... ..................... ................. 48

2-7. Graphical depiction of experimental parameters used to test the carbon dioxide
sensor for cross-sensitivity to humidity. ...................................... ............... 49

2-8. Graphical depiction of ammonia concentrations used in ramping tests ...................50

2-9. Graphical depiction of experimental parameters used to test the ammonia sensor for
cross-sensitivity to humidity while ammonia was present.................. .............51

2-10. Graphical depiction of experimental parameters used to test the ammonia sensor for
cross-sensitivity to humidity and carbon dioxide.................................................52









3-1. Diffuse Reflectance Infrared Fourier Transform Spectra (DRIFT-spectra) for a TiN
screen-printed film ....................................................... .. ........ .... ..... ...... 65

3-2. Mechanism for the reduction of NO by NH3 over a V205 sensing layer in the
presence of oxygen .......... ................. .. ............... .... ..... 65

3-3. The XPS N(ls) core-level spectra for ammonia............................................. 67

3-4. The XPS N(ls) core-level spectra for ammonia on (a) Si(100)-(2xl) and (b) Si( 11)-
(7 x 7 ) ............................................................................. 6 8

3-5. Surface species and desorption products from ammonia on Si(100) .......................69

3-6. Ball and stick models for the adsorption geometry for -NH2 and -H on two different
su rfa c e s ............. .. ............... ................. ..............................................7 0

3-7. The XPS N Is spectra of Ti/Si (100) surface .............................. ...................71

4-1. L ong-term drift test results .............................................................. .....................79

4-2. Average desorption data and curve fit for 100 ppm desorption tests......................82

4-3. Average adsorption data and curve fit for 100 ppm adsorption tests......................83

4-4. Actual 50 ppm average (N=5) desorption curve plotted along with the predicted
cu rv e ............. .......... .. ......... ................... ............................. 84

4-5. Actual 50 ppm average (N=6) adsorption curve plotted along with the predicted
c u rv e ............................................................................ ................ 8 5

4-6. The slope of the Gibbs energy changes as the reaction proceeds...............................87

4-7. Diffusion of ammonia in a semi-infinite region. .................. .................88

4-8. Average (N=6) desorption curve for the diffusion case compared with the forced
curve fit for 100 ppm data ...................................................................... 89

4-9. Average (N=6) adsorption curve for the diffusion case compared with the forced
curve fit for 100 ppm data ...................................................................... 89

4-10. Average desorption and adsorption curves at 50 ppm ammonia for heater voltages
of (a) (b) 2 V (N 10) and (c) (d) 3 V (N=7 and N 8)................................... 91

4-11. Single analyte test results (N 1) for the ammonia sensor................... ............93

4-12. Calibration curve and linear fit for ammonia sensor ............... ... ................ 93

4-13. Ammonia concentration ramped up and down with ammonia sensor in diffusion
b o x ................................................................................................9 5









4-14. Ammonia sensor tested for cross-sensitivity to humidity .....................................96

4-15. Ammonia sensor tested for cross-sensitivity to humidity with ammonia in the
sy ste m ...................................... ................................................... 9 7

4-16. Average (N = 6) ammonia sensor response to humidity, carbon dioxide, and
am m onia concurrently ..................................... ................................................. 100

4-17. Raw data (N 1) from carbon dioxide single analyte test ............... ...............101

4-18. Calibration curve for carbon dioxide sensor...................................... ..................101

4-19. The average (N=4) response time and respective curve fits of the carbon dioxide
se n so r .................................. .......................................................... ............... 1 0 3

4-20. Carbon dioxide sensor tested for cross-sensitivity to humidity.............................105

B-1. Schematic section through a hybrid flip chip field effect transistor ........................116

B-2. Schematic cross-section of a suspended gate field effect transistor ...................17

B-3. Scanning electron microscope cross-section of a hybrid flip chip field effect
tra n sisto r ................................................... .................. ................ 1 1 7

B-4. Printed circuit board for the suspended gate field effect transistor ......................118

B-5. Schematic for one channel of the suspended gate field effect transistor control board,
where the drain to source voltage is indicated by UDS. ........... ...............119

B-6. Detailed drawing of the suspended gate field effect transistor control board..........120

B-7. Simple drawing of the suspended gate field effect transistor control board............121

















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

MODEL FOR SUSPENDED GATE FIELD EFFECT TRANSISTORS USED IN
LABORATORY ANIMAL CAGE MONITORING

By

Karen E. Supan

December 2005

Chair: H.A. Ingley III
Major Department: Mechanical and Aerospace Engineering

Over the past century, great advances in medicine have been achieved through the

use of laboratory animals, specifically rodents. The quality of the animal environment is

important to the rodent's health and welfare, and their well-being directly affects the

quality of research involving their use. There can be significant variability in air quality

between cages depending on a number of factors such as population size and air flow. A

way to accommodate for the variability between cages is to monitor environmental

quality indicators within the cage, such as ammonia, carbon dioxide, temperature, and

relative humidity. Since rodent cages are approximately the size of a shoebox,

commercially available sensors would be too large for this application. Therefore, micro-

sensors, or field effect transistors, were investigated for application in a rodent cage.

Since these sensors were on the forefront of technology, a theoretical model was

developed for the ammonia sensor to further understand the chemical reaction taking

place on its surface.









The sensors were tested in a controlled environment, where the air quality was

known. The magnitude and time of the response to different levels of contaminants (e.g.,

ammonia and carbon dioxide) were determined. The study showed that the sensors can

detect changes in air quality in a sufficiently short amount of time (5 minutes) so that

corrective action could be taken to prevent the rodents from overexposure to harmful

levels of air contaminants. At the present development stage, the sensors used for this

investigation will require further improvements before implementation in a laboratory

animal cage. These improvements include but are not limited to eliminating drift of

baseline signal, increasing sensitivity of sensor, amplifying signal output, and coupling

each gas sensor with a humidity sensor.

The reaction mechanism selected for the model which was best supported by the

literature and the experiments was molecular adsorption of ammonia on a titanium nitride

surface. The experimental results were fitted to the model to obtain the adsorption and

desorption rate constants, the equilibrium concentration constant, equilibrium constant,

and Gibbs free energy, which respectively were 6.28 L/mol*s, 6.43 x 103 s 1, 976.7

L/mol, 39.04, and -9.25 kJ/mol. Based on these values, it was determined that the

forward reaction, or adsorption, occurs spontaneously. There was good correlation

between the theoretical model and the experimental results, indicating that the theoretical

model was sufficient for this application.















CHAPTER 1
BACKGROUND

Nearly every medical breakthrough in the last century has come as a result of

research with animals (1). In the nineteenth century, animals contributed to the

treatments for rabies, smallpox, and anthrax. The early 1900's saw breakthroughs in

cardiac catheterization techniques, treatment for rickets, and the discovery of insulin. In

the 1930s research with dogs contributed to the development of modern anesthesia, while

horses aided in the prevention of tetanus. The 1940s saw treatment of rheumatoid

arthritis and whooping cough, prevention of diphtheria, and development of antibiotics.

In the 1950s a cure for polio was found through the use of rabbits, monkeys, and rodents.

That decade also saw the development of open-heart surgery, cancer chemotherapy, and

tranquilizers. In the past 50 years, animals have helped find treatments for diseases such

as rubella, measles, and Hansen's disease. They have also further advanced research on

organ transplants, breast cancer, cystic fibrosis, multiple sclerosis, and Lou Gehrig's

disease. While much of this research was done with large animals, such as dogs,

monkeys, or sheep, today's medical research mainly utilizes rodents.

Mice have been utilized in a wide array of medical research. For example, mice

have been used in cancer research since 1894 (2). In 1921 inbred strains, which were

susceptible to tumors, were created. More strains were developed in 1929 with the

founding of one of the first animal laboratories, Jackson Laboratories. In 1962, a mutant

mouse with low immunity was discovered which led to human tumor transplantations.









The late 1980's saw a boom in mice research with the development of a transgenic mouse

whose genes were altered to produce a desired characteristic. From this research,

oncogenes, a gene that can cause a normal cell to become cancerous, could then be

studied. Genetically engineered mice are also used today to determine Vitamin C's role

in health and illness (3). Today, research is ongoing in such diverse areas as diabetes,

hearing loss, ovarian cancer, and glaucoma (4).

Mice are valuable for medical research because of their genetic similarity to

humans (1). Laboratory mice, Mus domesticus, belong to the family Muridae and are a

domesticated variant of the house mouse, Mus musculus (5). Adult mice are adapted to

live in groups and generally live no longer than two years. They typically weigh 20-40

grams (0.7-1.4 ounces), have a length of 12-15 centimeters (5-6 inches), and when

standing on hind legs achieve a height of 10-12 centimeters (4-5 inches).

The quality of the environment is imperative to the rodent's health and welfare, and

their well-being directly affects the quality of research collected from them (5). Rodent

cages have evolved over the years to better suit the needs of the animals.

Lisbeth Kraft, in the late 1950's, was the first to separate mice into isolator cages to

prevent the spread of communicable diseases through direct contact, specifically

epidemic diarrhea of infant mice (6).

Filter-top covers were added to Kraft's isolator system to prevent the transmission

of airborne diseases as well as to eliminate the exchange of feces, soiled bedding, and

hair between adjacent cages (7). Robert Sedlacek developed the modern filter top while

working with a large gnotobiotic mouse colony used in radiation biology research at

Massachusetts General Hospital. The filter top was a polycarbonate frame fitted with a









polyester filter medium held in place by a perforated aluminum plate. The rim at the

bottom of the filter top, where it fit over the underlying cage, formed a lip. One of the

first static isolator cages consisted of a cylindrical cage with a solid galvanized bottom

and a tight-fitting lid with metal mesh sides wrapped with fiberglass insulation that

filtered the incoming air. The contemporary static isolator caging system has given

institutions the ability to keep rodents "clean" while housing contaminated and clean

rodents simultaneously (6).

Further improvements were made to the static isolator cages with the onset of

ventilated caging systems (VCS). These systems were developed to improve intracage

ventilation and to increase housing capacity. The first ventilated cages were developed

under the direction of Dr. Ed Les at Jackson Laboratories in 1960, around the same time

as the filter-top cages. The earliest version deposited mouse odors and allergens out into

the room. Jackson Laboratories and Thoren Caging Systems collaborated to further

improve upon this design in the late 1970s by filtering the incoming and exhaust air. The

first systems were commercially available in the early 1980s. By the early 1990s VCS

had gained widespread popularity (6). Meanwhile, other companies such as Hazelton

Systems were developing a ventilated rack system to help reduce allergies of staff

members (8). In Hazelton's system, air was blown into each shelf row of cages or each

individual cage at a low velocity and removed by a main exhaust system.

In a modem VCS, high efficiency particulate arresting (HEPA) filtered air is blown

into cages through a manifold under positive-pressure. Air is either exhausted directly

from the cage or filtered before sending it to the room or venting outside (9).









In addition to the VCS, HEPA filtered flow work areas were developed for use

when changing the bedding or restocking food and water in a cage. The work area helps

maintain the microbial barrier created by the static isolator cage and VCS. Ambient air is

drawn into the station, moved through a pre-filter and HEPA filter, and is then blown

horizontally or vertically across the workstation.

Comparison of Static Isolator Cages to Ventilated Caging Systems

Static microisolator cages are still used in many animal laboratories. Isolator cages

are cost effective and allow containment at the cage level without expensive ventilation

(10). They are a proven technology for protecting valuable mice from microbial

contamination (5). Isolator caging systems provide a separate microenvironment and aid

in the development and upkeep of disease-free rodents for use in research. Static cages

are useful for studies where containment at the cage level is desirable, for example, in

vivo administration of hazardous agents (6). Photographs of static microisolator cages

are depicted in Figure 1-1.



,Nil-









(a) (b)

Figure 1-1 A static microisolator cage (a) side view and (b) top view.

While static cages may be cost effective, there are drawbacks that must be

considered. The use of these cages can be labor intensive depending on how often the









bedding is changed. Changing and handling of the cages is physically intensive

potentially leading to back or hand injuries. Due to frequent cage changing a large

supply of bedding is required. The cages must be washed frequently which uses a

significant amount of water and electricity and can accelerate the degradation of the cage

(9).

The static cages in use today are advantageous because they are durable,

transparent, and have a replaceable filter top and tight-fitting lid, which is not easy to

dislodge (11). When the filter top was first introduced, the advantage of containment

outweighed the effects on the air quality inside the cage. Research soon revealed that the

filter top was a barrier to air and moisture exchange increasing the intracage relative

humidity in one study by 38% compared to the macroenvironmental humidity (6). The

filter top impedes air exchange between the micro and macro environments. The only

ventilation in static cages comes from the rodents breathing patterns (8).

Individually ventilated caging systems (IVCS) have addressed the problem of little

to no air exchange between the micro and macro environments. The IVCS combines the

static cages with individual ventilation. The whole package, as depicted in Figure 1-2,

should include a microisolation cage, ventilated cage rack, and a Class 100 or Class II

change cabinet, which helps rodents remain disease-free in a healthy microenvironment

(5).

Microenvironmental air quality is better and the variability in air quality between

cages is reduced when using ventilated cages. Intracage ammonia, carbon dioxide, and

relative humidity are also lower. In addition, the day on which ammonia is first detected

can be delayed (6, 12, 13). One study revealed that direct ventilation, 23 air changes per









hour (ACPH) to each cage, improved the microenvironmental conditions in comparison

with a static isolator cage. The relative humidity, ammonia, and carbon dioxide in the

ventilated cages were 8% lower, 240 parts per million (ppm) lower, and 2900 ppm lower,

respectively (11). Ventilated systems also help prevent mouse urinary protein (MUP)

from spreading. A study showed that less than 0.05 ng/m3 (5x10-7 ppb) was detected in a

room with IVCS, whereas a high level, 4.6 ng/m3 (4.6x10-6 ppb), of MUP was measured

when mice were housed in open cages (14).



















(a) (b)
Figure 1-2 Equipment used to ensure a healthy microenvironment in an animal laboratory
should include (a) a cage changing station and (b) a ventilated cage rack.

Operational savings can also be achieved with IVCS. With the improved air

quality, cage changing can be delayed to weekly or longer, which translates into labor

savings. The time spent sanitizing cages and the quantity of bedding used are reduced as

well. Due to the decreased cage changing intervals the service life of the cages and

caging systems is increased. The stocking density per room can be increased, thereby









allowing more efficient use of space in a laboratory animal facility (6). Overall, an

improved microenvironment leads to lower operational costs.

Despite the improvements in micro and macro environmental air quality and

savings in operational costs, IVCS are not used in every laboratory animal facility. To

begin with, IVCS are expensive to acquire. Despite the advantages of IVCS, a new

system is decidedly capital intensive (6). Other costs accrue from electricity for

operating the system, maintenance, and replacement of filters. The IVCS are also more

difficult to sanitize than a standard rack holding isolator cages. Blowers, shelves, and

access panels must be removed before placing the IVCS in a rack washer. Extensive

washing by hand is required and access to all plenums and ducts may not be feasible.

The heat gain in the housing room may be increased due to the supply and exhaust

blowers. If stocking density is increased the heat gain due to the mice will rise causing

an increase in the total cooling load. Noise generation from the IVCS exhaust and supply

blowers may be an issue for employees working in the room and mice in the

microenvironment. Finally, excess intracage ventilation can cause chilling and

dehydration, especially with neonates and hairless mutants. Whichever system a

laboratory animal facility chooses to use, the most important issue is the health and

welfare of the animals and laboratory personnel.

The Macroenvironment

To better meet the needs of the animals and personnel, the environment in which

they work or reside must be examined. The overall facility or building is known as the

megaenvironment, the items in an animal's room are considered the macroenvironment,

and the items in an animal's cage or immediate surroundings are the microenvironment

(15). Each environment can be treated separately, but the couplings between each system









must also be considered. For example, a high-level exhaust system can improve thermal

ventilation efficiency for the mega and macro environments. This high-level exhaust

system may not provide enough circulation to properly ventilate the individual static

isolator cages and thereby does nothing to decrease the ammonia and carbon dioxide

concentrations in the microenvironment (6). Therefore improvements for the

macroenvironment will not necessarily improve the microenvironment.

One of the main reasons laboratory animal facilities are concerned with the

condition of the macroenvironment is for laboratory animal personnel. According to the

National Institute for Occupational Safety and Health (NIOSH), 33% of animal handlers

have allergic symptoms and 10% have symptoms of animal-induced asthma (5).

Laboratory workers can have laboratory animal allergies (LAA) to prealbumin and

albumin, which are derived from mouse urine and skin (16). Aeroallergens from mice

are highest during cage changing, while handling male mice on an unventilated table, and

while dumping the bedding from the cage without a dumping station.

A study by Sagakuchi et al. (16) revealed that using female mice, filter-top covers,

and corncob bedding could reduce LAA. Prealbumin and albumin levels were reduced

by 90% and 40%, respectively, when using female mice versus male mice. Using a filter-

top cover rather than no cover reduced prealbumin by 90% and albumin by 60%. When

wood shavings were replaced by corncob bedding prealbumin dropped 57% and albumin

by 77%. Other studies have shown that using IVCS and carrying out animal husbandry

and research procedures in ventilated cabinets can reduce exposure to aeroallergens (5).

A healthy macroenvironment is imperative for the overall facility. The

environmental conditions in the cage and room directly affect how an animal will









respond to laboratory procedures (17). Apposite housing and management are critical to

animal welfare, the quality of research data and teaching or testing programs where

animals are used, and the health and safety of employees (18).

The Microenvironment

The microenvironment should meet the various needs of the mice. First, the

primary enclosure should allow for normal physiological and behavioral needs of the

animals. These include allowing for maintenance of proper body temperature, urination,

defecation, normal movement, and postural adjustments. The cage should be large

enough so that the mouse can turn around and make typical movements. Second, the

primary enclosure should allow for social interaction and hierarchical development. In

addition, the cage should provide a clean, dry, safe area with adequate food, water, and

ventilation. Lastly, the personnel should be able to view the animal with minimal

disturbance (15). Taking into account all of these measures can help ensure a healthy

microenvironment.

The microenvironment's condition has been observed scrupulously due to the

advent of the filter-top. Before the filter-top, observations of the macroenvironment were

sufficient to maintain the health and well-being of the animals. In today's animal facility

it is inadequate to supply the macroenvironment with 15 ACPH, keep the temperature

and relative humidity at the recommended levels, and then completely disregard the

conditions in the microenvironment (11). This is because air exchange occurs at the

junction of the lid with the cage and not through the filter medium. A study by Keller et

al. showed that the lid reduced the air exchange rates within the cages to one ACPH

while the room was ventilated at 12 ACPH (19).









With such an airtight environment the cage can easily become contaminated. There

are four major environmental quality indicators within the cage. The first is ammonia,

which is produced from the urea found in the animal's excrement. Second is carbon

dioxide, which is generated as a metabolic waste product. Third is moisture, indicated by

relative humidity, from respiration, excrement, and the drinking water for the mice.

Additionally, thermal loads from metabolic activity can contribute to a rise in cage

temperature (11). In general, ammonia, carbon dioxide, relative humidity, and

temperature are used to assess the microenvironmental conditions.

The level of air quality is dependent on a number of factors, which include but are

not limited to population size, strain and stock of animal, location of a cage on the rack,

type of filter, airflow within the room, and relative humidity (12). Lipman identifies four

ways to address poor microenvironmental air quality (6):

1. Change cages at sufficient frequencies.
2. Use contact bedding with desirable performance characteristics.
3. Reduce the macroenvironmental relative humidity.
4. Increase the macroenvironmental temperature (dry-bulb) without altering the
moisture content in the air.

Choi et al. determined that restricting the number of animals per cage, regularly

changing soiled bedding, and increasing ventilation were techniques to prevent relative

humidity and concentrations of carbon dioxide and ammonia from increasing in static

cages (12).

Ammonia

The build-up of intracage ammonia concentrations is the primary reason for

ventilating a microisolation cage (5). Above a certain moisture threshold the urease-

positive bacteria grows, leading to ammonia production (15). The moisture threshold is









dependent on the type of bedding used, for example, pine shavings, recycled pulp, or

corncob bedding.

Enteric bacteria produce ammonia through two possible mechanisms of enzymatic

activity. In the first mechanism, bacterial urease acts as a catalyst for the hydrolysis of

urea to ammonia and carbamate:

Urea .e.se > Ammonia + Carbamate (1.1)

For the second mechanism, D- and L-amino acid oxidases remove the amino group

from amino acids to form keto acid and ammonia (12).

Ammonia levels. The human ammonia threshold limit value (TLV) used by

NIOSH of 25 ppm was established through the work of Gamble and Clough (20). A

TLV is the concentration to which humans can be exposed to for 8 hours a day 5 days a

week without any harmful effects (21). The Occupational Health and Safety

Administration (OSHA) standards for ammonia are 50-ppm time-weighted average

(TWA) and 35-ppm short-term exposure. The human TLV is accepted for animals,

although for mice the concentration capable of reducing respiratory rate by 50% (RD5o) is

approximately 300 ppm (22).

Effects of ammonia. Ammonia acts as an irritant and can alter or destroy the

tracheal epithelium. More specifically, when the epithelium becomes irritated, the cilia

are paralyzed, mucus flow is altered, and the surface layers of the epithelial lining are

destroyed (20). The epithelial thickness also changes, which increases the airflow and

deposition of airborne particles to that area (21). The abnormal increase in the tracheal

epithelium is dependent on the amount of ammonia and how long the animals are

exposed to it.









Gamble and Clough documented the effects of ammonia on the rat tracheal

epithelium (20). After four days of exposure at 200 + 50 ppm, there was a transitional-

stratified appearance to the epithelium and irregularities were noticeable. Some gross

changes were noted after eight days; the cilia disappeared, stratification increased, folds

formed on the surface of the lumen, and the amount of mucus increased. At twelve days

of exposure, the epithelial thickness increased and there was an acute inflammatory

reaction with increased cellularity and alteration of cell types.

Few studies showing long-term effects on mice exposed to ammonia exist and the

amount, which causes harmful effects, seems to vary for different studies done on rats.

For example, rats exposed to ammonia at 180 ppm for 90 days did not show any

problems, while in other studies a level greater than 25 ppm promoted growth of infective

agents in the respiratory tract (5).

Coon et al. completed rat inhalation studies on ammonia for both repeated and

continuous exposure (23). Repeated exposure (30 repeated exposures, 8 hours/day, 5

days/week) to 0.155 ppm of ammonia produced no adverse effects. At 0.770 ppm of

repeated exposure there were nonspecific inflammatory changes in the lungs of rats.

Under continuous 90-day exposure, 0.040 ppm led again to nonspecific inflammatory

changes in the lungs. At 0.127 and 0.262 ppm the same changes were seen in the kidneys

and lungs. At 0.455 ppm, 32 of 51 rats died by day 25 of exposure and 50 by day 65,

when the experiment was terminated. The rats showed mild signs of dyspnea and nasal

irritation. At 0.479 ppm, 13 of 15 rats died and the following conditions were found:

* Focal or diffuse interstitial pneumonitis
* Calcification of renal tubular and bronchial epithelia
* Proliferation of renal tubular epithelium
* Myocardial fibrosis and fatty changes of liver plate cells










Ammonia can also be harmful for rats that already have a weakened immune

system. For example, Broderson et al. found that ammonia plays an important role in

increasing natural murine respiratory mycoplasmosis (MRM) in rats infected with

mycoplasma pulmonis (24).

Despite the limited number of studies of chronic exposure on mice, ammonia has

been monitored in experiments where other parameters were the independent variable. In

an early experiment, ammonia concentrations were 400% higher in a cage with a punched

lid than one with a wire mesh lid (21). White and Mans (25) found that at low ammonia

concentrations in their experiments, there was little to no systemic accumulation of

environmental ammonia in the animals. After four days of exposure with an ammonia

concentration greater than 200 ppm, histological changes in the respiratory tract were

visible. Ammonia levels greater than 500 ppm were considered lethal. Choi et al. (12)

investigated the effect of population size on intracage ammonia levels and did not detect

ammonia in the ventilated cages for the duration of the study, 32 days.

According to NIOSH, mice exposed to ammonia for 6 hours/day for 5 days

showed signs of nasal lesions including hypertrophy, hyperplasia, epithelial erosion,

ulceration, and necrosis (22).

Ammonia production. The production of ammonia is affected by a number of

factors (21, 25):

* Number of animals in a cage
* Frequency of bedding changes
* Ambient temperature
* Relative humidity
* Time of day
* Type of caging
* Ventilation rate and air flow









Improving cage-washing procedures and animal room cleanliness can reduce the

concentrations of ammonia producing bacteria. Choosing a different strain of mice can

also help reduce ammonia concentration, as the formation of ammonia is strain

dependent. Using female mice instead of males can also lower ammonia levels. One

study showed that males produce noticeably more ammonia than females when housed

on vermiculite, pulp, and pine shaving beddings (21).

Carbon Dioxide

Carbon dioxide is a metabolic byproduct of respiration and is generally used as a

metric to determine whether there is enough fresh air in a conditioned space. The carbon

dioxide concentration in atmospheric air is roughly 300 to 350 ppm, so comparably low

levels are not harmful (26). The activity level of animals, population density, and air

exchange rate with the macroenvironment influence formation and accumulation of

carbon dioxide (12).

Carbon dioxide generation. A 25 g (0.88 oz) resting mouse consumes 1.65 ml

(0.06 fluid oz.) of oxygen per gram of bodyweight per hour and converts 1 ml (0.06 in3)

of oxygen to 1 ml of carbon dioxide. Five 30 g (1.05 oz.) mice housed in a filter-topped

Type II cage, 350 cm2 (54 in2) and 19 cm (7.5 in) high, generated 250 ml (15.26 in3),

approximately 37,000 ppm, if unventilated or undiluted, in one hour (27). In a study by

Krohn and Hansen (26), mice housed in a static filter top cage with a stocking density of

20 g/L stabilized at a carbon dioxide level of 5000 ppm after two hours. They also

measured carbon dioxide concentrations in IVCS cages without ventilation. The level

reached values between 20,000 and 80,000 ppm within two hours.

Carbon dioxide recommended levels. There are currently no official limits for

acceptable exposure of rodents to carbon dioxide. The guideline for humans of 5000









ppm is applied most often to rodents (26). However, since rodents are adapted to live in

tunnels where carbon dioxide levels can reach 14,000 ppm, higher values have been

investigated (5). It is recommended that intracage carbon dioxide for IVCS and static

cages not exceed 5,000 and 30,000 ppm, respectively (5). Lipman states that carbon

dioxide levels can be up to 4,000 ppm higher than those observed in the

macroenvironment when housing the maximum number of mice (6). It is also advised

that if animals are exposed to a level above 15,000 ppm, which is significantly higher

than the atmosphere, they should be used for experimental purpose with caution and

allowed a few days of recovery after exposure (27). Levels less than 30,000 are

acceptable for studies involving physiological or biochemical parameters, while when

between 30,000-50,000 ppm animals should be given ample time to recover. Exposures

greater than 50,000 ppm should not be accepted because the impact on the animals may

be harmful and irreversible (26).

Carbon dioxide effects. The reaction of animals exposed to carbon dioxide

mimics a stress reaction, with elevated serum corticosterone levels, increased respiration,

reduced numbers of eosinophils and lymphocytes, and a fight-or-flight reaction along

with the release of adrenaline. Humans exposed to less than 10,000 ppm showed only

minor effects, which then normalized after 10-15 days. Animals and humans exposed to

higher concentrations, 10,000-15,000 ppm, experienced the stress reaction described

above. Rodents exposed to levels greater than 30,000 ppm had elevated respiration rates

and high levels of circulating corticosterone, indicating physiological and hormonal

changes (27).









Other Contaminants

Ammonia and carbon dioxide are only two of many contaminants in laboratory

animal cages. Other contaminants found in cages include acetic acid, sulfur dioxide,

formaldehyde, dimethylamine, ethanol, ethylene glycol, methane, and hydrogen sulfide.

Experiments have found uncharacterized air contaminants in isolator cages (6).

Acetic acid. Perkins and Lipman in their comparison study of bedding materials,

detected acetic acid (mean = 0.86 ppm) in static isolator cages with and without mice

containing corncob bedding (10). Acetic acid was off-gassed, presumably from the decay

of vegetative material related with the corncob bedding, rather than from bacteria

associated with the mice (6).

Acetic acid, while not always present in static cages, has a low threshold limit

value. The permissible exposure limit was set by OSHA for acetic acid at 10 ppm for an

8-hour TWA. A 10 ppm 8-hr TWA was also set by NIOSH, as well as a 15 ppm short-

term exposure limit (15 minutes). Exposure can occur through inhalation, ingestion, eye

or skin contact, and absorption through the skin. The vapors cause eye, skin, mucous

membrane, and upper respiratory tract irritation. Mice exposed to 1,000 ppm of acetic

acid vapor had eye and upper respiratory irritation (28). Decreased lung mechanics were

observed in guinea pigs exposed to 5 to 500 ppm of airborne acetic acid for one hour

(10).

Sulfur dioxide. Perkins and Lipman found sulfur dioxide (mean = 0.42 ppm) in

static isolator cages, but only in the presence of mice and corncob bedding (9). The

OSHA standard for sulfur dioxide is 5 ppm averaged over an 8-hour work shift, while

NIOSH recommends that the limit be reduced to 0.5 ppm for the TWA for up to a 10-

hour work shift for a 40-hour week (29). Sulfur dioxide can affect the body if it is









inhaled or comes in contact with the eyes or skin. As a gas it is a severe irritant of eyes,

mucous membranes, and skin. It rapidly forms sulfurous acid on contact with moist

membranes (16).

Other contaminants. Mild inflammatory changes, primarily in the lungs, were

noted when rats were exposed continuously to formaldehyde (0.0046 ppm),

dimethylamine (0.009 ppm), and ethanol (0.086 ppm). When exposed repeatedly to

ethylene glycol at 0.010 and 0.057 ppm no changes were seen. After 8 days of

continuous exposure to 0.012 ppm of ethylene glycol, two out of fifteen rats suffered

corneal damage with apparent blindness (23). Methane and hydrogen sulfide

concentrations were evaluated in static isolator cages. Methane levels were greater than

500 ppm after seven days, while no increase in hydrogen sulfide was detected. The

physiological relevance and effects of these two gases on mice are still unclear (6).

The aforementioned contaminants are a sampling of what could be present in a

laboratory animal housing environment. The presence of any one of these contaminants

is dependent on the combination of factors such as type of bedding, cage type, and strain

of mice. A well-monitored cage is a key step to ensure a healthy and safe environment

for all involved.

Contact Bedding

Bedding in animal cages is a controllable environmental parameter, which can

influence animal welfare and research data (18). Beddings should be chemically and

biologically inert, contaminant free, highly absorptive, nontoxic, dust-free, compatible

with the research study, easily disposable, and inexpensive (10,15). Additionally, the

ideal contact bedding should enhance the physical and psychological well-being of the

animal, while not influencing it biologically (21).









Bedding materials produce environmental pollutants, such as ammonia. Ammonia

production is influenced by the following properties of bedding: particle size, absorption

properties, and the presence ofurease or a urease activator. The particle size plays an

important role in desiccating fecal pellets and thus reducing ammonia production.

Potgieter and Wilke state that because large particles have a larger exposure area they

may dry faster (21). Smaller particles, however, have greater surface to volume ratio and

it seems they would dry faster. Urease, an enzyme that catalyzes the hydrolysis of urea,

is widely distributed in plants, which are the source for most bedding materials (21).

Ammonia build-up in cages can be controlled by the frequency of bedding changes,

as the two are inversely related (21). The Guide for the Care and Use of Laboratory

Animals, henceforth referred to as The Guide, recommends that soiled bedding should be

removed as often as is necessary to keep animals clean and dry. The frequency and

intensity of cleaning and disinfection should depend on what is needed to provide a

healthy environment for an animal (18).

The kind of contact bedding chosen can affect air contaminants such as ammonia

(18). Common contact beddings include recycled paper, ground corncob, cellulose, and

wood chips. Perkins and Lipman evaluated several contact beddings in static isolator

caging, with 15 room ACPH, and four mice per cage (10). The beddings in the study

were ranked from lowest to highest mean ammonia concentration as follows:

* Corncob
* Virgin cellulose pelleted
* Recycled paper
* Hardwood chip
* Virgin pulp loose
* Reclaimed wood pulp
* Pine shavings
* Aspen shavings










Corncob proved to be the best bedding under these conditions as no ammonia was

detected after seven days of exposure. Corncob bedding is among the most popular

beddings currently used.

Relative Humidity

Relative humidity is the ratio of the partial pressure of water vapor, pv, in a given

moist air sample to the partial pressure in a saturated moist air sample, pg, at the same

temperature and total pressure,


P= .(1.2)


At room temperature, 250C (770F), the pressure of saturated water vapor is 3.2 kPa (0.5

psi).

Monitoring relative humidity is another way to help control the microenvironment.

Higher macroenvironmental relative humidity leads to higher intracage relative humidity,

which increases ammonia production within the cage (10, 30). The Guide recommends a

relative humidity range from 30-70% (18).

A relative humidity threshold has been found where above this level ammonia

production is independent of contact bedding. The threshold, however, varies with

contact bedding (30). For example, when the macroenvironmental relative humidity was

greater than 70%, the ammonia generation curve for a static isolator cage was similar

between pine shavings and corncob bedding. When the relative humidity was reduced to

60% the slope of the corncob reduced to zero for seven days while there was no change

in the pine shavings curve (6). Potgieter and Wilke discovered in their experiments that









when relative humidity and temperature reached 50% and 21C (700F) ammonia

production increased (21).

Ventilation

An average mouse (25-30 g) inhales approximately 35 liters (45 g) of air in a 24-

hour period, which is more than the total weight of its food and water. The quality,

quantity, and distribution of air are more directly associated with the animal's health,

comfort, and overall well-being than other environmental factors (7). In addition, air

associated pollutants can negatively affect the animals quality of life and general welfare

(21).

The Guide states that the purpose of ventilation is to (18):

* Supply adequate oxygen
* Remove thermal loads caused by animal respiration, lights, and equipment
* Dilute gas and particulate contaminants
* Adjust moisture content of room and cage air
* Create static-pressure differential between adjoining spaces

Ten to fifteen fresh ACPH are recommended for secondary enclosures and have

been the standard for many years. This guideline does not take into account possible heat

loads, species, size, number of animals, type of bedding or frequency of cage changing,

room dimensions, or efficiency of air distribution (18).

For individually ventilated cages, cage air change rates should be adjustable from

30 to 100 ACPH depending on the number of mice and changing frequency. The air

velocity at the inlet to the cage should be less than 15 m/min. (50 ft/min.), which is

consistent with still air. This reduces the risk of high-velocity air-cooling and

dehydration of cage occupants (5).









Reeb et al. (31) studied the impact of room ventilation rates on microenvironmental

parameters for static isolator cages. The study found that the microenvironment

maintained adequate levels of ammonia, carbon dioxide, and relative humidity at low (5

ACPH) room ventilation rates. They also discovered that increasing the room ventilation

rate had minimal effect on intracage ventilation except for cages on the highest row just

below the fresh air supply. Increased room ventilation did, however, decrease the

humidity in the room and cages. For example, with 5 room ACPH the relative humidity

was at 50%, but dropped to 22% as the room ACPH increased to 20.

Ventilation rate can be an important factor for controlling environmental ammonia

concentration. However, studies indicate that changes in ammonia concentration and

ventilation rate are not linear. White and Mans (25) found that the mean ammonia

concentration in unventilated cages did not vary in direct proportion to the room air

exchange rate. In a study by Besch (32), doubling and tripling the room ventilation rate

did not produce proportional decreases in ammonia concentration. Serrano also found

that increasing the ventilation rate did not proportionally decrease the ammonia

concentration (7). The lack of linearity is most likely correlated to physical limitations

placed on air motion patterns within a room, which is a function of the type of air

diffusion system and the face velocity of air from the diffuser at varying room air

exchange rates (25).

Additional Environmental Factors

In addition to the environmental factors previously described there are a few others,

which create a suitable living environment for mice. These include temperature, cage

space, noise, and light levels.









Temperature. Since rodents are warm-blooded animals, they must maintain their

body temperature within normal variation for their overall well-being (18).

Recommended dry-bulb temperatures for mice are 18-260C (64-790F). The cage

temperature may be higher than the macroenvironment due to animal heat load, heat

transferred from fan motors, and inefficient cage ventilation rates (5).

Cage space. The recommended cage space for lab mice as defined by the Guide is

listed in Table 1-1. Solid-bottom caging, with bedding is suggested as it is preferred by

rodents (18). It is important to follow the recommended space guidelines as the number

of animals in a cage positively influences ammonia production, carbon dioxide levels,

temperature, and relative humidity (12, 21).

Table 1-1 Recommended space for lab mice
Weight, g Floor Area/Animal, in2 Height, in.
<10 6 5
Up to 15 8 5
Up to 25 12 5
>25 >15 5

Noise and light levels. The control of noise and light levels is primarily for the

comfort of the mice. Mice can hear frequencies ranging from 80 to 100 kHz, but are

most sensitive to 15 to 20 kHz and 50 kHz. They hear high frequency and ultrasound,

which is why intracage ultrasound should be minimized. Mice have adapted to low light

levels of approximately 40-60 lux (3.7-5.6 candles). In comparison, ordinary office

lighting is less than 500 lux (46 candles) (5).

Previous Environmental Studies

There are many factors that contribute to the animal's environment as depicted in

the previous section; and different combinations offer varying results. For example, one

combination of contact bedding, bedding change frequency, and ventilation rate may









provide a suitable environment, while altering one of those factors may allow ammonia

production to occur earlier. The following is a review of previous environmental studies

illustrating the numerous combinations that have been tried to improve the animals'

environment.

Ammonia concentrations in filter-top cages. Serrano (7) was influential in the

evolution of the modern cage with his study on static isolator cage types in 1971.

Through his study he determined the effect of different types of covers on the distribution

of gases in cages with varying population sizes. Four types of filter tops were used:

fiberglass, molded laminated polyester, and two types of steel-wire mesh (40 by 40 and

20 by 20). In one experiment, eight mice were housed per cage with corncob bedding

and contaminant gas levels were measured after seven days. Carbon dioxide

concentration levels were less than 4000 ppm for all types of filter tops. For the polyester

type filter, which is similar to present day filter tops, the mean concentration on the

seventh day for cages with 4, 8, and 16 mice were <2, 21 + 18, and 90 28 ppm,

respectively. Ammonia concentrations varied significantly from cage to cage as seen in

Table 1-2 with the 40 by 40 mesh having the highest concentration.

Table 1-2 Ammonia concentration levels after seven days in static isolator cages
[NH3]
Filter Top (ppm)
Fiberglass 63 33
Molded Laminated Polyester 21 + 18
40 by 40 Mesh 177 64
20 by 20 Mesh 35 + 35

Comparison between macro- and microenvironment. Murakami (17) in 1971,

compared the environment within the cage to the ambient air. He found that changes in

temperature and relative humidity within the cage paralleled changes in the ambient air,









with negligible difference between the internal and external temperature. Relative

humidity and ammonia concentrations, however, were notably higher than in the room.

Decreases in room relative humidity. Lipman (11) conducted an environmental

study on filter-top cages, specifically Sedlacek-type. First, microenvironmental

parameters were measured while the room was held constant at 50% relative humidity.

Then, the room relative humidity was decreased to 20% and the same parameters were

measured.

Below 50% relative humidity, the cage relative humidity was 20% higher,

ammonia concentrations were 150 ppm higher, and carbon dioxide concentrations were

2300 ppm higher than in a cage without a lid. Ammonia was detected in the cages on day

4. When the room relative humidity was decreased to 20%, there was a 15% decrease in

cage humidity, the mean weekly relative humidity was 58%, and ammonia was less than

20 ppm and not detected until day 7.

Static isolator cages and strains of mice. Hasenau et al. (33) compared four

different static isolator cages for microenvironmental temperature, relative humidity, and

ammonia concentrations. Three cages had polycarbonate bases and lids with Reemay

2024 filter material, while the fourth was used as a control without a filter. Comparisons

were made of BALB/c and CD-1 same sex mice at four and two per cage under varying

microenvironmental conditions. The following parameters were used in the study:

Room ACPH: 20
Room Temp: 22.8+1.7C (standard RH)
24.10.70C (<40% RH)
Relative Humidity: 51.58.2% (standard)
22.77.7% (below normal)
Bedding: Autoclaved hardwood bedding
Bedding Change Frequency: Every 7-9 days









Under standard relative humidity (40-70%) conditions in the macroenvironment,

the relative humidity in the filter-top cages, each housing 4 mice, ranged from 17-28%

higher than the room levels. At below normal relative humidity (<40%) levels, the cage

humidity ranged from 25-38% higher than the room.

Ammonia concentrations varied significantly between strains of mice and cage

types. In the first study, where 4 BALB/c mice per cage were utilized, all cages

accumulated less than 5 ppm by day 9. With two mice per cage the concentrations

dropped to less than 3 ppm on day 9. When the CD-1 mice were housed 4 per cage under

standard relative humidity levels, ammonia concentrations after 8 days ranged from 1.9

ppm (control) to 117.1 ppm. The ammonia levels dropped significantly for all cages

when the macroenvironmental relative humidity dropped to below 40%; ammonia levels

varied from 0.1 to 8.7 ppm. At a stocking density of two mice per cage, all cages had an

ammonia concentration of less than 5 ppm at day 8.

In conclusion, Hasenau et al. (33) determined a number of factors that could be

altered to improve the environment. They found that reducing the macroenvironmental

relative humidity reduced the ammonia production in the cages. Also, reducing the

stocking density from 4 to 2 mice decreased ammonia levels more than decreasing the

room relative humidity. There was also a significant difference in ammonia levels

depending on the strain of mice.

Population size and cage type. Choi et al. (12) studied the effect of population

size on the buildup of ammonia and relative humidity in static and ventilated cages over

time. The pressurized individually ventilated cages received 50-60 ACPH, while the

room received 15 ACPH.









Environmental parameters varied between the static and individually ventilated

cages. Ammonia levels were less than 1 ppm throughout the 32-day study for the

individually ventilated cages housing both 2 and 4 mice. Relative humidity increased

slightly with the number of mice. In the static cages, no ammonia was detected in cages

with 1 or 2 mice after 8 days. The relative humidity increased with the number of mice,

which in turn increased the ammonia levels. Levels with 3 and 4 mice were 5.5 ppm

after 8 days. For all the cages, the corncob bedding appeared dry throughout the study

and there were no wet areas in one spot, indicating that the mice did not have one spot to

urinate.

Comparison of contact beddings. Potgieter and Wilke (21) investigated three

different contact beddings for dust content, dust generation, moisture absorption

properties, and ammonia production. The contact beddings: vermiculite, pine shavings,

and unbleached eucalyptus pulp were chosen because they were readily available in

South Africa, the location of the study. The room was kept at 24.71.1C (76.50F) and

51.35.3% RH and received eight ACPH. One hundred forty-four adult inbred

conventional BALB/c mice were divided among the three bedding types and housed in

static isolator breeding cages that were changed weekly.

Ammonia concentrations were surprisingly low throughout the study and never

exceeded 3.5 ppm. The lowest ammonia level (<1 ppm) measured on day 7 was from the

eucalyptus pulp. Potgieter and Wilke (21) do not recommend using the vermiculite as

contact bedding due to the quantity of dust it produced. They advise using the eucalyptus

pulp due to its moisture absorption properties and low levels of ammonia and dust.









IVCS and absorbent bedding. Huerkamp and Lehner (9) characterized and

compared microenvironments of three IVCS and a static isolator cage with ammonia-

inhibiting contact bedding to a standard static isolator cage containing corncob bedding.

The cage ACPH for the three IVCS were 74, 106, and 112, respectively. Each IVCS was

changed every 14 days, while the static cages were changed weekly. The room was held

constant at 22.4 0.30C (720F) and 42 6% relative humidity with 15 ACPH.

With the use of five and ten percent of the absorbent bedding in the static isolator

cages, the ammonia detection was delayed by one day. However, over the course of six

days the ammonia production was not altered. Ammonia was not detected in the three

IVCS after seven days and after 14 days the levels were still low. Conversely, ammonia

was detected after four days in the static cages and exceeded 100 ppm after seven days.

Carbon dioxide was reduced in the IVCS (1050 1650 ppm) compared to an average of

2050 ppm in the static cages. Methane was detected in all cages at an excess of 500 ppm,

while hydrogen sulfide was not detected.

Comparison of individually ventilated cages. Hoglund and Renstrom (34)

evaluated two different IVCS (BioZone VentiRack and Tecniplast Sealsafe) for ammonia

concentrations after two weeks, carbon dioxide build-up during a one-hour simulated

power failure, and the ability to maintain a positive or negative pressure differential for

long periods of time. Male mice, 10 weeks old, were used in the study, housed three per

cage. Aspen wood shavings were used for the contact bedding. The room was held at

221C (71.60F) and 555% RH and received 17 ACPH.

The VentiRack from BioZone provided a more uniform and balanced differential

pressure, but the systems exhibited similar behavior in all other areas. Under either the









negative or positive pressure differential the ammonia content in the cages was less than

10 ppm after 10 days when the bedding was not soaked. If the bedding was soaked, the

ammonia concentration remained high regardless of the ventilation rate.

Carbon dioxide did not build up to harmful concentration levels in the one-hour

simulated power failure due to the filter-top cages that were used with the IVCS.

Ventilation and frequency of bedding changes. Reeb et al. (13) evaluated the

microenvironment in pressurized individually ventilated (PIV) cages under two different

conditions: varying cage air change rates and reduced frequency of bedding changes.

Cage ventilation rates were held constant for 1 week at 30, 40, 60, 80, and 100 ACPH.

Bedding was not changed for 26 days. Two groups of mice were evaluated: 9-11 week

old males and trio groups for mating with pups less than 14 days old. The bedding was

autoclaved white pine shavings. The microenvironmental parameters measured were

temperature, relative humidity, ammonia, and carbon dioxide.

The results from increased cage ventilation show that the environment improves

with more circulation. Ammonia and carbon dioxide decreased significantly with

increased ventilation rates. For all ventilation rates the ammonia level was less than 3

ppm. Relative humidity was significantly higher at 30 and 40 ACPH, while it was not

significantly different for 60, 80, and 100 ACPH. For less than 60 ACPH the temperature

was 25.00.02C and dropped significantly at 80 ACPH to 23.30C.

Higher ventilation rates could reduce frequency of bedding changes to once every

two weeks. Mean ammonia concentrations stayed low for 21 days and increased to 12

ppm between days 21 and 26. The relative humidity was highest at day 21 (45%) and









decreased by day 26. Carbon dioxide and temperature fluctuated, but did not increase in

relation to the number of days with soiled bedding.

In a similar study, Reeb-Whitaker et al. (35) compared three different cage-

changing frequencies (7, 14, and 21 days) at three different cage ACPH (30, 60, and 100).

Twelve breeding pairs and twelve breeding trios were evaluated for general health over

seven months. Pressurized individually ventilated cages with white pine shavings were

used to house the mice. Ammonia was greater than 25 ppm at 30 ACPH at all

frequencies of bedding changes and at 60 ACPH after 21 days. The pup mortality rate

was higher when cages were changed every seven days. Reduced frequency of cage

changes had no effect on the following health areas: weanling weight, animal growth,

plasma corticosterone concentrations (important for carbohydrate and protein

metabolism), immune function, breeder mortality, and breeder productivity.

It is evident that many environmental studies have been done which combine a

wide range of environmental parameters. Appendix A includes a comprehensive matrix

classified by the dependent variables in each study.

Cost Analysis for Current Husbandry Practices

Besides maintaining a healthy environment, many of the previous studies were

driven by economics. One of the major cost savings areas is the frequency of bedding

changes. For example, at the University of Florida Animal Care Services static

microisolator cages are changed twice a week and not autoclaved, which translates into

$116 per cage per year. Agrawal (36) recommends decreasing the frequency from 3.5

days to 7 days and autoclaving the bedding, which will reduce the cost to $76, a savings

of $40 per cage per year. Likewise for individually ventilated cages, if the bedding

change frequency can be extended from 14 to 21 days a cost savings of $16 per cage per









year could be realized. Considering the Animal Care Services current housing needs,

$100,300 could be saved per year.

With real-time monitoring of cages, the bedding change frequency could be

reduced and savings realized without compromising the health and welfare of the animals

and laboratory personnel. The goal of this work was to explore the possibilities of

continuous monitoring of laboratory animal cages through the use of field effect

transistors.

Air Sampling Techniques

Presently, there is no standardized sampling method for rodent aeroallergens nor is

there a uniform procedure for measuring and quantifying rodent allergen exposure in

rooms with IVCS systems (5). Current husbandry practices rely on environmental

measurements taken in studies and regulated by the Guide (18) to control aeroallergens.

Environmental measurements in studies were typically taken with a gas-sampling pump

as seen in Table 1-3. While these sampling pumps have low measurement error, +2.5%

(35), sampling is often done on an infrequent basis due to cost and time restraints. For

example, in one study measurements were taken three times between 1:00 pm and 5:00

pm on days 6, 13, and 20 of a 21-day cage changing cycle. Both experiments and

husbandry practice would benefit through continuous or more frequent monitoring of

cages.

Table 1-3 Gas sampling pumps used in environmental studies
Type Manufacturer Reference
Gas Analyzer, Model 1302 and 1303 Briel and Kjar (35)
Aspirating Pump #8014-400A Matheson Gas Products (12)
Multi Gas Detector Drager (21)
Toxic Gas Monitor, Model SC-9 Riken Keiki (34)









One way to provide continuous monitoring in an animal cage is through low-power

miniaturized gas sensors offered as field effect transistors (FET). For background

purposes, the following section includes general information on semiconductor

properties, field effect transistors, and surface reactions between sensing films and

ammonia or carbon dioxide.

Semiconductors

In general FETs operate on the principle of electrical manipulation of fields. The

field manipulation is controlled by a gate, which acts on the conduction of carriers in a

semiconductor channel (37).

Semiconductors are materials consisting of elements from group IV of the periodic

table with electrical properties lying somewhere between insulating and conducting

materials. Conducting material is characterized by a large number of conduction band

electrons that have a weak bond with the basic structure of the material. Therefore an

electric field easily transmits energy to the outer electrons and allows the flow of electric

current (38).

Semiconductors act as conductors when the electrons possess enough energy to

exceed the energy gap, Eg, between the valence band, the energy level filled by electrons

in their lowest energy state, and conduction band, the unfilled energy level into which

electrons can be excited to provide conductivity as seen in Figure 1-3 (39).

Semiconductors have a lattice structure, which is characterized by covalent

bonding. Whenever a free electron leaves the lattice structure, it creates a positive charge

or hole. Electrons move to fill the holes, consequently creating more holes. When

voltage is applied electrons move towards the positive band, while holes shift towards the









negative band. The movement of electrons and holes conducts current. The number of

electron-hole pairs determines the conductivity according to the following relationship,

o" = nequ, + nhqh,, (1.3)

where ne and n are the number of electrons in the conduction and valence bands,

respectively, je and ph are the mobility of electrons and holes, respectively, and q is the

charge (39).

Conduction Band



Electron Energy
Gap


Valence Band


(D 3 SI
I ~- N I
*Q=ll


Figure 1-3 When excited, electrons move from the valence band to the conduction band
across the energy gap.

The number of charge carriers in semiconductors is controlled by temperature. At

absolute zero, all the electrons are in the valence band, while the conduction band is

empty. As the temperature increases it is more likely that an energy level in the

conduction band will be occupied. The number of electrons in the conduction band is

equal to the number of holes in the valence band and is related to temperature, Tby

(-E g
n = n=n= exp 2- (1.4)


where k is Boltzmann's constant and no is a constant (39).

If the voltage source or exciting energy is removed, the holes and electrons will

recombine over a period of time, designated as,









f-t'
n= nexp (1.5)


where t is the time after the field is removed and r is a constant known as the

recombination time (39).

The behavior of an intrinsic semiconductor cannot be accurately controlled due to

its sensitivity to slight variations in temperature. Therefore adding impurities or dopants

that determine the number of charge carriers can create an extrinsic semiconductor. An

n-type has an extra electron that lowers the energy level, whereas a p-type does not have

enough electrons and a hole is created (39).

Field Effect Transistors

A transistor is a three-terminal semiconductor device, which performs two

functions: amplification and switching (38). Transistors have three connections, where

the voltage on (current into/out of) switch has the effect of controlling the ease with

which current can flow between the other two terminals (40). The effect is to make a

resistance whose value can be altered by the input signal. The patterns of signal

fluctuation can be transferred from a small input signal to a larger output signal.

Specifically for a metal oxide semiconductor-FET (MOSFET), there are three terminals:

gate, drain, and source (Figure 1-4). The gate is a metal film layer that is separated from

the bulk by a thin oxide layer. When a voltage is applied to the gate an electric field is

created which repels positive charge carriers away from the surface of the bulk in which

the negative charge carriers dominate and are available for conduction. By increasing the

gate voltage, the depth of the channel can subsequently be increased.









drain

I P-Type substrate I M gate
source
View from above Sandardsymbol












Figure 1-4 Schematic of field effect transistor (FET) (40).

Upon gas exposure, the surface of the sensing film registers a work function

change, which is seen electronically as a gas sensitive potential. This potential is then

added to the gate voltage and operates the transistor (41).

There are two basic types of FETs used in gas sensing: a classical FET and a

suspended gate FET (SGFET) or hybrid flip-chip FET (HFCFET). In the classical FET

the sensing layer lies directly on the gate surface, whereas in the SGFET the sensing film

is separated from the gate surface by an air gap (Figure 1-5 and Figure 1-6). In order to

ensure adequate capacitive coupling the air gap must be no larger than a few

micrometers, but at least one micrometer to allow for sufficient gas diffusion through the

channel. There must be no contact between the sensitive material and channel-insulating

layer (41).

An FET consists of two parts: an alumina substrate that contains the conducting

structures for the flip-chip contacts and a sensitive layer on a separate electrode (Figure


a The smallest amount of energy, measured in electron volts, required to remove an electron from the
boundary of an element.









1-7). The sensor film is applied to a designated area using a freely selectable deposition

process. If a heater element is needed, it is on the backside of the alumina substrate. For

an HFCFET the etching steps in thin-film technology ensure that a defined air gap for gas

diffusion is formed when the FET is mounted. An advantage of using an air gap is that

the same substrate design can be used for all types of transducers because the geometry

does not change (41, 42). Additional schematics and detailed electrical drawings can be

found in Appendix B.

Gas

Gate oo Sensing Film
Drain Source
/ II

"-Channel






Figure 1-5 Classical FET configuration where the sensing film lies in the same plane as
the gate.

Sensing Film

Vg Drain '" Source








Figure 1-6 Suspended Gate FET configuration where the sensing film is separated from
the gate by an air gap.

Gas Sensing

Field effect transistors can be used to detect a wide variety of gases by choosing the

sensing layer that reacts with or catalyzes the specific gas. The reaction mechanism,









adsorption, and desorption rates will vary for each combination and creates a challenge to

the designer to find the best combination.

Silicon chip (FET)










alumina substrate with
sensitive layer

Figure 1-7 Photograph of a hybrid flip chip FET sensor device (left) mounted on a
printed circuit board (right). (Data of Simon et al. (46))

Hydrogen detection is the most basic reaction and aids in understanding more

complex reactions involved with ammonia and carbon dioxide sensing. For sensing of

hydrogen the reaction occurs in three steps. First, the hydrogen molecules dissociate on

the catalytic metal surface of the device. Secondly, the hydrogen atoms are transported

through the metal film. Lastly, adsorption of hydrogen occurs at the interfacial layer

between the metal and insulator where a dipole layer forms. The adsorbed hydrogen

disrupts the electric fields across the device structure and is detected through changes of

the electrical characteristics of the device (43).

Ammonia reactions. For ammonia detection, one surface that is used is titanium

nitride (TiN). Ostrick et al. (44) claim that water and ammonia adsorb on the TiN film

and change the work function of the film. The TiN surface is covered with hydroxide

(OH) and water and upon exposure to ammonia the OH species is reduced and

compounds related to ammonia are formed. Two reaction mechanisms are possible for









ammonia adsorption. In the first, the ammonia removes OH from its binding site. The

second mechanism proposes that the OH groups become binding sites for the ammonia,

as seen here

NH3 )g, + OH)LUe O NH4 ),uce (1.6)

The change of the work function, AO is due to the difference of dipoles on the surface of

OH and ONH4, given by,

AO (dONH -dOH)O (1.7)

where ( is the coverage of ONH4 and dis the dipole moment (44).

Ostrick et al. (44) found that the reaction of ammonia on TiN occurred reversibly at

room temperature and was not hindered by pre-adsorption. No cross sensitivities to the

following gases were found: carbon monoxide (30 ppm), carbon dioxide (3000 ppm),

nitrogen dioxide (1 ppm), hydrogen (10 ppm), methanol (10 ppm), and acetone (10 ppm).

Sensitivity to ammonia was found to be independent of relative humidity (5-80%).

At room temperature ammonia solves almost completely in water under formation of

NH4 and OH- ions. Since the ammonia concentration is low (<100 ppm) compared with

the surface water concentration (>5000 ppm), all of the ammonia reacts with the water.

In the next reaction step, the ammonia ions may react directly with the surface or with

adsorbed hydroxide ions. Due to the excess of water molecules, the found ammonia

sensitivity may be independent of the water concentration. At higher temperatures,

however, the sensitivity to ammonia decreases. The solubility of ammonia in water

decreases at higher temperatures as well as the amount of water on the surface. The

amount of ammonia on the TiN surface is also reduced, therefore the ammonia is less

likely to react with adsorption sites and the sensitivity decreases.









A reaction mechanism proposed for ammonia detection in metal-insulator-

semiconductor (MIS) field effect devices is that adsorbed NH4' is detected on the oxide

surface in the holes and cracks of the metal film. Another proposed mechanism is that

ammonia dissociates on the catalytic surface and reacts similarly to hydrogen. The

SGFET and MIS devices differ in that, for an SGFET the response results from

adsorption of the detectable species on the surface, whereas in an MIS it occurs by

adsorption in the interface between metal and oxide. Results from Abom et al. (43)

imply that when a porous TiN film is used for an SGFET, no NH3 response is seen. Even

when the surface is covered with NH4+, the molecules cannot reach the reactant surface

SiO2 and produce a response. However, it was found that NH3 is dissociated if Pt is

present and the resulting atomic hydrogen can be detected.

Carbon dioxide reactions. Ostrick et al. (45) outlined the temperature dependent

reaction mechanisms for barium carbonate (BaCO3) as the sensing film used to detect

CO2. At 500C the reaction of CO2 is dependent on water and occurs only if water is

present, yet it is independent of the partial pressure of oxygen. When the temperature

increases, above 200C, the reaction is more complicated and is dependent on oxygen

and humidity.

The reaction mechanisms at low temperatures are predicted as follows,

OH ).suce + C2)g, <> HCO )~,ce (1.8)

HCO ),+ce + OH )suc <-> H20)g .surfce + C3 2 )bulk (1.9)

At high temperatures, for CO2 in 02 as the dry carrier gas, the reaction of formation of

the carbonate appears as,


CO2 )ga + 2 2 )gs +2e )mel <-> CO 2 )buk


(1.10)









which results in an electromotive force (emf) change of

E = Eo + k,T/2e ln(pC,) + kT/4e ln(p,) (1.11)

where k is the reaction rate constant, e is the elementary charge, andp, is partial pressure.

In terms of work function change, the reaction appears as,

AO = AO, + Sco, log(p0, ) + So, log(p ) (1.12)

where S is the sensitivity. When humidity is present, the sensitivity to oxygen is reduced

and another reaction dominates, possibly the formation of hydrogen carbonate from

carbonate, carbon dioxide, and water,

CO32 L +CO2)ga + H20)gas -> 2HCO3 )surface (1.13)

The work function equation for this reaction is

AO = AO8 + Sco0 log(pC2 ) + SHo log(pHo). (1.14)

Another surface layer used to detect CO2 is barium titanate (BaTiO3) (46). This

compound can exist with a low excess quantity of barium. The excess of barium is not

compensated by Ti- or O- vacancies or by forming low amounts of a new barium titanate

phase with a higher stoichiometry of barium. A mixture of BaTiO3 with CuO has been

reported as a highly sensitive material for CO2 sensing using the capacitance change in a

temperature range of 200-10000C. Kelvin probe measurements of BaTiO3 indicate that it

has a fast response time, less than one minute, and a sufficient sensitivity to CO2 (20

mV/decade). Similar to the BaCO3 surface, the sensitivity of CO2 is dependent on the

presence of humidity. The sensor showed significant cross-sensitivity to NO2 and drift

effects tended to increase.

Ostrick et al. (47) also investigated the different reactions occurring in a multi-

layer system, Pt/NiO/BaCO3, used to detect CO2. It was found that besides the CO2









reaction, a separate reaction to NO2 could occur at the NiO interface. Inserting inert

metal oxide layers stopped this reaction.

Summary

As developed in this chapter, there is an evident need for continuous environmental

monitoring in laboratory animal cages. By continuously monitoring the cages, laboratory

animal personnel can determine when a cage needs to be changed so that the mice are

disrupted only when necessary. Prolonged cage changing intervals also translate into

labor savings for the animal care facility. Through real-time monitoring, individually

ventilated cages can be properly ventilated; increased ventilation during active times and

reduced ventilation if the mice are inactive and air contaminant levels are low. Field

effect transistors provide a way to achieve continuous monitoring of ammonia, carbon

dioxide, temperature, and relative humidity in a laboratory animal cage. Most

imperatively, continuous monitoring can provide a better environment for the animals

and laboratory personnel.

As a prelude to the following chapters, the goals of this project were to:

* Assess feasibility of applying field effect transistors for monitoring laboratory
animal cages through the following tests

o Single analyte in air

o Time-response

o Cross-contamination

* Theoretically model the chemical kinetics and catalysis between the sensing film
and contaminant gases and more specifically,

o Define reaction mechanisms

o Determine adsorption and desorption rates

o Find position of chemical equilibrium






41


o Determine equilibrium constants

o Explore the role of diffusion
















CHAPTER 2
EXPERIMENTAL FACILITIES AND METHODS

The purpose of the experimental portion of this study was to assess the feasibility

of applying field effect transistors for monitoring microenvironments in laboratory

animal cages used in animal research facilities; specifically the work focused on

ammonia and carbon dioxide sensors.

Experimental Setup

The experimental facilities included humidification, mixing, and sensor-testing


regions as seen in the schematic in Figure 2-1.

PID
Peristaltc Temp,
Humidification P"p Controller
section
Water to increase Fiberglass
relative humidity insulation
--Heating tape


Mixtu ass beads NH3 C Humidity Temp
(N2& to ad in Sensor Sensor Sensor





2)Figure 2-1 Schematic of experimental facilities, which included a mixing, humidification,
System C sensor-testing region.onneco
Mixing Section iLa hew
Software


Figure 2-1 Schematic of experimental facilities, which included a mixing, humidification,



In general, the desired gas mixture (e.g., air, ammonia, carbon dioxide, and water

vapor) was prepared by mixing the desired gases (see Table 2-1 for stock gas) and then

passing through the test chamber. Table 2-1 provides a comprehensive list of the

equipment for the experiments. If humidification was necessary, compressed synthetic









air (Praxair) was preheated to 1000C, and then water was added through a variable flow

pump (Fisher Scientific, Model No. 13-876-4), pictured in Figure 2-2. Once humidified,

ammonia or carbon dioxide (Praxair) was added to the air system through flow

controllers (Alicat Scientific, MC Series) as pictured in Figure 2-3.

Table 2-1 Sensor testing equipment
Sensor Testing Equipment Manufacturer Part No.
Vacuum Piping --
Flow meters Alicat Scientific
Fiberglass Cloth Tape Insulation Fisher Scientific 01-472A
Variable Flow Peristaltic Pump Fisher Scientific 13-876-4
Laptop Gateway
Data Acquisition System National Instruments NIDAQPad-6020E
Labview Software National Instruments Labview 7.1
Controller Omega CN1A-TC-24V
Handheld Thermometer Omega HH-26K
Heating Tape, 1/2"x4' Omega SRT-051-040
Heating Tape, 1/2"x6' Omega SRT-051-060
Multimeter Omega HHM-11
T/C TO ANALOG CONVERTER, "K" Omega SMCJ-K
Type K Connector Omega OST-K-MF
Type-K ex. Wire, 100' Omega EXPP-K-20-100
Type-K Probe Omega KQIN-116G-12
2% Ammonia in Nitrogen Praxair
75% Nitrogen, 25% Carbon Dioxide Praxair
Synthetic Air Praxair
DC Power Supply Protek 303

The combined gas stream flowed either directly over the sensor, Figure 2-4a, or

into a Plexiglas box, Figure 2-4b. The Plexiglas box was used to simulate cage

conditions where diffusion would be the main mode of mass transfer to the sensor. When

the gas stream flowed directly over the sensor, the flow rate was limited to between 1 to 2

L/min as specified by the sensor developers. Higher flow rates could induce baseline

drift and add noise to the signal. At a flow rate of 2 L/min the Reynolds number was

460, which indicates the flow was laminar.


























(a) (b)
Figure 2-2 Humidification section of experimental facilities: (a) PID temperature
controller heats the air before and after adding water, (b) Variable flow pump
adds water to the air stream.






Preheat

Humidified air




(a) (b)
Figure 2-3 Mixing section of experimental facilities: (a) Schematic of mixing section (b)
Flow meters used to control flow rate of air, ammonia, and carbon dioxide

In the diffusion case, the relevant flow rates for the calibration gases were

determined based on the air exchange rates and velocity of air leaving an individually

ventilated mouse cage. In a typical cage, air enters the cage through a small tube and

exits the cage by way of a small gap between the lid and cage bottom. The flow rate

calculations were based on the following cage parameters:

Cage air changes per hour 60
Volume of cage 7.067 L (431.25 in3)









Cage width 12.7 cm (5 in.)
Gap opening 1.27 cm (0.5 in.)
Area of gap 16.1 cm2 (2.5 in2)

Q 7.067L/min
Vce = 7.67Lmin =4.38m/ min (2.1)
age A 16.129cm2

where Q is the volumetric flow rate, A is the cross-sectional area, and Vcage is the velocity

of air leaving the cage.















(a) (b)
Figure 2-4 Two flow regimes used in sensor testing: (a) Gas stream flowed directly over
the sensor or (b) Gas stream diffused onto the sensor mounted in a Plexiglas
box used to simulate the cage environment.

The volumetric flow rate in the test section was calculated based on the velocity of

air leaving the cage. With a cross-sectional area of 9.62 cm2 (1.5 in2) the volumetric flow

rate was calculated as follows,


Qg = 9.62cm2 x4.38 m = 4.21L/min. (2.2)
mmn

In keeping with this value, the air flow rate was set to 5 L/min in the diffusion test

section. The stock ammonia and carbon dioxide were suspended in nitrogen with

concentrations of 2% and 25% by volume, respectively (see Table 2-1). The ammonia

and carbon dioxide streams were then diluted to the desired concentrations by the pure air









stream. The desired analyte concentration was entered into a Labview program (National

Instruments (2004) Labview 7.1), which subsequently calculated the necessary flow rates

and used analog output channels to control the flow controllers (Alicat Scientific). The

accuracy specification for the flow controllers was 1% of the full scale, where the full

scales for the air, ammonia, and carbon dioxide controllers were 5 L/min, 50 cc/min, and

100 cc/min, respectively. The accuracy for the compressed gas cylinders was 2% and

1.25% for the ammonia and carbon dioxide cylinders, respectively. The uncertainty

associated with the ammonia and carbon dioxide concentrations for each experimental

condition is listed in Table 2-2. Detailed uncertainty calculations can be found in

Appendix C.

The actual sensor was on a micro-scale as depicted in Figure 2-5. For ease of use

in the laboratory, the sensor was mounted on a larger box housing the electrical

connections, as shown in Figure 2-4b. In actual field operation, the sensor would be

mounted in a cage with wireless feedback to the electrical board. To prevent the signal

from drifting, the surface of the sensor was heated by means of an internal heater. The

internal heater was powered by voltage from an external regulated DC power supply

(Protek, Model No. 303).


Figure 2-5 Carbon dioxide sensor. (Scale bar = 1 inch (2.54 cm)).









Table 2-2 Ammonia and carbon dioxide concentrations and their associated uncertainties
for each experimental condition.
Air flow NH3 CO2
rate NH3 conc. uncertainty CO2 conc. uncertainty
Flow regime (L/min) (ppm) (+%) (ppm) (+%)
Diffusion 5 5 40.0
5 10 20.1
5 15 13.5
5 20 10.2
5 25 8.3
5 30 7.0
5 40 5.5
5 50 4.6
5 75 3.5
5 100 3.0
5 150 2.6
5 200 2.4
Forced 2 50 10.5
2 100 5.9
2 300 41.9
2 1000 13.6
2 2500 7.4
2 3000 6.9
2 5000 6.1
2 7000 5.8
Diffusion 5 50 4.5 3000 5.3
5 50 4.5 4800 5.2
5 100 3.0 3000 5.3

Signals from the sensors were collected through a data acquisition board (National

Instruments, NIDAQPad-6020E) and processed using Labview software (National

Instruments (2004) Labview 7.1). Inputting a two-volt signal from a regulated DC power

supply (Protek, Model No. 303) directly into the board for a six hour time period tested

the accuracy and precision of the board and software. The results of this test can be seen

in Figure 2-6.










2.04-

2.02-

2.00-

0)
50 1.98-

1.96-

0 1 2 3 4 5 6
time (hr)
Figure 2-6 Signal from DC power supply as collected in Labview, which shows precision
and accuracy of the data acquisition system over six hours.

Experimental Procedures

Before starting any of the other tests, the carbon dioxide and ammonia sensors were

tested for drift of the baseline signal. Output signals were collected every 30 minutes for

over 40 hours while air flowed over the sensor at a rate of 2.00 L/min. The sensors were

not exposed to either ammonia or carbon dioxide during these tests.

Carbon Dioxide Sensor

Single analyte tests, where the sensor was only exposed to mixtures of air and

carbon dioxide, were completed to establish a calibration curve. Since carbon dioxide

levels should not exceed 5000 ppm; the measurable range was set to 300 to 7000 ppm.

The sensor was exposed to concentrations between 300 and 7000 ppm carbon dioxide for

10 to 30 minutes at a time with a 10 minute purging cycle with air between exposures.

These tests were all conducted using the forced flow regime.

The other experiment completed with the carbon dioxide sensor tested for cross-

sensitivities to humidity. First, air flowed over the sensor for 30 minutes to establish a

baseline. Then the sensor was exposed to 3000 ppm carbon dioxide for 30 minutes. The






49


relative humidity was then increased from dry (2%) to humid (50 60%). After

humidifying the air, the sensor was again exposed to 3000 ppm carbon dioxide. A

graphical depiction of this test can be seen in Figure 2-7.


3500 100
--- CO conc.
E 3000- ---- rh
.. 75
2500 -
0
2000 -

2 1500 E

o 1000- M
0 25
en CO 500
(9
0 1 i-- -- t- 0
0 20 40 60 80 100 120 140
time (min)
Figure 2-7 Graphical depiction of experimental parameters used to test the carbon dioxide
sensor for cross-sensitivity to humidity.

Ammonia Sensor

A number of single analyte tests were conducted on the ammonia sensor in order to

establish a calibration curve. Since the recommended threshold limit value (TLV) for

ammonia is 25 ppm, the desired measuring range was 25 to 100 ppm. The sensor was

exposed to concentrations of 25, 50, 75, and 100 ppm ammonia for 10 to 20 minutes at a

time with a 10 minute purging cycle with air between exposures. These tests were all

conducted using the forced flow regime.

In actual operation the sensor will not have a purging cycle after it has been

exposed to air contaminants. Therefore, tests were conducted where the ammonia

concentration was ramped up or down to simulate gradual changes that may be seen in

the environment. The parameters used for the ramping tests are listed in Table 2-3, while







50


a graphical depiction is given in Figure 2-8. The ramping tests were completed using the

diffusion box.

Table 2-3 Ammonia concentrations used for ramping tests.
Ramp up Ramp down
NH3 Duration NH3 Duration
Concentration (min) Concentration (min)
(ppm) (ppm)
0 10 0 10
5 10 200 10
10 5 150 5
15 5 100 5
20 5 75 5
25 5 50 5
30 5 40 5
40 5 30 5
50 5 25 5
75 5 20 5
100 5 15 5
150 5 10 5
200 5 5 5
Increment back to 0 Increment back to 200

200- ,- 200-
E E
150- CL 150-
C C

100- 100-
C 0C
o 0 0 50
O 50- o 50
S o .. 1 o--

0 20 40 60 80 100 120 140 20 40 60 80 100 120 140
time (min) time (min)
(a) (b)
Figure 2-8 Graphical depiction of ammonia concentrations used in ramping tests.

Since the laboratory animal cage is not a homogeneous environment, the effects of

other air quality factors were tested. First, the ammonia sensor was tested for cross-

sensitivity to humidity. The sensor was exposed to air under dry conditions (2% rh) for

20 minutes, the relative humidity was increased to 40% rh for 20 minutes, and then this

cycle was repeated twice. Six repetitions of this cycle were completed for the forced and









diffusion flow regimes. Second, the sensor was tested for cross-sensitivity to humidity

while ammonia was present. The sensor was first exposed to ammonia for 30 minutes,

followed by 30 minutes of air, then the relative humidity was increased, and lastly the

sensor was exposed again to ammonia. A graphical depiction of the experimental

parameters used for this test can be seen in Figure 2-9.

100 NH3 conc.
RH

-E 75



o -
-E 50-



CO
( 25


0 -----
0 30 60 90 120 150
time (min)
Figure 2-9 Graphical depiction of experimental parameters used to test the ammonia
sensor for cross-sensitivity to humidity while ammonia was present.

Next the ammonia sensor was tested for cross-sensitivities to humidity and carbon

dioxide with ammonia in the system. A graphical depiction of the combination of

parameters used is shown in Figure 2-10. The introduction of each new parameter was at

least 10 minutes after the last change in experimental conditions to ensure that the effect

on the sensor was from the intended parameter. Six repetitions of this experiment were

conducted.

Time response tests were completed to determine the amount of time required by

the sensor to display 95% of a step change in gas concentration. This test also aided in

determining adsorption and desorption rates of the analyte on the sensor. To complete

this test, air flowed across the sensor for 10 minutes reaching steady state, and then









ammonia was added to the system for 10 minutes. The analyte stream was shut off and

air again flowed over the sensor for 10 minutes. This cycle was repeated six times for

each of the eight experimental conditions listed in Table 2-4.


100
,, CO
-NH3

7rh
1000 75


S50 E

S100- -

-U 25


10--0
10 O
0 30 60 90 120 150

time (min)
Figure 2-10 Graphical depiction of experimental parameters used to test the ammonia
sensor for cross-sensitivity to humidity and carbon dioxide with ammonia in
the system.

Table 2-4 Experimental parameters for time response tests (X indicates tests completed).
Flow Humidity 50 ppm 100 ppm
(%) NH3 NH3
Forced 2 X X
Diffusion 50 X X
Forced 2 X X
Diffusion 50 X X

In the last set of experiments, the surface temperature of the sensor was varied to

test for temperature effects on the response time of the sensor. The sensor was exposed

to 50 ppm ammonia for 10 minutes at 310C (1.0 V), 390C (2.0 V) and 520C (3.0 V). The

temperature of the surface was controlled by an internal heater, which was powered by an

external DC power supply (Protek Model No. 303). The output voltage was calibrated to







53


the above-mentioned temperatures. At each temperature, the experiment was repeated

six times.














CHAPTER 3
THEORETICAL MODELING

The theoretical modeling of the reactions between the sensing film and analyte

gases in the field effect transistors allowed for a number of questions to be investigated.

The first topic of research dealt with the reaction mechanism for the ammonia gas on a

titanium nitride surface. The second inquiry focused on the adsorption and desorption

rates for each reaction. The goal was to model the time response for adsorption and

desorption and then determine the position of chemical equilibrium. In the process of

conducting these queries, the equilibrium constants for each reaction were explored.

Diffusion was also investigated as it played an important role in transporting the

molecules to and from the surface so that adsorption or desorption may take place. The

experimental portion of the study was used to confirm the adsorption and desorption rates

found through the theoretical modeling.

To determine the adsorption and desorption rates for each reaction, the position of

chemical equilibrium, and the equilibrium constants, the following heterogeneous

chemistry theory was reviewed and applied to the analysis.

Gibbs Free Energy

Gibbs free energy, G, is essential in determining the driving force or spontaneity of

chemical reactions, the equilibrium constant, and the position of chemical equilibrium,

where the equilibrium position of a reaction is said to lie far to the right if almost all

reactants are used up and far to the left if scarcely any product is formed. Gibbs energy

can be described by the following equation,









G =H -TS, (3.1)

where H is enthalpy, T is the absolute temperature, and S is entropy. For an isothermal

process, the change in free energy as the process proceeds can be written as,

AG = AH- TAS. (3.2)

A chemical process will continue in the direction that decreases the free energy. For

example, if AG < 0 then the forward reaction will continue spontaneously. Likewise, if

AG > 0 then the reverse reaction will occur spontaneously. The process continues in the

direction to minimize the free energy of the system until AG = 0, at which point

equilibrium is achieved.

The free energy change can be related to the reaction equilibrium constant. First,

consider the elementary reaction

aA +bB xX + yY. (3.3)

If the reaction proceeds by a differential amount, d4, then the number of moles of each

chemical species changes according to

dnA = -ads, (3.4)
dnB = -bd (3.5)
dnx = xdE, (3.6)
dny = yd (3.7)
The total differential of mixture free energy is thus

K
dG = -SdT + Vdp + kdn (3.8)
k=l
where / is the chemical potential. At fixed temperature and pressure, this equation

converts to


dG = PJAdnA + BudnB + uxdnx + /rdn .


(3.9)








dG = -a/,d bkBd + xxd.d + yu, d (3.10)

dG = (x/.u + yu, a/ bu,)d (3.11)

)j =(xIX +yU a/A -b/B). (3.12)
d pT

Equilibrium is reached when

K =G 0. (3.13)
^p,T

Therefore at equilibrium,

x/x + yu/ a/A bB, = A/, = 0. (3.14)

Considering the chemical potential at standard state, Eq. 3.12 can be written as,

x(/u + RT n px)+ y(r + RT n p,)
) )\ t \ (3.15)
a(u + RT In PA) b(/ + RT In PB) = 0

With this result, a constraint is put on the pressures that the four gases can have at
equilibrium. Rearranging Eq. 3.15 leads to

-RTln (P) =x(PY x/ + y ao aA- bBu (3.16)
(PAY PB) eq

-RTln (PX)-PYY AGO. (3.17)


Because the right hand side of Eq. 3.17 is constant for a given temperature, the
logarithmic term must also equal a constant,

K = [(PX) (P Y (3.18)
(PA) (PB) e









known as the equilibrium constant. If K, is greater than one, the products of reaction are

favored over the reactants and the forward reaction proceeds.

Based on Eq. 3.17 and Kp, at equilibrium,

AG = -RT nK,, (3.19)

where it can be rewritten with the help of Eq. 3.2 as

AHo + TAS = RT In K (3.20)

or

H-
r+ +AS =RlnK (3.21)
T

When the reaction is exothermic, AH, /T relates to a positive change of entropy of the

surroundings and favors the formation of the products. As the temperature increases,

- AH/ IT decreases and the increasing entropy of the surroundings has a less powerful

effect. Resultantly, the equilibrium lies less to the right. If the reaction if endothermic,

the primary factor is the increasing entropy of the reaction system. The significance of

the unfavorable change of the entropy of the surroundings is lessened as the temperature

increases and the reaction can shift towards the products.

Looking at the forward and reverse rates of progress at equilibrium develops the

equilibrium concentration constant, Kc. If the reaction is at equilibrium, then,

K K Vk;
kff[Xk] k,, [xk ] (3.22)
k=1 k=1

or

k K ,Vk
K = -=1[xk] (3.23)
r 1 k=1 e









where e indicates equilibrium, and vk, is the net stoichiometric coefficient for species k in

reaction i. Concentration is related to pressure through the ideal gas law,


[Xk ] k Pk (3.24)
V RT

Equation 3.23 now becomes,


K, k= (3.25)


The equilibrium pressure constant for a general reaction is written as,


K,,. = n k) (3.26)
k=1 o e

where pO is the standard-state pressure. The two equilibrium constants can be combined

into one equation as follows,


Kc, KP,, i (3.27)


Gibbs free energy and associated thermodynamic terms were an essential part of

examining the reaction of ammonia on a titanium nitride surface.

Adsorption

The reactions between the titanium nitride surface on the field effect transistor and

ammonia occurred through adsorption, attachment of particles to the surface. The

substance that adsorbs is called the adsorbate while the underlying material is the

adsorbent. The reverse of adsorption is desorption. Adsorption can occur in two ways:

physical adsorption, physisorption, or chemical adsorption, chemisorption.

Physical adsorption is due to van der Waals interactions between the adsorbate and

the adsorbent. Van der Waals interactions have a long range, but are weak. The energy









released is on the same order of magnitude as the enthalpy of condensation, roughly 20

kJ/mol. These small energies are absorbed as vibrations of the lattice structure and

dissipated as thermal motion. A molecule will bounce around and finally adsorb to the

surface in a process called accommodation. The bonds do not break; therefore a

physisorbed molecule retains its identity (48).

In chemical adsorption the molecules adhere to the surface by forming a chemical

bond, typically covalent. In comparison with physisorption, the enthalpy of

chemisorption is ten times greater at approximately 200 kJ/mol. The distance between

the surface and closest adsorbate atom is shorter for chemisorption than for

physisorption. A chemisorbed molecule may be torn apart at the demand of unsatisfied

valences of surface atoms. The existence of molecular fragments on the surface is one

reason why solid surfaces catalyze reactions.

Chemisorption is most often exothermic, which can be proven by examining the

Gibbs equation, AG = AH TAS. For chemisorption to be a spontaneous process,

AG < 0 and because the translational freedom of the adsorbate is reduced when it is

adsorbed, AS < 0. Therefore, the enthalpy of adsorption, AH, must be negative, which

indicates an exothermic nature. The enthalpy of adsorption, however, is dependent on the

extent of surface coverage because the adsorbate particles interact with each other. For

example, if the particles repel each other like CO on palladium, adsorption becomes less

exothermic as the coverage increases. If the particles attract each other, such as 02 on

tungsten, the process becomes more exothermic as the particles cluster together.









The rate and extent to which a surface is covered are important when considering

heterogeneous reactions. Molecules will quickly cover a surface exposed to a gas. The

collision flux, Z,, can be expressed as,


= Y (3.28)
(2mnkT)

As an example, air with molecular weight of 29 g/mol at 1 atm and 250C, will have a

collision flux of 3 x 1027 m-2S-1. So for a one meter square metal surface containing 1019

atoms, each atom is struck approximately 108 times each second. The fractional

coverage, 0, is given by,

number of adsorption sites occupied V
S= (3.29)
number of adsorption sites available V,

where V, is the volume of adsorbate corresponding to complete monolayer coverage.

dO
The rate of adsorption, is determined by observing the change of fractional
dt

coverage over time.

Surface Reaction Rate Expressions

A number of classic rate expressions can be used to typify heterogeneous reactions.

These include adsorption isotherms, competitive adsorption, and dissociative adsorption.

Langmuir Adsorption Isotherm

The free gas and adsorbed gas are in dynamic equilibrium where the fractional

coverage of the surface depends on the pressure of the overlying gas. The variation of

the fractional coverage with pressure at a specified temperature is known as the

adsorption isotherm.









The Langmuir adsorption isotherm describes the equilibrium between a single-

component gas, A, and adsorbed species, A(s), at a surface. It is based on three

assumptions (48):

* Adsorption cannot proceed beyond monolayer coverage.
* All sites are equivalent and the surface is uniform.
* Ability of a molecule to adsorb at a given site is independent of the occupation of
neighboring sites.

The isotherm expression relates the fraction of surface, OA, covered by the adsorbed

species as a function of partial pressure, pA, exposed to the surface and is given as

follows,

KpA
OA KpA (3.30)
1+ KpA

At low partial pressures, the coverage of adsorbed species increases linearly with the

partial pressure. As the partial pressure of A increases, the amount of adsorbed A(s)

begins to saturate, and the coverage OA approaches unity. The monolayer has thus been

completed and further adsorption cannot take place.

An equivalent expression for the isotherm can be developed using mass-action

kinetics. For a gas molecule reacting with the surface, the adsorption process proceeds as

follows,

A + O(s) k A(s) (3.31)
k_

where O(s) represents an open site on the surface and kl and ki_ are the equilibrium

constants for the forward and reverse processes, respectively. At equilibrium the

concentration of the surface adsorbed species is constant and can be represented by









d[A(s)] 0 = k[AO(s)]- k1[A(s)]. (3.32)
dt

If the open site is related to the site density, F, by

[O(s)]= F- [A(s)], (3.33)

then at steady-state,

k,[A] = (k, + k, [A[A(s)]. (3.34)

The coverage (A then becomes,

F [A(s)] k, [A] K,[A] (335)
01 (3.35)
A F k- +k,[A] I+K,[A]

Using Eq. 3.2 the coverage can be expressed in terms of pressure as

+Kp(pA p)
0 + Kp pApO). (3.36)


Dissociative Adsorption

Some molecular species undergo dissociation upon adsorption, especially on metal

surfaces. For example, H2 dissociates on a metal surface into two surface adsorbed H(s)

atoms. Likewise, methane dissociates into CH3(s) and H(s). Dissociative adsorption is

assumed to require two open sites on the surface. The process of adsorption and

dissociation are thought to occur in a single step. The site fraction is given by

(Kp )L
A Kp (3.37)


Mass-action kinetics provides a surface reaction of

A2 + 20(s) 2A(s) (3.38)
kl


and leads to the surface site fraction on a concentration basis:









OA 1 [ (3.39)


Competitive Adsorption

In competitive adsorption, two gases A and B are considered when modeling. Both

gases are present above a surface and compete for available sites on the surface for

adsorption. The coverage of OA and OB are represented as follows

OA = KAPA
1+KAApA +KBBpB
(3.40)
B = KBPB
1+KAApA +KBBpB

From mass-action kinetics the reaction rate expressions for A and B are

A + O(s) A(s)
kB (3.41)
B + O(s) B(s)

The steady state analysis, similar to the development of the Langmuir isotherm, leads to

an expression for the surface coverage for both A and B:

K,,[A]
OA =K c [A [B]
I + Ko, [A]+ Kc,2[B]
S K2 B](3.42)

I + K,, [A]+ K,2z[B]

Adsorption of ammonia on the surface was assumed to occur by chemisorption.

The Langmuir adsorption isotherm was used as the basis for the reaction mechanism.

Competitive adsorption was important when considering ammonia in humid air, where

both ammonia and water were competing for available surface sites.









Proposed Mechanisms

Due to the limited amount of testing that could be done with the carbon dioxide

sensor, the model focused only on reactions occurring on the ammonia sensor. To detect

ammonia a titanium nitride (TiN) film was used as the sensing layer. The change in

ammonia concentration was detected by a change in the work function of the sensing

layer. Several hypotheses for the reaction mechanism, which causes the change in the

work function, were considered.

Ammonia and Hydroxide

The first reaction mechanism considered involved ammonia and hydroxide (OH).

Previous work by Ostrick et al. (44) indicated that ammonia may bond to OH groups or

OH-precovered sites on the sensing layer. Two mechanisms are possible, either the

ammonia removes the OH from its binding sites or the ammonia binds directly to the OH

groups already on the surface. The mechanism proposed was based on the following:

1. Peaks seen in Diffuse Reflectance Infrared Fourier Transform Spectra (DRIFT-
spectra), Figure 3-1,

2. Sensitivity of the TiN sensing layer at room temperature versus at higher
temperatures (1200C), and

3. Experiments by Takagi-Kawai et al. (49) where a similar mechanism was proposed.

Considering first the spectra of Figure 3-1, the peaks at 1450 cm-1 and between

3600 3800 cm-1 were attributed to OH groups, the deformation vibration and free or H-

bridge-bonded valence vibration, respectively. Increases in signal were attributed to

ammonia as follows: 1600 cm-1, asymmetric deformation band of NH3, 2700 2800 cm-1,

v (N-H) of ammonia ions, and 3000 3300 cm-1, NH stretching region.











10 2700-2800 Cm r 3850 cm-1




0,af cm 3 740 cm

1000 2000 3000
Wave number [cm ']
Figure 3-1 Diffuse Reflectance Infrared Fourier Transform Spectra (DRIFT-spectra) for a
TiN screen-printed film, where absorbance was used to distinguish between
the species. (Data of Ostrick et al. (44)).

Next Ostrick et al. (44) indicated that because of the high affinity between

ammonia and water, the small concentration of ammonia molecules (<100 ppm) will

react entirely with the high concentration of surface water (>5000 ppm). Also, in the

second reaction step, the ammonia ions may react directly with the sensing layer or with

adsorbed OH ions. When the temperature rises, the amount of water on the surface

decreases along with the solubility of ammonia in water. Resultantly, the amount of

ammonia on the TiN surface is reduced and the sensitivity decreases.

From Takagi-Kawai et al. (49) evidence of a reaction between ammonia and

hydroxide on a surface is depicted in Figure 3-2.


NO + (1/2)02 NO2(ad)


NH3 + -KOH -> r-- -NH4+(ad)


N02(ad) + O---0-NH4t(ad) -- N2 + 2H20 + O-0

Figure 3-2 Mechanism for the reduction of NO by NH3 over a V205 sensing layer in the
presence of oxygen. (Data of Takagi Kawai et al. (49).)









This mechanism was derived from infrared (IR) and x-ray photoelectron

spectroscopy (XPS). The spectra from several surfaces were compared as seen in Table

3-1.

Table 3-1 Infrared (IR) and XPS measurements and assignments for adsorbed species of
ammonia on surface catalysts. Degenerate deformation and symmetric
deformation are indicated by 6d and 6s, respectively. (Data of Takagi Kawai
et al. (49).)
IR (cm-1) XPS (ev)
Surface 6sNH4+ 6dNH3 6sNH3 6,sN4+ 6dNH3 6sNH3
V205 1413 --------- --------- 400.9 ---- ----
V205/A1203 1410 1610 1275 401.0 400.2 400.2
A1203 ---------- 1610 1275 ---------- 400.0 400.0
V205/SiO2 1435 1620 --------- 401.0 ---- ----
V205/TiO2 1424 1605 1238 401.0 399.4 399.4
TiO2 ---------- 1605 1177 ---------- 399.6 399.6
NH4V03 1410 --------- --------- 400.9 ---- ----

By comparison, only one peak from Figure 3-1, at 1600 cm-1, matches up with

those in Table 3-1. Shimanouchi (50) has also reported this mode, degenerate

deformation, with great certainty as 1627 cm-1. Ostrick et al. (44) indicated there was a

peak between 2700 2800 cm-1, which was evidence of NH4 Since this peak does not

match up with the values for s6NH4+ in Table 3-1, another look suggests that the

absorbance in this region is not large enough to warrant it as a peak. For example, the

absorbance seen in the region between 2700 2800 cm-1 is no greater than at wave

number 2000 cm-1, which was not considered by Ostrick et al. (44). With that in mind,

NH4+ may not even be a product and therefore another mechanism should be considered.

Ammonia Dissociation

The second mechanism considered was dissociation of ammonia on the TiN

surface. Several studies examining ammonia reactions on various surfaces have been

done. The earliest studies evaluated the reaction of ammonia on silicon surfaces. Later









investigations examined ammonia on other transition metal surfaces such as titanium and

silicon, titanium nitride and platinum, and nickel.

Hlil et al. (51) used x-ray photoemission spectroscopy (XPS) and ultra-violet

photoemission spectroscopy (UPS) to study chemisorption of ammonia on a Si(100)

surface at substrate temperatures from 100 to 700 K. When the surface was exposed to

ammonia at low temperatures (100 K), a line appeared at 400.0 eV binding energy, curve

(a) in Figure 3-3, which was attributed to molecular condensation on the substrate. At

room temperature another line emerged at 398.5 eV, curve (b) in Figure 3-3, which

correlates to NHx (X = 1,2). Additionally, the presence of weak silicon to hydrogen

bonds evidences partial dissociation and co-adsorption of NHx radicals and atomic H.







I V
S .b- N,HR
aI 100K
a N3 F

394 396 398 400
Bii,;,,i_- .: _-:. (eV)
Figure 3-3 The XPS N(ls) core-level spectra for ammonia over the Si(100) surface as a
function of substrate temperature. (Data of Hlil et al. (51)).

Bozso and Avouris (52) also studied reactions of two different silicon substrates

with NH3 and atomic nitrogen using XPS and UPS. On a Si(100)-(2xl) surface at 100 K

a peak at 400.1 eV, physisorbed molecular ammonia, was visible. Another peak was

evident at 398.5 eV, although not as clearly defined as the NH3 peak. As the temperature

increased to 300 K, this peak became more defined. Again this peak was most probably

due to NH2 or NH. Upon further annealing to higher temperatures, the peak broadened to









a lower binding energy indicating conversion of this species. When annealed to 950 K, a

peak appeared at 397.7 eV, atomic nitrogen on the silicon surface, which resulted from

complete dissociation of the NHx intermediate at 398.5 eV. Both the molecular ammonia

(400.1 eV) and N bonded to silicon (397.7 eV) species were also seen on the Si( 11)-

(7x7) surface. A surface species produced by dissociation was also apparent on this

surface, but at a slightly higher binding energy of 398.8 eV. Figure 3-4 illustrates the

transition from low to high temperature on both surfaces.

SI I I I -. Ic NI i









tOOK

950K

394 395 39 3SB 9 400 4 4042 34 395 396 397 3s9 e399 4400 4 4 403
BINDING ENERGY(eV) BINDING ENERGY(eV)
(a) (b)
Figure 3-4 The XPS N(ls) core-level spectra for ammonia on (a) Si(100)-(2xl) and (b)
Si(1 11)-(7x7) surfaces as a function of substrate temperature. (Data of Bozso
and Avouris (52).)

Bischoff et al. (53) used ultra high vacuum (UHV) multilayer preparations, which

combined the different species, to identify by XPS the nitrogen chemical environments in

the Si/NH3 system. The assignment of the binding energies was found as follows:

Nitride (Si, N) 397.4 eV
Imide (Si, = N H) 398.0 eV
Amide (Si N = H) 398.6 eV
NH3 physisorbed 400.1 eV










Zhou et al. (54) studied the decomposition of NH3 on Si(100) using static

secondary ion mass spectroscopy (SSIMS). This procedure probed the surface directly

and followed reaction intermediates in real time. At low temperatures, 100 K, adsorption

of NH3 on Si(100) was dissociative and produced NH2(a) and H(a). Some thermal

decomposition of NH2(a) to N(a) and H(a) occurred at 320 K, but most of the NH2(a) was

stable up to 630 K. NH2(a) decomposed rapidly between 630-730 K, with no evidence of

NH(a) at this temperature. Some NH2(a) recombined with H(a) at 685 K to liberate

NH3(g). Temperature programmed desorption (TPD) was also completed and revealed

that NH3(g) was desorbed at room temperature. Figure 3-5 illustrates the surface species

and desorption products found as the temperature increased from 100 to 1000 K.



Surface Species 3 Temperature
S IProgrammed
Adsorption Desorption
NH3(a, 2) 100 K (TPD)
NH3(a, 1) Products
NH2(a), H(a)
NH3- 190 K (

NH3(a, 1)
NH2(a), H(a)
Heat -- -2----190 400 K 4
H eat ---|a
H(a) 400 630 K

desorption
NH2(a)
H(a), N(a)
630 730K (NH3(g)
(small)
H(a), N(a)30 0 K
730 1000 K

Silicon nitride C

Figure 3-5 Surface species and desorption products from ammonia on Si(100). (Data of
Zhou et al. (54).)









Chen et al. (55) examined ammonia surface chemistry on two different surfaces,

Si(111) (7x7) and Si(100) (2x1) using a high resolution electron energy loss

spectrometer (HREELS) and TPD. On the Si( 11) surface dissociative adsorption

produced NH2(a) and H(a), and between 300 and 600 K further dissociation occurred to

produce NH(a). On the Si(100) surface, the NH2(a) species remained thermally stable

until approximately 600 K. The differing geometries of the two surfaces contributed to

the species that were stable between 300 and 600 K. Figure 3-6 shows a ball and stick

model of the two different surfaces, as well as the adsorption geometry of ammonia on

the surface.

J NH3
H 3


H~


Y//
N-H H







(a) (b)
Figure 3-6 Ball and stick models for the adsorption geometry for -NH2 and -H on two
different surfaces: (a) Si(l 11) (7x7) and (b) Si(100) (2xl). (Data of Chen
et al. (55).)

Siew et al. (56) used XPS to observe the adsorption and reaction of ammonia on a

titanium/silicon surface (100). At low temperatures, 120 K, three Nis peaks emerged:

397.8 398.1, 400.5 400.8, and 402.2 402.6 eV, attributed to NHx (X = 1 or 2),

molecular NH3, and NH4 respectively. At room temperature more NH3 molecules

desorbed and NH4" disappeared indicating the species was not stable at room










temperature. Some conversion of the NHx species occurred and the N diffused into the

film. Figure 3-7 shows the spectra at 120 and 300 K.


NIs


300K




-.1 1 .n i 120K

392 394 396 398 400 402 404 406 408
Binding Energy (eV)

Figure 3-7 The XPS N Is spectra of Ti/Si (100) surface at 120 and 300 K. (Data of Siew
et al. (56)).

Abom et al. (43) used titanium nitride as part of a sensing layer of a field effect

metal-insulator-semiconductor device. The three different layers investigated for

responses to ammonia and hydrogen were: TiN, a double layer with platinum on top of

TiN, and two-phase Pt TiN films formed by co-sputtering. The sensor response to

ammonia, hydrogen, propene, and acetaldehyde was measured and it was found that

devices containing platinum were responsive to all gases. Devices without platinum did

not respond to any of the test gases. Abom et al. (43) indicated that the change in signal

of the sensor was due to an interaction with hydrogen. It was assumed that only atomic

hydrogen diffused through the film and that atomic hydrogen was only created when Pt

was present. From the results it was apparent that the response to ammonia was due to

dissociation of ammonia molecules and detection of atomic hydrogen.

Laksono et al. (57) studied the adsorption of NH3 on clean and oxygen pre-treated

nickel (111) surfaces at room temperature using XPS. Without oxygen on the nickel









surface, no adsorption of ammonia was observed. Similarly, the surface reactivity was

strongly linked to the presence of adsorbed oxygen; it increased with increasing adsorbed

oxygen coverage. Two N adspecies were detected from Nis core level peak: 399.8

(molecular) and 397.8 (dissociated). Looking at the Ols core level peak revealed that at

low ammonia exposures, the hydroxyl component increased, while the main feature at

529.9 eV decreased. Laksono et al. (57) found that the concurrent transformation of both

peaks indicates that hydrogen is removed from ammonia by the adsorbed oxygen to

produce OH and NHx. Quantitative treatment of the XPS offered the following

stoichiometry for the reaction:

NH3 (a) + O(a) -> NH2 (a) + OH(a)

Additionally, the experiments indicated that the kinetics of desorption were faster than

the kinetics of dissociation.

Molecular Adsorption

This last piece of evidence leads into the next mechanism, molecular adsorption

of ammonia. While dissociation of ammonia was seen on many surfaces, it is possible

that in the present study the rate of ammonia desorption is faster than the rate of

dissociation.

Diebold and Madey (58) investigated adsorption and electron stimulated

desorption of NH3 on TiO2(1 10) by XPS and low-energy ion scattering (LEIS). Three

different surfaces were studied: (1) a stoichiometric surface, (2) a thermally treated,

slightly oxygen deficient surface, and (3) a sputtered, highly oxygen deficient surface.

Ammonia was seen to adsorb molecularly on all three surfaces and the saturation

coverage was governed by repulsive interactions between the molecules. Diebold and









Madey (58) proposed a model in which ammonia adsorbs at titanium cation sites and

where lateral repulsion between the ammonia molecules along one-dimension limits the

saturation coverage. For the stoichiometric surface, ammonia appeared to desorb in

molecular form, whereas for the highly oxygen deficient surface electron stimulated

dissociation of ammonia was seen. The final product of this dissociation process was

atomic nitrogen. The previous experiments were done at 160 K. On the highly oxygen-

deficient surface, heating the surface to 395 K desorbed all nitrogen-containing species

from the surface, with no evidence of thermally induced dissociation. Therefore, it was

concluded that ammonia was adsorbed as an intact molecule.

Karthigeyan et al. (59) studied an iridium oxide thin film integrated HSGFET

which was selectively sensitive to ammonia at room temperature. An increase in sensor

signal for ammonia at higher temperatures was seen, but the nature of the response to

ammonia was unchanged from room temperature to 1000C. A possible reaction

mechanism for ammonia on the sensing layer was based on the sharing of lone donor

bond to positively charged vacancies by chemisorption. The sensor showed no response

to concentrations of hydrogen up to 10,000 ppm. Additionally, the signal response was

negligible to CO, SO2, Cl2, and NO2.

Karthigeyan et al. (59) indicated that in the classical model, a charge transfer

between adsorbed molecules and the surface of the sensitive layer and/or dipole moments

of adsorbed molecules on the surface create a surface dipole layer which causes a work

function change. A decrease in the work function upon ammonia adsorption indicates

that there is a net electron transfer from ammonia to cation sites on the surface. The

strong nature of the ammonia reaction is further indicated by the effect of temperature on









the reaction and recovery properties seen in the transient response of the sensor. The

increase of the work function signal with temperature came from diffusion of the

ammonia molecule into the film. Diffused molecules attached to cation sites lying

underneath the surface. The desorption was complicated by residual gases in the

measurement chamber and changes in bonding nature of adsorbed molecules before

adsorption, i.e. surface diffusion to alternative bonding sites, lateral interactions between

ammonia molecules, or immediate compound formation due to ammonia dissociation.

Reaction Kinetics

Based on the preceding evidence, molecular adsorption of ammonia on the titanium

nitride surface was chosen as the reaction mechanism for the model. Therefore the

reaction mechanism was described as follows,

NH3 + O(s) e-k NH (s), (3.43)
kr

where O(s) is an open surface site and NH3 (s) represents ammonia adsorbed on the

surface. For the rate of the forward reaction or adsorption the forward and reverse

reactions were considered together as,

d[NH3( = kfpNH [O(s)]- k[NH3(s)]
dt (3.44)
= k [NH3 (g)]([S] [NH3 (s)]) k, [NH3 (s)

Upon integration,

[NH s)] = [S]i. 1 -e(kf[NH3(g)]+kr)t
[NH, (= k (3.45)
1+
k,[NH3(g)]

where [S],.s. is the total concentration of sites on the surface where the ammonia can

adsorb at steady-state. This is different than total site density, F, because [S]s.s. varies









with the free stream gas concentration. For example, the number of sites the analyte

occupied was less at a lower analyte free stream concentration than at a higher analyte

concentration.

For the reverse reaction the analyte was removed by forced flow rather than by

diffusion thereby quickly forcing the ammonia out of the free stream, which caused the

first term on the right hand side of Eq. 3.44 to be negligible compared with the reverse

reaction. Therefore, the rate of the reverse reaction or desorption is

d[NH3 ()= -k, [NH3 (s)]. (3.46)
dt

Integration leads to

[NH3 ()]= [S],, exp(- kt). (3.47)

At equilibrium,


[NH (S)]= [S], Kp N (3.48)
1 + KNH3

where K=kf/kr. Also at equilibrium,

K = exp(- AG/RT)
(3.49)
= exp(AS / R)exp(- AH/ RT)

The rate equations were made dimensionless to allow them to be used in other

applications where the physical parameters, such as the area of the sensing film, may be

different. The dimensionless variables were as follows,

[NH3 (S)
0= [ (3.50)


and

Dt
= (3.51)
W-L'









where D is the diffusion coefficient and W and L are the width and length of the sensitive

film. Equation 3.44 in dimensionless form now becomes,

dO =(kf[NH3(g)](1 ) kr)W) L (3.52)
drH) D(3.52)
dz D

As noted above, molecular adsorption was the focus of this study, however since

dissociation was not ruled out as a possible reaction mechanism, the rate equations for

dissociation were formulated and integrated. The reaction mechanism for ammonia

dissociation on a titanium nitride surface was as follows,

NH, + 20(s) k NH, (s) + H(s). (3.53)
k,

The dissociative adsorption was assumed to require two open surface sites and the

adsorption and breaking-apart of the molecule were taken to occur in a single step.

Assuming that NH2(ad) and H(ad) adsorb and desorb at the same rate on the surface,

Eq. 3.53 was simplified as follows

A, + 20(s) ef 2A(s), (3.54)
k,

where A2 represents NH3 and 2A signifies the two surface sites that NH2 (s) and H(s)

occupied. The rate of the forward reaction was given as

d[A(s)] k [A2 ][O()]2 k, [A(s)]2
dt (3.55)
= kf [A ([S]. [A(s)2 k, [A(s)]2

Integrating the rate equation and applying the initial condition, t = 0, [A(s)] = 0, provides,


[A(s)] k= tanh k (3.56)
k kf[A,








The rate of the reverse reaction or desorption was considered without the forward

reaction because the analyte was removed by forced flow rather than by diffusion thereby

preventing the particles from adsorbing to the surface again. Hence the gas phase

concentration [A2 ] was assumed to be negligible, yielding Eq. 3.57 from Eq. 3.55. The

rate of the reverse reaction or desorption was

d[A(s)] -k[A()]. (3.57)
dt

Integrating using the initial condition, t = 0, [A(s)] = [S], led to


[A(s)]= (3.58)
kryt + 1/[S]5.

Summary

The foundation for the theoretical modeling of the ammonia sensor was laid in this

chapter by first presenting the relationship between thermodynamic properties and

reaction kinetics. Then adsorption theory and surface reaction rate expressions were

discussed, followed by a literature review of mechanisms that led to the choosing of

molecular adsorption as the mechanism for the model. Finally, the reaction kinetics for

molecular adsorption were mathematically formulated to be used with the experimental

results presented in the next chapter.














CHAPTER 4
RESULTS AND DISCUSSION


The results from the experimental portion of this study are presented in this

chapter. Throughout the course of this project, the sensor developers were continuously

upgrading the performance of the field effect transistors; resultantly, three generations of

sensors were used in testing. The latest versions were always utilized in the experiments.

For the ammonia sensor, first presented are the response times fitted to the model to

obtain rate information. Then, the performance and feasibility of the sensor for use in a

laboratory animal environment are discussed. Detector reliability and sensitivity limited

the useful data obtained from the carbon dioxide sensor. Consequently, for the carbon

dioxide sensor, only a calibration curve, response times of the sensor, and cross-

sensitivity to humidity are reported. Experiments for modeling purposes were not

completed with the carbon dioxide sensor.

Drift Tests

Prior to conducting any tests with the analyte gases, the carbon dioxide and

ammonia sensors were tested for drift of the baseline signal. The sensors were exposed

to air at a flow rate of 2 L/min for over 40 hours. The results of these tests can be viewed

in Figure 4-1. Each sensor had two channels of output offering the same signal with

different amplification factors. Table 4-1 lists the rates of the signal drift. The carbon

dioxide sensor had a negative drift rate, while the ammonia sensor drift rate was positive.

The magnitude of the drift rate for the carbon dioxide sensor was slightly larger than for







79


the ammonia sensor. For both sensors the drift rate was much higher in the first five

hours than for the rest of the test, which shows that the stability of the baseline increased

with time. Therefore, for the following tests the sensor was turned on at least one hour

before each experiment to help stabilize the baseline signal.


ChA 4.8- Ch A
Ch B -- Ch B
2.00- 4.6-

4.4-
S1.75- V
S 4.2-
a 1.50- a 4.0-
0 0)
3.8
1.25- 3.6- ,^-, -,--,
3.4-
0 10 20 30 40 50 60 70 0 10 20 30 40 50
time (hr) time (hr)
(a) (b)
Figure 4-1 Long-term drift test results (sample size, N = 1) for the (a) carbon dioxide and
(b) ammonia sensors. Signal output recorded every 30 minutes.

Table 4-1 Rates of baseline signal drift for the carbon dioxide and ammonia sensor
exposed to air at 2 L/min for over 40 hours.
Slope (mV/hr)
Sensor Channel A Channel B
Carbon dioxide -7.3 -5.5
Ammonia 1.8 2.6

Ammonia Sensor Results

For the ammonia sensor, first presented are the time response tests fitted to the

model to obtain rate constants, the equilibrium constant, and Gibbs free energy. Second,

the role of diffusion is discussed along with the presentation of time response results in

the diffusion box. Third, results from tests for temperature dependency and the resulting

Arrhenius parameters are presented. Finally, the performance and feasibility of the

sensor for use in a laboratory animal environment are discussed.









Sensor Response and Mechanism

The reaction mechanism selected for the ammonia sensor model was molecular

adsorption of ammonia on the titanium nitride surface,

NH3 + O(s) NH (s), (4.1)
kr

where O(s) is an open surface site and NH3 (s) represents ammonia adsorbed on the

surface. The Langmuir adsorption isotherm was used when developing the mechanism.

In conjunction with the Langmuir model, the assumptions made were

* Adsorption does not proceed beyond single layer,
* All sites are equivalent and the surface is uniform,
* Ability of a molecule to adsorb at a given site is independent of the occupation of
neighboring sites.

The forward and reverse reactions were considered together to determine the rate of the

forward reaction,

dNH3 = k, fNH3 [O(s)] k, [NH (s)]
dt (4.2)
= k [NH3 (g)][S], [NH3 (s)D k [NH3 (s)]

where kf and kr are the forward and reverse rate constants, respectively, [NH, (s)] and

[NH3(g)] are the concentrations of ammonia on the surface and in the gas phase,

respectively and [S],.s. is the total concentration of sites on the surface where the

ammonia can adsorb at steady-state. This is different than total site density, F, because

[S]s.s. varies with the free stream gas concentration. For example, the number of sites the

analyte occupied was less at a lower analyte free stream concentration than at a higher

analyte concentration. Solving Eq. 4.2 by integration using the initial condition: t = 0,

[NH3 (s)]=0, led to the following,










[N3 ) = kr (4.3)
1+ ,
kf[NH3 (g)]

The rate of the reverse reaction or desorption was considered without the forward

reaction because the analyte was removed by forced flow rather than by diffusion

thereby, preventing the particles from adsorbing to the surface again. Hence the gas

phase concentration, [NH3(g)], was assumed to be negligible, yielding Eq. 4.4 from Eq.

4.2. The rate of the reverse reaction or desorption is

d[NH3 ) k[NH3 (s)]. (4.4)
dt

Integrating using the initial condition: t = 0, [NH3 (ad)] = [S],,, provided the following

[NH3(s)]= [S],, exp(- kt). (4.5)

To solve for the rate constants, the dimensionless concentration, or the fractional

surface coverage, 0, was used,

o= [NH3(s)] (4.6)
[S= (4.6)

To experimentally determine 0, the sensor was exposed to air for 10 minutes, then 100

ppm ammonia for 10 minutes, and then air again for 10 minutes. The surface coverage

for the forward reaction was calculated by dividing all data points from when gas was

introduced into the system until it was removed by the maximum voltage signal. Hence

the signal was normalized, yielding a scale from 0 to 1, with 1 being maximum surface

coverage. Similarly for the reverse reaction, 0 was determined by dividing all data

points from when gas was removed from the system by the maximum voltage signal.









The rate constants for the forward and reverse reactions were solved by curve

fitting the experimental data. The reverse reaction rate constant was analyzed first

because the forward reaction is a function of the reverse rate constant. For the reverse

reaction the surface coverage, 0, was plotted versus time and the following equation was

used for the curve fit,

0 = aebt, (4.7)

where the fitting parameter b yields the rate constant, b = -k, For the normalized data,

ideally a = 1. This was typically the nominal value, (see Table 4-2) although the curve fit

was not forced to a pre-exponential of unity. The desorption data with the curve fit can

be seen in Figure 4-2. Based on the curve fit, the reverse reaction rate constant, kr, is

6.43 x 103 s-1. The coefficients, R-squared values, and rate constants for both adsorption

and desorption are listed in Table 4-2.


1.0

S0.8

U 0.6

8 0.4

Cu 0.2

1) 0.0

0 100 200 300 400 500 600
time (s)
Figure 4-2 Average desorption data and curve fit for 100 ppm desorption tests. The error
bars represent +1 standard deviation at selected data points (N=5).

For the forward reaction, 0 was also plotted versus time with the following

equation used for the curve fit,









0 =a( -ebt)) (4.8)

where again a is a pre-exponential constant and b = kf [NH (g)] + k The curve fit for

the adsorption data can be seen in Figure 4-3. The constant b was used to solve for kf as

follows,

b-k
k ^
S- [NH(g)]
(0.04335 0.00643)s1 (4
~1 n1mgNH3 L 1gNH3 lmol
100 ppm x --x -- x --
lppm l000mgNH, 17gNH,
=6.28L/mol- s

1.2

1.0

aj 0.8-
0)
S0.6-
0
00.4-



0.0

0 20 40 60 80 100 120 140
time (s)
Figure 4-3 Average adsorption data and curve fit for 100 ppm adsorption tests. The error
bars represent 1 standard deviation for 6 runs at selected data points.

Table 4-2 Curve fit coefficients, R-squared values, and rate constants for desorption and
adsorption of 100 ppm ammonia on sensor.
a b R2 k
Desorption 0.915 -0.00643 0.96 6.43 x 10-3 s-1
Adsorption 1.01 0.04335 0.99 6.28 L/mol*s

Time response experiments were also conducted at 50 ppm, where the sensor was

exposed to ammonia for 10 minutes with 10 minutes of air before and after the exposure.

The rate constants calculated using 100 ppm data were then used to predict the adsorption









and desorption curves at 50 ppm ammonia. The actual and predicted curves for

desorption and adsorption are depicted in Figure 4-4 and Figure 4-5, respectively.

Percent error between the actual and predicted curves was calculated as follows,

Actual Predicted
%Error = x 100%. (4.10)
Predicted

For the desorption curve, the average percent error was 27%, where the error

significantly increased as the predicted curve approached zero, which was the limit of the

exponential curve fit. Neglecting the error at the limit, the average percent error from 0 -

450 seconds was 10%. The percent error for the adsorption curve was 38%, which is

expected when comparing the two curves in Figure 4-5. The actual curve for 50 ppm

adsorption has large standard deviations at the representative points, which may

contribute to the large error seen between the actual and predicted curves. It is noted that

the detection limit for the ammonia sensor is roughly 100 ppm (see below); therefore the

50 ppm data must be considered in this context.


1.0 Actual
Predicted

S0.8

(U 0.6

0.4

M 0.2

(1 0.0-

0 100 200 300 400 500 600
time (s)
Figure 4-4 Actual 50 ppm average (N=5) desorption curve plotted along with the
predicted curve based on the rate constant determined from 100 ppm curves.
Error bars represent +1 standard deviation for selected data points.










1.0- Predicted .

S0.8-

( 0.6-
> .-I
8 0.4

3 0.2
-Ar
CO 0.0-/

0 20 40 60 80 100 120 140 160 180
time (s)
Figure 4-5 Actual 50 ppm average (N=6) adsorption curve plotted along with the
predicted curve based on the rate constants determined from 100 ppm curves.
Error bars represent +1 standard deviation for selected data points.

Once kr and kfwere determined, the equilibrium concentration constant, Kc, was

determined from the following,

k,
kr
6.28L/mol s
-- (4.11)
6.43 x 103' S
= 976.7L/mol

The equilibrium constant, K, can be calculated from the equilibrium concentration

constant by,

( 1 Y pRTY
K = Kp I KCR


-KcP
RT
976L 1m3 105 N/m2
x xl .0lbars x (4.12)
mol 1000L lbar
mol K 1 1J
x--x--x-
8.315J 304K 1N*m
=39.04









where p is standard pressure and v = -1 for molecular adsorption. Also at equilibrium,

K =exp(- AGo/RT) (4.13)

where AG is the standard Gibbs energy of reaction, R is the ideal gas constant, and T is

the surface temperature of the sensor. Therefore, AG was calculated as follows,

AG =-RT nK
J
= -8.31-- x 304K x In 39.04. (4.14)
mol-K
= -9.26kJ/mol

In comparison, the standard Gibbs energy of reaction for ethylene hydrogenation on

platinum is -148.2 kJ/mol and for ammonia synthesis on iron it is 62.0 kJ/mol (60).

The Gibbs energy of reaction predicts the direction of spontaneous change of a

reaction at constant temperature and pressure. The reaction Gibbs energy is also defined

as the slope of the graph of the Gibbs energy plotted against the extent of the reaction, 5:


AG = (4.15)


and it is further derived as the difference between the chemical potentials, [a, of the

reactants and products at the composition of the reaction mixture,

AG = Uproducts reactantss (4.16)

Due to the fact that the chemical potentials vary with the composition, the slope of the

plot of Gibbs energy against extent of reaction varies as the reaction continues (Figure 4-

6). Further, because the reaction proceeds in the direction of decreasing G, it is apparent

that the forward reaction is spontaneous when preactants > Pproducts or ArG <0. The reverse

reaction is spontaneous when products > reactants ArG >0, whereas the system is at

equilibrium when ,products = reactants, ArG =0. Since ArG < 0 for the reaction of