<%BANNER%>

Development of a MEMS-Based Acoustic Energy Harvester

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 E20110115_AAAADL INGEST_TIME 2011-01-15T20:32:26Z PACKAGE UFE0008378_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 57381 DFID F20110115_AACJTI ORIGIN DEPOSITOR PATH horowitz_s_Page_009thm.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
a3503cba21cdd07bdddb270daac004d9
SHA-1
4a47caa4b3e9fe042398dc48afbb52625eb620b2
252 F20110115_AACJSU horowitz_s_Page_188.txt
c9dd66c7f42993d8c18a42fe56d62305
517b0746fd9710692ac364d4d1070d361a6ec259
1114 F20110115_AACIQG horowitz_s_Page_123.txt
9addc9eb42a08e99029d79cc9356880c
96d52e183e53cbc1ab145c5ec8344f7053cb2042
90267 F20110115_AACIPS horowitz_s_Page_147.QC.jpg
4262003d3c082797b5aed253d0918cea
4161bd1fd473c37ed286a339a55a69811743b487
35317 F20110115_AACJTJ horowitz_s_Page_190thm.jpg
f26f7afe503d8a183ce6b8fa46100653
fab62711ef72c572158ef15df0d4a084efdc76db
79174 F20110115_AACIQH horowitz_s_Page_094.QC.jpg
a5953a714f5001cf3a8b317d8e8fee8b
e3b931133e7d2ba2d10c347da2a22477dd8252e7
37905 F20110115_AACJTK horowitz_s_Page_172thm.jpg
16d9675ee4c137fe612f0844d96d6a9e
26ffdd70f5f103a649da445cb89da8b4e536b9bb
229 F20110115_AACJSV horowitz_s_Page_189.txt
87cd46eddb3d45838e7fe01088c0ff42
727a2c726b80a35fabdb4a259e21fde6eddb8e37
47446 F20110115_AACIQI horowitz_s_Page_164.pro
562c78930ed9516d00ce8810d65e7f56
5bd76ec377ca4884196b6ac0798e5780c19c8415
25271604 F20110115_AACIPT horowitz_s_Page_180.tif
aeddcbeec6def251d567ba2d0c98f5f7
8000c92cc407a3aa5b4d69a475c48092e6dc30ec
44001 F20110115_AACJTL horowitz_s_Page_087thm.jpg
208f7fb780089f3c42133174992512c6
697c9a04a964caceadc7e64d8e41932b85c80ea2
2633 F20110115_AACJSW horowitz_s_Page_202.txt
7709393a1c09b69033d55c821797dc36
b56c5bb5ae7a43b031c698756947f4ae325c5fd6
77907 F20110115_AACIQJ horowitz_s_Page_155.QC.jpg
d1aa57a111cbd33f825ce55913612385
347757bbbef123344d761aced6aeba5140a58813
42991 F20110115_AACIPU horowitz_s_Page_079thm.jpg
c2bf07f61bcb136d7213c033589bbfbe
cc4301a48e33cb95d692e8c0ecf18cfa6f12341c
78627 F20110115_AACJUA horowitz_s_Page_018.QC.jpg
fa74ab2ecd377c8c3f3084a8e23a08f7
0cc91358a7ca4c452b1584ddab2d4fa90306db43
84114 F20110115_AACJTM horowitz_s_Page_206.QC.jpg
1172dcb7dca2c26f64b09d28fe040fe5
4718cce557739eef70de92e5324927b27cfc620a
2027 F20110115_AACJSX horowitz_s_Page_205.txt
aedf33781b0dd94e27d925c16a70bcfe
58ced847a935a6b732eaa9c8c6833fce44e0887e
23917 F20110115_AACIQK horowitz_s_Page_057.pro
2a7cb6900d6640ccdf5c2f7f7e851c74
60d718816fd66cb20026d1c202278174f7d022c4
27953 F20110115_AACIPV horowitz_s_Page_086.pro
82854e3e8cfe4bcf73b02b9761398cd8
775021f3b7077d5d216aef31d36993a138aca5f0
83946 F20110115_AACJUB horowitz_s_Page_113.QC.jpg
d410bc8e7b0ddabededf8b1c57f82a89
3f5dc254a9e600d6be712f11dc9949d5e903327f
81311 F20110115_AACJTN horowitz_s_Page_137.QC.jpg
897618e575a9646ed7452b9b999bb9a2
c2159b19df258e5b6e7e7401234e489f5b866732
605 F20110115_AACJSY horowitz_s_Page_209.txt
7e50e1efd32867f8a0e17740a978b780
14af3462dba1f093f1869d91262a3cb63619b0b8
25652 F20110115_AACIQL horowitz_s_Page_206thm.jpg
ab2d740dc16b8058627c9705378b7459
7404ee47d3e436b83f16814c4f612f88d079715c
64216 F20110115_AACIPW horowitz_s_Page_060.QC.jpg
c128cb5b580619cdcc10b384c0820086
079cb55be1b858d98c51b54db39f718f3287f3b6
36252 F20110115_AACJUC horowitz_s_Page_192thm.jpg
de32c80e5dd427e05872a4694e813e83
2fbad33b8678eb820b6e431bc1054530fa6f93af
44524 F20110115_AACJTO horowitz_s_Page_084thm.jpg
15afe8794275c27016334e4840dde946
e278fa45f6498105a2723fb4ce7e0a5c82f35c0b
19314 F20110115_AACJSZ horowitz_s_Page_075thm.jpg
7b76fe960aa665f629659e5ca67b7b87
12f31eaf7d5dbc997ee4822338223dbc875bf9c7
1566 F20110115_AACIRA horowitz_s_Page_064.txt
5b91af330751535c754b04ee91e729ba
3a992dea563c47d1b23849c52f3d53c4eda8507d
97766 F20110115_AACIQM horowitz_s_Page_019.jp2
3cf454d464f6f7aa1fd69f862558e78e
8f17f5a0146dd2a1ae6ea38330c19cf089467f3c
1053954 F20110115_AACIPX horowitz_s_Page_163.tif
141d937bebd7df4602732885323288d9
d27c1b2b2a2624e97235ee106c3098187d5b9951
45081 F20110115_AACJUD horowitz_s_Page_185thm.jpg
a55a96524e39f74cfb3c5dd82c2e95e1
2c358f1d9b307887f5c070ec6110ae7b34f1e6db
56649 F20110115_AACJTP horowitz_s_Page_183.QC.jpg
f5dd1c8c60e7a7cbb9a0a2556b9c5722
0d2af734d271f33f12ed85bf09bbc1f93fc5ff92
89442 F20110115_AACIRB horowitz_s_Page_108.jp2
185ac70e7f6b6f2f419037b8a02af1c5
d49c191471dcf705a6afa8cf827a3a257154dd8c
1395 F20110115_AACIQN horowitz_s_Page_139.txt
7b5041a5b1f894af31050fb9692ee32e
313bfe203245eb48ab51ce1d5716176b758a128a
670851 F20110115_AACIPY horowitz_s_Page_016.jp2
a33f5c764615cbf1058d29c2854d89c0
9ec1c8dc52117bf386cc38f52a3ba6df0111bb08
120624 F20110115_AACJUE horowitz_s_Page_005.QC.jpg
fa7087bfd679766479c739f942a51d19
39174304a07162ab9647ae8e99b86f7b494c1c8e
78920 F20110115_AACJTQ horowitz_s_Page_022.QC.jpg
cbe2c27529ccfbc1c1ec0d3ea461e091
8594a23edb4d4827a76648e7ebe5f71b39398727
187730 F20110115_AACIRC horowitz_s_Page_115.jpg
09f56f20d1d93c99eb7e27294e436a43
725cbb5bd27fe7d0407b46af441bb2f1020fce40
41418 F20110115_AACIQO horowitz_s_Page_054thm.jpg
f24fccc089d9a8ff966508c28aeef444
c90f172ebd2fb9b84405fc51774331880d3445df
18262 F20110115_AACIPZ horowitz_s_Page_064thm.jpg
4c9a00d030e08ef7e9625fa73405be4f
b05885b24edccc948ac598bdb6b578bf7f36394d
25069 F20110115_AACJUF horowitz_s_Page_038thm.jpg
516c179ad2626834f41b2f8c0ef9771e
f3bf65f5b7b7b6553508a45ea047bd47a8ac523d
99414 F20110115_AACJTR horowitz_s_Page_053.QC.jpg
cd8bdf6138dce25d0264c245ab11e2b3
5981c684b0a953bc398eadbe5ae1e736fffaaf54
32969 F20110115_AACIRD horowitz_s_Page_046.pro
1ca85b24e6fe8ca9ca771fa2e0bd3714
f69e889d0962a2805eab714c677a5c9dd577cd14
146580 F20110115_AACIQP horowitz_s_Page_163.jpg
b5f5849a314a20581d91f3716043a6c7
d07d7c253ee44e9ab11a30434c3aecb1cb8bc72a
42957 F20110115_AACJUG horowitz_s_Page_080thm.jpg
7da46df85616449f4a13a30fc91e673f
6298c0cb9916e39c30b4ecb65e99a191976b4c37
26866 F20110115_AACJTS horowitz_s_Page_209.QC.jpg
ddb6ca7d593cb0b69a74e7a6505f46fe
fd1ee5ee94ae8849535d98d1506eaa1c00d9a21f
18767 F20110115_AACIRE horowitz_s_Page_067thm.jpg
459e5ced08c3e6e7e4e360804c0f9c5c
3151dbfed38ddefab9f0090f8c6039640e635ac4
176892 F20110115_AACIQQ horowitz_s_Page_071.jpg
5bced0488782545250200762ec72114f
74b2197dbd6fe0f9cbf043e8d91eb54eb973cc70
20959 F20110115_AACJUH horowitz_s_Page_119thm.jpg
7be9381f0cb38dbf7db7fa620858ba11
9c75f41479a604e473b20a874fa74d391cdb4e37
13761 F20110115_AACJTT horowitz_s_Page_030.QC.jpg
8883e5bd4b6414c88b73219ccfc79d63
270db8c73feeff3a46d45eb1b70c77ac48b434d1
F20110115_AACIRF horowitz_s_Page_193.tif
f1468f645acad7c4b7aacf65207fb8da
6b7bd36e0e4f677ace0efc664e9b4d570c047cf2
48148 F20110115_AACIQR horowitz_s_Page_017thm.jpg
48b7afe02cae232c93ebdc4134cd0205
c39bf2c3ba21de0b727f15bbb36e5bc4bfe5a927
70308 F20110115_AACJUI horowitz_s_Page_205.QC.jpg
36c547c0702fe2917788d7e6a0581087
8400db9f24e2c34ddd2fabde1553ddb36081b0bf
93188 F20110115_AACJTU horowitz_s_Page_104.QC.jpg
87ec7d895ccdfeadc3f31922ce134a94
ea60bd9e52eb2f31771a611e94da3fb69fe1776f
F20110115_AACIRG horowitz_s_Page_067.tif
5c424cfc87c5f05905ea60ff188ab0c8
99099afc8e3430b71bb318e6408fdf2c6f713bc8
23420 F20110115_AACIQS horowitz_s_Page_164thm.jpg
6bef27218841b7ebd8489f6780715a8b
8ccd419762c73bfe152785e428155d80a2372644
41862 F20110115_AACJUJ horowitz_s_Page_013thm.jpg
17bd35321039eece88a83d8acc5f98a7
dc6463a70e663b7524a796518defb87263df5d4f
17258 F20110115_AACJTV horowitz_s_Page_014thm.jpg
7e219ebad5ffa44770008f4f2ef0f4af
b9309f06c8e2fb1208ee70181facadedf88e9e57
F20110115_AACIRH horowitz_s_Page_081.tif
4c642c6931842fe37144fb00258efa00
7d004056ca3255f6d719a5a0d4444eabafdcadb4
42398 F20110115_AACIQT horowitz_s_Page_032thm.jpg
420e67f400b15f049ab6e0c3575d316c
00b5f4fe80f265e507a7a26cd5d9e5b2970ee56d
20937 F20110115_AACJUK horowitz_s_Page_037thm.jpg
5415225a52cb911af829403d75bdaf7f
06b36c359bc11bc7ec2ae1bfc60d7b1615411428
78703 F20110115_AACIRI horowitz_s_Page_024.QC.jpg
96fd9e8f00d1c4824dc827b1cf7df7b0
c6a4c075f0e436eb44f08db64cd671e8f1dc4bda
83725 F20110115_AACJUL horowitz_s_Page_129.QC.jpg
6ae6eb9379770f88a7eb862346ddc697
24d65c88068deb9f09c102f899d89a3ae589dfb9
43048 F20110115_AACJTW horowitz_s_Page_007thm.jpg
9ab48ccf8ad09c3ba64b6058ec4940be
2e6b0f2565df9f04962f65e2c584b13527e6fa7d
39341 F20110115_AACIRJ horowitz_s_Page_160thm.jpg
e64f93af4198959a55198c2290759db2
9b587441a2d0b10347186659c93be322106d13c6
6091 F20110115_AACIQU horowitz_s_Page_002.jp2
95099cad8c2eabf36fb096af269377e4
1c8382c1e9305a04d71bd686efad4ecb652d770d
88678 F20110115_AACJVA horowitz_s_Page_154.QC.jpg
2f3f543a24624c0d91090ef09e1db2b5
1d650bcda76c5bdc449fa7b9ebcd4a85365f326f
80959 F20110115_AACJUM horowitz_s_Page_180.QC.jpg
dabc95564c27c578ba6998b626bb1cbb
422bbd2559b6c641a44975b3307ef4a7e1c591c8
43608 F20110115_AACJTX horowitz_s_Page_144thm.jpg
50a344d3c729f6bff233bd32890d047a
10343edb65df19872aa32a171962ed2975850d38
48560 F20110115_AACIRK horowitz_s_Page_057.jp2
db83d70da4b860169c1a3bd0fce7b91e
bd4385afe0d71ed1227abb295bcf873a6daab2a7
50498 F20110115_AACIQV horowitz_s_Page_169.pro
a9a154282157db8dc970b6799462c3dd
31acf82b276ce5a72443c26096d43df231d67cf2
75611 F20110115_AACJVB horowitz_s_Page_156.QC.jpg
36869c806466235a5cd46fcdf82d89a6
39e4b440c9e524b729797453acee203c379fdde1
79838 F20110115_AACJUN horowitz_s_Page_034.QC.jpg
25c1f06361c82492661961917c6cc378
2f1af7c0c342a996c11ad89dab063946bd50f828
43745 F20110115_AACJTY horowitz_s_Page_140thm.jpg
82d4bd4c5e071d65f28d1b35152ee33c
ca280088848fe50fd424b0ce2b4e5b21653d0c82
2021 F20110115_AACIRL horowitz_s_Page_034.txt
3eec6bd4cdc4a9ba7c02b55fcac73361
4af20d8335251b7283f909815c3d2460e004fa46
92098 F20110115_AACIQW horowitz_s_Page_066.jpg
8e5c057346c011eee9935114abf105c2
049a1ce1b849c5c93d5a545a80065de2707963da
17743 F20110115_AACJVC horowitz_s_Page_162thm.jpg
9d1c6e68b74776f73061ddd68f1f4057
cf15dd7b8ac8d610da3ab3ac2c14f7099ae14c3e
93446 F20110115_AACJUO horowitz_s_Page_148.QC.jpg
fbb01cd9b292e2ca563069b6aa013b06
aa7ebd7c1dd9a92dcf844d9359eab9971f84e17d
41257 F20110115_AACJTZ horowitz_s_Page_036.QC.jpg
ad565a35c771028b2f6dcc119a4f8aa3
888bedf259202bbcf18e36633b47f1575f71a56a
46240 F20110115_AACIRM horowitz_s_Page_104thm.jpg
36d40a0dc05f0f69a0147d9a86d8a38a
9c45f90135a84250fe985416346c27f61c02d440
41562 F20110115_AACIQX horowitz_s_Page_035thm.jpg
9f51e33192ab3ef606e828cc05989b87
753efae38b16e8a5f5f430052ebfab601bc939d2
49630 F20110115_AACISA horowitz_s_Page_098thm.jpg
8dab0f37c090bab50ea2ba233d146cdd
ff5cd15c68d86e05ebc614817f09d4971bad85bb
24757 F20110115_AACJVD horowitz_s_Page_197thm.jpg
db13d110b3afb986718ebdb6d40454ba
1b64fdd184770ea80f9a875a4d48834a0ed6e952
47854 F20110115_AACJUP horowitz_s_Page_130thm.jpg
3e86944b1a6549f82c9456c30073cf75
f630d0e3edcdfcf55bc9b9a39ee12c3695d413c3
1051979 F20110115_AACIRN horowitz_s_Page_013.jp2
e2a04c818268d8754e0bbb3bd36f8c7e
ceebb78eda4c90a542b01340b599bba5c766d121
25550 F20110115_AACIQY horowitz_s_Page_202thm.jpg
0c9fc01b6ddbea86a86f0809becc3e78
2f79dedb4d81747ad529f3403e12ac60e93a2cc8
164721 F20110115_AACISB horowitz_s_Page_062.jpg
5ddeffba99c0fba3233a9a7a4ff394f9
23543cd91d1899900d6bba36e529ef34fa161778
25492 F20110115_AACJVE horowitz_s_Page_200thm.jpg
60c304399f74f8b7fdbc9c587f02bf79
ddfd67b01e48a0fbce1f29882531e572cc274ac4
24814 F20110115_AACJUQ horowitz_s_Page_083thm.jpg
270ab79a7bf4eb41220fdd5149ae0474
02f1af24621dea4697a66b40df8267bf4ad4d3db
333843 F20110115_AACIRO horowitz_s_Page_010.jpg
d06c519db5364974bd2af93224fbbde7
51ef42e1e029e7d4af9800c6da6c277f7e36f424
27657 F20110115_AACIQZ horowitz_s_Page_079.pro
109ecf76bfe70790b42da7adcc6297cc
49c330821437d45d6cb85de3f645b3c851cc7ab7
79637 F20110115_AACISC horowitz_s_Page_176.QC.jpg
4bb6e86d3fdd155a6fab4ce8327c6105
949f5f8318c1ec620b896b611d93fd9c600404a0
84895 F20110115_AACJVF horowitz_s_Page_202.QC.jpg
dd4c63b81501a1226e99daf672074dcd
61e5d7249bf34d02188bde403529221b3463b62a
47808 F20110115_AACJUR horowitz_s_Page_043thm.jpg
66706062572c55b7da21a9ac5e962dd8
c918158c26c657be6180885443493dd41d275c9d
82679 F20110115_AACIRP horowitz_s_Page_141.QC.jpg
fb14ad2481ec0aab3c22dbd801c74298
410449e85b65f53b99ceb93021d36b55a3d28eb0
1376 F20110115_AACISD horowitz_s_Page_150.txt
5a42a3ce38ef863a6f4d2c98944a4068
3bffa89e7eb30aa41eb7a7585c40870061643787
61442 F20110115_AACJUS horowitz_s_Page_015.QC.jpg
d51d63bfe68dc0313c6f9d98f75c0765
e88e3c601c562e97404393eb6ad46fc7c9dcd1db
767582 F20110115_AACIRQ horowitz_s_Page_140.jp2
2318c3fd845140847e12bfe359f5eea3
42f36ba9811ca2c9f8c2dcf9d7f32c0cb5d4f722
35203 F20110115_AACISE horowitz_s_Page_057.QC.jpg
e774b7afbea26bdf7b5b3301248893cf
7cc193cf0585c8c5d564c16f21f506054de34b99
314235 F20110115_AACJUT UFE0008378_00001.xml FULL
69e3c05b0eaf8e43b4e19491eac401bc
a01104025f5718394427f74e328d6d4fc672fd7e
F20110115_AACIRR horowitz_s_Page_009.tif
d18b3f8bca327a8ebe41c902c23ea0e9
b8cfb8f9b13f8a69f86a5c00251b50789d487d9b
53548 F20110115_AACISF horowitz_s_Page_207.pro
6ad2404454346e354008a05b20908032
15b3f5307ddb00dc39a52d4742e7442cd6856f72
8182 F20110115_AACJUU horowitz_s_Page_003.QC.jpg
989f76be542d0ddf881342f4919cd46b
4fd04899e46c9f04ff71a4da45cfd5ad90725e72
44346 F20110115_AACIRS horowitz_s_Page_103thm.jpg
af94c469be447246d605e5c831f58aae
98928b0191ee84062be6501b96e93b68fbbb5c5a
63244 F20110115_AACISG horowitz_s_Page_085.QC.jpg
cc315fdeb4478f65f86d751f2f74a98b
e45b94815e85c57f6515f0a7872d6766f5eb3244
43921 F20110115_AACJUV horowitz_s_Page_074thm.jpg
6d105e02b1f66fa6ec7f8dffe14433b0
6e363aa3ed9362839c6b1c19fb7f4afad325b84d
129104 F20110115_AACIRT horowitz_s_Page_191.jpg
e87d539560eb5577f7cf68b92c5f7735
0c114c1c4e47835fd2b13e695b8a278f77b452ca
295331 F20110115_AACISH horowitz_s_Page_190.jp2
b3f81fd6bc3fddc580672c0d7c75ba1e
193a159d31044109536cc9370f09637b747662a8
60154 F20110115_AACJUW horowitz_s_Page_075.QC.jpg
3c1f3f670d8ed3ecf8a073d70e5a29e1
0720a6131d8ca427e22a1f095ab976c14081059c
51259 F20110115_AACIRU horowitz_s_Page_067.QC.jpg
6eaffea7157aa2014f87574087e1f5f2
f9954ae2fd057376d789a08dfdd51ec729b8c435
1611 F20110115_AACISI horowitz_s_Page_168.txt
fb47f2e80ba7a80748a34a74926b8a01
b73c5c99e8cae1b27efedd8cc686d752f2be5387
45957 F20110115_AACISJ horowitz_s_Page_129thm.jpg
e14819c4b13561299e56a5704fed9427
d4fc2de9262d49a22ff44f2753860ee4b0df810d
74346 F20110115_AACJUX horowitz_s_Page_140.QC.jpg
04cce50f3a7dda63de6849d2503160ee
2f5f20aa214e44d83cb03f36b0ee551abbb3e587
90593 F20110115_AACIRV horowitz_s_Page_142.QC.jpg
a8ed55c74b3e733f7ba84ecb2ac04b59
fe1d1b560452abbb536a6eec8527a355b346eb6b
114019 F20110115_AACISK horowitz_s_Page_027.jp2
7c1d92bec849f13113a2bd78c060a3e8
d1d499ac9235a3ec04bb94751450c6c96acdb998
43582 F20110115_AACJUY horowitz_s_Page_149thm.jpg
6e040ae3c1affe01ffe051ebdecb99d2
fa28120b9d390bc9302715dbb7fe91b74c897348
112599 F20110115_AACIRW horowitz_s_Page_193.jpg
ac1e4b399d9fe4e1c116dd1d67247bfc
53070631f9b41f237178b418d983c56b64a7df6b
65054 F20110115_AACISL horowitz_s_Page_191.QC.jpg
d89a4f10808666120ca315fee579c1b5
d504e54abb2bacb4d2f747c5d0134e4236c21d02
45830 F20110115_AACJUZ horowitz_s_Page_154thm.jpg
0e2ad2b5b838def36fe9896b83cc2a40
11724635eb5af6626d2288f3c2f10c8b14358213
45006 F20110115_AACIRX horowitz_s_Page_088thm.jpg
b7ef71c4f37533ae9e4aec817e9b444e
aa7e2b48beb37d3ff725b513a2c55902d8e9f920
1051964 F20110115_AACITA horowitz_s_Page_008.jp2
a69b03259bc4b35a56109c03af378e08
b7b473a2a6b512eaf624f6f9785cbcbd0834b1cc
347743 F20110115_AACISM horowitz_s_Page_009.jpg
056e7316297252ab949b18129335ebbd
ab82028552208fec67854c06274b4a33b227c7a2
68969 F20110115_AACIRY horowitz_s_Page_019.QC.jpg
5e80a1e6e609b514f6ee0278c6174daf
79f059ba01965e58f1200928048751f4bab63c5e
1009187 F20110115_AACITB horowitz_s_Page_132.jp2
6d84b46c3be432a2080f649378b51c50
ef8c89d8bcef03f346a987ffa0059a4d8f01a21c
107566 F20110115_AACISN horowitz_s_Page_005.pro
1bb3aaad413f0cdbdeb72a1a32a31d68
4cb0ad0656b97aab66e39421df2bbd344d09e1ac
130286 F20110115_AACIRZ horowitz_s_Page_009.QC.jpg
36ad8ed1ce079d233111608223d67603
b3a53dc94f09781d30de0a671cf314071759b965
2446 F20110115_AACITC horowitz_s_Page_138.txt
93064a19cbaa327db10ad3c86cc60fe1
bd1d69b2847aaccea5a2ff62110201737ecbdbc4
F20110115_AACISO horowitz_s_Page_057.tif
234a701d6d6ef88117136882b754ca2d
8ed1746787e0c191fbb903efa420bb1b630237f8
20997 F20110115_AACITD horowitz_s_Page_124thm.jpg
b0b8bb9408c16e4586ee5b794636cb57
d1841840dc93978bf9c5dfa87e2b38aedb70b4a3
78090 F20110115_AACISP horowitz_s_Page_152.QC.jpg
6d8155cdfccaa0bacbd50bc441ceda21
d131b0b24d6bd9b879f63d34a60fb3e970d9c138
706462 F20110115_AACITE horowitz_s_Page_126.jp2
70c934ddeff54a84e45b799b502de087
8dfa18f6445e94d3f58a72b041044d279627f908
F20110115_AACISQ horowitz_s_Page_022.tif
e4a3b3e252af202d64492053e7fbde18
e6c7e024ab0af32c9875a810b0073951c61f007c
126804 F20110115_AACITF horowitz_s_Page_183.jpg
a1ff016928ca37efb3701b8507572c2b
dbdbf471d411739c32abc4f2c0afd2f83ea154a2
76903 F20110115_AACISR horowitz_s_Page_126.QC.jpg
6ee295a14f0927a16679429045fa6130
ab7653b89623666e27cb2210b4bbeb35f74fe958
122291 F20110115_AACITG horowitz_s_Page_056.jpg
faa3ed255611ccb7aab47de226aa9e46
9a06f1cfb18005279e78714477d800192ce6f16b
F20110115_AACISS horowitz_s_Page_207.tif
c1d4973774b85e05d7d6d4a306e5dae5
a2fb4d6575b706ca8b231abecbbcffa72dffb9b4
F20110115_AACITH horowitz_s_Page_011.jp2
81ddfe235050e861fdee3d8e6389c329
8f40971345656789544bc4c4800802bb70919671
F20110115_AACIST horowitz_s_Page_030.tif
db99e468f9a865fed39c2675852e2a83
5b4e16b31c56574988bcc093d68e9027fec6b781
83399 F20110115_AACITI horowitz_s_Page_071.QC.jpg
b3bfb2040d7f7064abd06f6b17c79870
ae18c4823416ab4ba0daecab0e11ad466e4a4c07
205563 F20110115_AACISU horowitz_s_Page_041.jpg
929a71277103c16b9d7ac4f4b26accf6
ae7047d485afc64031c0d5f05fbebc101d450172
881518 F20110115_AACITJ horowitz_s_Page_168.jp2
2287715bdfba21e1e8e7ba192b9d6976
031fd61ad82512308b0e4919fb4c6d37adb5cb1d
216461 F20110115_AACISV horowitz_s_Page_094.jpg
40a59ee0eb9ca1f5789aa0637f1da351
29224fdf311310539815fa7a99be297cb0f6cf6b
100208 F20110115_AACITK horowitz_s_Page_099.jp2
c08cdd7e3e088af2759e89adc411bca1
3c9e68c3caceebbf5c7ac37055c8ed3b96557a52
1788 F20110115_AACITL horowitz_s_Page_065.txt
c172789308294dca6cd9a3017cfbbe19
cb5299229a625dc0180c8c20b183539b31fdf21f
6879 F20110115_AACISW horowitz_s_Page_001thm.jpg
c5052f798f7101dc320e003d878c4e91
2513db79d86319639e51e195be3659b08e196dff
33942 F20110115_AACIUA horowitz_s_Page_111.pro
97aae4891afb43a51f9d2ef3bb83f0db
f5f86ea35af61e7652b9eca999848acf2ef83c9f
97315 F20110115_AACITM horowitz_s_Page_192.jpg
b2758360b996b6afdcb3925fe1d7b65c
5833c025458f4b14ea6cb6a057e03089040e4de2
F20110115_AACISX horowitz_s_Page_050.tif
81ce5db71b068cc3c4c11e02b3552e82
440557234c23c822ef578ae749ab2dbaabc4f35a
35833 F20110115_AACITN horowitz_s_Page_133.pro
97922cfc638801c8d4c25a896a171299
6017964f4c2e46c5c9bba155324e4c2ecfe806bd
69403 F20110115_AACISY horowitz_s_Page_010.pro
0348e492225599a8c93e2dc8b7c37331
50e8fb7563e193ef9516a636aa3a24a6bcad32bd
27169 F20110115_AACIUB horowitz_s_Page_036.pro
5f4f681408fb6e1fbe1215949483f13b
fe35429417aa8a8df5c268b71148a3321579abf6
F20110115_AACITO horowitz_s_Page_187.tif
f45a9ac7b67c1b849a6cb076ae29a25e
28df277b70ab84cf940d75e96cd66fec1b5e63d8
24364 F20110115_AACISZ horowitz_s_Page_207thm.jpg
a7bd2daf07804b67eea0e9baf807bc6e
06ba7da5828d8853a16e1d315b8dd855760fba25
635836 F20110115_AACHRA horowitz_s_Page_048.jp2
d4c8c0d12c6d880b85bebc1034e0ea09
648f838c5c47d4c73ad20cd8bb0a4a1b15dac869
50974 F20110115_AACIUC horowitz_s_Page_110.pro
3c7d97b6e3aa7dd0497c0bcb07ada161
37d1f74e32f9266f2af444fcae5d679223fdbb10
F20110115_AACITP horowitz_s_Page_137.tif
7df24cfcbfe6c199af627e3c1d4cd656
5dd0a652c3e15292c463b27cfd5dc3ebbc725a3d
F20110115_AACHRB horowitz_s_Page_045.tif
d1c146c288fb664fa309813b8d318413
37caa39c823a7eb31f38e2f1c66ac5bc1f1e0057
156517 F20110115_AACIUD horowitz_s_Page_178.jpg
f8de9b8422e1585b9df1c2eff2890fa6
2d0152bf60cf02dbd5335989820f5158e63f55b5
1117 F20110115_AACITQ horowitz_s_Page_076.txt
25d286f81352ee0de41dda866b176d44
18c1c16a80b4aaf872fd21286b4280a027f521d6
28396 F20110115_AACHRC horowitz_s_Page_163.pro
536b8e8c94efd7bfe64f441d52292daa
2c3e81abe26b9a3b6f9667aaa54aa1ff033dc7c4
F20110115_AACIUE horowitz_s_Page_010.tif
4eb096a5f39a6a822174ae28d7afc66c
9c86c7f4c93fb75a6ffea5fb640c289b7f6e793c
214879 F20110115_AACJAA horowitz_s_Page_205.jpg
1345b03b05874fa42aaf797f5ee1c005
e21bfdfd7d656059c5027cef7450f17bee5bfdc9
74892 F20110115_AACITR horowitz_s_Page_125.QC.jpg
e17118cff65b8adc0f46c9de90d80de7
6e241d45bae67e919db64103d18c21eda7ce90a8
65215 F20110115_AACHRD horowitz_s_Page_073.jp2
92c4acd4afd8a81601d7383ded5a6fd1
a769326b803d75af0a91b2fb232619c128f6ab1b
76960 F20110115_AACIUF horowitz_s_Page_169.QC.jpg
76ee89918af75a9392bcb4c1bbf461df
84851c3e8c5baa92d9aab11d7625902429236ba0
F20110115_AACJAB horowitz_s_Page_167.tif
575ca59cf99463da1330f225e033aad8
3aa9a10d4a3b991641b92677631ae8319dbbc53a
41962 F20110115_AACITS horowitz_s_Page_048thm.jpg
daafc3121e2d1c73c4b19a5080a3086f
9a655f75cd31461ecfaa7c188005a1374f00d2ca
61217 F20110115_AACHRE horowitz_s_Page_059.jp2
6a432b9cbff28a996dabcd030fad7c6e
eac58e5f2ccf6f2387b285e40e499033abc93036
24799 F20110115_AACIUG horowitz_s_Page_199thm.jpg
6d23eaa6236219a026da9308e824dd86
cea6039478996caf4b2d4ff6aae20dfcb8e68967
134173 F20110115_AACJAC horowitz_s_Page_162.jpg
4fba794d7089a8bc6eb504748774d380
981beaaa910f9da171bc7d268b6c086343e2f5f2
176571 F20110115_AACITT horowitz_s_Page_068.jpg
bd4978f61762b52cd3dc3187bdb7f62e
4ae02dc68eb53efc10316a42a9d01ca9013e6cfc
1946 F20110115_AACHRF horowitz_s_Page_026.txt
8582f616354e70371fb5d4c572153394
1977d1ca94bdaaf541b8c17dc8e7aef5712a154f
42169 F20110115_AACIUH horowitz_s_Page_077thm.jpg
95dadc25a12e1d211fcc90d1d80db497
e97e0a32afcd11a49c77222973cdce99c02e6e0c
16959 F20110115_AACITU horowitz_s_Page_072thm.jpg
3c0ee9114a7b139bf988b8cf13db32cd
993421ad37c1ff78e0357eeb247e827c7703d9c2
211709 F20110115_AACHRG horowitz_s_Page_034.jpg
46f6a5ea96d2cb9f504ad422431ac4ad
ff244fcdcc7fdc694e6fb429c4d5deb08c559b28
48652 F20110115_AACIUI horowitz_s_Page_028.pro
5cdf111bba0d2fb76d671e9c57b1f0a3
2cf99c5353e93ff39894d39662a7da2a177c2ae8
59477 F20110115_AACJAD horowitz_s_Page_200.pro
7eb05ff6d4e75321210a64863bc1429a
8d0fd07db899f528226a38dcfd67e2dad12074db
182621 F20110115_AACITV horowitz_s_Page_019.jpg
e92c81e7cd94d9252c858af8cc892bfa
9e2f4f7ddba62bb2aa9c8c9107d79aa8380690e1
F20110115_AACHRH horowitz_s_Page_020.tif
d4091f75197ac54543ec72112140abc4
b51b8f8e0e48d9c01fed58faafffd64ce647eb57
1524 F20110115_AACIUJ horowitz_s_Page_043.txt
5c2862c2c5f23d559431e7f67652e01d
47f68143787ad4d3bb89c77854d0be894e00da56
59974 F20110115_AACJAE horowitz_s_Page_186.QC.jpg
2f5af26f0e06930b151c029e9ced3a1e
214fd09999af6e89c7861279c81475d7544f5ce6
2315 F20110115_AACHQU horowitz_s_Page_008.txt
5bcadfd491a092147ad46384f8586562
93d3450e57bd525c7e902a0dc36af5b06da8ee60
159412 F20110115_AACITW horowitz_s_Page_075.jpg
51a505f456dd8c1e3d095b1cfcaae40f
89fb7663315eafd67105dbe7045880490961c807
F20110115_AACHRI horowitz_s_Page_175.tif
0c2e098bdbbde3a3b710bfa10d325380
9bb4c344080211d29e5458ac703499e6bb379be2
176693 F20110115_AACIUK horowitz_s_Page_088.jpg
cfd1c9c8bc6eb034288fcdff980fa028
855b4ef2ecafd7baaff54246009d3439f18af925
F20110115_AACJAF horowitz_s_Page_177.tif
125ebf05f54c568d6bd00cf913fccdef
da266d90590a3a0b80c37f884d725a801d0e4e0b
64937 F20110115_AACHRJ horowitz_s_Page_171.jpg
508df6d4ff7cd43f6a49248bc036638a
ae8dde3aeed330a9016453385280fa49d6116ee0
678379 F20110115_AACIUL horowitz_s_Page_184.jp2
53612d575a5e93fddbf91a9507913e84
086fa2566e9b401c8135f5ff329ea9886ccc3980
52313 F20110115_AACJAG horowitz_s_Page_111.QC.jpg
ce6fa25584c54a312f21c239118817ee
a0af3a28dc706c4f4e2e0a2771b0792c779ee232
25080 F20110115_AACIVA horowitz_s_Page_198thm.jpg
284ea4a2660ed89b31f8186be541cbc1
4e225913fc8f80e0185c8e62f6df0c49a5e1ff7d
25765 F20110115_AACHQV horowitz_s_Page_110thm.jpg
e05cf0d4ebb8403eb5ce05ac2ea20ba1
27b6fe3448bfd565d66d050b377cea9c83177dfb
43950 F20110115_AACITX horowitz_s_Page_050thm.jpg
5e56a4a77e4ed5dce02c3f50dd3cab85
b06c12fc9a0e041c42c45e91bda737b59a6733b9
144591 F20110115_AACHRK horowitz_s_Page_106.jpg
8154a4390ef9f6129631a30e3ad78433
ae5413d2aeccc28d90c28d9d24157dd901345e81
5015 F20110115_AACIUM horowitz_s_Page_030thm.jpg
d127cc727392359fd487d461ef52a330
b1393fce696cda76e78d6b0b72af805932860e5f
25419 F20110115_AACJAH horowitz_s_Page_158thm.jpg
2ad4a237377709c466605138bf0ecd0b
de8dafb94beb901a81f1f88295224cb58e4947d0
46183 F20110115_AACIVB horowitz_s_Page_132thm.jpg
4d1922278f8c1d2599afcb79859605bb
b3cefa6ce0d14a4c54459bef78f8a05f6e466dd2
4501 F20110115_AACHQW horowitz_s_Page_003thm.jpg
59b30b21d030bbaede8b01f63412cb89
cc2caeb3ca58c5e093deea7d62a879373c9f3085
70986 F20110115_AACITY horowitz_s_Page_114.QC.jpg
51ffa097c47a59155be1e97a7e037c76
e081e2a753b966b15d00fba2d737b133ebe6af1f
24351 F20110115_AACHRL horowitz_s_Page_138thm.jpg
9ba93671af3deee9ea66f7f8dfe4ac29
5f3f4e8c85afd694ffd66cccf18f0dd73617a608
F20110115_AACIUN horowitz_s_Page_049.tif
0966fe3a1e92fb03e98bbb604e771758
89878182552586bb80586f7a8bb9338c19bb975b
2294 F20110115_AACJAI horowitz_s_Page_024.txt
e7a81f9996221381982062488ae15dce
72798ff2c4dea020b00dc55f2dc8de874bc303be
200376 F20110115_AACIVC horowitz_s_Page_090.jpg
a53318149b4b0ea5bde67207edc71089
6add307bb24c0ce55c7493c5584d064b0f146fec
29362 F20110115_AACHQX horowitz_s_Page_139.pro
9b432015eff33f8e0b6bfa582dec8424
9dd13773688f263c52317538f2d1d49a7e1f3400
35920 F20110115_AACITZ horowitz_s_Page_189thm.jpg
460c1bb78937b6f9f48419cd93b7dac6
9c0d19a7c15a3e31011a9149303b3b73e5eeb921
220094 F20110115_AACHSA horowitz_s_Page_158.jpg
1d3585a67ee79ec930c25d49771bee15
40bc32e1454154c87b796fa45dc7b0731bd30400
42390 F20110115_AACHRM horowitz_s_Page_178thm.jpg
4efe30d837916c6bb6d3a2fa5786719d
350ccc16e0ce2fc22531fbaffb708cedd12f0b3a
86910 F20110115_AACIUO horowitz_s_Page_133.QC.jpg
51d4a48d6a29e9cac226ae4c8513deb7
155275335f1ba914ef34b1503be5b0fc6fdec5a1
1478 F20110115_AACJAJ horowitz_s_Page_148.txt
4081166b899595e9ee8218b006c9d037
9848e6654e1114fd5a452018441f3a127ff186fc
193746 F20110115_AACIVD horowitz_s_Page_099.jpg
7c8ea48e81a0afb57264ba39ec9ba6eb
13b4bdc42c2d6e86147834c9a16037fdf40ac8df
35024 F20110115_AACHQY horowitz_s_Page_071.pro
a9203d51e23fff0b1a9bcb5d2dc5c22c
695925d3b4299512e432491a9f0223ed6dba6808
38197 F20110115_AACHSB horowitz_s_Page_176.pro
4a403cc00657741545c041344179703c
05e497c21c4947709dd3b559a6b0e3e42682aedd
19066 F20110115_AACHRN horowitz_s_Page_159.pro
0be7dcdf5fc8d6ecebb959143940a2b6
aed90476c6d7084156d7ccfee61cce9e37673fc6
100361 F20110115_AACIUP horowitz_s_Page_025.jp2
aab11a2d80ea13c1aa3e6705749cbd50
d4a136f3f3bbc11737f184c42ecf4f2cda3597c5
83391 F20110115_AACJAK horowitz_s_Page_203.QC.jpg
fed82d4f40c281e5ba2709a33995e822
25d1890a0b716fae781f2c231329136d2a796d32
998602 F20110115_AACIVE horowitz_s_Page_130.jp2
55236b5198abc8233075619ed3315f2f
6dd355f39e260f91c8fb18be9a6b7b22fe82b388
F20110115_AACHQZ horowitz_s_Page_112.tif
f0c6a9e5acb713af2f93caf0b98d4836
28c119d1f1067a91c9c0679e236bf1c0541c91ed
51067 F20110115_AACHSC horowitz_s_Page_040.pro
d2e1eeec3497772f79d4f16fc246b829
119eab650e6305550792b4bd9d509289a4b6008e
45180 F20110115_AACHRO horowitz_s_Page_168thm.jpg
dc7f43ea76d43a9287a8de55e5e0638a
97f585815702d1142f8c9055071fa933105cc477
F20110115_AACIUQ horowitz_s_Page_071.tif
34fb7a58dae06a1ef79095be198b1484
9c4f14be609d4bd9b163357dab6bdb69ff3b11f7
23007 F20110115_AACJBA horowitz_s_Page_001.jp2
8225b02cde118bd5f1a777f6968e898a
c78f4bb728ce4c19e099574d8c0952b0735a122d
900059 F20110115_AACJAL horowitz_s_Page_131.jp2
fd0bceee45e4255b8d5d3e1bb52e24a1
348a4b241ba9d5fbf430d82b029b6283309174d0
83397 F20110115_AACIVF horowitz_s_Page_134.QC.jpg
698e3e48507b068c34664bf8a3b10a17
eaf064d4cc992c102337a4c80682ed2e418e8ecc
48530 F20110115_AACHSD horowitz_s_Page_056.QC.jpg
57f67c4d8c8c828363a6325726ba458c
8aa55c6b28bf14b7aec5d011404ec7060992a607
F20110115_AACHRP horowitz_s_Page_034.tif
8775e84f3c331a9b0cc5a6e9521b76f8
0adabacaaad0499a3112976eee8699cbfd1644b5
F20110115_AACIUR horowitz_s_Page_044.tif
d0565e53cf51f4196b4479ee4d2d9730
d9845c64f4acabf866ea1dba4ae616058cb72268
1469 F20110115_AACJBB horowitz_s_Page_053.txt
3e7923dc6cd89254c2de0cf59ae0f815
c14bc7bc23c33cddf54eb94711aafe0e2380584e
1573 F20110115_AACJAM horowitz_s_Page_175.txt
87c06881c4eda1e99c9a460e3c94a9ca
11923d3040f1e70dd38197acf8c2391006f7b10f
21902 F20110115_AACIVG horowitz_s_Page_161thm.jpg
3bfd9ab4ef1419c8c8870c12a40313e2
9c203be45b7ffffe81faa36f33184271639feb6e
45304 F20110115_AACHSE horowitz_s_Page_120thm.jpg
9a28b231aa39dfb57275d3bb27423cf7
869ba025e8b6d8991f75e1d0e82de0eeb88c2fd4
873420 F20110115_AACHRQ horowitz_s_Page_145.jp2
321fa8a3bac5fe386199f56e67af9bf2
8b4d99d8d278e92b249be7040251db1e11cabecb
24902 F20110115_AACIUS horowitz_s_Page_208.pro
5941eaaab3246a45980e954ac1de9c2a
b30f6a88d0b0305d2928cf70f8d725510d7237e4
75722 F20110115_AACJBC horowitz_s_Page_044.jp2
09721d861d0325e582ad4cb2c294f799
4bbadbf8535c6047619594b4d734e99f9db0ed5b
1051980 F20110115_AACJAN horowitz_s_Page_118.jp2
212655b667348e6620a63f96cb28e852
489962015d00e1b781b20e11996a6b69d1ad1a77
275754 F20110115_AACIVH horowitz_s_Page_004.jpg
4394888c119aca6d4690d636e3e476c3
880c27d7c50f262382b209b8bc26bef0b7799acf
F20110115_AACHSF horowitz_s_Page_095.tif
c23aa8eb5060c9a38ab7f97b5cf1f55a
12b94ff0ebad86283c707aeffeaf4d6f39f6a2c5
21678 F20110115_AACHRR horowitz_s_Page_101thm.jpg
d5de581751e18956a7b23f86485d73dc
edfcbec11fbdf8fd88254726d46e6aa1fefe32f5
35945 F20110115_AACIUT horowitz_s_Page_067.pro
e8788f8a1fa8370645e8578aa05c3791
c365fd6a45e56423b4fc01a9c2a84a757cec19f5
49387 F20110115_AACJBD horowitz_s_Page_026.pro
7eee6c139fee54aba0709c36350a4e96
27acd9ef53f1922f2367ef9e248b8acf0827fdeb
1915871 F20110115_AACJAO horowitz_s.pdf
ef82069174a6e3bd4e891090837900a4
5af51ef672743b9baf8b281495ac764583fdae83
211977 F20110115_AACIVI horowitz_s_Page_110.jpg
b881e591828415e31a29acf44f5f0ab3
9c269e1908cfac1a5d058b4b49571c67d9793e6c
42654 F20110115_AACHSG horowitz_s_Page_174thm.jpg
1ea6cb6d1af98cb57772ba8ef77a6668
17e5a832a084ef07407cd8e042a62c64bd85f0ee
56669 F20110115_AACHRS horowitz_s_Page_036.jp2
f39443a3c032fd9484369d10fac9c3d9
a791f6dc126752fd66b7d44ec419c10f9f06455b
137718 F20110115_AACIUU horowitz_s_Page_096.jpg
6c0c9189aabbb1ea443cf18c2e46ceab
ce7a9024395cdfb61a8f2f4c4ba02241d56888b6
162470 F20110115_AACJAP horowitz_s_Page_049.jpg
179ec9e242c799519a8f81f832883dd4
3af8ffea7c1311bdf5127d009ee481aa1a9a6357
1051892 F20110115_AACIVJ horowitz_s_Page_103.jp2
ae4c95fd7692582875f613396cedbce2
8fde9489067a2c6b68821b86de374884d6541f67
61478 F20110115_AACHSH horowitz_s_Page_195.pro
5219d170f3eaa4637d27e5ad19f842bd
467b8268a135c16b7d95cecc43c77f46bb2cdbb5
159943 F20110115_AACHRT horowitz_s_Page_126.jpg
89ee0802b79236d3885eec681e4b3d20
45f3dbd7dfa216e6e5272a136f914c5e016610d6
59529 F20110115_AACIUV horowitz_s_Page_100.QC.jpg
5571e2bfc49399080dd51dbef0f87e18
602da8d09f36b16881f77becb6c9512ee8194d72
F20110115_AACJBE horowitz_s_Page_166.tif
9699bfffb7d56d207160e9a5d54affb5
c492f19f95b253409e8470b20639aca9dec613a0
815463 F20110115_AACJAQ horowitz_s_Page_076.jp2
9335933428fb61c60f7a0e6b0020d577
5ef2459894bd69c288d3718ed21810a3618a72b6
158075 F20110115_AACIVK horowitz_s_Page_112.jpg
93d1bfb6e1634ce9b8554db56df4ab5d
808842b200283cb7f0dc6e50f7d66b31a44b95fd
1265 F20110115_AACHSI horowitz_s_Page_163.txt
f66c02b1bfb499e2ddf40ea3f3f1b59c
256fa07ce2c41a023cb8b56fe9f8c15f22699fdc
28045 F20110115_AACHRU horowitz_s_Page_178.pro
613b8eabc888d40abf951e0582b9e7bb
c5c0daaa3988014243f13e879264a5b5a3ec06af
53215 F20110115_AACIUW horowitz_s_Page_051.QC.jpg
7cd9ecd6bf7e6fed62ed9ee9166c1356
4f27d67971e256c2347965fd792e1740c45f7c8b
1652 F20110115_AACJBF horowitz_s_Page_045.txt
8767be4303f38b8f8d490858b688c6b5
4e2e748c8ac2c2b90139f7c697b8082aa85675c8
23832 F20110115_AACJAR horowitz_s_Page_024thm.jpg
3033dfc8ba0bdb38bd297993e9c51c0a
2d5d30d373b078c3e91c7b7df4ce8cbda386bb95
F20110115_AACIVL horowitz_s_Page_105.tif
b39f45277b45a601ea765df2a6861d2d
271ba971a6c7c499fcf713332a53c4652150c687
70286 F20110115_AACHSJ horowitz_s_Page_179.QC.jpg
4d49bfbc08538ea9b743872653e31c9c
fb4ba3700fa9a9b8b57177b25f706e1f3bede38c
681477 F20110115_AACHRV horowitz_s_Page_125.jp2
b875d81ec769381d41d47a8cd4cf1c0e
5945ec627e7265a272ae0bc7dcb529b3d27a16b0
91313 F20110115_AACIUX horowitz_s_Page_057.jpg
97e3e4e7b66367eaddadcd6d4fff0371
75d9711abd4a1b239793730f59e8a9da85af38b5
219558 F20110115_AACJBG horowitz_s_Page_182.jpg
a9d8700177de5cf8717541dc1932036b
5301843a50a0028c908737ea1c1be8d301a14a53
33788 F20110115_AACJAS horowitz_s_Page_089.pro
cb25a578263874c6d3f8a2d393cb71cf
1690f0ecd788c49da0037c1a741dfa8735f4fcec
970763 F20110115_AACIWA horowitz_s_Page_154.jp2
9f050679cdb4803fb83b4a71ec09c7da
d397874f8f4318b7a1765b5c338b287ad02070d6
64640 F20110115_AACIVM horowitz_s_Page_056.jp2
797653b8dfa93bab51211bd86cd0be45
087bed3344fff5ebe432f0c3c680e9470dde5ccf
1036 F20110115_AACHSK horowitz_s_Page_208.txt
06c9fc240e7fee3310be6d3b885437e3
c939e83d8ea5677be15adfcce2a67ead88d61801
8423998 F20110115_AACJBH horowitz_s_Page_165.tif
bd18bf6c647e68ce084966c0b0e510c5
ae897488b322e246d01ab9cc8e8c634ebde6f537
214442 F20110115_AACJAT horowitz_s_Page_103.jpg
460a19d93bc2ba6e8c0ab654aca9a3d5
b70f1e98d9321fa80897b20f18202d716d035dbf
75533 F20110115_AACIWB horowitz_s_Page_050.QC.jpg
16e2de45a2cabb652755e2c0ecaf8eba
9e5edbc4e0816da5a24111855da571f4b4a6c047
45192 F20110115_AACIVN horowitz_s_Page_151.pro
a71383afb12daace7883d15aa759ec95
0325dbbce9a95c44454e81ac299a872d6ce3d24e
F20110115_AACHSL horowitz_s_Page_113.tif
6e39374d1b95a72a5d1445ce568986b8
b80aab8c573fcf0979c71b30cb575d4bfd383b02
23613 F20110115_AACHRW horowitz_s_Page_145.pro
0485af09508b5c7e3006d4fc29165e3d
fac77e4c6ed978ce372f0116d57ef022b4f2612e
68154 F20110115_AACIUY horowitz_s_Page_042.QC.jpg
7249b8f4371b1c2dcd1ba95e5b9f87f6
f7e1b49314922d4db06c0de4d29fbe8f3fb72f97
59824 F20110115_AACJBI horowitz_s_Page_128.QC.jpg
a31fdc72a66dbf8d6373409e4dd29c10
43710bc771eb16e6244787e4819655204dbc8497
22898 F20110115_AACJAU horowitz_s_Page_205thm.jpg
8ee8b1157bf73e0dba5e50fad58e0362
d5e0e6b5e65f21dbfa11052770c76f852369da10
189023 F20110115_AACIWC horowitz_s_Page_114.jpg
dbb2fea2ace30e18f0ea62fc02e2282b
d3ead96b24e9942b53a88a81c8cb6b8d1eb78af9
198741 F20110115_AACIVO horowitz_s_Page_122.jpg
6386368ed6fea35f506d4dcaa7c0cbcd
cb9af90a0a5547eb4742699997111e0217ef0cba
F20110115_AACHTA horowitz_s_Page_027.tif
8537d49b359845caa6abcd44b8d978da
09ce2f36fb173f0d987f33a99a9036b0bcc2fad1
190704 F20110115_AACHSM horowitz_s_Page_118.jpg
cd532a45562f6a58a7b4a7644e318737
6b43c3c7dc814b15a25a501be74e56ba14689787
211972 F20110115_AACHRX horowitz_s_Page_022.jpg
61056b14ad479bcf04d72bc3ec254fd7
5971eccfeace24f412c5937e57c368876be253a0
184321 F20110115_AACIUZ horowitz_s_Page_143.jpg
a1b873c4174730c950f2848d250a5616
af0f483ecaea1776e8ad0753d46a47eb62ed3bc7
75523 F20110115_AACJBJ horowitz_s_Page_039.QC.jpg
27c79a0e6472138f9ac2d44758368442
340f00d469598f4fce7409981dbe833a97511958
74879 F20110115_AACJAV horowitz_s_Page_045.jp2
88ab60e18ee0c8e48d9c9abfceecc262
9744a154ff71a0e3530b9f5d3b72597597de4601
193475 F20110115_AACIWD horowitz_s_Page_105.jpg
f22c07ad7a2a8097f09400037cacdb1f
6841a57a55982f84ef6f6088108d2841860557b6
F20110115_AACIVP horowitz_s_Page_173.tif
5875914ed534a1258f9d42431a86fba9
d6cd472d97a5599fe54b6e1728a8fdf13d20b29c
196700 F20110115_AACHTB horowitz_s_Page_176.jpg
f08380c198174fdf09ae90ad10515f48
74278e842544dd48117f5a4e2dcb8e1203650296
1975 F20110115_AACHSN horowitz_s_Page_116.txt
0ad4d18e5ba7199cecb0768a3598809e
3cbb50843c81210b50982f3a7cb0938cf8dda8b2
1795 F20110115_AACHRY horowitz_s_Page_055.txt
82be002c0e318b303589b9d632d5efb7
a02cb14058e546eec15945a12ab4eca6cd1c2bc1
30015 F20110115_AACJBK horowitz_s_Page_073.pro
45d3dd2b9ef595227200a87d581ac1c0
d2a2efd7750aa2d94b674b959a5d60779f592aed
1051933 F20110115_AACJAW horowitz_s_Page_127.jp2
8f86043de3e2fd7f729ae46bb875c157
728ac19cd6d861e78895680729251bdfcd74ea3c
1835 F20110115_AACIWE horowitz_s_Page_082.txt
6388ca36114a9b318a1c9deacb0f0be5
1a8a4b79fc11e9a55aab2a5d1f667b6a77de5f52
1710 F20110115_AACIVQ horowitz_s_Page_089.txt
9dc0cff7bc8ea6a3dd942bcc0d18e295
13ed02bdb6ab953fbf67916fbb6d379c63ac5384
82760 F20110115_AACHTC horowitz_s_Page_112.jp2
a2223964e52b157012ce69125da01630
bfeedeee63d6844acbd104a86bfed49110103649
43348 F20110115_AACHSO horowitz_s_Page_095.pro
36ffa1f4b8c6ee441fb11b54a3b68ff0
0ddfeb4a274b223a0eb3f231af7f4031e043ffed
45480 F20110115_AACHRZ horowitz_s_Page_004thm.jpg
a7886e99395dc64007f8f70a86fec699
3be098a5eabc6c9808daa916972ef8b8202cfd37
1705 F20110115_AACJCA horowitz_s_Page_057.txt
c521d93eb4390aa286032b4f52f95420
a4ad9c233386a2bf8dc3eeecd6443a3a893a2d12
834968 F20110115_AACJBL horowitz_s_Page_069.jp2
a10c14104bb89076bf94ce0407b14c53
44d91a1b7d4bbf73846f1f43960f6c6f296af814
1141 F20110115_AACJAX horowitz_s_Page_160.txt
2d9369a4a7b3adb1edc0cd86dfbca1b9
ea255d8dfd650183c7b274878562004e41e0e57b
F20110115_AACIWF horowitz_s_Page_202.tif
2ec2bc1c783c2bfffe49edf594619eb3
d63edae2cda9858bf3bc437f523956fb6f28a37a
183732 F20110115_AACIVR horowitz_s_Page_093.jpg
291cb59074adce3e1893d881a0acf7a9
ac24fc566ad5dec53978c8a8f3fb2172bb1c639b
37433 F20110115_AACHTD horowitz_s_Page_191thm.jpg
15ea8d8963572369a36213eb3570597b
7c8377c094c094f61c2f93220bd6a51f8585e5e6
1662 F20110115_AACHSP horowitz_s_Page_128.txt
23e850682fff2ec222af070a9ad09f7a
a133d8f73eb3b764061c4cf2cc0650726f0d0632
20697 F20110115_AACJCB horowitz_s_Page_109.pro
ee18190ddcc8c729c67b529cd1219d3b
4a35ff8cb4a906782309320f1797c3c3e5bb494b
F20110115_AACJBM horowitz_s_Page_189.tif
816ccb8dd890097d35a1c61348b19e14
80bd807fbf8cd4434c6e61cde10b86e5a599bd40
66271 F20110115_AACJAY horowitz_s_Page_136.QC.jpg
31580a99d80282c24b044c52ace40667
9c3e78464e10dc7db5a529d12ab7b3eded4dc417
F20110115_AACIWG horowitz_s_Page_139.tif
7458450e9418fc61380a55322762505c
ea193fdfeb3357a5db1d469537b677b7e3320a18
141634 F20110115_AACIVS horowitz_s_Page_051.jpg
2b0af1e63c81d6688de01038bec44a0a
52949face34d775fa8ce17d1ea232e9c20fea2b5
18795 F20110115_AACHTE horowitz_s_Page_066.pro
cdee5c9c55692ff133d35a75ccad7200
281ec62185d83a22c86b006b3635b1a2a0bdfa79
42227 F20110115_AACHSQ horowitz_s_Page_155thm.jpg
676a32c701ba98e1d07a50b4a1fe1b0f
36ab8610cae85694c89d4cef2bae18ab6d583c9d
F20110115_AACJCC horowitz_s_Page_155.tif
c582589d667385499c692212b7dd9852
d4b196cbbf335cfe6106b683d394da36660e1ef4
37863 F20110115_AACJBN horowitz_s_Page_017.pro
f363b8782d23c09097be5ac4055ed446
10922e6437c91a34eb5795873d0b90449cce0ca9
394728 F20110115_AACJAZ horowitz_s_Page_191.jp2
10374b696deb1d022ce460e3db225cd9
5fc258545df02a697a381c7af49ec815f960de51
185949 F20110115_AACIWH horowitz_s_Page_055.jpg
cd828fdb5393eb5e0df34bb57c9759a0
3ef494d7ba9892d3eff7560cfc2a6a01d7ad05c1
1856 F20110115_AACIVT horowitz_s_Page_099.txt
b389525e8df60fa149fd26fddd7943b9
c371285d026ff851c52801f271d25e89f13fd340
143829 F20110115_AACHTF horowitz_s_Page_064.jpg
b2c7cd28093a6b62e8a89d099c939318
f08a2ea6b3e367dc1a2553f3b93eaa6682e9e542
25861 F20110115_AACHSR horowitz_s_Page_094thm.jpg
970c7246ad8094bfdc6aa626400d5502
ee69281724701a0b771829a55cff86b1f961e66c
44853 F20110115_AACJCD horowitz_s_Page_107thm.jpg
1619c9dfea191165b9b74840e9b67d0c
040b7455b1dca2adff987aa58d3bcc103c0b500e
31133 F20110115_AACJBO horowitz_s_Page_072.pro
b5129a3ad28f03e1fca096c3b8515dc0
85496151d6d8451ae12422df3551cd2ab9e68374
51696 F20110115_AACIWI horowitz_s_Page_063.QC.jpg
8a3770c9b3ec2027e014f9bada03e0ae
664fabead136f04dcffb6dded3080adce8af0e37
F20110115_AACIVU horowitz_s_Page_104.tif
181adca8e10b03cbbda388a139a8fb71
5f98a0b64c2551edf5afdf1ddd24afd89878b2f4
25640 F20110115_AACHTG horowitz_s_Page_040thm.jpg
47c2a4aa63871e16458a8f090af48942
40c272e919d32b1c18640ad8c3968b852d7f324b
42091 F20110115_AACHSS horowitz_s_Page_175thm.jpg
efba120303cd1be25d0ee23be59ff8bb
f438f9c43b39d60c93a9fe3c6c3fa555a2ea00ef
16021 F20110115_AACJCE horowitz_s_Page_059thm.jpg
cdf16212d0db2f4b11c69c0d2f723dc1
8751c326ddfc24e6b6b2127eb88e08e7d9f3e121
1051940 F20110115_AACJBP horowitz_s_Page_010.jp2
0500647ff9f15e028d09a70da2c1241a
0fa48b0d697a33ad18c4ebeec8f15f04779334a5
44462 F20110115_AACIWJ horowitz_s_Page_061thm.jpg
b859b4e42ceda3bae553e52a8bd383bc
f94b52370f982f88ad7e92678165698a91728176
73433 F20110115_AACIVV horowitz_s_Page_178.QC.jpg
fe9928598e7771ec3626991c761cf461
36d15efbbf152e0995a51441e55403b410579c84
84961 F20110115_AACHTH horowitz_s_Page_146.QC.jpg
1a7b22ef1593f7eaf7da57a97b88f67e
bf964a4ed6fcdc4ee4f923a7ef3c6865ea84f801
1882 F20110115_AACHST horowitz_s_Page_020.txt
acd4f4ee90b7c9885821885e5f394d4c
af11512d0dd08069f41e04f16f1e189ed0c6e859
206883 F20110115_AACJBQ horowitz_s_Page_152.jpg
96ac7d4b77800129bec949c209fa316b
771a7e7b5541ad011b3422d371e494ef960900f6
F20110115_AACIWK horowitz_s_Page_017.txt
05e7ae6633d14888d2cce1767b695c46
52f9c62158e2dba3266fe974b7bde2847cdf6e47
1022199 F20110115_AACIVW horowitz_s_Page_181.jp2
8c39a31b35dae6f6f06a6215796fe0bb
fcc9889ad7abd83e5ffa60299a4462563cad032c
99654 F20110115_AACHTI horowitz_s_Page_004.QC.jpg
fc664655a09d5ff773f6f5d006dfdf2c
e49cf8a3e07235c9224721237d782e99b03f442a
F20110115_AACHSU horowitz_s_Page_041.tif
1590fa7f4f6978ffae53f02d390b9511
f8c85c1b54d300389438ef93ad978c90c90aa92e
740987 F20110115_AACJCF horowitz_s_Page_084.jp2
0fb331808617009ee262be365db45eaf
aae540617981b2e18c2080303053a9ec884ba6b9
F20110115_AACJBR horowitz_s_Page_060.tif
77548f772e71bfd4a2d72252b9e6510c
fb08351d72cf557f7084b9831d9a553158fabaa6
42115 F20110115_AACIWL horowitz_s_Page_167.pro
63378a67113cbc57b345fd58e2084a2c
ded451e627793e99916a3d09a6287c917d85b49b
F20110115_AACIVX horowitz_s_Page_103.tif
eb04d792320de31da17b1d106fb2fea0
9c421ec9201f5f71875d9e38687807aa04c0147d
18295 F20110115_AACHTJ horowitz_s_Page_001.QC.jpg
1039cf94f86149f6547f6eeb01dad4cd
fc67b6f02c67c3f5f1f66102cdf3e70987210854
F20110115_AACHSV horowitz_s_Page_208.tif
08fd50d6cfc3559bab254f67b9fdcf19
9019fe2af8922c55d134bdb273ef67d8b5f3c80e
F20110115_AACJCG horowitz_s_Page_012.tif
a92a192413fbc40e37f79b77473c9a52
8ce094ee40632145f964869db693e4c46a35bb6c
47129 F20110115_AACJBS horowitz_s_Page_135thm.jpg
71f7513fddb5faace8c5a6e5b0bba20f
54032fe65966d88af4d6d6f9b528084d822b4531
79555 F20110115_AACIXA horowitz_s_Page_200.QC.jpg
e514a54c80c636ed0db8d33bb0fa8def
2a01725b92ca516737e9b01406c83fd29de88dfc
23710 F20110115_AACIWM horowitz_s_Page_152thm.jpg
694257707e353609305d41ddd6157219
509b70d2a3d9de2ac513d160009b23fabcee765d
1962 F20110115_AACIVY horowitz_s_Page_088.txt
aa3a7aff5b70d2dd0deb1d4822ea0a2d
9f7bbcb61f5d5307fe6b7b2629a4602a8d6732c6
F20110115_AACHTK horowitz_s_Page_068.tif
39d1bf0291e5a280c1111f7dc4340898
3b645a3257cb35cf2711bfbfd430e33889fd3c46
F20110115_AACHSW horowitz_s_Page_092.tif
8b816fbfce0cbc89d092549ed65c6454
ffdfee8ca5f328f037a58918e77bbb834b2a42da
1978 F20110115_AACJCH horowitz_s_Page_182.txt
d8c7d9aa6f0abc21220966557cacc2ea
70e8b252a1f05cd24d43091451eb35371ec18f46
75640 F20110115_AACJBT horowitz_s_Page_083.QC.jpg
f8950a477b1810d7921e80ea45a1346a
96e02d69e04becef3f67b4c1595316de0db28fed
254025 F20110115_AACIXB horowitz_s_Page_203.jpg
70c53f686467842f1a2c5d5875759c0b
5ba77dcdf7e4a60a1ba22d6c34235e26f1a6a356
2048 F20110115_AACIWN horowitz_s_Page_085.txt
e933391a39b836a5c991403fb8ae6a20
b23c40a717e2e361c2476f99c4ea68bfc1059fa3
22292 F20110115_AACHTL horowitz_s_Page_020thm.jpg
6188f6fde19ff2d01d3622f0dabbf182
8535fbac8202b7fba837ec140c1e1a8ac06fa5d5
1974 F20110115_AACJCI horowitz_s_Page_022.txt
5eb09c2025fdfdf7a886c314474af510
46c9cb52903a7fa791343dfe0fbb1be7be396eeb
669402 F20110115_AACJBU horowitz_s_Page_074.jp2
e32a9b877ef84f66de49111efa49b7f0
fb8d7ec8f14751e9e5d466e6d8935c9eb496c1d7
1811 F20110115_AACIXC horowitz_s_Page_019.txt
a808f59cdab87cb2b09f3f7e453987ab
33405c3f13aab3f9a000ad5940cf8cfda5c468f0
45877 F20110115_AACIWO horowitz_s_Page_066.jp2
9fc6ad724a9fffa57b2ec712bb02f58f
36d1c07578e455bd35c59b0807a51cdba542473c
87393 F20110115_AACIVZ horowitz_s_Page_204.QC.jpg
2911d1c5eecfef0283411bed1dd740cd
97862ccbf10924a0dd4580cd8f43138ba303e861
1013002 F20110115_AACHTM horowitz_s_Page_147.jp2
6277eb1795dc51fd6700931cb992743b
d542b4fe567c9ebf124775003631ee2bef95f14f
63220 F20110115_AACHSX horowitz_s_Page_172.QC.jpg
8478a8a4925b45e2a55a8c4b19d58859
b26efaac854ea91be31e76a8a783d226407b9486
31601 F20110115_AACHUA horowitz_s_Page_148.pro
0c0ba969e9580b72e366178b87dc4ee6
3d8f12af7dae502db8641bf15a9af9cfdf4817a5
77952 F20110115_AACJCJ horowitz_s_Page_116.QC.jpg
4b6c33510e2592c568da478d888e7d56
bbd9d048376aa8d0df250f5ca9f151dd0d4a2b64
166437 F20110115_AACJBV horowitz_s_Page_156.jpg
92ce32b75cd852fde06655659e09d8a3
1356a537107ed623f82482a23094c92b6f75751b
F20110115_AACIXD horowitz_s_Page_132.tif
50b2ece2709e9cc122bdef8247954645
ab69fb92dcc0a0dd327135253b20f50265b7653e
F20110115_AACIWP horowitz_s_Page_099.tif
ca3debfa4078ddcdaef042fea31d7e27
1c1efcaee8ef0de7517ff227a642ab142524677b
1787 F20110115_AACHTN horowitz_s_Page_184.txt
aad8fe983452df5e7eef9fda06863448
2d30086582ed0f6df0129f59fd85f1ed1340a877
83737 F20110115_AACHSY horowitz_s_Page_201.QC.jpg
c1563e8939db6a970112c7713807eacd
2dd86164f5d57227ba6720393a46247b107318c7
26399 F20110115_AACHUB horowitz_s_Page_171.QC.jpg
5bec8724038938e12a13c32c32822a93
768878555cd48cb9737ad64e33bdeafc45d3b10a
1934 F20110115_AACJCK horowitz_s_Page_033.txt
875638f07f4613a53dae15115a46f87f
c36a5b78ee7da832217842a9c6bb1342e196a5dd
1775 F20110115_AACJBW horowitz_s_Page_161.txt
21c0935f187333706dbfc241ad0c624b
99cbdb89e8358b0baed7ec978a083c4e82eabd4c
107577 F20110115_AACIXE horowitz_s_Page_169.jp2
0f16cd05aefa4bff98d5c05675381879
eb30083e434f4c68e64c68918a53a3bc3065a4cf
F20110115_AACIWQ horowitz_s_Page_152.tif
7aa31db0024ea51d8c91176469b1fc7f
9c5d3c6439ac6f9a7ff2f20c2c7e6c21fa0cdcaa
14867 F20110115_AACHTO horowitz_s_Page_065thm.jpg
76e1a793dbc44d975dfb4908043c9bfb
06c9374f69f9c309032d87915ded3103afa29ce6
92790 F20110115_AACHSZ horowitz_s_Page_101.jp2
1bb9c7250ff17d3bfa8822a58f0607e3
21cbb5d2ff44385f73d73be45f5bf74d11c1c900
38031 F20110115_AACHUC horowitz_s_Page_130.pro
9d17f9a89414afad38cf544a1983068d
3d6bc1a2159cb8001870d330a7fdaf112d4de3f5
F20110115_AACJDA horowitz_s_Page_076.tif
6920a24012948c1c5a40ca9b52ea1a41
0716d40e951aa254e253d8f11b548d4cc251c3d5
80289 F20110115_AACJCL horowitz_s_Page_110.QC.jpg
fd0b9324a7373cfe551ff8a25d4cceca
381432c0e231b2bcbcf671c9a8dcdf77de7c6a9f
63982 F20110115_AACJBX horowitz_s_Page_202.pro
0afba2229b361cb3ab9dd5d48d292158
934d505d0a128567ebc424fce406a7ad177fd30b
84509 F20110115_AACIXF horowitz_s_Page_085.jp2
a4c1525ea31c8717bb1ca68d1042de35
83e8168738771566559169d79e2f66497b91b59a
1181 F20110115_AACIWR horowitz_s_Page_097.txt
078d78ce3b1367640ad9c6c1ac5d2fce
6392ddab220e1997a181277f9eb01bcdab4178b2
30889 F20110115_AACHUD horowitz_s_Page_030.jpg
76972717911e4d6059aa3fd1b6bfaf8f
eff35895e03951582e4399fb6b4d14b68b7dfa3c
124 F20110115_AACHTP horowitz_s_Page_002.txt
e02e6cd0bda0ab3a915d37c6c33f512f
ca616803427475cb349882b121375e1456cf94be
17213 F20110115_AACJDB horowitz_s_Page_073thm.jpg
05e08980c7f00c9e3318aea7906db5ca
9e72f16dd6fda683b3a56c627c50675d6c80b4fc
1051912 F20110115_AACJCM horowitz_s_Page_148.jp2
a6cff6387d733c2513cc1b07a7bd6748
cba9207b2a22f31ac202a90696183883822cd377
176575 F20110115_AACJBY horowitz_s_Page_100.jpg
feb81f6d3cd449992a733ac6709ef803
080a79c80aa3cfab1775480bbffe305bb6083b43
954975 F20110115_AACIXG horowitz_s_Page_151.jp2
307729ccb6f0e16583b5ffd19ff1eb08
fde37ae4659f90dfd5200fa610c84e5c8ccb6292
940433 F20110115_AACIWS horowitz_s_Page_052.jp2
a6ab9b1d10b02bf8ad75efd4760c646b
b3d875f460710147dc082633dae55251167a9f41
61679 F20110115_AACHUE horowitz_s_Page_203.pro
459f14b25bf556564cec0d908ab5be8d
1537e6caae67e1fa028152391812b14b0086d493
602 F20110115_AACHTQ horowitz_s_Page_016.txt
843ed6a4bb2250c8812cc0ba19e4c126
d566001299d1a1dd2c29339064d05c072bc4626f
20178 F20110115_AACJDC horowitz_s_Page_085thm.jpg
0fd015be6855aed40ad4516b44795fd3
b0aa9a2f80729246dd1594c632183e96be379dce
58847 F20110115_AACJCN horowitz_s_Page_198.pro
e2613dc86c4040aefe28dbdd81f3c3e3
cce392c715ec503bb9cf104bc5f146cedd079aa2
41265 F20110115_AACJBZ horowitz_s_Page_188.QC.jpg
021d9dce87e9281d08bbfefd855e5a47
b0a1449ad8440a9c853f7376fe6c61a5b2e786f0
192237 F20110115_AACIXH horowitz_s_Page_170.jpg
c458d5625318f1afe429c3c83b50350f
f1bf3b5fbc7d626bac1a2bf6e051eb9654188959
847498 F20110115_AACIWT horowitz_s_Page_078.jp2
54d8b1d17f4198a4bb810472f5330365
c4e79dd903cb583ca1a62ee661c6cebf797b7ecb
25284 F20110115_AACIAA horowitz_s_Page_165thm.jpg
3fe6c15d1b9c4b390a569c38bd4b5955
8d34c424cb07ee0805029e6836b071a52288d3ad
F20110115_AACHUF horowitz_s_Page_127.tif
ac2927cf4903882810b673eed9748004
dad16cff23863e48c58bfba1fb14cfca513c16c8
35599 F20110115_AACHTR horowitz_s_Page_183thm.jpg
465f86976f5dcf218d7bafeb293bf545
328ef1bd81310b1e5d2874684e8844a75476184f
43830 F20110115_AACJDD horowitz_s_Page_058.pro
46d5f70c0ec0627d86d54737e03569ea
64c999eeb9669a1be06f8f09b293ca8c65cff01a
27688 F20110115_AACJCO horowitz_s_Page_144.pro
be2d73fe9ecc0d1576df033a6fa49d02
aa9e7ee27e2a9e0828d1d16f5913cfc26c979f92
F20110115_AACIXI horowitz_s_Page_125.tif
14c43515f624fcf9b51c6c6ecbd12867
49e70e95de13a7c41ae7d0bf277ae23764c604fe
F20110115_AACIWU horowitz_s_Page_123thm.jpg
2c6af165d6e99c65af33eeccff5f3284
1aae69843245ec4f79ef21d114d7695eb2848c67
128292 F20110115_AACIAB horowitz_s_Page_010.QC.jpg
0ef2b73348398becc6c1df2d62b0707e
e8f041cf8501e18e4cfbfd8f3b27914778ede51a
77533 F20110115_AACHUG horowitz_s_Page_120.QC.jpg
08cbb056e7457bc257b1591c26ecaa6e
5ec2501667b652cd78acb858dde6a22d7fd3d433
1319 F20110115_AACHTS horowitz_s_Page_155.txt
9a11cfd0c4d88e48d14d6220d309f159
346d63bc2f2665c30f03033518de97a566c855fb
55438 F20110115_AACJDE horowitz_s_Page_119.QC.jpg
602dad657c1ede214aedec400f76dd0f
ff1e46b73d269d08e28f94f8033492f157257ef0
751950 F20110115_AACJCP horowitz_s_Page_120.jp2
c6a5a6e9e0b3adb54e9069dfa83dde4b
b24d08f031dce6d6d5c022654977c7d1d8094901
13324 F20110115_AACIXJ horowitz_s_Page_066thm.jpg
8f4ec35940e02f21f2b60314478f0f33
a9a2a695606923432f37da02d87501ba8a3eb524
43314 F20110115_AACIWV horowitz_s_Page_180thm.jpg
72bbef7418404ee0f11ed5b5b7dccebc
5a557404b84c38827640e0d0d73c4dd0074998d6
95501 F20110115_AACIAC horowitz_s_Page_042.jp2
6a7d9f7fe090fbb2e72b3c23c94058af
275a3d26e474ec9e0346a4254183c2326474e7d4
80844 F20110115_AACHUH horowitz_s_Page_007.QC.jpg
47d76fd8128cf4b09ffb4ee312e7dd02
62c976cec12f66249d67c7a0c7f76afefaca2a1e
883019 F20110115_AACHTT horowitz_s_Page_134.jp2
ae229a847a6d41c41c39de557af9a15e
c198cd21b052f6b401649df73761b473f5bb48f4
48362 F20110115_AACJDF horowitz_s_Page_029.pro
a9a1e6903d4cc67bccebd700df3b7877
c16dd08e9af2e29584235259bb66bf582ff4f2af
86360 F20110115_AACJCQ horowitz_s_Page_006.QC.jpg
b4401ede68b084f82c594ffa97786a55
2750e9dd11f63fae325b56cf1a3d14bc2f626a50
31344 F20110115_AACIXK horowitz_s_Page_069.pro
718737e8643b5fa364f64745b5b46960
e43b9910a4990c5dab4d7672889ac8926d5a6fa9
742354 F20110115_AACIWW horowitz_s_Page_156.jp2
c264b52d713258f9b40758e3dd7efbfe
9d3898d8132acb137ba8200e4469a29d79a21939
757419 F20110115_AACIAD horowitz_s_Page_172.jp2
fb002d642869b5741dfd802f0011dbc6
489a8c4cc3c821b66a78cee24a54a9b5845acd4c
24415 F20110115_AACHUI horowitz_s_Page_026thm.jpg
604f69d876661e81bb77715ce8b8005a
c2ee83e2ebc25498287535678b76c45b05ff0e12
F20110115_AACHTU horowitz_s_Page_087.tif
32f8995e1211c1c4bad6179491651da5
cce69125a63eb08241198fbe62ae35273971a682
198089 F20110115_AACJCR horowitz_s_Page_132.jpg
b96769be38f4398ddfb60f162d8b7f50
6b9726022e7599bd698370ff71b79e0c9f7465f6
190404 F20110115_AACIXL horowitz_s_Page_092.jpg
508282b061f29c3dbbb208a4d5f4cd06
4af7380e524ea3b5de9781eb7d32db2e07dac60c
272989 F20110115_AACIWX horowitz_s_Page_201.jpg
7ce37775e96e6716477ac9ccb9e8b733
8755bceaa84aa5d0116472f1c40325d4fca09ff1
F20110115_AACHUJ horowitz_s_Page_016.tif
5a612a2072b25f0645847ea3a780f908
e2e78968298965e7639535c185f7f9372a81c827
784716 F20110115_AACHTV horowitz_s_Page_155.jp2
4d8850f0d86c5d21e28bd9a55b36d662
98081d77a00fce32ce894e027f28ad4f4c806161
754197 F20110115_AACJDG horowitz_s_Page_173.jp2
d2cc616cf4d9983176dbfea6e7a25c51
9dba4d969d749af505cba3d28afa7b18d3eb5d7f
39117 F20110115_AACJCS horowitz_s_Page_166.QC.jpg
82b54b947617d2163a58ccc8e8809234
00bc2b527144cca8ba8eb487f5fd754dfc55d821
59546 F20110115_AACIYA horowitz_s_Page_124.QC.jpg
e352ecaa2704c6963b01585c0641b35f
29111d39990ffde670079370fc7b988333148c05
F20110115_AACIXM horowitz_s_Page_131.tif
b226846f96c3a3dd8f6d42f7f3175d3e
22a7cbda228dccfd40615bbb3e01b0f584a77964
772859 F20110115_AACIWY horowitz_s_Page_088.jp2
b72e076d7a1040d7009ea36eccef2f08
ba979d1e178bce2c7c77100c027f79f9ebe12bcd
25071 F20110115_AACIAE horowitz_s_Page_204thm.jpg
df4191bca6df83150acc1700e8a18b07
ad49ccf57018c148b95de8c428441da02d4d6aba
F20110115_AACHUK horowitz_s_Page_178.tif
767d29ae6245b7f5bc5c231e9bc734c1
e95902a118030bf6aebc24111276d10e82df475c
F20110115_AACHTW horowitz_s_Page_158.tif
6ce60570bf7ab8e5881401f307ab8c74
3908c651093238772b63854146775654ea559875
F20110115_AACJDH horowitz_s_Page_021.tif
3dab7c1f051f497182c88d16192fca3e
4e2748fa95bd93c1e1501d2186a4eb26f6a36014
69939 F20110115_AACJCT horowitz_s_Page_188.jpg
3f3647014ef6831f08093416022e350c
020c048c6f09b0c646837cd1f2a1e38965500d04
27769 F20110115_AACIYB horowitz_s_Page_125.pro
03927f3be30edd07504a915b77a6deb0
b87aa26a833dd6999aaf702a268f0f3c3cf47d59
690036 F20110115_AACIXN horowitz_s_Page_062.jp2
13b922f021de0487033c3e02d4267f13
fb468852768079e090d655ee62e7d7ad52681189
185917 F20110115_AACIWZ horowitz_s_Page_168.jpg
ea50df8d8592e4f1e8424308df31df1f
8159cb791ec5670c3301882cc28834aa4e5f44ae
44875 F20110115_AACIAF horowitz_s_Page_089thm.jpg
657b64569886379529601f7163e8b358
58241ee203407b14b90c7312f0862e65d9643416
974354 F20110115_AACHUL horowitz_s_Page_122.jp2
b7aed2765bd3b32e847054eee9a44903
03f5efa3f19fb4c9f45ab68c5bf95799e1d133fd
74587 F20110115_AACHTX horowitz_s_Page_067.jp2
638dc0920723dc9f449b7321cebe6516
2f37c0904b19754c159342aac0a7c8382702226c
57784 F20110115_AACJDI horowitz_s_Page_008.pro
58e42e17c390a9b14f1c9d2d3da3769f
a3eb87a550b8a8b80e07594aadb0d0cdc72e8e62
8184 F20110115_AACJCU horowitz_s_Page_187.pro
39a04ee82da269551d01a95a9ab0cc9d
a73fc9e2e4f51da9564a858d4b3602f2131fdf07
17160 F20110115_AACIYC horowitz_s_Page_095thm.jpg
ee4f873c41ee15a5739185a1cf6427ea
2c27c4fdf468d18ffe35c44c12cceb5307f447ac
F20110115_AACIXO horowitz_s_Page_123.tif
d054efcad5d193182a85550e25bf3192
e1a33a90637a5281d52a8146e4f1c9c785e2002e
25057 F20110115_AACIAG horowitz_s_Page_033thm.jpg
6d597f9fb7e204adc6b13570a57bb6a9
ba68ef9ace3bed562d469dc8d9e400cc6a5e5cc4
541540 F20110115_AACHVA horowitz_s_Page_159.jp2
6e7118849f2d63b6ac5886eecd18572c
c899d754173d50717aa8bc9e3ada7461d4cafb07
78240 F20110115_AACHUM horowitz_s_Page_074.QC.jpg
7e42e79d2158cbfd7b76f352290ad0db
59eacb85a50c556f23789336dd1df1fb8d35711b
25981 F20110115_AACJDJ horowitz_s_Page_122.pro
f8983fdebdec33e2772cb9473900c824
eb807e9884faf7454065cea0ae331ad043d17d85
F20110115_AACJCV horowitz_s_Page_077.tif
4e95c4d4b2cbd94a604e2f69d4d05ad7
df98db5e859e5c29e6ba102df255e5dfafb18aa5
50076 F20110115_AACIYD horowitz_s_Page_116.pro
44aa8570db730045add624f52297492a
9ca6f85b914615bf31e36feb5aacf0d9439c6a47
F20110115_AACIXP horowitz_s_Page_039.tif
b2f7f57a8e7c68a3ea19b27fca08224d
be4bd5edbf0a1ec8597a06e5a9910a3bb19bc5da
572 F20110115_AACIAH horowitz_s_Page_171.txt
0f47dc24f752ed8cf13afd2084c21d89
8fbf9d59b63d6d46ebd40b62151d36a012cb5aee
1711 F20110115_AACHVB horowitz_s_Page_149.txt
69f5bfc57596a7676f5f747fe045f1c4
6d164db91d38b179cb19efc0ba54359edeb3958d
2950 F20110115_AACHUN horowitz_s_Page_006.txt
8cf0c412b7a1c4157c1146c136a9eed1
15e1febf12d85a0907ce01afe172679de1edaf6e
110720 F20110115_AACHTY horowitz_s_Page_065.jpg
0d8aa5ad8241951ccdc01ee21c61e51a
2d89d4ec40e12a5edce9cbde0676391a5c4d558f
F20110115_AACJDK horowitz_s_Page_154.tif
7d1fa655cffba93381f04f58421d71b5
599125088dd378825a91f8ce73826a64ef53aad8
F20110115_AACJCW horowitz_s_Page_203.tif
916f6fb3e1a7002607893b5def76068e
b54a369be1f67c6f7c56f50b709b15aac611e8a1
78605 F20110115_AACIYE horowitz_s_Page_195.QC.jpg
3b5da56c416d55e0bfb34a9d1df01a0f
960de4a269fc5f1eaf884ee530c08fa77b521274
42245 F20110115_AACIXQ horowitz_s_Page_016thm.jpg
df65d6268d571340bdbdfaad8f200c99
052df810bcc7a446f26b07457935629305670362
88120 F20110115_AACIAI horowitz_s_Page_153.QC.jpg
dd5f679d7da2cb6084c834529bdf60ce
82ef9af85283e5ab3f343f579e98a63253956455
151458 F20110115_AACHVC horowitz_s_Page_014.jpg
75dd9d3ca3475d60a1e5326ee578b745
3a092bc553aec9abb692876bb8e9dfd73ffe68fa
138317 F20110115_AACHUO horowitz_s_Page_172.jpg
f37894f1f22ae8960ccd21d733b7f541
7efa3aa45322c9adb202985aeffa7dfa0995bcfe
220923 F20110115_AACHTZ horowitz_s_Page_138.jpg
6e05d687840c30cddf4c66d6899183bd
f49849cb8d3bd70f1b2022cbaa2bf4de52611344
21058 F20110115_AACJEA horowitz_s_Page_160.pro
6dde5a48b449bf2c228ee994a27e32cf
a326eb897cffb320b6ea23489ea9d3b7395ece0c
253332 F20110115_AACJDL horowitz_s_Page_200.jpg
7386a3b1aea653724d337e613556f521
d1ab25e7ceb4e7c3bba5ea8ada5a783f32812eec
44460 F20110115_AACJCX horowitz_s_Page_071thm.jpg
84183bf83ebff5451cf7c2addaec3516
96065ad4a3a8d594d9fe19fe5374cd794bd9d6cd
4853 F20110115_AACIYF horowitz_s_Page_190.pro
30eb8846210d07f3b2f345174a2bf9a6
00605f483287a281402bb6a45e633d913bf5d348
1279 F20110115_AACIXR horowitz_s_Page_177.txt
d65e77d1860706a941fad50c0aef3ae7
98a74e1a4035e0d5865735464a2dbbb69a156227
164867 F20110115_AACIAJ horowitz_s_Page_150.jpg
c8390d51cb58a95e87aff966812e193f
08abeb3ecc5143c2f497614a676d19e7e94827c4
42955 F20110115_AACHVD horowitz_s_Page_125thm.jpg
0b76cfcf480a87121fc3f137c130a767
01176f3a7225e4255d399f55aabe33790950518f
4432 F20110115_AACHUP horowitz_s_Page_005.txt
afd0188ba993267ec0eef97fa1b62168
71911c620e5ae24caba4de630bbbc0a3b8933119
F20110115_AACJEB horowitz_s_Page_008.tif
a65a5966d992f7568f1540fa59cf6b3d
0d013bd32227da6edca99d71e96bff1aa493cf53
24344 F20110115_AACJDM horowitz_s_Page_039thm.jpg
d34c350c3742e6921f02429ae3de92e9
bf948a7e5b6f8cac111d4b0558280b07665f0384
24373 F20110115_AACJCY horowitz_s_Page_103.pro
20193ed5c25929bc844c22c8c12ab05e
58cc03055b5837cd4e5d5c0507504607d038f4f0
1051981 F20110115_AACIYG horowitz_s_Page_146.jp2
30c03fc7e37d274773574a37738ba6a7
b23555af8b44d987666d44ae3af72dd7f0f1c310
F20110115_AACIXS horowitz_s_Page_186.tif
bb8f26e19d0a7feca9dc3e5f421d723e
9f1405d24204f248da0dadc95a62d4dca8dfff69
19503 F20110115_AACIAK horowitz_s_Page_044thm.jpg
fef0d665c1d1494b01a2e23b3f197245
a431c9777e87682d3e970950c87b5cb862c3184a
57952 F20110115_AACHVE horowitz_s_Page_138.pro
f419e28228256541df3c3d9b063b2ae8
e6665b09671d3fc6d2e110c5ea6c04f01fb647f6
F20110115_AACHUQ horowitz_s_Page_094.tif
1ae8277d449144b9bb5fa7dad6586178
98657716f463663d7d751541083d0b98f7cd43a4
104763 F20110115_AACJEC horowitz_s_Page_127.QC.jpg
63f3482b90b90b4eed439b5cfeab8629
0e33e7cee63df24c104202cf76ec44f60b6fcfe0
158620 F20110115_AACJDN horowitz_s_Page_095.jpg
ad845bdeb9e0cf6e592dbedf0d8a84ba
f66ce276a9ea85e849fcec3afb586a0591c4092e
46190 F20110115_AACJCZ horowitz_s_Page_117.pro
ff32bd3d6360739207f77ac9637570e1
ac436b74cb8a6f84241fc4afefebb51f95e725e3
32658 F20110115_AACIYH horowitz_s_Page_188thm.jpg
265dfc9ca9b1630dcfadbc2c265f063c
ee8039355fcf73abd0f45ee1218cc773caa5fb2f
60051 F20110115_AACIXT horowitz_s_Page_047.QC.jpg
ac9dfab96e6aaf911dcb2f34cc2acfe8
92664cc85148e31a170c31344a09d8c6254e9a8e
59876 F20110115_AACIAL horowitz_s_Page_194.QC.jpg
c19ea7ff1d300161d94905785a7f3d15
80fa8a2446e60e6ea2e006c2ba9e2e5f1d85d9b1
123210 F20110115_AACHVF horowitz_s_Page_160.jpg
e6eda0fbedc86ec7f1eb701724c55669
1d82525840008ac6c197103e24566ca08ed5c861
2209 F20110115_AACHUR horowitz_s_Page_207.txt
4de827043020d6de77c2e12bef05d39b
e53cb11469c9378ed29af63717a5b7c95a9f950d
F20110115_AACIBA horowitz_s_Page_061.tif
c148162aa8b297b5500bbc623f48a1cb
8ab72bc882efa416bc3e18c5425259fa98cd2bd0
1148 F20110115_AACJED horowitz_s_Page_105.txt
321ba66b493e1865a23475fd187d2e8b
d65d698ad750d296f702b20942929b71c191edb6
29682 F20110115_AACJDO horowitz_s_Page_126.pro
6e1c32c76ad41ce3757610e742c92080
d8666380c3a6f700298212db9b189cf68b35b33a
141863 F20110115_AACIYI horowitz_s_Page_044.jpg
7beba854236f4a76fef1eab1884f8549
91fb6af8a6ebe699e57cf8b6302458fcb3a99739
855722 F20110115_AACIXU horowitz_s_Page_129.jp2
3135189965b09d841888803d3b747de3
c15a6c317b175a93328b8fd6c508c2fc5a972c47
F20110115_AACHVG horowitz_s_Page_056.tif
8e656f4477a9631bd3e8d334ac6fa8cb
ac4c16c42892dcef278e0b5981804a790781f058
1532 F20110115_AACHUS horowitz_s_Page_081.txt
f681d501afe34da8179ed6dc9ddf42da
7fa5d10d78be0319a63eb0878336e056b3127a05
F20110115_AACIBB horowitz_s_Page_197.tif
4ee27a466504ce6ac04bb724e04c3c66
8892cea2ba195ba8590336940657afbc3255fcf1
38748 F20110115_AACIAM horowitz_s_Page_124.pro
ca6b24c26c4694c7f95b9e2a6fbc9845
3fcadeb5eb0d728eae1e9f97d0b8b54010b63c81
24148 F20110115_AACJEE horowitz_s_Page_169thm.jpg
4ecb1663af04779d016e88ebd4af801d
d93ec765f4f721c8a5a35414237c1b8374883e98
2478 F20110115_AACJDP horowitz_s_Page_200.txt
9972963e46c48c51e4d1d83cfcd96503
4ca3f52932d677d277fdd4ba9ac8e1ba3eea21fe
144103 F20110115_AACIYJ horowitz_s_Page_016.jpg
45d7cdc19399592ffc2d09b2ad5f88ed
8a2715557e288dcbd131838e412fc4b83f334d17
76403 F20110115_AACIXV horowitz_s_Page_123.QC.jpg
6c86c2e956a27f23989a128492ce87cd
6d5e5e393e93cce4b7cbd46dc4703a265a898765
263242 F20110115_AACHVH horowitz_s_Page_206.jpg
703c262645494b6b1ba6382a5b4fdf4f
a61ed97661226b0c52580d1f3208ad2f88d8d487
134635 F20110115_AACHUT horowitz_s_Page_012.QC.jpg
70fe4f93b09d80eacff337b98c761a81
a15cd7543711475249ed89c081bcfdcf5142c088
3235 F20110115_AACIBC horowitz_s_Page_188.pro
f8da4f141dace5b301ea7e00299c0168
442a863484b0a23e2a310c1b2c697696f7d5ff5e
1283 F20110115_AACIAN horowitz_s_Page_066.txt
f2ad2186ddb3448d23b6f29ee8bce5b1
104d42b49ad609c8589fadedc6a8f3650432309e
274038 F20110115_AACJEF horowitz_s_Page_202.jpg
8c8712ee12c5b02618cff084489b2681
d5ce4d8af4f4c328226f26dea9bfcf8afeacd3a0
68183 F20110115_AACJDQ horowitz_s_Page_070.QC.jpg
d3ce100f709b923adbd070a10cd46302
08bd6ee981daaa96d89d209df9129bcfa5299ea9
1727 F20110115_AACIYK horowitz_s_Page_130.txt
55f076f42c51966f42225dcf67671f16
7db443892985cdb5bd9cea16ff54e6f86c4c2030
519928 F20110115_AACIXW horowitz_s_Page_160.jp2
af01fed34c0fd46a759c72b6cac7aafc
b4c2783dd18157c10db4e23899e4821ffe98d06d
132460 F20110115_AACHVI horowitz_s_Page_054.jpg
b0e8493c8897120cf8deb2e6468e7360
3ab2a12ed2addd66b723a88ebb3bd5e4bde35836
76580 F20110115_AACHUU horowitz_s_Page_086.QC.jpg
e3e477e0c83808564901e3271408ae7d
7be0ea6578889389784caf384bc76f23e042114f
F20110115_AACIBD horowitz_s_Page_181.tif
20b99c965a399a06a29e813976568fd8
280dc2f747dddeb64f68e36eb13bcf36616e9f00
146194 F20110115_AACIAO horowitz_s_Page_048.jpg
c768593b9ed9e65dc896a076e102de57
f0e3cbd4d5441ffd729529407f73e10bc0d4d322
127263 F20110115_AACJEG horowitz_s_Page_195.jp2
79b3008749bfa8c67953a8942dc41a3e
f3342cbc3fb5e15f95cd791a68d6cfcfbabfccd6
212800 F20110115_AACJDR horowitz_s_Page_040.jpg
2c3e8fb9b96081d476f7844368bb86c6
2b6d5f0631b96b4da65cb2ac0bb9ceeafd07f172
82165 F20110115_AACIYL horowitz_s_Page_105.QC.jpg
78fe851be45e50b1c69b68c5aa1a94ee
963bdf0191784ea67f7896da60f1db6961d4a887
1372 F20110115_AACIXX horowitz_s_Page_084.txt
46bf00e274329cc5ad20698cde2c4b2c
45e96d029ea7761d2e5a931eecddc0495b4d624c
42328 F20110115_AACHVJ horowitz_s_Page_052.pro
1f5f4dc4f6f8dc0e1332d598e35215c0
5b85bf8cba8a75435531d981a49eb4888614f9bc
659808 F20110115_AACHUV horowitz_s_Page_032.jp2
08c3726ba917e3f223696ac3a92446d4
409f3cc06c9f7a5f108f78f93140ed6b519410dc
32521 F20110115_AACIBE horowitz_s_Page_087.pro
215753275382bc68630543f777df41be
29a44f83610cf0afc981b76f7fd8114e2ac24922
98685 F20110115_AACIAP horowitz_s_Page_098.QC.jpg
cddcbf2d114eca03a671f072ced738a2
be2fd22bd361675e565888b215e9722779c0a2b3
177928 F20110115_AACJDS horowitz_s_Page_087.jpg
14cccdf411fcac01831963a253377c12
0ba88905ba2dd559ed3a5cde76b5003685f496bf
F20110115_AACIZA horowitz_s_Page_028.tif
ba853a8f0689c9cb8cc590d0e9e84f08
4535c7c6ebb6fe2cf2cabab46b62e95e88df6da5
F20110115_AACIYM horowitz_s_Page_115.tif
631e44a38811445202f7341b1d75c362
bf7298c4bf1c0571617d1f4f6cbe44ade787cd3d
159623 F20110115_AACIXY horowitz_s_Page_184.jpg
dfe3a3afd1c6ecadebb2d83d3a9fae5d
132a08c793ef02de26210344285886c7189f7010
2403 F20110115_AACHVK horowitz_s_Page_140.txt
b32fbf34619fde5e4ec84389e5e16471
6f922b62fa2198fa0991e832b42e73c1241d2bac
1733 F20110115_AACHUW horowitz_s_Page_167.txt
4811bbbfae596eb4821011fe17f2f324
e4070b047d72a962e3262030b17bfe2def97a68a
1360 F20110115_AACIAQ horowitz_s_Page_079.txt
dd15856df3555513fd636002a7815bd9
de79f9dad48ee016aa84a808ed4422290c2a5b37
57479 F20110115_AACJEH horowitz_s_Page_011thm.jpg
19469901ce5e8a3ddc1608cfad78841f
9537ccd34a4e43d46292a86e420fa908910782bf
24338 F20110115_AACJDT horowitz_s_Page_203thm.jpg
f017d455927928497a29c27bedf3c6ea
3a2d17ec7aa3db8fdd1234ebbe5b948dfffd3f16
50013 F20110115_AACIZB horowitz_s_Page_022.pro
3fd5a5bab36a77c455901394500b4d1e
9d68638e5a2a1531f7c39948d0977b2f1d73e75d
882801 F20110115_AACIYN horowitz_s_Page_185.jp2
b0795fe956adc0bbc648f8c5f06989df
74b66bce28cf308bea40b2ccb845a2911cc56e74
57969 F20110115_AACIXZ horowitz_s_Page_199.pro
6948dbfe650eaf0c6412b3badd7d3f62
f73e23eb1ad955ab0c16218d13f98f5ce1420694
4927 F20110115_AACHVL horowitz_s_Page_192.pro
a0699331f693594300777181222c88b2
96b74c4cb732c3679607f71f64e1c9e49c07977f
F20110115_AACHUX horowitz_s_Page_025.tif
ef228eac759c3c998798729895f2dfed
85d55dbb7e64e37c14233f8f5af038902ca8d1db
F20110115_AACIBF horowitz_s_Page_043.tif
5178f4d57f5a83f2b534e7fd336a3885
db48a9405e7af7618df01e9ba051548d008c2930
F20110115_AACIAR horowitz_s_Page_118.tif
244f5cb381b3a3a2362d0e3172cb02c0
6b75ed34b3018374a85d7b4574d6ac709790e39c
150420 F20110115_AACJEI horowitz_s_Page_077.jpg
1360be28630896328747b71fafa15a80
cce26a879ebcc54978012675be409ba51395d9a5
1641 F20110115_AACJDU horowitz_s_Page_133.txt
3cd9d8b395a08b199438a7f6c1ed6c80
1b012489c8a637dfd725c6c3536b21a6c7a3ac48
80066 F20110115_AACIZC horowitz_s_Page_069.QC.jpg
338274629f070eb85b71e42a7edd8195
8c3b9542092f35fb16aaa89f7771d5ccd51acb2f
15602 F20110115_AACIYO horowitz_s_Page_002.jpg
08143d63ff506a56a8bf5fa97aa36229
6d330796c4f2113d11007d09e9b792b4f79f4a65
2242 F20110115_AACHVM horowitz_s_Page_100.txt
11563c88c71a0c3e901d21d2e12096a3
c9a18742c785917348e1e5137662905c54c59057
245993 F20110115_AACHUY horowitz_s_Page_199.jpg
91347b21b6958cac27c9a6d095c177ae
85e72f5af0e3e300f39f293290e4a600102cf635
258302 F20110115_AACIBG horowitz_s_Page_008.jpg
3f0edd30b55caf61c4ab86f26c9772f3
98d2828d29ac4ec15f0fe689f42b5eea011921ca
194911 F20110115_AACHWA horowitz_s_Page_142.jpg
1236262ab24e3e57a50fa33fd9839116
dde2ef9e02eff50757b43d38b3f34f4aa6308c43
83558 F20110115_AACIAS horowitz_s_Page_027.QC.jpg
d7ac2b819d9c3ca0bbb92cea3879d2a9
be9099e43d1f53415266ac39d620b32ce6645174
110965 F20110115_AACJEJ horowitz_s_Page_116.jp2
ddaaa068ddce770d22af552f5e635021
8878f214b865affaf06618c15032e6d3670c2930
168508 F20110115_AACJDV horowitz_s_Page_108.jpg
41def39eb30489766ac2fd4e7d5c28b6
af58219c3b360afe272aef52397754a9f249fcb6
872120 F20110115_AACIZD horowitz_s_Page_179.jp2
d4080b038da49f77788556f42d2c2a9f
78317721e55b64260de5e01215bb08fa77943be7
59109 F20110115_AACIYP horowitz_s_Page_165.QC.jpg
9aca1eae6b71e7bec3ec89eb6744989b
52505df8c8212cbb0d09a90ddf04697cd9e8d54b
777760 F20110115_AACHVN horowitz_s_Page_050.jp2
6ec54acd45bbdd66ecd6c0dde2b8c0d5
b98122a467acd6f5d327ae4a0b7bbd5286b814b2
68109 F20110115_AACIBH horowitz_s_Page_177.QC.jpg
8d575f710580080b6b65b80aae3fc91c
d7f3cafd20e7e389b29d7a95d3a99c22dec587f9
102900 F20110115_AACHWB horowitz_s_Page_117.jp2
1b4bc528cbe2ce87314b1725238746b8
036fad73bad10e5a19d092c08f3585d1a968c02d
28302 F20110115_AACIAT horowitz_s_Page_177.pro
d696d4ebddf148b0482f4eadf33d685b
a92969c0f541d306e29e6d1edadb9513832afb70
143106 F20110115_AACJEK horowitz_s_Page_111.jpg
b63c34c8874a417ec880cf1efdf48910
0abee8544664fe70b49918837c1c0e9d4166d142
78423 F20110115_AACJDW horowitz_s_Page_138.QC.jpg
27e3cd49eacfa9fbcd4582bbc1ad46bf
1dd10976d40d5846390b47a49916fc1532fd0ead
86144 F20110115_AACIZE horowitz_s_Page_181.QC.jpg
487f152c7cd7126070144e4340120fbb
a6c25f286fbbde976968ed153de72410ae11f122
2704 F20110115_AACIYQ horowitz_s_Page_009.txt
c625c2b347ae78e7f31fa77923f8eb7f
e05d58c58b8e59510f83fc3eb6966cf429fe5223
F20110115_AACHVO horowitz_s_Page_114.tif
74cb10c72b3c66352460d1a144ac1038
194d3f9eba0f586cfb335397d6636a1ffc11bcd8
938292 F20110115_AACHUZ horowitz_s_Page_017.jp2
2e914cee58644a3e5b00bfe1de8570fa
c37ec3245187d35f59fdf8d7a95e04fd4960878b
146633 F20110115_AACIBI horowitz_s_Page_109.jpg
fe9b9fe37342a9fdd549efc1422d27cc
766eca1eb81450b6f0f42dc5e5aaaec8e251a1b2
201547 F20110115_AACHWC horowitz_s_Page_117.jpg
78d44fd2beb3d9c1bf44fb6f35b6de50
bc9e2a38c18befaa7b6991a3b5bb0ab6b6474492
F20110115_AACIAU horowitz_s_Page_204.tif
9c59d439a0a4454e34591cf81f9d5912
f8b38dcd34ba73f1f9c44efa6978245533b58d41
11058 F20110115_AACJFA horowitz_s_Page_003.jp2
dd7a8bf9a7ffbc93668c546539a3f099
908051aa335ffdaa771ad5a151b9a6930599a801
1013 F20110115_AACJEL horowitz_s_Page_183.txt
7f1b1920f16c79d3c415eb3aa2b70b37
139c10e1c5fe7561631d62a86bd81a4d272e6f51
F20110115_AACJDX horowitz_s_Page_030.txt
e18a59e06d57bf0eb4898a4f1aaf431c
2833648d50e5bfaefc8aff8f61beea90824b8614
88423 F20110115_AACIZF horowitz_s_Page_182.QC.jpg
03b54b5ce354797ed927393c72d83356
969b40c06f848cb1ee8566e169d77a972c7d5f4e
89856 F20110115_AACIYR horowitz_s_Page_122.QC.jpg
fdbcaa9a4e01fb56494926ffacff38cc
ada666c7c615156e7ffb5c5fb03cc6ef741672e7
32917 F20110115_AACHVP horowitz_s_Page_142.pro
abb3bb03ae7d37657fc596eaa706219e
6aa71985f6ff455f01494b655ef7715a65d26c1c
184791 F20110115_AACIBJ horowitz_s_Page_089.jpg
19f20c6722dea5e1f75179cb92390a4a
7b7e6b87336bbb42ddcfbfbd628a7fd8249ec045
F20110115_AACHWD horowitz_s_Page_170.tif
4f2f9a123f12cb6026f46c02dc1eb957
7f983dad5bd831e62f54ea68896841498b0f8b8c
F20110115_AACIAV horowitz_s_Page_058.tif
793c24e365f6f2e21b4bf4f1e6521e72
6513127786bf05e126ed5fce9d509a8018069a8a
94122 F20110115_AACJFB horowitz_s_Page_017.QC.jpg
9b350fef2a30eac86d45d07b413a59d2
64bcf9b69d0ca5ca0e29ef2dba0cb9b4077b4c7a
F20110115_AACJEM horowitz_s_Page_051.tif
46da356b7e537c014b9c6c9160f345e0
7fc9f29b6cc91bf5176e21f7fe625f571af1d5c8
34558 F20110115_AACJDY horowitz_s_Page_045.pro
105febf7bf204b54b56f53851cd9c14d
541a7653447e9bf08c479fca0d2ee1cfe8e3592f
20084 F20110115_AACIZG horowitz_s_Page_054.pro
658e8ab9ab686eeb8171093eb2e29bb3
ad416240fffc9144b5dc58f86bbe4acf8da85536
46541 F20110115_AACIYS horowitz_s_Page_181.pro
15432c20887fd4e59d317c56efd5b64e
465ee9e36e255f6f30d3fceb485a24fcbee44418
F20110115_AACHVQ horowitz_s_Page_093.tif
52473fe2f2227a37b7ff919a8566fef5
d63fb318776000ee75a3a41083de85d7e79deb5d
86939 F20110115_AACIBK horowitz_s_Page_082.QC.jpg
faf8761a636b68f3bbb14c6251398b38
8f5faeba51d96917f46997e9b716ef1b8e9f7a70
60940 F20110115_AACHWE horowitz_s_Page_187.QC.jpg
fb899e61e1139a265363f7e873570fd5
b3265316f437118693ece4d559279cf2b7f41e23
39229 F20110115_AACIAW horowitz_s_Page_173thm.jpg
3007a54ceaf25d40a9424de1a84d9f2e
d0c0843f3b0cdfe0f515f7703aaa88946099a146
16528 F20110115_AACJFC horowitz_s_Page_121thm.jpg
e5b0b8a660d87f731b97ad2da596e801
5924dc1099cb56b38db3387478747edc72e2503c
26455 F20110115_AACJEN horowitz_s_Page_123.pro
3b85e289f564584ba5c990c0b593d2ca
4bdc1c58508f8ef8a114a5b5154f16bc16f98f9d
87675 F20110115_AACJDZ horowitz_s_Page_015.jp2
f4096be84c4c812f21bdab023e810921
a1fb7840311d184387ec36ab5f796fa64e1fe550
22806 F20110115_AACIZH horowitz_s_Page_070thm.jpg
65d4647e6f249766753c3556a990b10f
1caf6c0692f789d538e6e3147acde21aad2ac6ae
1827 F20110115_AACIYT horowitz_s_Page_117.txt
f2a0ebafde10e56b7f639d22b66b2db2
d53342697baa1480219550ccfcade924a901b0c2
99653 F20110115_AACHVR horowitz_s_Page_170.jp2
c46e3f5e8259bfadcb25ee54333eed2a
0b43198ea8d97eeff9c65dc5eb11b765b0ee419f
94921 F20110115_AACICA horowitz_s_Page_068.jp2
ed168decf83180bb7e44353c3c0c0be6
cfd1eea8474527b88b40064d3829f6f7c895ffb5
196091 F20110115_AACIBL horowitz_s_Page_174.jpg
25851f966d83f65bd052f1722defc2b5
2181136000e9269896edc3d6115cbbe5f10cc033
F20110115_AACHWF horowitz_s_Page_162.tif
ff30b3563fb8f0f0b843ff2e116094a9
53724096b7ec0d5098f7775cf43abe5f85b21f5c
23373 F20110115_AACIAX horowitz_s_Page_105.pro
d4a0667cf400e78dff055009973637d1
a8af1fec688ea6ce173014d474740e15494eff39
33946 F20110115_AACJFD horowitz_s_Page_091.pro
0ff7b6d02d3e426307a2702d9ffa6552
df8be6be4ddeea92f3f801d165ae1e294e9ee1ee
4543 F20110115_AACJEO horowitz_s_Page_189.pro
c83f8c4444e4721f4317ca63143419a1
a2c419e96265a1dd2f90fb8ee4c672792d923e27
53657 F20110115_AACIZI horowitz_s_Page_096.QC.jpg
1cc9cc3a724154268af2aba9d9f91772
c25e270eb8810d73d99282a52d98e732da9ed487
101775 F20110115_AACIYU horowitz_s_Page_102.QC.jpg
b34d4eac5dd468d741d6601a81f29c57
ad9cd7f902270afeb3b3fa6fc0937d81b53f4e0d
891800 F20110115_AACHVS horowitz_s_Page_080.jp2
1299ff80fa9fe856b978fc602b00605f
94745a3b5760b4e52c4b25339f6b2a0d9df2b642
43731 F20110115_AACICB horowitz_s_Page_091thm.jpg
57dc24449c54ca05a7e7b1d408a2bd24
43c9c24fb20b1f0a15209054cb38586aa59ca61e
37572 F20110115_AACIBM horowitz_s_Page_193thm.jpg
bcbfc37d070cfc6559ccc3f460082f61
aea186a011d4f2072b58614a0fd8014391ba9ee3
49085 F20110115_AACHWG horowitz_s_Page_039.pro
8740854c2d9901ce4471ab31de481743
2cb2a1f19eb659d49fac21bdfca72a50589d0487
34561 F20110115_AACIAY horowitz_s_Page_134.pro
9e7745bba3928cb1964bebcea026c89d
bb000f4f831bd11c907e6d83bcc7116519404d1a
41784 F20110115_AACJFE horowitz_s_Page_060thm.jpg
9a2e704c2fb81c7e8c3645da37d7d305
9e0dbc82c2079d743203e200b737949b915a114b
75250 F20110115_AACJEP horowitz_s_Page_023.QC.jpg
5c6e8e790b3fddfdc18f6c65d7ae85e4
4e5ca6f33a5b8d1e0063bb636a91bb0702660d80
3551 F20110115_AACIZJ horowitz_s_Page_004.txt
3ed8da6b5739b27c40369907f7fbb59e
e51c6f46cee7de2666c66aad0862cacb388026e8
111533 F20110115_AACIYV horowitz_s_Page_036.jpg
bad91b915dbf9a6960ac57c889f10d51
86024e35f474e20eb1253337fb79361171c37e58
24823 F20110115_AACHVT horowitz_s_Page_023thm.jpg
0c8a98a5140ebe3943589cc2f677d883
b47e0740e79cd3dfeb807598716b956a6a26c362
F20110115_AACICC horowitz_s_Page_097.tif
2d4d251218d566daf2dec4ae35661cdd
055908fe418881880af44bb3d6951ed2cb21fbdf
217114 F20110115_AACIBN horowitz_s_Page_024.jpg
dd4e065ca9803a418a3aa57067694d3c
a5b9725cf87110e0234958aa7437d85852667af8
44634 F20110115_AACHWH horowitz_s_Page_019.pro
ac2996028cf3f374fa94287304a300e9
6c7c9de05e90799bcc0702342df7a35800778806
96380 F20110115_AACIAZ horowitz_s_Page_161.jp2
6cd4e95a95fc828328d15b226673aeeb
6e1fa851519e91e81ba1561443251b56adc01a85
77015 F20110115_AACJFF horowitz_s_Page_028.QC.jpg
9e3d7a292a5d597e36731f92387f2a3c
8664dea43301af9224ea7d6f849e078387580c5d
39364 F20110115_AACJEQ horowitz_s_Page_128.pro
8a0571a800c8719fd2fe1babb105084c
db02b771afc8af902de66e61e60eb208ef62897b
24281 F20110115_AACIZK horowitz_s_Page_029thm.jpg
effb4af960ca04309d3c82b423d1e5e4
30ef0fba8ae1c15c3deccc1b118f695baa3aa791
2561 F20110115_AACIYW horowitz_s_Page_197.txt
5d02069d7d5a49dea3980d2e06a871fb
951b550bbba69c16ec7478b1722fb31247ea897b
40279 F20110115_AACHVU horowitz_s_Page_015.pro
51c909c74f342cfe4879638ce3ac9fcb
cd06e3538691cea647ca3e4162e435ae08c86de0
2733 F20110115_AACICD horowitz_s_Page_010.txt
d64bcf34fceaaccdb592b56e2145d72e
758819e0316084db024fdb7f44bda9200ee1120b
77980 F20110115_AACIBO horowitz_s_Page_097.QC.jpg
ac5a1e419a47d112febd230dd52bdb4e
f248c5ea2b9cffcfa203c2160d30eb0750e6ced1
2777 F20110115_AACHWI horowitz_s_Page_191.pro
e0ab5183e16d0efb259172ecdeee69e0
fc8136c2c09f6a4e05dc61e8e6ba348d28cf071b
105809 F20110115_AACJFG horowitz_s_Page_024.jp2
a724de66ef9c508661ebb7598534d1e2
fbe2539380d4a75352f1fa8b03e1d055baf55bde
18095 F20110115_AACJER horowitz_s_Page_163thm.jpg
477e38d297725c7a4ebdd624867bff0a
6442995b1333fa73f52238abf0ededfb8cddd9dd
1666 F20110115_AACIZL horowitz_s_Page_174.txt
89bb0a82d1909f4f40c4bde1ad488e96
1f6a72696ee299f912d629741f41947c64ba7812
150841 F20110115_AACIYX horowitz_s_Page_186.jpg
2b362236c93a6eeb45aae141b6e8a720
193ff1fbc9ef30e709c716bb4171ebfda3f262a5
50600 F20110115_AACHVV horowitz_s_Page_072.QC.jpg
130f48bb1a752de8f875a12272e023d0
aa0fd6e56a2ec20b7dc469654b0fb0fd6b1d049f
73203 F20110115_AACICE horowitz_s_Page_035.QC.jpg
1dceece42c1aaece4c8e85ad9658b608
e7022311ab230bf03c57111d8c26638cfebade5c
71472 F20110115_AACIBP horowitz_s_Page_184.QC.jpg
7256551c2be1239f360b679d77f61f79
b71804894e98d324a3f460fe6c17dd031fe9d6fd
32347 F20110115_AACHWJ horowitz_s_Page_062.pro
ac3a332da25253824b88443d40f6988d
6d228cc534feca00ae564e23209f75f2a221d185
193957 F20110115_AACJFH horowitz_s_Page_135.jpg
f9f957a887181312422475d93fbd8bc8
f5f53f0f561d23c9dcf9a23381f3460862a7023e
16009 F20110115_AACJES horowitz_s_Page_030.jp2
d2e68df946c30f285b0aca149856594f
dcffa1db3844bae059d3e69358c02de481228d78
46070 F20110115_AACIZM horowitz_s_Page_147thm.jpg
71e241374a42a86f4e3863c3ee7541a2
27a775ab3be95011ef235bb1ca0e49480bc5707a
44190 F20110115_AACIYY horowitz_s_Page_021.pro
a569555d6f8b4fb9be25677319837fcf
746bf5300330db60d752beac1ad44a359fa08fab
39082 F20110115_AACHVW horowitz_s_Page_159thm.jpg
7fba59a93b2ddff728722f6bbe272269
5036407b9e124a826a1579c53cf9aca6a5175bb7
23342 F20110115_AACICF horowitz_s_Page_097.pro
83b27e6d1ad0672935a3e0afb7fc2fa4
e8be86a4655a050e6cf1a30255ac13564505f070
33772 F20110115_AACIBQ horowitz_s_Page_137.pro
5415c08d2f29ffa6b8f8270d0964602c
3e1d859a8c9b59861efe3253970fe50b7d2a58f3
200779 F20110115_AACHWK horowitz_s_Page_039.jpg
4e303942580c68d219106b029a12dbac
4c6fe3a8d33fa740cef009abe75f1da9b339f083
F20110115_AACJET horowitz_s_Page_121.txt
d0c4e436062f4b21468fc07d65e894f1
299e791280e3af48424cb8207157427bd93b92ac
42807 F20110115_AACIZN horowitz_s_Page_176thm.jpg
5724c83d1a5b06feb0144540061c565c
a8ca34c2a56d35e3c1b9755ac77a0ffee327a9d7
24467 F20110115_AACIYZ horowitz_s_Page_156.pro
a2525807f8666145f6017f50e653971c
8186761fbfc364a3cce6a4c7cc883fb3f96cd174
F20110115_AACHVX horowitz_s_Page_116.tif
a2c34bda55eb242d17d7971797b9c5a5
9f17beb52f664b01d7101b4b31e269fe4cc1c7c9
F20110115_AACIBR horowitz_s_Page_153.tif
3c9f1b0b1440421b847fb7e1d54ba47b
dd7533162784d1b1db508e1b845cf0a5e219ebee
44481 F20110115_AACHWL horowitz_s_Page_181thm.jpg
7b18917250182bb22192ca57d6e111ea
dca31b6ec2a190c4cc45e86a0948551baf8acf75
1676 F20110115_AACJFI horowitz_s_Page_124.txt
be56d7da01168eeee5c5dd2562f22f4c
00bb43320d21e659052ebc54967a3eecad020a0e
29423 F20110115_AACJEU horowitz_s_Page_155.pro
d356288d30ae32ff1b1bc49510b44147
077129da3d02754e1f92cddedd141771d783b881
F20110115_AACIZO horowitz_s_Page_040.tif
7f9fa8b68c4327f4a6c0950374aef020
a7a45e40270a65126e767b419a8fd75afc1797e4
77701 F20110115_AACHVY horowitz_s_Page_014.jp2
335726c1e9c0876bdbf57b5846ba6b30
584a5c82248da0c2ab7cd6194b47f6eb0445d833
F20110115_AACICG horowitz_s_Page_091.tif
996e3f0f4afbed991c9bd152f1d43e75
50366946a4885d3b7fbbf4aa54906bf4aaaf9934
1437 F20110115_AACHXA horowitz_s_Page_078.txt
5d9d229089860ee111a7be706cd1e367
fe9298ddef740696a341fd637994b7dd4a19f06c
F20110115_AACIBS horowitz_s_Page_096.tif
7349002fb40be7ec013cae6e46c24bf6
2ef18b8b1df2324529c4c66856fcf1cd1ec0325b
172592 F20110115_AACHWM horowitz_s_Page_080.jpg
8724e7faba222f5faf963eb6a5df526e
328d9d5402a17a2a74e9a631146f7de86bd8b5ad
F20110115_AACJFJ horowitz_s_Page_183.tif
9fa382bb337ab5a57d64c086cad7b379
6439bb952c75055f843086ad7fe2a3027fb8e756
F20110115_AACJEV horowitz_s_Page_046.tif
48f2fcb7b5d1a12f33ec9148e1ecc1ee
6d9c0128ae30c47de642b4561163666e75fa46f2
166037 F20110115_AACIZP horowitz_s_Page_128.jpg
3dda73b8865b5f49d2a1d98b964b8d36
15befa3a24e1f4fad264146ea7b636a889af2dad
84623 F20110115_AACHVZ horowitz_s_Page_037.jp2
263acf22e250053ed189da0a76ad74e1
ad2072766446ffc3e47898c982ad525ce4b67616
53642 F20110115_AACICH horowitz_s_Page_046.QC.jpg
c7b4f5b71a5389509722979ef19345d2
4fb70c1fa462b4e9bd075f00be4245d2102e5f40
43276 F20110115_AACHXB horowitz_s_Page_105thm.jpg
568d12ad142722151c5e553c133076e6
1c3556d499ae21fb034fc872c0231279b15f181b
107256 F20110115_AACIBT horowitz_s_Page_141.jp2
95aa60132e1375e4e8f3459874052688
5c6093075d7dc57555f1112dac7ce28d9a0536ea
86148 F20110115_AACHWN horowitz_s_Page_168.QC.jpg
1388a88cd08aa5cd5a19480d8bee8fad
8c3aa432a5bcd8c6f4a6bcd57d696817fbbea402
94547 F20110115_AACJFK horowitz_s_Page_167.jp2
6f07f5305d7f18616045f450f2a2ba0d
98a86f0f56bb8856e1b5d681397b2a25097d72e4
24967 F20110115_AACJEW horowitz_s_Page_065.pro
c12298fa7fd61a82f1cf2059d82e4bfa
7d154d6bbba85ec1ed0863ec9fc71261d5af9f18
F20110115_AACIZQ horowitz_s_Page_169.tif
232a955459f83d280331abac8f8e63c3
0c0934a1180a1fa7ae7b6e46a50d7e725e46ac8b
F20110115_AACICI horowitz_s_Page_055.tif
0f5abd94fb18e21500fbd5ad3b0b5b0e
251e95e771ec0a4e7d5c104f0568e8e18b149063
F20110115_AACHXC horowitz_s_Page_168.tif
9d835706601ca10830c518251d6ac1b0
98b77ca84e4656deb3d2985995b8dd6f59a4bcd8
207794 F20110115_AACIBU horowitz_s_Page_026.jpg
e77198877e40788d9337c62b69741c65
6be3dba20c039b82a12bfd2f4941a4692f310d31
F20110115_AACHWO horowitz_s_Page_138.tif
59304a91d0171f5490595841f7e49de0
957651935cbb5285c688dbc622c833969832eb9b
1724 F20110115_AACJGA horowitz_s_Page_154.txt
3f7e032b648c73b32087dc34284e09f8
b416bb59470f3e52bbebc0f944061aeaf904bb04
72986 F20110115_AACJFL horowitz_s_Page_025.QC.jpg
97fdd1a7ab4c9371f832b960626421e4
99c653ceca0ede5085faf464cb3607ebc5e7baa2
F20110115_AACJEX horowitz_s_Page_126.tif
daab8e96acafe52e161db3ef588582b7
e31cda43f840a56327843a3440099a88a880115b
2373 F20110115_AACIZR horowitz_s_Page_132.txt
7afc330b9501afee8a4896f0f2575541
7ee8ee092a78e4fad7b22a230ec61f9959e748e9
180601 F20110115_AACICJ horowitz_s_Page_031.jpg
f80519b5feb6c024d156337d037ec69a
7e384300e23ac0e803983e5601b79249e489381d
80546 F20110115_AACHXD horowitz_s_Page_093.QC.jpg
e2be67b8b6e4c74550419f165fdc5cec
67fd4fcb78281d3faa406a64e0d5c06b4605cea4
409 F20110115_AACIBV horowitz_s_Page_001.txt
4926355dbd8d1d7ada7f9cb814f83297
1180f750692224bbd1e046f984c0a3970a22d005
22033 F20110115_AACHWP horowitz_s_Page_115thm.jpg
3ad8b3fd622e5a203ff97f3c9c1eb17a
96f0dd2253a8a2985d8724a7cf2b6c6383e95899
2655 F20110115_AACJGB horowitz_s_Page_201.txt
376ca959d60605435b663ed347ae0a96
c2c76d29ea69346c843e080d3ac881a3ce4f5645
929764 F20110115_AACJFM horowitz_s_Page_081.jp2
2518aeb0c26dc8c041de04b327534bd2
1df04c37b0a9681fef86824442305f5f2949a64f
45013 F20110115_AACJEY horowitz_s_Page_093thm.jpg
19215608a53fe3ac294ddd7f03e74b54
0e060d3d4a2118f53124e380873975c69451bbf3
200 F20110115_AACIZS horowitz_s_Page_193.txt
254877bb15f4ce107250014cc6f4e1b0
0171705f63b9c9580db79cec0fa905ba7c782b8c
40864 F20110115_AACICK horowitz_s_Page_174.pro
b8ca373c1734114a123f178c6d68e3d6
9eb65dd3ca49657fa598f703f2aabf2432a42c95
1358 F20110115_AACHXE horowitz_s_Page_125.txt
0211e9ef3344a2526c63c0f4f64b70ee
ea392a0268626f848dccc74cd8c9612d298ea9c4
69522 F20110115_AACIBW horowitz_s_Page_055.QC.jpg
f7d85e4c4488df51c7e2c4e18a76f2d2
f8cd6f5861ffbd63c6efed13523f7b7fb02a426c
39719 F20110115_AACHWQ horowitz_s_Page_082.pro
b1f8054dd509cac108ec64d8c3b3be0c
bdbcd2b37b716093b266cc467f8583b70bac8976
F20110115_AACJGC horowitz_s_Page_005.tif
903383165a0c9e326731e00971b17287
2e65c27cb6130be47134a61071b80507fcb706f9
239094 F20110115_AACJFN horowitz_s_Page_098.jpg
1d134208e7438ce1b658fb64e176d058
2b288aefb1ed631edaf8d8b8e073ce7594159f8e
14565 F20110115_AACJEZ horowitz_s_Page_166thm.jpg
2bcb1b645c82eb9383291f997bacc18d
88ae719e3b12eb686383ac5a1d19d07ba84e4368
45191 F20110115_AACIZT horowitz_s_Page_131thm.jpg
fd39de6dcbbb206adfd7c4cf2b4941d4
b428e0adfb831083b542f5a68b92506431d0a79a
20147 F20110115_AACICL horowitz_s_Page_100thm.jpg
2080036cd36df167c39f04160cd4c55e
9afb31c1d575634453e5d62d9e8f7a09d522718a
51030 F20110115_AACHXF horowitz_s_Page_185.pro
e3e2e6fd6e0bc4f15666c4ba74999777
b86267458978b73ca8e65d4d0a23b5730eef1c65
75526 F20110115_AACIBX horowitz_s_Page_117.QC.jpg
a1df3bf4b18ba93551f14305916ff74c
b2c04db0fa48da2b09b6ea6034559a4cbb290a49
16570 F20110115_AACHWR horowitz_s_Page_036thm.jpg
cabf03978dc40ed6a15efbe3e14f100c
ce933b08a138f2ce3b6129ead5d6558cda4e9896
1687 F20110115_AACIDA horowitz_s_Page_069.txt
4a6ed6384332e1ee644f2c9c8a5f21b1
8a7d6d82de980987e1ad1a496200f72ee00fbdc0
63851 F20110115_AACJGD horowitz_s_Page_196.pro
c11fc5a867d9e21c9f5fbc498498d249
5a05242c50b3c87c16aaf97e389575272ccb2210
43421 F20110115_AACJFO horowitz_s_Page_145thm.jpg
f302dbb3caad219c1960bf9620767ef0
5c630043e9cdadbb21d17b735e429799554648d1
27679 F20110115_AACIZU horowitz_s_Page_179.pro
576c6f9f870eb0e8cc3ef1e32ba92663
1b6cd8b9367fc79db5976b4740f20236a7cc2816
21505 F20110115_AACICM horowitz_s_Page_031thm.jpg
ef5fffd6c4141f96a5f99361ae0be37a
dfc4437c0fec788ddb9f9a8429c0cbe1a7b0232f
31162 F20110115_AACHXG horowitz_s_Page_096.pro
99bd409106aa75519d3680d76b0c73cd
77778098be0dd6798448a0fef3e42fd4c40eb47a
49649 F20110115_AACIBY horowitz_s_Page_141.pro
b40f6dc81c8c8b03b4ff9d6c45fbc951
e5474afc2965a2f66b39f120659901b1b6486a73
24953 F20110115_AACHWS horowitz_s_Page_041thm.jpg
2d766b596dcc311ae477f622c501b271
d6ad82f8f762e1ef86142bfb65ae98fbc581cc7b
46945 F20110115_AACIDB horowitz_s_Page_143thm.jpg
9e48ea9cd5f33999c4dd460352204080
656b5be0f1c2ab05462953d73e8e89d4b4ca993d
704666 F20110115_AACJGE horowitz_s_Page_018.jp2
2d73278135d9b41154c86ab438f696bb
c85932baabc47cf9e20cb4650624201fb29736a9
140712 F20110115_AACJFP horowitz_s_Page_063.jpg
36a361abcf2892e44e5d5c408f95f5a8
74d8f7ef92906c20519a0ecebea2995614a53d71
81407 F20110115_AACIZV horowitz_s_Page_047.jp2
92b5fdf5be043a4ba5806631846bf5a2
9d6f5eecda107c7425bf7d2edad29a5cfeeeee22
78743 F20110115_AACICN horowitz_s_Page_087.QC.jpg
27308e5cc8d52c787f8be13da6c502ab
332fed8d9f6f79a49380d296dfbdc8168afc40a9
378985 F20110115_AACHXH horowitz_s_Page_011.jpg
e43f370bd81e9fe318130a55dcbfa804
3b6fe0bbce00b3d286d745cba79cb98e83a0eca2
F20110115_AACIBZ horowitz_s_Page_191.tif
df6e08692e9ab8dd1f90db1ddd77e142
c9df4fbbb9b06b5bbbb5e29219811fc627fb7c8c
212538 F20110115_AACHWT horowitz_s_Page_141.jpg
e023ac00bedc1892957d20195036e9aa
9abca7280b8677f5630fd9de826704ba3758db35
F20110115_AACIDC horowitz_s_Page_182.tif
0bbb0ff96d5a2d076eac1e04f987be4b
3dbc736b4cfbaa364bbc54c1d960c76dd1657a1f
81312 F20110115_AACJGF horowitz_s_Page_119.jp2
6a3a3af33ee4c809dc3da1d59d075d8b
f1c0421a01248963dc81fce15bf2929d09ae5eb2
89895 F20110115_AACJFQ horowitz_s_Page_151.QC.jpg
9856a68287af4f50eed771cec55a3be2
3e6d87ce769cda835cda784da7af19c61ff0391d
12294 F20110115_AACIZW horowitz_s_Page_208thm.jpg
d7303eaa0d25c307b5ed246fe2c710af
aa9b5fd8ebfb0621f1aa868d25a2f258c091302c
77227 F20110115_AACICO horowitz_s_Page_026.QC.jpg
f95dc0cf5ffaab1b5a25277a7dae575b
34cc879d536ac7c7730b253a1df40a9ac423d291
88665 F20110115_AACHXI horowitz_s_Page_081.QC.jpg
1b151e417ecda8db95fa33bab0c66642
2d7581e1f5a771c01fcafae83403ba3586926962
967 F20110115_AACHWU horowitz_s_Page_103.txt
1eb80c04d00cd85150d5db1371de26a9
bc067a290a042d39bb8477b3cc20bbfae8298614
F20110115_AACIDD horowitz_s_Page_122.tif
07e276427b8740116a5e78c3a737dd54
6ec6c15956e658f335813987e6ca712cd4d4a1be
773398 F20110115_AACJGG horowitz_s_Page_071.jp2
24264e05d44cb71c685f6f9ec9042fc8
3cab963579aaa3500c168c353daa1db2d6f3fa7e
198294 F20110115_AACJFR horowitz_s_Page_130.jpg
fe830fc616316eac00a7bd515f5c101d
4643052341e2c499eb10aa3ee8f11615aa6fd8f4
124685 F20110115_AACIZX horowitz_s_Page_198.jp2
9c6847177ad9550ef117f87ee1ce210e
5ef259b1f3078e0f9a7fd2ae4a4073af7ac1c12a
44619 F20110115_AACICP horowitz_s_Page_059.QC.jpg
74c9778a3d279c457c9b4ba6b31e2999
7dc06c9f5f8bbd8b6c90bcc1104ce555f3f0d18b
46500 F20110115_AACHXJ horowitz_s_Page_146thm.jpg
7157af451e9fde84c65c3e77fa352aa8
3ebdea0d7223ca2681ecdb54d5379a8fbd8c3dbd
F20110115_AACHWV horowitz_s_Page_029.tif
27b832932c0b1a19dac7705a4b353205
78244df7d634f536a7a0fdce8388b8f7a500320f
2019 F20110115_AACIDE horowitz_s_Page_110.txt
2cdeb2c510d2af23b785b6deec3281cf
ad5c09e511bdbb8c8ce7b024397cfed0972328d2
54865 F20110115_AACJGH horowitz_s_Page_014.QC.jpg
13034d4b3f30bda0ef8b8d3d7b18e996
2b6f43b6e3f1f0e256146bd9509d209c0034b9b7
1963 F20110115_AACJFS horowitz_s_Page_028.txt
08ffde745480bd7d08d884f21e8d89a5
ae3ebe2b21c9d2081d7ff0097e4f1f05bcd9e1bb
921065 F20110115_AACIZY horowitz_s_Page_178.jp2
6ead1392ce9095abbf12fa6858186045
3a01d1cbc3aff579e456bf636dff8d4bc4a02775
214305 F20110115_AACICQ horowitz_s_Page_146.jpg
32b6ad26fa29e95ab50e44faff4269d4
f62c29a934d32f8c9a430ebf68b4058192859efd
76912 F20110115_AACHXK horowitz_s_Page_049.QC.jpg
c49975639abe9a13d8014dc97ae2533d
64919a1d1e5665e360bd96f3667fc40d304cafa3
45234 F20110115_AACHWW horowitz_s_Page_055.pro
f1fb54ff0e62f7c53e79f6edf7be0a9d
0b9e0b7ad0d6ca26d0ecd0c52a368679c478877d
F20110115_AACIDF horowitz_s_Page_184.tif
33c8e45b3d6857af0d3d1626f58fae21
b4e9d2bb3fdf594c7a7264f5b817ad49a9fe09bb
218 F20110115_AACJGI horowitz_s_Page_003.txt
801db08954e35f48f37ce759fbefc433
0725d4afe66ed4551c2ddca4ae1f4107fe6b3829
1051955 F20110115_AACJFT horowitz_s_Page_175.jp2
45250b21cf552700381bc013d7bd2644
73fe4519c599e14722b5bf0406a5fcd2a6fc8360
1366 F20110115_AACIZZ horowitz_s_Page_165.txt
159f3dbb7de86e47800288020f299a74
696c526985389cf8c9b107e8a6a85bf677982711
79213 F20110115_AACICR horowitz_s_Page_040.QC.jpg
324bddbfc31a0c5ae7721ebfa2c081ca
726842009f1f9c1efb15537f51eda4ecf8c08dff
76454 F20110115_AACHXL horowitz_s_Page_013.QC.jpg
de07735fa7a1531df1de8bae4ce19981
6f8e3d9e038896ca1b6107b13f9896ae055139e0
684387 F20110115_AACHWX horowitz_s_Page_150.jp2
03b4b6a2e6a25bbfe10610af36fe48a6
69a3546a13ee4ab06f880dc118c06eeb02c53c4a
224702 F20110115_AACIDG horowitz_s_Page_104.jpg
ed1d18fe2bf90453df3a77cb8110851c
30a810867d5873a6f8731b7e166fa3cc5fb96218
53002 F20110115_AACJFU horowitz_s_Page_027.pro
308902aa8100b51e383067d4a66a031e
da51c1bbaf7c1ed2dc273e6f461005edee6896cd
174823 F20110115_AACHYA horowitz_s_Page_137.jpg
979a919cec80c65733269e20920271a8
17acd6f9c03e7d8784baaddf930ce65242bff629
170790 F20110115_AACICS horowitz_s_Page_050.jpg
96b6d2452bce4319415b6adf6d279306
ab8cc4f549a6856ba0854326e8d67e3bd1894ca3
1784 F20110115_AACHXM horowitz_s_Page_114.txt
8b85e9e0a26a47806a2922caf964dec4
b276851505a5f6dee373e9efc8805aa9c83ae0d9
180989 F20110115_AACHWY horowitz_s_Page_149.jpg
c298de2c5fe9d074d303496053361308
fd892376dd80a047c930058db4e2edf87185b747
31843 F20110115_AACJGJ horowitz_s_Page_143.pro
2ebbb33ec823357e2777c7cc1151b442
987d05446a24394d565c65e2f1eb922760515b86
2590 F20110115_AACJFV horowitz_s_Page_204.txt
7a931d68db1e87c254d8ec9a1085e362
2e0d5492605613989618db019a68566aaf064c60
706334 F20110115_AACHYB horowitz_s_Page_086.jp2
a7cf6493f7431f0cbf1a62c8aefcde9f
2e560bde2513e2ac224c36eec14de6a4c8229590
43032 F20110115_AACICT horowitz_s_Page_018thm.jpg
3445cada17bacb12b175fc22e195e3d8
c08ca1999dbb38365de2fe1306fe969cb27be398
F20110115_AACHXN horowitz_s_Page_120.tif
1fd348c3745294772ecc9c85fdd1eb0f
7586e24092a8724ea73211028c48ceabbf7e9968
43176 F20110115_AACHWZ horowitz_s_Page_086thm.jpg
bdc3d2932b01f6462599517944faea88
a13a31473801361f721578a9ee60b808e41ff9f4
8702 F20110115_AACIDH horowitz_s_Page_171thm.jpg
619496b9547da6da66c2f5b328d04aba
2df8e67b85699bb8fcad8f72e454662d2e9d848b
F20110115_AACJGK horowitz_s_Page_032.tif
3c42834c26d8901334eddab8828182bc
eada474cb8551b88ce4d1fef441cd13433726082
1830 F20110115_AACJFW horowitz_s_Page_025.txt
1dbf4ad35ddfccbb671796ee8e3a5c7e
33acf04291586ed63adf5c72e01f2c5c75008809
F20110115_AACHYC horowitz_s_Page_144.tif
a13548a8e47f8ff4fa937755365e5393
5d7509bbab3146d44bcfdb17a33eaa88a335c439
174131 F20110115_AACICU horowitz_s_Page_013.jpg
d5d7347813d9399866bc947888187327
2a2402de6b5a631233352a93eea9aaea0fc0ccd2
64770 F20110115_AACHXO horowitz_s_Page_101.QC.jpg
fa7625981e58e1aea982cae367b8fb7f
7517eb04ecfbec175de8c79015630477ada04115
F20110115_AACIDI horowitz_s_Page_140.tif
d74c55d52e98a518dc8adc20b5a7728f
fa51e0d3a74038b5daf79aa2504e1928aac5ec77
39846 F20110115_AACJHA horowitz_s_Page_007.pro
8fd588ef055191613334b54f163fa5bf
ce7041ec0f6da39a69152bfa0cced9d6a677d666
46268 F20110115_AACJGL horowitz_s_Page_098.pro
8e9668b215adee7502e7f5b76c3e7828
010448e1fd6bcf8d07618805b7a09a1311b771a0
39925 F20110115_AACJFX horowitz_s_Page_184.pro
c47b95f9479d125760738927d2215174
c241c2486ae75c8001386d379908f700f4d9b94f
F20110115_AACHYD horowitz_s_Page_019.tif
a29ebb4e9732aba08f6359c991b13995
5ed7096351e1c8b3e6d5b87c9bf6417507b57613
18992 F20110115_AACICV horowitz_s_Page_194thm.jpg
346f90b229883e97ad4af1deaf6f6598
87ea9f75768985b615ee6fab8a57f89d0ecdb5ae
2082 F20110115_AACHXP horowitz_s_Page_027.txt
585f0dace777b8237d721e9b37b2fbf2
bc16258183b8d0bd80d12c697a19dab493dcc905
1593 F20110115_AACIDJ horowitz_s_Page_126.txt
d600120b0dbdcd284615fb55b52831b9
9b1fbe8292dd6fa2e1ffc657b712b8f2a3f4c8ac
17279 F20110115_AACJHB horowitz_s_Page_136.pro
656c6424b9098da399649547773e8b56
d2652065d22b6921dd236970e24888a1fb6c71e2
F20110115_AACJGM horowitz_s_Page_134.tif
8ae3b3d3db87a7e36ad7717a39eb1702
92751d9e8c0404e343df613bcf3cb690b1748199
719878 F20110115_AACJFY horowitz_s_Page_187.jp2
4463c4655a58dbd14803289043fef684
4ab728eefbff49a120a26bc21cfc9d9384d96ad5
F20110115_AACHYE horowitz_s_Page_004.tif
6c842f0d3f6fa074518d7ce6a72c880f
9809e627bf272b5a990901945337e9f87cb26777
202636 F20110115_AACICW horowitz_s_Page_033.jpg
205f4947f10dd6a6d861cc8df3afcea9
8b613e2b2f8f7ebd068b41c9437e900da99addcf
1059 F20110115_AACHXQ horowitz_s_Page_172.txt
1c4c8a485c806d4ad8c709713430addf
b11264f5c9c4994087794a57e7a5a93b36d1094e
87894 F20110115_AACIDK horowitz_s_Page_132.QC.jpg
7a69a9af23fe14f9c31d33e48b0ea5a4
16e2a5b5aa5792512f3780bbbc275e97d1432939
60274 F20110115_AACJHC horowitz_s_Page_127.pro
5630b6eeb35b3f7775bd0466f6defc5d
b03569d75df80a6362367c92dedaea36424d3c24
32197 F20110115_AACJGN horowitz_s_Page_081.pro
90676b94b9e3b2d87ee76826f2504e56
4049682824d311721b2cfe76fdfda438c4dba938
66381 F20110115_AACJFZ horowitz_s_Page_167.QC.jpg
58d9fc26abfe4a1a9451229ffc4e3b90
84becbdc64988523bea041e6a5a9b451b9cae71c
69493 F20110115_AACHYF horowitz_s_Page_021.QC.jpg
016b35a466485fda658a13bdabe1a3bd
2cfed2d67d86d967385d66782eefeba02f3b17f6
79221 F20110115_AACICX horowitz_s_Page_038.QC.jpg
7e249d8ddbc3af3b0c886890d209e3c3
acd9d25fb2c49178e0b5d026d203941093725727
1753 F20110115_AACHXR horowitz_s_Page_047.txt
66ee4db0351aa9ff54dccf6b407584cc
bd8dc0a08ec84d068ec55552852e21cf77a97aed
F20110115_AACIEA horowitz_s_Page_157.tif
d99204b85975f721b09491aaddf57cfe
23d483f1c6f14412dcaa4c0b31e1198e1ad92e19
2411 F20110115_AACIDL horowitz_s_Page_199.txt
3b98bc5b0dcc518eeb40b3eb0750b441
d5b1068c97c278e1ca4266d2416f8dbc108d1b96
34253 F20110115_AACJHD horowitz_s_Page_014.pro
6519e08277a12b8721682105ac99860b
124e8e3aa70f6b9fd3a0345a3f61e463955250ce
615208 F20110115_AACJGO horowitz_s_Page_136.jp2
e82423d9222b189288054f208139312e
e2848c5f11f75ae5a88f6ea58662b1a2cd86faba
23902 F20110115_AACHYG horowitz_s_Page_195thm.jpg
1ec594560ea3b23412e5f700d3efbe45
61fd80aeb25d4a2ca09dff37a827801d84fe313f
24353 F20110115_AACICY horowitz_s_Page_201thm.jpg
806e6f209be5ffa6258e3a10b96387a9
aa5fa3d1a4141480e91e58e60d21e6010f468ecb
32390 F20110115_AACHXS horowitz_s_Page_066.QC.jpg
14e0c0659c31e2f6f7e5a4273e446e98
e306f189b89c9c310613b5b1430d209582956f05
23913 F20110115_AACIEB horowitz_s_Page_172.pro
e83d8cd97ee0622d2f0aa0d080cf91e1
c460b7cb5f88c203372022842676991cd3c37057
198355 F20110115_AACIDM horowitz_s_Page_175.jpg
0e6e7d8fbd178a52d33c6d4a596ab36e
9a984d4a2710ec3caf69fbe98ea75633837c034a
56243 F20110115_AACJHE horowitz_s_Page_095.QC.jpg
4df8b05fe97d019836c41d38dab3d08f
dc450bce70f12cc217d220e41ba86854676f70f5
536766 F20110115_AACJGP horowitz_s_Page_054.jp2
266ef7b9d29754a1c539bf260f2e83b4
aad4b8562efcb6d9adaa5c1ec4342b46d06df1b0
262937 F20110115_AACHYH horowitz_s_Page_195.jpg
c4e1fb593f3adc2744405e9666744770
30bbdd5796fd3ba34b70ea1d641ed2b74787c63b
2138 F20110115_AACICZ horowitz_s_Page_151.txt
b581256abf8db81bdfbaf376ba0885a3
1eca25c23714f2b386ae360e875cd2d0f81e91ce
110215 F20110115_AACHXT horowitz_s_Page_034.jp2
6cdea40b667ac686bc7f4028cefcda25
6cf268bf4649c048bc5b461bb0c75155235ca1f8
93719 F20110115_AACIEC horowitz_s_Page_090.QC.jpg
1a919a8a096de5ea31b56c04b101a27c
d35d2e67e3d3438be50348cbf498c17ad4eff87a
F20110115_AACIDN horowitz_s_Page_005.jp2
4e7bf4248dc322083e74e82f6ec9e4e0
d31136bc412f361689b00ead35ae269f47b270e6
94600 F20110115_AACJHF horowitz_s_Page_115.jp2
40d5a37c6c37c6193abbe6ff5d420c47
a33f52fe8d8b211242351b4d448a51b0fb9eb11a
1722 F20110115_AACJGQ horowitz_s_Page_074.txt
60068bc77b72d470049a3f094fa02cb1
51299f58e433c07bccb5348804e51db018cb66ba
203302 F20110115_AACHYI horowitz_s_Page_029.jpg
a7a24d8a49e3948536d388eabc404224
5996c71fcf91fc42a2d7d1e8b1a0f5490ad8fe72
91349 F20110115_AACHXU horowitz_s_Page_070.jp2
0b4464f321630f8978e2b04f128b252c
e2fd086b2b071528a9784febf27e79118da62eb4
F20110115_AACIED horowitz_s_Page_188.tif
3032d649c9cba91595e9368470e0cdcd
5b9202da52bec670de59d0dd1381e97cb2a5c5ba
75992 F20110115_AACIDO horowitz_s_Page_111.jp2
00c9bd48003813dde8a38e69caaf7c87
87b92739042badb83486c056458ad1c4108448db
2531 F20110115_AACJHG horowitz_s_Page_206.txt
fbd6146833bfbed998d0251a8c700a53
43afb4c1b6287e84ee0ed6b02a154e06748bd021
F20110115_AACJGR horowitz_s_Page_054.tif
93c75a18d821e3d2806cb5d41ec1b774
ca5b8c2640ff9395ea0125959618f421dbfbef8c
142705 F20110115_AACHYJ horowitz_s_Page_201.jp2
82828312dec43d1582cdbde15560874f
c1ee0544b7feab6e6eee357301f69b760d57b843
72072 F20110115_AACHXV horowitz_s_Page_121.jp2
542147374cbedd62b77e187c53270f79
5735527e2a50fe6cd4f63f858b1bdb1674eac8f4
71635 F20110115_AACIEE horowitz_s_Page_164.QC.jpg
8b9cea0a7c722da683b46f3fa74dfa07
5e228687c13d5020a95f4ccfbc176994edd350c0
31403 F20110115_AACIDP horowitz_s_Page_121.pro
e91f28930f057d2018b73a4298f57b12
467364f59e9e52127704873713561fa23b75b311
1202 F20110115_AACJHH horowitz_s_Page_145.txt
9ee3f7f5129b1197ad468ff28d4b060a
6e4e7848307ed8c946613fd0a4f367f833f572c3
39865 F20110115_AACJGS horowitz_s_Page_184thm.jpg
97b02276de473ef7212ea258392c3aec
22cce062e4f63deb0fb84aa41f9e09e96c998c47
185054 F20110115_AACHYK horowitz_s_Page_042.jpg
9e3f7fdad611081d45fe9f13eb8ef711
fca12c313a89208ae125223ed51e6f56963b7488
110265 F20110115_AACHXW horowitz_s_Page_138.jp2
81e4a321a6d5e24af32f60eabc07108e
64d81dd529bd38870d189acd28ddfdbaccf5b867
22022 F20110115_AACIEF horowitz_s_Page_021thm.jpg
0da8f36e93fb72417c12007a36dfb9ef
188b7b6314ae1940086acc0bf5f4683af8eb673b
71461 F20110115_AACIDQ horowitz_s_Page_068.QC.jpg
adc808fe8760bc034ab187a27da6d0e3
a8282e77f3c19df2271a011adc6d6898f1825fa4
18472 F20110115_AACJHI horowitz_s_Page_046thm.jpg
ccce87a7c6ef7d2c2ed945547fcacb2c
17e45f982085a479a3421dfe5b10a4f61ce0eba0
82446 F20110115_AACJGT horowitz_s_Page_084.QC.jpg
51764b0f4f5702f5cf9ad59b5c488056
5f67ab7d1cdccd6c291a738def9a962a601c0f5b
F20110115_AACHYL horowitz_s_Page_088.tif
098448f6b1b52b4f5798af86a7a9a3d9
4a84bdf68b6d663da425c43b84095c654f832fa4
22861 F20110115_AACHXX horowitz_s_Page_183.pro
08213fd3ef1df30f2f3d27f06dd22df9
221153f586715fb36c36b8a4f28718db9eb246f7
158252 F20110115_AACIEG horowitz_s_Page_079.jpg
cc95aa039556d3bac346b499e5d1287e
fa98522c90f68b855819349b57ce2c8b9d9e3e45
24728 F20110115_AACIDR horowitz_s_Page_150.pro
3fa4319b52599352b35743f4961f521e
d0861f87d39f6097f28d557b7e16917a59cedcca
1359 F20110115_AACJHJ horowitz_s_Page_002.pro
cff18d022bff6efa7bb27adbf4599cec
b99902ffdb3913e4ae3623a84c0b75e40b469b39
45068 F20110115_AACJGU horowitz_s_Page_100.pro
bf34dcb63dff6e33b7d09e28b00e6a34
ad2b234b57adff6bef171f340253da16ac48f55a
120904 F20110115_AACHXY horowitz_s_Page_059.jpg
9657c6e3fd986b34f9e2a96f26337093
8aaaa8f14e50b3428f885dfa7ed47cd3842758b0
969486 F20110115_AACIEH horowitz_s_Page_053.jp2
01c44255b518ea74bd975993e8fbc82c
a7a0d99b8a4e50d1b2202174af36478ee7426217
1409 F20110115_AACHZA horowitz_s_Page_147.txt
3e7155dda1fd4499d7835233db94b8d8
7fdb218c2de56147dba071a2212683a19674108b
77092 F20110115_AACIDS horowitz_s_Page_041.QC.jpg
32c10f6874ebb02340186f7a42af15df
16daf414f39299667f2b105166f5272cc781484c
1316 F20110115_AACHYM horowitz_s_Page_113.txt
0156f827fe79b72702f3833a31da8f98
506fed09a1739e718d36881214b75838de7689ea
2077 F20110115_AACJGV horowitz_s_Page_070.txt
b47f89368527b2630340535e3e035214
fca2137a11788b2c369e955e2be6a5db84a4fd96
55553 F20110115_AACHXZ horowitz_s_Page_010thm.jpg
13271f58dc82d6d6ee147384d5bc7ead
e8ec5616940dd0b12669ac59220a5661172ee475
48266 F20110115_AACHZB horowitz_s_Page_053thm.jpg
292748f147dd644cef88205299d2db4c
8bea06ff9edd48a4f65ff2021865f6a663d62c76
98043 F20110115_AACIDT horowitz_s_Page_114.jp2
88c49825ef183e5090a4ba0d82113d28
e443526d8be3fbb12ecc80f5ae062901bbb1bfbd
71376 F20110115_AACHYN horowitz_s_Page_063.jp2
8615233f892d96b417d29dd074f971d1
26d6335e8d54b01135ef2e4b0ab61d3c1fc8d943
1592 F20110115_AACJHK horowitz_s_Page_144.txt
4a5955af7cfc8bd23f4aad8a6a6ad1c1
c79c023b91f0f6a541a97f5c51397839cf3107d2
44049 F20110115_AACJGW horowitz_s_Page_139thm.jpg
8582426e8ebfb299a3c9770c15878c17
20e2dc17d03711f9ef819a6217e90e02ca378d5b
119886 F20110115_AACIEI horowitz_s_Page_207.jp2
5365af26c20d626fb1d33f1ad7e8abce
2752d42d11c9e527224f4d96e8c9af17da44779c
F20110115_AACHZC horowitz_s_Page_133.tif
297b5801b3d549949eaa05169c899898
5b75d0b40dfeba64f40043428b1a1f2d0a6cd3fb
2563 F20110115_AACIDU horowitz_s_Page_195.txt
180a2b70ec09428dae0af124a81a45f8
fe0eaa3ece476959cb01c86ff66b7426512eaeaf
68185 F20110115_AACHYO horowitz_s_Page_161.QC.jpg
79bd322c63c5d715776edc3f6dce446c
87d68abce858bcb8663e718e07e07c1bfda138f0
165007 F20110115_AACJIA horowitz_s_Page_037.jpg
1e0d4c99b9c1d38e81e253e28cc42908
c5b1c4bc1515d864709ce7518883b6f742f9644d
61769 F20110115_AACJHL horowitz_s_Page_206.pro
69fe31e168408f92fad9a58014b3bd77
ca923aafe01e4b61d5b6347d9656e963a70e69ee
24984 F20110115_AACJGX horowitz_s_Page_034thm.jpg
8212c5b85c312f89f901120132ee5af2
56ccfbc935b90feeb500da6224b972443185853b
1998 F20110115_AACIEJ horowitz_s_Page_152.txt
12404f483c85e22d5d85d4fef2bbd08c
2711e4ac25d937df870ac6be8aca5edc29ebf051
32125 F20110115_AACHZD horowitz_s_Page_049.pro
7884e3a619d87bf92ad9456814b3cd17
38cd85b4e98bc0296a580ceaa10f6ba8d4d60a73
39990 F20110115_AACIDV horowitz_s_Page_037.pro
6bfa2350926b1564c8c8c5ca5369901b
f9007930e19853f061ada9ac70a5cdbf8b9b940d
F20110115_AACHYP horowitz_s_Page_083.tif
76b0fbef6c3388c52320b080874ddaea
62c74ea3081987bfc8bfda8f190d4cc4548d5471
F20110115_AACJIB horowitz_s_Page_089.tif
5b85dceb44ceb8b2a8f1460b71603bd3
0a8b5a03960689de9492af4b485418c72e79e307
43998 F20110115_AACJHM horowitz_s_Page_182thm.jpg
6e858ca3b1578e59ac57488add7a754e
0b7bb2dd9f2ed45664dca61fbd2dd9bf08619e70
1993 F20110115_AACJGY horowitz_s_Page_193.pro
d1ff2da6fac21dcea2f97c55b4ef61f9
598dc8c5c7ef0174dd50044ac45aa8cd2b941df1
376 F20110115_AACIEK horowitz_s_Page_187.txt
eb25c0d8e0aba7b8a0a02db41ae5e3a5
a646f0a9b4bdefa777e39fa0fa18f539e23aa709
33440 F20110115_AACHZE horowitz_s_Page_051.pro
652d670779839ed4e8a00b7d193f36d9
eaf35da677fa57baf1ce64f9e52b5cde5c5c6cb1
80356 F20110115_AACIDW horowitz_s_Page_199.QC.jpg
662de88e860d67b693a1cffdb2e9c11e
f2086b2d0a8faf59ea6fa2f926d2a9da515ad53b
1900 F20110115_AACHYQ horowitz_s_Page_092.txt
ed6d6cec21ab2b0e6d9db97234df2433
0d3a3ad4f737e14b1c2dd75a8ebbefddf176d576
17579 F20110115_AACJIC horowitz_s_Page_096thm.jpg
ab6abf75dc7d8b5a75274c6e9f69d43c
941e29af4df9ba78abcbea6cab366a780e24e850
2643 F20110115_AACJHN horowitz_s_Page_196.txt
c2fb130f82d029cf1e7c82915ff2425e
9c9f0bf73d04939ac0839b6e528405b8a9468c03
23386 F20110115_AACJGZ horowitz_s_Page_003.jpg
c2aab5280defd14e6a14e01b6a0a9767
ad394dc4b7e2bfa69abec78f89ac2be6d8a462a6
F20110115_AACIFA horowitz_s_Page_078.tif
4c2057fd6a4b19a8f8cc2b7a2968ab23
3a9d194c21e5e8163236cafca8ba27d5d2bd3cde
F20110115_AACIEL horowitz_s_Page_179.tif
040bc80393fdfac7b0c3503e741129db
eb69e8acf304a813d65b6dc00391b80b39c19737
1851 F20110115_AACHZF horowitz_s_Page_052.txt
f14a3de10c7561a913a100249e0ef904
b1e0a91478daf71e0f4dd99d9b2fc624831cc906
24092 F20110115_AACIDX horowitz_s_Page_022thm.jpg
09cfc2a7783c51bb45ce5e0e9bfc86de
c2823d530d5d5d1eb668bf81a05722668228f808
F20110115_AACHYR horowitz_s_Page_196.tif
87773fe23cfef472d7864b73474582f9
c4315bdfacfc08fce9dfdd9086c446fa755d32ec
139363 F20110115_AACJID horowitz_s_Page_067.jpg
1016461a2b3cfe93ca81d0d1a62251dc
06f0032041148ee0c76c730f82e2306acff7c939
F20110115_AACJHO horowitz_s_Page_102.tif
3d90182b81d129b482392a4ba00540d7
810f5c2ab7de8b4545e427e0bcc4e57f651ba1e4
367671 F20110115_AACIFB horowitz_s_Page_005.jpg
1fe17738dbcb7b0925409a6d0ab10221
fd92769af8e50d2012eb6278a395ea71ec8d25e2
1165 F20110115_AACIEM horowitz_s_Page_173.txt
36cb20c3064ddd50fd88925237876396
070ede05d6fd28043fc5b378569c592f8e2e526b
887622 F20110115_AACHZG horowitz_s_Page_153.jp2
fd521260a568f6141472396b1e01b96e
426113adc1fe6485005f2bb5bc8354026f9d3cf6
F20110115_AACIDY horowitz_s_Page_111.tif
1bd84ace52fb3930c3f2dcb694705e4f
49df0786a18b1850ec497021ee99c40ea793e2c6
1528 F20110115_AACHYS horowitz_s_Page_014.txt
8cbdefdb7a812bd5f8a9f27ad5d1570a
e14c507249319b04c85522ecec9e10b6e9006119
44069 F20110115_AACJIE horowitz_s_Page_092thm.jpg
4c45d5d2730557028a3a34a31449c00c
f03defeea318f0ad5afe3fac9f0051d2807b1c53
1203 F20110115_AACJHP horowitz_s_Page_104.txt
b543269d8721af7c0d3dd6e6d56fe5c9
48a0b2ebfde27bd4bdef96e83271d476e492294c
17837 F20110115_AACIFC horowitz_s_Page_056thm.jpg
93f6132bb6c12cb105a17fab844a1ce8
64167b336e6c7770213214128467a4e934b6a4ec
221569 F20110115_AACIEN horowitz_s_Page_027.jpg
6af6ed4bf047119ee60e4280248e43bf
b23c05f58a4ef6707af888b2d3dd7ba7599dc9b3
245 F20110115_AACHZH horowitz_s_Page_191.txt
a54e04705060b20b91e23396f7f86a86
3f94941c6da5c63b19548575190545404996a15e
F20110115_AACIDZ horowitz_s_Page_065.tif
c60511b52d776bbd8449d254d14177a3
6076877885b657008ef9b3a69ce1b4e9ab458fb0
158635 F20110115_AACHYT horowitz_s_Page_018.jpg
895a46b4a0a93559e7f854747eecc8cf
0440a888e9a2808cfa346d606cc4d16625d41093
F20110115_AACJIF horowitz_s_Page_001.tif
4f33af29e94b53c7f74886cc6c7ac951
93928ab2c0dfc13276ac79f87506fab6fbdc344f
64275 F20110115_AACJHQ horowitz_s_Page_108.QC.jpg
84d97527224630a43425db4398989291
03c5736a1e199a1c715036e83b24a65bf70959fa
42059 F20110115_AACIFD horowitz_s_Page_109thm.jpg
b654c815a5af8e94fdd46cf3d4331000
50d7eb8291ee2be7eb43006113f06c04a1190c66
53718 F20110115_AACIEO horowitz_s_Page_045.QC.jpg
2fc49a224c0a2e07ce5b8a57b500803c
26e0e9463e2e95789ae935bc5c5ca7d24a5eb70b
F20110115_AACHZI horowitz_s_Page_066.tif
69240016f1fbcf573246fa4fc84280b8
db3c568160021c25dc5fcfb36d5eadd5080121ec
76684 F20110115_AACHYU horowitz_s_Page_062.QC.jpg
a4a77393872b348d7722056ade60351e
547107dea8ad6deb7e0fbe2a805c965e5326fda3
21246 F20110115_AACJIG horowitz_s_Page_108thm.jpg
db76950a248a65864758e0de4e6f4986
736662ac368fbdc91ed64b529213145f283312ee
64750 F20110115_AACJHR horowitz_s_Page_173.QC.jpg
635eab60cbd0d601fcdbf89ab2e5332e
0edd0a1e27b534eb8542781ade8bfbbd7564bd8a
1051983 F20110115_AACIFE horowitz_s_Page_009.jp2
74b99284998b11fe7e78fb6144b5f9c9
7665c9b981fab214d5bc65d1dc38c3ef6b8f19ca
56821 F20110115_AACIEP horowitz_s_Page_190.QC.jpg
da2f04937bc2f3aa7592c0ca2cac1f90
b4acfa3126c6e8c0962a8779645b575c4635e067
F20110115_AACHZJ horowitz_s_Page_061.txt
f9798d2659c388bd7b7f4d27061852c6
80d78b7ad432be52c762b6e0ff0878627fe9de1a
F20110115_AACHYV horowitz_s_Page_101.txt
d2b132ed65de96cb18d491f4ddfffeba
7ab687795be1369f58cdb350cc44f93a3c7b793f
176721 F20110115_AACJIH horowitz_s_Page_107.jpg
c7148c5c3b4e06529a7a23953da148c6
457f9cb366c014b9e85b0ab83ed8c74f62bbeba6
1985 F20110115_AACJHS horowitz_s_Page_095.txt
32bda927779e61be65ddd125b7abbfe6
58174ef1fd656cd4f5e9919014a997df0c4a802f
782379 F20110115_AACIFF horowitz_s_Page_137.jp2
41d24b6f59122cb8ea91d03ad213f97a
c1f956ac921ae8b0b12071c7871835a28665ee12
26210 F20110115_AACIEQ horowitz_s_Page_060.pro
e414fb3012d9cc673a13cc40d7ceeb36
5cdcd1ad7050bfcb44209b5c9a72286d5d321fae
F20110115_AACHZK horowitz_s_Page_143.tif
9118af34e9076c63b41741a365600867
6d659fb6458779b1d730f9e4821de56d7c73c5fc
23644 F20110115_AACHYW horowitz_s_Page_019thm.jpg
d383be9bcadef7bc2c76e168d11b9ef9
c789ed57688334d0683369c1688150e3f57aa0bc
49982 F20110115_AACJII horowitz_s_Page_152.pro
dec242644f08fa14f7c2feb3c06c27a5
842b60fe25a68ee7bab2f1d4bf7ef39622104d1e
183414 F20110115_AACJHT horowitz_s_Page_139.jpg
e1e2f6ca7c9664928cd484780ed936a0
b46dbc74df17e4c9c84d338b91051ac1c8b2dfe5
2015 F20110115_AACIFG horowitz_s_Page_040.txt
d87762123ea95aaf25f48f28eb304024
b83ff1d3572aaad53af220ce4d03fd606fdb114d
91004 F20110115_AACIER horowitz_s_Page_058.jp2
bf2d34488055efcec206269c65091746
8d7daa378d06c1d2a4d1e8fb9a4a9a481f47e415
187438 F20110115_AACHZL horowitz_s_Page_134.jpg
aa0ae7afce6f4a669b3f997df41cdc39
26f15dcf7fe75b632b4124002d0bfd9b60625c0b
30129 F20110115_AACHYX horowitz_s_Page_102.pro
bc1149a31458b02b99790cd5fe05e58a
6fb23a764dc94fd4cd68de6f213b69fc75ef4885
F20110115_AACJIJ horowitz_s_Page_075.tif
c93f2a4ce82b74bac94e92f8f208e9d2
3563d31a536b6dfcd27836e130c27d41a55075f6
2976 F20110115_AACJHU horowitz_s_Page_011.txt
0607587498ede7cd24d6ed544564639a
2ba1291655bebeca027c34c9e6514cd778dbffe6
31926 F20110115_AACIFH horowitz_s_Page_093.pro
be490768e6688072e36e138153a10272
92247e70b7b98357e18e88b5863e353aca68face
883 F20110115_AACIES horowitz_s_Page_146.txt
03d4168f3cdfec6845170d814ecb7702
80e874e33b947bb9e8a2f81b27aded399a45d32c
844731 F20110115_AACHZM horowitz_s_Page_091.jp2
c50598c864e9287e5ac64eccdde0f7fd
22395a920d44633bafdd12a8083f199832312126
157625 F20110115_AACHYY horowitz_s_Page_047.jpg
f572a9341cf9f8aec4674c355280ed63
c9bd69d9ae3575c2b6ac29731157c3c45cd93bfd
F20110115_AACJIK horowitz_s_Page_053.tif
ba6e00d1d0b8cf465222ba999eec20ce
0a55caa3f1b03210297da333e15160e7c52ca24e
58558 F20110115_AACJHV horowitz_s_Page_189.QC.jpg
6b3108b010ba91bdc39360f4ad4cea56
88a33f74495d9c84c327ba9b60f59d634f574ce8
27113 F20110115_AACIFI horowitz_s_Page_077.pro
39bfd13893eb3b982c3dcec4ecfc7a29
cfef7bb74e9aa9de5fcea5868278959ad07951ce
F20110115_AACIET horowitz_s_Page_149.tif
01e0325cfe8a5532dcfe29d0248506d2
b38e442214f5ffc7d6729fc2a8b512037ff583e3
23400 F20110115_AACHZN horowitz_s_Page_117thm.jpg
96925c9ef496c62b307742ce309e6a23
970436090ada010194575f2305ac91d612a8ae63
88010 F20110115_AACHYZ horowitz_s_Page_135.QC.jpg
67f11458c4f1d60b8fdfac908939d4db
ecfe58146fc3176d8f833359eab766a3bdb1a278
42514 F20110115_AACJHW horowitz_s_Page_073.QC.jpg
ea92f6d220e554c73c3cffe8bfdc40c8
36a5b2d57ffa8c8e34f253c262482896c5d0b7e3
F20110115_AACIEU horowitz_s_Page_072.txt
a2c349a627ebf6aa363baff2d3efba31
545052a7d33ddfa8db7ba56fce78bc03317bd80c
49277 F20110115_AACHZO horowitz_s_Page_132.pro
045b1bf4cda705637a336ba980444e73
c414b22f5f685da472aa28a03ae197bd67f16fac
31358 F20110115_AACJJA horowitz_s_Page_078.pro
0d612378067bd70bac66172f0c505a0a
508102855c76ab7b2afae511db689437acf05bd9
41977 F20110115_AACJIL horowitz_s_Page_150thm.jpg
0c1e18aeb6c055d9c39f06541c26dc1d
66b9d27ca9663a6f0c7eba9de807f0911c287989
812221 F20110115_AACJHX horowitz_s_Page_093.jp2
f1d4c0f3390c08255564c28f95a79731
beca6028119347367dab8a4d4a9670ca705b6a25
F20110115_AACIEV horowitz_s_Page_205.tif
6457e3645b4060654faf0151b9617ea8
03ed71900acdb9262b3019cf6775cb0302b34d11
179919 F20110115_AACHZP horowitz_s_Page_058.jpg
647f50afadd9157a11e5f1017c5d852d
aba71e4666581c0a47eafdb3cdd7925787ce51e9
F20110115_AACIFJ horowitz_s_Page_011.tif
98ace4a52b494f8db8f62d68e6accbd8
bc87080023d52dc761ae5af1ce4672d769639e8c
74240 F20110115_AACJJB horowitz_s_Page_079.QC.jpg
e9807600baf5ec123da6ac032b54e929
c2a3622646555f42e5e00125bff88b514754654b
894724 F20110115_AACJIM horowitz_s_Page_123.jp2
eb64333a39c53089bb87ef6661ec9337
69f62f2b2ee5a3a2c8afe42482e1f3b253c0c3ce
150057 F20110115_AACJHY horowitz_s_Page_179.jpg
05c1404a22b647785e872f2e5834ee51
d8f71f650044a650422677409ae68edd3c8e8020
1712 F20110115_AACIEW horowitz_s_Page_048.txt
51909a6861baaaf82480d0cc5daef89d
d8bd9a4b193d1812bd4d8798908cdc44cbc76d14
173004 F20110115_AACHZQ horowitz_s_Page_155.jpg
c229c00d19f8d9fd714e6fec00e53891
e187db91b2f9f64f1b10add54e1a5c1cf2825987
45344 F20110115_AACIFK horowitz_s_Page_140.pro
9ed8408aa1a45c3d680c3a84f1fa7a26
88801f861b29a711b33ef7e1a5a9cf0a41b79f3b
2028 F20110115_AACJJC horowitz_s_Page_038.txt
99346a495236a5144b6abb5121517bd8
1044f70c3745ac7f34ba8df589f04c6c9cfadde4
84017 F20110115_AACJIN horowitz_s_Page_139.QC.jpg
15bc2ae95391a768e2ca3dde5bb21570
604fc9ca35bab028e5d87e7d397c77f7f11af3fe
F20110115_AACJHZ horowitz_s_Page_098.jp2
8d306870e9ef94f3f7fe421ffef52ae7
92f2843f4dfffaf4220371f981b2e213248e68ce
31788 F20110115_AACIEX horowitz_s_Page_074.pro
4ae38bba594778dc090a1daf12e61d85
cdb537b8013bd9c9f0d75cc8f69d2ec27b59bb31
197740 F20110115_AACHZR horowitz_s_Page_043.jpg
ac72ff3d07f832043d3bc0e1bc21d0ca
ca882fee5fee7be27d92e53164b58c7e3c28e058
1476 F20110115_AACIGA horowitz_s_Page_086.txt
e85b97fb8689a58f1dc23e439242003c
0bbaab6187a395d339012a60b5c463230a295d7c
36954 F20110115_AACIFL horowitz_s_Page_180.pro
e785c9ff76eae283cd8c772b36a9e2a3
e738e7d697c1201b4eedb6389158bf768330fb16
179941 F20110115_AACJJD horowitz_s_Page_076.jpg
f242e5a9c38b93be2ade834276f50c89
e7af784ee4cf64c9ce01062a2f74caf795dc85e2
76050 F20110115_AACJIO horowitz_s_Page_207.QC.jpg
e181a49a7c28f3575344f4d381b81627
869cc064f0c9a2800f17e85055f17a7b5235eded
911 F20110115_AACIEY horowitz_s_Page_106.txt
2bd6b8f5c82e2c7003c8abd89459a997
566348190f85225f8b19be1ea9f7fa77a1de2432
93264 F20110115_AACHZS horowitz_s_Page_021.jp2
3bc03743228a03ac727e4184af9ff0c6
e44e631c949d250029aa95d984afd0064fde193c
188237 F20110115_AACIGB horowitz_s_Page_153.jpg
79895603f33a1869f6395da5a8a35857
cdd8f6b31f11c3c523a159cbd59d75ea7d468c67
F20110115_AACIFM horowitz_s_Page_013.tif
9b3465122f3af148e9d5ecc394882c84
95c3ded505bb910699d3e4f56e3925d50f59586c
2067 F20110115_AACJJE horowitz_s_Page_094.txt
7255467a89dc815fe3156a8f643f8d57
e9c24a4d96f3f0790414b0178fa977fbd537f9b5
99947 F20110115_AACJIP horowitz_s_Page_020.jp2
33dd63f55ab1833178490fdf267272bf
4bcc4b509cad2a3ab926d0c400b391248645c606
2045 F20110115_AACIEZ horowitz_s_Page_042.txt
3fabceeefa3de4bd4ee4bd37b2098207
821832bff0566b9acbb58b9a6546d9b958c5c4f8
62094 F20110115_AACHZT horowitz_s_Page_197.pro
2f65efe31a272a277e9a65d2a3eb94ba
6167460aff3e8925c7510b142332d44cb498e0a0
38646 F20110115_AACIGC horowitz_s_Page_119.pro
fa1be455598a498e63e8f43ce1664d49
c5c4ecaf907f6f59ec4449b64385989155ad2bbb
5530 F20110115_AACIFN horowitz_s_Page_030.pro
65402e3d627680e3c1ce91ba37cd8e8e
03a9bbe936096474256b54dce78cf56f5aa3faed
17127 F20110115_AACJJF horowitz_s_Page_111thm.jpg
8babd63edb034f70f1d0188a69b60f04
b3a693eb266db9725458edd022fcd888c75494eb
86876 F20110115_AACJIQ horowitz_s_Page_092.QC.jpg
e8d3a43e227dc94c90b0a28263da457a
bbb35046ab8bc247ef7f007c26aa223c385b5877
1893 F20110115_AACHZU horowitz_s_Page_087.txt
44d793724a96abd3e37795cad0e92c55
894f571d22dff40ddf0fdee28f06b0928068e9ac
51804 F20110115_AACIGD horowitz_s_Page_121.QC.jpg
71bb7cda402da39da0b339c5b02363ce
bb87ea727490c1caedb084a71c68763137e340b4
111111 F20110115_AACIFO horowitz_s_Page_158.jp2
a0ace65a485a33ff3d74058dbb86ede2
abd6c4454557de3683efa5eac21bbc29ee5c9fce
187536 F20110115_AACJJG horowitz_s_Page_180.jpg
b1a1818e00390c0ee51f587250d03d8b
76d3eef81e4b2ab414400400ea07ac6d54dc50a7
52836 F20110115_AACJIR horowitz_s_Page_005thm.jpg
572a6261de1b0d685e68bf84bc365646
3de3ac271237c1fccdd937f173750257f91e40aa
48610 F20110115_AACHZV horowitz_s_Page_148thm.jpg
8e82b315cbf0ce14fe1ec2e2f31ca83f
1e187a7abb8f880ecf8627339311bc6415575d2c
F20110115_AACIGE horowitz_s_Page_015.tif
d442da1411551f3aecf2fb724fedcd7e
f4667a226587440f67a390baaaf5fe1d6e15be56
44479 F20110115_AACIFP horowitz_s_Page_069thm.jpg
80485d59f0d354ca5d9c8bb6986c3f57
bdb01cc24b66f002881a708f979b60647ac68ec8
F20110115_AACJJH horowitz_s_Page_006.tif
157d3faa6b7f96128e33a1bf1cc3cccc
c9d3a484b09f3f9d0da33701be3aebadc05821a5
44299 F20110115_AACJIS horowitz_s_Page_076thm.jpg
fdf068cdbe5c16bf936a0b3de6eabd10
d418d1e4b1ab4c03d26f69f0de829d18fc8ea253
43589 F20110115_AACHZW horowitz_s_Page_097thm.jpg
0a789ca27803461b7f62de8fa0e225ee
3250424cd82fb3902603ce5739f0f83e35812c88
52757 F20110115_AACIGF horowitz_s_Page_044.QC.jpg
56455fcdd6b28ba2be7b118d0b4088ec
555e92cf2c183a9e3e396d36ae35f1688753e897
F20110115_AACIFQ horowitz_s_Page_073.tif
647ca640457d31331d4c07764eb3504c
8cfb24e29984d04a4db4afffb94ba79f7a307df0
F20110115_AACJJI horowitz_s_Page_198.tif
c584c89f437c72821acb145856ade310
7f38fe560708796fc0c93316a93b4bafcb9e0b8d
110081 F20110115_AACJIT horowitz_s_Page_157.jp2
592f4c2de6ed2c898836b04684094756
4f5f66cac5ae492903aa462a398c54a1b19a45a1
188183 F20110115_AACHZX horowitz_s_Page_161.jpg
cbf98dcb766add5859368d0e9693f1a0
cc5277f46d446a58c9425d4a3e68318a99f179c6
80219 F20110115_AACIGG horowitz_s_Page_088.QC.jpg
8b2c678788a5a50282e3f218dcf7f208
e5980ec84e5afdba2dc2acdf9cf53c2857f057dd
891924 F20110115_AACIFR horowitz_s_Page_082.jp2
b3c4bb378515732fe9dadae8d3d02018
5f70ad396e96f642baa609cc8139960c42736fff
68812 F20110115_AACJJJ horowitz_s_Page_009.pro
c6c28365d0d6ccf396816da8201ecfd2
c6a6ac3434339d97acdc89282afd6d330367dbbf
1051948 F20110115_AACJIU horowitz_s_Page_174.jp2
f9b9424c7351f93c9a17a905ea929449
59a1007b385e3bbe8671820d9bd85d0b4ab4fac4
34565 F20110115_AACHZY horowitz_s_Page_209.jp2
a521766c61ed18964051e6aeab184044
baf5d6a2250373992d539dbb10ffe4bb60e79e9b
196693 F20110115_AACIGH horowitz_s_Page_133.jpg
73f0ed2049f868743b6b183b8e6fbb74
ae1bf4674424d818c9c587868509f13b32eb1f69
71637 F20110115_AACIFS horowitz_s_Page_106.QC.jpg
d489cdee6e87e0009d743bdc8e85daa9
bd76c2bcf63535d26585f3f78bb27ec3d9b7c23b
1671 F20110115_AACJJK horowitz_s_Page_051.txt
0609ee5d17786766d752ae7f0cdc0eec
a916cb872436fb3bf68496112b66caa8e3668da9
2071 F20110115_AACJIV horowitz_s_Page_058.txt
d3d7c2750302657050bc76fd46a03ce3
b3b063ba6e1d54169b457ee759f7f9f422d1ea6b
F20110115_AACHZZ horowitz_s_Page_085.tif
e25050a3b96b0ea27cd7eff74274ba3c
76757ba117d0ee677a392fad355d7ebcccaed030
87131 F20110115_AACIGI horowitz_s_Page_128.jp2
4e7dc93ae7587e298d5bfc35919e871c
49e2203e256454a60ba5e6e5bd060c6b9db559bb
F20110115_AACIFT horowitz_s_Page_037.tif
4c29cced11498b6e48aa8b12a8137d5d
a4f5a00b49bc2f76faf0e129ff43ad7a7ea7ab60
F20110115_AACJJL horowitz_s_Page_033.tif
6ced2d3a495b63fc8f89fe5b1a181bc8
d50d9f33ee79087e9ec4175ea7d658887b404b7c
68374 F20110115_AACJIW horowitz_s_Page_058.QC.jpg
51efeea46613962b151125dfa7844394
caff3eb8ab7434666bfc8ec943f2fa6b382462d3
81034 F20110115_AACIGJ horowitz_s_Page_145.QC.jpg
8d31a3f42d120a4900024af935fc6c08
e829266b40e4597aa49bf5cfa48331718533d222
F20110115_AACIFU horowitz_s_Page_079.tif
6e7bc1e243ba56345d1a0a57d9128dfe
75d02d79b5184711a097d75623b8a5124140ec88
1007302 F20110115_AACJKA horowitz_s_Page_182.jp2
ba6501b48d11d355c281aae63d731467
297b81ad15f8f98e6fa2c60df9fc5b1e2d4de98a
47318 F20110115_AACJIX horowitz_s_Page_122thm.jpg
edca424e8e8806d04949b1578d4372a4
cde08ccc76451fb72b13bc38f106c326afb391b4
782 F20110115_AACIFV horowitz_s_Page_136.txt
323dba350bb654e469419cbd0c970fff
b963982157956dc93fed6f5e044c925c2364427c
475824 F20110115_AACJKB horowitz_s_Page_183.jp2
2b9b467edc39bebb4ad5d743e8c147ea
ae851f9b4d353cd1d674c3494484cd60c5568a7a
783254 F20110115_AACJJM horowitz_s_Page_107.jp2
6f2a955f2e512b4b5ada30022e901e50
a8b7aacb067b527945bcafc7c77af658ae09b818
88872 F20110115_AACJIY horowitz_s_Page_130.QC.jpg
84d1d4cabd080efa29756c3970a7b2e4
b3136ad10681e11fa4925074423658c79c20682e
44535 F20110115_AACIGK horowitz_s_Page_068.pro
e979bf3bff8c99c3181083c08031f80a
afd3314efc38a2812dfcfca48df61acf71aba526
151733 F20110115_AACIFW horowitz_s_Page_177.jpg
5b2697d5e426428dfce02726c8c71895
07f438477dbb86d1a65b8da86da0c22935197f05
1626 F20110115_AACJKC horowitz_s_Page_142.txt
8480ac9a0aaabb6622b44984f36f7a90
4367197dfe0289a106a77372271eb142d02fdb2d
72930 F20110115_AACJJN horowitz_s_Page_051.jp2
9a465a0f34352397c85dcdbfcae6c5bd
9e3f09cb496a768e1d635ed26e3cea5360befe17
44370 F20110115_AACJIZ horowitz_s_Page_101.pro
790f6e757b8b287714a6727875b90ebc
258c47133edc97433973b72eabc8f35fd42fdbc3
F20110115_AACIHA horowitz_s_Page_176.tif
d90ed0288292d5c52b98aa2e2fad744c
f0c69c4f957d33917a5a371f39c00e69eeddb3b1
F20110115_AACIGL horowitz_s_Page_038.tif
1f20e609bb3f12a4d8726a5a8da795bb
8f75ba47dd4573ff026241379f18e150d04b63df
2525 F20110115_AACIFX horowitz_s_Page_203.txt
4152a6a159c7eea7b52bcf137e66727d
90e45441e6413405a8bbe021f097390ae76ec3e9
1051982 F20110115_AACJKD horowitz_s_Page_180.jp2
728bed6db116cdbfa2871fd3fa1e3905
b025223df4fd9bd09cc11cf1353ebe37afe1f0ab
49026 F20110115_AACJJO horowitz_s_Page_162.QC.jpg
3b6afcc238da568949f447b84a6094df
97e4c27df5ce3bce505b02d2e9bd9fd7d5e9bf2a
F20110115_AACIHB horowitz_s_Page_082.tif
b60e7a2cae332578655e6c3181ceb63c
9a0e13887a4764ad547f9dfc457524b524c6f884
1553 F20110115_AACIGM horowitz_s_Page_013.txt
a16fbb466c20b3d840fb3381814b2fc3
5e61d3e1460b9a7ef2a73f9d62bb1b0c9c4332b8
25636 F20110115_AACIFY horowitz_s_Page_116thm.jpg
793dfea05e045297df37dc7c1b147ee6
140b4efc40d0742b3439da4cfc7bb26efc5ccd59
43559 F20110115_AACJKE horowitz_s_Page_137thm.jpg
e606993fb0279c44754501f319fe1ac6
9e28f31a5efcb8947724c3e50fb2f9dcbd1f188e
24623 F20110115_AACJJP horowitz_s_Page_196thm.jpg
419f890075238df472963b0a94fc2bc5
d2314ac98ec66c50e5e230d61b8e87d07b7f388b
410 F20110115_AACIHC horowitz_s_Page_192.txt
26b2742c4d0886dba6ee689a4418246d
7f0984dc4bc868fff0834e8967a04496f480ea91
F20110115_AACIGN horowitz_s_Page_101.tif
120a2623d29f13489d75d0b5ca9147a9
c737fb844b45641a739e5803e4dffe60d53c0d94
108992 F20110115_AACIFZ horowitz_s_Page_152.jp2
41390634ad3396e0c5dd4ac0cacbdfa5
81b7532bbb4d08044a02b4abf394441e724056af
64705 F20110115_AACJKF horowitz_s_Page_159.QC.jpg
a08e414c31d5904dc1ab222b5deb951f
d1f445cbcfe02bb87905aa92c110229525c700b9
83805 F20110115_AACJJQ horowitz_s_Page_107.QC.jpg
d3140ccc1e8fc6902a0f8cdf90221f67
d80b73b07fea15095bc8ee41b67f970b595f957f
902 F20110115_AACIHD horowitz_s_Page_159.txt
5acd5226f6fa5925a826fe84c532477e
2cb1eb7ac461b1809f05a23b930bee3f618517e7
2056 F20110115_AACIGO horowitz_s_Page_157.txt
c3d92835f32bdcbf516b265d34c72ce4
131e266de977f43d629ae460952a0e5b41dea26e
42779 F20110115_AACJKG horowitz_s_Page_187thm.jpg
62bcb3f9779518d1fd8c66dafdfc8c36
efa4159bf9a9f3a0c12e670d253517b777080db2
98 F20110115_AACJJR horowitz_s_Page_032.txt
278719e746955aa867d99dc2ec89b8b4
7c0530b336a3f8c35ebd46c296b8ef4a6c05935b
1674 F20110115_AACIHE horowitz_s_Page_134.txt
9c3381cfbc641bc947cb00a6937830f5
a8c86c13bcdd14c580ce05f919755341498c56cd
34549 F20110115_AACIGP horowitz_s_Page_064.pro
89be51bd3a0919ec2d3444b91b7e3932
58c8095a8cc770f78339966f519e64b95be272e0
169218 F20110115_AACJKH horowitz_s_Page_015.jpg
08bcdcd947b8d485a358a28ceebc93ec
13b315eae86e11b47e2285c1e1ad6b55fed276b0
29146 F20110115_AACJJS horowitz_s_Page_080.pro
831110087dc4d46ae8eec054c02f737e
3144bd3caf2ae5b0e8923ca3916301eaf26474cc
F20110115_AACIHF horowitz_s_Page_017.tif
b4bac2926954bb17e87ca778e79552f0
e460583385763c7181e6244dd1133856c72a1d2f
F20110115_AACIGQ horowitz_s_Page_090.txt
3ee2555549ec19d7595785b77100856f
b86c44892621530ae417eec319840a70ba1de2f1
937 F20110115_AACJKI horowitz_s_Page_050.txt
56ff74e6bcbc96786cd5838413822438
a28fedc127b8b39b9542d7f794c57a7a5b48cadf
1737 F20110115_AACJJT horowitz_s_Page_108.txt
32d6540130e8320592387f4d0755eedf
69bef66e298ef7dfd147b8fb8252b19c39ee68e4
39232 F20110115_AACIHG horowitz_s_Page_090.pro
d67e6cd3bc45cdf0ef8c7fe6768a15ae
9666285e62e12b5c457412c40e1ca7a334ca761b
1365 F20110115_AACIGR horowitz_s_Page_111.txt
45e8f61cff687cd499ff675f7c5ca945
5bc148437c1208f6e9617d1d24a68821d005b22b
771396 F20110115_AACJKJ horowitz_s_Page_079.jp2
21993554d53c28b922343de0257e22c4
8c29abb466a76fee169a7e44a9d54762f15322f8
40033 F20110115_AACJJU horowitz_s_Page_136thm.jpg
9e804ed2532eccdd776f80e5935639df
4bbd28f4f09c3404ce816c7a591e59cf479b3a70
107387 F20110115_AACIHH horowitz_s_Page_041.jp2
2d789de9f39f93249e3180062670fff5
ea1172e8f70edf5aef8899c8dcc6ae906e2e32f8
45939 F20110115_AACIGS horowitz_s_Page_142thm.jpg
f27cf6388769df48a4d1ed1dda917902
617b014bad06a6187ea42b904dcd6cb97c778f65
38549 F20110115_AACJKK horowitz_s_Page_075.pro
8fd75783623b89dce19b5c3acae8f56a
fd73cef7c87053fa06743877eed998e9629b6523
1682 F20110115_AACJJV horowitz_s_Page_046.txt
86d7a06a61c37ad94a656dfcbca73258
5696f40485343c5b666054f864c67b14a9c14f91
158560 F20110115_AACIHI horowitz_s_Page_119.jpg
8be146bdb7b3c7f975d634e2aa52a113
bcc1db7266ab223ea6553758308e0d018623385e
23151 F20110115_AACIGT horowitz_s_Page_068thm.jpg
253350ab2009a7b487250ec35ae76142
c27ec9df55615daf2cfc9a0775a6f7d44b404115
1272 F20110115_AACJKL horowitz_s_Page_166.txt
5d6f6112718ccba927b712dab502515a
7517fc54a0863b8a761ab3e127f333014284a8f9
1992 F20110115_AACJJW horowitz_s_Page_169.txt
df7c7257f5091edf9c909926a6d4df8f
70e94c6abe839508a008e990bdad88649330781c
106847 F20110115_AACIHJ horowitz_s_Page_026.jp2
914626c6078b55952623bb36dbdbb03c
6043f552c513bda69e411fb1681de4cb06990781
60576 F20110115_AACIGU horowitz_s_Page_208.jp2
93892563aa7fa2446171dce4051ff736
9656920c1de5003ebf1e0d966d52cdf9db3b1778
158765 F20110115_AACJLA horowitz_s_Page_140.jpg
759dc1d8914f09648fff43a06b7d1468
2358509e67ac3514483dd9fcc59f1d1d57b990df
586014 F20110115_AACJKM horowitz_s_Page_109.jp2
8e661379d835f75cc911c50229896344
a35b25a2c5ee662a27bc5584e4eeee5a0d3afa12
F20110115_AACJJX horowitz_s_Page_014.tif
c8c1c8caf387a4a3678bc0709ec93fb5
fedf3f5dda9b10cb40722a310358c3ec2d8e077c
326199 F20110115_AACIHK horowitz_s_Page_193.jp2
3da3b06d6d987090cbfec123d46800f2
185ebff2b789711f0a4107865654aafeeba9b4d3
2030 F20110115_AACIGV horowitz_s_Page_063.txt
e379e65de3d273e551502e62dc041a69
b82fc12fe5f623c1bb3b66e8e013e7c340a68a80
107125 F20110115_AACJLB horowitz_s_Page_033.jp2
ac7e59a5b6bb2dc617a01b80e2c1d94b
193c664bc23d95d0d60c1d9c236a3d571d86c038
F20110115_AACJJY horowitz_s_Page_069.tif
870d27949d8397293a9395c0f1350951
3ea00dd826b5b9a97d988361711c387d8f9d716e
49869 F20110115_AACIGW horowitz_s_Page_041.pro
1ad7bc477c1e5087c195c173a28bb0a4
c85320eb7d8bf458a7ac88ae4047dd7ff3e797a2
49010 F20110115_AACJLC horowitz_s_Page_127thm.jpg
e34f4e776987b6f3924d3c6a95457504
358d26a93a821183302799ed9870cb0c3865fff7
F20110115_AACJKN horowitz_s_Page_107.txt
8f1c6d98b79cff459323108ba669cbc2
0d47aa7753f4b977e098e5ff3e7ec2d7e4048698
105835 F20110115_AACJJZ horowitz_s_Page_029.jp2
1aaa23578e06d894c8d955c476af9110
1d3287b6267cbd6a776ca36ff7a0b90e9944e4fe
128373 F20110115_AACIHL horowitz_s_Page_200.jp2
a0ff3234b7befdbbf2fb2460c6342dec
6eb1ace942be8ed36379f7c686683f14986c2c35
1269 F20110115_AACIGX horowitz_s_Page_179.txt
c50c2eb621083caca894c957fde4e7b4
fe2d62571e75eea3b34d3ae34e4fe587cae72817
48994 F20110115_AACIIA horowitz_s_Page_033.pro
85c269bef4ba997387299dbe33dc37fe
d98941f17ed83e844db93fb5a1728d291dc332dc
44640 F20110115_AACJLD horowitz_s_Page_153thm.jpg
93e0937216dcb8973c2fe7ec4daa5baf
2ed72e434f5cc5abdd5540ed6cad94ecd1587f04
144241 F20110115_AACJKO horowitz_s_Page_035.jpg
f44c3369643dd0b4abd60a5d3e3f59e1
cee7be7eeae8de51b0ddddbc7ada5820cca173b8
21041 F20110115_AACIHM horowitz_s_Page_050.pro
0e94756b1b995e17f70f01cc5b6ebfb1
1906748b3edf7317a850de16d7188fb00bf9de7c
43175 F20110115_AACIGY horowitz_s_Page_031.pro
1d01ac743a4d95f28e8951616c0705be
b8f4ec988670ffdde774da5db113288bfbe9c40b
149463 F20110115_AACIIB horowitz_s_Page_060.jpg
2e7bba2c4f3493d1276e0305ce85ac76
535cf660b2ad3d849a4be8681e13428b9bd7fe98
44567 F20110115_AACJLE horowitz_s_Page_078thm.jpg
49238006696c3f73b920b1f51ba49c03
72a12748dd90ef603bfcdde2d272cb34daaa26fe
36893 F20110115_AACJKP horowitz_s_Page_065.QC.jpg
f9f3b0aebff17a465365a07cbb6f1357
35bc61daa85dc24bfa54f03aeb24c5927feb07c3
F20110115_AACIHN horowitz_s_Page_129.tif
886a7fb01aabcceb6829a96f2d43bb2a
03d64725c23998f2c99738f6149b7586d2c3fb2b
903545 F20110115_AACIGZ horowitz_s_Page_133.jp2
9faef42880e0506a226f6ec2d5b3fe25
8bd1f30322a7781f9b626af9d10af04aad342cfa
64335 F20110115_AACIIC horowitz_s_Page_201.pro
38e9b9f6a930931a63bb440cd31bb294
8cca6915186624e19d5346631faee5f6dc926329
29612 F20110115_AACJLF horowitz_s_Page_048.pro
319c24ffcad745fb903c438db363b626
abef00fb7f40e217bcf515c8adb7f9327a8c7304
F20110115_AACJKQ horowitz_s_Page_042.tif
01c1e052354c20406f98303e6725ea0c
8f47114caff51f9d69a44fe0a7f28a25952da95c
39712 F20110115_AACIHO horowitz_s_Page_208.QC.jpg
f73b38daa3a46bc7c881255d79c01971
a818c75ad96b597dc707111b15745a54bddd883f
84695 F20110115_AACIID horowitz_s_Page_143.QC.jpg
b971b3bcf8f87a26b6b6606b6b93c253
89ec6eff5a13b7b53580253d6af1f92fc17d5142
F20110115_AACJLG horowitz_s_Page_209.tif
cf63c26e83e3f182bd0c6e803ce54473
92f9c849c3f6da38a125dfff4dfd18e7a9edee78
72765 F20110115_AACJKR horowitz_s_Page_048.QC.jpg
bd8811a838833406db29a630bd08045d
c3f5f1b34956bf2ef3be7e24c166b7e845b80f99
275689 F20110115_AACIHP horowitz_s_Page_204.jpg
41ce261893f9683268b23da25aa84654
2d0a461a683795bbbfa7d8836d40d56bfc00a57b
25206 F20110115_AACIIE horowitz_s_Page_027thm.jpg
ef5a9e24701e59dd2a6d4e4119dd910c
2b6821109821b84f4392e89e0b3b3406a7964e63
72802 F20110115_AACJLH horowitz_s_Page_096.jp2
e06190315a65954b1ac210c9057a9edb
ce0553ef593e5778a94b734e46f433b7fa3ff63d
27599 F20110115_AACJKS horowitz_s_Page_018.pro
d35782ef57678efdb43f13d674f099d5
776944ec180e10d9ce784165f972d5a61e593d03
52720 F20110115_AACIHQ horowitz_s_Page_064.QC.jpg
cf6f02a7e6799e561c628148e81dffdd
3f225dd18213bf5a7865c3d453ce101c7c0c9a9d
71704 F20110115_AACIIF horowitz_s_Page_144.QC.jpg
4c775c498449fbf976b27bba2b0ea1b3
2f4e6425f7a6bd5412850053c2a9c37732464e80
160047 F20110115_AACJLI horowitz_s_Page_187.jpg
773325fa9df8b13bc2fd22f62b5a8cac
24f5618acbbd090f507a98efbb24e3c6bf6a137f
44495 F20110115_AACJKT horowitz_s_Page_118thm.jpg
ea4c931b20c5ae1b3e304013ac9f30a5
7f29d4dd8d398f703b26786368b2ff2770a03bcc
36608 F20110115_AACIHR horowitz_s_Page_112.pro
785b24b41ca73742a603e2f7c62b188a
557b99b815f633c90195bd0a81db05feb2e03017
44982 F20110115_AACIIG horowitz_s_Page_081thm.jpg
3046c669e31c6587f5dc50bc9ab5ad9e
7dfd966f5be0eb3b91fa29d13cf9331b2319ec9f
18156 F20110115_AACJLJ horowitz_s_Page_051thm.jpg
be060e8c078606fb0f25772878913a42
e6579628b2adc815b2ba7aa9d3c415c73a5ba0ce
141240 F20110115_AACJKU horowitz_s_Page_173.jpg
5299bac82718a33f04e110443b12852c
16400c316f4d5d2ffd1365284ccb41924420fb80
3145 F20110115_AACIHS horowitz_s_Page_002thm.jpg
64667c501053297d2845994570ca5530
9245d64bbb1e518d5b0b7509d066fff2c4183786
28492 F20110115_AACIIH horowitz_s_Page_084.pro
bfa7050a98a5f22a668927c44a72bf8d
f4576162b38b1f08eff11633b005b3a08a9bf05f
123619 F20110115_AACJLK horowitz_s_Page_072.jpg
578cf1ead269d9826e74cdc36495a082
d9933b523f0929d68fccbeb08d2121b6124fee9d
55388 F20110115_AACJKV horowitz_s_Page_192.QC.jpg
7edee1f915ffb627fbdde3a884b68464
a6c5b1a70cea8a12ccb9936908412019770fbe6e
170941 F20110115_AACIHT horowitz_s_Page_061.jpg
9f707effcf5c9c709e4ac977c2123322
b1ebaa27bc84a2feed8ee61efa5777bef91ec5d5
F20110115_AACIII horowitz_s_Page_080.tif
89d5df1053dff9e5f1cb279f9eef6a8a
f8928483399bebe8ccb790172f18ce353d524ae6
74023 F20110115_AACJLL horowitz_s_Page_006.pro
559e8a7866c2f215e400b944a3dea5b6
12c916dc8145314c4255cb505e39701bf0b1d64e
20692 F20110115_AACJKW horowitz_s_Page_106.pro
85823e8faec39a3a617dd961135b6aef
a4d3177dc668bfc9bb85ed96e9c6b903dccae477
37857 F20110115_AACIHU horowitz_s_Page_092.pro
37e0c71fe492b2f9ac9d125a3006688b
47a72b2e98f524600b436f0cfa781e7a08e60ffb
F20110115_AACIIJ horowitz_s_Page_084.tif
fbffb78d0a97d455d7e1e5c04cdcb229
d1200e8cd1c488565a46d65a5ed9153745615778
136750 F20110115_AACJMA horowitz_s_Page_011.QC.jpg
9023b2fbca2db3f6df5ed373619d76a2
7ebedc104835242167fcb0d4fa1f0e255c87440e
58618 F20110115_AACJLM horowitz_s_Page_112.QC.jpg
b06a3e64f321f28d0c899329a1ce63e9
a0afd9a07dedebb6534112f93796d6cf4ae6d926
86643 F20110115_AACJKX horowitz_s_Page_118.QC.jpg
e7f0a1ba12c0c77a4ad6c02951a7976a
43ebdafcb4293f756d88afb8ef30706985315302
54196 F20110115_AACIHV horowitz_s_Page_037.QC.jpg
a80e408a7ce67c3f7c43f014465f77c8
e3738e1fff2c6f13ac10b1b16420318bf29d1a1e
F20110115_AACIIK horowitz_s_Page_174.tif
0c89a86e17953e929784d0f27e74c503
05348aac0f1b2a4ff160a704d2b6e04555e19234
84763 F20110115_AACJMB horowitz_s_Page_004.pro
1c0950e60f76a64c1c09373893f7d719
da5f1eb08c60cfc7140d2e345dd842e192a0a3f5
103817 F20110115_AACJLN horowitz_s_Page_008.QC.jpg
654a4b1b94830effc0cdae569066ca6f
858cb4d4a8e9951dd66942585a3a37969ffa193c
1604 F20110115_AACJKY horowitz_s_Page_059.txt
007c526a87d02a7ce15881d365fac562
efbe27b1e0f84a2742918d920c7e74f80d78f2b1
52713 F20110115_AACIHW horowitz_s_Page_094.pro
c48f0d8b4f04bd6c316d920fe8b05f5f
857c4ae5b967d952ab03bda1ae8211e39c57863a
1546 F20110115_AACIIL horowitz_s_Page_176.txt
f5a3f76b23889f526017c79866f4c085
adcc37f579515dd3a1ffd4bae1e6f17c8b7af24f
71120 F20110115_AACJMC horowitz_s_Page_054.QC.jpg
c1119b4b7ce184d679be74ebd02110c2
ca34cc5ecd73f6591fa963b0f49f83c0372c2dfe
95902 F20110115_AACJKZ horowitz_s_Page_055.jp2
bda88489c88b1b8e8440a0c0ea43882c
97f22cb358d055ba756e46cd5d7d18990c8d23ae
132930 F20110115_AACIHX horowitz_s_Page_197.jp2
29e16f0d31b363063886f036aeab99e1
a0297c5cdb94edeae9d73066f26d52a433aec96f
30568 F20110115_AACIJA horowitz_s_Page_056.pro
dc892fadd600d4db0caca5872b40d488
11ef41fb35896fcb1396f23732f5a2ef4d3a7e81
173562 F20110115_AACJMD horowitz_s_Page_084.jpg
38d6c32980cd26437cd501ebfad59dd6
60877581166228676b90146c3675ce02c9718948
F20110115_AACJLO horowitz_s_Page_146.tif
0a7ab817c6dfc2690feae17133455802
0f1965f3c8d440f6c81c2b2885e6d33296278327
8664 F20110115_AACIHY horowitz_s_Page_209thm.jpg
919e7e19fcd03a7ed026068d2daad894
910968634507f30f8791946974c5985389158d7c
162444 F20110115_AACIJB horowitz_s_Page_144.jpg
b457d2671c6ca4cfb291f76e87cb29f5
de485579f93cdb6404f13108fe6a66f5c04ad0cc
F20110115_AACIIM horowitz_s_Page_148.tif
456626fe9950584e7b8747bf469c101e
bdd5fadba40ebfc64be477d441c697f19fd61d24
85978 F20110115_AACJME horowitz_s_Page_124.jp2
d363f955b901c2717443a0fc461a54ec
2df81b89743170faf47118ff8d1046804db83480
39313 F20110115_AACJLP horowitz_s_Page_013.pro
e88d88a777e5c39d33d7108e6f7038e6
c966bfe55e2be3af7b2992eb9274d8ecb5661183
72614 F20110115_AACIHZ horowitz_s_Page_109.QC.jpg
cb4da05ce87729954f1fd112e2b348a6
a4eb66d166fbb2a1dcdc5e60bb22008e28c8356e
F20110115_AACIJC horowitz_s_Page_031.tif
a482607bfcc68f0e46201dc59950f4fe
759fc41e9517474b4099feaba0d03303380a4fef
F20110115_AACIIN horowitz_s_Page_091.txt
cc6f84a4599930145e08eb6e0577a86b
f8e145119c0bb75b18742dfc2ec0ce511d6e34fd
F20110115_AACJMF horowitz_s_Page_199.tif
fbaa5533baa3d6db1755863b3d5ccc97
2e47ac46482c8ced0d1b0e815d8bfae0cb44d551
1751 F20110115_AACJLQ horowitz_s_Page_073.txt
2d7df69718231b8a9f7c21f6a2ae41d5
21b0ff942dd4733aed6a8fe639dd59af9eb3ff01
88360 F20110115_AACIJD horowitz_s_Page_103.QC.jpg
28c7cfe3382ffb41acbd21aed93411de
59159bbf838bb35d1317f3f3181bf39c0f63f2da
F20110115_AACIIO horowitz_s_Page_052.tif
29bba92731d723a086218392e8e4a2cf
949a3551174ea04f9a5e6d7291e052d07da69e66
63410 F20110115_AACJMG horowitz_s_Page_204.pro
2231d167f8de3005761170025b931cce
611a7c023f0fbc7543e5c30cd997b22ec4e68032
161307 F20110115_AACJLR horowitz_s_Page_097.jpg
a325f4b9541834176cb3d0e3e3e42a30
ad3c4ff5fa6798133fcd3ccf806070844f5f669e
899599 F20110115_AACIJE horowitz_s_Page_143.jp2
f75b7640696e5f1c915dbee75207f638
da32dc12f436ba41c0b12716aeb6d8ee452e0258
47064 F20110115_AACIIP horowitz_s_Page_090thm.jpg
b44d4e409aa531c7aa4430199125257c
a7f4d54acc72aa929d7e3949ad2221fae835d25f
15578 F20110115_AACJMH horowitz_s_Page_186.pro
6421ec3154a57f3fd705cd6cbe450a21
dbe55e8eff0d60416015458377b5a7a46f6233f5
251937 F20110115_AACJLS horowitz_s_Page_127.jpg
0447cde130b1747b6bc688a7417f6001
967873cb39489ae979b9c94b7e4cee6c2313ce3a
45638 F20110115_AACIJF horowitz_s_Page_134thm.jpg
94ef594dadfdd4163bd36cc2684d1146
eb60a3e549394b6899daff2772a3ec3f04eee67e
43203 F20110115_AACIIQ horowitz_s_Page_062thm.jpg
df738e321de3d77992a8e1cbce4dbe38
a1a73f3b9adc93ae4238dc552dc67327d20fb86e
73889 F20110115_AACJMI horowitz_s_Page_099.QC.jpg
722d302e21ed3a9000d55998a999e8cf
d7d6b790c8b685e16ffdd36d9c42bdefad55c53c
67926 F20110115_AACJLT horowitz_s_Page_170.QC.jpg
203ce21c1bf6a336811b08c5e9f025b9
3c584429fe51de523b70586e253c87d224bfd1be
226 F20110115_AACIJG horowitz_s_Page_190.txt
d66dc613af47dbd35ca70a2f0ef128d7
38621818f7d9fa582d03f559122b547e68bc0310
28304 F20110115_AACIIR horowitz_s_Page_035.pro
4f327267801f754a5af9aff33e9d93db
9b673c01bfc45a9e20e7c0d0fe411ea51d3252b7
52717 F20110115_AACJMJ horowitz_s_Page_158.pro
9a36eda1107adac3c1fc94643c06137d
dc2c79e034dc5765e400477d0ecaef5321a9f1d1
F20110115_AACJLU horowitz_s_Page_172.tif
adf7f2be391c806fc19d2249212126df
8f581a11d39968da3ea8e2a065e6918a943bb74e
1816 F20110115_AACIJH horowitz_s_Page_119.txt
ad22cd59f9e2783ecbb197001940d316
0ec1b8981e8535d9c07e6e4f029c0ed2f8328e4a
1579 F20110115_AACIIS horowitz_s_Page_093.txt
ca621c6b93a454773c61f138c6e21aec
449a989f72efe3515e007a692944b44a83be9644
50437 F20110115_AACJMK horowitz_s_Page_023.pro
73b679219f273775eb98b5c9f93c5b9d
d87093eb2ab1252d07e61ada1c2e6c5fab047b47
19887 F20110115_AACJLV horowitz_s_Page_112thm.jpg
1d9e5199285b5d4a8672c2d212ea446a
1eaf73d3357600219da69fe8f5274fcb56bdcee1
1430 F20110115_AACIJI horowitz_s_Page_109.txt
dd14ee9100a8dc9337678c3c3fec5344
a129a1a15f3f31068ab9dfd1f4b578e400767c4d
1051944 F20110115_AACIIT horowitz_s_Page_102.jp2
cc74d55de798cd9a02c8e4bcf7afa44b
f1e7a9d21b3d6fa9ce2401d18f1b76a38c4ce5ac
273632 F20110115_AACJML horowitz_s_Page_197.jpg
b5a6d9af2e539f47f0d9742a2297a276
a16d3e389ca6543a398efa07a5ae5695d37c20ac
F20110115_AACJLW horowitz_s_Page_145.tif
1b4d385d2faef07ae3c94a29af398b20
dabe3467af8141f844a0a56a618429377857da87
195287 F20110115_AACIJJ horowitz_s_Page_082.jpg
b7edc0ec033e941d16c00232168d5e48
598f4d32ac60da5dee8c9e0b2ed25b0c3d761aa9
22661 F20110115_AACIIU horowitz_s_Page_076.pro
1a49aaff9aa5bf6ee87ba6d175d67fe5
fd9410bc10cdcd9ac45e1c4f2b590a40a70cafa1
172486 F20110115_AACJNA horowitz_s_Page_120.jpg
d3ebe2388ad420a6ff7a27cd4c113406
04d3ae6143f574a7f5bd655f27ccb93289ee7908
1708 F20110115_AACJMM horowitz_s_Page_118.txt
26d17dffc3ce34816274abe79188372a
e9dcf97c8dd9281625d59be7b8092c9ab314ad82
F20110115_AACJLX horowitz_s_Page_064.tif
ba7cc60173277c42d5ad3332acb887d9
bf57f0b82187060ef18f683fd36730192e437a6e
1829 F20110115_AACIJK horowitz_s_Page_044.txt
de9877f2bcc465b0f52990c476ef3951
f250b86d80d326afac9f11eb15d12fa528d309a9
600644 F20110115_AACIIV horowitz_s_Page_106.jp2
9970ab306062df4da85ff758d387bc92
49694f5152802939bd5990dd64d842477b3abda8
1968 F20110115_AACJNB horowitz_s_Page_039.txt
2c81f84ee4f666ac788f7a2bda53da30
94d78c830ed51093e414ceac0e41ef2cca1b52cc
108753 F20110115_AACJMN horowitz_s_Page_022.jp2
a5d733977933de04e505df0fd238b13a
ff569d87cf2a4fd1d35dabc66f43f1e1c62c90bf
80087 F20110115_AACJLY horowitz_s_Page_150.QC.jpg
9416489ea7aa7054c698fe11b148f3bb
2a464ac8e459233330f7155b2d5c60f3cd4fdfe9
40701 F20110115_AACIJL horowitz_s_Page_108.pro
c58b07a32fc9a28450a6300b45cf7f08
b3e2e6a061f2975ed3dc004047feb8a642f6d32b
41887 F20110115_AACIIW horowitz_s_Page_179thm.jpg
0d86ba7c450a4ea670f10433b760cff3
cecd456d5603f0476bf0ab17f91481f902e572a7
25195 F20110115_AACJNC horowitz_s_Page_157thm.jpg
8ae9d72b158ea13e56d9a08c6689d9de
0121dc40c0bcbc0fd28f7cd4396883b451fb25cc
14935 F20110115_AACJMO horowitz_s_Page_016.pro
974af1bb536575004f060969c9d2683d
56485e485f90cf3e16a8c6881e6f7ffb2edfd9cb
85296 F20110115_AACJLZ horowitz_s_Page_091.QC.jpg
808bd31a244eaf21f523cf9cef03a39d
2357f0b002e133e244efde98834659b511a639c9
93357 F20110115_AACIKA horowitz_s_Page_031.jp2
437b4e6bd1075c2649eda61e65591f45
c882e2f4c741181f350fa19ba021f594ed14c1ac
F20110115_AACIJM horowitz_s_Page_192.tif
335115daade43aff6b18cba398b6549a
31b432c9410161262f0e50f53c779593ea3448df
1569 F20110115_AACIIX horowitz_s_Page_032.pro
7a683527eb58978ef622a4f614b3d09e
4a9e8c9358b5149aed35d1c1b510d7d3466ccc06
80917 F20110115_AACJND horowitz_s_Page_149.QC.jpg
9f3e959d232d965e2a445057aacdaeb3
de5e8a6087836d5829ff195691825cc989180599
F20110115_AACIKB horowitz_s_Page_136.tif
f7e0f8fde1384e65c96919b4e16aeb38
21f704a25d20318dd745975cd3c00bd8ead0380d
F20110115_AACIIY horowitz_s_Page_070.pro
c81b4fd4f93d96a42b8b716087d04b5c
527238f1a4506b3e4e674183391a2e25dd1a50ab
242056 F20110115_AACJNE UFE0008378_00001.mets
d6ba1e4ca858a75dad3830c5b49c9d21
97e197d9032cf748f39889e17365f5d970878381
48548 F20110115_AACJMP horowitz_s_Page_083.pro
f6ea1d9b3a6ffccaaef6a20777d21145
96d71986cc6ca54693f64a49876031c866825d7f
74775 F20110115_AACIKC horowitz_s_Page_115.QC.jpg
38cb6be72d61bee57d2653fb57a402d4
8703e2d2c1c45f95c030391419e42a4da4c15d9c
43098 F20110115_AACIJN horowitz_s_Page_156thm.jpg
dbc91c4f97acf0a83f5f327c5c22d4d2
fad04a259c237de634d29b33206a65e733b307b3
1735 F20110115_AACIIZ horowitz_s_Page_067.txt
cb85dce5cebdd871fdd9e49518216b4c
55df61f98a6a6fe9066d066488f2d6941bc8f512
1942 F20110115_AACJMQ horowitz_s_Page_181.txt
e6110e04ed480896e3d2fa079af23e56
e00b0095b920c7ef79c36d8514b9e1c90ae88daa
93664 F20110115_AACIKD horowitz_s_Page_194.jp2
fbae90f4f4e3f962dad6afcea16deae8
f14d02b334f10e48203853f00e6d9159853feeac
112451 F20110115_AACIJO horowitz_s_Page_094.jp2
4663f460696adc306ed074d5aabcc17f
7ff9ed2393d3c0b5854a1358a1eacbf1322b8b77
1462 F20110115_AACJMR horowitz_s_Page_077.txt
59b8db518de84bd7acbba7582f7b7814
4d251beee33fb4e7e065bbda6988adf2ab4a6920
209858 F20110115_AACIKE horowitz_s_Page_169.jpg
4bdd92318dc2e337fd9763e9dece8a0e
bf9004e15351095591bd9ebaf3275612bbf9d13a
23941 F20110115_AACIJP horowitz_s_Page_099thm.jpg
0c4da0621ebf08c7888e640675f7775f
50a4035c2883bd512353a70257bea3fa72b29273
51137 F20110115_AACJNH horowitz_s_Page_001.jpg
ccadda8d241fe2255af9f3c5a1959ce5
0fe520387df8f856dd907ec76228b9ce2ca0fb7f
83068 F20110115_AACJMS horowitz_s_Page_075.jp2
73904bdf028be51254294d308148056e
4df4acf112faf9a9643b2cc4766318147afabd4e
920348 F20110115_AACIKF horowitz_s_Page_177.jp2
ec0ed632f666f8c5e82e65604d379085
fd69ea3e48554ec829018394e9da451cfd560401
859 F20110115_AACIJQ horowitz_s_Page_186.txt
68c7ed53b20074296668d4d944365e6a
1d08bd3257c6bc45340d0dd208cddcb6b5e9e6c6
241792 F20110115_AACJNI horowitz_s_Page_006.jpg
1c0fecb24c1c61713f2d0c8d4acc283d
b9b87728fba240bb0f8cd82ccc5b2099ebe2f3b7
F20110115_AACJMT horowitz_s_Page_194.tif
5d2338588bb8f85ff8fe7b1c8ccfe794
7a94d360bd38ff716c7792d84932b181e06842bd
153522 F20110115_AACIKG horowitz_s_Page_074.jpg
fcff22926d946afadec2656efad3526a
84fc93084d5ae90a14dbe3575753d0b5acb38f16
F20110115_AACIJR horowitz_s_Page_121.tif
73f3817c7fcb0aa88410964907f8c4fd
7bacb8a8900255345d4ca46b3bc539f7e12e2cb6
362560 F20110115_AACJNJ horowitz_s_Page_012.jpg
9299eb6f3bd1b4e72079d44af87df6e0
97f2771ec55f79f9a3315eccd2a56df156591460
42127 F20110115_AACJMU horowitz_s_Page_106thm.jpg
9436366c807ff7b6996a533b4521d603
49685af8924474031b850b35b1eca63e4dfb6d50
70547 F20110115_AACIJS horowitz_s_Page_016.QC.jpg
8229fc2a947b6131a3b379808e5b4dee
3b0d31cb1ff4073adb8e518401e7cd63b7de96d7
46464 F20110115_AACIKH horowitz_s_Page_099.pro
d31c4fe4b5fdba9105216720b89a097f
d85210a0a47cdefe10f019ac3004495f151915d6
197381 F20110115_AACJNK horowitz_s_Page_020.jpg
24291f2da47565a44f4e1bc1a4132f55
71dc46f8e74a6cbd3009ae24148a055217fee5f2
31612 F20110115_AACJMV horowitz_s_Page_120.pro
c4949daf61fb2a9d2cb8fd0f9091d3da
b5545f51abd6778106f5990a79dbd22693894e55
1199 F20110115_AACIJT horowitz_s_Page_122.txt
fb5cf119bf5ea5e9415ac7098848f982
6863cf87c4ca715552c063dc2711875bf2bb152b
F20110115_AACIKI horowitz_s_Page_100.tif
7023295672bd1e1dc6b0ec6f8ebf8052
9fe6b64b13623151e7c0b0ca0f8f0c98af022989
176458 F20110115_AACJNL horowitz_s_Page_021.jpg
de1e0b5c5ab1630e295fa5d5d083ba5c
1dbfe5810d9d49c134f85a098f940f22166a04df
F20110115_AACJMW horowitz_s_Page_124.tif
1e40eea92c8b3ca4aa154268465b846d
1824664c22e1dd1ae2bd7764a3e492f55e786bc9
139796 F20110115_AACIJU horowitz_s_Page_202.jp2
729c41d5beac7e141b914d01b62a58a2
a512843a9f2f6c05df127a29d739efd2ef6c0473
45722 F20110115_AACIKJ horowitz_s_Page_042.pro
7abc90f6654a0f820ef2426971fc45b0
e4d7793d8c71e9994138e513d6ecc0e65badfec9
138753 F20110115_AACJOA horowitz_s_Page_121.jpg
d49c6def7d0838884ebcf4233743d4ed
4444f1dfe9ec7fe79bf678f2211662b0cd3d155a
205428 F20110115_AACJNM horowitz_s_Page_028.jpg
764ea35d0ea380b072c6b6c749476b08
575a27db8f4376cc8ba0cc1df7d191762f586712
2462 F20110115_AACJMX horowitz_s_Page_198.txt
0ecb573ce4fcfb8d42842e3750e0c2f6
4b325d8226d33d653eaf0e65a6c5645859fa234a
111774 F20110115_AACIJV horowitz_s_Page_166.jpg
d44a85cf848c30ad41c8fface6a1a378
73c9272b01f6d40f9bb192c94341b9a68fd5942c
37064 F20110115_AACIKK horowitz_s_Page_154.pro
9f27a3bca48cf384a7291fbf7e7546c3
ad8aab60ec1839ab429792cc812b7ebdf19e6c78
178866 F20110115_AACJOB horowitz_s_Page_123.jpg
a10caaeee4967b174ef072ffed302939
900f0bbecddf7a740ed9b27035240acae06a1d8c
143175 F20110115_AACJNN horowitz_s_Page_045.jpg
45d3d5ac6bb3f336635020518ee69962
d7d5d5cdad422efde436a8385868ec4d6b9f6b9b
44967 F20110115_AACJMY horowitz_s_Page_006thm.jpg
8072b313ca9f95b77ed15faa3ab62675
ad9366e012e71348fbaa7a1068987c84fc159502
81145 F20110115_AACIJW horowitz_s_Page_158.QC.jpg
e00a56ffd279ac790114492c2bfc4fdf
09186d2b54dc2645e27d4295b57b18b263cdb420
82193 F20110115_AACIKL horowitz_s_Page_197.QC.jpg
cb3d4ca4889e463c81f152e1a4baaedb
cb3a3e971862f494412adbc48cd5b71dadfa908e
191817 F20110115_AACJOC horowitz_s_Page_129.jpg
886102fc6e199ae94a8727050a8318f5
75d576ac40e7581fec10b44a264ec1fe6d797b63
214819 F20110115_AACJNO horowitz_s_Page_052.jpg
d05e6c13602c1a13757f6e5030cdfaba
a247caafe1d5482c1e03d9a5c82c4bbdfae6e6d7
189935 F20110115_AACJMZ horowitz_s_Page_185.jpg
f2f22deb2535697e8adb7218bd84d83f
b173ef2a0966987cfda19cc99798f47c8b3920de
42427 F20110115_AACIJX horowitz_s_Page_126thm.jpg
bbd162890b7ffef9288ac45f176f5bc0
6ef43fd7d7e84d4c711e31c9d0a9c7b0a6eabe84
23599 F20110115_AACILA horowitz_s_Page_055thm.jpg
05142e4eda212bd0af7620411126568d
17c41f18ff2c67bd5d40e7b47522dde8e97ac537
133649 F20110115_AACIKM horowitz_s_Page_203.jp2
adb54437cad6a1f05634ffbbe85cd2f2
2844dcc715a54eb87c872a660ec1c3e7ddb3fac6
176128 F20110115_AACJOD horowitz_s_Page_131.jpg
be253517ec265c4c8a590a970784dd51
f2a0965f97ad6e3b8258d1e54b19b68a261a655d
227674 F20110115_AACJNP horowitz_s_Page_053.jpg
6a02f9841a329c20f6ec9bc9291b7b61
695721f508ff0c0b9a37ca981505099bd870e1dc
1051978 F20110115_AACIJY horowitz_s_Page_004.jp2
3b8c8e82e8732189203ed18b0f209f38
58b6f10fc72dad1d8db0a14d20416e308685680c
44107 F20110115_AACILB horowitz_s_Page_049thm.jpg
412c14a827e6f2546bbae967ef9b5dbc
607973ea5da6436bea0bd0c646c8e8bebf29e349
17470 F20110115_AACIKN horowitz_s_Page_045thm.jpg
7ca7a7570e03d1d81bf216ac1eae5169
93bac7202fd0b30a506a3402620083c7bf4069b4
136071 F20110115_AACJOE horowitz_s_Page_136.jpg
b74f67e2098ca45fcf2d208ac65fc7f5
811bb7061b5e3a2ee70ea7855e3e343f54a159b2
30510 F20110115_AACIJZ horowitz_s_Page_162.pro
7c53481afca8ec40f069f884776b9b28
afc8959a372a8bb1ebc23f2d350ba95030e70fc6
F20110115_AACILC horowitz_s_Page_006.jp2
b0aabf241cf4835f30cf4c0510a3ce3c
0b21645d96b5f90a35494ec3f2ec5479625a71fa
183460 F20110115_AACJOF horowitz_s_Page_145.jpg
12bbb969c30d305fc9bdf386bbcb7480
3b2edb082108020ed567b4c5aca72a1682cbd4f3
180942 F20110115_AACJNQ horowitz_s_Page_069.jpg
46bac9c0291e4fa50727728a97003a45
dc4226c37b98d0387e0d4a6cc4ffa004aaba6716
211074 F20110115_AACILD horowitz_s_Page_038.jpg
12af1e54ce9dfae3ba3d79fc25fa3793
7f3cc5d20f72f6c05015fc808a3cbb024474a98f
102216 F20110115_AACIKO horowitz_s_Page_164.jp2
367c411751c2a9a95c424f05797837a7
d909ffc92aa64740d1369477090d8b5378ab895a
210334 F20110115_AACJOG horowitz_s_Page_147.jpg
7eb92da288d9b78dd0e6c0e6b6e001b7
0121fe9c8c9aa274e30389d8ec0d972b71c28b51
119936 F20110115_AACJNR horowitz_s_Page_073.jpg
3ed271465d509a828eebf73ecdcef7ed
6e375fb8b4bfbf0637f423723a3a7bfa36c7c023
109891 F20110115_AACILE horowitz_s_Page_040.jp2
0c3c4c88cfe625c16908c12a13c50eca
e543f243f29de5c13e09f578c3f203c8479348a2
F20110115_AACIKP horowitz_s_Page_059.tif
70cd92ca7d7a69104df994fa6a7c91d3
47abaabf7ee8701b9237ea6bba8f3dd6473df090
218255 F20110115_AACJOH horowitz_s_Page_148.jpg
05d1db3a8548f5ae20513e223a1221a2
8f8a78b71be65b2cdef15b00c30d89164a13712a
180664 F20110115_AACJNS horowitz_s_Page_078.jpg
e74c6a345fb0bceb1176a7c6dd4d52a2
33b1e646c2c887c8e0c8c15fd8aa592d6f699d13
F20110115_AACILF horowitz_s_Page_086.tif
90e89e13dc4f8e2b352de7b684905736
e382bcf48e0f17955d68d90542e6597db3ac3e1c
48221 F20110115_AACIKQ horowitz_s_Page_102thm.jpg
ce70d8bd896b3fd3be3d7bffd8b2b941
267515dc75387418593138471b6e1de46ae80755
211157 F20110115_AACJOI horowitz_s_Page_154.jpg
6e6e6d7ec91a77a1e24f0d6453eb647e
a3f9768527c43050a8e906802adcd1fcde4a6960
202244 F20110115_AACJNT horowitz_s_Page_083.jpg
d8b1b14741f17803235677df3576c204
47ea5863e30cc444f3bea25a32f883134a7b157a
74624 F20110115_AACILG horowitz_s_Page_064.jp2
b7b087034e32bec334d9134ae7979708
333db3bef323094d081ba6a8820b3d0e4d5fe6b1
1576 F20110115_AACIKR horowitz_s_Page_036.txt
2af14e3af48153a84c2056648028f15a
7f096d7223079d5a70a8e65e55132cd8db87c4e6
211763 F20110115_AACJOJ horowitz_s_Page_157.jpg
f83df087e14a8ea4705c6a2796015159
63c135b96e2628cab7c7d695eb7f0108002c7b19
167497 F20110115_AACJNU horowitz_s_Page_085.jpg
091b7781990ee633bb8a723a04075056
6a2bb6d58f58e2de7cc20ba3cc4a4039194035fe
43601 F20110115_AACILH horowitz_s_Page_161.pro
f419f9e5e3e7897ee6934ffcece0cb32
f38e1beb013012a512dcd3c9d0c69ca5ff1b5312
F20110115_AACIKS horowitz_s_Page_128.tif
a1dbd331409ac94f71f15f4ad9167044
82571aab6266620fbdeef6204d2bfb8e82188316
197414 F20110115_AACJOK horowitz_s_Page_164.jpg
80f60e6c58c81e9a53496ca16efb1e73
bc28e35389908a51bc949f7d79960a450bde7cc8
170649 F20110115_AACJNV horowitz_s_Page_086.jpg
b6e851f896269b6ce223b15cc414ce6e
a68bad165ed0bf0f3dbd673aa5ffde4e4983df9b
3762 F20110115_AACILI horowitz_s_Page_003.pro
de95e43124118ef93730153ee5d0d04d
4d61350c61610e070ec28d81c82095b7480fc1ad
40374 F20110115_AACIKT horowitz_s_Page_177thm.jpg
4717f8454c0f016201157d80e41c940c
20f61a02ee0f9c242e013939796b8a6f3605127a
144822 F20110115_AACJOL horowitz_s_Page_165.jpg
ff1da8728135910259dde9d6257ca239
c73a9fc682821d8564157b35726655f967826076
182911 F20110115_AACJNW horowitz_s_Page_101.jpg
739d9441720f725a5edd00c4dfec8132
1dbe6fdeddc16b51598a0ecaab6608f998567ac8
165726 F20110115_AACILJ horowitz_s_Page_124.jpg
6ea741dada99bb1dd98b1e8a34187f31
c6e7555f994a8aa527cdfb9a78c129b878d29ec2
272674 F20110115_AACIKU horowitz_s_Page_196.jpg
bc2039ccbf7fe7b59ce11853ab8a287c
38870ace1d04af2cad823fe2048eae4fcd021179
756161 F20110115_AACJPA horowitz_s_Page_061.jp2
cf2dd29d7d6d5c5b0303903ca7b785ec
d01880ffd039e25fc3d6cad96b5c1b1db50ce8fb
182304 F20110115_AACJOM horowitz_s_Page_167.jpg
4a428fa9fff21ffd62ba50e0173afe75
94f06fd414a8882f81606aeffe3c7bca2c09660b
245884 F20110115_AACJNX horowitz_s_Page_102.jpg
3eeb01ea1674646c66644d55e49df6f5
a8d0cc2448b7d61e3d4db4cb5e89fb6b41e21215
23561 F20110115_AACILK horowitz_s_Page_025thm.jpg
8c8079928d4756f21914b78647b93842
7bf4e9b35d52ab63fe29d0cff2c8181cb9e8ab33
138038 F20110115_AACIKV horowitz_s_Page_046.jpg
b3ed3a02248c1ebb24c14a2d1b948a19
d1932fa77c7324a60da16293750db2087003c421
53178 F20110115_AACJPB horowitz_s_Page_065.jp2
ffb8a6fa9d3352776b4fa98c3bed99c7
b7a1cfcddb1b4c575e092b83ced11b3efc449d4f
116284 F20110115_AACJON horowitz_s_Page_189.jpg
bc1cbe9ba1a4e32dfa7c34168c4183c8
0ccfd3d6a21ed210193baa62d6cb1b34a3540438
193828 F20110115_AACJNY horowitz_s_Page_113.jpg
fd664bd598a14d834071d096a98e2940
e97b83bf3f454e40bf268f66484918ecfabf6098
1955 F20110115_AACILL horowitz_s_Page_141.txt
22857ab541fb9ca64584e73d984cc19f
0f0f6c8c99d22d916684d48cf9a0186796e019f8
67408 F20110115_AACIKW horowitz_s_Page_031.QC.jpg
ed1780ba88c5e6be4e7511cd460203d6
fcac1190f0690c0e4382f0049403fefd881b8fdc
64797 F20110115_AACJPC horowitz_s_Page_072.jp2
a3f10899d238ef370205fa677bae7d10
052a0fc0add251cd7b134716cbf2d2b28716e991
103541 F20110115_AACJOO horowitz_s_Page_190.jpg
21bb91dabf71ce802295f42162fc1cda
0c0a3094d2ff8db2e2eafc2de4dcc029a20df493
213641 F20110115_AACJNZ horowitz_s_Page_116.jpg
8ad76349fef0ed2ff9867d2e407a8b2f
35644c4a4142ebf9a7d14f1a436a0c2282488463
39267 F20110115_AACIMA horowitz_s_Page_175.pro
6bc98a12cc1d42d4741c5a6fd5bdcaea
c985765eca1ab7332861f25872ac0e324af026ea
161970 F20110115_AACILM horowitz_s_Page_125.jpg
bb8e81ae396266ca785686014f9531ad
ad5a6b958272053f292f257fe93193fe50551d6e
F20110115_AACIKX horowitz_s_Page_074.tif
5625ea43f1b7d2b66761d67ec58897ed
5a4797e8e5cb96b739ec41abe010b25816b29197
746983 F20110115_AACJPD horowitz_s_Page_077.jp2
73b12213adda727c014b21315fd386d5
540f4310187912eb1c445c9b48979953a673ad8f
239914 F20110115_AACJOP horowitz_s_Page_198.jpg
0d309a729e332b4476be6c387db69b05
8053bac1b4b0b40e5fb1170d1515b0c1a9811ac9
22729 F20110115_AACIMB horowitz_s_Page_166.pro
38d0fd4b9c1f19df55aa6647bc9306dc
0f482c912352b5ef4e57be2a9269bd9bfb6a9fa3
45519 F20110115_AACILN horowitz_s_Page_170.pro
bf71636611eb3ce421e75d80c18165df
a4574eb9546c33b3e034c8449f35778cdb45ffb7
47169 F20110115_AACIKY horowitz_s_Page_151thm.jpg
2cb2943563b3d9ffe47f9e93b451bcb1
9ab481679936f7d3d2355a1124b831c8c84966ee
784432 F20110115_AACJPE horowitz_s_Page_087.jp2
2f42533bb2179a158470a530cef0a8fd
ed0ab5c7b9cbbe1a8fdf777e4e8a22d4a8193c2e
69308 F20110115_AACJOQ horowitz_s_Page_209.jpg
f129886ed70209bb6e0b4b6e2a3260a9
d6448c8ca80d06d20a4afd400aec36d42f1382eb
607028 F20110115_AACIMC horowitz_s_Page_186.jp2
fa043e1c3e00f22ab0437cc46e5c9c54
351a2c06661e455a28ccdba297d76f812fe1c65d
19919 F20110115_AACILO horowitz_s_Page_128thm.jpg
7a98a8c1e9fd4ab751cc964af858d5e3
4f8bf3d54eb7f980725e88cfb2a640fd52e6c9cb
906170 F20110115_AACIKZ horowitz_s_Page_043.jp2
37f1962f27f1e2800392d30957e5c724
f03b15d165547019f9e23b547308064c1574aa31
825482 F20110115_AACJPF horowitz_s_Page_089.jp2
bf2d613cba6fa7683095bc8e205ffef5
6d8772bd7b9f6524b17cf9f34e1c2ad2b13f5858
1606 F20110115_AACIMD horowitz_s_Page_120.txt
1be74304ab2376fe94f93ac8a64d191a
cbb98d2bbbf082f5615890c5a4b0b60ff0add358
945473 F20110115_AACJPG horowitz_s_Page_090.jp2
962ee7eeedc0b0509c6888c860dd80c3
2671af275b3caf172ecf230f5690251299f5ee55
1051977 F20110115_AACJOR horowitz_s_Page_007.jp2
b34df4882d4c71dbb76c55515fe52db1
73428bef6a6b6a639ae93dc4623d3e47aed40c4f
879262 F20110115_AACIME horowitz_s_Page_113.jp2
a1c9d69a6005dbacc607831596057810
f5ecb74e6eed79ca060c27d48f8f23f100dc4964
40947 F20110115_AACILP horowitz_s_Page_186thm.jpg
f1ed35db39b0f7c0b18eabfe4de36503
7fbf916fd1b076b0ba738649238679716d2276ad
880379 F20110115_AACJPH horowitz_s_Page_092.jp2
b1e9b1c0bd8025deeaaa6606cabee055
970aa4998036346a90b9775d73ce833e82aff6b8
1051976 F20110115_AACJOS horowitz_s_Page_012.jp2
683c69cde8c11896dbf3e89aee7207a7
018b0205b4797bccc35c89a3528f6c309af653b2
1505 F20110115_AACIMF horowitz_s_Page_049.txt
318dba045646daf0635b1025fe0472cc
94582caa66018975c1404ec6c7a466ce233d8f7e
60145 F20110115_AACILQ horowitz_s_Page_193.QC.jpg
209ba23e841021d43361c2e0c4852f8d
9965540bdbdb7a6208a913b291bb07f789d56022
719551 F20110115_AACJPI horowitz_s_Page_097.jp2
15225f6ab62aec55f9c2d45eee9559dd
214153d4ea7b46ce10b953914deb2695c4797c18
105918 F20110115_AACJOT horowitz_s_Page_023.jp2
34a9577a0f1bf66da68ea3236be8c5e4
28dd3bcd8ede61bba1ee35c3aca53b910efff808
203212 F20110115_AACIMG horowitz_s_Page_017.jpg
23bfe94119624b5e0f4134d3bfbe78b6
dabea9fe97a4278f9864883fed02904db3054a96
85294 F20110115_AACILR horowitz_s_Page_100.jp2
2d42de92d30ba1fcf1efb8c7dde13ae2
255948a9bee001c2b0e39a1a496a4c7de6837a6d
1051888 F20110115_AACJPJ horowitz_s_Page_104.jp2
50a0015c25af57b09dca3e9c097b468d
2c298f5e79febf58e7c4997a41c294bf00f8d8a5
104764 F20110115_AACJOU horowitz_s_Page_028.jp2
77a076b111c72c196bc00a48b73d222e
9d0672744ccfa7b3dd22c6aa0c652013598b238e
46326 F20110115_AACIMH horowitz_s_Page_025.pro
03fb997b300a2eee82e680a01698ba35
ac86c412431c7604ad2afb324a89031aca31ea44
F20110115_AACILS horowitz_s_Page_072.tif
0d1a03a3bbcc5b2d1bdf983dc2369b23
4ed855148c590d6c70923a106fed1280d06a546a
1051954 F20110115_AACJPK horowitz_s_Page_105.jp2
d200f45f3bb807ab36480cf86e222856
98b2a91f4a9d67853f7823ba866faff92c6789c4
611678 F20110115_AACJOV horowitz_s_Page_035.jp2
53c71c6615922968c6f9ed45a75e187b
de424f158452cd7336edbd9865a6a8ee785f4f34
595146 F20110115_AACIMI horowitz_s_Page_060.jp2
18f5bb6d71c5ff52ab620bd17b7c8eec
c8f6bf4ceb4e16b45503af40cb2949389f3e04ac
F20110115_AACILT horowitz_s_Page_158.txt
2ffc7fe010c6655d37c8bb9e83ce56fc
96ff5fca0ea71d8fba77e502aa861cf28cedcccd
933390 F20110115_AACJPL horowitz_s_Page_135.jp2
d295758b6df945a367d0aa3611ed640e
9b20acdc85d4ffceeb378098ba471c9e59381630
109742 F20110115_AACJOW horowitz_s_Page_038.jp2
884f0deac932c99388ef7dee6e5048e2
79266bcf755d6fb6b36539c90d12436b4ede0425
1349 F20110115_AACIMJ horowitz_s_Page_060.txt
cb08623968c7ac0bafabb0c80fc8807f
28efb46258b2edc0fd274b1136c6f4e068196f70
196487 F20110115_AACILU horowitz_s_Page_025.jpg
8512de7708a672ab2404db509b517543
c4bb75d1b0501830edc56473738bb6c0bd42ed09
F20110115_AACJQA horowitz_s_Page_048.tif
82d0e0ca6116942320976c5cc03c01d0
bb08a95c1605fa1f82c8a565149293631f41805b
789587 F20110115_AACJPM horowitz_s_Page_149.jp2
6a1f1d3696db096e05118e62855b39a7
acd4df43b829d0270af734a943520acc232f61a8
106350 F20110115_AACJOX horowitz_s_Page_039.jp2
e427c9a3046290f76da8f4ae2e1b2836
dac4f66840fd3a89de1542f6a7ef336560cdb272
84008 F20110115_AACIMK horowitz_s_Page_078.QC.jpg
97edadaed6328b33a92ebfd2fe75cb8a
95cef2ba5c92cd384ed3434164d9bf5f10ef1def
73383 F20110115_AACILV horowitz_s_Page_020.QC.jpg
3ff073d92cf97524d2da2c7e91d88c9e
56e5503cba4e5fc563337142e693269726417663
F20110115_AACJQB horowitz_s_Page_062.tif
bfb5203e069ecba4bc4f382e6218fe4e
bcb99ad25e656b4a165ddb333fb7bea5c63ee629
67572 F20110115_AACJPN horowitz_s_Page_162.jp2
edd90f00758c3d05f78cba2b07c66a0d
f50db2d3f7ebd8151f0c87e8785cc4412a5b4104
73243 F20110115_AACJOY horowitz_s_Page_046.jp2
774bfe5cf2df269b1b86521fce64c6bc
be78a12fff7cf700c56eaa5f329b2a413b52b087
F20110115_AACIML horowitz_s_Page_082thm.jpg
3e6f854c5cf0205af43373ba146fd81f
ba6653051902d1e2e38d317353c7426649a32116
927772 F20110115_AACILW horowitz_s_Page_142.jp2
ffe39b9a781fac07063b74ec0456b859
d10ac1fe0aab790451e04e8ae508c2419c822afe
F20110115_AACJQC horowitz_s_Page_070.tif
457249cd55a9120ed4b3630966cc5ce7
2980b4423cd849f64de83a690872cb64a3398dcf
34543 F20110115_AACJPO horowitz_s_Page_171.jp2
e548719717d9066de4e21e7539d92e9d
7cedd785e0165523470adef5f2c2b6ef2bea5200
734169 F20110115_AACJOZ horowitz_s_Page_049.jp2
c7f343a930ae9792cec03aecf14397f7
09d6d9228acf16c06e56c2902689272147c107f5
89215 F20110115_AACIMM horowitz_s_Page_043.QC.jpg
e8248bb559d9bc70142ae811e00fa495
715592096e61c27279988342da35800fe5988dff
7512 F20110115_AACILX horowitz_s_Page_001.pro
e5515b88deb9933e4e7f3a80d945f50c
deae496c8a32b2c85251eea4f6ce684f8b89b29c
57171 F20110115_AACINA horowitz_s_Page_012thm.jpg
6c8f18fbddd10927e89511dcdcdaec51
09343023be8cd23e3402206a89fc8dd63d974d7f
F20110115_AACJQD horowitz_s_Page_090.tif
4a79dba0e8dc1924196e272efd0a9f7b
3457d43f1230cb8c83fb63778e41e87681c0313e
364357 F20110115_AACJPP horowitz_s_Page_189.jp2
5c10ffccd0b9ef1410cf39294c839155
f753260af32408756406efcddb2f1a3b4c179ebb
179523 F20110115_AACIMN horowitz_s_Page_070.jpg
7c9a0f26cfdb735be245eb55019ff992
c9e7fe98fdb4025ccbc174220b58aa7f0b02e2ae
F20110115_AACILY horowitz_s_Page_185.tif
d6a4071ed34b74301bc6d832e5353bfd
317922497a3f77eb1af599089a06df7e13a2aa86
1417 F20110115_AACINB horowitz_s_Page_080.txt
88e03692f2e0f067d4a6fdd9423eaf12
6d6d47f9a1abac06fe41ff9d0f21bdccd586f57a
F20110115_AACJQE horowitz_s_Page_098.tif
f0442637855366920fa98f72ac3aa262
6ca2b76332a872fba169b94ab8d6497396841e73
317114 F20110115_AACJPQ horowitz_s_Page_192.jp2
a853163c5d4b19bb73da6bdcad20bdf7
0090e09bdf6cf702f5e0f9de8100318ba69f4e83
1742 F20110115_AACIMO horowitz_s_Page_115.txt
a8f9c289770da5d3ba49b63e3c2ef974
89e6d0ac192edd035957c7a162eb5050ff23f805
729218 F20110115_AACILZ horowitz_s_Page_165.jp2
475901cd3588f72d4eddd31e33e6e74a
a13cfecb716f9fbed08e7d1915345fd64a4d1e29
17704 F20110115_AACINC horowitz_s_Page_146.pro
60fc4880a3aed1a43714421d805c8870
314bca63bf4f3c15b24843468ac0fe87d87956fe
F20110115_AACJQF horowitz_s_Page_106.tif
3664207ef2a0a1142cb9bed693c769db
6feef7107f83e0f9c20c9e57f24484fdf7501e50
127548 F20110115_AACJPR horowitz_s_Page_199.jp2
f0af55a295949f4b9caaa26ef4f01d3e
152ff6cf6a95c7c40bef2638369a46e6c97d9e33
20891 F20110115_AACIMP horowitz_s_Page_015thm.jpg
1a6b5c69c1014d8e294be9f1806c8f39
b791308c9a624b64ac619dff2894be74d32dfb08
F20110115_AACIND horowitz_s_Page_159.tif
55e75063eb6643464bd7fb6465625109
e33c5b8d684d7ec277434cbae07dea7797564f8e
F20110115_AACJQG horowitz_s_Page_107.tif
d2d63a5b5bd7c14bcd95d99be1c39d98
79934d8f78e0557071f1be205b9fa1191c191568
146984 F20110115_AACINE horowitz_s_Page_032.jpg
d22d8f071f98049679b3716d79a598ec
d1c87de053d1fec41abfbb55eac42ca54d21a9de
F20110115_AACJQH horowitz_s_Page_108.tif
7f0ded807f12cdc4cba11683c41f8023
06df68892062e50e0795b84a98004be259417a2a
136846 F20110115_AACJPS horowitz_s_Page_204.jp2
bd13d3d0ea674f35b8d106eecebd1b8e
c8fc95485801d80df473dd96d6343ecfacdc125c
23109 F20110115_AACIMQ horowitz_s_Page_131.pro
a282fe9805d3324417ff9c0e271249d9
7d6ebe69189c2d485aa74525d6c162ae8b6c05b0
F20110115_AACINF horowitz_s_Page_176.jp2
00a3bf51fa1352b52cf1b08a1617f088
5b3da06eee8467ca0f65c8cf290e2abc320d31ed
F20110115_AACJQI horowitz_s_Page_109.tif
52b741c61983fd21552e064fd28eeed6
ddba5c8c9adf2af4a8c6292cd6eec6d490659a19
111659 F20110115_AACJPT horowitz_s_Page_205.jp2
458c63ea77f03719bc1e279182d20408
2f973b6d38b7d88bcf2a5110acd8bc8896faebc1
131444 F20110115_AACIMR horowitz_s_Page_159.jpg
7cae780ea80b5cf77d6b85eed2a3e070
9a2a957db2e4b16498b5bad2973ad6cd6fd2c956
23925 F20110115_AACING horowitz_s_Page_028thm.jpg
7477efca1e08f4aabf09da9cb0f5b4ab
bc83918f28006987d91b52d4b3933d23c20b9223
F20110115_AACJQJ horowitz_s_Page_110.tif
f32c7cc4e3f7ece88172460a7a480de0
45d914adea14fc05e5fbf2b50fdfa13d3ca24f87
F20110115_AACJPU horowitz_s_Page_003.tif
5459144144016c2fde6bcb611a27bc66
a4a4f2f1a6d9b78aa497f8ec23e4fce06b857e41
80799 F20110115_AACIMS horowitz_s_Page_175.QC.jpg
3ce1e4305c47a9ed22d310802000990b
7855f086bbd7bc185acc174ecac012d5cbb6ea49
104801 F20110115_AACINH horowitz_s_Page_083.jp2
d6eb0ea702bcbd066e7cd7d2a4dc5f59
76bc113bd33c29f03904adc6a8cf6a18a1b33fc6
F20110115_AACJQK horowitz_s_Page_119.tif
d169c6cbfec8bdc4bedb3e23f9ca362f
2a938066f40f674a69f713fb9fa9a86d758da13f
F20110115_AACJPV horowitz_s_Page_007.tif
78fe3fad78f90f6a800483ef15f0e1e6
d9e3782277ae3307af4d88c75f1f13139dc3e0e1
223804 F20110115_AACIMT horowitz_s_Page_207.jpg
7da61d7939164c9a98afe299a2c54cb9
3f7fc0d3e6f1380d58639aa9dd153a125d7b6dc6
33083 F20110115_AACINI horowitz_s_Page_088.pro
e9d39c1d32b9e5e7b705fe8ad9fee48d
1fdd507b8d831511f0898c47e9951907b88c76dc
F20110115_AACJQL horowitz_s_Page_130.tif
2ba5cc7744aea34f482843c336d1fd5a
df85587adda848e2670b09842c752d27a7e373fc
F20110115_AACJPW horowitz_s_Page_023.tif
ecfb04789659d53fd016a8b88d26969a
400163f3ae99409536323f89f405ddf98e558bf3
1897 F20110115_AACIMU horowitz_s_Page_021.txt
2547a174d6f3663e22061077f2723daa
f8660d008a7a86a8b87ce8b385a7a3c6ed6ac4c1
97479 F20110115_AACINJ horowitz_s_Page_052.QC.jpg
092f5fe1aabba6afa4d1908c41b4d2d1
a30bf7a0dd765470bf48957512463035db563796
34909 F20110115_AACJRA horowitz_s_Page_043.pro
ca65411d54f6cb240b45d1f061650074
3d6b11b93fb38ef7d398af92aee6dbd68150ab02
F20110115_AACJQM horowitz_s_Page_135.tif
174a131152b32f7623dc8f5e37840999
0c6d0d35e4d2f7aeec72c5cbc6fd70b0e042c070
F20110115_AACJPX horowitz_s_Page_026.tif
c7c5bbbc29f7950b7c4d75127a3fffba
1b2f7bac371ff08fb7bf8d7e68e0bffb6d3fb90a
85309 F20110115_AACIMV horowitz_s_Page_157.QC.jpg
1858e252e7f7147b4d4d78b08080acd8
fafa41b336b14e2f06b2b05c4d5f3c518c074555
F20110115_AACINK horowitz_s_Page_164.tif
9ff05f74cfa437710bf42cf2d19a956f
01dd7df15938380e48ebd7d61b52f7b3676c44c3
35980 F20110115_AACJRB horowitz_s_Page_044.pro
39debe84214039ef4e6bb83cc412b3cd
8b4ae764b173777f6c1799d8633bd5ffe3d67798
F20110115_AACJQN horowitz_s_Page_147.tif
efadccf3fa625af578a0908a983a96b8
4bca9c7256b3ac600f694c3168813415639ac825
F20110115_AACJPY horowitz_s_Page_036.tif
9e246d745e454a974c82137fcfaec225
144455ef0320b96e0448ebeb0e9c0c27036dc5c2
6170 F20110115_AACIMW horowitz_s_Page_002.QC.jpg
b29315aff9a663639c4ffec38271ea09
e037d71de2d8a9e169326d984908763118538f10
F20110115_AACINL horowitz_s_Page_024.tif
91be4dda87412b57deb5d39836259e4f
ebcbd262136c33a43f61aee443d383de1e45f247
28953 F20110115_AACJRC horowitz_s_Page_059.pro
b48499bac297a91c465908175877ff05
477f9674fbf3241245d84bee4e9960d2a9969de6
F20110115_AACJQO horowitz_s_Page_150.tif
9370d2ed7467e87cb72994f2fd3c232c
b61c1e441a339eed32cc2e68fbcfabe5d608ef2a
F20110115_AACJPZ horowitz_s_Page_047.tif
22d3637046417ffe59c7044de294953e
b6e62b7e49bffb569ccad588b166a1d2dd339bbd
81950 F20110115_AACIMX horowitz_s_Page_076.QC.jpg
632d46a058e570bcf1ecc0f97103f551
32a710e6b264b767b81a04377f3edb55c42dab0f
35287 F20110115_AACIOA horowitz_s_Page_061.pro
526a733c1ec76ec4e881ed231aef0588
bf2bc4609c8b3e4d3acf06e57eec20c3561441ff
221588 F20110115_AACINM horowitz_s_Page_151.jpg
be9447e035ed24df22a8428a6a09ccbb
7191352859cd4b8a2135e33ec7d1d2af37d5372b
35871 F20110115_AACJRD horowitz_s_Page_063.pro
c33cc5d85b89c824e87c7ffdd48de167
296e449a347af78bd0ff13304f1eadcb624397e9
F20110115_AACJQP horowitz_s_Page_151.tif
739a033027bc03bd7815e1e34803f82d
1d561bf2075209cd19cfe4c6021e683e43bc04d8
F20110115_AACIMY horowitz_s_Page_002.tif
5190f8fac1843ec774b93dc750c8770a
4c39e1b48acc57880f8bf0c5195acae741a1f275
74784 F20110115_AACIOB horowitz_s_Page_011.pro
f14d6196c66eb190a17cbf9ddf349604
54231eacd9fc9c38fc0ec1137a36580a15616442
1769 F20110115_AACINN horowitz_s_Page_194.txt
0c060ce3d0d80c0d0a590a0f7e5d8799
6a9f2bb23f8bdb9427578dce23c19c98146a2fad
40955 F20110115_AACJRE horowitz_s_Page_085.pro
965e6c7a76bab0ecbaf1848bee20f72f
59cfc2ac062c5e7159c9c406bbede5f86efe8e5d
F20110115_AACJQQ horowitz_s_Page_160.tif
63524e71e297707749508d4139e00710
6654fa2ed987ba2425e6e0314f46e9b9c6ea2d81
1971 F20110115_AACIMZ horowitz_s_Page_041.txt
7305796e4fb1547cfd0374211cb73dab
b4a06e7a18b3fbcc25079a1c8314a49b1fdb6d5f
128702 F20110115_AACIOC horowitz_s_Page_206.jp2
372225f3b2ef36d3621155bf5fae45d7
e100fb6f0e997ab07f5ca3d05b69a1abed2d4285
F20110115_AACINO horowitz_s_Page_018.tif
5b77b17c9cbeec37780e73d7c4986c49
34a3b71fb9c48fc8807d6bb405c8bf2ef383f2a8
32639 F20110115_AACJRF horowitz_s_Page_107.pro
726f147c3f6b560ab3d748aed35a5150
0b4c6629b62f1c4f6102735175c8f99fcb990e1d
F20110115_AACJQR horowitz_s_Page_171.tif
f3ec8dd8690b336801bf19cd749535fd
54466598c7d48e83ee5ec0c4e7ea6dfb6dae65ff
77338 F20110115_AACIOD horowitz_s_Page_198.QC.jpg
bdbea99ed6d5d530a1554fd4686daaec
06c4185d1ce9d039bf0cbd219cb706f8a49ef9c5
21971 F20110115_AACINP horowitz_s_Page_167thm.jpg
6874bf8339aee6616054ffffba648737
93586eb7cb67421cb32c0c74e7035d98ba1e115e
30607 F20110115_AACJRG horowitz_s_Page_113.pro
5418646b00de258bd1216977a3e6ead6
0345d5379d851535503edf7ebbdcda644a4ede7c
F20110115_AACJQS horowitz_s_Page_190.tif
4c482bf5886b827fd4633fb561ef6fd1
98c3400fc111cbc8b641a811a0d6c525bfbf976b
63734 F20110115_AACIOE horowitz_s_Page_160.QC.jpg
ea85cf33227b5c86c5be101bb5d5e88a
6cf63fa66bf60b6a2e3c6272e5501b50747dff34
82597 F20110115_AACINQ horowitz_s_Page_131.QC.jpg
66764a1ce501cd57a6dc090575c488c6
189d4836cbe22ea24249a0917e32061700ed08bd
44947 F20110115_AACJRH horowitz_s_Page_114.pro
7a4c30f9926c4595e74836251996b3e5
c98ea36e40259b2899b6c2a833040e7c7174e4b4
1799 F20110115_AACIOF horowitz_s_Page_031.txt
db59155107eda294b29f23e25a283b2e
6910f388a851248b1ec6a7cf938c18a11d4b26de
43028 F20110115_AACJRI horowitz_s_Page_115.pro
6a02cebbe951865720c7895a4d0856f8
8f4c50228878886e683b3be168cce45e65621b27
F20110115_AACJQT horowitz_s_Page_200.tif
19416ada094996f265b29b906b3c4906
d7990df7c7244e32f0162836e0c850d4d86226bd
133883 F20110115_AACIOG horowitz_s_Page_196.jp2
b09322b3fb901e519b1f097970039a75
b6c88b5c142380ded7eb65f35061f9a5fafddc8b
728347 F20110115_AACINR horowitz_s_Page_144.jp2
a573494cdee87bf1b78f6dc3e0c54fc8
1a3b18d051bfdbd8ab416df952c88649e244ce1e
32667 F20110115_AACJRJ horowitz_s_Page_118.pro
3eb5512cf822243ccdde54a6f4f6be09
70d8999e664735c02c89d8d26760ac23fb22aa8b
F20110115_AACJQU horowitz_s_Page_206.tif
69969965723169d7a057770f9db1b5a1
72da317fbd18adb616f604b6c079cac9c0ff0522
34576 F20110115_AACIOH horowitz_s_Page_153.pro
0fecc653b97ec2660844e30f3e1d0f29
86775b3fb10b4459898fab8576f36781b5a8d2a1
68908 F20110115_AACINS horowitz_s_Page_163.jp2
f73958a3d327a323e8f256b67b572197
aa3e66622fd01cf1bc3f175419a9b2b43009693d
37325 F20110115_AACJRK horowitz_s_Page_129.pro
edc6a8870e1a2df158a14c894fb39b94
1772e1e1634463a37822667019fae9b9bd242403
75787 F20110115_AACJQV horowitz_s_Page_012.pro
2782cc229e8b6fff74214b5be676b5d4
86efdc0936ba8ea2993584bc3d83591be6285cfa
14122 F20110115_AACIOI horowitz_s_Page_171.pro
98ace557809989ae34eb46d158a7d35e
e43a0c93da529b2ba5080ac268639b3780d09bbe
18054 F20110115_AACINT horowitz_s_Page_063thm.jpg
9157e9cdc3d1a3e0f3296ec1d3dc0823
4739e23114dba2cf7b952548e887328575bb07d2
30476 F20110115_AACJRL horowitz_s_Page_147.pro
974397bc60834bc81da02a3f142a3784
0aa25be68c964dabdaf74e86d0b9bcf8def55650
46524 F20110115_AACJQW horowitz_s_Page_020.pro
f113071603124f120f2ba298e2a1c5f5
2dde70df8f62a557cecaaa3e62d770e2cdddfc80
34993 F20110115_AACIOJ horowitz_s_Page_053.pro
1037982685f5c31bd05c70dd60ec8ce2
46f1f570f2f39f27f7abe33673c1c08c930e857b
1248 F20110115_AACINU horowitz_s_Page_178.txt
70b122c0317194ab7f1a015245cc2ac9
b857dbd496ce1b07caad1afc827569a69da07d22
824 F20110115_AACJSA horowitz_s_Page_054.txt
bd6f5a44bdb5080a7f400db83a74fc3e
b14adcd8296d250fbbbbb487872b8dcce137cc58
34910 F20110115_AACJRM horowitz_s_Page_149.pro
163ad436e2dea4441e2d86cde22bc831
c677d44c1260d5e8ff883dd47f272b8302a496fe
51964 F20110115_AACJQX horowitz_s_Page_024.pro
080ae5b72d7a59cea615d831a3bce487
382b9cc3eeb61a7884522df975e906df4fe4f36b
F20110115_AACIOK horowitz_s_Page_156.tif
73475c31c551cd70e4314535599ca15f
6dcff8f1551e96f4d76d39d3a5869d8c67321cac
48035 F20110115_AACINV horowitz_s_Page_052thm.jpg
0f4e3a774b9767bf7a51e648da431130
570c3b897d465cc3499714dc63e1f5e0e2205927
1726 F20110115_AACJSB horowitz_s_Page_062.txt
15937d8ae6644e2f694b362713afe36e
d0702293eaa12196270a88fa93da729d741184d2
52133 F20110115_AACJRN horowitz_s_Page_157.pro
0dba7e92f8912b12cfc0a30dfb7106ff
488312b7268dd90f5b119653817c6e2771807449
51186 F20110115_AACJQY horowitz_s_Page_034.pro
a2aa189d315507fb25a3e2ca6394facd
631d1c1a4b944209adaef606ace0f3480b923ca6
22735 F20110115_AACIOL horowitz_s_Page_114thm.jpg
a36e7c2be2a15b21d3c817ca66b910df
bff8a934a13d5b26297fe0aad582fb95354a57ab
F20110115_AACINW horowitz_s_Page_117.tif
fe9a0ba3557be5bae29b33ecd9da7c60
3fdc193dfa5a3048d4468e5659e1819eaab6dc80
1759 F20110115_AACJSC horowitz_s_Page_068.txt
9c98d85e426276999a967dbbdbe7e19a
7cabc2d7c1a96b68f0971686a209ba8347dc3587
28908 F20110115_AACJRO horowitz_s_Page_165.pro
8339e41a84c694a84950e8c1025a387e
f2dbe347deee8a6d29469d9ba1dbbdb6712282e2
51416 F20110115_AACJQZ horowitz_s_Page_038.pro
9cf56a7386a88f6ce172948af7a549ef
114182f211a31abe01beeccc975d52c89f7202e9
74051 F20110115_AACIPA horowitz_s_Page_077.QC.jpg
f95cc0274aed8d1df51b8a759fb123ba
480f44c7ed09092de29f7681d89884fc53866024
184480 F20110115_AACIOM horowitz_s_Page_007.jpg
454cf9670c95e664906c587cf4dfc63f
4e483d1dcbfb60a90b409bdf41c459dccd7faca3
47753 F20110115_AACINX horowitz_s_Page_166.jp2
c227df4cca51d69f23b3c29bd77e9203
acd8eb1796216b8a11faf35527a9cede8fedad4d
1770 F20110115_AACJSD horowitz_s_Page_071.txt
e5034e8bd7a3683fe4285c7ed75a0dd5
d5426d3fe07aaa04defea4157e65020bbc25a0ae
36364 F20110115_AACJRP horowitz_s_Page_168.pro
a2fd1808127d284ad9cb3e23c257fe81
c032ce5544ca71fcafa413165bbc7fddd42d1419
24647 F20110115_AACIPB horowitz_s_Page_141thm.jpg
b688a00588bbe8f8355903ecd5aa4353
94e37c6da2f6c61022085c7d2111aa9b34ef83c6
F20110115_AACION horowitz_s_Page_195.tif
78eff443098b105067dd34e9bbd95f19
9e12531ad30ccc8993b7c8dc023f5068bb8bac21
79267 F20110115_AACINY horowitz_s_Page_174.QC.jpg
248cba6049bf2636049c6e60b325623f
2e17aad427e4452b0cfc87557ae98b4854916c23
1912 F20110115_AACJSE horowitz_s_Page_075.txt
71f98c4cbb19008de7b4bd2277366f74
0923200f24eef6dcc0a6819e10e210f0c99db601
47506 F20110115_AACJRQ horowitz_s_Page_182.pro
3fcc237ca2859564ff35e6a164f81a71
39177c8b269165beb6489d03d85bc9eb9e00760a
55954 F20110115_AACIPC horowitz_s_Page_163.QC.jpg
64a56916bf5280b7922e2b802c32fb92
fcaae4ac236aae117bfef2fb14a5957c9c0a5d49
1063 F20110115_AACIOO horowitz_s_Page_131.txt
65857d8ead62bd9ab5a57510625727ba
415c3f24c922ea74ed78a32d1fa6ee495e194909
73490 F20110115_AACINZ horowitz_s_Page_095.jp2
45d716f2731da07d9d3b864aea3a5b90
8c2f3223fce4592af990d1cba3f47fb7a913052a
1911 F20110115_AACJSF horowitz_s_Page_083.txt
8b1a57ddfc16ef07f200c9bbf44168ff
b82505cddc65911e82dbf62cb7a74e68d8c93d6e
42596 F20110115_AACJRR horowitz_s_Page_194.pro
a984f1cd193d28e0df87c5f24638ae18
4995278953c8f98487d1ec1106a676d9c0a6525c
180261 F20110115_AACIPD horowitz_s_Page_091.jpg
aa93ec41269d4ed836df58b6a8052ca7
acd29367a145ba5899291012dbf40fe178a782c1
48551 F20110115_AACIOP horowitz_s_Page_008thm.jpg
0d176a4cfbd6bd6443120f9ead11cca5
069f00fbc797f02c6e8095efb074d07f9c7fa354
F20110115_AACJSG horowitz_s_Page_098.txt
4c0c0a04342edb0e0010f0d41eb6cb2b
93f857e5e4732f6bbdfa2410f5e6faa3e01af0ed
49528 F20110115_AACJRS horowitz_s_Page_205.pro
df5b48003a7daebe36b0aaeaef808fcb
980f41ee8db61af7c8f9c5aac7e353d80c81a5fc
18325 F20110115_AACIOQ horowitz_s_Page_047thm.jpg
b33baa992764c5c103633fb15f68fa6f
6fb16360dbb5a4022cd1e7655c07f2c71ee43576
29691 F20110115_AACIPE horowitz_s_Page_104.pro
cd200f19346ed49841e4d331005357f0
a16df1e444bfb2cb936556ccf1f967d5f031e2e8
1188 F20110115_AACJSH horowitz_s_Page_102.txt
1d67a87b62177148f22ef66a17d6efb4
45123774644e6d93351cb314193914be8ca78d97
1628 F20110115_AACJRT horowitz_s_Page_007.txt
2dfa1d6dbb0c7adee24bca6a7506aac5
25af7e30bd194af4311a3427757e0e4ccf24e9a6
13670 F20110115_AACIOR horowitz_s_Page_057thm.jpg
8a1d0fbe9e74b8ed896b8545526218a8
74c5e3d914f4b528a1b3bd29e841ddce2d3efa49
F20110115_AACIPF horowitz_s_Page_142.tif
a778f99f8ed53eb2c0a173cf1bc3204a
af0e8917cd9908a0eacb69b66b91ff061c960a4b
1551 F20110115_AACJSI horowitz_s_Page_112.txt
e7fd76aed2e75ef5538094ca565ad4ca
20c82c3921c02714f00b8107eeac2ab6c253acb6
876482 F20110115_AACIPG horowitz_s_Page_139.jp2
d18b85fa5e55946d201d72d0d2f74e57
dc65a7538e0e213b03bb427e04468bb1cde5aa24
2729 F20110115_AACJSJ horowitz_s_Page_127.txt
dca807253e0d1e94da179e0baa4b5c38
9842ed2dd1126258ab06fd9174554ca47753b6b7
3005 F20110115_AACJRU horowitz_s_Page_012.txt
5fa8d57cd3570ee23a735a5c79ad9167
353462f3da7649f6a6bc2058d663952fcc654d70
F20110115_AACIOS horowitz_s_Page_141.tif
71fb4c5befaabd291c48de892ddd4f66
970b70175691872285b110c33f66b83c3eb46193
1393 F20110115_AACIPH horowitz_s_Page_096.txt
1b45150a77cedc18d5f33643d034d271
5a085509a177de5508c0031e2deda340e123c81f
1746 F20110115_AACJSK horowitz_s_Page_129.txt
63ef95c089be1a45afc028e5aa7a4adb
3b43faca1898f2f8264ee12dbe0d53840889f458
1700 F20110115_AACJRV horowitz_s_Page_015.txt
b3bf373727f51750ad279cdfc0222fca
c8141e914b55d64f2858741ca3806fdc7ebca082
190428 F20110115_AACIOT horowitz_s_Page_081.jpg
ca0db7087d26258432c5484996aea0b3
b956519e1d8fd8cfd0d633b966c3504d348ea454
109017 F20110115_AACIPI horowitz_s_Page_110.jp2
fbe985437d742a4e44d3df7fb6089ed0
c8d501ce7af233f411973aaf8d6b9cff51784194
1766 F20110115_AACJSL horowitz_s_Page_135.txt
27a997e30794d1f6f477c6b8b1668214
0e1ce69eb44320a5351d32990dde7ab29f4567b1
F20110115_AACJRW horowitz_s_Page_018.txt
613bf20a20a437b3d23f15cbc623293f
f93f4cef58888d5b417753653b283ee511426e51
45315 F20110115_AACIOU horowitz_s_Page_133thm.jpg
89b56308985041d9d6f7806c6e24438a
f2545254841cf887cd2ef47608ae00c2b91d2410
38250 F20110115_AACIPJ horowitz_s_Page_047.pro
b4ef6346f0e1dce7ea81ad46055820aa
bb2c08044be0736f1d6eac333b3cc960c71178bc
71848 F20110115_AACJTA horowitz_s_Page_032.QC.jpg
c8e4b888f62e80e571e40c332022499e
4ce4f3bf4f302ef1ad20dfe8d123fc58ae561a39
1561 F20110115_AACJSM horowitz_s_Page_137.txt
b8aac99e324055498f5ae16922a0db38
62cdb55d12bf66f7f4cfa0570a2309c06a687353
1928 F20110115_AACJRX horowitz_s_Page_029.txt
5d5200b4c715e05713009092a5fc4c9b
cd038eaf6fc33e8daa0836e16795840590637e05
181941 F20110115_AACIOV horowitz_s_Page_194.jpg
5aa8062a9744feb7022f4bee4531e22b
08bc6154c15b0f4694dc8a33661fce68ac91f649
174159 F20110115_AACIPK horowitz_s_Page_188.jp2
39489754668e8f31a7d76db8194ee0fd
c2392a78f7d27d42ce9aaad6c02ac55a7da80394
46562 F20110115_AACJTB horowitz_s_Page_113thm.jpg
196c1a818d699b8bce19ba67b742c88f
c989b8ed5dc6a635d7e87c62b08c27d2af907a72
1404 F20110115_AACJSN horowitz_s_Page_143.txt
9320f8dad2e1a32ae0ee426a0668fbbd
ab2905164f650f930b70c397c9181e988b51deb5
1529 F20110115_AACJRY horowitz_s_Page_035.txt
ab44509227961b61f1c898fc62b3b4bc
a2811542d43b3373bdb214128318c1bc8f58fc8d
14029 F20110115_AACIOW horowitz_s_Page_209.pro
892f7125e073de0abf37f46bd0e37193
4f8c6c946d3933710e81a959b1f8363a026663ed
F20110115_AACIPL horowitz_s_Page_161.tif
9357812d22366376506e3c37c8b78456
485ada1a5cbcf01ff88da2f1fd7d5730022fe8dd
81773 F20110115_AACJTC horowitz_s_Page_080.QC.jpg
d1e3d86880f34a22ef12c7adacb520cc
bf1eb0bb9e2831466c3ec1040eb13aeeae18aedd
F20110115_AACJSO horowitz_s_Page_153.txt
b154da16a5f1d09b5c558e589ee69188
3a9b711ff2912eefe0102a7cbd06e0cabe602db9
1790 F20110115_AACJRZ horowitz_s_Page_037.txt
da43ca366290599c33e9e4b0bffa84c6
2012853efbfafbef71a35db1b3623bb1905687c5
38366 F20110115_AACIOX horowitz_s_Page_135.pro
eb2e56aa40c0983803e2d94eed46956b
5e37d73934956a6724f8d0f5246940ca47119471
208832 F20110115_AACIQA horowitz_s_Page_023.jpg
3f8155a85c2958c7cb33eada78c3c2a3
d22d91962589df7ab08b24e5ec53fe72f202193a
81528 F20110115_AACIPM horowitz_s_Page_185.QC.jpg
266cd16dcf1b65c36eda72e0518cbd57
1d8e668dc777a15b68d1297d4759d795dc1ed9bd
21858 F20110115_AACJTD horowitz_s_Page_042thm.jpg
4c8532910e8331118652ec57bc504e64
85aaddf494a20c700db50573e90cadb112889a2d
1246 F20110115_AACJSP horowitz_s_Page_156.txt
de939e1e93101f2a5776cd3678ad0dc3
4b116a75fcda788609d235bd848a20582c1778ec
22133 F20110115_AACIOY horowitz_s_Page_058thm.jpg
6dc94473fdb7dce2d126c0af99e337bd
e9fb983b8dd3293e21c1eaeb8a040cde962dbaa0
27412 F20110115_AACIQB horowitz_s_Page_173.pro
476a55c1f41cab0960509c2aa320b852
9095e76cf51691f4351cea420ac96534756d291c
213851 F20110115_AACIPN horowitz_s_Page_181.jpg
203d5d7dbef04af9dc241f8f7061a33e
100898d25d71e8c998c83ccebaa9d87a5d70a7e5
82080 F20110115_AACJTE horowitz_s_Page_089.QC.jpg
ae87976c7d3940374c34831cefe85149
a88c686bc12260ab03406a7fc0f5b719e0b98d2e
1548 F20110115_AACJSQ horowitz_s_Page_162.txt
96846d88f31846c1fde69ccb20de49c3
1f4257d7ba636b713c7a9bcf7b1f6c9d93ec359a
F20110115_AACIOZ horowitz_s_Page_035.tif
192389c360a4841e78a275326ede3d16
8823e20cb3197798baa69aed71790e1afa288167
F20110115_AACIQC horowitz_s_Page_063.tif
d903c75723a158f86df56689ee62199d
c196f4c4aace545a837f4f4e984c82306af4577e
F20110115_AACIPO horowitz_s_Page_180.txt
66cf05f91969acddff8a24eda8b31e07
e77335809900da62dea04fce241e6840fae5ece7
78331 F20110115_AACJTF horowitz_s_Page_033.QC.jpg
7158aaffe790829aad5e71c018b029d8
0f27f14caa063e3b886e810b771cd9bda2ff3917
1873 F20110115_AACJSR horowitz_s_Page_164.txt
9bebaee014351328b1ac75cde8e61f8b
7b84c0df88c6380a50cb939e4cf423a56a0699fe
84097 F20110115_AACIQD horowitz_s_Page_061.QC.jpg
2ba6017d0f68893df3b0df7cf98d657d
9b64af1377c9b3caa8d66bce85bdc9315e494895
86462 F20110115_AACIPP horowitz_s_Page_196.QC.jpg
73a3ace48d74d9215832bab5d9df7079
aefdd550814c73098a1d9e50a56cebe2ae1bfad8
75204 F20110115_AACJTG horowitz_s_Page_029.QC.jpg
4513c40d26f0179711c6b413c9012cff
1ca2a9ebf6a26f65e7e98d7539a56336a2c29111
1837 F20110115_AACJSS horowitz_s_Page_170.txt
7629057e2ae4154289ddcc39d4ec2222
6d3f863abd78392360b51fbd924cc26821345b99
F20110115_AACIQE horowitz_s_Page_056.txt
22dfe6cc35d165ec50ec8b26c61f0b80
b5e3fb18aa137c865360efc68838c3abfb6d0859
117123 F20110115_AACIPQ horowitz_s_Page_208.jpg
94aae487d8800757b400df439167f230
db26f01355660e2942ebd780a62ffbca3e2572c6
23044 F20110115_AACJTH horowitz_s_Page_170thm.jpg
c69315712f4e1b43b2db182a13a9cbf6
4cead5572923726fee60143bd63280aafa6aebdd
2202 F20110115_AACJST horowitz_s_Page_185.txt
786752c0c0839d25c3a70a5f168b0f45
9bd9cd43882ab3c05fe0c50d244f646d14421f9d
F20110115_AACIQF horowitz_s_Page_201.tif
fec31d198c069b24abe1e3fe579116e4
442e7788c295108688fcd1e5c383d795cbe8bc3e
1988 F20110115_AACIPR horowitz_s_Page_023.txt
b32dc5aaa6677b9789c8f1f90c747bd7
c9f35135d67d5b5a410b182631ea28daa1452ca1



PAGE 1

DEVELOPMENT OF A MEMS-BASED ACOUSTIC ENERGY HARVESTER BY STEPHEN BRIAN HOROWITZ A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2005

PAGE 2

Copyright 2005 by Stephen Brian Horowitz

PAGE 3

I dedicate this dissertation to my wife, Megan. Youre my sweetheart, my wife and my best friend. I love you more than words can say.

PAGE 4

iv TABLE OF CONTENTS Page LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.....................................................................................................................xi v CHAPTER 1 INTRODUCTION........................................................................................................1 Acoustic Energy Harvester Concept..........................................................................1 Motivation..................................................................................................................... 5 Background...................................................................................................................6 Piezoelectricity......................................................................................................6 Piezoelectric Materials in MEMS.........................................................................6 Materials and properties.................................................................................7 Deposition methods........................................................................................8 Piezoelectric integration with MEMS............................................................9 Reduced-Order Modeling of Piezoelectric Materials..........................................11 Energy Absorption/Vibration Damping and Energy Harvesting........................13 Approach.....................................................................................................................14 2 THEORETICAL BACKGROUND...........................................................................17 Overview Energy Harvesting..................................................................................17 Maximum Average Power Transfer....................................................................20 Electromechanical Transducers...........................................................................23 Lumped Element Modeling........................................................................................25 Helmholtz Resonator..................................................................................................29 Lumped Element Model......................................................................................29 Equivalent Circuit................................................................................................33 Piezoelectric Composite Plate....................................................................................35 Piezoelectricity....................................................................................................36 Composite Plate Modeling..................................................................................41 Lumped element model................................................................................41 Equivalent circuit.........................................................................................46 Case 1: Well below the short-circu it mechanical resonant frequency.........50 Case 2: At the short-circuit m echanical resonant frequency........................52 General expression vs. simplified cases.......................................................54

PAGE 5

v Effects of including the radiation impedance...............................................55 Equivalent circuit parameters of piez oelectric composite circular plates....61 Acoustic Energy Harvester Dynamic Behavior..........................................................68 Acoustical Input Behavior...................................................................................68 Electrical Output Behavior..................................................................................71 Operation at an Optimal Frequency....................................................................76 Device Configurations................................................................................................79 3 FABRICATION AND PACKAGING.......................................................................82 Overview.....................................................................................................................82 Process Flow...............................................................................................................82 Packaging....................................................................................................................92 Packaging Scheme...............................................................................................92 Quarter-wave resonator package..................................................................92 Sealed cavity package..................................................................................94 Chip to Package Mounting..................................................................................94 Poling...................................................................................................................96 4 EXPERIMENTAL SETUP........................................................................................98 Parameters and How They are Obtained....................................................................98 Experimental Setup Details......................................................................................101 Ferroelectric Measurements..............................................................................101 Electrically Actuated Response Laser Scanning Vibrometer.........................102 Acoustic Characterization Plane Wave Tube.................................................103 Electrical Impedance Characterization..............................................................107 Initial Deflection Measurements Wyko Optical Profilometer........................107 5 EXPERIMENTAL RESULTS AND DISCUSSION...............................................108 Ferroelectric Measurements......................................................................................108 Electrical Measurements...........................................................................................110 Electrical Output Impedance.............................................................................110 Electrically Actuated Response Laser Scanning Vibrometer.........................114 Frequency response....................................................................................114 Linearity.....................................................................................................124 Acoustical Measurements.........................................................................................126 Acoustically Actuated Frequenc y Response Plane-Wave Tube....................126 Sealed cavity package................................................................................126 Quarter-wave resonator package................................................................127 Acoustic Input Impedance Measurements Plane-Wave Tube........................130 Quarter-Wave Resonator Package.....................................................................133 Initial Deflection Measurements Wyko Optical Profilometer...............................134 Energy Harvesting Measurements............................................................................138 Optimal Resistance............................................................................................138 Optimal Energy Harvesting...............................................................................139

PAGE 6

vi 6 ALTERNATIVE APPLICATION PIEZOELECTRIC MICROPHONE.............147 7 CONCLUSION AND FUTURE WORK.................................................................153 Development Issues..................................................................................................154 Future Work..............................................................................................................156 APPENDIX A MATLAB CODE.....................................................................................................158 lem.m........................................................................................................................15 8 eh_setup.m................................................................................................................163 silicon_sweep.m........................................................................................................163 platinum.m................................................................................................................163 tio2.m........................................................................................................................1 63 piezo_sweep.m..........................................................................................................164 initialise.m................................................................................................................164 totaldeflection_forP.m..............................................................................................164 totaldeflection_forV.m..............................................................................................164 initialdeflection.m.....................................................................................................165 formmatrix_CMFa.m................................................................................................165 frommatrix_dA.m.....................................................................................................166 abdpiezo.m................................................................................................................166 constants1234.m.......................................................................................................167 solver1.m..................................................................................................................168 B DETAILED PROCESS FLOW................................................................................170 Energy Harvester Process Traveler........................................................................170 Graphical Representation of Process Flow...............................................................172 Mask Layouts............................................................................................................174 Package Drawings....................................................................................................178 LIST OF REFERENCES.................................................................................................180 BIOGRAPHICAL SKETCH...........................................................................................195

PAGE 7

vii LIST OF TABLES Table page 1-1: Typical material pr operties of select piezoelectric materials.......................................7 1-2: Selected thin film piezoelectric materi als and corresponding prope rties as reported in the literature..............................................................................................................10 2-1: Equivalent lumped elements in several common energy domains............................28 2-2: Conjugate power variables.........................................................................................29 2-3: General materi al properties used in modeling............................................................55 2-4: PZT properties used in modeling...............................................................................55 2-5: Chosen device configur ations for fabrication.............................................................81 3-1: Wafer bow measurements and the resulting calculated stress....................................86 3-2: Measured wafer mass and calculated density of PZT................................................87 5-1: Summary of electrical impedance measurements....................................................113 5-2: Comparison of ferroelectric and diel ectric properties of thin film PZT...................113 5-3: Summary of electrically actuated frequency response measurements.....................124 6-1: Summary of experiment al results of microphone...................................................152

PAGE 8

viii LIST OF FIGURES Figure page 1-1: Roadmap to Chapter 1..................................................................................................2 1-2: Schematic of overall energy harvesting concept..........................................................3 1-3: Conceptual schematic of the acoustic energy harvester...............................................4 1-4: Basic LCR electrical circuit..........................................................................................4 2-1: Roadmap to Chapter 2................................................................................................18 2-2: Thvenin equivalent circuit for purposes of maximizing power transfer...................21 2-3: Diagram showing side view of a Helmho ltz resonator and its mechanical equivalent of a mass-spring-damper system..............................................................................29 2-4: Equivalent circuit representa tion of a Helmholtz resonator.......................................34 2-5: Theoretical pressure amplification of a conventional Helmholtz resonator. (a) magnitude (b) phase.................................................................................................35 2-6: Cross-sectional and 3-D schematic of piezo electric composite circul ar plate. (not to scale)......................................................................................................................... 36 2-7: Notation of axes used in piezoelectric transduction...................................................37 2-8: Idealized perovskite crys tal structure for PZT. a) centrosymmetric structure prior to poling. b)non-centrosymmetric structure after poling.............................................38 2-9: Schematic of the poling process:...............................................................................39 2-10: Polarization vs. electri c field hysteresis loop...........................................................40 2-11: Electro-acoustic equivalent circuit representation with ebC asC and .................46 2-12: Electro-acoustic equivalent circuit representation with efC, asC and ................47 2-13: Dynamic electro-acoustic equivalent circuit............................................................48

PAGE 9

ix 2-14: Sensitivity vs. frequency for a pi ezoelectric composite circular plate.....................55 2-15: Equivalent circuit of piezo-composite plate includi ng the radiation impedance.....57 2-16: Equivalent circuit with resistive load.......................................................................60 2-17: Conceptual 3-D and cr oss-sectional schematic of th e circular composite plate.....62 2-18: Effective acoustic short-circ uit compliance as a function of 12 R R and pstt.......63 2-19: Effective acoustic mass as a function of 12 R Rand pstt.......................................64 2-20: Short-circuit resonant frequency as a function of 12 R R and pstt........................65 2-21: Electro-acoustic transducti on coefficient as a function of 12 R Rand pstt............66 2-22: Electromechanical coupling coefficient, k, as a function of 12 R R and pstt.........67 2-23: Equivalent circuit for acoustic en ergy harvester with resistive load........................68 2-24: Magnitude of the acoustical input im pedance for the piezoelectric composite diaphragm and in combination with the Helmholtz resonator.................................70 2-25: Magnitude and phase of th e acoustical input power for an input acoustic pressure of 94 PdB .................................................................................................................72 2-26: Electrical output impedance for the piezoelectric composite diaphragm by itself and in combination with the Helmholtz resonator..........................................................73 2-27: Magnitude and phase of the output voltage for the piezoelectric composite diaphragm and in combination with the Helmholtz resonator.................................74 2-28: Magnitude and phase of the electric out put power delivered to the load resistor....75 2-29: Magnitude and phase of th e energy harvester efficiency.........................................76 2-30: Open-circuit voltage as a function of the applied acoustic pressure for the piezoelectric diaphragm by itself and packaged with a Helmholtz resonator..........77 2-31: Input and output power as a function of the load resistance placed across the electrodes of the piezoelectric material....................................................................78 2-32: Input and output power as a function of the applied acoustic input pressure, while using an optimal load resistor...................................................................................79 3-1: Roadmap for Chapter 3.............................................................................................83

PAGE 10

x 3-2: Condensed process se quence in cross-section. ........................................................84 3-3: Early DRIE results showi ng significant sidewall damage.........................................88 3-4: SEM image showing black silicon at the base of a DRIE-etched trench...................89 3-5: SEM image of a successful DRIE etch through the thickness of a silicon wafer......90 3-6: Illustration of a single diaphragm device...................................................................91 3-7: Block layout illustrating numbering scheme for devices...........................................91 3-8: Wafer layout illustrating numbering scheme for blocks............................................92 3-10: Quarter-wave resonator package consisting of acrylic plug, copper leads and vent channel.....................................................................................................................93 3-11: Sealed cavity package...............................................................................................95 3-12: Optical photograph of a packaged energy harvester................................................95 4-1: Roadmap for Chapter 4..............................................................................................99 4-2: Experimental setup for impe dance and power measurements.................................104 4-3: Schematic of incident, reflected and input power....................................................106 5-1: Roadmap for Chapter 5...........................................................................................108 5-2: Hysteresis loop for Device 1A-5-4 in a sealed cavity package. The device has a radius of 1200 m and a central mass..................................................................109 5-3: Measured paralle l output capacitance, pC vs. outer radius, 2 R overlaid with a theoretical capacitance curve based on the average extracted dielectric constant.111 5-4: Parallel output resistance, p R vs. outer radius, 2 R overlaid with a theoretical resistance curve based on the average measured conductivity...............................112 5-5: Device 1A-6-1, electrically actuated center deflection for the device with a radius of 900 m and no central mass, packaged in the quarter-wave resonator package..115 5-6: Device 1A-6-2, electrically actuated center deflection for the device with a radius of 900 m and a central mass, packaged in th e quarter-wave resonator package....116 5-7: Device 1A-11-3, electrical ly actuated center deflection for the device with a radius of 1200 m and no central mass, packaged in the quarter-wave resonator..........117

PAGE 11

xi 5-8: Higher order mode image of Device 1A11-3, taken using scanning laser vibrometer at 120.9 kHz............................................................................................................117 5-9: Device 1A-10-4, electrical ly actuated center deflection for the device with a radius of 1200 m and a central mass, packaged in the quarter-wave resonator............118 5-10: Device 1A-6-5, electrica lly actuated center deflection fo r the device with a radius of 1500 m and no central mass, packaged in the quarter-wave resonator..........119 5-11: Device 1A-10-6, electricall y actuated center deflection for a device with a radius of 1500 m and a central mass, packaged in the quarter-wave resonator package..120 5-12: Device 1A-3-7, electrica lly actuated center deflection fo r the device with a radius of 1800 m and no central mass, packaged in the sealed-cavity package............121 5 -13: Electrically actuated sensitivity at low frequency (well below resonance) ( 120.95RR )........................................................................................................122 5-14: Summary of electrically actuated resonant frequencies........................................122 5-15: Drop in resonant frequency due to the addition of the central mass......................123 5-16: Magnitude of the resonant center deflection versus driving voltage......................125 5-17: Mechanical sensitivity at re sonance versus driving voltage...................................126 5-18: Device 1A-5-4, magnit ude and phase of the acousti cally actuated frequency response in a sealed cavity package fo r the device with a radius of 1200m and a central mass............................................................................................................128 5-19: Device 1A-10-4 in a quarter-wave resona tor package. Magnitude and phase of the acoustically actuated frequency response for the device with a radius of 1200 m and a central mass..................................................................................................129 5-20: Device 1A-6-5 in a quarter-wave resona tor package. Magnitude and phase of the acoustically actuated frequency response for the device with a radius of 1500 m and no central mass................................................................................................130 5-21: Device 1A-10-2, Normali zed acoustic impedance in a sealed cavity package for the device with a radius of 900 m and a central mass..............................................131 5-22: Device 1A-5-4, Normalized acoustic impe dance in a sealed cavity package for the device with a radius of 1200 m and a central mass............................................132 5-23: Device 1A-3-7, Normalized acoustic impe dance in a sealed cavity package for the device with a radius of 1800 m and no central mass..........................................132

PAGE 12

xii 5-24: Device 1A-6-1, Normalized acousti c impedance in a quarter-wave resonator package for the device with a radius of 900 m and no central mass..................133 5-25: Device 1A-11-3, Norma lized acoustic impedance in a quarter-wave resonator package for the device with a radius of 1200 m and no central mass................134 5-26: Device 1A-11-3, Initial st atic deflection resulting from residual stresses for a device with a radius of 1200 m and no central mass.....................................................135 5-27: Device 1A-6-6, Initial static deflection resulting from residual stresses for a device with a radius of 1500 m and a central mass.......................................................136 5-28: Device 1A-3-7, Initial static deflection resulting from residual stresses for a device with a radius of 1800 m and no central mass.....................................................136 5-29: Device 1A-3-8, Initial static deflection resulting from residual stresses for a device with a radius of 1800 m and a central mass.......................................................137 5-30: Measured power delivered to a load as function of the load resistance for Device 1A-4-4 and Device 1A-3-8 as compar ed against theoretical values......................140 5-31: Measured output voltage across the load as a function of applied acoustic pressure and compared to theoretical values........................................................................140 5-32: Measured power delivered to load as a function of applied acoustic pressure and compared to theoretical values...............................................................................141 5-33: Measured overall efficiency of each de vice overlaid with theoretical values for comparison.............................................................................................................142 5-34: Resonant frequency versus applied acoustic pressure, resulting from nonlinear response..................................................................................................................146 6-1: Linearity of the microphone device at 1 kHz............................................................148 6-2: Frequency response spectrum in terms of magnitude and phase.............................149 6-3: Noise floor spectrum of output voltage wh en no acoustic signal is applied, as well as noise floor due to measurement setup alone..........................................................151 6-4: Electrically actuated fre quency response of microphone device.............................152 7-1: Roadmap to Chapter 7..............................................................................................154 B-1: Step 1: Deposit Ti on SOI Wafer and oxidize to TiO2............................................172 B-2: Step 2: Deposit Ti/Pla tinum(30nm/170nm) -Lift-Off w/ Mask (ElectrodeBot)......172

PAGE 13

xiii B-3: Step 3: Spin PZT 6 times to achieve desired thickness.........................................172 B-4: Step 4: Deposit Plat inum(180nm)-Lift-off w/ Mask (ElectrodeTop)......................172 B-5: Step 5: Wet Etch PZT using Pt as etch mask...........................................................173 B-6: Step 6.1: Spin thick photo resist on bottom (7um)..................................................173 B-7: Step 6.2: Pattern using mask (Cavity).....................................................................173 B-8: Step 6.3: DRIE to BOX...........................................................................................173 B-9: Step 6.4: Ash Resist.................................................................................................173 B-10: Step 6.5: BOE back side to remove oxide..............................................................174 B-11: Backside metal mask.............................................................................................174 B-12: Bottom electrode mask..........................................................................................175 B-13: Top electrode mask................................................................................................176 B-14: Cavity mask...........................................................................................................177 B-15: Detailed schematic drawing of quarter wave resonator package and mounting....178 B-16: Detailed schematic of sealed ca vity package and mounting plate........................ 179

PAGE 14

xiv 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 DEVELOPMENT OF A MEMS-BASED ACOUSTIC ENERGY HARVESTER By Stephen Brian Horowitz December 2005 Chair: Toshikazu Nishida Cochair: Mark Sheplak Major Department: Electrical and Computer Engineering In this dissertation, I discuss the devel opment of an electromechanical acoustic resonator for reclamation of energy using microelectromechanical systems (MEMS) technology. The MEMS device converts acousti cal energy to electri cal energy that can then either be stored for later use or utilized directly for a pa rticular circuit application. The work presented in this dissertation takes a first step to wards that goal, by designing, fabricating and test ing several electromechanical acoustic resonators of varying size. The resonators are fabri cated using MEMS processing techniques on a silicon wafer. Each re sonator consists of a 3 m thick silicon diaphragm with a circular ring of piezoelectric material. The dia phragm transduces an acoustical pressure fluctuation into a mechanical deformation, wh ile the piezoelectric material transduces that mechanical deformation into an el ectrical signal (charge or voltage).

PAGE 15

1 CHAPTER 1 INTRODUCTION The goal of my research was to utilize microelectromechanical systems (MEMS) based fabrication techniques to develop an electromechanical acoustic resonator for harvesting of acoustic energy. The MEMS devi ce converts acoustical en ergy to electrical energy that can then either be stored or used directly for a particular circuit application. The relatively small geometries possible in ME MS make such a device useful for small, portable devices, where battery requirements ar e often difficult to m eet. Additionally, the device is well suited to applications where wiri ng is difficult such as remote sensors or unwieldy as in the case of large arrays of devices requiring power. This chapter begins with an overview of the acoustical to electrical energy reclamation process that is employed in th is research. Following this, the motivation behind this research is presented. Then, an in-depth literature survey is presented to familiarize the reader with the technologica l and theoretical developments related to energy harvesting and piezoelec tricity, followed by a short discussion of the technical approach that was employed. A graphical roadmap for this chapter is shown in Figure 1-1. Acoustic Energy Harvester Concept The overall concept behind the energy harv ester is illustrated in Figure 1-2. The main idea is to convert acoustic energy into a usable form of el ectrical energy. The figure illustrates a plane wave tube as the sour ce of acoustic power. Some of the incident

PAGE 16

2 acoustic power is reflected at the end of the tube; however, a portion is transmitted through to the energy harvester that is ci rcled in Figure 1-2. The energy harvester performs the actual transduction of energy fr om the acoustical to the electrical energy domains. More details on this transduc tion technique will be provided later. Overview Motivation Background Piezoelectricity Piezoelectric Materials in MEMS Impedance Modeling of Piezoelectrics Energy Absorption/ Vibration Damping Energy Harvesting Approach Materials and properties Deposition methods Piezoelectric integration with MEMS Figure 1-1: Roadmap to Chapter 1.

PAGE 17

3 The electrical energy that exits the ener gy harvester has the same temporal form as the input acoustic signal, which we assume for now is sinusoidal. The sinusoidal signal is then routed to the energy harvesting circuitry that serves to rectify and condition the signal before sending it on to the end a pplication. In the example illustrated in Figure 1-2, the end application is a battery th at is charged by the out put of the harvesting circuitry. Figure 1-2: Schematic of overa ll energy harvesting concept. A conceptual close-up schematic of the acoustic energy harvester is shown in Figure 1-3. This device cons ists of a Helmholtz resonator possessing a piezoelectric composite backplate. A Helmholtz resonator, which is a type of acoustic resonator, consists of a cavity connected to the envir onment through a small neck. When excited by an acoustic input, a single resonance is see n, whereby the acoustic pressure inside the cavity is amplified to a level much greater th an the incident acoustic signal [1, 2]. The Helmholtz resonator is very similar to an LCR resonant circuit in electrical engineering, as shown in Figure 1-4. Both systems have a single degree-of-freedom, with a single resonant peak, whereby the amplit ude of the forcing function is greatly

PAGE 18

4 amplified. In the LCR circuit, the voltage is equivalent to the acoustic pressure. Both systems operate through the os cillation of energy between generalized potential and kinetic forms. In electrical systems this storage occurs via cap acitors and inductors respectively. These electroacoustic analogies wi ll be explained in more detail in Chapter 2. neck cavity piezoelectric ring compliant plate neck cavity piezoelectric ring compliant plate Figure 1-3: Conceptual schematic of the acoustic en ergy harvester. Figure 1-4: Basic LCR electrical circuit. The large acoustic cavity pressure created by the resonance can then be exploited for energy reclamation by converting the acous tic energy to electr ical energy. This conversion is performed by the piezoelectric composite plate. First, acoustical to mechanical transduction is accomplished vi a the compliant diaphragm, followed by mechanical to electrical transduction, due to the piezoelectric response of the composite, whereby a mechanical strain creat es an electrical voltage.

PAGE 19

5 The voltage that is created through the electromechanical transduction can be harvested by the energy harvester circuit, which is a necessary part of the energy harvesting process. The necessity arises due to the need to match impedances at interfaces. As will be derived in Chapter 2, the power transfer across an interface is maximized when the impedances on each side of the interface are complex conjugates of each other. This is often referred to as impedance matching. One role of the energy harvesting circuit is to provide an input impe dance that matches the output impedance of the piezoelectric structure. Additionally, th e harvesting circuit conve rts the form of the electrical energy to a more a ppropriate form for storage, su ch as converting from AC to DC. In one implementation, a rectifying switched capacitor circuit is used to efficiently convert the AC signal to DC that is then st ored across a low-loss cap acitor [3]. Another possibility is based on the Kymi ssis circuit approach [4]. This approach utilizes passive storage elements for impedance matching coupled with a regulated output, for improved energy harvesting performance. Motivation Acoustic energy is commonly present in a variety of everyday situations. The motivation behind this research is to enable the reclamation of th at energy that would otherwise be lost. Although the availa ble acoustic energy may be small in many situations, the energy requirements for certain applications such as microsensing are also correspondingly small. The ability to reclai m acoustic energy and store it in a usable electrical form enables a novel means of supplying power to relatively low power devices.

PAGE 20

6 Background Piezoelectricity In 1880, the brothers Pierre and Jacques Curie discovered that some crystals developed surface charges when compressed [5 ]. They furthermore found that these charges were proportional to the applied pressure. This phenomenon was later named piezoelectricity by Wilhelm Gottlieb Hankel, and is historically referred to as the direct piezoelectric effect [5]. Add itionally, in these same crysta ls, a strain is produced under application of an electrical field. This is commonly referred to as the converse piezoelectric effect. The material constant relating strain and char ge in a piezoelectric material is known as the piezoelectric charge modulus, d, and is typically quoted in units of p CN or p mV. In order for a material to be piezoelectric, it must have a noncentrosymmetric crystal structure. It was not until 1921, that a useful applica tion was developed for piezoelectricity. It came in the form of a quartz crystal osc illator that was developed by Walter Cady to provide good frequency stabil ity for radio systems [5, 6]. Then, in 1947, the first commercial phonograph pickups based on barium titanate (BaTiO3), a piezoelectric ceramic, were introduced [6]. Modern appl ications of piezoelec tric materials now include high voltage ignition systems, piezoel ectric motors, ink-jet printer heads, acoustic speakers, sonar, ultrasonic transducers, fre quency filters, acoustic delay lines, electrical transformers, and a wide range of physical sens ors, such as acoustic, force, pressure, and acceleration sensors [6]. Piezoelectric Materials in MEMS A number of papers have been publ ished on the use of piezoelectric and ferroelectric materials in ME MS [6, 7]. Piezoelectric materials commonly used include

PAGE 21

7 zinc oxide (ZnO), aluminum nitride (AlN), a nd lead zirconate titanate (PZT). The choice of piezoelectric material depends on se veral selection fact ors including deposition methods, process complexity, integrated ci rcuit (IC) compatibility, and material parameters. Materials and properties Of the three commonly used materials, PZT offers the largest piezoelectric modulus [6]. In comparison to PZT, AlN a nd ZnO have relatively weak piezoelectric coefficients and coupling factors, however they tend to have low diel ectric constants and low dielectric losses, making them more attrac tive for certain applications [8]. A table of typical material properties is shown in Table 1-1 for comparison; however these properties are highly dependent on the deposi tion method, as will be described in more detail shortly. Table 1-1: Typical material properties of select piezoelectric materials. 31d 33d 33, r E p CN p CN GPa 3kgm PZT[6] -130 290 1300 96 7.7 AlN[8] n/a 3.4 10.5 330 3.26 ZnO[9-11] -4.7 12 12.7 210 5.6 In Table 1-1, 31d is the piezoelectric coefficient re lating electrical displacement in the direction (z-axis) to a mechanical st ress in the direction (x-axis) and is the relevant coefficient for bending-mode transducers, whereas 33d relates the electrical displacement in the direction (z-axis) to a mechanical stress in that same direction. The coefficient, 33d, is thus the relevant coefficient for compression-mode transducers. Additionally, 33, r E, and are the relative dielectric constant, Youngs modulus and mass density of the materials, respectively.

PAGE 22

8 Of the three materials, only AlN is fully compatible with standard IC processing, eliminating the integration difficulties present with PZT and ZnO. PZT poses a particular integration challenge, as lead contaminati on is always a big concern; however, PZT has been integrated successfully into ferroe lectric random-access memory (FRAM) [12-14], typically using sputtered platinum as a bo ttom electrode and diffusion barrier layer. Deposition methods Piezoelectric deposition t echniques include various sputtering methods, photoablation [15], hydrothermal and chemical va por deposition (CVD) techniques, and spinon sol-gel processing. For deposition of P ZT, three popular methods are sputtering, solgel and hydrothermal techniques. The earliest work used various forms of sputtering to deposit PZT [16-22] followed soon after by CV D [23]. Castellano and Feinstein [16] used an ion-beam sputtering technique to de posit PZT as did Tro lier-McKinstry et al., [22] while Sreenivas et al [17] employed DC magnetr on sputtering in their PZT deposition. Additionally, sputter depositi on of PZT thin films was performed by Kawabata et al.[24] and Li et al. [25]. Dubois and Muralt [8] deposited PZT using two different methods. One method involved a sol-gel process and produced an average thickness of 900 nm. The other method used reactive sp uttering and achieved a thickness of approximately 500 nm. Hydrothermal deposition relies on a chemical reaction between a titanium bottom electrode and ionic solution under high temperature and pressure. Deposition of 10 m thick PZT by hydrothermal techniques was perf ormed by Kanda et al. [26] for use in a touch probe sensor. Additionally, Morita et al. [27-30] chose to deposit PZT by the hydrothermal method due to the relatively large thicknesses achiev able and the selfalignment of the poling di rection during deposition.

PAGE 23

9 More recently, numerous researchers ha ve utilized the so l-gel process for deposition of PZT [31-47]. In particular, Bahr et al. [31] used a sol-gel PZT process and investigated the reliability and piezoelectric properties of the resulting material. Using this process for various film thickness es, the relative dielectric constant, 33, r varied between 700 and 1000 and delamination occu rred for an indentation load of 1250 N for a film sintered for 5 minutes. Higher load s were found to be achievable using longer sinter times. Bernstein et al [32] and Xu et al. [33] used a sol-gel process to achieve crack-free PZT films with thicknesses of up to 12 m that yielded piezoelectric properties close to the bulk values for PZT. A dielectric constant of 1400 and a piezoelectric coefficient, 33d, of 246 pC/N were reported on a 4 m thick film. The PZT films were used on an array of membranes for acousti c imaging. Kunz et al. [37] report a piezoelectric coefficient, 31d, of 110 pC/N for a sol-gel deposited PZT film used in a triaxial accelerometer. Zurn et al. [48] report similar ma terial properties for sol-gel deposited PZT on a micro cantilever, including a 31d of 120 pC/N for a PZT film thickness of 0.5 m deposited on a low-stress silicon ni tride layer. A summary table of deposited thin film piezoelectric materials and their properties, as reported in the literature, is given in Ta ble 1-2. In this table, 31e is a piezoelectric coefficient relating mechanical stress and an electric field, tan is called the loss tangent and is a measure of the relative losses in the material, and res is the residual stress in the material, while the other properties are as previously defined. Piezoelectric integration with MEMS Some of the earliest integration of piezo electric materials with MEMS focused primarily on ZnO [49, 50] and AlN [49] as th e piezoelectric material. More recently,

PAGE 24

10 Devoe and Pisano [10] developed and charac terized surface micromachined piezoelectric accelerometers that utilized thin films of ZnO for sensing. Deposition of the ZnO was performed using single-target RF sputtering and exhibited a piezoelectric coefficient of 2.3 p CN. Also, Devoe [9] investigated micromech anical beam resonators that use ZnO and a three-mask fabrication process. The resonators are intended for use as electromechanical filters. Table 1-2: Selected thin film piezoelect ric materials and corresponding properties as reported in the literature. Properties Ref. Material Deposition Method 33, r 31d pC N 33d pm V 31e 2C m E GPa 3kg m tan res M Pa [9, 10] ZnO RF Sputt. 12.7 2.3 --161 5605 -1 [11] ZnO Sputtered ----210 5700 --[8] AlN DC Sputt. 10.5 -3.4 -1.02--0.002-700 [8] PZT (45/55) React. Sputt.900 -55 -5.12--0.03150 [8] PZT (45/55) Sol-Gel 1100 -50 -8.28--0.0370 [8] PZT (53/47) React. Sputt.1300 -70 -6.83--0.05230 [26] PZT Hydrothermal--34.2--0.13----[31] PZT (52/48) React. Sputt.700-1000-------[32] PZT React. Sputt.1400 -246 ---0.03-[37] PZT (53/47) React. Sputt.-110------[42] PZT React. Sputt.1100 -------[44, 45] PZT (52/48) React. Sputt.800 --200-56 -0.02-[48] PZT React. Sputt.--120 -60 7600 --[16] PZT IB Sputt. 125 -------[20, 21] PZTWCd RF Sputt. 60-460 -30 -----[51] PZT(X/1-X) React. Sputt.200-600 -----0.02-The integration of ZnO with micromachin ing was also investigated by Indermhle et al. [52], where it was used in an array of silicon micro cantilevers. The end application for the array was parallel atomic force mi croscopy. Another application of ZnO was performed by Han and Kim [11] in the fabr ication of a micromachined piezoelectric ultrasonic transducer. They sputter deposited ZnO on Al, followed by a layer of parylene

PAGE 25

11 for insulation. Dubois and Muralt [8] fabri cated thin films of RF sputtered AlN and performed measurements of the effective transv erse piezoelectric coe fficient that yielded a value of 3.4 p mV. Reduced-Order Modeling of Piezoelectric Materials The development of accurate, practical m odels of the piezoelectric transduction and associated structural interact ions is critical to the design and behavioral prediction of piezoelectric based devices, particularly when micromachining is involved. For a micromachined device, considerable time and expense must be invested in the fabrication, and it is therefor e desirable to have some ability to predict the device behavior before proceeding with fabrication. Behavioral prediction can be accurately accomplished using complete analytical methods. However, these methods, which ofte n involving partial differential equations, can often be unwieldy and physically unintuitive. Furthermore, this functional form is not readily conducive to a fu ll systems level design that links the transducer to electronics. Similarly, finite element m odeling (FEM) techniques are often used to predict system behavior, numerically. The re sults produced by this technique can very precisely follow the physical system; however, the physical insight that can be gleaned is limited. Additionally, it is very difficu lt to determine scaling behavior from FEM results. The scaling behavior, i.e., the cha nge in the system performance as the entire system is scaled up or down, is a critical design issue in the creation of devices using micromachining technology. In order to faci litate a physics-based approach for design, a simplified, reduced-order model is necessary that accurately captur es the geometric and

PAGE 26

12 material dependencies. This reduced-order m odel uses lumped elements to represent the key components that dominate the device behavior. In 1915, Butterworth [53] first showed that any mechanical oscillator, when driven by a periodic voltage across a capacitor, woul d have an equivalent electrical circuit consisting of a resistance, inductance and capac itance in series, and all in parallel with another capacitance. Working independen tly, the earliest equivalent circuit model specifically for piezoelectric materials wa s developed by Van Dyke in 1925 [54-56]. Later, Dye [57] proved that Van Dykes ci rcuit could be derive d from Butterworths theorem. Additionally, Mason [58, 59] and Cady [5] provided thorough reviews of the equivalent circuit model and associated equa tions for quartz oscillator applications. Finally, Fischer [60] extensively covered equi valent circuit models for electromechanical oscillators. Significant research has been performed on the modeling of st ructures containing piezoelectric materials. [38, 41, 61-94] C ho et al. [61, 65] developed a five-port generalized equivalent circu it for a piezoelectric bimorph b eam. The generalized circuit can be used under a variety of boundary c onditions. In these papers, three boundary conditions are specifically analyzed free, simply supported, and cantilevered. For these boundary conditions, the equivalent circuit was found to produce the exact expressions for the beam vibration. Othe r equivalent circuits were developed by Martin [68] for resonators with low Q values, Sheritt et al [69] for thickness vibrators, Lin [67] for coupled ceramic disk resonators, and Chen et al. [68] for coupled resonant filters. Tilmans [94] also presented an equivalent circuit approach fo r modeling distributed parameter systems. Liang et al. [66] de veloped a generalized electromechanical

PAGE 27

13 impedance model that was then used to address issues of en ergy conversion, power consumption, and dynamic response. The approach described can be used for any system for which the driving-point impedance can be fo und either analytically or experimentally. Additionally, van de Leur [91] provided a critical interpreta tion of equivalent circuit models obtained from impedance measurements indicating that care must be taken in identifying individual compone nts contributing to an impe dance as resulting from particular structures in a gi ven device. This can be further understood as an example of the non-uniqueness of equivalent circuit representations, i.e ., more than one equivalent circuit can represent the same impedance. Additionally, Lesieutre and Davis [87] provided insight into the piezoelectric c oupling coefficient, including a technique whereby an outside stress is applied to increase the effective device coupling coefficient beyond the coupling coefficient of the material itself. Research into coupled resonators and thei r unique properties is a related and also relevant topic since the acoustic energy ha rvester involves mating the piezoelectric composite diaphragm to a Helmholtz resonator, resulting in a coupled resonant system. Fischer [60] provided an early treatment of coupled resonant systems. Chen et al., [72], Lin [71, 95], and Li et al. [96] also disc uss modeling and design of coupled resonant systems. These papers address some of the issues involved in coupled resonators, including the shift in res onant frequencies away from their uncoupled values. Energy Absorption/Vibration Damping and Energy Harvesting The absorption of acoustical and mechanic al energy via piezoelectric coupling is closely related to the harvesting of electri cal energy from acoustical and mechanical energy sources. From the acoustical point of view, acoustical energy that is converted to the electrical domain and dissipated across a resistive load can be viewed as being

PAGE 28

14 absorbed. From the electrical point of view, the incident acoustical energy can be viewed as an energy source that may theoretically be harvested and used immediately for electrical subsystems or stored for later use. Because of the relate d nature of these two fields, papers that address eith er of these fields are direc tly relevant to acoustic energy harvesting. Some of the earliest work in piezoelec tric vibration damping was performed by Hagood and von Flotow [97] who us ed resistive and inductive elements in various shunt networks designed to dissipate mechanical energy. A few years later, Hollkamp [98] discussed the use of resonant electrical shunt circuits for multi-modal vibration suppression. In addition, numerous other re searchers have inves tigated piezoelectric means of vibration damping [99-108]. More recently, Caruso [109] discusses electrical shunt circuits for damping of vibrations, as does Wu et al. [110] for damping of panels on an F-15 aircraft. Additionally, a number of pape rs directly address the i ssue of obtaining electrical energy from piezoelectric conve rsion of mechanical energy [3, 4, 111-117]. Early work in this area was performed by Lomenzo et al. [114] and Stein et al. [116]. Later, Kymissis [4] and Smalsers [3] work focuse s on the electrical circuitry necessary for storage of piezoelectrically generated en ergy, while Meninger [115] discusses energy harvesting from an electrostatic transducer. Goldfarb and Jones, [111] Giurgiutiu and Rogers [112, 113] and Zhou and Rogers [ 117] utilize impedance modeling of the piezoelectric and associated structure to analyze the ability to harvest energy. Approach The approach taken in this research is to divide the concept of an acoustical energy harvester into three distinct components as shown previously in Figure 1-2. The first

PAGE 29

15 component consists of a piezoelectric compos ite diaphragm that is responsible for the acoustical to electrical tr ansduction of energy. The second component, which is the packaging, serves as a means to improve the acoustical coupling to the environment. The third and final component is the electrical circuitry that take s the alternating current (AC) electrical signal from the piezoelectric dia phragm and converts it into a direct current (DC) output that can be stor ed for later use. The first two components of the energy harvester were developed in this researc h, while the conversion circuitry remains an ongoing research topic in our gr oup [118-120]. Furthermore, a macroscale version of an acoustic energy harvester was previously deve loped [120] and, through miniaturization, led to the present MEMS-based energy harvester. Chapter 2 presents a theore tical background on various aspects of this research including lumped element modeling of the complete system as well as individual components along with predictions of the dyna mic behavior. Chapter 3 discusses the fabrication of the piezoelectric diaphragm and the packaging scheme employed. Following this, the experimental setup is provi ded in Chapter 4, including a discussion of the measurements that were taken. Then, in Chapter 5, the experimental results are presented. Chapter 6 addresses an alternate application that has been explored, while Chapter 7 presents the conclusions and a statem ent of the contributions to this research. Finally, Appendix A presents the Matlab pr ogram code for computing the diaphragm deflection and calculating the lumped elem ent values, while Appendix B provides a detailed process flow, mask layout, and packaging design The main contributions of this dissertation are as follows: Acoustic energy harvesting concept

PAGE 30

16 Comprehensive lumped element model (LEM) of the acoustic energy harvester First reported integration of thin -film PZT with SOI-based MEMS First aeroacoustic capable piezoelectric microphone.

PAGE 31

17 CHAPTER 2 THEORETICAL BACKGROUND This chapter focuses on the theory and modeling of the acoustic energy harvester and each of its components. The chapter begi ns with an overview of energy harvesting, followed by an introduction to lumped elemen t modeling, by discussi ng both its utility and limitations. This is followed by the development of a lumped element model for a Helmholtz resonator and the corresponding equi valent circuit. Next, the modeling of piezoelectric composite circular plates is addre ssed. To achieve this, an equivalent circuit is presented and general equations are di scussed. Then specific limiting cases are addressed that simplify the anal ysis. Finally, the overall dyna mic behavior of the system is theoretically investigated, including the ac oustical input an d electrical output behavior. A graphical overview of this chapter is provided by th e roadmap of Figure 2-1. Overview Energy Harvesting The general concept behind energy harv esting is to convert energy from the environment that is in an otherwise unusable form into a more useful form. Often the form of energy that is most useful in mode rn applications is electrical energy, where it can be stored in a battery or used to power el ectrical circuitry. The initial form of energy can originate from any number of energy domai ns, such as optical, thermal, mechanical, acoustical, fluidic, chemical, and biological. So me form of transducer is then required to convert that energy to a usable form of elect rical energy. Depending on the situation, that also may require passage of the energy through an intermediary energy domain..

PAGE 32

18 Figure 2-1: Roadmap to Chapter 2.

PAGE 33

19 There are pros and cons to each of these di fferent local energy sources [121]. Some energy sources such as solar power utilize the optical energy doma in to achieve high power densities on the order of 15,000 2Wcm in direct sunlight. The main downside is that direct sunlight is not always available. Vibr ational energy, on the other hand, offers power densities up to 250 2Wcm While this is considerably lower than solar energy, it is useful in pla ces without sunlight but wher e high vibrational energy is available. Acoustic energy, in a manner similar to vi brational energy, offers power densities on the order of 1 2Wcm for a 100 dB acoustic signal [121], or approximately 964,000 2Wcm at 160 dB. While most signals are t ypically much lower than 160 dB, there are applications where such high levels are presen t. Additionally, as w ith vibrational energy, acoustic energy, does not require th e presence of sunlight. Chemical energy sources are commonly em ployed today in the form of batteries and fuel cells. Batteries typically o ffer power densities in the range of 45 3Wcm for non-rechargables, and 7 3Wcm for rechargeable lithium ba tteries [121]. Fuel cells employing methanol, on the other hand, offer power densities as high as 280 3Wcm leading to the current interest in fuel ce ll development [121]. Additionally, a microcombustion engine that employs hydrocarbons as a fuel source realizes power densities of 333 3Wcm [122]. The downside to these types of chemical energy sources is the limited supply of energy. Used batteries must either be replaced or recharged and fuel cells require refueling for continual operat ion. By contrast, the scavenged energy sources, such as vibrational, solar, and acoustical do not theoretically have a limited supply of energy, given the ri ght operating environment.

PAGE 34

20 For the particular application considered in this dissertation, the initial energy is in the form of acoustical energy. In order to convert acoustical en ergy into electrical energy, a diaphragm based transducer is utili zed. The diaphragm tr ansducer utilizes the mechanical energy domain as an inte rmediary to transfer energy. Maximum Average Power Transfer Regardless of the route through which th e energy passes, certain fundamental issues must be addressed in order to maximi ze the amount of energy that is harvested. Whenever a change in impedance is encounter ed by a traveling wave, a reflection of all or part of the energy in that wave occurs, while the remaining energy is transmitted. The procedure for maximizing the energy focuse s on minimizing the reflected component and thus maximizing the transmitted component. This is achieved by matching the impedance along the route traveled by the ener gy. As long as the impedance at a given interface is matched, complete transmission of the energy will occur, regardless of whether or not the two sides of the interface are in the same energy domain. Often it is not possible to match an impedance exactly, du e to external and physical constraints on the system, however it is genera lly still desirable to match the impedances as closely as possible in order to optimize for maximum energy within those constraints [2, 123]. Additionally, one must consid er that power is a comple x quantity. The total power, which is generally complex, is composed of real power and reactive power. The real power is the physical power that is deliver ed to the load, while the reactive power represents energy that is tempor arily stored in the load, before it returns to the source. When designing an energy harvesting system, it is generally preferable to maximize the real power rather than the total power, as it is only the real power that is ultimately retained by the load.

PAGE 35

21 It is useful to look at a Thvenin equivale nt circuit, as shown in Figure 2-2, in order to determine the maximum achievable power transfer to a load [124]. Assuming sinusoidal voltage and current, the average real power delivered to the load is 1 cos 2LLLVi LLVI (2.1) where LV and L I are the peak magnitudes of the vo ltage and current at the load, and L V and L i are the phase angles of the voltage and current respectively. This can be rewritten as 1 2LLLVIpf (2.2) where p f is the power factor, defined by the ra tio of average power to apparent power and is given by L rmsrmspf VI (2.3) where rmsV and rms I are the root-mean-square voltage and current, respectively. Figure 2-2: Thvenin equivalent circuit for purposes of maximi zing power transfer. Using basic circuit analysis, the load vo ltage and current can be expressed in terms of the open circuit voltage, ocV, the Thvenin equivalent impedance TH Z and the load impedance, L Z The expressions are given as

PAGE 36

22 ocL L THLVZ V Z Z (2.4) and oc L THLV I Z Z (2.5) where the Thvenin equivalent impedance, THTHTH Z RjX and the load impedance can be written as LLL Z RjX Plugging these expressions b ack into Eq. (2.4) and Eq. (2.5) and separating out the magnitude components yields 1 22 2 1 22 2 ocLL L THLTHLVRX V RRXX (2.6) and 1 22 2 oc L THLTHLV I RRXX (2.7) The phase angle between the voltage a nd current can also be expressed as L LLViZ where L Z is the phase angle of the impedance. The power factor can be defined as 1 22 2cosL Z L LLR pf RX (2.8) Plugging this expression into Eq. (2.1), toge ther with Eq. (2.6) and Eq. (2.7) yields 2 221 2ocL L THLTHLVR RRXX (2.9) As, the quantity THL X X is only in the denominator, any nonzero value reduces the overall power absorbed by the load, thus LTH X X is optimal and reduces Eq. (2.9) to

PAGE 37

23 2 21 2ocL L LTHVR RR (2.10) This expression can be maximized by differe ntiating this expression with respect to L R and equating the derivative to ze ro, yielding an optimal value of LTH R R. The average power to the load can be maximized, th en, by setting the load impedance to LLLTHTHTH Z RjXRjXZ (2.11) where TH Z is the complex conjugate of the Thvenin impedance. This derivation assumes total freedom in the choice of the load impedance. Under the constraint of a purely resistive load, where 0LX the equation given in Eq. (2.9) can be differentiated di rectly and set equal to zero with 0LX, in order to find the maximum average power transfer. Th is procedure yields an optimal value for the load resistance of 22 L THTHTH R RXZ (2.12) under the constraint of a purely resistive load. Electromechanical Transducers A transducer can be broadly defined as an interface between any two energy domains, including optical, thermal, mechanical electrical, acoustical fluidic, chemical, and biological. Typically, transducers are br oadly classified into two main categories: sensors and actuators. The breakdown among thes e classifications relates to the direction of information flow. Sensors are trans ducers that convert information from the environment into a form that is more easily analyzed, recorded or processed. Actuators, on the other hand, take information from this form and convert it into a form that more

PAGE 38

24 readily interacts with the intended environment. Both types of transducers are mainly concerned with the way in which the information is utilized. The classifications described above focu s on the information aspect of a given signal that is encoded in either the effort or flow, but not their product. As such, these types of transducers are optimized for transm ission of information in one direction or another. Besides information, a signal carries power. A third type of transducer exists that falls under neither categor y and can be best described as an energy harvester. The concept behind the energy harvester is to achieve an optimal am ount of average power transfer between energy domains. Instead of optimizing sensitivity or dynamic range, an energy harvester is optimized for efficiency and maximum average power transfer. Additionally, actuators and especially sensors are designed as linear devices in order to facilitate a straight forward relationship between in put and output. The linear input-output relationship ensures that the problem of calculating their behavior is tractable, and is necessitated by the information contained in the signal. Energy harvesters, on the other hand, do not necessari ly require a linear i nput-output relationship as only the power is of interest. Some deta ils of the input may get lost in the nonlinear transduction, but only the power at the output is needed. There are a wide range of transduction techniques available that couple the mechanical and electrical energy domains, including electrostatic, electrodynamic, piezoresistive and piezoelectric [1, 125]. Each technique offers va rious tradeoffs among performance, reliability, and material integrat ion. Transducers can be separated into two types, direct and indirect, based on the way they interact with energy. Direct transducers, as the name implies, directly transduce ener gy between the two energy domains. Indirect

PAGE 39

25 transducers, however, rely upon a second en ergy source that modulates the primary energy source as it passes through the transd ucer. The inherent nature of indirect transduction leads to a lower transduction efficien cy as compared to direct transduction. Electrodynamic transduction, a direct transduction mechanism, yields high coupling efficiencies. Additionally, this met hod uses no outside power source in order to operate. The transduction is essentially powered by the input signal itself. Despite these advantages, electrodynamic transduction suffers from difficulty in integrating the particular required materials, such as copper or magnetic materials, into a standard IC process line. Additiona lly, the necessary fabrication is often more complex than the other techniques discussed here. Piezoelectric transduction requires no outside source of power ot her than the input signal, and offers a high sensitivity, with a potentially low noise floor. The main disadvantages are the difficulty in integrati ng the piezoelectric materials into a standard process flow, and the limited coupling efficien cy due to indirect transduction of energy. Lumped Element Modeling The most accurate, complete, mathematical description of a physical system is a physics-based model, supported by an exact analytical expressi on for the system behavior. Why then are alternative m odeling techniques commonly used? These alternative methods, such as lumped elem ent modeling (LEM), and finite element modeling (FEM), excel in predicting system behavior in situations where an exact, analytical approach is unwieldy or impossible. Additionally, it is not uncommon to have an exact mathematical description of a system that allows for precise prediction of device behavior, but is so complex that it offers lit tle in the way of physical insight into the

PAGE 40

26 design and scaling system. Ofte n, the exact solution is in a mathematical form that, while convenient for behavioral prediction, conceals the underlying physics. Similarly, FEM techniques can accurately predict system behavi or, in this case via a numerical approach. The results produced by this technique can very precisely follow the physical system; however, the physical in sight that can be gleaned is limited. Furthermore, the results depend on the numer ical mesh and convergence of the iterative calculations. In particular, it is very difficult and/or cumbersome to determine scaling behavior from FEM results. The scaling be havior, i.e. the change in the system performance as the entire system is scaled up or down, is a critical design issue in the creation of devices using micromachining technology. In order to efficiently understand the physics behind a given systems response and also to fully appreciate the scaling laws for that system, it is necessary to employ LEM [1, 60]. The concept behind LEM is to reduce the complexity of an analytical or numerical expression by breaking down a given distributed system into discrete elements based on how the system interacts with ener gy [60, 126]. More sp ecifically, the total energy going into any given system is divide d among three distinct t ypes of interactions: the storage of kinetic energy, the storage of potential energy, a nd the dissipation of energy. All systems are composed of these three energy processes. In some systems, known as distributed syst ems, the storage of kinetic and potential energy occurs over a distributed region in sp ace [90, 127]. To accura tely represent these systems mathematically requires a partial differential equation, as spatial and temporal components are inherently coupled. Physi cally, the distribution occurs because the wavelength is on the order of the physical syst em or smaller. At different points along

PAGE 41

27 the period of the wave, differing amounts of energy are split among the energy storage and dissipation mechanisms. As the wavelength of the signal (e. g. acoustical, fluidic, mechanical, optical, etc.) gets larger, to where it is significantl y larger than the length scale of interest, very little variation occurs in the distributi on of energy as a function of space. Mathematically, under these conditions, the spatial and temporal components can be decoupled, allowing for the use of ordinary di fferential equations to solve the problem, rather than partial differential equations. Physically, it means each energy storage or dissipation mechanism can be equa ted to the energy stored or dissipated in an equivalent element that is lumped to a chosen spatial location. In various energy domains, the names for the types of lumped elements vary; however, the concept and mathematics remain th e same. In lumped mechanical systems, kinetic energy is stored via mass, potential energy via the compliance of a spring (i.e. inverse of stiffness), and dissipation of ener gy through the losses of a damper. Similarly, in electrical systems, where lumped element systems are commonplace due to the extremely long wavelengths of el ectrical signals, kinetic ener gy is stored in the magnetic field of an inductor, while pot ential energy is stored in th e charge across a capacitor. Additionally, dissipation of energy is modeled via the re sistor. Finally, in lumped acoustical systems, the kinetic and potential energy is stored in an acoustical mass and acoustical compliance, respectively, while di ssipation of energy is represented by an acoustic resistance. These elements are summarized in Table 2-1. The commonplace nature of lumped elements in electrical systems has led, over the years, to a large growth in graphical and anal ytical techniques to so lve large networks of

PAGE 42

28 interconnected elements. These networks are most commonly represented using electrical circuit notation. A significant benefit of lumped element modeling is that in all of the energy domains, the lumped elements can be represented using an equivalent circuit form. Thus, masses can be repres ented using inductors, compliances using capacitors, and dissipativ e components using resistors. Once the complete equivalent circuit is constructed, standard circuit an alysis techniques (e.g. Kirchoffs current and voltage laws) can be applied to find the solution of interest. Table 2-1: Equivalent lumped elements in several common energy domains. Kinetic Energy Storage Potential Energy Storage Energy Dissipation Acoustical Acoustic Mass 4kgm Acoustic Compliance 3mPa or 5mN Acoustic Resistor 4ms Mechanical Mass kg (pt mass) Compliance mN (spring) Frictional Damper ms Electrical Inductance H Capacitance F Resistance Whenever dealing with more than one lump ed element, the concept of power flow between the elements must be considered. If we define the power flow from element A to element B as AB and the reverse as BA then the net power flow from A to B is netABBA (2.13) Now, since each of the power flows must be gr eater than or equal to zero, then they can each be written as the square of a real number, 1r and 2r [128]. The net power flow can then be rewritten as 22 21212 netlrrrrrr (2.14) It is therefore seen that the net power flow can be written as th e product of two real numbers, which are referred to as conjugate power variables. Moreover, these quantities

PAGE 43

29 are more specifically referred to as an effort, e, and a flow f where the product ef is the power. A table of conjugate power variables, divi ded into effort and flow, is given in Table 2-2 for a number of energy domains. Table 2-2: Conjugate power variables Energy Domain Effort Flow Mechanical translation Force, F Velocity, v Fixed-axis rotation Torque, Angular velocity. Electric circuits Voltage, V Current. I Magnetic circuits MMF, M Flux rate, Incompressible fluid flow Pressure, P Volumetric flow, Q Thermal Temperature, T Entropy flow rate, S (after Senturia, pg 105 [128]) Helmholtz Resonator Lumped Element Model LEM was applied to the Helmholtz resona tor in order to better understand the system [1, 125]. A schematic diagram of a Helmholtz resonator is shown below in Figure 2-3, where V is the cavity or bulb volume, l and 2Sa are the length and cross-sectional area of the neck, respectively, where a is the radius of the neck, 1P is the incident acoustic pressure, and 2P is the cavity acoustic pressure. Both 1P and 2P are considered to be functions of the radian frequency, Figure 2-3: Diagram showing si de view of a Helmholtz resonator and its mechanical equivalent of a mass-spring-damper system.

PAGE 44

30 A conventional Helmholtz resonator can be lumped into three distinct elements. The neck of the resonator constitutes a pipe through which frictional losses are incurred. Additionally the air that is moving through the neck possesses a finite mass and thus kinetic energy. Therefore the neck has both di ssipative and inertial components. The air in the cavity is compressible and stores pot ential energy, and is th erefore modeled as a compliance. The acoustic compliance of the cavity and effective mass of the neck can be derived from first principles. [129] As mass flows into the bulb, the volume, V, remains constant, assuming the walls are rigid, and so the pressure must rise, by conservation of mass. kg sodt dM VQmassflowrate dtdt (2.15) where M is the mass in the bulb, o is the mean density of the air, and 2Qua is the volumetric flow rate or volume velocity, where u is the velocity. If the disturbance is harmonic and isentropic then 2 2 00 200cQ Pc jV (2.16) where oc is the isentropic speed of sound of the medium and 1j The linearized momentum equation fo r a lossless medium is given by 0u P t (2.17) where P is the acoustic pressure. Assuming a linear pressure gradient yields 120u PPl t (2.18)

PAGE 45

31 where l is the length of the neck. Substituting for 2P yields the following equation. 2 000 1QcjQl P jVS (2.19) Factoring Q, this can be rewritten as 11aN aCPQjM jC (2.20) where the effective acoustic compliance, aCC, of the cavity is 3 2 00 aCVm C cPa (2.21) and effective acoustic mass, aN M of the air in the neck is given by 4 o aNl kg M Sm (2.22) The notation for the lumped elements has been determined through the use of three components. First, the primary variable name is determined by the element type. Next, the first subscript represents, the energy domain in which the element has been defined, where the subscript a represents the acousti c energy domain, e represents the electrical enegy domain, and m represents the mechanical energy domain. The final subscript, which has been capitalized for easier reading, represents the actual structure that is represented at least in part by the lumped element. In this instance, C represents, the cavity, while N stands for the resonator neck. Later, a compliant diaphragm will be introduced that will be represented by a D subscript. The e xpressions above do not account for any viscous damping effects that occur in the He lmholtz resonator neck. The

PAGE 46

32 viscous damping represents a resistance, whos e value can be approximated from pressure driven, laminar pipe flow as 248 aNlkg R Sms (2.23) where is the dynamic viscosity of the air. Furthermore, the viscous damping produces a non-uniform axial velocity profile in the n eck that ultimately leads to an additional factor of 4/3 in the expression for the e ffective acoustic mass [ 130]. The corrected effective mass is then given as 44 3o aNl kg M Sm (2.24) Additionally, the effec tive resistance and ma ss of the neck are, in fact, non-linear and frequency dependent due to turbulence a nd entrance/exit effect s, [131] however for simplicity in modeling, these are not considered here. The expression for cavity compliance given by Eq. (2.21) can be compared to an approximation based on the exact ex pression for the impedance in a short closed tube [2]. The exact expression is given by 2cotoo injc Z kl a (2.25) where 0k c is the wavenumber. Using a Maclauri n series expansion of the cotangent function yields 3111 cot... 345 klklkl kl (2.26) For kl << 1, the impedance can be approxima ted by keeping only the first couple of terms in the expansion, yielding

PAGE 47

33 2 2 22 23 3ooooooo incklccV Zjj jklaajV a (2.27) From this expression, we once again see that 3 2 00 aCVm C cPa (2.28) We now also have an additional mass term, given by 0 2 4 2kg m 3aCV M a (2.29) which is equal to one-third the acoustic mass of the cavity. This correction term is small for 1kl but becomes more prominent as kl increases. At 1kl the correction term is 33.3% of the primary term, while at 0.1 kl the correction term is only 0.33% of the primary term. Equivalent Circuit To create an equivalent circuit model fo r the Helmholtz resonator, knowledge of how to connect these lumped elements is ne eded. Connection rules between elements are defined based on whether an effort-type vari able or a flow-type variable is shared between them [132]. Whenever an effort variab le, such as force, voltage or pressure, is shared between two or more elements, those elements are connected in parallel in the equivalent circuit. Conversely, whenever a common flow (i.e., velocity, current, or volume velocity) is shared be tween elements, those elements are connected in series. These connection rules are used to obtain th e equivalent circuit representation for the Helmholtz resonator, as shown in Figure 2-4. The connection rules, as given, are assuming that what is known as an impeda nce analogy is employed. If an admittance

PAGE 48

34 analogy were used instead, then the connecti on rules would be reve rsed from what is described above. Figure 2-4: Equivale nt circuit representation of a Helmholtz resonator. The frequency response function 21PP represents the pressure amplification of the resonator. It is the ratio of cavity pre ssure to incident pressure, and is given by 2 11 1aC aNaN aCsC P P RsM sC (2.30) where sj From an analysis of the above circuit, a single resonant peak is expected in this frequency response function, when the sum of the reactances is zero, and is given by 1 2res aNaCf M C Hz. (2.31) At the resonant frequency, the pressu re amplification reaches a value of 2aN resres aNM PAf R. (2.32) This is shown in Figure 2-5, for an arbi trary Helmholtz resonator having a neck length and diameter of 3.18 mm and 4.72 mm respectively, and a cavity volume of 1950 mm3. The single peak in the pr essure amplification freque ncy response represents the single degree of freedom present in the system.

PAGE 49

35 Figure 2-5: Theoretical pressure amplification of a conve ntional Helmholtz resonator. (a) magnitude (b) phase. Piezoelectric Composite Plate On the back wall of the Helmholtz res onator, a piezoelectric diaphragm can be placed such that the acoustic pressure in the cavity can be exploited to cause a deflection of the plate and ultimately an electrical signal via piezoelectric transduction. The piezoelectric composite plate, consisting of a ci rcular layer of silic on and an annular ring of PZT, is shown in crosssection in Figure 2-6, where E is theYoungs modulus, is Poissons ratio and is the density of the silicon and P ZT as indicated by the subscript. Additionally, s t is the thickness, and 2 R is the outer radius of the silicon, while for the PZT, pt is the thickness, is the relative permittivity, 31d is the transduction coefficient for a voltage applied across the piezoelect ric causing a displacement in the radial (a) (b)

PAGE 50

36 direction, and p is the residual stress. A region of high stress exists in a circular plate near the clamped boundary during deflection of the plate. By placing the PZT in this region, the electromechanical c oupling is increased due to th e stress concentration. In addition, the annular structure facilitates the connection of elec trodes and bond pads, as the metal lines can be run from the bond pads to the outer radius of the annular structure and therefore do not have to be placed across the surface of the diaphragm. Figure 2-6: Cross-sectional a nd 3-D schematic of piezoelectric composite circular plate. (not to scale) Piezoelectricity Mathematically, the linear piezoelect ric effect is expressed as [133] E ijijklklkijkSsTdE (2.33) and

PAGE 51

37 T iiklklikkDdTE, (2.34) where ijS is the mechanical strain, E ijklsis the elastic compliance 2mN at constant electric field, klT is the mechanical stress 2/Nm, and kijd is the piezoelectric coefficient / or /CNmV, iD is the electric displacement 2/Cm, T ik is the electric permittivity at constant stress Fm, and kE is the electric field /Vm [133]. These equations can also be rewritten using a reduced notation convention, by replacing ij or kl by p or q where i, j k, and l can only have values of 1,2,or 3 while p and q can have any value between 1 and 6. Using this notation, the resulting equations are E ppqqkpkSsTdE (2.35) and T iiqqikkDdTE. (2.36) The subscripts in the reduced notation equatio ns refer to the compone nt of each variable in a specified direction as define d by Figure 2-7. For example, 31d is the piezoelectric coefficient relating electrical displacement in the direction (z-a xis) to a mechanical stress in the direction (x-axi s). From Eq. (2.33), it is also seen to relate a mechanical strain in the direction (x-axis) with an electric field in the direction. 2 (y) 3 (z) 1 (x) 4 5 6 Figure 2-7: Notation of axes us ed in piezoelectric transduction.

PAGE 52

38 One class of materials, ferroelectrics, e xhibits the property th at the direction of their polar axis can be change d through application of an exte rnal electric field [5, 6, 134]. The polar axis is the direction along wh ich a polar molecule exhibits an internal electric field. When the external electric fi eld is then removed, the polar axis remains in an altered direction. This process that cau ses a long term rotation of the polar axis is commonly referred to as poling. All ferro electric materials are also piezoelectric, however they are not naturally piezoelectri cally active, as the crystal structure is centrosymmetric, as in Figure 2-8a. PZT is a typical example of a ferroelectric material, and has a perovskite crystal structure as show n in Figure 2-8. By applying an external electric field to rotate the polar axis, a non-centrosymmetric crystal structure is created and the ferroelectric material becomes piezoel ectrically active, as in Figure 2-8b. Additionally, as a piezoelectrically active material is brought close to its Curie temperature, it loses its piezoelectric properties as the dipoles relax to their unpoled state. a) b) Figure 2-8: Idealized perovskite crystal stru cture for PZT. a) centrosymmetric structure prior to poling. b)non-centrosymmetric structure after poling.

PAGE 53

39 A piezoelectric ceramic, such as PZT, consists of individual domains. The polarization within each domain is in th e same direction; however, the polarization direction varies randomly from domain to do main, leading to a net polarization of zero. This is illustrated in Figure 2-9(a), for th e unpoled material. After poling the material with a sufficiently high electric field, the di poles are rotated so that the polarization of each domain is in the same general direction, as defined by the poling direction, as is illustrated in Figure 2-9(b). Raising the temperature during poling enhances the polarization and resulting piezoel ectric properties, as the dipoles rotate more readily at higher temperatures [6]. Figure 2-9: Schematic of the poling process: a) An unpoled piezoel ectric material, where the polarization in each domain is randomly oriented. B) The same material, after poling, where the domains are now polarized mostly along the direction of the applied field. (after Setter, pg 6 [6]) An additional property of ferroelectric materi als is the double-valued nature of their response to an electrical excitation, resulti ng in a hysteretic behavi or. Hysteresis is a property of systems that do not react instantly to an applied force and may not return to

PAGE 54

40 their original state when the force is removed. In ferroelectrics, when an electric field via a voltage is applied, a polarization is in curred. Upon removal of the voltage, the polarization decreases but does no t return to zero. The result is a remanent polarization, rP. If the voltage is swept up and down, the result will be what is known as a hysteresis loop, shown in Figure 2-10. Figure 2-10: Polarization vs. electric field hysteresis l oop (after Xu, pg 10 [134]). The graph is displayed here as polarization, P, typically given in 2Ccm versus applied electric field, E although other variations do exist. There are four figures of merit shown in the illustration. mP is the maximum polarization measured, while rP is

PAGE 55

41 the remanent polarization (i.e. the polarizatio n which remains when the field is returned to zero). Furthermore, s P is known as the spontaneous pol arization and is defined as the straight line extrapolation of the line defined by the upper saturation region. Finally, the coercive field, cE, represents the magnitude of th e field required to cancel out the remanent polarization. Composite Plate Modeling In order to accurately model this structure, the material properti es of each of the layers were considered. These include the m echanical properties of the layers such as Youngs Modulus, E and Poissons ratio, and the dielectric constant, and piezoelectric coefficient, 31d, of the PZT. The subscript p or s indicates the layer as PZT or silicon respectively. Furthermore, th e geometry of the structure is accounted for in the values for inner PZT radius, 1 R outer PZT radius, 2 R PZT thickness, pt, and silicon thickness, s t. Lumped element model A pressure applied to the plat e creates a deflection of the plate, resulting in a stress in the piezoelectric layer. The stress de forms the piezoelectric layer, creating an electrical charge on the elec trodes, thus generating a vo ltage across the piezoelectric layer. Although the stress and charge are dist ributed over a finite regi on of the plate, it is useful to first look at one-dimensional (1-D) piezoelectric transduction. This 1-D analysis can then be extended to incorporat e effective lumped element values that are calculated from the actual distributed case. In the 1-D piezoelectric transduction, the piezoelectric material displaces longitudinally due to the application of a force, F, and/or a voltage, V, applied in the

PAGE 56

42 thickness direction. Additionally, a charge, q, is generated by the application of the same force and/or voltage. The constitutive equations, Eq. (2.33) and Eq. (2.34), can be extended to this situation by modifying their form. When looking at the resulting strain and el ectric displacement in the direction, for an applied stress and electric field in the same direction, the equations reduce to 333333 ESsTdE (2.37) and 333333 TDdTE (2.38) Then this set of equations can be converted by multiplying both sides by the thickness of the piezoelectric material. 333333 3333 E EtSsTdE tSstTdtE (2.39) Now, the mechanical compliance of a ma terial under compression in the thickness direction can be defined as 33 E mst C A, (2.40) where t is the thickness, and A is the area over which the force is applied. Additionally, for a constant electric field, E, V E t (2.41) Substituting (2.40) into second part of (2.39) yields, 33 m x CATdV (2.42) Finally, the stress, T, when applied over the area, A can be equated to a force, F, given by

PAGE 57

43 FTA (2.43) leading to the final equation given by 33 m x CFdV (2.44) A similar procedure can then be performed on Eq. (2.38), by multiplying both sides by the area, and making similar su bstitutions as before, giving 333333TADdTE, (2.45) 3333TqdFAE, (2.46) 33 33TA qdFV t, (2.47) so 33efqdFCV (2.48) Thus a pair of equations can be written desc ribing the 1-D piezoelec tric transduction, and is given by 33 33ms efCd x F dC qV (2.49) In this pair of equations, msC, the mechanical compliance when a short is placed across the electrodes of the piezoelectric, is defined as 0|V msx C F, (2.50) while efC, the electrical capacitance when the piezoelectric is free to move, is defined by 33 0p F ef pq A C Vt, (2.51)

PAGE 58

44 where q is the resulting charge from the applied voltage V, 33 is the electrical permittivity in the thickness direction, pA is the area of the piezoelectric, and ptis the thickness of the piezoelectric. Furthermore, 33d is the piezoelectric coefficient relating the displacement in the thickness direction for an applied voltage in the same direction, when no mechanical force is applie d across the piezoelectric, defined by 0 33|Fx d V (2.52) For the case of the piezoelec tric composite plate that is presented here, though, there is a distributed, nonuniform deflection acr oss the plate. In or der to apply the 1-D model developed above to this situation, it is necessary to lump th e actual distributed deflection to a single point and compute eff ective values by equating the energy in the distributed system to the energy in a corresponding lumped component. The 1-D electromechanical transduction described above can then be extended to relate the acoustical and electrical energy domains by integrating over the areas and switching to acoustic conjugate power values, as well as incorporating time-harmonic signals [64]. The time-harmonic, electroacoustical equi valent of Eq. (2.49) is given by efA AasjCjd I V jdjC QP 3A ms (2.53) whereP is the acoustic pressure, Q is the volume velocity of the plate, I is the electrical current. Furthermore, Ad is the piezoelectric coefficient relating the volumetric displacement to the applied voltage, when pressure equals zero, and defined by 20 0 02R P A Pwrrdr Vol d VV 3m V (2.54)

PAGE 59

45 where wr is the deflection as a func tion of the radial position, 2 R is the radius of the plate and Vol is the volumetric displacement defined by 202RVolrwrdr3m (2.55) Additionally, asC the acoustical compliance when a s hort circuit is placed across the piezoelectric, is defined by 20 0 02R V as Vwrrdr Vol C PP 3m Pa (2.56) From two-port network theory, a generali zed electro-acoustic, reciprocal, two-port network can be written as ef asYG I V GY QP (2.57) where efY is the electrical admittance when the aco ustic terminal is free to move (i.e no pressure exists, 0P ), asY is the acoustical admittance when the electrical terminal is shorted (i.e. no voltage exists, 0V ), and G is the electro-acoustic transduction admittance, given both by the ratio of current, I to pressure, P when 0V, and the ratio of volume-velocity, Q, to voltage, V, when 0P Comparing Eq. (2.53) and Eq. (2.57), it can be seen that AGjd (2.58) efefYjC (2.59)

PAGE 60

46 and asasYjC (2.60) Equivalent circuit Using Eq. (2.57), an equivalent electro-ac oustic circuit can be drawn, as shown in Figure 2-11, where ebC the electrical capacitance when the plate is blocked from moving is given by 21ebefCCk F, (2.61) the transduction factor, is given as AA asasas j dd G YjCC Pa V (2.62) and the coupling factor, k, is given by 2 2 A efasefasd G k YYCC (2.63) Physically, the coupling factor, k, represents the fraction of energy that is coupled between the acoustical and electr ical energy domains. It is se en here to be a function of the piezoelectric coefficient, Ad, which represents coupled energy and the two elements which store potential energy asso ciated with the transduction, efC and asC. The coupling factor is therefore related to the ratio of the coupled energy to the stored energy. Figure 2-11: Electro-acous tic equivalent circuit representation with ebC, asC and

PAGE 61

47 An alternative equivalent circuit can also be drawn, as shown in Figure 2-12. The alternative circuit is simply another repr esentation for the same physical process. Figure 2-12: Electro-acous tic equivalent circuit representation with efC, aoC and In Figure 2-12, aoCrepresents the acoustic complia nce when an open-circuit is placed across the piezoelectric, and is given by, 21aoasCCk 3m Pa (2.64) Additionally, the transduction factor, is given by AA efefefjdd G YjCC V Pa (2.65) The two circuits above repres ent the transduction under sta tic conditions, thus they do not take into account the dynamics of the sy stem (i.e. mass). At frequencies greater than zero, a mass must be added to the equiva lent circuit. As the mass stores kinetic energy, it is represented by its electrical equi valent, which is an i nductor. Furthermore, this mass must be in series with the complian ce, as both experience the same motion (i.e displacement or velocity). The acoustic mass,aD M is determined by equating the lumped kinetic energy of a point mass moving with the center velocity to the total kinetic energy of the vibrating diaphragm and is therefore given by 2 2 0 0 02R V aDA Vwr M rdr V (2.66)

PAGE 62

48 where A is the areal density of the piezoe lectric composite plate defined by, 2 12 z A zdzkgm (2.67) where is the density of th e corresponding layer. In addition to the above mentioned elemen ts, it is also necessary to include an additional element that represents the di electric losses that are incurred during transduction, yielding the circuit shown in Figure 2-13. Generall y this is expressed through a term known as the loss tangent, also known as th e dielectric loss factor. Represented by the ratio of the parallel re actance to the parallel resistance in the dielectric, the loss tangent is formally given by 1 tan 2p pppX R fCR (2.68) where p R is a resistor in para llel with a capacitance, pC and f is the frequency at which p R and pC were measured. The loss tangent is also equal to the inverse of the quality factor, Q Figure 2-13: Dynamic electro-ac oustic equivalent circuit. In this circuit, aD M represents the acous tic mass of the diaphragm. The input impedance from the acoustical side is then given by 21 1p inaD asebpR ZsM sCsCR (2.69)

PAGE 63

49 By voltage division we get 22 211 1 1pp ebpebp p in aD asebpRR sCRsCR VPP R Z sM sCsCR (2.70) which, after rearranging yields, 3221 1asp aDasebpaDaspebasV sCR P sMCCRsMCsRCC (2.71) This is the general expressi on for the open-circuit sensitivit y of the circular composite piezoelectric plate. From a physical standpoi nt it is also useful to look at certain conditions that allow for additi onal insight. The presence of the dielectric loss resistance, p R complicates some of the underlying behavi ors of the device, by adding an additional R-C time constant. Looking at a situation in which this parallel loss resistance is infinite (i.e.pR ), the above expression can be reduced to 221 1as as eb aDas ebC V C PC sMC C (2.72) Several important things can be gleamed from this expression. First of all, the equation describes the behavior of a sec ond order system, with a primary resonance occurring when the denominator goes to zero at 21as eb res aDasC C MC (2.73) From this expression, the resonanc e is seen to depend on the ratio 2 asebCC in addition to the standard dependence on the product aDas M C If 0 this reduces to a

PAGE 64

50 simple plate. For 0 electrical energy is stored acro ss the piezoelectric, resulting in a stiffer device possessing a larger res The full expression that includes the dielec tric loss, given by Eq. (2.71), is the general expression for the sensitiv ity of the circular composite piezoelectric plate. It can be further simplified for two important cases: 1. Well below the short-circuit mechani cal resonant frequency (compliance dominated) 2. At the short-circuit mechanical resona nt frequency (resistance dominated) Case 1: Well below the short-circu it mechanical resonant frequency When a short is placed acrossebC, it is equivalent to assigning 0pR Physically, this implies that no electrical energy stor age is allowed, as the voltage across the capacitor, ebC, must be zero. The expression for th e short-circuit mechanical resonant frequency is given by 1scres aDas M C, (2.74) which is the same resonant fre quency as described above when 0 Note that both cases describe a situation where no electrical energy is stored. The general expression for sensitivity can then be rewritten in terms of s cres as 22 21 1scscasp ebppebas resresV jCR P jj jCRjRCC (2.75) Well below resonance, we have

PAGE 65

51 21scresj (2.76) This term can then be dropped from the expression, yielding 21 1 1as eb as ebpebC V j PC C j CRC (2.77) To simplify this further, the transduction factor, can be written in terms of the effective piezoelectric modulus, Ad, by recalling Eq. (2.62), yielding 2 21 1 1 1 1 1as A aseb as A asebpeb A eb A asebpebC jd V PCC C d j CCRC d C d CCjRC (2.78) Furthermore, the blocked electrical capacitance, ebC, can be written in terms of the free electrical capacitance,efC and the coupling coefficient, k, using Eq. (2.61) and Eq. (2.63) to yield the an expre ssion for sensitivity, given by 2 2 22 2 221 1 1 1 1 1 11 1 1 1 11A eb A asebpeb A eb A asefpef A eb pefd V PC d CCjRC d C d CCkjRCk d C k kjRCk (2.79) which can be simplified further to yield

PAGE 66

52 22 21 11 11 11 1 1 1 1A eb pef A eb p A ef p efpd V PC kjRCk d C jR k d C jR jCR (2.80) Now under the condition of a high value for the loss resistor, p R Eq. (2.80) can be reduced approximately to A efd V PC (2.81) This corresponds to the alternative trans duction coefficient, given by A efd C (2.82) and thus under the conditions of a high loss resistor and operation well below s cres the sensitivity can be given by V P (2.83) Case 2: At the short-circuit mechanical resonant frequency The general expression for sens itivity is once again given by 22 21 1scscasp ebppebas resresV jCR P jj jCRjRCC (2.84)

PAGE 67

53 Now, since we are only considering the syst em at the short-circuit resonance, where s cres we have 21scresj (2.85) yielding an exact expression for the sensitivity as 211 11asp ebppebasV jCR P jCRjRCC (2.86) We therefore find that the sensi tivity at resonance reduces to 21V Pk (2.87) as 2k (2.88) This can be understood from the equi valent circuit of Figure 2-13. At s cres the impedance of the mass is canceled by the impedance of the compliance, permitting a direct transduction of energy fr om the acoustic to electrical energy domain. Note that s cres exists whether or not 0pR therefore Eq. (2.86) is always valid whenever s cres In other words, Eq. (2.86) holds whenever the operating frequency coincides with the frequency that is defined by th e short-circuit resonance, but makes no requirement for an actual short-circuit load. As the load across ebC is increased, the resonance frequency will shift towards the open-circuit resonance. R ecalling Eq. (2.64), the open and short-circuit acoustic compliances are related by the coupling factor and define the limits of the resonant frequency as all resistive loads fall somewh ere between open and short-circuit. The

PAGE 68

54 larger the coupling factor, th en the larger the range betw een open and short-circuit resonant frequencies. General expression vs. simplified cases Figure 2-14 plots the opencircuit and near short-circuit sensitivity versus frequency overlaid with the values of and 1 The material properties used in calculating these results are shown in Tabl e 2-3 and Table 2-4. Notice that below resonance, the general open-circuit sensitiv ity expression asymptot es to the value of Additionally, notice that the genera l expression reaches the value of 1 at a frequency below where the peak amplitude occurs. This is because 1 is the simplified expression for the sensitivity at the short-circuit mechan ical resonant frequency, whereas the general expression represents the open-circuit frequency response. Thus the peak amplitude occurs at the open-circuit resonant frequenc y. The frequency at which the general expression for sensitivity reaches the value of 1 coincides with the short-circuit resonant frequency, as expected. Also note th at this short-circuit sensitivity calculation was performed with an almost-short circuit cond ition, primarily to illustrate the effect on the resonant frequency. To achieve the almost-short circuit condition a resistive load of 1 was used as the value is much less than the impedance due to ebC. One final feature of the frequency response is the low frequenc y rolloff that is visibl e in the open circuit case. This rolloff is due to the dielectric loss resistor,p R and the resulting time constant due to the combination of that resistor and the electrical cap acitance, as evidenced in Eq. (2.80) for the low frequency regime.

PAGE 69

55 Table 2-3: General material properties used in modeling. Material E GPa 3kgm tm Silicon 150 0.27 2300 3 Platinum 170 0.38 21440 0.170 TiO2 283 0.28 2150 0.100 [135, 136] Table 2-4: PZT properties used in modeling. pEGPa p 3 pkgm 31dpmVtan p M Pa 30 0.3 7600 1000 -50 0.02 30 [9, 10, 16, 31] 100 101 102 103 104 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 Frequency [Hz]|Sensitivity| [V/Pa] 1/ Near short circuit frequency response Open circuit frequency response Figure 2-14: Sensitivity vs. frequency for a piezoelectric composite circular plate. Effects of including th e radiation impedance Since the clamped circular plate is vibrat ing in a medium, the radiation impedance of the plate must be taken into account, a nd consists of a radiation mass and radiation resistance. The radiation mass accounts for the in ertial mass of the fluid that is vibrating in unison with the plate, while the radiation resistance accoun ts for the acoustic radiation

PAGE 70

56 of energy away from the plate. In addition to the radiation resistance, energy is lost via structural radiation to the supports. The radiation resistance and radiation mass is determined to first-order by approximating the ba ckplate as a piston in an infinite baffle, as given by Blackstock [2] in term s of specific acoustic impedance, p Z as 11 112222 122 22pooooJkaKka Z cjcRkajXka kaka (2.89) where 12 R ka and 12 X ka indicate that they are functions of 2ka, 2aR is the radius of the piston,1J is a Bessel function of the first kind of order one, and 1K is a firstorder Struve function. The Maclaurin expansio ns of Eq. (2.89) are also given by [2]as 246 1 222121231234kakaka R (2.90) and 35 1 22222 42 335357kaka ka X (2.91) For small values of ka, where 1ka, the resistance and reactance can be approximated by keeping only the first terms of Eq. (2.90) and Eq. (2.91). The radiation impedance as given in Eq. (2 .89) is in terms of specific acoustic impedance. This is converted to an acoustic impedance via the effective area, effA, of the piezoelectric plate. As the ci rcular composite diaphragm doe s not deflect uniformly over the entire surface (i.e. the deflection is a function of radial distance), the volume displaced by the deflection of the diaphragm is less than that of a circular piston of the same area. The effective area, effA, is therefore defined to represent the equivalent area that a circular piston with uniform deflecti on would need to have to create the same

PAGE 71

57 volumetric displacement as the composite diaphr agm. This is necessitated by the need to maintain continuity of volume velocity ac ross the interface between the mechanical and acoustical domains. It can be calculated by integrating the distri buted deflection of the diaphragm over the entire surface and then dividing by the center deflection of the diaphragm. By using the center deflection as the reference value, the distributed deflection is then considered to be lumped to the center. For an axisymmetric deflection, such as would occur for the fundamental mode, the effective area is given by 22 0effwrrdr Am w (2.92) The radiation resistance can then be approximated, for low values of kaas 2 4,1 2o aDrad effkac kg Rka Ams (2.93) while the radiation mass is approximated as 48 ,1 3o aDrad effkac kg Mka Am (2.94) These two elements add in series to create a radiation impedance, rad Z that is defined by aDradaDradaDradZRsM (2.95) The equivalent circuit of the composite plat e including the radiati on impedance is shown in Figure 2-15. Figure 2-15: Equivalent circ uit of piezo-composite plat e including the radiation impedance.

PAGE 72

58 When the radiation impedance of the diaphr agm is added to the equivalent circuit, the analysis becomes only slightly more co mplex. The acoustical input impedance,in Z defined as PQ, can be represented in terms of the equivalent circuit parameters, and is now given by 21 1p inaDaDradaDrad asebpR ZsMsMR sCsCR (2.96) Then, the general expression for sensitivity is now given via the voltage divider method as 22 211 1 1pp ebpebp p in aDaDradaDrad asebpRR sCRsCR VPP R Z sMsMR sCsCR (2.97) or 21 1 1p ebp p aDaDradaDrad asebpR sCR V R P sMsMR sCsCR (2.98) From Figure 2-15, it can be s een that the radiation mass,aDradM adds directly to the acoustical mass of the plate,aD M as is evidenced in Eq. (2.97), while aDradR provides damping to this second-order system. Note that here, the acoustical resistance, aDradR damps the resonance, while the electrical resistance, p R leads to a low frequency rolloff. The different effects arise because aDradR is in series with the reactive elements, asC and

PAGE 73

59 aD M while p R is in parallel with the capacitance, ebC The general expression for the undamped resonance frequency is now given by 21as eb res aDaDradasC C M MC (2.99) When a short is placed acrossebC it is effectively removed. This is mathematically equivalent to assigningebC as the impedance of a capacitor is inversely proportional to the capacitance. Th e short-circuit resonance frequency then reduces to 1res sc aDaDradas M MC (2.100) Upon comparison with Eq. (2.74), it can be seen that the shor t-circuit resonance frequency has now been shifted downward by the radiation mass, as is also the case with the open-circuit resonance frequency. At the radial frequency of the short-circuit resonance, the short-circuit input impedance reduces to inaDradZR (2.101) because ebC and p R are effectively removed by the short-circuit. The volume velocity, Q is then given by aDradP Q R (2.102) Via the piezoelectric tr ansduction, the current, I in the piezoelectric material, is then given by aDradP IQ R (2.103)

PAGE 74

60 The output voltage in this case, however, rema ins at zero due to the short circuit placed across the output capacitance. To achieve real power at the output, a finite resistive load is necessary. Adding a resistive load, load R across the output yields the equivalent circuit shown in Figure 2-16. Figure 2-16: Equivalent circ uit with resistive load. The input impedance, in Z is then given by 21loadp inaDaDradaDrad asloadploadpebRR ZsMMR sCRRRRsC (2.104) Now, if we choose load R such that 1loadeb R sC and loadp R R, then the parallel combination of these three elements can be approximated with just load R In addition, the resonant frequency will be very close to the short-circuit resonant frequency. At this frequency, the input impedance reduces to 2 inaDradload Z RR (2.105) and the volume velocity, Q is given by 2inaDradloadPP Q ZRR (2.106) Through piezoelectric tran sduction, a current, I is created in the piezoelectric given by 2aDradloadP IQ R R (2.107)

PAGE 75

61 As the resistance,load R is much less than 1ebsC most of the current goes through it, leading to a voltage drop given by 2load load aDradloadPR VIR RR (2.108) Thus the power absorbed by the load resi stance is purely r eal and is given by 22 2 2ReReload aDradloadPR IV RR (2.109) An optimal solution to this equation is found by setting 2 22 3 20aDradload load aDradloadRR d P dR RR (2.110) Solving for the optimal load resistance yields 2aDrad loadR R (2.111) which is just the impedance matching condition at the interface. Note that is for the special case of 1loadeb R sC, loadp R R and operation at resonance. Equivalent circuit parameters of piez oelectric composite circular plates In order to obtain the equivalent circ uit parameters, an analytical model was developed for the piezoelectric composite circ ular plate by Wang et al. [62, 63]. Using this approach, analytical modeling was acco mplished by dividing the problem of Figure 2-6 into two portions, an inne r circular plate, surrounded by an annular composite ring with matching boundary conditions at the interface, as shown in Figure 2-17. The boundary conditions consist of equal moments a nd forces at the interface as well as equal slope and transverse displacement. After so lving for the deflection in each region, the deflection equation for each region can then be combined [62, 63].

PAGE 76

62 Figure 2-17: Conceptual 3-D and cross-sec tional schematic of th e circular composite plate. (Not to scale.) The deflection equation can then be utili zed to determine the potential and kinetic energy stored in the plate, leading to expressions for the acoustic compliance and mass of the composite plate. Similarly, expre ssions can be found for the electro-acoustic transduction coefficient and the blocked elec trical capacitance. Using the parameters shown in Table 2-3 and Table 2-4, plots were obtained for the lumped elements as a function of both 12 R R and pstt As many of the material parameters for PZT are highly dependent on actual processing conditi ons and techniques, typical values were chosen as a best guess estimate [8, 12-14, 1648]. The effective acoustic short-circuit

PAGE 77

63 compliance and mass are functions of both 12 R R and pstt and are shown in Figure 2-18 and Figure 2-19, respectively. 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.5 1 1.5 2 2.5 3 3.5 x 10-12 R1/R2Cas [m5/N] tp=0.6 m tp=3.0 m Figure 2-18: Effective acoustic shortcircuit compliance as a function of 12 R R and pstt For these calculations, 3stm and 22 R mm while the piezoelectric layer has thicknesses of 0.6, 1.2, 1.8, 2.4, 3.0 ptm The acoustic compliance is found to increase with increasing 12 R R and decrease with increasing pstt however, the acoustic mass is f ound to generally decrease with 12 R R and increase with pstt It is also useful to look at the physical limits and effects of 12 R R As 120 RR the piezoelectric ring covers the entire surface of the diaphragm and the compliance is at a relative minimum for a given thickness, while the mass is at relative maximum. Furthermore, note that the acoustic mass is most sensitive

PAGE 78

64 to changes in the piezoelectric layer thic kness under this condition. Meanwhile, as 121 RR the piezoelectric ring would have in finitesimal width and so would be essentially nonexistent. U nder this condition, the piezoel ectric layer thickness has no effect on the compliance, as would be expected. 0 0.2 0.4 0.6 0.8 1 500 1000 1500 2000 2500 3000 3500 R1/R2MaD [kg/m4] tp=0.6 m tp=3.0 m Figure 2-19: Effective acoustic mass as a function of 12 R R and pstt For these calculations, 3stm and 22 R mm while the piezoelectric layer has thicknesses of 0.6, 1.2, 1.8, 2.4, 3.0 ptm Shown in Figure 2-20 is a graph of the resonant frequency as a function of both 12 R R and pstt From this figure, it can be seen that the resonant frequency increases as the thickness of the piezoelectric layer incr eases or generally as the inner radius of the piezoelectric layer decreases. The resonant frequency trend with respect to thickness arises because the effective acoustic mass is found to increase with the thickness while the effective acoustic compliance decreased, but by a much larger amount. Again, it

PAGE 79

65 should be noted that at the limit of 121 RR the resonant frequency is no longer affected by the thickness of the piezoelectric layer. Also, for a given pt there is a maximum resonant frequency at 120.4 RR. 0 0.2 0.4 0.6 0.8 1 2000 4000 6000 8000 10000 12000 14000 R1/R2Resonant Freq [Hz] tp=0.6 m tp=3.0 m Figure 2-20: Short-circuit resona nt frequency as a function of 12 R R and pstt For these calculations, 3stm and 22 R mm while the piezoelectric layer has thicknesses of 0.6, 1.2, 1.8, 2.4, 3.0 ptm The electro-acoustic transduction coefficient, was also found as function of these relative dimensions, and is shown in Figure 221. The transduction coefficient represents the conversion of the voltage across the piezoelectric to th e acoustic pressure produced by the resulting motion of the composite plate. It should be noted that the transduction coefficient is negative, implying a 180 phase shift between pressure and voltage, as

PAGE 80

66 computed relative to an assumed poling direct ion. As can be seen in Figure 2-21, a maximal magnitude occurs when 12 R R is around 0.45. 0 0.2 0.4 0.6 0.8 1 -6 -5 -4 -3 -2 -1 0 R1/R2a [Pa/V] tp=0.6 m tp=3.0 m Figure 2-21: Electro-acoustic transduc tion coefficient as a function of 12 R R and pstt For these calculations, 3stm and 22 R mm while the piezoelectric layer has thicknesses of 0.6, 1.2, 1.8, 2.4, 3.0 ptm Additionally, the magnitude of the transduction coefficient increases with increasing piezoelectric thickne ss. Looking at the limit as 121 RR the transduction is seen to decrease to zero, as would be expected of a st ructure with no piezoelectric material. Furthermore, as 120 RR and the piezoelectric material covers the entire diaphragm, the transduction factor is seen to go to zero as well. Physically, this results from cancellations between different regi ons of the diaphragm that are undergoing opposite polarities of stress, such that the net electr ic displacement (charge) is zero.

PAGE 81

67 Another parameter of interest is the coupling coefficient, k. It is defined as the ratio of energy converted by the transducer to the energy supplied to the transducer, thus providing a measure of the coupling. For the purposes of harvesting energy, this parameter is more important than the transduction coefficient as the primary interest is in maximizing the coupled energy, rather than maximizing only the output voltage. A plot of k as a function of 12 R R and pstt is shown in Figure 2-22. 0 0.2 0.4 0.6 0.8 1 0 0.002 0.004 0.006 0.008 0.01 0.012 0.014 R1/R2k tp=0.6 m tp=3.0 m Figure 2-22: Electromechanical coupling coefficient, k, as a function of 12 R R and pstt For these calculations, 3stm and 22 R mm while the piezoelectric layer has thicknesses of 0.6, 1.2, 1.8, 2.4, 3.0 ptm From the figure, the maximum k of 0.013 is obtained for an 12 R R ratio of 0.95 and a pstt ratio of unity. Physically, the increasing trend in k as 12 R R increases, arises from the stress distribution within the diaphragm. There is a stress concentration near the clamped boundary of the diaphragm that provi des for a high level of coupling. By

PAGE 82

68 concentrating the piezoelectric material in th is high stress region, the averaged coupling factor over the ring will be hi gher than a piezoelectric ring that is spread over a larger area. Acoustic Energy Harvester Dynamic Behavior When the piezoelectric diaphragm is mounted in the wall of a Helmholtz resonator, the equivalent circuit for the diaphragm (Fi gure 2-16) is combined with that of the Helmholtz resonator (Figure 2-4). By usi ng the same lumped element connection rules previously described, the complete circuit can be obtained, as shown in Figure 2-23. MaDCaD :1 Ceb MaDrad RaDrad + P + V Q I Rload RaNMaN CaC Figure 2-23: Equivalent circ uit for acoustic energy harvester with resistive load. Notice that the equivalent circuit for the diaphragm is in parallel with the cavity compliance, aCC This occurs because the pressure in the cavity is responsible both for compressing the air in the cavity as well as de flecting the diaphragm, i.e. the cavity and the diaphragm both see the same pressure Additionally, it is assumed that loadp R R and therefore the parallel combination can be approximated as //loadpload R RR Acoustical Input Behavior An expression for the acoustical input impedance can be obtained from the equivalent circu it and is given by 2 211 1 11 1load aDaDradaDrad aCaDloadeb inaNaN load aDaDradaDrad aCaDloadebR sMMR sCsCRsC ZRsM R sMMR sCsCRsC (2.112)

PAGE 83

69 From this equation, it can be seen that the total input impedance is simply the impedance of the Helmholtz resonator neck in series wi th a parallel combination of the Helmholtz resonator cavity impedance and the piezoelect ric composite diaphragm impedance with a resistive load attached. Many energy harvesting devices with vary ing geometries were designed for this dissertation, but for illustrati ve purposes, I will only expl ore the behavior of one, which serves as a typical example. More explicitly eight different devices were designed and created to give a range in performance, but to illustrate the qualitative and typical behavior of the devices, only a single representativ e device was plotted for the remainder of this chapter. A plot of th e acoustical input impedance versus frequency is shown in Figure 2-24 for the piezoelectric co mposite diaphragm, both with and without the Helmholtz resonator, to elucidate the beha vior of the individual components. For this plot and those that follow, the Helm holtz resonator has a neck length of 3.18 L mm, and a radius of 2.36 n R mm along with a cavity volume of 31950 cavVmm in addition to a diaphragm with a thickness of 3 sitm and a outer and inner radius of 21.95 R mm and 21.85 R mm respectively. These dimensions were chosen as typical values in the range of what was expected of the final device and package design. From this plot, it can be seen that, by itself, th e piezoelectric composite diaphragm has a single resonance near 3.6 kHz where the impedance reaches a lo cal minimum. When combined with the Helmholtz resonator, two minima are seen. The lower resonance that occurs near 1.8 kHz is dominated by the Helmholtz resonato r that has an uncoupled resonance of 2 kHz as evidenced by the peak in Figure 2-5 The upper resonance at 3.9 kHz is

PAGE 84

70 dominated by the piezoelectric composite di aphragm. Additionally, an antiresonance, where the impedance reaches a local maxima, occurs be tween the two resonances. Figure 2-24: Magnitude of the acoustical input impedance for the piezoelectric composite diaphragm and in combination with the Helmholtz resonator. (3.18 Lmm 2.36 n R mm 31950 cavVmm 3 sitm 21.95 R mm 11.85 R mm ) The acoustical input power can be obtain ed from the input acoustic pressure, P, and is given by 2Re Re{}in inP Z (2.113) The input acoustic pressure, P is measured inside the planewave tube near the end-face of the tube. The microphone that measures th is pressure is placed as close as possible ( ~116 ) to the end-face so that it se rves as a measure of the pressure that is incident on the energy harvesting device. In the case of the device that includes a Helmholtz

PAGE 85

71 resonator, the incident pressure is the pre ssure incident upon the resonator neck, while for the diaphragm only device, it is the pressure incident upon the face of the diaphragm. A plot of the magnitude and phase of th e input power is shown in Figure 2-25. This plot assumes an ideal ac oustic source, i.e. that the applied acoustic pressure is constant and independent of the input impedance. In reality, this is not the case, and the applied acoustic pressure will vary as a function of frequency due to a frequency dependent impedance as seen by the speaker. Th e important point to take away from this plot, however, is that there are frequencies at which the input power will be maximized. These frequencies occur when the input impedance is minimized thus allowing for maximum power flow from an ideal sour ce. These do not necessarily correspond to frequencies where the output power is maximized as will be shown in the next section. Electrical Output Behavior The electrical output impeda nce can be found in a similar manner and is given by 21111 //////outloadaDaDradaDradaNaN ebaDaCZRsMMRRM sCsCsC .(2.114) Expanding this equation out yields the full e xpression for the output impedance, given by 2 21 111 1 1 1111 1aNaN aC loadaDaDradaDrad ebaD aNaN aC out aNaN aC loadloadaDaDradaDrad ebebaD aNaN aCRM sC RsMMR sCsC RM sC Z RM sC RRsMMR sCsCsC RM sC .(2.115)

PAGE 86

72 Figure 2-25: Magnitude and phase of the ac oustical input power for an input acoustic pressure of 94 PdB. (3.18 L mm 2.36 n R mm 31950 cavVmm 3 sitm 21.95 R mm 11.85 R mm ) A graphical plot of the electrical output impedance is shown in Figure 2-26 in terms of real and imaginary components. Th e solid curve in the figure corresponds to the expression given by Eq. (2.115), while the dotte d curve represents the impedance of the piezoelectric composite diaphragm by itself. Notice that only the real component shows any significant difference between the two cases Physically, this is due to the poor coupling that is typical of indirect transducer s. The result is that acoustical components have minimal effect on the electrical impedance.

PAGE 87

73 Figure 2-26: Electrical output impedance fo r the piezoelectric composite diaphragm by itself and in combination with the Helmholtz resonator. (3.18 L mm, 2.36 n R mm 31950 cavVmm 3 sitm 21.95 R mm 11.85 R mm ) The output voltage can easily be found from the equivalent circuit to be 21 1 1 1 1 1load eb load eb aNaN load aDaDradaDradaNaN aDloadeb aCR sC P R sC V RsMR sMMRRsM sCRC sC .(2.116) The output voltage is displayed graphical ly in Figure 2-27 for the piezoelectric composite diaphragm by itself and in combination with the Helmholtz resonator. For both cases, the magnitude has a similar shap e to the output impedance, although a few differences remain. In the output impedance curve, the second resonant peak is higher than the first, whereas for the output voltage curve shown below, the first resonant peak

PAGE 88

74 is higher than the second. The voltage curve shown was computed for a load resistance, load R equal to the electrical output resistance at the diaphragm resonance frequency, and an input acoustic pressure of 1 Pa. Figure 2-27: Magnitude and phase of the outpu t voltage for the piezoelectric composite diaphragm and in combination with the Helmholtz resonator. (3.18 Lmm 2.36 n R mm 31950 cavVmm 3 sitm 21.95 R mm 11.85 R mm 94 PdB1) The electrical power delive red to the resistive load, load R can be found from 2Re{}out loadV R (2.117) A plot of the electrical out put power delivered to the resistive load is shown in Figure 2-28 for the piezoelectric composite diaphrag m by itself and in combination with the Helmholtz resonator, for the case of an incident acoustic pressure of 1 Pa 1 re 20dBPa throughout this document.

PAGE 89

75 Figure 2-28: Magnitude and phase of the el ectric output power delivered to the load resistor. (3.18 Lmm 2.36 n R mm 31950 cavVmm 3 sitm 21.95 R mm 21.85 R mm 94 PdB ) The overall power conversion efficiency, can then be found as the ratio of output electrical power to i nput acoustical power, given by Re{} Re{} Re{}out in (2.118) The magnitude and phase of the efficiency is shown in Figure 2-29 for the piezoelectric diaphragm by itself and in combin ation with the Helmholtz resonator. As seen in the plot, the efficiency of the composite diaphragm reaches a peak at the diaphragm resonance, while the efficiency of the coupled system reaches a peak at the second resonance, which is dominated by the di aphragm resonance. It is important to note that the peak magnitude is similar both with and without the Helmholtz resonator, suggesting that little benefit is gained through the use of the Helmholtz resonator, as this

PAGE 90

76 Helmholtz resonator was not optimally designe d for impedance matching. This need not always be the case, as the relative benefits of impedance matching are somewhat offset in varying amounts by the increased resistive loss es due to the Helmholtz resonator. The benefits can thus be enhanced by mainta ining the improved impedance matching while minimizing the additional losses. A perfect impedance match between the plane wave tube and the Helmholtz resonator would im prove the efficiency by approximately 40% over the values shown in the figure. This value is estimated based on comparison to a Helmholtz resonator that is perfectly impedance matched to the plane-wave tube. Figure 2-29: Magnitude and phase of the energy harvester efficiency. (3.18 L mm, 2.36 n R mm 31950 cavVmm 3 sitm 21.95 R mm 21.85 R mm 94 PdB ) Operation at an Optimal Frequency To get a better feel for these devices in th eir intended application as harvesters of energy, it helps to look at the theoretical behavior for a single frequency. For the purposes of illustration, a device with the geom etry of Device 8, as given by Table 2-5 on

PAGE 91

77 page 81 was used and the frequency at which the efficiency is maximized was chosen. This optimal frequency is different betw een the diaphragm mounted on the Helmholtz resonator and the diaphragm strictly by its elf. For the diaphragm mounted on the Helmholtz resonator, a sinusoidal signal at 1966 Hz was applied, while for the diaphragm mounted by itself, a sinusoidal signal at 13.11 kHz was used. For both cases, the signal was swept over a range of acoustic pressures while the open circuit output voltage was determined. A plot of the open circuit voltage versus acoustic pressu re for both cases is shown in Figure 2-30. 115 120 125 130 135 140 145 150 155 160 10-2 10-1 100 Input Acoustic Pressure [dB]Vout [V] Membrane Membrane/HR Figure 2-30: Open-circuit volta ge as a function of the appl ied acoustic pressure for the piezoelectric diaphragm by itself and packaged with a Helmholtz resonator. (Device = 8 from Table 2-5, H R f =1966 Hz mem f =13.11 kHz ) Note the steadily increasing output voltage as the acoustic pressure increases, as would be expected for a linear system such as this. This model, however, does not take into account any nonlinearities that may occur at high acoustic levels. Such

PAGE 92

78 nonlinearities may include, but are not limited to, piezoelectric satura tion effects on the output voltage and large deflec tion effects on the diaphragm compliance, in addition to acoustic nonlinearities introduced by the Helmholtz resonator orifice at high amplitudes. These nonlinearities will act to limit the maxi mum voltage in a real world application. Operating at these same optimal frequencies, a load resistor was placed across the output terminals and the power was determined through that resistor. Repeating this over a range of load resistances yields the plot s hown in Figure 2-31. Note that while the input power remains constant, the output power has a p eak at a particular load resistance. This corresponds to the optimal resi stance as given by Eq. (2.12) The input and output power using an optimal load are also shown on this graph and are indicated by an x for both the diaphragm with the Helmholtz resonator and without. 102 103 104 10-10 10-8 10-6 10-4 Load Resistance [Ohms] [W] in Membrane in Membrane/HR out Membrane out Membrane/HR Figure 2-31: Input and output power as a function of the load resistance. Shown for the piezoelectric diaphragm by itself and packaged with a Helmholtz resonator. (Device = 8 from Table 2-5, H R f =1966 Hz mem f =13.11 kHz, and 114 PdB )

PAGE 93

79 Using the optimal values for the load resi stance in each case, the input power and output power was determined as a function of the applied acoustic i nput pressure, and is shown in Figure 2-32. Once again, it is impor tant to note that nonlinear effects are not included in the model used to calculate the output power. 115 120 125 130 135 140 145 150 155 160 10-10 10-8 10-6 10-4 10-2 100 102 Acoustic Input Pressure [dB]Power [W] in Membrane in Membrane/HR out Membrane out Membrane/HR Figure 2-32: Input and output power as a function of the a pplied acoustic input pressure, while using an optimal load resist or. (Device = 8 from Table 2-5, H R f =1966 Hz mem f =13.11 kHz ,486 loadMemR ,/3242 loadMemHRR ) Device Configurations The devices were designed based on the theo retical plots, along with estimates of the output power based on the lumped element m odel. Eight configurations were chosen based on their resonant frequency and maxi mum power before the onset of nonlinear behavior. The criteria for choosing the desi gns were to keep the resonant frequencies

PAGE 94

80 inside of the testable range, below 6.4 kHz, while maximizing the power. Four outer radii, R2, were chosen, consisting of 0.9 1.2, 1.5, and 1.8 mm while the inner radii were all chosen to be 0.95 times their outer counter part. The second four configurations had the same dimensions as the first four, with the addition of a central pillar of silicon designed to act as a rigid mass. It should be noted that th e addition of the central mass invalidates the structural model that was us ed, however, the model is still expected to provide some guidance over the behavior of th ese particular devices as well. The central mass was added in order to reduce the resona nt frequency of the diaphragms. It was fixed at 0.2 times the outer radius of each diaphragm. This value was chosen such that the central mass was located primarily over a region of the diaphragm that experiences little stress. By restricting the mass to this region, only a minimal increase in the diaphragm stiffness was expected, while the mass of the diaphragm could be greatly increased, thereby lowering the resonant fre quency. The addition of this central pillar was predicted to lower the resonant frequency of each device by an average of 57.76 %. Additionally, the diaphragm thickness was chosen to be 3 um. Only one thickness for the diaphragm was chosen as all the devices were going to be made on the same wafer. The chosen device configurations are summ arized in Table 2-5. Overall, the geometries were chosen as to create a proof of concept set of devices. Manual parameter adjustment was performed in order to impr ove performance, however, strong constraints were placed on the geometry in order to ensure fabrication ability and improve the chances of first-run success. Also listed in the table are the predicted resonant frequencies and maximum power before the ons et of nonlinearities. Note that this does not define an upper limit for the output power but merely provides a figure of merit upon

PAGE 95

81 which to compare devices. Also, for the pur poses of predicting the behavior of these specific devices and comparing them to experimental data, the lumped element parameters were computed using the equations from Chapter 2 and the Matlab code given in Appendix A. The calculated parameters are given in Table 2-6. Table 2-5: Chosen device conf igurations for fabrication. Device sitm ptm 2Rm 1Rm massRm res f kHz outPownW 1 3 0.5 900 830 0 65.68 15.56 2 3 0.5 900 830 180 28.08 36.12 3 3 0.5 1200 1115 0 48.14 1.73 4 3 0.5 1200 1115 240 20.37 3.99 5 3 0.5 1500 1400 0 38.01 0.302 6 3 0.5 1500 1400 300 15.98 0.667 7 3 0.5 1800 1685 0 31.41 0.070 8 3 0.5 1800 1685 360 13.15 0.162 Table 2-6: Lumped element paramete rs used for theoretical models. k Ad aDC aD M aDmassM ebC Device Pa V V Pa 3m V 3m Pa 4kg m 4kg m nF1 -1.81 -5.80E-07 0.001 1. 01E-14 5.58E-15 5663.3 0 17.4 2 -1.81 -5.80E-07 0.001 1.01E -14 5.58E-15 5663.3 19649 17.4 3 -9.29 -7.16E-06 0.008 1. 93E-13 2.07E-14 2602.5 0 26.9 4 -9.29 -7.16E-06 0.008 1.93E -13 2.07E-14 2602.5 11052 26.9 5 -12.93 -1.94E-05 0.016 7. 43E-13 5.74E-14 1521.8 0 38.2 6 -12.93 -1.94E-05 0.016 7.43E -13 5.74E-14 1521.8 7073.6 38.2 7 -16.92 -4.71E-05 0.028 2. 42E-12 1.43E-13 1009.8 0 51.3 8 -16.92 -4.71E-05 0.028 2.42E -12 1.43E-13 1009.8 4912.2 51.3

PAGE 96

82 CHAPTER 3 FABRICATION AND PACKAGING Overview This chapter focuses on the device fabrication and process flow, as well as the device packaging design and implementation. First, a detailed discus sion of the process flow is presented, including fabrication me thods, equipment, and recipes. This is followed by an overview of the packaging schemes that were employed. The packages were designed to flush mount the device in or der to expose the diaphr agm directly to the acoustic input, thereby enabling direct measurement of the diaphragm parameters. Two variations in package design were employed, namely a quarter-wave resonator package and a sealed cavity package. A graphical overview of this chapter is provided by the roadmap of Figure 3-1. Process Flow The devices were batch fabricated on 4 silicon-on-insulator (SOI) wafers. The process sequence is given in detail in Appendix B. A ll of the processing steps up through the deposition of the top electr ode were performed at Sandia National Laboratories. The remainder of the steps were formed at the University of Florida, with the exception of the PZT etch step which was performed at the Army Research Laboratory.

PAGE 97

83 Overview Process Flow Packaging Packaging Scheme Chip to Package Mounting Sealed Cavity Package Quarter-Wave Resonator Package Poling Figure 3-1: Roadma p for Chapter 3. A diaphragm thickness of 3 um was desired in order to achieve a testable resonant frequency and therefore required a top silicon thickness in that range, as the top silicon layer of the SOI wafer ultimately forms the bulk of the diaphragm. Unfortunately, the only available SOI wafers contai ned a top silicon layer of 12 um thickness. In order to reduce this thickness, a timed KOH (Potassi um Hydroxide) etch [137] was performed, however a residue was left behind on most of the wafers, that was difficult to remove. The composition of the residue was not iden tifiable, however a 1 minute dip in 10% Nitric acid, followed by a 2:1 Piranha etch fo r 5 minutes had no noticeable effect.

PAGE 98

84 Figure 3-2: Condensed process sequence in cr oss-section. a) Depos it 100 nm of Ti on SOI Wafer and oxidize to form TiO 2 b) Deposit Ti/Pt (40/180 nm) and liftoff with ElectrodeBot mask. c) Spin coat PZT 52/48 solution & pyrolize (6 layers for 400 nm total). d) De posit Pt (180 nm) and liftoff with ElectrodeTop mask. e) Wet Etch PZT in 3:1:1 of (NH 4 )HF 2 /HCl/DI water. f) Spin & pattern thick photoresist on bottom with Cavity mask. g) DRIE (deep reactive ion etch) to buried oxide layer. h) Ash resist and BOE (buffered oxide etch) backside to remove to buried oxide layer. Two of the original KOH-etched wafers we re replaced with two new wafers that were etched in an HNA bath to avoid the re sidue issue [138]. The advantage of the HNA etch was a smooth, mirror finished surf ace to the wafer with no visible residue. The disadvantage was a less consistent et ch rate than the KOH procedure. For comparison the HNA etch had an average etch rate of 2.6 minm with a standard deviation of 0.33 minm while the KOH etch had an etch rate of 1.69 minm with a standard deviation of 0.03 minm a) b) c) d) e) f) g) h)

PAGE 99

85 Following the wafer thinning, the process fl ow proceeded as given by Figure 3-2. The first step was to deposit a 100 nm thick layer of Ti, and oxi dize in a tube furnace to create a TiO2 layer that serves as a diffusion barri er for the PZT [139-142] as shown in Figure 3-2a. A liftoff process was then used to pattern the bottom electrodes with 40 nm of Ti followed by 180 nm of Pt as shown in Figure 3-2b. The Ti layer served as an adhesion layer in this step. The next step was the deposition of a previous ly mixed 52/48 sol-gel solution of PZT [139, 143, 144] as shown in Figure 3-2c. The solution was spin-cast at 2500 RPM for 30 sec, then pyrolized at 350C in air for 2 min on a hot plate. Repeating the spin and pyrolize step 6 times yielded a total PZT thickness of approximately 400 nm The wafers were then furnace annealed at 650C for 30 min to achieve a perovskite phase of PZT that has the most attractive piezoelectric properties. Wafer bow measurements were performed after deposition and oxidation of the TiO2 as well as after annealing of the PZT. By recording the wafer curvature in terms of radius or bow, the stress in the layers can be determined [145-147]. Table 3-1 shows the results and calculations from these measurements for all of the fabricated wafers. Stress levels were first calculated for the combination of TiO2 and backside SiO2 that was formed during the titanium oxidation step. Then the stress contribution from the TiO2 was calculated. Finally, using the wafer curvature measurements after deposition and annealing of the PZT, the stress in the PZT layer was calculated. In order the calculate the stress from th e wafer curvature, a relationship known as Stoneys equation [145-147] was used and is given by

PAGE 100

86 3 2611ss f s sf fEh h Rh h (3.1) where 1 R is the curvature, s fhh is the substrate and film thickness respectively, s E is Youngs modulus of the substrate and s is Poissons ratio of the substrate. Table 3-1: Wafer bow m easurements and the resulting calculated stress. Initial After TiO2 Dep. + Ox. TiO2 + SiO2 TiO2 After PZT PZT Wafer # Radius m Bow m Radius m Bow m Stress M Pa Stress M Pa Stress M Pa Radius m Bow m Stress M Pa 1 -43.8 18.51 -137.8 6.15 1367.1 844.7 1217.0 -149.9 5.65 12.8 1A -40.1 20.5 -103.2 8.4 1335.3 1239.7 1809.6 -1144 1.52 176.3 2 -38.9 20.45 -99.0 8.03 1371.1 848.8 1223.1 -1338 0.72 187.3 3 -60.8 13.05 -291.4 3.12 1140.6 678.6 967.9 n/a n/a n/a 4 -44.2 18.24 -121.2 6.93 1259.8 772.9 1109.4 279.3 -2.82 249.9 5A -63.7 12.84 -717.3 1.48 1254.5 776.3 1114.5 156.3 -4.74 159.4 6 -43.0 18.86 -109.3 7.6 1235.4 769.5 1104.2 809.9 -0.77 214.5 7 -44.3 18.3 -106.1 7.84 1154.2 714.8 1022.2 662.9 -1.27 233.5 Before the above equation can be us ed to find the stress in the TiO2 layer, the thickness of the SiO2 layer that was formed during the oxidation step must be determined. For silicon oxidation, the relationship be tween oxide thickness and oxidation time is approximated by [148] 2 oxoxtAtBt (3.2) where A and B are coefficients which depe nd on material properties and operating conditions, oxt is the total oxide thickness, t is the oxidation time, and is a time shift due to an initial oxide thickness For short oxidation times, where 24 tAB, this equation reduces to oxB tt A (3.3)

PAGE 101

87 For <100> silicon oxidized in dry O2 at 650 C, 0.0001 B Amhr, and 20.0001 B mhr Under these conditions, the a pproximation of short oxidation times holds, as the total oxidation time wa s 30 min, yielding an oxide thickness of approximately 50 nm. This thickness was then utilized in Stoneys equation to help determine the contribution due to each of the layers. In addition to wafer bow measurements, the density of the PZT was calculated by measuring the mass of the wafers both before and after deposition. The results of these measurements are shown in Table 3-2. Table 3-2: Measured wafer mass and calculated density of PZT. Mass [g] PZT Density Wafer Pre-PZTPost-PZT 3kgm 1A 9.752 9.7734 10217 2 9.48169.5129 9963 4 9.50499.5369 10185 5A 9.74459.7616 8164 6 9.556 9.5771 10074 7 9.55659.5764 9501 Avg. 9684 Following the PZT deposition, the top electr odes were deposited and patterned in a liftoff process similar to the bottom electr odes as shown in Figure 3-2d. The top electrodes also served as a hard etch mask for etching the PZT. The PZT was etched using a 3:1:1 solution of ammonium biflouride ((NH 4 )HF 2 ), hydrochloric acid (HCl), and deionized water as shown in Figure 3-2e. The etchant leaves a residue that then must be removed with a dilute nitric acid and hydrogen peroxide solution. The next step was to deposit a thick phot oresist on the backside, and pattern the backside release etch as shown in Figure 3-2f. Then the sili con wafer is backside etched using a deep reactive ion etch (DRIE) proces s, also known as a Bosch etch [148-152] as

PAGE 102

88 shown in Figure 3-2g. This is a plasma-bas ed etch process that alternates between a plasma etch, and a polymer deposition. The polymer is easily removed from the bottom of the etch holes during the ensuing etch cycl e, but serves to significantly slow down any lateral etching, leading to highaspect ratio etching. The ratio of etch time to passivation time during each cycle, along with a number of other process parame ters, greatly affects the resulting etch quality. Shown in Figure 3-3 is a scanning electron micrograph (SEM) image of the sidewall of one of the test devi ces after an early stage, through-wafer, DRIE test, prior to working out the optimal parameters In this figure, signi ficant erosion of the sidewall can be seen, in the form of both pitting and undercutting. For comparison, the upper right region of the image shows what a smooth silicon sidewall looks like. The damage shown here is a result of an etch -to-passivation ratio that is too high. The passivation layer is then insufficient to protec t the sidewalls during the entire etch cycle. Figure 3-3: Early DRIE results s howing significant sidewall damage.

PAGE 103

89 A different problem occurs when the etch-topassivation ratio is too low. In this instance, some portions of the passivation la yer are not removed from the bottom of the etch holes during the etch cycle, leading to a micromasking effect and the ultimate formation of what is known as silicon grass or black silicon [153]. This was also encountered during the early tria l stages of our DRIE attempts. Shown in Figure 3-4 is an SEM image showing the formation of black silicon at the base of a trench. Figure 3-4: SEM image showi ng black silicon at the base of a DRIE-etched trench. Finally, after adjustment of the various process parameters a successful DRIE was performed through the thickness of the silicon wafer. This is shown in Figure 3-5. Notice, a nearly vertical sidewall was achieved with little cratering and erosion, leaving a smooth surface with the exception of the large debris particle.

PAGE 104

90 Figure 3-5: SEM image of a successful DRIE etch through the thickness of a wafer. Once the DRIE recipe was worked out, the DRIE process was used here to etch through the bulk of the silicon wafer, stopping on the burie d oxide layer, leaving a thin diaphragm as well as creating channels that served to later separate the wafer. Additionally, the central pillar that serves as the point mass was created in this process by masking the central portion of the diaphr agm backside. Following the DRIE, the photoresist was stripped from the backside. A composite top view schematic and crosssection of the device are shown in Figure 3-6. The devices were arranged on the wafer in blocks as shown in Figure 3-7. Each block c onsists of the eight device configurations. The blocks were then arranged on the wafer as shown in Figure 3-8. The blocks were arranged across the wafer such that the wafer would maintain structural stability after the DRIE release of the die with in the blocks. A photograph of a completed energy harvester chip taken under a microscope with a 1.25x lens is shown in Figure 3-6c.

PAGE 105

91 c) Figure 3-6: Illustration of a single diaphragm device. a)Composite top view schematic of single diaphragm device showing ring shaped electrodes, with bottom electrode extending to lower left and top electrode extending to lower right. Also shown is the die separation cha nnel around the perimeter. b) Completed cross-section showing the Pt/PZT/Pt sa ndwich. c) Optical photograph of a completed device, taken under a micros cope with a 1.25x magnification. Figure 3-7: Block layout illu strating numbering scheme fo r devices, per Table 2-5. a) b) Piezoelectric Ring Diaphragm Bottom Electrode Top Electrode

PAGE 106

92 Figure 3-8: Wafer layout illustrati ng numbering scheme for blocks. Packaging Following the processing steps listed above each die was separated from the wafer by scribing across the small tab holding each die and cleaving at the scribe line. Then each die was mounted in a Lucite package, described next, and bonded to a pair of electrical leads. The devices to be tested inside of an acoustic plane wave tube were packaged accordingly so that a tight seal was created with the plane wave tube while allowing external electri cal access to the devices. Packaging Scheme Quarter-wave resonator package In order to provide a convenient pack age for measurement of the diaphragm parameters, an acrylic plug for the acoustic waveguide was designed and constructed.

PAGE 107

93 The plug allows for flush mounting of the de vice on the topside, and acoustic venting on the backside to prevent a cavity stiffening e ffect. The cavity stiffening effect describes an overall lowering of sensitivity due to the low compliance (i.e. high stiffness) of a small cavity (Eq. (2.21)). This however introduced a quarter-wave resonance effect that will be demonstrated in the experimental results. Additionally, two, 18 gauge, copper wires are embedded in the plug to allow for electri cal connection to th e electrodes of the piezoelectric ring. A schematic of the package is shown in Figure 3-9. Figure 3-9: Quarter-wave resonator package consisting of acrylic plug, copper leads and vent channel. The quarter-wave resonance effect arises because of the impedance mismatch that occurs at the interface between the backside vent channel and the environment on the exterior of the package. An acoustic wave th at is traveling down the channel is reflected upon encountering the open end of the channel. This sets up a standing wave within the

PAGE 108

94 channel. Both the pressure, P, and velocity, U, exhibit this spatial pattern, and the ratio of PU, which is the impedance, Z shows a spatial dependency as well. Mathematically, the impedance at a distance x along the length of the channel can be written as ()tano Z xjZkx (3.4) where o Z is the specific acoustic impedance of the medium. When the distance, x, is equal to of the wavelength, this equation reduces to 2 ()tantan0 444oofc ZjZkjZ cf (3.5) Thus, the impedance goes to zero at fr equency-position combinations where 4 x This zero impedance allows for large velocity responses for small pressure inputs, i. e. a resonance. Similarly, ther e are frequency-position combinations where the impedance becomes infinite. In practice, various losses serve to limit the impedance from becoming exactly zero or infinity. Sealed cavity package This package was designed so that the parameters of the piezoelectric composite diaphragm could be measured independently of effects due to the quarter-wave resonator. The design is identical to the quarter-wave res onator except the vent channel is sealed. A cross-sectional schematic of the sealed cav ity package is given in Figure 3-10 and a corresponding optical photograph of the devi ce and package is shown in Figure 3-11. Chip to Package Mounting The chip was flush mounted within the provided recess in the package, and epoxy bonded to the package to provide an airtight seal. The el ectrical connections to the

PAGE 109

95 package leads were accomplished via short le ngths of flattened bare wire that were bonded using two-part, silver epoxy (Epotek H20E) on the chip bond pads as well as to the copper pads on the package. The silver epoxy was then cured at 90 C for one and a half hours. A relatively low temperature, long curing time recipe was chosen so as to remain below the glass transition temperat ure of the Lucite, which occurs at 100 C. Figure 3-10: Sealed cavity package. Figure 3-11: Optical photograph of a packaged energy harvester.

PAGE 110

96 The silver epoxy method, as outlined above, was used because of problems encountered when trying to wire bond from the package to the die. The problems were believed to have arisen from overhanging top electrode metal on the individual die. The liftoff step that produced the top electrode was found to have left stringers around the edges of the electrode. These stringers we re the result of performing a liftoff using positive resist, as the positive sidewall slope of the resist does not allow for a clean break in the metal deposition, thereby leading to tear ing during the liftoff portion of this step. The overhanging metal was found to cause s hort-circuits on all of the devices of several wafers. On two of the wafers, no short-circuits were evident initially; however, after wire bonding, all devices would then exhibit a short-circ uit. It was concluded that the vibrations and impact of the wire bonding tool were causing the overhanging top electrode to fall into contact w ith the bottom electrode directly beneath it. This problem was avoided by utilizing a more gentle appr oach to bonding, namely silver epoxy. The entire issue, however, could be avoided in th e future by using a negative resist and an image reversal mask to create a negative slopi ng photoresist sidewall fo r this liftoff step. Poling Following the completion of the packaging steps, the devices were poled so as to align the domains along the poling axis. It should be noted that the devices were found to exhibit a small amount of pi ezoelectric behavior prior to poli ng. This is believed to be due to a partially pre-aligned structure resulting from the deposition process. Poling was accomplished by applying an electric field across the device for a specified amount of time. The devices were poled at room temperature at 20 Vm that corresponds to an appl ied voltage of 5.34 V for the given device thickness. This poling

PAGE 111

97 field was held for 10 min A typical maximum polarization of 16.9 2Ccm was achieved with a remanent polarization of 6.26 2Ccm. More details of this measurement are provided in Chapter 5. Typically, poling is accomplished at higher temperatures for improved piezoelectric properties [8, 145], although Bernst ein et al [28] reported acceptable poling results at room temperature. While it wa s desirable to try a hi gher temperature poling procedure, temperature constraints caused by the glass transition temperature of the acrylic package prohibited the use of temperatures above 100 C. Prior the successful poling of the devices, several othe r variations of the poling method were tested. For instance, initially, 10 Vm was utilized as a poling field, however piezoelectric properties were found to be relatively unchanged from the unpoled state. Additionally, higher elect ric fields were tested, above 20 Vm however this consistently led to the development of a s hort-circuit in the device. The short was believed to be a result of the many stringer s that overhung from the top electrode layer down to the bottom electrode. At sufficientl y high voltages, the force of attraction between these stringers and the bottom elect rode is hypothesized to pull them into contact.

PAGE 112

98 CHAPTER 4 EXPERIMENTAL SETUP This chapter discusses the various setups for the experimental work. A series of measurements were performed to experimenta lly determine the various parameters of the piezoelectric composite diaphragm. This allo ws for further analysis and verification of the models. The measurements were performed using four different experimental setups. First, the electrical output impedance of each device was measured using an impedance meter. Then an electrically actuated re sponse was measured using a scanning laser vibrometer (LV), to determine a voltage driven frequency res ponse and corresponding mode shapes. Following this, an acoustically actuated response was measured using an acoustic plane wave tube. This was used to determine a pressure driven frequency response, input and output power, output voltage and efficiency. Finally, the static mode shapes were measured using an optical prof ilometer to determine the initial deflection due to residual stresses. A graphical overvie w of this chapter is provided by the roadmap of Figure 4-1. Parameters and How They are Obtained A structured methodology was develope d for obtaining all of the desired measurement parameters for the composite plat e. Using a predefined set of experiments and a few relations between various paramete rs, a method was derived for obtaining each of the parameters.

PAGE 113

99 Figure 4-1: Roadma p for Chapter 4. The series electrical output capacitance, s C and resistance, s R can be experimentally obtained through use of an impe dance meter. This measurement needs to be performed after poling the de vice, as the poling procedure can affect the capacitance. The measured value can then be compared to theory, given by sCAd if totally free, (Eq. (2.51)) and 21sCAdk if totally blocked, (Eq. (2.61)). Now in the case of the energy harvesters, 1 k so that the free and blocked cases are approximately equal. For the series resistance, the theoretical value can be obtained from a predicted loss tangent for the material. The resonant frequency of the devices can be obtained through use of an acoustic plane-wave tube (PWT) that is used to excite the diaphragm acoustically, while the output voltage from the device is measured The measured frequency can then be compared to a theoretical value that is based on the lumped element model for aDC and aD M

PAGE 114

100 There are four readily definable sensitivitie s of interest for the energy harvester. Two of these sensitivities are defined at reso nance, while the other two represent static sensitivities. EA resSens is the electrical-to-acoustical se nsitivity at resonance and can be experimentally determined by applying a volta ge signal to the device at the resonant frequency while the diaphragm deflection is measured via a laser Doppler velocimeter (LDV). AE resSens is the acoustical-to-electrical sens itivity at resonance and can be found experimentally by acoustically exciting the diaphragm at resonance via the PWT while measuring the voltage produced by the device. Both of these resonant sensitivities can then be compared to theory which is comput ed directly from the lumped element models forrad R aDC ,aD M s R and The static electrical to acoustical sensitivity, EA lowSens is approximated using the LDV at low frequencies. The LDV requires a finite velocity, and so a low frequency approximation is necessary for estimation of th e static value. Similarly, the static acoustical to electrical sensitivity, AE lowSens is measured in the PWT using an acoustic input signal at low frequencie s. The same limitation to th e static case occurs in this measurement as well. Both of these values can then be compared to theoretical values computed directly from the lumped element model using the equations for aDC ,ef R ,efC and The initial deflection of the diaphragm th at results from residual stress in the composite can be measured using a Wyko opti cal profilometer without any excitation to the device. Theoretical values are obtained by numerical calculations accounting for the residual stress.

PAGE 115

101 The coupling factor, k, is measured in two different ways. First it may be obtained from the difference between the open and short ci rcuit resonant frequencies, or it can be found via k Theoretical values are obtained from and or from from Ad aDC and efC Finally, the electroacoustic, and acoustoelectric, transduction factors are extracted from a combination of EA lowSens and aDC and a combination of AE lowSens and efC respectively. The theoretical values can be found directly from the numerical model. Experimental Setup Details Ferroelectric Measurements The ferroelectric properties of the devi ces were measured by attaching the two electrodes to a Precision LC, Precision Ma terials Analyzer manufactured by Radiant Technologies. This ferroelectric tester appl ies a bipolar triangular waveform, while it measures the resulting charge, from which it can calculate the polarization of the material as a function of applied voltage. The bipol ar waveform excitation to the ferroelectric material leads to the familiar hysteresis loop. For these measurements, the voltage was swept first between -0.5 V and 0.5 V Subsequent loops swept to higher voltages in steps of 0.25 V up to a maximum of 5 V. For each loop, 101 points were measured and the total loop time was set to 10 ms In addition, for each loop, a pre-loop was performed, where the same sweep was performed without measurements taken. Then the loop was repeated with the actual measurements taken. This process is de signed to preset the conditions for the measurement. A delay of 1 ms was used between the preloop and the measurement loop.

PAGE 116

102 Electrically Actuated Response Laser Scanning Vibrometer In order to determine the voltage actua ted mode shapes and frequency response, laser scanning vibrometry was performed. The packaged energy harvesters were mounted under an Olympus BX60 Microscope, [154] with a 5x objective lens. The composite diaphragms were excited using either an HP 33120A Arbitrary Waveform Generator [155] or a Stanford Research Systems SRS-785 Dynamic Signal Analyzer. Displacement data were obtained over the su rface of the composite diaphragms using a Polytec OFV 3001S laser scanning vibromet er, with a Polytec OFV-074 Microscope Adapter. [156] The laser vibrometer was e quipped with a velocity decoder, from which the displacement was calculated. To determine the frequency response, a wh ite noise input was applied. The center deflection was then measured as a function of frequency, while the reference voltage was simultaneously recorded. From this meas urement, the resonant frequencies were determined. A Hanning window was applied to all incoming data and 6400 FFT lines, with 100 averages were taken. Depending on the particular device, the measurement bandwidth was adjusted as was the laser sens itivity and range in order to optimize the signal level and resulting data quality. A slow tracking filter was applied to all incoming laser measurements. Additionally, due to an inherent time delay in the laser vibrometry channel, there is a fixed linear phase trend in any measured frequency response data. All resulting plots were detrended by taking this time delay into account. The amount of the delay varies, depending on the gain setting, how ever for a given setting it is fixed and defined by the manufacturer. The input was then changed to a sine wave at the primary resonant frequency and the laser was scanned over the surface to m easure the primary mode shape. Higher order

PAGE 117

103 modes were also measured, although primarily to identify the nature of the higher resonances, and will not be presented in much detail. Acoustic Characterization Plane Wave Tube The packaged energy harvesters were mounted one at a time, to the end of the 1 x 1 cross section plane wave tube (PWT). Figure 4-2 shows this experimental setup for obtaining acoustical input impedance and power to the energy harvester. A Bruel and Kjr (B&K) microphone type 4138 [157] was mounted on the wall of the PWT near the end of the tube to record the incident sound pressure level (SPL). Additionally, two B&K microphones were mounted in a rotating plug at 3.19 cm and 5.26 cm from the end of the tube. These two microphones were used to record the acoustic field in the tube, extract the standing wave component, and de termine the correspondi ng impedance of the energy harvester. The data was recorded by a B&K PULSE Acoustic Characterization System that also supplied the periodic random source signal through a Techron 7540 Power Supply Amplifier [158] to the BMS 50.8 mm coaxial compression driver (Model 4590P) mounted at the far end of the PWT. The tests were then repeated with a 1 k load shunted across the electrode s of the piezoelectric compos ite diaphragm. For all of the tests, the equipment was setup to measure 6400 FFT lines from 0 kHz to 12.8 kHz with uniform windowing, no overlap, and 500 averages. To determine the normal incidence acoustic impedance, frequency response measurements (later defined as 12H ) were taken using the two microphones in the rotating plug. [159, 160] The plug was then rotated and the measurement repeated. The two measurements were then averaged to re move any differences due to the individual

PAGE 118

104 microphone calibrations. This averaged freque ncy response was then used with the twomicrophone method (TMM) to determine the acoustic impedance [159, 161-163]. Figure 4-2: Experimental setup for impedance and power measurements. The TMM is a method for estimating the re flection coefficient of an object or material mounted at the end of an acoustic wave guide. The TMM requires less than a half-wavelength spacing between the two microphone locations [159-161, 163, 164]. Although the waveguide used is physically capable of suppor ting plane waves up to 6.7 kHz testing was limited to 6.4 kHz because of this microphone spacing constraint. An alternative, single-microphone method [157] exists but was not used due to greater errors associated with the technique. Additionally the uncertainties were estimated for the acoustic impedance measurements using a t echnique presented by Schultz et al [165]. Under the assumption of planar waves in the waveguide, the sound field inside the waveguide can be described by ReRjtkdjtkd ippee (3.6)

PAGE 119

105 where the j tkde term represents the right-propagating waves, the j tkde term represents the left-propagating waves, and d is the axial distance from the specimen surface. From Eq. (3.6), an expression for the reflection co efficient is found by taking the ratio of two pressure measurements at different axial locations, 12 12 21 12jkdjkd jkdjkdHee R eHe (3.7) where 1d and 2d are the distances from the specimen test surface to the measurement locations, is the radian frequency, k is the wavenumber, R is the complex reflection coefficient, and 12H is an estimate of the frequenc y response function between the two microphones. The frequency response function is defined as [166] 12 12 1212 11jG HHe G (3.8) where 12G is the cross-spectral density function and 11G is the autospectral density function. After calculation of the reflection coefficient from the measured data, the normalized acoustic impedance is then found to be a function of the reflection coefficient and is given by 1 01 1 ZR j ZR (3.9) where 0 Z and 1 Z are the specific acoustic impedance of the acoustic medium and the test specimen, respectively and and are the normalized resistance and reactance, respectively. Using the same test setup as desc ribed above, the input acoustic power, input can also be found. A schematic showing the incide nt and input power is given in Figure 4-3.

PAGE 120

106 Figure 4-3: Schematic of incide nt, reflected and input power. There are two main methods for obtaining the input acoustic power. First it can be calculated from the power reflection coefficient, 2rR and the incident acoustic power, 2tube p Z as 2 211input tubep rR Z (3.10) where tube Z is the acoustic impedance of the plane-wave tube, and p is the acoustic pressure of the right-propagating (incident) waves. The second method for obtaining the input acoustic power is to u tilize the input impedance, in Z that was also calculated from the two-microphone measurements via R Using this method yields 2tr input inP Z (3.11) where trPPP is the transmitted acoustic pressu re. Under ideal conditions, both methods yield equivalent answers; however, when R approaches 1, the value for R is known with significantly greater certainty than the value for in Z In addition to measuring the input impe dance and power, the frequency response function of the devices can be measured. Th e resonant frequency of the diaphragm can

PAGE 121

107 be found from the frequency response functi on, as can the maximum output voltage and power. As the input power is also known, th e efficiency can be calculated. Additionally, a load can be placed across the electrical term inals of the energy harvester and the effects monitored. The resistance of the load can be varied to experimentally determine the optimal loads for maximum power and maximum efficiency. Electrical Impedance Characterization The electrical impedance was measured using a HP 4294A Impedance Analyzer [155] after first poling the de vices. Measurements of the parallel capacitance and resistance were obtained at 1 kHz for each of the packaged devices. Additionally, the impedance was measured over a sufficient sp ectral bandwidth to capture the first resonance that varied from device to device. Initial Deflection Measurements Wyko Optica l Profilometer Due to residual stresses in the composite diaphragms and compliant boundary conditions, the devices exhibited an initial static deflection. This deflection was measured by mounting the devices under a Wyko Optical Profilometer [131] with a 5x optical lens and a 0.5x reducer, yieldi ng a total magnification of 2.5x.

PAGE 122

108 CHAPTER 5 EXPERIMENTAL RESULTS AND DISCUSSION This chapter discusses the experimental resu lts that were obtained from the various measurements. First, the electrical impedan ce results are shown and explained. This is followed by the electrically actuated freque ncy response measurements. Following this, the acoustically actuated frequency response and impedance results are discussed for both package configurations. Finally, the initial deflection measurements are presented for several devices. A graphical overview of th is chapter is provided by the roadmap of Figure 5-1. Figure 5-1: Roadma p for Chapter 5. Ferroelectric Measurements The ferroelectric properties of several de vices were measured using the Radiant Technologies, Precision LC Analyzer. Show n in Figure 5-2, is a graph of the polarization versus applied voltage over severa l loops of steadily in creasing voltage. For

PAGE 123

109 this measurement, the voltage maxima were swept from 1 V to 5 V in 0.25 V increments, corresponding to a maximum applied electric field of 187.95 kVcm.. -200 -150 -100 -50 0 50 100 150 200 -20 -15 -10 -5 0 5 10 15 20 Electric Field [kV/cm]Polarization [ C/cm2] Pr=6.26 Pm=16.9 Figure 5-2: Hysteresis loop for Device 1A-5-4 in a sealed cavity pack age. The device has a radius of 1200 m and a central mass. The resulting polarization reached a maximum of 16.9 2Ccm with a remanent polarization of 6.26 2Ccm for the 5 V sweep. A typical hysteresis curve shows flattened regions near the upper and lower ends of the sweep, where the response becomes saturated. It was not possible to ex tend the measured hysteresis curve into those regions as the maximum voltage that c ould place across the piezoelectric was approximately 5 V Beyond this voltage, unexpected eff ects occurred in the material that produced a short circuit between the two electrodes. The nature of this short circuit was

PAGE 124

110 not resolutely identified, but it is believed to be related to the ove rhanging top electrode resulting from problems in the fabrication. Electrical Measurements Electrical Output Impedance Electrical measurements were taken using a HP 4294A Impedance Analyzer. Data was generally recorded from 1 kHz to 50 kHz however, the results shown below were extracted from the 1 kHz values. The parallel capacitan ce and parallel resistance were measured in this manner, and the correspondi ng relative dielectric constants and loss tangents were extracted, by taking into account the respective geometries of each device. Shown in Figure 5-3 is the measured parall el capacitance as a function of the outer radius, 2 R All eight device configurations, as given in Table 2-5, were measured, resulting in two data points for each outer radius. The capacitance was found to vary between 8.65 nF and 27.4 nF and increases with increasing radius as expected. The individual corresponding relative diel ectric constants were calculated using, 22 21pp r otC RR (5.1) where pt is the piezoelectric thickness, 2 R and 1 R are the outer and inner radii respectively, and o is the permittivity in a vacuum. The relative dielectric constant was found to have an average value of 483.47 with a standard deviation of 23.48. Others have reported dielectric constants for PZT from 125 [16] to 1400 [32], although the average value in the literature is approximately 839.

PAGE 125

111 Figure 5-3: Measured parallel output capacitance, pC vs. outer radius, 2 R overlaid with a theoretical capacitance curve based on the average extracted dielectric constant. The error bars indicate th e 95% confidence intervals based on the variance across multiple devices. In a manner similar to the capacitance measurements, the parallel output resistance was measured at 1 kHz for each of the devices and is shown in Figure 5-4. As can be seen from the figure, it was found to decrease with increasing radius. This effect was investigated further by looking at the conductance per unit ar ea of each of the devices, given by 2222 21211pG R RRRR (5.2) where G is the conductance. It was found th at the conductance per unit area was relatively constant with respect to radius, with the exception of one outlier point measured for the device with an 1800 m outer radius. The average value was found to

PAGE 126

112 be 2.40 2Siemensm, with a standard deviation of 0.46 2Siemensm, when excluding the outlier point. Figure 5-4: Parallel output resistance, p R vs. outer radius, 2 R overlaid with a theoretical resistance curve based on the average measured conductivity. The error bars indicate the 95% confidence intervals of the variance across multiple devices. The loss tangent was then calculated base d on the measured electrical reactance, p X and resistance, p R and is given by, tanp p X R (5.3) The average loss tangent, as measured at 1 kHz and once again ignoring the outlier, was found to be approximately 0.024, which matche s well with previously published values ranging from 0.02 to 0.05 for thin film PZT [12, 26, 34, 35, 37]. A summary of all the electrical impedance measurements is provide d by Table 5-1. The average values are given at the bottom of the table and were co mputed without the outlier points that are marked in the dotted box.

PAGE 127

113 Table 5-1: Summary of electrical impedance measurements. Device # 1Rm 2Rm massRm pCnF pRk r tan 2Gm 1A-5-1 830 900 0 8.72 665.89 507.91 0.027 2.867 1A-6-1 830 900 0 8.65 837.09 503.77 0.022 2.281 1A-6-2 830 900 180 8.45 1120.20 492.42 0.017 1.704 1A-10-2 830 900 180 7.76 900.73 451.84 0.023 2.120 1A-11-3 1115 1200 0 12.68 652.73 478.21 0.019 1.895 1A-5-4 1115 1200 240 12.14 730.94 458.04 0.018 1.692 1A-10-4 1115 1200 240 12.59 414.63 474.99 0.030 2.983 1A-6-5 1400 1500 0 18.33 348.04 486.75 0.025 2.501 1A-6-6 1400 1500 300 17.64 306.31 468.60 0.029 2.842 1A-10-6 1400 1500 300 18.32 367.91 486.67 0.024 2.366 1A-3-7 1685 1800 0 26.70 52.60 527.59 0.113 12.314 1A-6-7 1685 1800 0 22.82 237.78 450.94 0.029 2.724 1A-3-8 1685 1800 360 27.40 233.55 541.54 0.025 2.773 Average n/a n/a n/a n/a n/a 483.47 0.024 2.40 The devices are numbered according to the device type and location on the wafer, in the form of WW-BB-DD, where WW is th e wafer number, BB is the block number as defined in Figure 3-8 and DD is the device number as defined in Figure 3-7. Note that all the devices that were tested came from the sa me wafer, referred to here as wafer 1A. Shown in Table 5-2 is a comparison of the properties of the PZT presented in this dissertation along with the proper ties of thin-film PZT as repor ted in the literature. Upon comparison, the relative permittivity and loss tangent are within range of previously reported values, however the remanent polarizati on is lower. This is most likely a direct result of the room temperature poling proces s, and future work will seek to understand the precise nature of this result. Table 5-2: Comparison of ferroelectric and dielectric properties of thin film PZT. Author rP 2Ccm r tan Tuttle and Schwartz [167] 27 1000 -Morita et al. [169] 9.2 --Nunes et al. [160] 20 436 0.07 Xia et al. [135] 25 860 0.03 Kueppers et al. [161] 22 480 -Horowitz 6.3 500 0.03

PAGE 128

114 Electrically Actuated Response Laser Scanning Vibrometer Frequency response The results of the electri cally actuated frequency res ponse in terms of magnitude and phase are shown in Figure 5-5 to Figure 512 for a number of devices. Along with each frequency response plot is the corres ponding uncertainty for each measurement, based on the measured coherence. A minimu m of one resonance was detected for each device. Some of the larger devices also show a second resonance within the measured frequency range. Note that the phase shifts by 180 at the resonance, close to the observed peak in the magnitude response. At low frequencies, well below resonance, the devices exhibit a relativel y flat response, although more significant scatter is evident in the data resulting from poor coherence due to signal levels close to the noise floor. Also note that the frequency range varies from graph to graph to focus in on the features of interest. All of the devices tested below have the same materials and layer thicknesses, and only vary with respect to the radius as well as the presence or absence of a central mass. The electrically actuated frequency res ponse that is defined here to be the frequency dependent center deflection, 0 w, that results from an applied voltage, can be found from the circuit of Figure 2-15 to be 0effaDaDradw VsAZZ (5.4) where 1aDaDaD Z sMsC aDradZ is defined in Eq. (2.95), and effA is the effective area of transduction.

PAGE 129

115 Shown in Figure 5-5 is the electrically act uated frequency response for the smallest device, with a radius of 900 m and without a central mass. Note the resonant peak occurring around 50 kHz with a peak response of 0.71 mV 20 40 60 80 10-4 10-2 100 Frequency ResponseMag(EAFR) [ m/V] 20 40 60 80 10-4 10-2 100 Uncertainty 20 40 60 80 -150 -100 -50 0 Freq [kHz]Phase(EAFR) [Deg] 20 40 60 80 5 10 15 Freq [kHz] Data Theory Figure 5-5: Device 1A-6-1, elec trically actuated center deflection for the device with a radius of 900 m and no central mass, pack aged in the quarter-wave resonator package. Frequency response is shown in the left column with the corresponding uncertainty gi ven in the right column. For an identical size device, except with the addition of a central mass, the electrically actuated frequency response is shown in Figure 5-6. With the addition of the central mass, the resonant frequency drops from 50 kHz down to roughly 23 kHz, while the peak response increases to 6.66 mV This demonstrates that the central mass works as intended. Namely, it increases the overall effective mass without a

PAGE 130

116 correspondingly large d ecrease in effective compliance, thereby leading to an overall decrease in the resonant freque ncy. There had been some a pprehension prior to testing as to the net effect of the centra l mass, mainly because significant modeling of the effect of the mass had not been performed prior to fabrication. 10 20 30 40 10-2 100 Frequency ResponseMag(EAFR) [ m/V] 10 20 30 40 10-2 100 Uncertainty 10 20 30 40 -300 -200 -100 0 Freq [kHz]Phase(EAFR) [Deg] 10 20 30 40 10 20 30 40 Freq [kHz] Data Theory Figure 5-6: Device 1A-6-2, elec trically actuated center deflection for the device with a radius of 900 m and a central mass, packaged in the quarter-wave resonator package. Frequency response is sh own in the left column with the corresponding uncertainty gi ven in the right column. Shown in Figure 5-7 is the frequency response for the second smallest device, with an outer radius of 1200 m Similar to Figure 5-5, it does not have a central mass. Note two resonant frequencies are visible. The first resonance occurs near 34 kHz with peak value of 0.44 mV and represents the fundamental m ode of the device, as were the resonances in the previous graphs. The second resonance occurs near 120 kHz and

PAGE 131

117 corresponds to a higher order mode with a peak value of 0.98 mV This was verified through use of the scanning feature on the lase r vibrometer, and is shown in Figure 5-8. 50 100 150 10-3 10-2 10-1 100 Frequency ResponseMag(EAFR) [ m/V] 50 100 150 10-3 10-2 10-1 100 Uncertainty 50 100 150 -300 -200 -100 0 Freq [kHz]Phase(EAFR) [Deg] 50 100 150 20 40 60 80 Freq [kHz] Data Theory Figure 5-7: Device 1A-11-3, el ectrically actuated center defl ection for the device with a radius of 1200 m and no central mass, packaged in the quarter-wave resonator package. Frequency response is shown in the left column with the corresponding uncertainty gi ven in the right column. Figure 5-8: Higher order mode image of Device 1A-11-3, taken using scanning laser vibrometer at 120.9 kHz

PAGE 132

118 Once again, the addition of a central ma ss lowers the resonant frequency, as illustrated in Figure 5-9. In this instan ce, the resonant frequency drops from 34 kHz as measured above to 14 kHz as shown below. It should be noted here that these are physically different devices and the use of th e phrase addition of a central mass is not meant to convey that the same physical stru cture was modified, but simply that the central mass is the main distinguis hing feature between the two devices. 10 20 30 40 10-2 100 Frequency ResponseMag(EAFR) [ m/V] 10 20 30 40 10-2 100 Uncertainty 10 20 30 40 -300 -200 -100 0 Freq [kHz]Phase(EAFR) [Deg] 10 20 30 40 100 200 300 400 Freq [kHz] Data Theory Figure 5-9: Device 1A-10-4, el ectrically actuated center defl ection for the device with a radius of 1200 m and a central mass, pack aged in the quarter-wave resonator package. Frequency response is shown in the left column with the corresponding uncertainty gi ven in the right column. Shown in Figure 5-10 is the frequency res ponse for the second largest radius device without a central mass. This devi ce has an outer radius of 1500 m It exhibits a resonant frequency of around 25 kHz, compared to 34 kHz for the next smaller device.

PAGE 133

119 10 20 30 40 10-4 10-2 100 Frequency ResponseMag(EAFR) [ m/V] 10 20 30 40 10-4 10-2 100 Uncertainty 10 20 30 40 0 50 100 150 Freq [kHz]Phase(EAFR) [Deg] 10 20 30 40 2 4 6 Freq [kHz] Data Theory Figure 5-10: Device 1A-6-5, electrically actu ated center deflection for the device with a radius of 1500 m and no central mass, packaged in the quarter-wave resonator package. Frequency response is shown in the left column with the corresponding uncertainty gi ven in the right column. With the addition of a central mass to th is size device, with outer radius of 1500 m the resonant frequency drops from 25 kHz to 10.5 kHz. This is shown in Figure 5-11 for Device 1A-10-6 and corroborates that the addition of the mass does not increase the stiffness enough to cancel out the lowe ring effect on the resonance frequency. The largest device is shown in Fi gure 5-12, with a radius of 1800 m without a central mass. A resonant frequency of around 20 kHz was measured for this device, as compared to 23.5 kHz for the next smaller device. Also note the smaller resonances at around 26 kHz and 45 kHz that are most likely effects from higher order modes.

PAGE 134

120 5 10 15 10-4 10-2 100 Frequency ResponseMag(EAFR) [ m/V] 5 10 15 10-4 10-2 100 Uncertainty 5 10 15 -100 0 100 Freq [kHz]Phase(EAFR) [Deg] 5 10 15 2 4 6 8 Freq [kHz] Data Theory Figure 5-11: Device 1A-10-6, el ectrically actuated center de flection for a device with a radius of 1500 m and a central mass, pack aged in the quarter-wave resonator package. Frequency response is shown in the left column with the corresponding uncertainty gi ven in the right column. Overall, the electrically act uated frequency response test s demonstrated the general trends that were expected, namely that of a decrease in resonant frequency with each increase in outer radius, as well as a decrease in resonant frequency due to the addition of central pillar in the center of the diaphragm that acts like a point mass. Shown in Figure 5-13 is a summary graph of the low frequency sensitivity as measured well below each individual resona nce. Note the general trend towards increasing sensitivity with increasing radius as expected from a physical standpoint, although the variation in sensitivity among id entical devices is on the same order of

PAGE 135

121 magnitude. These results as well as the meas ured resonant frequenc ies and sensitivities at resonance are summarized in Table 5-3. 10 20 30 40 10-4 10-2 100 Frequency ResponseMag(EAFR) [ m/V] 10 20 30 40 10-4 10-2 100 Uncertainty 10 20 30 40 -300 -200 -100 0 100 Freq [kHz]Phase(EAFR) [Deg] 10 20 30 40 2 4 6 8 10 12 Freq [kHz] Data Theory Figure 5-12: Device 1A-3-7, electrically actu ated center deflection for the device with a radius of 1800 m and no central mass, packaged in the sealed-cavity package. Frequency response is sh own in the left column with the corresponding uncertainty gi ven in the right column. A graph of the measured resonant frequenc ies under electrical excitation is shown in Figure 5-14. Two series of data are show n, both with and without a central mass. The central mass was measured to have the effect of reducing the resonant frequency as was expected. These results show that, as desired, the effect of the addi tion of mass due to the central pillar was dominant over any decrea se in compliance that resulted from the presence of the pillar. Also, as the radi us is increased, the resonance frequency decreases. Overall, the range of re sonant frequencies varies from 6.73 kHz to 60.34 kHz

PAGE 136

122 800 1000 1200 1400 1600 1800 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 Outer Radius (R2) [ m]Low Frequency Sensitivity [ m/V] Open Cavity w/ central mass Open Cavity w/o central mass Sealed Cavity w/ central mass Sealed Cavity w/o central mass Theory Figure 5-13: Electrically actuated sensitivity at low frequency (well below resonance) ( 120.95 RR ). 800 1000 1200 1400 1600 1800 10 20 30 40 50 60 70 Outer Radius (R2) [ m]Mechanical Resonant Frequency [kHz] Exp. w/o mass Exp. w/ mass Theory w/o mass Theory w/ mass Figure 5-14: Summary of electrically actuated resonant frequencies.

PAGE 137

123 The effect of the central mass is illustrated more explicitly in Figure 5-15, where the negative shift in resonant frequency due to the additional mass is plotted versus the original resonant frequency. The values ar e expressed in percenta ge change from the original values, where a positive number indicates a drop in resonant frequency. The mean percentage change was experimentally found to be 58.93 %, as compared to a predicted shift of 57.76 %. Also the shift wa s found to be more prominent for the larger devices with the lower resonant frequencies 20 30 40 50 60 70 0 20 40 60 80 100 Resonant Frequency [kHz]Frequency Shift [%] Theoretical Experimental Figure 5-15: Drop in resonant frequency due to the addition of the central mass. Overall, significant numerical differences were seen between theory and experiment, as evidenced in Table 5-3. Qual itatively, however the device behavior did generally follow the expected trends, particular ly for the resonant frequencies, as seen in Figure 5-14 and Figure 5-15. On the othe r hand, while the low frequency sensitivity did generally increase with radius, as exp ected, the experimental behavior was not as

PAGE 138

124 predictable and varied greatly in response to packaging a nd the presence of the central mass. The measured behavior was most consis tent with the theory in the sealed cavity packages, suggesting that unwanted acoustical interactions related to the package geometry, may be altering the low frequency sensitivity of the open cavity packaged devices. Table 5-3: Summary of elect rically actuated frequency response measurements. Experimental Theoretical %Difference Device # 2R m mR m mresf kHz lowEASens mV resEASens mV mresf kHz lowEASens mV %mresf %lowEASens 1A-5-1 900 0 n/a n/a n/a 65.68 0.011 n/a n/a 1A-6-1 900 0 50.8 0.005 0.30 65.68 0.011 22.54 54.55 1A-6-2 900 180 23.30 0.015 6.66 28.08 0.011 17.02 -36.36 1A-10-2 900 180 25.20 0.007 0.47 28.08 0.011 10.26 36.36 1A-11-3 1200 0 34.25 0.025 0.45 48.14 0.013 28.85 -92.31 1A-5-4 1200 240 16.69 0.005 1.54 20.37 0.013 18.07 61.54 1A-10-4 1200 240 14.12 0.030 19.70 20.37 0.013 30.68 -130.7 1A-6-5 1500 0 25.2 0.015 0.88 38.01 0.014 33.70 -7.143 1A-6-6 1500 300 9.53 0.050 8.18 15.98 0.014 40.36 -257.1 1A-10-6 1500 300 10.56 0.010 6.60 15.98 0.014 33.92 28.57 1A-3-7 1800 0 19.84 0.008 4.20 31.41 0.015 36.84 46.67 1A-6-7 1800 0 22.72 0.020 0.62 31.41 0.015 27.67 -33.33 1A-3-8 1800 360 6.74 0.078 12.20 13.15 0.015 48.75 -420.0 Linearity Once the frequency response was measured for each device, experiments to characterize the linearity of the response were performed. The measurements were made at the resonant frequency of each device, in order to maximize deflection. A zero DC offset sinusoidal voltage signal was used to dr ive the diaphragm while the velocity at the center of the diaphragm was measured using th e laser vibrometer. The center deflection was then derived from this measurement a nd plotted versus the driving voltage. The resulting plot is shown in Figure 5-16 for all of the tested devices. In the figure, most of the measured devices exhibited a linear response up to about 50 mV of input voltage.

PAGE 139

125 Increasing the voltage beyond that did not lead to a corresponding increase in deflection for these devices. From these results, the ma ximum input voltage fo r a linear response appears to be independent of the device radi us. This suggests that it is not due to a mechanical factor that would increase when the device is scaled. Instead, it is most likely due to a depoling of the piezoelectric materi al, leading to a reduc tion in piezoelectric properties and the re sulting deflection. 101 102 101 102 103 AC Driving Voltage [mV]Center Deflection [nm] Device 1A-6-1 Device 1A-10-2 Device 1A-5-4 Device 1A-6-5 Device 1A-10-6 Device 1A-3-7 Figure 5-16: Magnitude of the resonant center deflection versus driving voltage. Another convenient way of looking at th e nonlinear behavior is by calculating the mechanical sensitivity to voltage inputs as th e driving voltage is varied. The resulting plot is shown in Figure 5-17. From this plot it is easy to see where the sensitivity starts to drop for most of the devices, indica ting the onset of nonlinear behavior.

PAGE 140

126 101 102 102 103 104 AC Driving Voltage [mV]Mechanical Sensitivity [nm/V] Device 1A-6-1 Device 1A-10-2 Device 1A-5-4 Device 1A-6-5 Device 1A-10-6 Device 1A-3-7 Figure 5-17: Mechanical sensitivity at resonance versus driving voltage. Acoustical Measurements Acoustically Actuated Frequency Response Plane-Wave Tube Sealed cavity package The devices that were packaged in the sealed cavity packages (Device 1A-10-2, Device 1A-5-4 and Device 1A-3-7) were tested in the plane-wave tube, with a nominally average incident acoustic pressure of 92.6 dB. The frequency response results shown below in terms of magnitude and phase ar e plotted in Figure 5-18 over the frequency range of 250 Hz to 6700 Hz for Device 1A-5-4, which has a radius of 1200 m and a central mass. The plots for the other devices were not shown to avoid needless repetition,

PAGE 141

127 as they are all highly similar in behavior. The lower frequency limit was set to avoid speaker inefficiencies while the upper limit wa s due to the onset of higher order modes in the tube. The important thing to note from all of these results is the relatively flat magnitude over the frequency range, which is to be expected well below resonance. A slight dip in the low frequency region is mo st likely due to a vent resistance arising from a small leakage in the packaging. The other factor that will lead to a low-frequency rolloff is due to the p eB R C time constant of the piezoelect ric. For the device shown in Figure 5-18, p eB R C =0.0089, and therefore the relate d corner frequency, given by 1 p eB R C was found to be 112.7 Hz As can be seen in Figure 5-18, the corner is closer to 1792 Hz and therefore unrelated to the dielectric losses. Among all of the tested devices, a relative ly flat magnitude of the response was seen, with the numerical value of the ma gnitude as the only significant difference between devices. This value was found to in crease with increasing radius and occurs because the testable frequency range was we ll below all of the resonant frequencies, leading to a compliance dominated testing regi me. Once again, this frequency range was limited by the cut-on conditions for higher order modes in the plane wave tube. Quarter-wave resonator package The same experiment was then performed on the devices that were packaged in the quarter wave resonator package (Device 1A -5-1, Device 1A-11-3, Device 1A-10-4, Device 1A-6-5, Device 1A-10-6, Device 1A-3-7 and Device 1A-3-8). The results for Device 1A-10-4 with a radius of 1200 m and a central mass and Device 1A-6-5 with a radius of 1500 m and no central mass are shown in Figure 5-19 and Figure 5-20 in terms of magnitude and phase. Note that a dip occurs in the magnitude at around 3.694

PAGE 142

128 kHz instead of the flat responses of the sealed cavity packages. The dip arises due to a quarter wave anti-resonance that causes a maximum in the acoustic impedance of the vent channel, as seen by the diaphragm. At this frequency, the diaphragm sees a significantly higher impedance, and the resul ting load acts to reduce the deflection, thereby leading to a stiffening effect on the diaphragm. 1000 2000 3000 4000 5000 6000 10-6 10-5 Mag(AE FR) [V/Pa]Freq [Hz] 1000 2000 3000 4000 5000 6000 -100 -50 0 50 Phase(AE FR) [Deg]Freq [Hz] Data Theory Figure 5-18: Device 1A-5-4, magnitude and pha se of the acoustically actuated frequency response in a sealed cavity package fo r the device with a radius of 1200 m and a central mass. Using an analytical approach to calculate the impedance of a tube with an open end yields a theoretical maximum in impedance (and hence minimum in sensitivity) occurring at 3.882 kHz leading to a difference of 4.84% between the experimental and theoretical values. The radiation impedan ce from the open end of the tube was also considered, however it was found to have only a small impact on the response, by

PAGE 143

129 extending the effective le ngth of the tube to efflll where l is the effect due to the radiation mass. 1000 2000 3000 4000 5000 6000 10-8 10-5 10-3 Mag(AE FR) [V/Pa]Freq [Hz] 1000 2000 3000 4000 5000 6000 0 50 100 150 200 Phase(AE FR) [Deg]Freq [Hz] Data Theory Figure 5-19: Device 1A-10-4 in a quarter-wave resonator package. Magnitude and phase of the acoustically actuated frequency re sponse for the device with a radius of 1200 m and a central mass. For all of these devices, the antiresonant frequency is the same while the overall magnitude increases with each larger device The constant antiresonant frequency is indicative of this packaging do minated effect, rather than a device dominated effect. The packaging is identical for all of these de vices, meanwhile the geometry of the device itself varies. Furthermore, the antiresonant frequency has a quarter wavelength that corresponds closely to the length of the backsi de hole in the resonator package. This quarter-wave resonator behavior was not initially intended, but became obvious in retrospect, considering the geometry of the package.

PAGE 144

130 1000 2000 3000 4000 5000 6000 10-8 10-5 10-3 Mag(AE FR) [V/Pa]Freq [Hz] 1000 2000 3000 4000 5000 6000 -200 -100 0 100 200 Phase(AE FR) [Deg]Freq [Hz] Data Theory Figure 5-20: Device 1A-6-5 in a quarter-wave resonator package. Magnitude and phase of the acoustically actuated frequency re sponse for the device with a radius of 1500 m and no central mass. Acoustic Input Impedance Mea surements Plane-Wave Tube The results of the acoustical input impedance measuremen ts on the devices in the sealed cavity packages are shown in Figure 5-21 to Figure 5-23 in terms of normalized resistance and reactance. The normalization was performed with respect to the acoustic impedance of the plane wave tube, and can be found as inin oo tube tubeZZ j c Z A (5.5) where tubeA is the cross-sectional area of the tube.

PAGE 145

131 2000 2500 3000 3500 4000 4500 5000 5500 6000 0.94 0.96 0.98 1 Frequency [Hz]Mag(R) 2000 2500 3000 3500 4000 4500 5000 5500 6000 -20 -10 0 10 20 Frequency [Hz]Phase(R) [Deg] Theory Data Figure 5-21: Device 1A-10-2, Normalized acous tic impedance in a sealed cavity package for the device with a radius of 900 m and a central mass. The impedances were shown over the same range of 250 Hz to 6700 Hz. The reactance of these devices was generally negative, as would be expected below resonance, where the behavior is domina ted by the compliance of the diaphragm. Additionally, the resistance s hows some unusual and incons istent features below 2 kHz; however, due to the close proximity to unity of the reflection coefficient at these frequencies, small measurement errors and un certainties are magnified when calculating the acoustic impedance.

PAGE 146

132 2000 2500 3000 3500 4000 4500 5000 5500 6000 0.9 0.95 1 Frequency [Hz]Mag(R) 2000 2500 3000 3500 4000 4500 5000 5500 6000 -20 -10 0 10 20 Frequency [Hz]Phase(R) [Deg] Theory Data Figure 5-22: Device 1A-5-4, Normalized acoustic impedance in a sealed cavity package for the device with a radius of 1200 m and a central mass. 2000 2500 3000 3500 4000 4500 5000 5500 6000 0.9 0.95 1 Frequency [Hz]Mag(R) 2000 2500 3000 3500 4000 4500 5000 5500 6000 -30 -20 -10 0 10 20 Frequency [Hz]Phase(R) [Deg] Theory Data Figure 5-23: Device 1A-3-7, Normalized acoustic impedance in a sealed cavity package for the device with a radius of 1800 m and no central mass.

PAGE 147

133 Quarter-Wave Resonator Package The same measurements were then performe d on the devices that were packaged in the quarter-wave resonator package (Device 1A-6-1, Device 1A-11-3, Device 1A-10-4, Device 1A-6-5, Device 1A-10-6, Device 1A-3-7 and Device 1A-3-8). The results of some of these measurements are shown in Figure 5-24 and Figure 5-25, for Device 1A6-1 with a radius of 900 m and no central mass and Device 1A-11-3 with a radius of 1200 m and also no central mass, respectively. The plots for the other devices were not shown to avoid needless repetition, as they are all highly similar in behavior. Note how the reactance is no longer solely negative, due to the presence of the quarter wave resonator. Furthermore, a peak was consis tently observed in the resistance near 1500 Hz. 2000 2500 3000 3500 4000 4500 5000 5500 6000 0.8 0.85 0.9 0.95 1 Frequency [Hz]Mag(R) 2000 2500 3000 3500 4000 4500 5000 5500 6000 -20 0 20 Frequency [Hz]Phase(R) [Deg] Theory Data Figure 5-24: Device 1A-6-1, Normalized acoustic impedance in a quarter-wave resonator package for the device with a radius of 900 m and no central mass.

PAGE 148

134 The important thing to note from Figure 5-24 and Figure 5-25 is the decidedly different character of the impedance as compar ed to the previous experiments with sealed cavity devices. With the sealed cavity devi ces, the reactance was primarily negative and increased more or less monotonically with fr equency, however the introduction of the quarter-wave resonator package alters the impedance such th at positive reactances occur more frequently and the reactance in general ha s a more variable natu re. The significant changes in the impedance highlight the role of the packaging in the overall behavior. 2000 2500 3000 3500 4000 4500 5000 5500 6000 0.8 0.9 1 Frequency [Hz]Mag(R) 2000 2500 3000 3500 4000 4500 5000 5500 6000 -20 -10 0 10 20 Frequency [Hz]Phase(R) [Deg] Theory Data Figure 5-25: Device 1A-11-3, Normalized acoustic impedance in a quarter-wave resonator package for the device with a radius of 1200 m and no central mass. Initial Deflection Measurements Wyko Optica l Profilometer Using the Wyko Optical Profilometer, the in itial static deflection was measured for four devices. The initial deflection that aris es due to residual stresses in the composite

PAGE 149

135 structure is plotted versus radi us in Figure 5-26 to Figure 5-29. Note that a small step is evident in the region near the clamped edge, resulting from the height difference between the piezoelectric ring and diaphragm surface. Also note that the two even numbered devices, Figure 5-27 and Figure 5-29, have a fl atter central region due to the presence of the central pillar. Figure 5-26: Device 1A-11-3, Ini tial static deflection resulting from residual stresses for a device with a radius of 1200 m and no central mass. There are several key points of note from the initial static deflection plots. First of all, the static deflection was found to be si gnificant and on the same order of magnitude as the diaphragm thickness. The models that were used to predict the device behavior are linear models that depend on the assumption of small deflections. The large initial deflections violate this assumption and leads to a device that behaves differently than predicted using a linear theory.

PAGE 150

136 Figure 5-27: Device 1A-6-6, Initial static defl ection resulting from residual stresses for a device with a radius of 1500 m and a central mass. Figure 5-28: Device 1A-3-7, Initial static defl ection resulting from residual stresses for a device with a radius of 1800 m and no central mass. Knowledge of the deflection curv e is critical to accurate prediction of the device, as many of the lumped element parameters, such as the piezoelectric coefficient, Ad, are fundamentally derived from the curve. Because of the nonlinear nature of the deflection, the models cannot accurately predict the device behavior. Therefore, in order to compare

PAGE 151

137 the experimental data against a more r easonable theoretical prediction, measured deflection curves were used to compute new values for the piezoelectric coefficient, Ad. Additionally, the measured dielectric loss resistance, pR, and damping factor,m R were included into the model, are shown in Table 54 In the theoretical curves, previously shown in Chapter 2, these were not include d. The energy harvesting measurement section utilizes these new predictions for comparis on to the measurements. Notice from Table 5-4 that the measured Ad values are much lower than the predicted values of 131.9310 and 122.4210 for Device 1A-4-4 and Device 1A-3-8 respectively. The explanation behind this phenomenon will be discussed later in this chapter. Table 5-4: Revised lumped element parameters used for comparing experimental data to theory. New and revised columns are marked with an asterisk. Device k Ad* aDC aD M aDmassM ebC p R *m R Pa V V Pa 3m V 3m Pa 4kg m 4kg m nF k 4kg ms 1A-4-4 -9.29 -7.16E-6 0.0083.08E-1 42.07E-142602.5 11052 26.9 414.63.24E7 1A-3-8 -16.92 -4.71E-5 0.0282.15E-1 41.43E-131009.8 4912.2 51.3 233.55.09E7 Figure 5-29: Device 1A-3-8, Initial static defl ection resulting from residual stresses for a device with a radius of 1800 m and a central mass.

PAGE 152

138 Energy Harvesting Measurements The primary application of the work presente d in this dissertation is as a harvester of acoustic energy. For this application, it is useful to look at the device performance at resonance as a function of the applied acousti c pressure. In order to accomplish this, a load is placed across the output of the resona tor and the voltage is measured across that load as the pressure is varied. Before va rying the pressure, however, it is useful to optimize the load in order to maximize the energy transfer. Recall from Eq. (2.12), that under the constraint of a purely resistive load, the power transfer to the load is maximized when the load resistance is equal to the magnitude of the output impedance. This optimal value was determined experimentally by measur ing the power while the load is varied. The resulting value was then compared against the measured output impedance of the device, in order to check for c onsistency with Eq. (2.12). Optimal Resistance Using a Stanford Research Systems SRS 785 Dynamic Signal Analyzer, the optimal resistance for maximum power transfer wa s determined. First however, the exact resonant frequency values had to be determ ined, therefore the SRS785 was first set to record 500 averages of 800 frequency bins over a frequency span of 25.6 kHz. For Device 1A-4-4, no window was needed as a ch irp input was used, while for Device 1A-38, a Hanning window was used in conjuction with a band-limited white noise signal input. This measurement resulted in m easured resonant frequencies of 13.568 kHz and 5.232 kHz for Device 1A-4-4 and Device 1A-3-8, respectively. Once the resonant frequencies were determined, the source was changed to a sinusoidal signal at each of these frequencie s for the respective devices. The output voltage was then measured while the load resistance was varied from 46.4 to 1.003

PAGE 153

139 M The output power was then calculated according to Eq. (2.117) and the results were then plotted as shown in Figure 5-30. The figure is overlaid with the theoretical values for comparison. As can be seen, ther e is less than one order of magnitude between data and theory, which overpredicted the pow er for Device 1A-4-4 and underpredicted the power for Device 1A-3-8. The experimentally determined, optimal resistance for both of these devices was found to be 982.9 Optimal Energy Harvesting Using the optimal frequency and load re sistance measured pr eviously, the input signal was steadily increased in amplitude wh ile the output voltage was measured. The power was then calculated based on the volta ge and resistance measurements, again according to Eq. (2.117). The resulting output voltage for each device is shown in Figure 5-31 and is overlaid with the theoretical predictions for the output voltage. 101 102 103 104 105 106 10-16 10-15 10-14 10-13 10-12 10-11 10-10 Load Resistance [ ]Power Delivered to Load [W] Data-Device 1A-4-4 Data-Device 1A-3-8 Theory-Device 1A-4-4 Theory-Device 1A-3-8

PAGE 154

140 Figure 5-30: Measured power delivered to a lo ad as function of the load resistance for Device 1A-4-4 and Device 1A-3-8 as compared against theoretical values. The input acoustic pressure was 92 dB for these measurements. As can be seen in the figure, th e output voltage was linear up to 125 dB for Device 1A-4-4 and up to 133 dB for Device 1A-3-8. The output voltage was seen to range between 22 V and 4.6 mV. Using the voltage data, the output power was then calculated and is shown in Figure 5-32. 90 100 110 120 130 140 150 100 101 102 103 104 105 Input Acoustic Pressure [dB]Voltage across Load [ V] Data-Device 1A-4-4 Data-Device 1A-3-8 Theory-Device 1A-4-4 Theory-Device 1A-3-8 Figure 5-31: Measured output voltage across the load as a function of applied acoustic pressure and compared to theoretical values. The load resistance was fixed at the optimal value of 982.9 (f = 13.568 kHz and 5.232 kHz for Device 1A4-4 and Device 1A-3-8, respectively) The power density was then calculated ba sed on a square unit cell with lateral dimensions equal to the diameter of the di aphragm. The resulting values are shown in Figure 5-33 and are again overlaid with theoretical values. Note from the graph that the maximum power density measured was around 1 2Wcm at around 150dB which is

PAGE 155

141 considerably lower than the avai lable acoustic power density of 100 2mWcm at the acoustic pressure of 150 dB. 90 100 110 120 130 140 150 10-14 10-12 10-10 10-8 10-6 Input Acoustic Pressure [dB]Power Delivered to Load [W] Data-Device 1A-4-4 Data-Device 1A-3-8 Theory-Device 1A-3-8 Theory-Device 1A-4-4 Figure 5-32: Measured power delivered to load as a function of applied acoustic pressure and compared to theoretical values. The load resistance was fixed at the optimal value of 982.9 (f = 13.568 kHz and 5.232 kHz for Device 1A-4-4 and Device 1A-3-8, respectively) This can be seen more clearly by looking at the efficiency, The efficiency was calculated according to Eq. (2. 118) for each of the devices, and is shown in Figure 5-34. As can be seen, the efficiency was fairly constant near 0.012 % for Device 1A-4-4 in the linear regime and 4410 % for Device 1A-3-8 in the linear regime. For comparison, the theoretical model predicts efficiencies of 0.056 % and4310 % respectively.

PAGE 156

142 90 100 110 120 130 140 150 10-14 10-12 10-10 10-8 10-6 10-4 Acoustic Input Pressure [dB]Power Density [W/cm2] Theory-Device 1A-3-8 Theory-Device 1A-4-4 Data-Device 1A-4-4 Data-Device 1A-3-8 Figure 5-33: Measured power density delivered to load as a function of applied acoustic pressure and compared to theoretical values. (f = 13.568 kHz and 5.232 kHz for Device 1A-4-4 and Device 1A-3-8, respectively) 90 100 110 120 130 140 150 10-5 10-4 10-3 10-2 101 Input Acoustic Pressure [dB]Efficiency [%] Data-Device 1A-4-4 Theory-Device 1A-4-4 Data-Device 1A-3-8 Theory-Device 1A-3-8 Figure 5-34: Measured overall efficiency of each device overlaid with theoretical values for comparison. (f = 13.568 kHz and 5.232 kHz for Device 1A-4-4 and Device 1A-3-8, respectively).

PAGE 157

143 A question then arises regarding the orig in of the low efficiency and whether improvements in either the manufacturing or design of the devices could improve the efficiency, and thereby the output power. Seve ral issues were faced during the fabrication process that limited the efficiency of the devices. First, the materials and processes led to a large residual stress in several layers of th e device, most notably the titanium dioxide. The effect of this large tensile stress, 2TiO is twofold: (1) it al ters the deflection mode shape of the devices and (2) it leads to a large, non-linear, initial deflection of the diaphragm. The end result is a modifica tion to the piezoelectric coefficient, Ad and ultimately the coupling coefficient, k. It is difficult to asce rtain the exact magnitude of this effect since the deflection lies in the nonlinear regime and the analytical model is only linear. Nevertheless, a 1st order estimate of the effect can be made by using the linear model and comparing the results for the existing high-stress case to an ideal zerostress case. Using this method, the power density decreased by a factor of 4.3 over the ideal zero-stress case. A second issue arising from the fabrica tion is the overha nging metal on the electrodes that created short circuits unde r high electric fields. The field limitation imposes a maximum poling voltage that can be a pplied. As the piezoelectric coefficient, Ad, is proportional to the polarization, the full potential poling was not realized. The effects of this issue, along with the unmode led (non-linear) portion of the stress induced effects, lead to a reduction in power dens ity by a factor of 24.6 below the ideal case. The final fabrication issue concerns the c hoice of the starting piezoelectric material and the process by which it is deposited on the wafer. The particular sol-gel technique that was employed in this research produces a typical 31d of -50 p CN, whereas other

PAGE 158

144 variations of the sol-gel process as well as techniques such as sputtering can produce a film with a 31d of -120 p CN. The output power density is sm aller by a factor of 2.8 as a result of the lower value of 31d as compared to other reported PZT thin films. The final issue that reduces the output power density concerns the quality factor, Q, of the device. The measured quality fact or for Device 1A-3-8 is approximately 10, leading to a damping factor of 0.05. This damping accounts for radiation resistance and internal damping, as well as radiation to the supports. This damping factor is much higher than a more typical 0.02, leading to a reduction in power density by a factor of 2.5 Through improvements in the fabrication and design, leading to a reduction in the damping factor, the output power density would increase by a factor of 2.5. Additionally, provided that all of the othe r sources of inefficiency are addressed, including the PZT material quality, the poling capability, and th e residual stress, the overall power density would increase by a factor of 740. The resu lting power density curv e is shown in Figure 5-35, overlaid with the currently achievable results. With these improvements, at 150 dB, the output power densit y would be on the order 2500 Wcm. Additionally, the individual contributions to this improvement are illustra ted in Figure 5-36, clearly showing the prominent role of improving Ad to the overall results. Nonlinear behavior is eviden t in Figure 5-31 through Figur e 5-34. Physically, there are several possible sources for the nonlinea r behavior. The overall response may be affected by one or all of these sources. Fi rst of all, a mechani cal nonlinearity arises, when the deflection of a plate is on the order of the thickness of that plate. Under large deflections, the neutral axis of the plate st retches by a finite amount, requiring energy in addition to the energy required for bending of the plate. The resulting effect is an

PAGE 159

145 increase in stiffness above that predicted by linear, small-deflection plate theory. The increase in stiffness leads to a lower sensitiv ity and an increased resonant frequency. 90 100 110 120 130 140 150 10-14 10-12 10-10 10-8 10-6 10-4 102 Acoustic Input Pressure [dB]Power Density [W/cm2] Realistically Optimal Current Results Figure 5-35: Currently achievable power dens ity for Device 8 and potentially achievable power density using an impr oved fabrication process. 72% 13% 8% 7% dA TiO2 Ideal PZT Q Figure 5-36: Breakdown of the individual cont ributions to the overall improvement in power that is achievable under better fabrication conditions.

PAGE 160

146 The effect on the resonant frequency was measured for one particular device that exhibited strong nonlinear behavior and is show n in Figure 5-37. The resonant frequency was seen to approximately shift linearly as the acoustic pressure was increased. Additionally, in the frequency domain, the nonlinearity generates harmonics at integer multiples of the fundamental signal frequency, thereby dumping some of the energy that would have otherwise resided at the fundamental frequency. 0 200 400 600 800 1000 10.15 10.2 10.25 10.3 10.35 10.4 Input Acoustic Pressure [mPa]Resonant Frequency [kHz] fres = 250*Pin + 10.15E3 [Hz] data 1 linear fit Figure 5-37: Resonant frequenc y vs. acoustic pressure, arising from nonlinear response.

PAGE 161

147 CHAPTER 6 ALTERNATIVE APPLICATION PIEZOELECTRIC MICROPHONE The previous chapter on experimental characterization measured the overall characteristics of the device in terms of im pedance, acoustical response, and electrical response in a general sense, as well as an evaluation of the performance of the device as an energy harvester. The goal was to unde rstand the actual behavior and to try and explain the difference between the actual behavior and the predicted behavior. This chapter serves to enhance our u nderstanding of the device performance by looking at an alternative application of the piezoelectric composite diaphragm, by using it as a piezoelectric microphone. For the ener gy harvester devices, we are using the piezoelectrically generated charge as an energy source, whereas in a microphone application, this charge (or corresponding voltage) is a signa l for transmitting information about the acoustical environment. For this set of experiment s, a piezoelectric diaphragm with 1.8 mm diameter and no central mass was used. Experiments were performed both as a sens or (acoustically ac tuated) and as an actuator (electrically actuated) in order to obtain the performance specifications. The packaged device was mounted to the end of a 1 x 1 cross section plane wave tube (PWT) for obtaining the frequency response a nd linearity. A Bruel and Kjaer (B&K) 1/8 Type 4138 microphone was mounted on the wall of the PWT near the end of the tube to record the incident sound pressure level (SPL). The signal was provided by a PULSE Multi-Analyzer system to a BMS 2 coaxial co mpression driver, via a Techron amplifier.

PAGE 162

148 The PULSE analyzer was also used to reco rd the signals from the B&K microphone and the piezoelectric microphone. First, the linearity was tested using a sinusoidal signal at 1 kHz. Measurements were taken with a binwidth of 1 Hz and the signal falling direc tly on the center of a bin to avoid leakage. Additionally, 100 averages were taken at each point. The measured linearity of the sensor at 1 kHz is shown in Figure 6-1. The device was experimentally found to be linear up to (at least) the ma ximum testable sound pressure level of 169 dB. The sensitivity was measured over th is range to be approximately 11.0 VPa Furthermore, the response flattens out below 45 dB as the signal level approaches the noise floor. The maximum testable level is limited here by the out put capacity of the compression driver. The measured sensitivity was found to be 0.75 VPa 75 100 125 150 175 10-8 10-7 10-6 10-5 10-4 10-3 10-2 Output Voltage [V]Input Acoustic Pressure [dB] Data Fit R2=0.9995 Figure 6-1: Linearity of the microphone device at 1 kHz.

PAGE 163

149 The frequency response was measured us ing a periodic random noise acoustic signal of 94dB and 1000 spectral averages over a band width from 0 to 6.4 kHz with a 1 Hz bin width. Once again, the testable frequency range was limited on the upper end by the onset of higher order modes. The resu lting frequency response of the sensor is shown in Figure 6-2 in terms of magnitude and phase. 101 102 103 10-7 10-6 10-5 Frequency [Hz]Magnitude [V/Pa] 101 102 103 -50 0 50 100 Frequency [Hz]Phase [Deg] Figure 6-2: Frequency res ponse spectrum in terms of magnitude and phase. The response is flat over the testable frequency range up to 6.7 kHz, except for a low frequency rolloff at approximately 100 Hz. This is most likely due to the vent resistance; however, the magnitude of this c ontribution is unknown, as the value for the vent resistance due to packaging leaks has not been measured. Additionally, there is a low-frequency rolloff due to the parallel combination of the capacitance and resistance in

PAGE 164

150 the piezoelectric material, however the corner frequency due to this combination is 6.7 Hz, well below that of the measurable corner frequency of 100 Hz. The noise floor spectrum of the sensor is shown in Figure 6-3 from 0 Hz to 12.8 kHz, along with the noise floor due to the experimental setup alone. The noise floor was found to generally decrease with increasing frequency. The noise spectrum of the experimental setup alone wa s obtained by shorting the inpu ts to the preamplifier and recording the resulting output voltage signal, using the same conditions and setup as the sensor noise floor measurement. This noise spectrum is characteristic of a resistance shunted by a capacitance. The resistance in this case is the sensor dielectric resistance, p R and the capacitance is the blocked electrical capacitance, ebC. The resistance and capacitance form a low-pass tran sfer function that spectrally shapes the thermal noise. When integrated over an infinite bandwidth, the resulting mean square noise voltage is given by ebkTC, where k is Boltzmanns constant and T is the ambient temperature. For this device, the low pass filter has a cutoff frequency of 6.7 Hz, determined by 12peb R C There are three characteristic regions in this spectrum. Below 6.7 Hz, there exists a relatively flat region where the noise is dominated by the resist ance of the sensor. Then, from 6.7 Hz up to approximately 2 kHz a 1 f rolloff in the noise is observed that is consistent with a capacitive dominated noise source. Above 2 kHz, the noise spectrum flattens out again as the sensor noise approaches the setup noise. For a 1 Hz bin width centered at 1 kHz, the output voltage with no acoustic signal applied is 3.69 nV that corresponds to an equivale nt acoustic pressure of 4.93 mPa or 47.8 dB.

PAGE 165

151 100 102 104 45 50 55 60 65 70 75 80 85 Freq [Hz]Magnitude [dB] Sensor Setup Figure 6-3: Noise floor spectrum of output voltage when no acous tic signal is applied, as well as noise floor due to measurement setup alone. Because the device is a linear, reciprocal electromechanical transducer, it can operate in both an electrically driven and ac oustically driven modes and the electrically driven frequency response can be used to gain additional information about the sensor, such as the bandwidth. In particular, as the acoustically actuated frequency response could only be measured up to 6700 Hz due to testing limitations of the plane wave tube, the bandwidth can be approximated via us e of the electrically actuated frequency response. The electrically actuated frequency res ponse is shown in Figur e 6-4 in terms of magnitude and phase. The resona nt frequency was found to be 50.8 kHz and provides an estimate of the usable bandwidth as a microphone, as the electrically and acoustically

PAGE 166

152 actuated resonance frequencie s are related. A summar y of the overall performance specifications of the microphone is given in Table 6-1. 10 20 30 40 50 60 70 80 90 -20 -18 -16 -14 Magnitude [dB re 1 V/Pa]Freq [kHz] 10 20 30 40 50 60 70 80 90 -150 -100 -50 0 Freq [kHz]Phase [Deg] Figure 6-4: Electrically actuated frequency respons e of microphone device. Table 6-1: Summar y of experimental results of microphone. Specification Measured Value Diaphragm Size 1.8 mm diameter, 3 m thick Die Size 5 5 mmmm Measured Resonant Frequency 50.8 kHz Sensitivity 0.75 VPa Dynamic Range 47.8169 dB

PAGE 167

153 CHAPTER 7 CONCLUSION AND FUTURE WORK The acoustic energy harvesting technique th at was developed in this dissertation provides a unique and innovative method for extracting energy from an otherwise inaccessible source. For the development of acoustic energy harvesters, piezoelectric composite diaphragms were designed based on analytical models of circular piezocomposite structures. Lumped element mode ling was then used to further understand the dynamics of the system. The modeling served as a basis around which the geometry and dimensions of numerous diaphragms were determined. The lumped element model developed in this dissertation serves as a design basis for other types of piezoelectric composite structures as well as energy harv esting devices employing other transduction methods. Furthermore, it is directly applicab le to those interested in the scaling and design of piezoelectric composite diaphragms. Following the design stage, a fabrication pr ocess flow and mask set was developed to synthesize the devices. Fabrication was then performed as prescribed in the process flow. Following this, the devices were packag ed into two different packages for testing purposes. Finally, a number of experiments were set up and performed on the devices to better understand the real behavior as compar ed to the theory. A graphical overview of this chapter is provided by th e roadmap of Figure 7-1. The main contributions of this disser tation included the development of the acoustic energy harvesting concept. While prev ious researchers had looked at harvesting

PAGE 168

154 energy from vibration, solar, chemical and thermal energy among others, a thorough investigation into the harv esting of acoustic energy had not been performed. Additionally, in order to deve lop the concept, a comprehensive lumped element model (LEM) of the acoustic energy harvester was ut ilized. This model allowed for a single representation for the dynamics of the system as a whole, as well as the individual components. Development Issues Future Work Figure 7-1: Roadmap to Chapter 7. The implementation of the acoustic energy harvester led to the first reported integration of thin-film PZT with SOI-based MEMS. While numerous researchers had previously looked at thin-f ilm PZT, or MEMS using SOI wafers, none had sought to combine them. The high transduction coeffici ent of PZT along with a simpler fabrication route to a diaphragm via SOI wafers led to the decision to combin e these technologies. Finally, although the circular piezoelectric membrane was initially intended as a harvester of acoustic energy, it was discovered that it also served as fairly capable microphone. This is the first reported inst ance of an aeroacoustic capable piezoelectric microphone. Development Issues Several issues arose during the development of these de vices. First of all, the initial process flow that was developed cont ained more steps and masks than the process

PAGE 169

155 flow that was presented here Initially, gold bond pads and an insulating oxide were incorporated to enable on-chip arrays of devices; however, this process flow was later simplified prior to fabrication to reduce cost, time and fabrication complexity. The next issue arose during fabrication of the devices, but was not discovered until packaging of the devices was attempted. Th e problem involved a thin strip of platinum that overhung the piezoelectric ring from the top electrode. Normally, for a metal liftoff step, a image reversal mask is used on positive photoresist to obtain a negative sloping sidewall prior to metal deposition. The negative slope permits a clean break in the metal for improved liftoff results. However, due to processing constraints, we had to use a negative photoresist for this pattern, resulti ng in a positive sloping sidewall. Then during the metal liftoff, a clean break was not ach ieved, resulting in a few microns of overhanging metal around the e dges of the ring. This was not noticed until electrical testing was performed on the devices after fabrication wa s completed. The testing showed that four of the six wafers had el ectrical short circuits between the top and bottom electrodes. Furthermore, the resistance of the short circuits was proportional to the perimeter of the ring, indicating it was due to an edge effect rather than a surface effect. The two remaining wafers still f unctioned correctly, though, despite the same overlapping metal issues, and were solely us ed for the experiments presented here. Another fabrication issue that arose concerne d the residual stress in the devices. It was discovered that the measured resi dual stress of the titanium dioxide (TiO2), at 1800 GPa was significantly higher than anticipated. This led to a large initial static deflection of the diaphragms. Additi onally, the buried oxide laye r on the backside of the

PAGE 170

156 diaphragms proved difficult to remove. This residual stress of this layer further exacerbated the initial static deflection. Unfortunately, the large initial static defl ection placed the devices into a nonlinear regime of plate mechanics and led to a stiffe r device than was anticipated. Furthermore, the models that were developed for the design of the diaphragms were no longer applicable under this nonlinear condition. This caused numerous difficulties in fully understanding the overall behavior of the devices from a quantitative standpoint. Qualitatively, however, the devices did behave as expected, following anticipated trends and dependencies. Ultimately, however, the increased stiffening opened up a new application for the devices as a microphone. Because the stiffer devices had a higher resonance frequency, the usable bandwidth was extended be yond our expectations. Experimental characterization indicates a sensitivity of 11.0 VPa with a dynamic range from 44.1 dB to 153 20dBrePa and a resonant frequency of 22.72 kHz, suggesting a potential suitability as a microphone. Future Work A number of improvements can be made in the design of these devices. The fabrication process flow need s to be modified to avoid the platinum liftoff issue and resulting short circuits. Furthermore, more analysis and optimization of the residual stress in the various layers is needed, in particular for the TiO2. A reduction in this residual stress would improve performance by reducing the diaphragm stiffness, lowering the resonant frequency to a more testable range and ultimately allowing for comparison to a well understood linear model. Alterna tively, a modified process allowing for the

PAGE 171

157 removal of most of the TiO2 would achieve the same results Additional work is also needed on developing a nonlinear ex tension to the model to facilitate the analysis of the existing structures. Other future work involves a more thor ough investigation of this device as a microphone, including a rigorous microphone calibration. Additionally, other applications need to be explored furthe r, including the use of the devices as a piezoelectric actuator and speaker, as well as the potential for active acoustic impedance control.

PAGE 172

158 APPENDIX A MATLAB CODE lem.m clear all ; close all ; eh_setup; %Helmholtz Resonator Ma=(rho*Leff)/(pi*r_neck^2); %acoustic mass of neck Ca=volume/(rho*c^2); %acoustic compliance of cavity Ra=(40*8*mu*L)/(pi*r_neck^4); %acoustic resistance of neck % Readinput; silicon_sweep; %Reading shim properties platinum; %Reading platinum properties tio2; %Reading silicon doxide properties piezo_sweep; %Reading piezo props from console R21= R1/R2; % Give the values PP = 1; VV = 1; h_ca = 2e-4; density_air = 1.165; sound_velocity = 344; initialise; %Initialising sampling etc. N_in = .001; %Initial in-plane stresses in the inner region N_out = sigma02*t02+sigmap*tp; %Initial in-plane stresses in the outer region % Calculate the total deflection for the pressure loading totaldeflection_forP; % Calculate the total deflection for the voltage loading totaldeflection_forV;

PAGE 173

159 % Calculate the initial deflection initialdeflection; % Incremental deflection for pressure loading Wi = W1p W0vp; formmatrix_CMFa; % Voltage loading only Wiv = W1v W0vp; frommatrix_dA; k = sqrt((dA^2)/(Cefp*CaD)); %Calculate the value of 1-k^2 PHI =-dA/CaD; %Electro-acoustic Xduction coefficient PHIprime =-dA/Cefp; Ceb =Cefp*(1-k^2); Rrad_RES=(FREQ*2*pi*R2/c)^2*rho*c/(2*Area); %rad resistance at resonance RloadA=Rrad_RES; %Equivalent Circuit Zrad=tf([0 Rrad_RES],1); %radiation impedance ZaD=tf([MaD 0],1)+tf(1,[CaD 0]); %impedance of diaphragm Zeb=tf(1,[Ceb 0]); %blocked electrical impedance Zload1=tf(RloadA+Ra,1); ZL=tf([0 0],1); ZRopen=tf(10^9,1); ZRshort=tf(1,1); Zload2=tf(RloadA,1); Zneck=tf([Ma Ra],1); %impedance of neck Zcav=tf(1,[Ca 0]); %impedance of cavity Ztube=rho*c/Atube; wnHR=1/sqrt(Ma*Ca); fHR=wnHR/(2*pi); wnD=1/sqrt(MaD*CaD); fD=wnD/(2*pi); % alpha=Ma/MaD; Pmax=2.3*Esi/(R2/tsi)^4 PmaxdBSPL=20*log10(Pmax/20e-6) P=1; Ze1=Zeb*Zload1/(Zeb+Zload1); Ze1short=Zeb*ZRshort/(Zeb+ZRshort);

PAGE 174

160 Ze2=Zeb*Zload2/(Zeb+Zload2); Zin1=Zneck+Zcav*(ZaD+Zrad+PHI^2*Ze1)/(Zcav+ZaD+Zrad+PHI^2*Ze1); Zin2=ZaD+Zrad+PHI^2*Ze2; Zin2short=ZaD+Zrad+PHI^2*Ze1short; Zout1=Zeb*((ZaD+Zrad+Zcav*Zneck/(Zcav+Zneck))/PHI^2)/(Zeb+(ZaD+Zrad+Zca v*Zneck/(Zcav+Zneck))/PHI^2); Zout2=Zeb*((ZaD+Zrad)/PHI^2)/(Zeb+(ZaD+Zrad)/PHI^2); Powin1=P^2/Zin1; Powin2=P^2/Zin2; %Frequency response Vout1=PHI*Ze1*P/(((Zneck/Zcav)+1)*(ZaD+Zrad+PHI^2*Ze1)+Zneck); %voltage generated by diaphragm on HR Vout2=PHI*Ze2*P/(ZaD+Zrad+PHI^2*Ze2); %voltage generated by diaphragm alone Voutmax1=PHI*Ze1*Pmax/(((Zneck/Zcav)+1)*(ZaD+Zrad+PHI^2*Ze1)+Zneck); %max voltage generated by diaphragm on HR Voutmax2=PHI*Ze2*Pmax/(ZaD+Zrad+PHI^2*Ze2); %max voltage generated by diaphragm alone Sens1Open=PHI*Ze1/(((Zneck/Zcav)+1)*(ZaD+Zrad+PHI^2*Ze1)+Zneck); %sensitivity of diaphragm on HR Sens2Open=PHI*Ze1/(ZaD+Zrad+PHI^2*Ze1); %sensitivity of diaphragm alone Sens1Short=PHI*Ze1short/(((Zneck/Zcav)+1)*(ZaD+Zrad+PHI^2*Ze1short)+Zne ck); %sensitivity of diaphragm on HR Sens2Short=PHI*Ze1short/(ZaD+Zrad+PHI^2*Ze1short); %sensitivity of diaphragm alone PA1=Zcav/(Zneck+Zcav); Pow1=Vout1^2/Ze1; Pow2=Vout2^2/Ze2; PowAvail=P^2/Ztube; Eff1=Pow1/Powin1; Eff2=Pow2/Powin2; PowMax1=Voutmax1^2/Ze1; PowMax2=Voutmax2^2/Ze1; bode(Vout1,Vout2); F=linspace(0,7000,100); W=2*pi*F; V1=freqresp(Vout1,W); V2=freqresp(Vout2,W); VMAX1=freqresp(Voutmax1,W); VMAX2=freqresp(Voutmax2,W); P1=freqresp(Pow1,W); P2=freqresp(Pow2,W); POWMAX1=freqresp(PowMax1,W); POWMAX2=freqresp(PowMax2,W); PowIn1=freqresp(Powin1,W); PowIn2=freqresp(Powin2,W);

PAGE 175

161 ZIn1=freqresp(Zin1,W); ZIn2=freqresp(Zin2,W); ZIn2short=freqresp(Zin2short,W); PAmp1=freqresp(PA1,W); SENS1OPEN=freqresp(Sens1Open,W); SENS2OPEN=freqresp(Sens2Open,W); SENS1SHORT=freqresp(Sens1Short,W); SENS2SHORT=freqresp(Sens2Short,W); EFF1=freqresp(Eff1,W); EFF2=freqresp(Eff2,W); PHIarray=ones(length(F),1)*PHI; PHIprimearray=ones(length(F),1)*PHIprime; ZOUT1=freqresp(Zout1,W); ZOUT2=freqresp(Zout2,W); figure(2); subplot(2,1,1) semilogy(F,abs(V1(1,:)),F,abs(V2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Mag(V_o_u_t) [V]' ); subplot(2,1,2); plot(F,angle(V1(1,:))*180/pi,F,angle(V2(1,:))*180/pi); xlabel( 'Frequency [Hz]' ); ylabel( 'Phase(V_o_u_t) [Deg]' ); legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ); figure(3); subplot(2,1,1) semilogy(F,abs(P1(1,:)),F,abs(P2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Mag(\pi_o_u_t) [W]' ); subplot(2,1,2); plot(F,angle(P1(1,:))*180/pi,F,angle(P2(1,:))*180/pi); xlabel( 'Frequency [Hz]' ); ylabel( 'Phase(\pi_o_u_t) [Deg]' ); legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ); hold on ; figure(4); subplot(2,1,1); semilogy(F,real(ZOUT1(1,:)),F,real(ZOUT2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Re(Z_o_u_t) [Ohms]' ); subplot(2,1,2); plot(F,imag(ZOUT1(1,:)),F,imag(ZOUT2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Im(Z_o_u_t) [Ohms]' ); legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ); figure(5) subplot(2,1,1); semilogy(F,abs(PowIn1(1,:)),F,abs(PowIn2(1,:)));

PAGE 176

162 axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Mag(\pi_i_n) [W]' ); subplot(2,1,2); plot(F,angle(PowIn1(1,:))*180/pi,F,angle(PowIn2(1,:))*180/pi); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Phase(\pi_i_n) [Deg]' ); legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ); figure(6); subplot(2,1,1); semilogy(F,abs(EFF1(1,:)),F,abs(EFF2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Mag(Efficiency)' ); subplot(2,1,2); plot(F,angle(EFF1(1,:))*180/pi,F,angle(EFF2(1,:))*180/pi); axis tight ; xlabel( 'Frequency [Hz]' ); ylabel( 'Phase(Efficiency) [Deg]' ); legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ); figure(7) subplot(2,1,1) semilogy(F,real(ZIn1(1,:)),F,real(ZIn2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ) ylabel( 'Input Impedance Real [kg/m^4*s]' ) legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ) subplot(2,1,2); plot(F,imag(ZIn1(1,:)),F,imag(ZIn2(1,:))); axis tight ; xlabel( 'Frequency [Hz]' ) ylabel( 'Input Impedance imag [kg/m^4*s]' ) legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ) figure(8) subplot(2,1,1); semilogy(F,abs(PAmp1(1,:))); xlabel( 'Frequency [Hz]' ); ylabel( 'Magnitude [dB]' ); subplot(2,1,2); plot(F,angle(PAmp1(1,:))*180/pi); xlabel( 'Frequency [Hz]' ); ylabel( 'Phase [deg]' ); figure(9); semilogy(F,abs(SENS2OPEN(1,:)),F,abs(SENS2SHORT(1,:)),F,1./abs(PHIarray ),F,abs(PHIprimearray)); figure(10) semilogy(F,imag(ZOUT1(1,:)),F,imag(ZOUT2(1,:))); axis tight ; xlabel('Frequency [Hz]' )

PAGE 177

163 ylabel( 'Output Impedance [kg/m^4*s]' ) legend( 'with Helmholtz resonator' 'without Helmholtz resonator' ) eh_setup.m clear all ; %Constants rho=1.165; %density of air mu=1.789E-5; %viscosity of air c=344; %speed of sound in air eps0=8.85E-12; %free-space permittivity %Vent/Neck L=3.18e-3; r_neck=(4.72E-3)/2; Leff=L+1.6*r_neck; %effective neck/vent length %Resonator Cavity r_cav=5E-3; depth=2E-3; area_cav=pi*r_cav^2; volume=1950E-9; %Piezo-Diaphragm tsi = 3E-6; %thickness of silicon R2 = 1.95E-3; % radius of silicon %R2=3e-3; tp = 0.6E-6; %thickness of piezo %tp = 10e-6; R1 = .95*R2; %inner radius of piezo %Impedance tube Atube=(1*2.54e-2)^2; silicon_sweep.m Esi = 169e9; %youngs modulus of shim vsi = 0.3; %poissons ratio of shim mu = 1; %const of proportionality of shear stress vs shear strain densitysi = 2500; %Density of the shim platinum.m %all distances in um %http://www.webelements.com/webelements/elements/text/Ag/phys.html Epl=170E9; vpl=0.38; tpl=170e-9; densitypl=21440; tio2.m % tio2 %all distances in um

PAGE 178

164 %http://www.webelements.com/webelements/elements/text/Ag/phys.html E02=283E9; v02=0.28; t02=.1e-6; density02=2150; sigma02=1200e6; piezo_sweep.m %Material Properties of the piezo %all distances in um Ep = 30e9; %Youngs modulus of the piezo vp = 0.3; %Poissons ratio of the piezo material d31 = -50e-12; %electromechanical transduction const of the piezo densityp = 7600; %Density of the piezo epsilon0 = 8.85E-12; %permitivity of free space in F/m dielectricconstant = 1000; %relative permitivity of the piezo epsilon=dielectricconstant*epsilon0; %absolute permitivity of the piezo sigmap = 100E6; initialise.m num1 = 125; r = linspace(0,1,num1+1); deltar= r(2)-r(1); sr=size(r); theta=zeros(sr); totaldeflection_forP.m %%%% totaldeflection_forP.m %%%%%%%%%% P = -PP; V = 0; Ef = V/tp; %pressure loading only abdpiezo; %Calculate the stiffness matrix A B D constants1234; %Obtain the constants c1 c2 c3 c5 solver1; W1p = w; totaldeflection_forV.m %%%%%%%% totaldeflection_forV.m %%%%%%%%%% V = VV; %electric field in V applied in direction 3(Z)

PAGE 179

165 Ef = V/tp; P = 0; %voltage loading only abdpiezo; %Calculate the stiffness matrix A B D constants1234; %Obtain the constants c1 c2 c3 c5 solver1; W1v = w; initialdeflection.m P = 0; V = 0; Ef = V/tp; %no pressure and voltage abdpiezo; %Calculate the stiffness matrix A B D constants1234; %Obtain the constants c1 c2 c3 c5 solver1; W0vp = w; formmatrix_CMFa.m %%%%%% formmatrix_CMFa.m %%%%%%%% % Intergrate the deflection to obtain energy, WW and total deflection jj=floor(R21*num1)+1; for i=1:num1+1 rad1=r(i)*R2; if (i
PAGE 180

166 MaD = 2*pi*wwtot/(wtotal^2); % acoustical mass FREQ = sqrt(1/(CaD*MaD))/2/pi; % resonant frequency %R1byR2(index) = R21; % for the graphs Areap = pi*(R2^2-R1^2); % area of the pzt layer Cefp = epsilon*pi*(R2^2-R1^2)/tp; frommatrix_dA.m %%%%%%% frommatrix_dA.m %%%%%%%%%% % Intergrate the deflection to obtain energy, WW and total deflection jj=floor(R21*num1)+1; for i=1:num1+1 rad1=r(i)*R2; if (i
PAGE 181

167 zout6 = t02 + tpl + tp + tpl + tsi/2; % A B D Matrixs for inner and outer A_in = Qsi.*(zin2-zin1) + Q02.*(zin3-zin2); B_in = Qsi.*((zin2^2-zin1^2)/2) + Q02.*((zin3^2-zin2^2)/2); D_in = Qsi.*((zin2^3-zin1^3)/3) + Q02.*((zin3^3-zin2^3)/3); A_out = Qsi.*(zout2-zout1) + Q02.*(zout3-zout2) + Qpl.*(zout4-zout3) + Qp.*(zout5-zout4) +Qpl.*(zout6zout5); B_out = Qsi.*((zout2^2-zout1^2)/2) + Q02.*((zout3^2-zout2^2)/2) + Qpl.*((zout4^2-zout3^2)/2)+ Qp.*((zout5^2-zout4^2)/2)+Qpl.*((zout6^2zout5^2)/2); D_out = Qsi.*((zout2^3-zout1^3)/3) + Q02.*((zout3^3-zout2^3)/3) + Qpl.*((zout4^3-zout3^3)/3)+ Qp.*((zout5^3-zout4^3)/3)+Qpl.*((zout6^3zout5^3)/3); % Computing D Mark(determinant of matrix mapping defined variables y1,y2 to U0 theta Dstar_in = D_in(1,1)-(B_in(1,1)^2)/A_in(1,1); Dstar_out = D_out(1,1)-(B_out(1,1)^2)/A_out(1,1); % Computing fictitious forces due to piezo Mp_in = 0; Np_in = 0; Mp_out =-Ef* (Ep/(1-vp)) d31 (zout5^2-zout4^2)/2; Np_out =-Ef* (Ep/(1-vp)) d31 (zout5-zout4); constants1234.m %%%%% constants1235.m %%%%%%% k_in = sqrt(N_in*(R1^2)/Dstar_in); k_out = sqrt(N_out*(R2^2)/Dstar_out); a_R = R1/R2; b_jy = besseli(1,k_out)/besselk(1,k_out); Gamma_out = B_out(1,1)/A_out(1,1); Gamma_in = B_in(1,1)/A_in(1,1); AB11_in =-(Gamma_in*A_in(1,1)-B_in(1,1))*k_in/2/R1; AB11_out =-(Gamma_out*A_out(1,1)-B_out(1,1))*k_out/2/R2; AB12_in =-(Gamma_in*A_in(1,2)-B_in(1,2))/R1; AB12_out =-(Gamma_out*A_out(1,2)-B_out(1,2))/R1; BD11_in =-(Gamma_in*B_in(1,1)-D_in(1,1))*k_in/2/R1; BD11_out =-(Gamma_out*B_out(1,1)-D_out(1,1))*k_out/2/R2; BD12_in =-(Gamma_in*B_in(1,2)-D_in(1,2))/R1; BD12_out =-(Gamma_out*B_out(1,2)-D_out(1,2))/R1; x_p = P*R2/2/N_out/besselk(1,k_out); % Compute the matrix A11 = besselj(1,k_in); A12 = 0; A13 = b_jy*besselk(1,k_out*a_R) besseli(1,k_out*a_R); A14 = 0;

PAGE 182

168 A21 =-Gamma_in*besselj(1,k_in) ; A22 = R1; A23 = (-b_jy*besselk(1,k_out*a_R) + besseli(1,k_out*a_R))*Gamma_out; A24 = R1*(1/(a_R^2)-1); A31 = AB11_in*(besselj(0,k_in) besselj(2,k_in)) + AB12_in*besselj(1,k_in); A32 = A_in(1,1)+A_in(1,2); A33 = AB11_out*((besseli(0,k_out*a_R)+besseli(2,k_out*a_R))+b_jy*(besselk(0,k _out*a_R)+besselk(2,k_out*a_R))) ... AB12_out*(besseli(1,k_out*a_R)-b_jy*besselk(1,k_out*a_R)) ; A34 = A_out(1,2)*(1/(a_R^2)-1) A_out(1,1)*(1+1/(a_R^2)); A41 = BD11_in*(besselj(0,k_in)-besselj(2,k_in)) + BD12_in*besselj(1,k_in); A42 = B_in(1,1) + B_in(1,2) ; A43 = BD11_out*((besseli(0,k_out*a_R)+besseli(2,k_out*a_R))+b_jy*(besselk(0,k _out*a_R)+besselk(2,k_out*a_R))) ... BD12_out*(besseli(1,k_out*a_R)-b_jy*besselk(1,k_out*a_R)) ; A44 = B_out(1,2)*(1/(a_R^2)-1) B_out(1,1)*(1/(a_R^2) + 1); b1 =-P*R2*besselk(1,k_out*a_R)/2/N_out/besselk(1,k_out) + P*R1/N_out/2 + P*R1/2/N_in; b2 = (P*R2*Gamma_out/2/N_out)*(besselk(1,k_out*a_R)/besselk(1,k_out)1/a_R); b3 = x_p*AB11_out*(besselk(0,k_out*a_R)+besselk(2,k_out*a_R)) x_p*AB12_out*besselk(1,k_out*a_R) + (P/2/N_out)*(B_out(1,2)+ B_out(1,1) ... +(Gamma_out/(a_R^2))*(A_out(1,1)-A_out(1,2))) + (P/2/N_in)*(B_in(1,1)+B_in(1,2)) Np_out + Np_in + (N_out); b4 = x_p*BD11_out*(besselk(0,k_out*a_R)+besselk(2,k_out*a_R)) x_p*BD12_out*besselk(1,k_out*a_R) + (P/2/N_out)*(D_out(1,2)+D_out(1,1) + ... (Gamma_out/(a_R^2))*(B_out(1,1)-B_out(1,2))) +(P/2/N_in)*(D_in(1,1)+D_in(1,2)) Mp_out + Mp_in + (N_out)*(tp+tsi)/2; A = [A11 A12 A13 A14; A21 A22 A23 A24; A31 A32 A33 A34; A41 A42 A43 A44]; b = [b1 b2 b3 b4]'; c1234 = inv(A)*b; c1=c1234(1); c2=c1234(2); c3=c1234(3); c4=c1234(4); solver1.m % Composite plates; % Computing deflections and Forces at the interface by superposition % Finding Deflection in the central region and the annular region sr=size(r);

PAGE 183

169 w=zeros(sr); jj=floor(R21*num1)+1; for i=1:num1+1 rad1=r(i)*R2; if (i
PAGE 184

170 APPENDIX B DETAILED PROCESS FLOW Energy Harvester Process Traveler Table B-1: Process Traveler Wafer 4" n-type <100> SOI 6 wafers Step # Lab/Equip Process Description 0 Start with SOI wafer with 4000 oxide + 12 um of silicon overlayer. 1.1 Sandia National Labs (SNL) Reduce silicon overlayer using KOH (~1.7 minm ) and HNA (~2.6 minm ) etches. KOH: (Temp ~ 85 C) 3 wafers 320 s etch expected thickness of Si = 2.933 m 2 wafers 270 s etch expected thickness of Si = 4.350 m 3 wafers 212 s etch expected thickness of Si = 5.993 m HNA: Recipe 160 ml acetic acid, 60 ml nitric acid, 20 ml hydrofluoric acid 1 wafer 211 s etch expected thickness of Si = 2.86 m 1 wafer 176 s etch expected thickness of Si = 4.37 m 1.3 SNL Deposit 100 nm Ti Temescal BJD-1800 Evaporator (modified) Conditions: 772.1102.310atmPTorr Ti deposition rate = 3 sec Total thickness = 1000 1.4 SNL Oxidize in tube furnace at 650 C for 30 mins with dry 2O ambient to form 2TiO 2.1 SNL spin positive photoresist (Clariant AZP4330-RS)on back surface (1 m) positive resist and pattern backside alignment marks, (BS_Metal) Mask DARK FIELD evaporated HMDS onto wafer for adhesion spin @ 4000 RPM for 30 s baked @ 90 C for 90 s on hot plate yields 3.75 m thick resist 2.2 SNL Sputter deposit 15 nm of Cr on backside followed by 40 nm of Pt Conditions: 76.310atmPTorr Deposition rate = 3 sec

PAGE 185

171 2.3 SNL Strip resist in Acetone for 2 min to liftoff Cr Ash resist residue in Oxygen plasma stripper 2.4 SNL spin negative photoresist (JSR) on front surface (1 m) negative resist and pattern bottom electrode, (ElectrodeBot) Mask DARK FIELD spin @ 3000 RPM for 30 s bake @ 120 C for 1 min pattern and expose ( Front to back alignment ) using Karl Suss MA6/BA6 Backside Aligner develop in AZ400K for 90 s bake @ 100 C for 45 s descum in Oxygen plasma (PDS/PDE-301) for 5 min 2.5 SNL Evaporate 30 nm of Ti in Temescal BJD-1800 Evaporator for adhesion of Pt to TiO2 @ 3 sec for 100 s 2.6 SNL Evaporate 170 nm of Pt in Temescal BJD-1800 @ 2.5 sec for 680 s 2.7 SNL Strip resist to liftoff Ti/Pt in acetone 3.1 SNL Spin coat 52/48 PZT solution @ 3000 RPM for 30 sec 3.2 SNL Pyrolize @ 300 C for 5 min on a hot-plate in an air ambient 3.3 SNL Repeat spin/pyrolize 4 times to achieve 400 nm thick PZT 3.4 SNL Furnace anneal at 650 C for 30 min 4.1 SNL spin negative photoresist (JSR) on front surface (1 m) negative resist and pattern top electrode, (ElectrodeTop) Mask DARK FIELD spin @ 3000 RPM for 30 s bake @ 120 C for 1 min pattern and expose ( Front to back alignment ) using Karl Suss MA6/BA6 Backside Aligner develop in AZ400K for 90 s bake @ 100 C for 45 s descum in Oxygen plasma (PDS/PDE-301) for 5 min 4.2 SNL Deposit 180 nm Pt in Temescal BJD-1800 @ 2.5 sec for 720 s. 4.3 SNL Strip resist to liftoff Pt in acetone 5.1 Army Research Lab (ARL) Etch PZT in 3:1:1 ammonium biflouride/hydrochloric acid/DI water uses Pt top electrode as etch mask 5.2 ARL Etch PZT residues left behind by previous etch dilute nitric acid/hydrogen peroxide etchant 6.1 UF Spin thick photoresist on backside positive resist and pattern cavity, (Cavity) Mask DARK FIELD 6.2 UF Plasma etch (DRIE) 500 um of silicon from backside

PAGE 186

172 stops on BOX layer (STS MESC Multiplex ICP) 6.3 UF Strip resist in acetone. 6.4 UF Etch oxide (BOE) on backside to remove oxide mask and BOX Graphical Representati on of Process Flow 1000A BOX Figure B-1: Step 1: Deposit Ti on SOI Wafer and oxidize to TiO2. 1800A Figure B-2: Step 2: Deposit Ti/Pla tinum(30nm/170nm) -Lift-Off w/ Mask (ElectrodeBot). 4000A Figure B-3: Step 3: Spin PZT 6 times to achieve desired thickness. 1800A Figure B-4: Step 4: Depos it Platinum(180nm)-Lift-off w/ Mask (ElectrodeTop).

PAGE 187

173 Figure B-5: Step 5: Wet Etch PZT using Pt as etch mask. 500um Figure B-6: Step 6.1: Spin th ick photo resist on bottom (7um). 500um Figure B-7: Step 6.2: Patt ern using mask (Cavity). 500um Figure B-8: Step 6.3: DRIE to BOX 500um Figure B-9: Step 6.4: Ash Resist.

PAGE 188

174 500um Figure B-10: Step 6.5: BOE backside to remove oxide. Mask Layouts Figure B-11: Backside metal mask

PAGE 189

175 Figure B-12: Bottom electrode mask

PAGE 190

176 Figure B-13: Top electrode mask

PAGE 191

177 Figure B-14: Cavity mask

PAGE 192

178 Package Drawings Figure B-15: Detailed schematic drawing of quarter wave resona tor package and mounting plate.

PAGE 193

179 Figure B-16: Detailed schematic of seal ed cavity package and mounting plate.

PAGE 194

180 LIST OF REFERENCES 1. M. Rossi, Acoustics and Electroacoustics, Artech House: Norwood, MA, p. 245308. 1988. 2. D.T. Blackstock, Fundamentals of Physical Acoustics, John Wiley & Sons, Inc.: New York, p. 153-156. 2000. 3. P. Smalser, Power Transfer of Piezoelectric Generated Energy, USA: Ocean Power Technologies, West Trenton, NJ, 1997, #5,703,474. 4. J. Kymissis, C. Kendall, J. Paradiso and N. Gershenfeld. "Parasitic Power Harvesting in Shoes." in 2nd International Symposium on Wearable Computers, Pittsburgh, Pennsylvania p. 132-139, 1998. 5. W.G. Cady, Piezoelectricity, Dover Publications, Inc. : New York, p. 1-8. 1964. 6. N. Setter, ed. Piezoelectric Materials in Devices. N. Setter, EPFL Swiss Federal Institute of Technology: Lausanne. 2002. 7. J. Baborowski, "Microfabric ation of Piezoel ectric MEMS." Journal of Electroceramics. 12: p. 33-51, 2004. 8. M.-A. Dubois and P. Muralt, "Measu rement of the Effective Transverse Piezoelectric Coefficient e31 of AlN and Pb(Zr,Ti)O3 Thin Films." Sensors and Actuators, A: Physical. 77(2): p. 106-112, 1999. 9. D.L. DeVoe, "Piezoelectric Thin Fi lm Micromechanical Beam Resonators." Sensors and Actuators, A: Physical. 88(3): p. 263-272, 2001. 10. D.L. DeVoe and A.P. Pisano, "S urface Micromachined Piezoelectric Accelerometers (Pixls)." Journal of Microelect romechanical Systems. 10(2): p. 180-186, 2001. 11. C.H. Han and E.S. Kim. "Micromach ined Piezoelectric Ultrasonic Transducers Based on Parylene Diaphragm in Silicon Substrate." in IEEE International Ultrasonic Symposium, San Juan, Puerto Rico p. 919-923, 2000.

PAGE 195

181 12. B.L. Lee, K.H. Lee, and B.K. Ju, "C ompositional Dependence of the Properties of Ferroelectric Pb(Zrti)O3 Thin Film Cap acitors Deposited on Single-Layered Ptrho Electrode Barriers." Ceramics International. 30(7): p. 1543-1546, 2004. 13. K.T. Kim, C.I. Kim, and S.G. Lee, "Ferroelectric Properties of Pb(Zr,Ti)O3 Heterolayered Thin Films for Fram Applications." Microelectronic Engineering. 66(1-4): p. 662-669, 2003. 14. G.R. Fox, F. Chu, and T. Davenpor t, "Current and Future Ferroelectric Nonvolatile Memory Technology." Journal of Vacuum Science and Technology, Part B: Microelectronics and Nanometer Structures. 19(5): p. 1967-1971, 2001. 15. P. Verardi, M. Dinescu, and F. Craciun, "Pulsed Laser Deposition and Characterization of PZT Thin Films." Applied Surface Science. 154-155: p. 514518, 2000. 16. R.N. Castellano and L.G. Feinstein, "Ion-Beam Deposition of Thin Films of Ferroelectric Lead Zirconate Titanate (PZT)." Journal of Applied Physics. 50(6): p. 4406-11, 1979. 17. K. Sreenivas, M. Sayer, and P. Garrett "Properties of D.C. Magnetron-Sputtered Lead Zirconate Titanate Thin Films." Thin Solid Films. 172: p. 251-67, 1989. 18. L. Blossfeld, W.L. Larson, L. Kamme rdiner, B.T. Banko, and C.E. Wickersham, "Ferroelectric Pb(Zr,Ti)O3 Thin Films by Reactive Sputtering from a Metallic Target." Vacuum. 41(4-6): p. 1428-30, 1990. 19. M. Huffman, J.P. Goral, M.M. Al-Jassim, A.R. Mason, and K.M. Jones, "Compositional and Microstructural Characterization of Thin Film Lead Zirconate Titanate Ferroelectrics." Thin Solid Films. 193-4(1): p. 1017-22, 1990. 20. Z. Surowiak, M. Loposzko, I.N. Zakharchenko, A.A. Bakirov, E.A. Marchenko, E.V. Sviridov, V.M. Mukhortov, and V. P. Dudkevich, "Thin Ferroelectric Films of the Lead Zirconate-Titanate Type Obtained by R.F. Sputtering." Thin Solid Films. 205: p. 76-84, 1991. 21. Z. Surowiak, D. Czekaj, A.M. Margolin, E.V. Sviridov, V.A. Aleshin, and V.P. Dudkevich, "The Structure and the Piezoelectric Properties of Thin Pb(Zrtiwcd)O3 Films." Thin Solid Films. 214: p. 78-83, 1992. 22. S. Trolier-McKinstry, H. Hu, S.B. Krupanidhi, P. Chindaudom, K. Vedam, and R.E. Newnham, "Spectroscopic Ellipsometry Studies on Ion Beam Sputter Deposited Pb(Zr, Ti)O3 Films on Sapphire and Pt-Coated Silicon Substrates." Thin Solid Films. 230: p. 15-27, 1993.

PAGE 196

182 23. S.-G. Yoon, H.-Y. Lee, and H.-G. Kim, "Compositional Analysis of Lead Titanate Thin Films by Auger Elect ron Spectroscopy and Their Electrical Properties." Thin Solid Films. 171(2): p. 251-62, 1989. 24. T. Kawabata, M. Ikeda, H. Goto, M. Matsumoto, and T. Yada, "2-Dimensional Micro Scanner Integrated with PZT Thin Film Actuator." International Conference on Solid-State Sensors and Actuators, Proceedings. 1: p. 339-342, 1997. 25. X.S. Li, K. Yamashita, T. Tanaka, Y. Suzuki, and M. Okuyama, "Structural and Electrical Properties of Hi ghly Oriented Pb(Zr,Ti)O3 Thin Films Deposited by Facing Target Sputtering." Sensors and Actuators, A: Physical. 82(1): p. 265-269, 2000. 26. T. Kanda, T. Morita, M.K. Kurosawa, and T. Higuchi, "Flat Type Touch Probe Sensor Using PZT Thin Film Vibrator." Sensors and Actuators, A: Physical. 83(1): p. 67-75, 2000. 27. T. Morita, M.K. Kurosawa, and T. Higuchi, "A Cylindri cal Shaped Micro Ultrasonic Motor Utilizing PZT Thin Film (1.4 Mm in Diameter and 5.0 Mm Lon Stator Transducer)." Sensors and Actuators, A: Physical. 83: p. 225-230, 2000. 28. T. Morita, Y. Wagatsuma, H. Moriok a, and H. Funakubo, "Ferrelectric Property of an Epitaxial Lead Zirconate Titanate Thin Film Deposited by a Hydrothermal Method." Journal of the Material Research Society. 19(6): p. 1862-1868, 2004. 29. T. Morita and Y. Cho, "A Hydrotherm ally Deposited Expitaxial Lead Titanate Thin Film on Strontium Ruthenium Oxide Bottom Electrode." Applied Physics Letters. 85(12): p. 2331-2333, 2004. 30. T. Morita, M. Kurosawa, and T. Higuc hi, "Cylindrical Micr o Ultrasonic Motor Fabricated by Improved Hydrothermal Method." International Conference on Solid-State Sensors and Actuators, Proceedings. 1: p. 49-52, 1997. 31. D.F. Bahr, J.C. Merlino, P. Bane rjee, C.M. Yip, and A. Bandyopadhyay, "Reliability and Properties of PZT Th in Films for MEMS Applications." Materials Research Society Symposium Proceedings. 546: p. 153-158, 1999. 32. J. Bernstein, K. Houston, L. Niles, S. Finberg, H. Chen, L.E. Cross, K. Li, and K. Udayakumar, "Micromachined Ferroelectric Transducers for Acoustic Imaging." International Conference on Solid-State Sensors and Actuators, Proceedings. 1: p. 421-424, 1997. 33. B. Xu, L.E. Cross, and J. Bernstein, "Ferroelectric and Antifer roelectric Films for Microelectromechanical Systems Applications." Thin Solid Films. 377-378: p. 712-718, 2000.

PAGE 197

183 34. T. Itoh, J. Chu, I. Misumi, K. Ka taoka, and T. Suga, "New Dynamic Scanning Force Microscopes Using Piezoelectric PZT Microcantilevers." International Conference on Solid-State Sensors and Actuators, Proceedings. 1: p. 459-462, 1997. 35. C. Lee, T. Itoh, R. Maeda, and T. Suga, "Characterization of Micromachined Piezoelectric PZT Force Sensors for D ynamic Scanning Force Microscopy." Review of Scientific Instruments. 68(5): p. 2091-2100, 1997. 36. J.M. Jungk, B.T. Crozier, A. Ba ndyopadhyay, N.R. Moody, and D.F. Bahr, "Mechanical Properties and Adhesion of PZT Thin Films for MEMS." Materials Research Society Symposium Proceedings. 594: p. 225-230, 2000. 37. K. Kunz, P. Enoksson, and G. Stem me, "Highly Sensitive Triaxial Silicon Accelerometer with Integrated PZT Thin Film Detectors." Sensors and Actuators, A: Physical. 92(1-3): p. 156-160, 2001. 38. B. Piekarski, D.L. Devoe, R. Kaul and M. Dubey. "Design, Modeling, and Testing of Micromachined Piezoelectric Clamped-Clamped Beam Resonators." in Microelectromechanical Systems (MEMS) ASME 2000, 2000. 39. B. Piekarski, D. DeVoe, M. Dubey, R. Kaul, and J. Conrad, "Surface Micromachined Piezoelectric Resonant Beam Filters." Sensors and Actuators, A: Physical. 90(3): p. 313-320, 2001. 40. J.S. Pulskamp, A. Wickenden, R. Po lcawich, B. Piekarski, and M. Dubey, "Mitigation of Residual Film St ress Deformation in Multilayer Microelectromechanical Syst ems Cantilever Devices." Journal of Vacuum Science and Technology, Part B: Microe lectronics and Nanometer Structures. 21(6): p. 2482-2486, 2003. 41. A.T. Ferguson, L. Li, V.T. Nagaraj, B. Balachandran, B. Piekarski, and D.L. DeVoe, "Modeling and Design of Compos ite Free-Free Beam Piezoelectric Resonators." Sensors and Actuators, A: Physical. 118: p. 63-69, 2005. 42. Y. Yee, H.-J. Nam, S.-H. Lee, J.U. Bu, and J.-W. Lee, "PZT Actuated Micromirror for Fine-Tracking Mechan ism of High-Density Optical Data Storage." Sensors and Actuators, A: Physical. 89(1-2): p. 166-173, 2001. 43. J.-C. Yu and C.-B. Lan, "System Modeling of Microaccelerometer Using Piezoelectric Thin Films." Sensors and Actuators, A: Physical. 88(2): p. 178-186, 2001. 44. E. Zakar, M. Dubey, B. Piekarski, J. Conrad, R. Piekarz, and R. Widuta, "Process and Fabrication of a Lead Zirconate T itanate Thin Film Pressure Sensor." Journal

PAGE 198

184 of Vacuum Science and Technology, Pa rt A: Vacuum, Surfaces and Films. 19(1): p. 345-348, 2001. 45. E. Zakar, M. Dubey, R. Polcawich, B. Piekarski, R. Piekarz, J. Conrad, and R. Widuta, "Study of PZT Film Stress in Mu ltilayer Structures for MEMS Devices." Materials Research Society Symposium Proceedings. 605: p. 287-292, 2000. 46. L.M.R. Eakins, B.W. Olson, C.D. Ri chards, R.F. Richards, and D.F. Bahr, "Influence of Structure and Chemistr y on Piezoelectric Properties of Lead Zirconate Titanate in a Microelecr omechanical Systems Power Generation Application." Journal of the Material Research Society. 18(9): p. 2079-2086, 2003. 47. S. Whalen, M. Thompson, D.F. Bahr, C. D. Richards, and R.F. Richards, "Design, Fabrication and Testing of th e P3 Micro Heat Engine." 2003. 104: p. 290-298, 2003. 48. S. Zurn, M. Hsieh, G. Smith, D. Markus, M. Zang, G. Hughes, Y. Nam, M. Arik, and D. Polla, "Fabrication and Struct ural Characterization of a Resonant Frequency PZT Microcantilever." Smart Materials and Structures. 10(2): p. 252263, 2001. 49. T. Shiosaki and A. Kawabata, "Piezoel ectric Thin Films for Saw Applications." Ferroelectrics. 42(1-4): p. 219-232, 1982. 50. M. Royer, J.O. Holmen, M.A. Wurm, a nd O.S. Aadland, "Zno on Si Integrated Acoustic Sensor." Sensors and Actuators, A: Physical. 4: p. 357-362, 1983. 51. H.N. Al-Shareef, K.R. Bellur, O. Auciello, X. Chen, and A.I. Kingon, "Effect of Composition and Annealing Conditions on the Electrical Properties of Pb(Zrti)O3 Thin Films Deposited by the Sol-Gel Process." Thin Solid Films. 252: p. 38-43, 1994. 52. P.F. Indermuhle, G. Schurmann, G.A. Raci ne, and N.F. de Rooij, "Tip Integration on Arrays of Silicon Micro Cantilevers with Self Exciting Piezoelectric Sensor for Parallel Atomic Force Mi croscopy Applications." International Conference on Solid-State Sensors and Actuators, Proceedings. v 1: p. 451-453, 1997. 53. S. Butterworth, Proceedings of the Physical Society. 27: p. 410-24, 1915. 54. K.S. Van Dyke, "The Electrical Ne twork Equivalent of a Piezoelectric Resonator." Physical Review. 40: p. 1026, 1932. 55. K.S. Van Dyke, "The Electric Network E quivalent of a Piezoelectric Resonator." Physical Review. 25: p. 895, 1925.

PAGE 199

185 56. K.S. Van Dyke, "The Piezoelectric Resonator and Its Equivalent Network." Institute of Radio Engineers (IRE) Proc. 16: p. 742-64, 1928. 57. D.W. Dye, "Piezoelectric Quartz Res onator and Equivalent Electrical Circuit." Proceedings of the Physical Society. 38: p. 399-458, 1926. 58. W.P. Mason, "An Electromechanical Repr esentation of Piezoelectric Crystal Used as a Transducer." Institute of Radio Engineers (IRE) Proc. 23: p. 1252-1263, 1935. 59. W.P. Mason, Piezoelectric Crystals and Th eir Application to Ultrasonics. The Bell Telephone Laboratories Series, D. Van Nostrand Company, Inc.: New York. 1950. 60. F.A. Fischer, Fundamentals of Electroacoustics, Interscience Publishers, Inc.: New York, p. 21-22. 1955. 61. Y.S. Cho, Y.E. Pak, H.M. Jeong, and S.K. Ha, "Five-Port Equivalent Electric Circuit of Piezoelectric Bimorph Beam." Proceedings of SPIE The International Society for Optical Engineering. 3673: p. 261-277, 1999. 62. G. Wang, Modeling and Design of a MEMS-B ased Piezoelectric Microphone, in Mechanical and Aerospace Engineering. 2003, University of Florida: Gainesville, FL. Last accessed 63. G. Wang, B.V. Sankar, L. Cattafesta, and M. Sheplak. "Analysis of a Composite Piezoelectric Circular Plate with Initial Stresses for MEMS,." in International Mechanical Engineering Congress & Exposition, New Orleans, LA: ASME Paper IMECE2002-39337, 2002. 64. S.A.N. Prasad, B.V. Sankar, L.N. Cattafesta, S.B. Horowitz, Q. Gallas, and M. Sheplak. "Two-Port Electroacoustic Model of an Axisymmetric Piezoelectric Composite Plate." in 43rd AIAA/ASME/ASCE/AHS/AS C Structures, Structural Dynamics, and Materials Conference, Denver, CO, 2002. 65. Y.S. Cho, Y.E. Pak, C.S. Han, and S.K. Ha, "Five-Port Equivalent Electric Circuit of Piezoelectric Bimorph Beam." Sensors and Actuators, A: Physical. 84(1): p. 140-148, 2000. 66. C. Liang, Sun, Fanping, Rogers, Cr aig A., "Electro-Mechanical Impedance Modeling of Active Material Systems." Smart Materials and Structures. 5: p. 171-186, 1996. 67. C. Liang, Sun, F.P., Rogers, C.A., "An Impedance Method for Dynamic Analysis of Active Material Systems." Journal of Vibration and Acoustics. 116: p. 120128, 1994.

PAGE 200

186 68. G.E. Martin, "Determination of Equiva lent-Circuit Constants of Piezoelectric Resonators of Moderately Low Q by Absolute-Admittance Measurements." The Journal of the Acoustical Society of America. 26(3): p. 413-420, 1954. 69. S. Sherrit, H.D. Wiederick, B.K. Mukhe rjee, and M. Sayer, "Accurate Equivalent Circuit for the Unloaded Piezoelectri c Vibrator in the Thickness Mode." Journal of Physics D: Applied Physics. 30(16): p. 2354-2363, 1997. 70. S. Sherrit, Wiederick, H.D., Mukherjee, B.K. "Accurate Equivalent Circuits for Unloaded Piezoelectric Resontators." in IEEE Ultrasonics Symposium, Toronto, Can.: IEEE p. 931-935, 1997. 71. S. Lin, "Analysis of the Equivalent Circuit of Piezoelectric Ceramic Disk Resonators in Coupled Vibration." Journal of Sound and Vibration. 231(2): p. 277-290, 2000. 72. D.P. Chen, S. Jen, and C.S. Hartmann, "Resonant Modes in Coupled Resonator Filters and the Unique Equivalent Circuit Representation." Proceedings of the IEEE Ultrasonics Symposium. 1: p. 59-62, 1995. 73. Q. Meng, M. Mehregany, and K. Deng, "Modeling of the Electromechanical Performance of Piezoelectric Laminated Microactuators." Journal of Micromechanics and Microengineering. 3(1): p. 18-23, 1993. 74. P. Seshu and N.G. Naganathan, "Analy sis of Strain Induced in Piezoelectric Actuated Beams." Proceedings of SPIE The International Society for Optical Engineering. 2715: p. 658-669, 1996. 75. H.T. Banks, Smith, R.C., Wang, Yun, "The Modeling of Piezoceramic Patch Interactions with Shells, Plates, and Beams." Quarterly of Applied Mathematics. 53(2): p. 353-381, 1995. 76. S.H. Chang and B.C. Du, "Optimi zation of Asymmetric Bimorphic Disk Transducers." The Journal of the Acoustical Society of America. 109(1): p. 194202, 2001. 77. C.C. Cheng and P.W. Wang, "Applica tions of the Impedance Method on Multiple Piezoelectric Actuators Driven Structures." Journal of Vibration and Acoustics. 123: p. 262-268, 2001. 78. J.L. Delany, "Bender Transducer Design and Operation." The Journal of the Acoustical Society of America. 109(2): p. 554-562, 2000. 79. E.M. Flint, Rogers, Craig A., Lia ng, Chen, "Electro-Dynamic Transduction Equations for Piezoelectric Stack Actuators." Adaptive Structures and Composite

PAGE 201

187 Materials: Analysis and Application Amer ican Society of Mechanical Engineers, Proceedings of the 1994 International Mechanical Engineering Congress and Exposition Chicago, IL. AD-45: p. 201-210, 1994. 80. V. Giurgiutiu, Rogers, Craig A. "N ew Results in the Application of E/M Impedance Method to Machinery Health Monitoring and Failure Prevention." in 53rd Meeting of the Society for Machinery Failure Prevention Technology, Virginia Beach, VA, 1999. 81. M. Goldfarb, Celanovic, Nikola, "A Lu mped Parameter Electromechanical Model for Describing the Nonlinear Behavi or of Piezoelectric Actuators." ASME Journal of Dynamic Systems, Measurement, and Control. 119(3): p. 478-485, 1997. 82. C. Hu, Kawiecki, Grzegorz, Chandra, Ramesh, Chopra, Inderjit. "Structural Modeling of Beams with Skewed Piezoactuators." in 36th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference and AIAA/ASME Adaptive Structures Forum, New Orleans, LA: AIAA p. 3423-3430, 1995. 83. F. Lalande, Chaudhry Zaffir, Rogers, Cr aig A., "Impedance Modeling of in-Phase Actuation of Actuators Bonded on Ring Structures." Adaptive Structures and Composite Materials: Analysis and App lication American Society of Mechanical Engineers, Proceedings of the 1994 In ternational Mechanical Engineering Congress and Exposition Chicago, IL. AD-45: p. 193-200, 1994. 84. F. Lalande, C. Zaffir, and C.A. Roge rs, "Impedance-Based Modeling of Induced Strain Actuators Bonde d on Ring Structures." Journal of Sound and Vibration. 201(2): p. 169-187, 1997. 85. C.K. Lee, "Theory of Laminated Piezoelec tric Plates for the Design of Distributed Sensors/Actuators. Part I: Governing E quations and Reciprocal Relationships." Journal of the Acoustical Society of America. 87(3): p. 1144-1158, 1990. 86. M. Leibowitz, Vinson, Jack R. "The Use of Hamilton's Principle in Laminated Piezoelectric and Composite Structures." in ASME Winter Annual Meeting, New Orleans, LA: ASME p. 257-267, 1993. 87. G.A. Lesieutre and C.L. Davis, "Can a Coupling Coefficient of a Piezoelectric Device Be Higher Than Those of Its Active Material?" Proceedings of SPIE The International Society for Optical Engineering. 3041: p. 281-292, 1997. 88. S.E. Miller and H. Abramovich, "Self-Sensing Piezolaminated Actuator Model for Shells Using a First Order Shear Deformation Theory." Journal of Intelligent Material Systems and Structures. 6(5): p. 624-638, 1995.

PAGE 202

188 89. A. Sermoneta, Liang, C., Rogers, C.A. "Investigation of the Electro-Mechanical Coupling in Active Structures with Multiple Actuators Determination of Optimal Actuator Locations." in 36th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference and AIAA/ASME Adaptive Structures Forum, New Orleans, LA: AIAA p. 3524-3528, 1995. 90. H.A.C. Tilmans, "Equivalent Circui t Representation of Electromechanical Transducers: I Lumped-Parameter Systems." Journal of Micromechanics and Microengeneering. 6: p. 157-176, 1996. 91. R.H.M. van de Leur, "Critical Consid eration on the Interpre tation of Impedance Plots." Journal of Physics D: Applied Physics. 24(8): p. 1430-1435, 1991. 92. K. Yao and K. Uchino, "Analysis on a Composite Cantilever Beam Coupling a Piezoelectric Bimorph to an Elastic Blade." Sensors and Actuators, A: Physical. 89(3): p. 215-221, 2001. 93. S. Zhou, Liang, C., Rogers, Craig A. "Impedance Modeling of Two-Dimensional Piezoelectric Actuators Bonded on a Cylinder." in ASME Winter Annual Meeting, New Orleans, LA: ASME p. 247-255, 1993. 94. S.-W. Zhou, Liang, Chen, Rogers, Craig A., "Modeling of Distributed Piezoelectric Actuators Integrated with Thin Cylindrical Shells." Journal of the Acoustical Society of America. 96(3): p. 1605-1612, 1994. 95. S.Y. Lin, "Coupled Vibration Anal ysis of Piezoelectric Ceramic Disk Resonators." Journal of Sound and Vibration. 218(2): p. 205-217, 1998. 96. X. Li, R. Lin, and K.W. Leow, "Performance-Enhanced Micro-Machined Resonant Systems with Two-Degrees-of-Freedom Resonators." Journal of Micromechanics and Microengineering. 10(4): p. 534-539, 2000. 97. N.W. Hagood, von Flotow, A., "Dampi ng of Structural Vibrations with Piezoelectric Materials and Pa ssive Electrical Networks." Journal of Sound and Vibration. 146(2): p. 243-268, 1991. 98. J.J. Hollkamp, "Multimodal Passive Vibration Suppression with Piezoelectric Materials and Resonant Shunts." Journal of Intelligen t Material Systems and Structures. 5(1): p. 49-57, 1994. 99. K.W. Wang, Lai, J.S., Yu, W.K., "Struc tural Vibration Control Via Piezoelectric Materials with Real-Time Semi-Active Electrical Networks." Adaptive Structures and Composite Materials: Analysis a nd Application American Society of Mechanical Engineers, Proceedings of the 1994 International Mechanical Engineering Congress and Exposition Chicago, IL. AD-45: p. 219-226, 1994.

PAGE 203

189 100. D.J. Leo, Inman, Daniel J., "Pointin g Control and Vibrati on Suppression of a Slewing Flexible Frame." Journal of Guidance, Control, and Dynamics. 17(3): p. 529-536, 1994. 101. D. Sciulli and D.J. Inman, "Comparis on of Singleand Two-Degree-of-Freedom Models for Passive and Active Vibration Isolation Design." Proceedings of SPIE The International Societ y for Optical Engineering. 2720: p. 293-304, 1996. 102. C.L. Davis, Lesieutre, George A., Dosch, Jeffrey. "A Tunable Electrically Shunted Piezoceramic Vi bration Absorber." in SPIE Smart Structtures and Materials Conference p. 51-59, 1997. 103. J. Hwang, T. Ueno, Y. Tomabechi, and K. Matsumara, "Variation of Resonant Frequency of Disk Resonator C oupled with External Circuit." Electronics & Communications in Japan, Part II: Elect ronics (English translation of Denshi Tsushin Gakkai Ronbunshi). 80(11): p. 48-57, 1997. 104. G.A. Lesieutre, "Vibration Damping and Control Using Shunted Piezoelectric Materials." The Shock and Vibration Digest. 30(3): p. 187-195, 1998. 105. C.J. Cross, Fleeter, Sanford, "Passive Control of Flow Induc ed Vibrations Using Shunted Piezoelectrics." AIAA-98-1952/A98-25205: p. 2021-2033, 1998. 106. A. Curtis, Piezoelectric Materials for Ener gy Transfer Vibration Control, Materials Systems Inc. 1-19. 1999, ONR BAA 99-023 107. C.L. Davis, Lesieutre, George A., "An Actively Tuned Solid State Piezoelectric Vibration Absorber." Proceedings of SPIE The International Society for Optical Engineering. 3327: p. 169-182, 1998. 108. J. Kim, Y.-H. Ryu, and S.-B. Choi, "New Shunting Parameter Tuning Method for Piezoelectric Damping Based on Meas ured Electrical Impedance." Smart Materials and Structures. 9(6): p. 868-877, 2000. 109. G. Caruso, "A Critical Analysis of Electric Shunt Circuits Employed in Piezoelectric Passive Vibration Damping." Smart Materials and Structures. 10: p. 1059-1068, 2001. 110. S.-y. Wu, T.L. Turner, and S.A. Rizzi, "Piezoelectric Shunt Vibration Damping of F-15 Panel under High Acoustic Excitation." Proceedings of SPIE The International Society for Optical Engineering. 3989: p. 276-287, 2000. 111. M. Goldfarb and L.D. Jones, "On the Efficiency of Electric Power Generation with Piezoelectric Ceramic." ASME Journal of Dynamic Systems, Measurement, and Control. 121(3): p. 566-571, 1999.

PAGE 204

190 112. V. Giurgiutiu, Rogers, Craig A., "Dyn amic Power and Energy Capabilities of Commercially-Available Electro-A ctive Induced-Strain Actuators." Journal of Intelligent Material Systems and Structures. 7(November): p. 656-667, 1996. 113. V. Giurgiutiu, Rogers, Craig A., "Power and Energy Characteristics of Solid-State Induced-Strain Actuators for Static and Dynamic Applications." Journal of Intelligent Material Systems and Structures. 8(September): p. 738-750, 1997. 114. R.A. Lomenzo, Sumali, Hartono, Cudne y, Harley H. "Maximizing Mechanical Power Transfer from Piezoceramic Stacked Actuators to Structures." in ASME Winter Annual Meeting, New Orleans, LA: ASME p. 229-235, 1993. 115. S. Meninger, Mur-Miranda, Jose Oscar, Amirtharajah, Rajeevan, Chandrakasan, Anantha, Lang, Jeffrey, Vibration-to-Electric Energy Conversion, MIT 116. S.C. Stein, Liang, Chen, Rogers, Crai g A., "Power Consump tion of Piezoelectric Actuators Driving a Simply Supported Beam Considering Fluid Coupling." Journal of the Acoustical Society of America. 96(3): p. 1598-1604, 1994. 117. S.-W. Zhou, Rogers, Craig A., "Power Flow and Consumption in Piezoelectrically Actuated Structures." AIAA Journal. 33(7): p. 1305-1311, 1995. 118. R. Taylor, F. Liu, S.B. Horowitz, K. Ngo, T. Nishida, L. Cattafesta, and M. Sheplak. "Technology Development for Elect romechanical Acoustic Liners." in Active 04, Williamsburg, VA, Paper #A04-093, 2004. 119. A. Kasyap, S.B. Horowitz, J.-S. Lim, K. Ngo, A. Kurdila, T. Nishida, M. Sheplak, and L.N. Cattafesta. "Energy Reclam ation from a Vibrating Piezoceramic Composite Beam." in 9th International Congr ess on Sound and Vibration (ICSV9), Orlando, FL, 2002. 120. S.B. Horowitz, A. Kasyap, F. Liu, D. Johnson, T. Nishida, K. Ngo, M. Sheplak, and L.N. Cattafesta. "Technology Development for Self-Powered Sensors." in 1st AIAA Flow Control Conference, St. Louis, MO, 2002. 121. S. Roundy, P.K. Wright, and J. Rabaey "A Study of Low Level Vibrations as a Power Source for Wireless Sensor Nodes." Computer Communications. 26: p. 1131-1144, 2003. 122. A. Mehra, X. Zhang, A.A. Ayon, I.A. Waitz, M.A. Schmidt, and C.M. Spadaccini, "A Six-Wafer Combustion Syst em for a Silicon Micro Gas Turbine Engine." Journal of Microelect romechanical Systems. 9(4): p. 517-526, 2000. 123. R.K. Wangsness, Electromagnetic Fields. 2nd ed, John Wiley & Sons, Inc.: Hoboken, NJ. 587. 1986.

PAGE 205

191 124. J.D. Irwin. "Basic Engineering Circu it Analysis." in, Upper Saddle River, NJ: Prentice-Hall, Inc. p. 466-7, 1996. 125. F.V. Hunt, Electroacoustics: The Analysis of Transduction, and Its Historical Background, Acoustical Society of America: New York. 1982. 126. J.F. Lindsay and S. Katz, Dynamics of Physical Circuits and Systems, Matrix Publishers, Inc.: Champaign, Illinois. 1978. 127. H.A.C. Tilmans, "Equivalent Circui t Representation of Electromechanical Transducers: Ii Distributed-Parameter Systems." Journal of Micromechanics and Microengineering: p. 285-309, 1997. 128. S.D. Senturia, Microsystem Design, Kluwer Academic Publishers: Boston, MA. 689. 2001. 129. A.P. Dowling and J.E. FFowcs Williams, Sound and Sources of Sound. Ellis Horwood Series in Engineeri ng Science, Ellis Horwood Limited: Chichester, Eng. 1983. 130. A.W. Nolle, "Small-Signal Impedance of Short Tubes." The Journal of the Acoustical Society of America. 25(1): p. 32-39, 1953. 131. A.S. Hersh and B. Walker, Fluid Mechanical Model of the Helmholtz Resonator, NASA Contractor Repor t. 1977, NASA CR-2904 132. R.C. Rosenberg and D.C. Karnopp, Introduction to Physical System Dynamics, McGraw-Hill Book Company: New York, N.Y. 1983. 133. IEEE, IEEE Standard on Piezoelectricity, IEEE. 1987, ANSI/IEEE Std 176-1987 134. Y. Xu, Ferroelectric Material s and Their Applications, North-Holland: Amsterdam. 391. 1991. 135. J.F. Shackelford and W. Alexander, eds. Crc Materials Science and Engineering Handbook. 3 ed. CRC Press: Boca Raton, FL. 2001. 136. P. Hess, "Laser Diagnostics of Mechan ical and Elastic Propertis of Silicon and Carbon Films." Applied Surface Science. 106: p. 429-437, 1996. 137. J.B. Price. "Anisotropic Etching of S ilicon with Koh -H20-Is opropyl Alcohol." in Semicond Silicon 1973, 2nd Int Symp on S ilicon Material Sci and Technology, Chicago, IL p. 339-353, 1973.

PAGE 206

192 138. D. Resnik, D. Vrtacnik, U. Aljancic, M. Mozek, and S. Amon, "Different Aspect Ratio Pyramidal Tips Obtained by Wet Etching of (100) and (111) Silicon." Microelectronics Journal. 34: p. 591-593, 2003. 139. B. Tuttle, J.A. Voigt, T.J. Headley, B.G. Potter, D. Dimos, R.W. Schwartz, M.T. Dugger, J. Michael, R.D. Nasby, T.J. Garino, and D.C. Goodnow, "Ferrolectric Thin Film Microstructure Development and Related Property Enhancement." Ferroelectrics. 151: p. 11-20, 1994. 140. L.I. Vergara, M.C.G. Passeggi Jr., and J. Ferron, "The Role of Passivation in Titanium Oxidation: Thin Film and Temperature Effects." Applied Surface Science. 187: p. 199-206, 2002. 141. M. Harting and G. Fritsch, "Struc tural Analysis and Residual Stress Measurements in a Natural Titanium Oxide Layer." Materials Science Forum. 166-169: p. 343-348, 1994. 142. S. Norasetthekul, P.Y. Park, K.H. Baik, K.P. Lee, J.H. Shin, B.S. Jeong, V. Shishodia, L. E.S., D.P. Norton, and S.J. Pearton, "Dry Etch Chemistries for Tio2 Thin Films." Applied Surface Science. 185: p. 27-33, 2001. 143. K. Yao, S. Yu, and F. Eng-Hock Tay, "Residual Stress Analysis in Ferroelectric Pb(Zrti)O3 Thin Films Fabricated by a Sol-Gel Process." Applied Physics Letters. 82(25): p. 4540-4542, 2003. 144. D. Xia, M. Liu, Y. Zeng, and C. Li, "F abrication and Electrical Properties of Lead Zirconate Titanate Thick Films by the New Sol-Gel Method." Materials Science and Engineering. B87: p. 160-163, 2001. 145. G.G. Stoney, "The Tension of Thin Metallic Films Depos ited by Electrolysis." Proceedings of the Royal Society of London. A 82: p. 172, 1909. 146. X. Zhang, K.S. Chen, R. Ghodssi, A. A. Ayon, and S.M. Spearing, "Residual Stress and Fracture in Thick Tetraethylor thosilicate (Teos) and Silane-Based Pecvd Oxide Films." Sensors and Actuators, A: Physical. 91: p. 373-380, 2001. 147. J.S. Kim and K.W. Paik, "The Multilayer-Modified Stoney's Formula for Laminated Polymer Composites on a Silicon Substrate." Journal of Applied Physics. 86(10): p. 5474-5479, 1999. 148. L. Fu, J.M. Miao, X.X. Li, and R.M. Lin, "Study of Deep Silicon Etching for Micro-Gyroscope Fabrication." Applied Surface Science. 177: p. 78-84, 2001. 149. K.S. Chen and A.A. Ayon, "Effect of Process Parameters on the Surface Morphology and Mechanical Performan ce of Silicon Structures after Deep

PAGE 207

193 Reactive Ion Etching (Drie)." Journal of Microelectromechanical Systems. 11(3): p. 264-275, 2002. 150. A.A. Ayon, S. Nagle, L. Frechette, A. Epstein, and M.A. Schmidt, "Tailoring Etch Directionality in a Deep Reactive Ion Etching Tool." Journal of Vacuum Science and Technology, Part B: Microe lectronics and Nanometer Structures. 18(3): p. 1412-1416, 2000. 151. F. Laermer and A. Urban, "Challenges Development and Applications of Silicon Deep Reactive Ion Etching." Microelectronic Engineering. 67-68: p. 349-355, 2003. 152. A.M. Hynes, H. Ashraf, J.K. Bhardw aj, J. Hopkins, I. Johnston, and J.N. Shepherd, "Recent Advances in Silicon Etching for MEMS Using the Ase Process." Sensors and Actuators, A: Physical. 74: p. 13-17, 1999. 153. H. Jansen, M. de Boer, J. Burger, R. Legtenberg, and M. Elwenspoek, "The Black Silicon Method Ii: The Effect of Mask Ma terial and Loading on the Reactive Ion Etching of Deep Silicon Trenches." Microelectronic Engineering. 27: p. 475-480, 1995. 154. Olympus America Inc., www.Olympusamerica.com. Microscopes & Imaging Systems, 1-800-446-5967. Last accessed June 2005 155. Agilent Technologies Inc., www.Agilent.com. Test & Measurement Center, 1 (800) 829-4444. Last accessed December 2004 156. Polytec PI Inc., www.Polytecpi.com. (508)832-3456. Last accessed August 2005 157. Bruel and Kjaer Inc., www.Bkhome.com. (800) 332-2040. Last accessed May 2005 158. AE Techron Inc., www.Aetechron.com. (574)295-9495. Last accessed December 2004 159. M.G. Jones and P.E. Stiede, "Comparison of Methods for Determining Specific Acoustic Impedance." Journal of the Acoustical Society of America. 101(5 Pt. 1): p. 2694-2704, 1997. 160. M.G. Jones and T.L. Parrott, "Eva luation of a Multi -Point Method for Determining Acoustic Impedance." Mechanical Systems and Signal Processing. 3(1): p. 15-35, 1989. 161. W.T. Chu, "Extension of the TwoMicrophone Transfer Function Method for Impedance Tube Measurements." Journal of the Acoustical Society of America. 80(1): p. 347-349, 1986.

PAGE 208

194 162. J.Y. Chung and D.A. Blaser, "Transfe r Function Method of Measuring in-Duct Acoustic Properties. I. Theory." Journal of the Acoustical Society of America. 68(3): p. 907-921, 1980. 163. ASTM, Standard Test Method for Impedance and Absorption of Acoustical Materials Using a Tube, Two Microphone s, and a Digital Frequency Analysis System, ASTM. 871-876. 1990, E 1050-90 164. ASTM, Standard Test Method for Impedance and Absorption of Acoustical Materials by the Impedance Tube Method, ASTM. 113-121. 1995, C 384-95 165. T. Schultz, L. Cattafesta, T. Nishida, and M. Sheplak. "Uncertainty Analysis of the Two-Microphone Method." in 8th AIAA/CEAS Aeroacoustics Conference, Breckenridge, CO, 2002. 166. J.S. Bendat and A.G. Piersol, Random Data: Analysis and Measurement Procedures. 3 ed, John Wiley & Sons, Inc.: New York. 594. 2000. 167. B. Tuttle and R.W. Schwartz, "Solution Deposition of Ferroelectric Thin Films." MRS Bulletin. 21(6)1996.

PAGE 209

195 BIOGRAPHICAL SKETCH Stephen Brian Horowitz was born on August 3, 1977, in Pompton Plains, N.J. He attended Marjory Stoneman Douglas High School in Parkland, FL, graduating in 1995. He received his bachelors degr ee in electrical engineering from the University of Florida in 1999 and his Master of Science degree in electrical engineeri ng in 2001. He is completing his doctoral degree at the University of Florida, concen trating his research efforts in the area of microelectromechanic al systems (MEMS) and energy harvesting.


Permanent Link: http://ufdc.ufl.edu/UFE0008378/00001

Material Information

Title: Development of a MEMS-Based Acoustic Energy Harvester
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: UFE0008378:00001

Permanent Link: http://ufdc.ufl.edu/UFE0008378/00001

Material Information

Title: Development of a MEMS-Based Acoustic Energy Harvester
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: UFE0008378:00001


This item has the following downloads:


Full Text












DEVELOPMENT OF A MEMS-BASED ACOUSTIC ENERGY HARVESTER


BY

STEPHEN BRIAN HOROWITZ













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

Stephen Brian Horowitz


































I dedicate this dissertation to my wife, Megan. You're my sweetheart, my wife and my
best friend. I love you more than words can say.
















TABLE OF CONTENTS
Page

L IS T O F T A B L E S .............................................................. .......................................... v ii

L IST O F FIG U R E S .............. ............................ ............. ........... ........... viii

ABSTRACT ........ .............. ............. ...... ...................... xiv

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

Acoustic Energy Harvester- Concept....................................................... ................
M otiv action ............................................. 5
B ack g rou n d ................................................................................................. . .6
Piezoelectricity ................................................ ........6
Piezoelectric M materials in M EM S ........................................ ...... ............... 6
M materials and properties ....................................................... .... ........... 7
D position m methods .......................................................................... .... .8
Piezoelectric integration w ith M EM S ........................................ .................9
Reduced-Order Modeling of Piezoelectric Materials.........................................11
Energy Absorption/Vibration Damping and Energy Harvesting ......................13
A p p ro a c h ....................................................................... 14

2 THEORETICAL BACKGROUND ........................................ ....................... 17

Overview Energy H harvesting ............................................................................ 17
Maximum Average Power Transfer .............. ...........................................20
Electrom echanical Transducers................................ ................................... 23
Lum ped Elem ent M odeling ......................................................... ............... 25
H elm holtz R esonator .......................................... ................... ........ 29
Lum ped Elem ent M odel ............................................. ............................. 29
E qu iv alent C ircu it.......... ............................................................... .... .... .... .. 33
Piezoelectric Com posite Plate ............................................................................. 35
P iezoelectricity .....................................................................36
Com posite Plate M odeling ............................................................................ 41
Lum ped elem ent m odel ........................................ .......................... 41
E equivalent circuit ............................... ........ ... .... .. ........ .... 46
Case 1: Well below the short-circuit mechanical resonant frequency .........50
Case 2: At the short-circuit mechanical resonant frequency.....................52
General expression vs. simplified cases.................................................54









Effects of including the radiation impedance.............................................55
Equivalent circuit parameters of piezoelectric composite circular plates ....61
Acoustic Energy Harvester Dynamic Behavior .....................................................68
A coustical Input B ehavior.............................................................. ... ............ 68
E electrical O utput B behavior ........................................... ................. ............... 71
Operation at an Optim al Frequency ....................................... ............... 76
D vice C configurations .............................................................................. 79

3 FABRICATION AND PACKAGING .............................................. ...............82

O v e rv ie w ............................................................................................................... 8 2
P process F low .......................................................82
P a c k a g in g .............................................................................................................. 9 2
Packaging Scheme ................................. .......................... ..........92
Quarter-w ave resonator package ............................................................... 92
Sealed cavity package ................. ................................94
Chip to Package M counting ............................................................. ............ 94
P o lin g .......................................................9 6

4 EXPERIMENTAL SETUP ........................................ ........ .........98

Parameters and How They are Obtained ........................... ... ...............98
E xperim ental Setup D details ................................................................................ 10 1
Ferroelectric M easurem ents ....................................................................... 101
Electrically Actuated Response-Laser Scanning Vibrometer ......................102
Acoustic Characterization Plane Wave Tube .................. ..............................103
Electrical Im pedance Characterization.............................. .............................107
Initial Deflection Measurements Wyko Optical Profilometer......................107

5 EXPERIMENTAL RESULTS AND DISCUSSION............... .............108

Ferroelectric M easurem ents ......................................................................................108
Electrical M easurem ents .......................................................... ............... 110
Electrical Output Im pedance ....................................... ............ ..................110
Electrically Actuated Response-Laser Scanning Vibrometer ......................114
Frequency response ....................................................................... 114
L in e arity ...............................................................12 4
A acoustical M easurem ents .. ......................................................... ................... 126
Acoustically Actuated Frequency Response Plane-Wave Tube ....................126
Sealed cavity package ............................................... ............... 126
Quarter-w ave resonator package .............. ..... .....................................127
Acoustic Input Impedance Measurements Plane-Wave Tube........................130
Quarter-Wave Resonator Package ............ ......................................... 133
Initial Deflection Measurements -Wyko Optical Profilometer................. ...............134
Energy Harvesting Measurements ................................................................. 138
O ptim al R resistance ....... ................................... ............... ............... 138
O ptim al Energy H harvesting ..................................................... .. .. ............ .. 139


v










6 ALTERNATIVE APPLICATION-PIEZOELECTRIC MICROPHONE............ 147

7 CONCLUSION AND FUTURE WORK ............ .................... .... .............153

D evelopm ent Issues ........1.. ... ...... ........ .... .................... .. ............ ... 154
F future W ork .............. ................................................................................ 156


APPENDIX

A M A T L A B C O D E ......................................................................... .......................158

le m .m ................................................................................................................... 1 5 8
e h setu p .m ........................................................................................................... 1 6 3
silicon_sw eep.m ............................................................................................. .....163
p latin u m .m .....................................................................................................1 6 3
tio2.m ........... .......... ................................ ............... 163
piezo_sw eep.m ............................................................................................. .... .....164
in itia lise .m ........................................................................................................... 1 6 4
totaldeflection forP.m ............................ ......... 164
totaldeflection forV .m .................................................. 164
initialdeflection.m ................................................................. ... ......... 165
form m atrix C M Fa.m .....................................................165
from m atrix dA .m ...............................................................166
abdpiezo.m ................................................. 166
con stantsl2 34 .m ................................................................167
so lv e r l .m .................................................................................................................. 1 6 8


B DETAILED PROCESS FLOW .................................... .................. 170

Energy Harvester Process Traveler ......... .........................170
Graphical Representation of Process Flow .........................................................172
M ask Layouts .......................................................................... ...... .............. ......... 174
P ack ag e D raw in g s ............................................................................................... 17 8

LIST OF REFERENCES ......................... ......... .........180

BIOGRAPHICAL SKETCH ............... ......... ........ ........195
















LIST OF TABLES


Table p

1-1: Typical material properties of select piezoelectric materials. ......................................7

1-2: Selected thin film piezoelectric materials and corresponding properties as reported in
th e lite ratu re ...................................... ............................... ................ 1 0

2-1: Equivalent lumped elements in several common energy domains............................28

2-2: C onjugate pow er variables. ............................................................. .....................29

2-3: General material properties used in modeling ........................................................55

2-4: PZT properties used in m odeling. ........................................ .......................... 55

2-5: Chosen device configurations for fabrication.................................. ............... 81

3-1: Wafer bow measurements and the resulting calculated stress ....................................86

3-2: Measured wafer mass and calculated density of PZT. ...................... .............. 87

5-1: Summary of electrical impedance measurements .............................. ...............1.13

5-2: Comparison of ferroelectric and dielectric properties of thin film PZT.................13

5-3: Summary of electrically actuated frequency response measurements ...................124

6-1: Summary of experimental results of microphone ................................. .............152
















LIST OF FIGURES


Figure pge

1-1: Roadm ap to Chapter 1. ...................... ..................................... ...... ... .

1-2: Schematic of overall energy harvesting concept ................ ............................... 3

1-3: Conceptual schematic of the acoustic energy harvester.................. ..................

1-4: B asic L C R electrical circuit......... ...................................................... ............... 4

2-1: R oadm ap to Chapter 2. ............. .................................................................... 18

2-2: Thevenin equivalent circuit for purposes of maximizing power transfer ..............21

2-3: Diagram showing side view of a Helmholtz resonator and its mechanical equivalent
of a m ass-spring-damper system ........................................................................... 29

2-4: Equivalent circuit representation of a Helmholtz resonator.......................................34

2-5: Theoretical pressure amplification of a conventional Helmholtz resonator. (a)
m magnitude (b) phase ...................................... .... .. ...... ........... 35

2-6: Cross-sectional and 3-D schematic of piezoelectric composite circular plate. (not to
sc ale) ...................................................... .... ................. 3 6

2-7: Notation of axes used in piezoelectric transduction ................................................37

2-8: Idealized perovskite crystal structure for PZT. a) centrosymmetric structure prior to
poling. b)non-centrosymmetric structure after poling..........................................38

2-9: Schematic of the poling process: ........................................ ......... ....39

2-10: Polarization vs. electric field hysteresis loop. .................................. ............... 40

2-11: Electro-acoustic equivalent circuit representation with Cb, Ca, and .................46

2-12: Electro-acoustic equivalent circuit representation with Ce C, and '. ................47

2-13: Dynamic electro-acoustic equivalent circuit. ................................ .................48









2-14: Sensitivity vs. frequency for a piezoelectric composite circular plate ...................55

2-15: Equivalent circuit of piezo-composite plate including the radiation impedance. ....57

2-16: Equivalent circuit with resistive load. ........................................... ............... 60

2-17: Conceptual 3-D and cross-sectional schematic of the circular composite plate .....62

2-18: Effective acoustic short-circuit compliance as a function of RIR,/ and tP/t, .......63

2-19: Effective acoustic mass as a function of R,/R2 and t/t, ......................................64

2-20: Short-circuit resonant frequency as a function of R,/R2 and tp/t, ......................65

2-21: Electro-acoustic transduction coefficient as a function of R,/R] and t /t, ............66

2-22: Electromechanical coupling coefficient, k, as a function ofRK/ R and t/t( .........67

2-23: Equivalent circuit for acoustic energy harvester with resistive load......................68

2-24: Magnitude of the acoustical input impedance for the piezoelectric composite
diaphragm and in combination with the Helmholtz resonator .................................70

2-25: Magnitude and phase of the acoustical input power for an input acoustic pressure of
P = 94 dB ......... .......................................................... .......... ..... 72

2-26: Electrical output impedance for the piezoelectric composite diaphragm by itself and
in combination with the Helmholtz resonator. ......................... ............... ......73

2-27: Magnitude and phase of the output voltage for the piezoelectric composite
diaphragm and in combination with the Helmholtz resonator. .............................74

2-28: Magnitude and phase of the electric output power delivered to the load resistor....75

2-29: Magnitude and phase of the energy harvester efficiency. ...................................76

2-30: Open-circuit voltage as a function of the applied acoustic pressure for the
piezoelectric diaphragm by itself and packaged with a Helmholtz resonator..........77

2-31: Input and output power as a function of the load resistance placed across the
electrodes of the piezoelectric material .............................................................. 78

2-32: Input and output power as a function of the applied acoustic input pressure, while
u sing an optim al load resistor ...................................................................... ....... 79

3-1: R oadm ap for C chapter 3. .................................................................... ..................83









3-2: Condensed process sequence in cross-section. ................................................84

3-3: Early DRIE results showing significant sidewall damage. ........................................88

3-4: SEM image showing black silicon at the base of a DRIE-etched trench ...................89

3-5: SEM image of a successful DRIE etch through the thickness of a silicon wafer. .....90

3-6: Illustration of a single diaphragm device.. ...................................... ...............91

3-7: Block layout illustrating numbering scheme for devices. ........................................91

3-8: Wafer layout illustrating numbering scheme for blocks. ........................................92

3-10: Quarter-wave resonator package consisting of acrylic plug, copper leads and vent
channel. ....................................................................93

3-11: Sealed cavity package...................................................................... ...................95

3-12: Optical photograph of a packaged energy harvester. .............................................95

4-1: R oadm ap for C chapter 4. .................................................................... ...................99

4-2: Experimental setup for impedance and power measurements. .............................104

4-3: Schematic of incident, reflected and input power. ............. .................................... 106

5-1: R oadm ap for C chapter 5. ........................................... ......................................... 108

5-2: Hysteresis loop for Device 1A-5-4 in a sealed cavity package. The device has a
radius of 1200 ium and a central mass. ...................................... ............... 109

5-3: Measured parallel output capacitance, Cp vs. outer radius, R, overlaid with a
theoretical capacitance curve based on the average extracted dielectric constant. 111

5-4: Parallel output resistance, Rp, vs. outer radius, R,, overlaid with a theoretical
resistance curve based on the average measured conductivity............................112

5-5: Device 1A-6-1, electrically actuated center deflection for the device with a radius of
900 /um and no central mass, packaged in the quarter-wave resonator package.. 115

5-6: Device 1A-6-2, electrically actuated center deflection for the device with a radius of
900 /um and a central mass, packaged in the quarter-wave resonator package....116

5-7: Device 1A-11-3, electrically actuated center deflection for the device with a radius
of 1200 /um and no central mass, packaged in the quarter-wave resonator..........117









5-8: Higher order mode image of Device 1A-11-3, taken using scanning laser vibrometer
at 12 0 .9 k H z ............... ............ ......................................... .................................. 1 17

5-9: Device 1A-10-4, electrically actuated center deflection for the device with a radius
of 1200 um and a central mass, packaged in the quarter-wave resonator ..........118

5-10: Device 1A-6-5, electrically actuated center deflection for the device with a radius
of 1500 um and no central mass, packaged in the quarter-wave resonator.......... 119

5-11: Device 1A-10-6, electrically actuated center deflection for a device with a radius of
1500 um and a central mass, packaged in the quarter-wave resonator package. .120

5-12: Device 1A-3-7, electrically actuated center deflection for the device with a radius
of 1800 um and no central mass, packaged in the sealed-cavity package............ 121

5-13: Electrically actuated sensitivity at low frequency (well below resonance)
(R,/R = 0.95).................................................................. ....... 122

5-14: Summary of electrically actuated resonant frequencies. ....................................... 122

5-15: Drop in resonant frequency due to the addition of the central mass. .....................123

5-16: Magnitude of the resonant center deflection versus driving voltage......................125

5-17: Mechanical sensitivity at resonance versus driving voltage.............................126

5-18: Device 1A-5-4, magnitude and phase of the acoustically actuated frequency
response in a sealed cavity package for the device with a radius of 1200 /m and a
central m ass. ........................................................................128

5-19: Device 1A-10-4 in a quarter-wave resonator package. Magnitude and phase of the
acoustically actuated frequency response for the device with a radius of 1200 pm
and a central m ass. ........................................ ................... ........ 129

5-20: Device 1A-6-5 in a quarter-wave resonator package. Magnitude and phase of the
acoustically actuated frequency response for the device with a radius of 1500 um
and no central m ass. ........................... ................ ................... .. .... .. 130

5-21: Device 1A-10-2, Normalized acoustic impedance in a sealed cavity package for the
device with a radius of 900 pm and a central mass. ..................... ...............131

5-22: Device 1A-5-4, Normalized acoustic impedance in a sealed cavity package for the
device with a radius of 1200 pm and a central mass .........................................132

5-23: Device 1A-3-7, Normalized acoustic impedance in a sealed cavity package for the
device with a radius of 1800 pm and no central mass ............... .................132









5-24: Device 1A-6-1, Normalized acoustic impedance in a quarter-wave resonator
package for the device with a radius of 900 pum and no central mass .................133

5-25: Device 1A-11-3, Normalized acoustic impedance in a quarter-wave resonator
package for the device with a radius of 1200 pum and no central mass ..............134

5-26: Device 1A-11-3, Initial static deflection resulting from residual stresses for a device
with a radius of 1200 pm and no central mass. ............. ..................................... 135

5-27: Device 1A-6-6, Initial static deflection resulting from residual stresses for a device
with a radius of 1500 pum and a central mass ...................................................... 136

5-28: Device 1A-3-7, Initial static deflection resulting from residual stresses for a device
with a radius of 1800 pm and no central mass. ............. ..................................... 136

5-29: Device 1A-3-8, Initial static deflection resulting from residual stresses for a device
with a radius of 1800 pum and a central mass ...................................................... 137

5-30: Measured power delivered to a load as function of the load resistance for Device
1A-4-4 and Device 1A-3-8 as compared against theoretical values ....................140

5-31: Measured output voltage across the load as a function of applied acoustic pressure
and com pared to theoretical values.. ........................................... ...............140

5-32: Measured power delivered to load as a function of applied acoustic pressure and
com pared to theoretical values .................................................................. ....... 14 1

5-33: Measured overall efficiency of each device overlaid with theoretical values for
com prison ...................................................... ................. 142

5-34: Resonant frequency versus applied acoustic pressure, resulting from nonlinear
response..................................... ........................... ..... .......... 146

6-1: Linearity of the microphone device at 1 kHz....................................................... 148

6-2: Frequency response spectrum in terms of magnitude and phase. ............................149

6-3: Noise floor spectrum of output voltage when no acoustic signal is applied, as well as
noise floor due to measurement setup alone. ................................. ............... 151

6-4: Electrically actuated frequency response of microphone device. ............................152

7-1: R oadm ap to C chapter 7. ............. ..................... ......... .................................... 154

B-l: Step 1: Deposit Ti on SOI Wafer and oxidize to TiO2. .......................................172

B-2: Step 2: Deposit Ti/Platinum(30nm/170nm) -Lift-Off w/ Mask (ElectrodeBot) ......172









B-3: Step 3: Spin PZT 6 times to achieve desired thickness ............... ................. 172

B-4: Step 4: Deposit Platinum(180nm)-Lift-off w/ Mask (ElectrodeTop)......................172

B-5: Step 5: Wet Etch PZT using Pt as etch mask................................... ..................173

B-6: Step 6.1: Spin thick photo resist on bottom (7um). ............................................173

B-7: Step 6.2: Pattern using mask (Cavity). ....................................... ............... 173

B -8: Step 6.3: D R IE to B O X ............................ ...................................................... 173

B -9: Step 6.4: A sh R esist. ...................... .................... .......................... 173

B-10: Step 6.5: BOE backside to remove oxide. .................................. .................174

B -11: B ackside m etal m ask .................................................. ............................... 174

B-12: Bottom electrode mask ......... .. ..... .. ......... .. ........................ 175

B -13 : T op electrode m ask ................................................................................ .... ... 176

B -14 : C av ity m a sk ..................................................................................................17 7

B-15: Detailed schematic drawing of quarter wave resonator package and mounting.... 178

B-16: Detailed schematic of sealed cavity package and mounting plate........................ 179















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

DEVELOPMENT OF A MEMS-BASED ACOUSTIC ENERGY HARVESTER

By

Stephen Brian Horowitz

December 2005

Chair: Toshikazu Nishida
Cochair: Mark Sheplak
Major Department: Electrical and Computer Engineering

In this dissertation, I discuss the development of an electromechanical acoustic

resonator for reclamation of energy using microelectromechanical systems (MEMS)

technology. The MEMS device converts acoustical energy to electrical energy that can

then either be stored for later use or utilized directly for a particular circuit application.

The work presented in this dissertation takes a first step towards that goal, by

designing, fabricating and testing several electromechanical acoustic resonators of

varying size. The resonators are fabricated using MEMS processing techniques on a

silicon wafer. Each resonator consists of a 3 um thick silicon diaphragm with a circular

ring of piezoelectric material. The diaphragm transduces an acoustical pressure

fluctuation into a mechanical deformation, while the piezoelectric material transduces

that mechanical deformation into an electrical signal (charge or voltage).















CHAPTER 1
INTRODUCTION

The goal of my research was to utilize microelectromechanical systems (MEMS)

based fabrication techniques to develop an electromechanical acoustic resonator for

harvesting of acoustic energy. The MEMS device converts acoustical energy to electrical

energy that can then either be stored or used directly for a particular circuit application.

The relatively small geometries possible in MEMS make such a device useful for small,

portable devices, where battery requirements are often difficult to meet. Additionally, the

device is well suited to applications where wiring is difficult such as remote sensors or

unwieldy as in the case of large arrays of devices requiring power.

This chapter begins with an overview of the acoustical to electrical energy

reclamation process that is employed in this research. Following this, the motivation

behind this research is presented. Then, an in-depth literature survey is presented to

familiarize the reader with the technological and theoretical developments related to

energy harvesting and piezoelectricity, followed by a short discussion of the technical

approach that was employed. A graphical roadmap for this chapter is shown in Figure

1-1.

Acoustic Energy Harvester-Concept

The overall concept behind the energy harvester is illustrated in Figure 1-2. The

main idea is to convert acoustic energy into a usable form of electrical energy. The

figure illustrates a plane wave tube as the source of acoustic power. Some of the incident







acoustic power is reflected at the end of the tube; however, a portion is transmitted
through to the energy harvester that is circled in Figure 1-2. The energy harvester
performs the actual transduction of energy from the acoustical to the electrical energy
domains. More details on this transduction technique will be provided later.


Piezoelectiicity


Piezoelectiic Nhiterials
in NIENIS

ImpedaC(lnI1ce NIOde(iIng
of Piezoelecriics


I
h

r
I-


Eiieiry Absorpfioii/
Eine.tion DHamin2
Enei-2- H. jj~ll


Figure 1-1: Roadmap to Chapter 1.









The electrical energy that exits the energy harvester has the same temporal form

as the input acoustic signal, which we assume for now is sinusoidal. The sinusoidal

signal is then routed to the energy harvesting circuitry that serves to rectify and condition

the signal before sending it on to the end application. In the example illustrated in

Figure 1-2, the end application is a battery that is charged by the output of the harvesting

circuitry.

Plane Wave Tube
fNo Ac
% I. > DC





Circuitry

Power B
i Battery
Energy Harvester


Figure 1-2: Schematic of overall energy harvesting concept.

A conceptual close-up schematic of the acoustic energy harvester is shown in

Figure 1-3. This device consists of a Helmholtz resonator possessing a piezoelectric

composite backplate. A Helmholtz resonator, which is a type of acoustic resonator,

consists of a cavity connected to the environment through a small neck. When excited by

an acoustic input, a single resonance is seen, whereby the acoustic pressure inside the

cavity is amplified to a level much greater than the incident acoustic signal [1, 2].

The Helmholtz resonator is very similar to an LCR resonant circuit in electrical

engineering, as shown in Figure 1-4. Both systems have a single degree-of-freedom, with

a single resonant peak, whereby the amplitude of the forcing function is greatly









amplified. In the LCR circuit, the voltage is equivalent to the acoustic pressure. Both

systems operate through the oscillation of energy between generalized potential and

kinetic forms. In electrical systems this storage occurs via capacitors and inductors

respectively. These electroacoustic analogies will be explained in more detail in Chapter

2.

neck




cavity





piezoelectric compliant
ring plate

Figure 1-3: Conceptual schematic of the acoustic energy harvester.

R L






Figure 1-4: Basic LCR electrical circuit.

The large acoustic cavity pressure created by the resonance can then be exploited

for energy reclamation by converting the acoustic energy to electrical energy. This

conversion is performed by the piezoelectric composite plate. First, acoustical to

mechanical transduction is accomplished via the compliant diaphragm, followed by

mechanical to electrical transduction, due to the piezoelectric response of the composite,

whereby a mechanical strain creates an electrical voltage.









The voltage that is created through the electromechanical transduction can be

harvested by the energy harvester circuit, which is a necessary part of the energy

harvesting process. The necessity arises due to the need to match impedances at

interfaces. As will be derived in Chapter 2, the power transfer across an interface is

maximized when the impedances on each side of the interface are complex conjugates of

each other. This is often referred to as impedance matching. One role of the energy

harvesting circuit is to provide an input impedance that matches the output impedance of

the piezoelectric structure. Additionally, the harvesting circuit converts the form of the

electrical energy to a more appropriate form for storage, such as converting from AC to

DC.

In one implementation, a rectifying switched capacitor circuit is used to efficiently

convert the AC signal to DC that is then stored across a low-loss capacitor [3]. Another

possibility is based on the Kymissis circuit approach [4]. This approach utilizes passive

storage elements for impedance matching coupled with a regulated output, for improved

energy harvesting performance.

Motivation

Acoustic energy is commonly present in a variety of everyday situations. The

motivation behind this research is to enable the reclamation of that energy that would

otherwise be lost. Although the available acoustic energy may be small in many

situations, the energy requirements for certain applications such as microsensing are also

correspondingly small. The ability to reclaim acoustic energy and store it in a usable

electrical form enables a novel means of supplying power to relatively low power

devices.









Background

Piezoelectricity

In 1880, the brothers Pierre and Jacques Curie discovered that some crystals

developed surface charges when compressed [5]. They furthermore found that these

charges were proportional to the applied pressure. This phenomenon was later named

piezoelectricityy" by Wilhelm Gottlieb Hankel, and is historically referred to as the direct

piezoelectric effect [5]. Additionally, in these same crystals, a strain is produced under

application of an electrical field. This is commonly referred to as the converse

piezoelectric effect. The material constant relating strain and charge in a piezoelectric

material is known as the piezoelectric charge modulus, d, and is typically quoted in units

of pC/N or pm/V. In order for a material to be piezoelectric, it must have a non-

centrosymmetric crystal structure.

It was not until 1921, that a useful application was developed for piezoelectricity.

It came in the form of a quartz crystal oscillator that was developed by Walter Cady to

provide good frequency stability for radio systems [5, 6]. Then, in 1947, the first

commercial phonograph pickups based on barium titanate (BaTiO3), a piezoelectric

ceramic, were introduced [6]. Modern applications of piezoelectric materials now

include high voltage ignition systems, piezoelectric motors, ink-jet printer heads, acoustic

speakers, sonar, ultrasonic transducers, frequency filters, acoustic delay lines, electrical

transformers, and a wide range of physical sensors, such as acoustic, force, pressure, and

acceleration sensors [6].

Piezoelectric Materials in MEMS

A number of papers have been published on the use of piezoelectric and

ferroelectric materials in MEMS [6, 7]. Piezoelectric materials commonly used include









zinc oxide (ZnO), aluminum nitride (A1N), and lead zirconate titanate (PZT). The choice

of piezoelectric material depends on several selection factors including deposition

methods, process complexity, integrated circuit (IC) compatibility, and material

parameters.

Materials and properties

Of the three commonly used materials, PZT offers the largest piezoelectric

modulus [6]. In comparison to PZT, A1N and ZnO have relatively weak piezoelectric

coefficients and coupling factors, however they tend to have low dielectric constants and

low dielectric losses, making them more attractive for certain applications [8]. A table of

typical material properties is shown in Table 1-1 for comparison; however these

properties are highly dependent on the deposition method, as will be described in more

detail shortly.

Table 1-1: Typical material properties of select piezoelectric materials.
31 33 33,r E P
[pC/N] [pC/N] [GPa] [kg/m3]
PZT[6] -130 290 1300 96 7.7
A1N[8] n/a 3.4 10.5 330 3.26
ZnO[9-11] -4.7 12 12.7 210 5.6

In Table 1-1, d31 is the piezoelectric coefficient relating electrical displacement in

the '3' direction (z-axis) to a mechanical stress in the '1' direction (x-axis) and is the

relevant coefficient for bending-mode transducers, whereas d33 relates the electrical

displacement in the '3' direction (z-axis) to a mechanical stress in that same direction.

The coefficient, d33, is thus the relevant coefficient for compression-mode transducers.

Additionally, E33,r, E, and p are the relative dielectric constant, Young's modulus and

mass density of the materials, respectively.









Of the three materials, only A1N is fully compatible with standard IC processing,

eliminating the integration difficulties present with PZT and ZnO. PZT poses a particular

integration challenge, as lead contamination is always a big concern; however, PZT has

been integrated successfully into ferroelectric random-access memory (FRAM) [12-14],

typically using sputtered platinum as a bottom electrode and diffusion barrier layer.

Deposition methods

Piezoelectric deposition techniques include various sputtering methods, photo-

ablation [15], hydrothermal and chemical vapor deposition (CVD) techniques, and spin-

on sol-gel processing. For deposition of PZT, three popular methods are sputtering, sol-

gel and hydrothermal techniques. The earliest work used various forms of sputtering to

deposit PZT [16-22] followed soon after by CVD [23]. Castellano and Feinstein [16]

used an ion-beam sputtering technique to deposit PZT as did Trolier-McKinstry et al.,

[22] while Sreenivas et al. [17] employed DC magnetron sputtering in their PZT

deposition. Additionally, sputter deposition of PZT thin films was performed by

Kawabata et al.[24] and Li et al. [25]. Dubois and Muralt [8] deposited PZT using two

different methods. One method involved a sol-gel process and produced an average

thickness of 900 nm. The other method used reactive sputtering and achieved a thickness

of approximately 500 nm.

Hydrothermal deposition relies on a chemical reaction between a titanium bottom

electrode and ionic solution under high temperature and pressure. Deposition of 10 um

thick PZT by hydrothermal techniques was performed by Kanda et al. [26] for use in a

touch probe sensor. Additionally, Morita et al. [27-30] chose to deposit PZT by the

hydrothermal method due to the relatively large thicknesses achievable and the self-

alignment of the poling direction during deposition.









More recently, numerous researchers have utilized the sol-gel process for

deposition of PZT [31-47]. In particular, Bahr et al. [31] used a sol-gel PZT process and

investigated the reliability and piezoelectric properties of the resulting material. Using

this process for various film thicknesses, the relative dielectric constant, 833,., varied

between 700 and 1000 and delamination occurred for an indentation load of 1250 pN for

a film sintered for 5 minutes. Higher loads were found to be achievable using longer

sinter times. Bernstein et al. [32] and Xu et al. [33] used a sol-gel process to achieve

crack-free PZT films with thicknesses of up to 12 um that yielded piezoelectric properties

close to the bulk values for PZT. A dielectric constant of 1400 and a piezoelectric

coefficient, d3, of 246 pC/Nwere reported on a 4 ,um thick film. The PZT films were

used on an array of membranes for acoustic imaging. Kunz et al. [37] report a

piezoelectric coefficient, d,,, of 110 pC/Nfor a sol-gel deposited PZT film used in a

triaxial accelerometer. Zurn et al. [48] report similar material properties for sol-gel

deposited PZT on a microcantilever, including a d,3 of 120 pC/Nfor a PZT film

thickness of 0.5 um deposited on a low-stress silicon nitride layer. A summary table of

deposited thin film piezoelectric materials and their properties, as reported in the

literature, is given in Table 1-2. In this table, e3, is a piezoelectric coefficient relating

mechanical stress and an electric field, tan 8 is called the loss tangent and is a measure

of the relative losses in the material, and r,, is the residual stress in the material, while

the other properties are as previously defined.

Piezoelectric integration with MEMS

Some of the earliest integration of piezoelectric materials with MEMS focused

primarily on ZnO [49, 50] and A1N [49] as the piezoelectric material. More recently,






10


Devoe and Pisano [10] developed and characterized surface micromachined piezoelectric

accelerometers that utilized thin films of ZnO for sensing. Deposition of the ZnO was

performed using single-target RF sputtering and exhibited a piezoelectric coefficient of

2.3 pC/N. Also, Devoe [9] investigated micromechanical beam resonators that use ZnO

and a three-mask fabrication process. The resonators are intended for use as

electromechanical filters.

Table 1-2: Selected thin film piezoelectric materials and corresponding properties as
reported in the literature.


Ref Material Deposition
Method


[9, 10] ZnO RF Sputt.
[11] ZnO Sputtered
[8] A1N DC Sputt.
[8] PZT (45/55) React. Sputt.
[8] PZT (45/55) Sol-Gel
[8] PZT (53/47) React. Sputt.
[26] PZT Hydrothermal
[31] PZT (52/48) React. Sputt.
[32] PZT React. Sputt.
[37] PZT (53/47) React. Sputt.
[42] PZT React. Sputt.
[44, 45] PZT (52/48) React. Sputt.
[48] PZT React. Sputt.
[16] PZT IB Sputt.
[20,21] PZTWCd RF Sputt.
[51] PZT(X/1-X) React. Sputt.


Properties
-33,r 31 d33 e3


12.7

10.5
900
1100
1300

700-1000
1400

1100
800

125
60-460
200-600


2.3






-34.2


110


3.4
55
50
70


246


-200
120

30


C -1
-- )

-1.02
-5.12
-8.28
-6.83
-0.13


E P tan ( o-r
(GPa)k (Pa)

161 5605 -- 1...80
210 5700 -- -
-- 0.002 -700
-- 0.03 150
-- 0.03 70
-- 0.05 230


-- -- 0.03


56 -- 0.02
60 7600 --


-- -- -- 0.02


The integration of ZnO with micromachining was also investigated by Indermihle

et al. [52], where it was used in an array of silicon micro cantilevers. The end application

for the array was parallel atomic force microscopy. Another application of ZnO was

performed by Han and Kim [11] in the fabrication of a micromachined piezoelectric

ultrasonic transducer. They sputter deposited ZnO on Al, followed by a layer of parylene









for insulation. Dubois and Muralt [8] fabricated thin films of RF sputtered A1N and

performed measurements of the effective transverse piezoelectric coefficient that yielded

a value of 3.4 pm/V.

Reduced-Order Modeling of Piezoelectric Materials

The development of accurate, practical models of the piezoelectric transduction and

associated structural interactions is critical to the design and behavioral prediction of

piezoelectric based devices, particularly when micromachining is involved. For a

micromachined device, considerable time and expense must be invested in the

fabrication, and it is therefore desirable to have some ability to predict the device

behavior before proceeding with fabrication.

Behavioral prediction can be accurately accomplished using complete analytical

methods. However, these methods, which often involving partial differential equations,

can often be unwieldy and physically unintuitive. Furthermore, this functional form is

not readily conducive to a full systems level design that links the transducer to

electronics. Similarly, finite element modeling (FEM) techniques are often used to

predict system behavior, numerically. The results produced by this technique can very

precisely follow the physical system; however, the physical insight that can be gleaned is

limited. Additionally, it is very difficult to determine scaling behavior from FEM

results. The scaling behavior, i.e., the change in the system performance as the entire

system is scaled up or down, is a critical design issue in the creation of devices using

micromachining technology. In order to facilitate a physics-based approach for design, a

simplified, reduced-order model is necessary that accurately captures the geometric and









material dependencies. This reduced-order model uses lumped elements to represent the

key components that dominate the device behavior.

In 1915, Butterworth [53] first showed that any mechanical oscillator, when driven

by a periodic voltage across a capacitor, would have an equivalent electrical circuit

consisting of a resistance, inductance and capacitance in series, and all in parallel with

another capacitance. Working independently, the earliest equivalent circuit model

specifically for piezoelectric materials was developed by Van Dyke in 1925 [54-56].

Later, Dye [57] proved that Van Dyke's circuit could be derived from Butterworth's

theorem. Additionally, Mason [58, 59] and Cady [5] provided thorough reviews of the

equivalent circuit model and associated equations for quartz oscillator applications.

Finally, Fischer [60] extensively covered equivalent circuit models for electromechanical

oscillators.

Significant research has been performed on the modeling of structures containing

piezoelectric materials. [38, 41, 61-94] Cho et al. [61, 65] developed a five-port

generalized equivalent circuit for a piezoelectric bimorph beam. The generalized circuit

can be used under a variety of boundary conditions. In these papers, three boundary

conditions are specifically analyzed free, simply supported, and cantilevered. For these

boundary conditions, the equivalent circuit was found to produce the exact expressions

for the beam vibration. Other equivalent circuits were developed by Martin [68] for

resonators with low Q values, Sheritt et al. [69] for thickness vibrators, Lin [67] for

coupled ceramic disk resonators, and Chen et al. [68] for coupled resonant filters.

Tilmans [94] also presented an equivalent circuit approach for modeling distributed

parameter systems. Liang et al. [66] developed a generalized electromechanical









impedance model that was then used to address issues of energy conversion, power

consumption, and dynamic response. The approach described can be used for any system

for which the driving-point impedance can be found either analytically or experimentally.

Additionally, van de Leur [91] provided a critical interpretation of equivalent circuit

models obtained from impedance measurements, indicating that care must be taken in

identifying individual components contributing to an impedance as resulting from

particular structures in a given device. This can be further understood as an example of

the non-uniqueness of equivalent circuit representations, i.e., more than one equivalent

circuit can represent the same impedance. Additionally, Lesieutre and Davis [87]

provided insight into the piezoelectric coupling coefficient, including a technique

whereby an outside stress is applied to increase the effective device coupling coefficient

beyond the coupling coefficient of the material itself.

Research into coupled resonators and their unique properties is a related and also

relevant topic since the acoustic energy harvester involves mating the piezoelectric

composite diaphragm to a Helmholtz resonator, resulting in a coupled resonant system.

Fischer [60] provided an early treatment of coupled resonant systems. Chen et al., [72],

Lin [71, 95], and Li et al. [96] also discuss modeling and design of coupled resonant

systems. These papers address some of the issues involved in coupled resonators,

including the shift in resonant frequencies away from their uncoupled values.

Energy Absorption/Vibration Damping and Energy Harvesting

The absorption of acoustical and mechanical energy via piezoelectric coupling is

closely related to the harvesting of electrical energy from acoustical and mechanical

energy sources. From the acoustical point of view, acoustical energy that is converted to

the electrical domain and dissipated across a resistive load can be viewed as being









absorbed. From the electrical point of view, the incident acoustical energy can be viewed

as an energy source that may theoretically be harvested and used immediately for

electrical subsystems or stored for later use. Because of the related nature of these two

fields, papers that address either of these fields are directly relevant to acoustic energy

harvesting.

Some of the earliest work in piezoelectric vibration damping was performed by

Hagood and von Flotow [97] who used resistive and inductive elements in various shunt

networks designed to dissipate mechanical energy. A few years later, Hollkamp [98]

discussed the use of resonant electrical shunt circuits for multi-modal vibration

suppression. In addition, numerous other researchers have investigated piezoelectric

means of vibration damping [99-108]. More recently, Caruso [109] discusses electrical

shunt circuits for damping of vibrations, as does Wu et al. [110] for damping of panels on

an F-15 aircraft.

Additionally, a number of papers directly address the issue of obtaining electrical

energy from piezoelectric conversion of mechanical energy [3, 4, 111-117]. Early work

in this area was performed by Lomenzo et al. [114] and Stein et al. [116]. Later,

Kymissis' [4] and Smalser's [3] work focuses on the electrical circuitry necessary for

storage of piezoelectrically generated energy, while Meninger [115] discusses energy

harvesting from an electrostatic transducer. Goldfarb and Jones, [111] Giurgiutiu and

Rogers [112, 113] and Zhou and Rogers [117] utilize impedance modeling of the

piezoelectric and associated structure to analyze the ability to harvest energy.

Approach

The approach taken in this research is to divide the concept of an acoustical energy

harvester into three distinct components as shown previously in Figure 1-2. The first









component consists of a piezoelectric composite diaphragm that is responsible for the

acoustical to electrical transduction of energy. The second component, which is the

packaging, serves as a means to improve the acoustical coupling to the environment. The

third and final component is the electrical circuitry that takes the alternating current (AC)

electrical signal from the piezoelectric diaphragm and converts it into a direct current

(DC) output that can be stored for later use. The first two components of the energy

harvester were developed in this research, while the conversion circuitry remains an

ongoing research topic in our group [118-120]. Furthermore, a macroscale version of an

acoustic energy harvester was previously developed [120] and, through miniaturization,

led to the present MEMS-based energy harvester.

Chapter 2 presents a theoretical background on various aspects of this research

including lumped element modeling of the complete system as well as individual

components along with predictions of the dynamic behavior. Chapter 3 discusses the

fabrication of the piezoelectric diaphragm and the packaging scheme employed.

Following this, the experimental setup is provided in Chapter 4, including a discussion of

the measurements that were taken. Then, in Chapter 5, the experimental results are

presented. Chapter 6 addresses an alternate application that has been explored, while

Chapter 7 presents the conclusions and a statement of the contributions to this research.

Finally, Appendix A presents the Matlab program code for computing the diaphragm

deflection and calculating the lumped element values, while Appendix B provides a

detailed process flow, mask layout, and packaging design

The main contributions of this dissertation are as follows:

Acoustic energy harvesting concept






16


* Comprehensive lumped element model (LEM) of the acoustic energy
harvester

* First reported integration of thin-film PZT with SOI-based MEMS

* First aeroacoustic capable piezoelectric microphone.















CHAPTER 2
THEORETICAL BACKGROUND

This chapter focuses on the theory and modeling of the acoustic energy harvester

and each of its components. The chapter begins with an overview of energy harvesting,

followed by an introduction to lumped element modeling, by discussing both its utility

and limitations. This is followed by the development of a lumped element model for a

Helmholtz resonator and the corresponding equivalent circuit. Next, the modeling of

piezoelectric composite circular plates is addressed. To achieve this, an equivalent circuit

is presented and general equations are discussed. Then specific limiting cases are

addressed that simplify the analysis. Finally, the overall dynamic behavior of the system

is theoretically investigated, including the acoustical input and electrical output behavior.

A graphical overview of this chapter is provided by the roadmap of Figure 2-1.

Overview-Energy Harvesting

The general concept behind energy harvesting is to convert energy from the

environment that is in an otherwise unusable form into a more useful form. Often the

form of energy that is most useful in modem applications is electrical energy, where it

can be stored in a battery or used to power electrical circuitry. The initial form of energy

can originate from any number of energy domains, such as optical, thermal, mechanical,

acoustical, fluidic, chemical, and biological. Some form of transducer is then required to

convert that energy to a usable form of electrical energy. Depending on the situation, that

also may require passage of the energy through an intermediary energy domain..

























it e Plate


Figure 2-1: Roadmap to Chapter 2.









There are pros and cons to each of these different local energy sources [121]. Some

energy sources such as solar power utilize the optical energy domain to achieve high

power densities on the order of 15,000 uW/cm2 in direct sunlight. The main downside

is that direct sunlight is not always available. Vibrational energy, on the other hand,

offers power densities up to 250 /W/cm2 While this is considerably lower than solar

energy, it is useful in places without sunlight but where high vibrational energy is

available.

Acoustic energy, in a manner similar to vibrational energy, offers power densities

on the order of 1 uW/cm2 for a 100 dB acoustic signal [121], or approximately 964,000

t/W/cm2 at 160 dB. While most signals are typically much lower than 160 dB, there are

applications where such high levels are present. Additionally, as with vibrational energy,

acoustic energy, does not require the presence of sunlight.

Chemical energy sources are commonly employed today in the form of batteries

and fuel cells. Batteries typically offer power densities in the range of 45 uW/cm3 for

non-rechargables, and 7 uW/cm3 for rechargeable lithium batteries [121]. Fuel cells

employing methanol, on the other hand, offer power densities as high as 280 /W/cm3,

leading to the current interest in fuel cell development [121]. Additionally, a micro-

combustion engine that employs hydrocarbons as a fuel source realizes power densities of

333 /jW/cm3 [122]. The downside to these types of chemical energy sources is the

limited supply of energy. Used batteries must either be replaced or recharged and fuel

cells require refueling for continual operation. By contrast, the scavenged energy

sources, such as vibrational, solar, and acoustical do not theoretically have a limited

supply of energy, given the right operating environment.









For the particular application considered in this dissertation, the initial energy is in

the form of acoustical energy. In order to convert acoustical energy into electrical

energy, a diaphragm based transducer is utilized. The diaphragm transducer utilizes the

mechanical energy domain as an intermediary to transfer energy.

Maximum Average Power Transfer

Regardless of the route through which the energy passes, certain fundamental

issues must be addressed in order to maximize the amount of energy that is harvested.

Whenever a change in impedance is encountered by a traveling wave, a reflection of all

or part of the energy in that wave occurs, while the remaining energy is transmitted. The

procedure for maximizing the energy focuses on minimizing the reflected component and

thus maximizing the transmitted component. This is achieved by matching the

impedance along the route traveled by the energy. As long as the impedance at a given

interface is matched, complete transmission of the energy will occur, regardless of

whether or not the two sides of the interface are in the same energy domain. Often it is

not possible to match an impedance exactly, due to external and physical constraints on

the system, however it is generally still desirable to match the impedances as closely as

possible in order to optimize for maximum energy within those constraints [2, 123].

Additionally, one must consider that power is a complex quantity. The total power,

which is generally complex, is composed of real power and reactive power. The real

power is the physical power that is delivered to the load, while the reactive power

represents energy that is temporarily stored in the load, before it returns to the source.

When designing an energy harvesting system, it is generally preferable to maximize the

real power rather than the total power, as it is only the real power that is ultimately

retained by the load.









It is useful to look at a Thevenin equivalent circuit, as shown in Figure 2-2, in order

to determine the maximum achievable power transfer to a load [124]. Assuming

sinusoidal voltage and current, the average real power delivered to the load is


n =VL Cos( -L ), (2.1)


where VL and IL are the peak magnitudes of the voltage and current at the load, and 0.

and 0, are the phase angles of the voltage and current respectively. This can be

rewritten as

1
nL = IL (pf ) (2.2)

where pf is the power factor, defined by the ratio of average power to apparent power

and is given by

n
pf =-r (2.3)
V I

where V,, and I, are the root-mean-square voltage and current, respectively.


I,
V TH

V oO VL Z



Figure 2-2: Thevenin equivalent circuit for purposes of maximizing power transfer.

Using basic circuit analysis, the load voltage and current can be expressed in

terms of the open circuit voltage, Vo, the Thevenin equivalent impedance ZH, and the


load impedance, Z The expressions are given as









VL- ZL (2.4)
TH + Z

and


IL = V1 (2.5)
z +z
ZTH + ZL

where the Thevenin equivalent impedance, ZTH = RH + jXTH, and the load impedance

can be written as ZL = RL + jXL Plugging these expressions back into Eq. (2.4) and Eq.

(2.5) and separating out the magnitude components yields




[(H RL2 (XTH X)2
VL = t. -.-- ^ ,Voo~z +AL2(2.6)



and


Ic = Vo 1 (2.7)
[(H )2+(XTH +XL)2 2

The phase angle between the voltage and current can also be expressed as 6, 0,L = Oz,

where Oz, is the phase angle of the impedance. The power factor can be defined as


pf =cosZL = RL1 (2.8)
(R2 + lx)2

Plugging this expression into Eq. (2.1), together with Eq. (2.6) and Eq. (2.7) yields

1 /V 2 R
nL = 2- c2L- (2.9)
2 (RH + R)2 +(XTH + XL )2

As, the quantity (XTH + X,) is only in the denominator, any nonzero value reduces the

overall power absorbed by the load, thus X = -XTH is optimal and reduces Eq. (2.9) to









1 V 2RL
HL = oc 2 L (2.10)
2(R +TH)2

This expression can be maximized by differentiating this expression with respect to RL

and equating the derivative to zero, yielding an optimal value of RL = RH The average

power to the load can be maximized, then, by setting the load impedance to

ZL = RL + jXL = RTH jXTH = ZTH*, (2.11)

where ZH* is the complex conjugate of the Thevenin impedance. This derivation

assumes total freedom in the choice of the load impedance.

Under the constraint of a purely resistive load, where XL = 0, the equation given

in Eq. (2.9) can be differentiated directly and set equal to zero with XL = 0, in order to

find the maximum average power transfer. This procedure yields an optimal value for

the load resistance of


RL = RTH2 +XTH2 = ZTH (2.12)

under the constraint of a purely resistive load.

Electromechanical Transducers

A transducer can be broadly defined as an interface between any two energy

domains, including optical, thermal, mechanical, electrical, acoustical, fluidic, chemical,

and biological. Typically, transducers are broadly classified into two main categories:

sensors and actuators. The breakdown among these classifications relates to the direction

of information flow. Sensors are transducers that convert information from the

environment into a form that is more easily analyzed, recorded or processed. Actuators,

on the other hand, take information from this form and convert it into a form that more









readily interacts with the intended environment. Both types of transducers are mainly

concerned with the way in which the information is utilized.

The classifications described above focus on the information aspect of a given

signal that is encoded in either the effort or flow, but not their product. As such, these

types of transducers are optimized for transmission of information in one direction or

another. Besides information, a signal carries power. A third type of transducer exists

that falls under neither category and can be best described as an energy harvester. The

concept behind the energy harvester is to achieve an optimal amount of average power

transfer between energy domains. Instead of optimizing sensitivity or dynamic range, an

energy harvester is optimized for efficiency and maximum average power transfer.

Additionally, actuators and especially sensors are designed as linear devices in

order to facilitate a straightforward relationship between input and output. The linear

input-output relationship ensures that the problem of calculating their behavior is

tractable, and is necessitated by the information contained in the signal. Energy

harvesters, on the other hand, do not necessarily require a linear input-output relationship

as only the power is of interest. Some details of the input may get lost in the nonlinear

transduction, but only the power at the output is needed.

There are a wide range of transduction techniques available that couple the

mechanical and electrical energy domains, including electrostatic, electrodynamic,

piezoresistive and piezoelectric [1, 125]. Each technique offers various tradeoffs among

performance, reliability, and material integration. Transducers can be separated into two

types, direct and indirect, based on the way they interact with energy. Direct transducers,

as the name implies, directly transduce energy between the two energy domains. Indirect









transducers, however, rely upon a second energy source that modulates the primary

energy source as it passes through the transducer. The inherent nature of indirect

transduction leads to a lower transduction efficiency as compared to direct transduction.

Electrodynamic transduction, a direct transduction mechanism, yields high

coupling efficiencies. Additionally, this method uses no outside power source in order to

operate. The transduction is essentially powered by the input signal itself. Despite these

advantages, electrodynamic transduction suffers from difficulty in integrating the

particular required materials, such as copper or magnetic materials, into a standard IC

process line. Additionally, the necessary fabrication is often more complex than the other

techniques discussed here.

Piezoelectric transduction requires no outside source of power other than the input

signal, and offers a high sensitivity, with a potentially low noise floor. The main

disadvantages are the difficulty in integrating the piezoelectric materials into a standard

process flow, and the limited coupling efficiency due to indirect transduction of energy.

Lumped Element Modeling

The most accurate, complete, mathematical description of a physical system is a

physics-based model, supported by an exact analytical expression for the system

behavior. Why then are alternative modeling techniques commonly used? These

alternative methods, such as lumped element modeling (LEM), and finite element

modeling (FEM), excel in predicting system behavior in situations where an exact,

analytical approach is unwieldy or impossible. Additionally, it is not uncommon to have

an exact mathematical description of a system that allows for precise prediction of device

behavior, but is so complex that it offers little in the way of physical insight into the









design and scaling system. Often, the exact solution is in a mathematical form that, while

convenient for behavioral prediction, conceals the underlying physics.

Similarly, FEM techniques can accurately predict system behavior, in this case via

a numerical approach. The results produced by this technique can very precisely follow

the physical system; however, the physical insight that can be gleaned is limited.

Furthermore, the results depend on the numerical mesh and convergence of the iterative

calculations. In particular, it is very difficult and/or cumbersome to determine scaling

behavior from FEM results. The scaling behavior, i.e. the change in the system

performance as the entire system is scaled up or down, is a critical design issue in the

creation of devices using micromachining technology.

In order to efficiently understand the physics behind a given systems response and

also to fully appreciate the scaling laws for that system, it is necessary to employ LEM

[1, 60]. The concept behind LEM is to reduce the complexity of an analytical or

numerical expression by breaking down a given distributed system into discrete elements

based on how the system interacts with energy [60, 126]. More specifically, the total

energy going into any given system is divided among three distinct types of interactions:

the storage of kinetic energy, the storage of potential energy, and the dissipation of

energy. All systems are composed of these three energy processes.

In some systems, known as distributed systems, the storage of kinetic and potential

energy occurs over a distributed region in space [90, 127]. To accurately represent these

systems mathematically requires a partial differential equation, as spatial and temporal

components are inherently coupled. Physically, the distribution occurs because the

wavelength is on the order of the physical system or smaller. At different points along









the period of the wave, differing amounts of energy are split among the energy storage

and dissipation mechanisms.

As the wavelength of the signal (e. g. acoustical, fluidic, mechanical, optical,

etc....) gets larger, to where it is significantly larger than the length scale of interest, very

little variation occurs in the distribution of energy as a function of space.

Mathematically, under these conditions, the spatial and temporal components can be

decoupled, allowing for the use of ordinary differential equations to solve the problem,

rather than partial differential equations. Physically, it means each energy storage or

dissipation mechanism can be equated to the energy stored or dissipated in an equivalent

element that is lumped to a chosen spatial location.

In various energy domains, the names for the types of lumped elements vary;

however, the concept and mathematics remain the same. In lumped mechanical systems,

kinetic energy is stored via mass, potential energy via the compliance of a spring (i.e.

inverse of stiffness), and dissipation of energy through the losses of a damper. Similarly,

in electrical systems, where lumped element systems are commonplace due to the

extremely long wavelengths of electrical signals, kinetic energy is stored in the magnetic

field of an inductor, while potential energy is stored in the charge across a capacitor.

Additionally, dissipation of energy is modeled via the resistor. Finally, in lumped

acoustical systems, the kinetic and potential energy is stored in an acoustical mass and

acoustical compliance, respectively, while dissipation of energy is represented by an

acoustic resistance. These elements are summarized in Table 2-1.

The commonplace nature of lumped elements in electrical systems has led, over the

years, to a large growth in graphical and analytical techniques to solve large networks of









interconnected elements. These networks are most commonly represented using

electrical circuit notation. A significant benefit of lumped element modeling is that in all

of the energy domains, the lumped elements can be represented using an equivalent

circuit form. Thus, masses can be represented using inductors, compliances using

capacitors, and dissipative components using resistors. Once the complete equivalent

circuit is constructed, standard circuit analysis techniques (e.g. Kirchoff s current and

voltage laws) can be applied to find the solution of interest.

Table 2-1: Equivalent lumped elements in several common energy domains.
Kinetic Energy Potential Energy Energy
Storage Storage Dissipation
Acoustical Acoustic Mass Acoustic Compliance Acoustic Resistor
[kg/m4] [n'/Pa] or [m'/N] [Mns]
Mechanical Mass [kg] Compliance [m/N] Frictional Damper
(pt mass) (spring) [m s]
Electrical Inductance [H] Capacitance [F] Resistance [0]


Whenever dealing with more than one lumped element, the concept of power flow

between the elements must be considered. If we define the power flow from element A

to element B as nAB and the reverse as nBA, then the net power flow from A to B is

"net = nAB IBA. (2.13)

Now, since each of the power flows must be greater than or equal to zero, then they can

each be written as the square of a real number, r, and r2 [128]. The net power flow can

then be rewritten as

net = r2 r =(r, +rz)(r, -r2). (2.14)

It is therefore seen that the net power flow can be written as the product of two real

numbers, which are referred to as conjugate power variables. Moreover, these quantities









are more specifically referred to as an effort, e, and a flow f, where the product e f is

the power. A table of conjugate power variables, divided into effort and flow, is given in

Table 2-2 for a number of energy domains.


Table 2-2: Conjugate power variables.
Energy Domain Effort
Mechanical translation Force, F
Fixed-axis rotation Torque, r
Electric circuits Voltage, V
Magnetic circuits MMF, 97N


Incompressible fluid flow
Thermal
(after Senturia, pg 105 [128])


Pressure, P
Temperature, T


Flow
Velocity, v
Angular velocity. o)
Current. I
Flux rate,
Volumetric flow, Q
Entropy flow rate, S


Helmholtz Resonator

Lumped Element Model

LEM was applied to the Helmholtz resonator in order to better understand the

system [1, 125]. A schematic diagram of a Helmholtz resonator is shown below in

Figure 2-3, where V is the cavity or bulb volume, / and S = ra2 are the length and

cross-sectional area of the neck, respectively, where a is the radius of the neck, P' is the


incident acoustic pressure, and P,' is the cavity acoustic pressure. Both PI'and P,' are

considered to be functions of the radian frequency, a).


neck
PH

R
7PP


cavity


Figure 2-3: Diagram showing side view of a Helmholtz resonator and its mechanical
equivalent of a mass-spring-damper system.









A conventional Helmholtz resonator can be lumped into three distinct elements.

The neck of the resonator constitutes a pipe through which frictional losses are incurred.

Additionally the air that is moving through the neck possesses a finite mass and thus

kinetic energy. Therefore the neck has both dissipative and inertial components. The air

in the cavity is compressible and stores potential energy, and is therefore modeled as a

compliance.

The acoustic compliance of the cavity and effective mass of the neck can be

derived from first principles. [129] As mass flows into the bulb, the volume, V, remains

constant, assuming the walls are rigid, and so the pressure must rise, by conservation of

mass.

dM dp(t) F kg7
=V Qp mass flow rate kg, (2.15)
dt dt T s

where Mis the mass in the bulb, po is the mean density of the air, and Q = u(ra2) is the

volumetric flow rate or volume velocity, where u is the velocity. If the disturbance is

harmonic and isentropic then


P2'= c02 = (2.16)
jco V

where co is the isentropic speed of sound of the medium and j = -1 .

The linearized momentum equation for a lossless medium is given by

Ou
o = -VP', (2.17)

where P' is the acoustic pressure. Assuming a linear pressure gradient yields


P' P' = p 1, (2.18)
at









where / is the length of the neck. Substituting for P2 yields the following equation.


S= 2 (2.19)
jcoV S



Factoring Q, this can be rewritten as


P = O-? I + jMO (2.20)


where the effective acoustic compliance, C, of the cavity is


C, = PoC2 (2.21)


and effective acoustic mass, MA,, of the air in the neck is given by


M N (2.22)
LS m

The notation for the lumped elements has been determined through the use of three

components. First, the primary variable name is determined by the element type. Next,

the first subscript represents, the energy domain in which the element has been defined,

where the subscript 'a' represents the acoustic energy domain, 'e' represents the electrical

enegy domain, and 'm' represents the mechanical energy domain. The final subscript,

which has been capitalized for easier reading, represents the actual structure that is

represented at least in part by the lumped element. In this instance, 'C' represents, the

cavity, while 'N' stands for the resonator neck. Later, a compliant diaphragm will be

introduced that will be represented by a 'D' subscript. The expressions above do not

account for any viscous damping effects that occur in the Helmholtz resonator neck. The









viscous damping represents a resistance, whose value can be approximated from pressure

driven, laminar pipe flow as


R 8;L/ kg (2.23)
S2S m4S

where / is the dynamic viscosity of the air. Furthermore, the viscous damping produces

a non-uniform axial velocity profile in the neck that ultimately leads to an additional

factor of 4/3 in the expression for the effective acoustic mass [130]. The corrected

effective mass is then given as

4pM o kg (2.24)
3S m

Additionally, the effective resistance and mass of the neck are, in fact, non-linear

and frequency dependent due to turbulence and entrance/exit effects, [131] however for

simplicity in modeling, these are not considered here.

The expression for cavity compliance given by Eq. (2.21) can be compared to an

approximation based on the exact expression for the impedance in a short closed tube [2].

The exact expression is given by


Z,,n = c cot(kl), (2.25)
na


where k =- is the wavenumber. Using a Maclaurin series expansion of the cotangent
CO

function yields


cot(ki)- 1 kl-1 (kl)3.... (2.26)
kl 3 45

For kl << 1, the impedance can be approximated by keeping only the first couple of terms

in the expansion, yielding










Poco klpoco poc2 poV
Z+ kl= p + P + pV (2.27)
jkla 2 3ra2 jcoV 3(sra2)


From this expression, we once again see that


C, C = a. (2.28)


We now also have an additional mass term, given by


M = poV [kg] (2.29)
3( ;ra2) m 4

which is equal to one-third the acoustic mass of the cavity. This correction term is small

for k <<1 but becomes more prominent as kl increases. At kl= 1, the correction term is

33.3% of the primary term, while at kl = 0.1 the correction term is only 0.33% of the

primary term.

Equivalent Circuit

To create an equivalent circuit model for the Helmholtz resonator, knowledge of

how to connect these lumped elements is needed. Connection rules between elements are

defined based on whether an effort-type variable or a flow-type variable is shared

between them [132]. Whenever an effort variable, such as force, voltage or pressure, is

shared between two or more elements, those elements are connected in parallel in the

equivalent circuit. Conversely, whenever a common flow (i.e., velocity, current, or

volume velocity) is shared between elements, those elements are connected in series.

These connection rules are used to obtain the equivalent circuit representation for the

Helmholtz resonator, as shown in Figure 2-4. The connection rules, as given, are

assuming that what is known as an impedance analogy is employed. If an admittance









analogy were used instead, then the connection rules would be reversed from what is

described above.

RN My


PI P2CLL

=-2
Figure 2-4: Equivalent circuit representation of a Helmholtz resonator.

The frequency response function P2/P represents the pressure amplification of the

resonator. It is the ratio of cavity pressure to incident pressure, and is given by

1
P2 SC'C
P= sac (2.30)
P 1
R RN + sM, +
sC,

where s = jco.
From an analysis of the above circuit, a single resonant peak is expected in this

frequency response function, when the sum of the reactances is zero, and is given by

1
fre. = Ic [Hz]. (2.31)


At the resonant frequency, the pressure amplification reaches a value of

PA. = (2;fref). (2.32)
RaN

This is shown in Figure 2-5, for an arbitrary Helmholtz resonator having a neck

length and diameter of 3.18 mm and 4.72 mm, respectively, and a cavity volume of 1950

mm3. The single peak in the pressure amplification frequency response represents the

single degree of freedom present in the system.










102

(a) 1






10-
10 -1----- "----------------

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Frequency [Hz]

(b)

-50 -
CD


C -1 50

-200
0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Frequency [Hz]

Figure 2-5: Theoretical pressure amplification of a conventional Helmholtz resonator. (a)
magnitude (b) phase.

Piezoelectric Composite Plate

On the back wall of the Helmholtz resonator, a piezoelectric diaphragm can be

placed such that the acoustic pressure in the cavity can be exploited to cause a deflection

of the plate and ultimately an electrical signal via piezoelectric transduction. The

piezoelectric composite plate, consisting of a circular layer of silicon and an annular ring

of PZT, is shown in cross-section in Figure 2-6, whereE is theYoung's modulus, v is

Poisson's ratio and p is the density of the silicon and PZT as indicated by the subscript.

Additionally, t is the thickness, and R2 is the outer radius of the silicon, while for the

PZT, tp is the thickness, e is the relative permittivity, d,3 is the transduction coefficient


for a voltage applied across the piezoelectric causing a displacement in the radial









direction, and rp is the residual stress. A region of high stress exists in a circular plate

near the clamped boundary during deflection of the plate. By placing the PZT in this

region, the electromechanical coupling is increased due to the stress concentration. In

addition, the annular structure facilitates the connection of electrodes and bond pads, as

the metal lines can be run from the bond pads to the outer radius of the annular structure

and therefore do not have to be placed across the surface of the diaphragm.


Es ots
Figure 2-6: Cross-sectional and 3-D schematic of piezoelectric composite circular plate.
(not to scale)

Piezoelectricity

Mathematically, the linear piezoelectric effect is expressed as [133]

S, = s Tkd +dkEk (2.33)


EP v P, t r F,~









D = dklTkl + lEk, (2.34)

where S, is the mechanical strain, s'1 is the elastic compliance (m2/N) at constant

electric field, Tk is the mechanical stress (N/m2), and dj is the piezoelectric

coefficient (C / N or m / V), Dr is the electric displacement (C / m2), k is the electric

permittivity at constant stress (F/m), and Ek is the electric field (V/m) [133]. These

equations can also be rewritten using a reduced notation convention, by replacing ij or

kl by p or q where i, j, k, and I can only have values of 1,2,or 3 while p and q can

have any value between 1 and 6. Using this notation, the resulting equations are

S, = fT +dEk (2.35)

and

D = dqTq + Ek. (2.36)

The subscripts in the reduced notation equations refer to the component of each variable

in a specified direction as defined by Figure 2-7. For example, d31 is the piezoelectric

coefficient relating electrical displacement in the '3' direction (z-axis) to a mechanical

stress in the '1' direction (x-axis). From Eq. (2.33), it is also seen to relate a mechanical

strain in the '3' direction (x-axis) with an electric field in the '1' direction.

3 (z)

6

4 A O 2 (y)
~-- 5
1(x)


Figure 2-7: Notation of axes used in piezoelectric transduction.









One class of materials, ferroelectrics, exhibits the property that the direction of

their polar axis can be changed through application of an external electric field [5, 6,

134]. The polar axis is the direction along which a polar molecule exhibits an internal

electric field. When the external electric field is then removed, the polar axis remains in

an altered direction. This process that causes a long term rotation of the polar axis is

commonly referred to as "poling". All ferroelectric materials are also piezoelectric,

however they are not naturally piezoelectrically active, as the crystal structure is

centrosymmetric, as in Figure 2-8a. PZT is a typical example of a ferroelectric material,

and has a perovskite crystal structure as shown in Figure 2-8. By applying an external

electric field to rotate the polar axis, a non-centrosymmetric crystal structure is created

and the ferroelectric material becomes piezoelectrically active, as in Figure 2-8b.

Additionally, as a piezoelectrically active material is brought close to its Curie

temperature, it loses its piezoelectric properties as the dipoles relax to their unpoled state.




I I I I
I I I
- ---- -- - T - - -





SIII

-- --- --
4 ------------ 4 -i3





a) Zr/Ti Pb 0 0 b) @ Zr/Ti Pb 0 0

Figure 2-8: Idealized perovskite crystal structure for PZT. a) centrosymmetric structure
prior to poling. b)non-centrosymmetric structure after poling.









A piezoelectric ceramic, such as PZT, consists of individual domains. The

polarization within each domain is in the same direction; however, the polarization

direction varies randomly from domain to domain, leading to a net polarization of zero.

This is illustrated in Figure 2-9(a), for the unpoled material. After poling the material

with a sufficiently high electric field, the dipoles are rotated so that the polarization of

each domain is in the same general direction, as defined by the poling direction, as is

illustrated in Figure 2-9(b). Raising the temperature during poling enhances the

polarization and resulting piezoelectric properties, as the dipoles rotate more readily at

higher temperatures [6].


















(a) (b)

Figure 2-9: Schematic of the poling process: a) An unpoled piezoelectric material, where
the polarization in each domain is randomly oriented. B) The same material,
after poling, where the domains are now polarized mostly along the direction
of the applied field. (after Setter, pg 6 [6])

An additional property of ferroelectric materials is the double-valued nature of their

response to an electrical excitation, resulting in a hysteretic behavior. Hysteresis is a

property of systems that do not react instantly to an applied force and may not return to






40


their original state when the force is removed. In ferroelectrics, when an electric field via

a voltage is applied, a polarization is incurred. Upon removal of the voltage, the

polarization decreases but does not return to zero. The result is a remanent polarization,

PI. If the voltage is swept up and down, the result will be what is known as a hysteresis

loop, shown in Figure 2-10.


Figure 2-10: Polarization vs. electric field hysteresis loop (after Xu, pg 10 [134]).

The graph is displayed here as polarization, P, typically given in [I C/cm2],

versus applied electric field, E, although other variations do exist. There are four figures

of merit shown in the illustration. Pm is the maximum polarization measured, while PI is









the remanent polarization (i.e. the polarization which remains when the field is returned

to zero). Furthermore, PI is known as the spontaneous polarization and is defined as the

straight line extrapolation of the line defined by the upper saturation region. Finally, the

coercive field, Ec, represents the magnitude of the field required to cancel out the

remanent polarization.

Composite Plate Modeling

In order to accurately model this structure, the material properties of each of the

layers were considered. These include the mechanical properties of the layers such as

Young's Modulus, E, and Poisson's ratio, v and the dielectric constant, e, and

piezoelectric coefficient, d31, of the PZT. The subscript p or s indicates the layer as

PZT or silicon respectively. Furthermore, the geometry of the structure is accounted for

in the values for inner PZT radius, R,, outer PZT radius, R2, PZT thickness, t and

silicon thickness, t.

Lumped element model

A pressure applied to the plate creates a deflection of the plate, resulting in a stress

in the piezoelectric layer. The stress deforms the piezoelectric layer, creating an

electrical charge on the electrodes, thus generating a voltage across the piezoelectric

layer. Although the stress and charge are distributed over a finite region of the plate, it is

useful to first look at one-dimensional (1-D) piezoelectric transduction. This 1-D

analysis can then be extended to incorporate effective lumped element values that are

calculated from the actual distributed case.

In the 1-D piezoelectric transduction, the piezoelectric material displaces

longitudinally due to the application of a force, F, and/or a voltage, V, applied in the









thickness direction. Additionally, a charge, q, is generated by the application of the

same force and/or voltage. The constitutive equations, Eq. (2.33) and Eq. (2.34), can be

extended to this situation by modifying their form.

When looking at the resulting strain and electric displacement in the '3' direction,

for an applied stress and electric field in the same direction, the equations reduce to

S33 = s3T+d33E (2.37)

and

D33 =d33T + 3T3E, (2.38)

Then this set of equations can be converted by multiplying both sides by the thickness of

the piezoelectric material.

t(33 =sT+d33E) (2.39)
(2.39)
tS = sE tT+d33tE

Now, the mechanical compliance of a material under compression in the thickness

direction can be defined as

s3jt
C st (2.40)
SA'

where t is the thickness, and A is the area over which the force is applied. Additionally,

for a constant electric field, E,


E = (2.41)
t

Substituting (2.40) into second part of (2.39) yields,

x= CAT+d33V (2.42)

Finally, the stress, T, when applied over the area, A, can be equated to a force, F, given

by









F = TA, (2.43)

leading to the final equation given by

x= CF+d33V. (2.44)

A similar procedure can then be performed on Eq. (2.38), by multiplying both sides

by the area, and making similar substitutions as before, giving

A(D33 =33T T3E), (2.45)

q= d33F+ 3T3AE, (2.46)

E3T A
q=d33F +33 (2.47)
t

so

q= d33F + CV. (2.48)





Thus a pair of equations can be written describing the 1-D piezoelectric transduction, and

is given by

x Cs, d3I F
[)]= [C-, IF (2.49)
q d33 C_ V

In this pair of equations, C,,, the mechanical compliance when a short is placed across

the electrodes of the piezoelectric, is defined as


Cm (2.50)
F

while Ce the electrical capacitance when the piezoelectric is free to move, is defined by


Eq o 33 AP
C = F- _33Ap (2.51)
Vtp









where q is the resulting charge from the applied voltage V, E33 is the electrical

permittivity in the thickness direction, Ap is the area of the piezoelectric, and tp is the

thickness of the piezoelectric. Furthermore, d33 is the piezoelectric coefficient relating

the displacement in the thickness direction for an applied voltage in the same direction,

when no mechanical force is applied across the piezoelectric, defined by


d33 F= (2.52)


For the case of the piezoelectric composite plate that is presented here, though,

there is a distributed, nonuniform deflection across the plate. In order to apply the 1-D

model developed above to this situation, it is necessary to lump the actual distributed

deflection to a single point and compute effective values by equating the energy in the

distributed system to the energy in a corresponding lumped component. The 1-D

electromechanical transduction described above can then be extended to relate the

acoustical and electrical energy domains by integrating over the areas and switching to

acoustic conjugate power values, as well as incorporating time-harmonic signals [64].

The time-harmonic, electroacoustical equivalent of Eq. (2.49) is given by


SL] d=oC jodA][V] [ AS] (2.53)
Q jod, jcoCa P M1 S

whereP is the acoustic pressure, Q is the volume velocity of the plate, I is the electrical

current. Furthermore, dA is the piezoelectric coefficient relating the volumetric

displacement to the applied voltage, when pressure equals zero, and defined by

R2
fw(r)\ 2,rrdr
AVol PJw ( 3
d, = (2.54)
SP->O V v












where w (r) is the deflection as a function of the radial position, R2 is the radius of the

plate and AVol is the volumetric displacement defined by

R2
AVol= 27rrw(r)dr [m3. (2.55)
0

Additionally, Ca, the acoustical compliance when a short circuit is placed across the

piezoelectric, is defined by

R2
w (r) 2Irrdr
AVol o m
CI o V "- o .Pm ,a(2.56)


From two-port network theory, a generalized electro-acoustic, reciprocal, two-port

network can be written as

I Y G V
[ f G (2.57)
Q G Y, P_

where YeJ is the electrical admittance when the acoustic terminal is free to move (i.e no

pressure exists, P = 0), Y, is the acoustical admittance when the electrical terminal is

shorted (i.e. no voltage exists, V = 0), and G is the electro-acoustic transduction

admittance, given both by the ratio of current, I, to pressure, P, when V = 0, and the

ratio of volume-velocity, Q, to voltage, V, when P = 0. Comparing Eq. (2.53) and Eq.

(2.57), it can be seen that

G= jcodA, (2.58)


Yef = j 0CeC,


(2.59)









and

Y = jo)C". (2.60)

Equivalent circuit

Using Eq. (2.57), an equivalent electro-acoustic circuit can be drawn, as shown in

Figure 2-11, where Ceb, the electrical capacitance when the plate is blocked from moving

is given by

Ceb C (1- k) [F], (2.61)

the transduction factor, q, is given as

-G -jcdA -d, PaJ
~= (2.62)
Y, jmC C V

and the coupling factor, k, is given by


k = = (2.63)


Physically, the coupling factor, k, represents the fraction of energy that is coupled

between the acoustical and electrical energy domains. It is seen here to be a function of

the piezoelectric coefficient, dA, which represents coupled energy and the two elements

which store potential energy associated with the transduction, Cef and Ca,. The coupling

factor is therefore related to the ratio of the coupled energy to the stored energy.


Cr v



Figure 2-11: Electro-acoustic equivalent circuit representation with Cb, Ca, and .








An alternative equivalent circuit can also be drawn, as shown in Figure 2-12. The

alternative circuit is simply another representation for the same physical process.

1:+' Co,

P V
T-_

Figure 2-12: Electro-acoustic equivalent circuit representation with Ce,, C"o and q'.

In Figure 2-12, C,, represents the acoustic compliance when an open-circuit is

placed across the piezoelectric, and is given by,


Cao =C( 1-k2) [ (2.64)

Additionally, the transduction factor, 0', is given by

-G -jod -dA V (2.65)
Yf j oCef Cef Pa

The two circuits above represent the transduction under static conditions, thus they

do not take into account the dynamics of the system (i.e. mass). At frequencies greater

than zero, a mass must be added to the equivalent circuit. As the mass stores kinetic

energy, it is represented by its electrical equivalent, which is an inductor. Furthermore,

this mass must be in series with the compliance, as both experience the same motion (i.e

displacement or velocity). The acoustic mass, M,, is determined by equating the lumped

kinetic energy of a point mass moving with the center velocity to the total kinetic energy

of the vibrating diaphragm and is therefore given by ,

R2r, ( rAV 2 (2.
M = 27r pA \rdr (2.66)
0 AV=0









where p, is the areal density of the piezoelectric composite plate defined by,


PA =J pdz [kg/m2], (2.67)

where p is the density of the corresponding layer.

In addition to the above mentioned elements, it is also necessary to include an

additional element that represents the dielectric losses that are incurred during

transduction, yielding the circuit shown in Figure 2-13. Generally this is expressed

through a term known as the loss tangent, also known as the dielectric loss factor.

Represented by the ratio of the parallel reactance to the parallel resistance in the

dielectric, the loss tangent is formally given by

X 1
tan = P (2.68)
R, 2zrfC ,R

where R is a resistor in parallel with a capacitance, Cp, and f is the frequency at

which Rp and Cp were measured. The loss tangent is also equal to the inverse of the

quality factor, Q.


"a Cas

+ [ +
P =


ZID
Figure 2-13: Dynamic electro-acoustic equivalent circuit.

In this circuit, MaD represents the acoustic mass of the diaphragm. The input

impedance from the acoustical side is then given by

1 R
Z= = sMD + + 2 P (2.69)
sC.s 1+ SCebR









By voltage division we get


02 Rp 02 Rp
R R
1+ sCRp 1+ sCR
Ve = P b= P (2.70)
Z 1 R
SsM^ + + -p
sCD 1 + SCebR,

which, after rearranging yields,

V 1
S= sCSR 2 (2.71)
P s MaDC sCebR, +S MDCs + sRp (C + ) +1


This is the general expression for the open-circuit sensitivity of the circular composite

piezoelectric plate. From a physical standpoint it is also useful to look at certain

conditions that allow for additional insight. The presence of the dielectric loss resistance,

Rp, complicates some of the underlying behaviors of the device, by adding an additional

R-C time constant. Looking at a situation in which this parallel loss resistance is infinite

(i.e. R -> oo), the above expression can be reduced to

V C 1
-== (2.72)
S^XsMC "+1
P Cb S2M oDC",-- 2 CIS +
Ceb

Several important things can be gleamed from this expression. First of all, the

equation describes the behavior of a second order system, with a primary resonance

occurring when the denominator goes to zero at



Sre^ Cb (2.73)
r MaD Cos

From this expression, the resonance is seen to depend on the ratio 2C ,/Ce in

addition to the standard dependence on the product MADC, If q = 0, this reduces to a









simple plate. For 0 # 0, electrical energy is stored across the piezoelectric, resulting in a

stiffer device possessing a larger o), -

The full expression that includes the dielectric loss, given by Eq. (2.71), is the

general expression for the sensitivity of the circular composite piezoelectric plate. It can

be further simplified for two important cases:

1. Well below the short-circuit mechanical resonant frequency (compliance
dominated)
2. At the short-circuit mechanical resonant frequency (resistance dominated)


Case 1: Well below the short-circuit mechanical resonant frequency

When a short is placed across Ceb, it is equivalent to assigning R -> 0 Physically,

this implies that no electrical energy storage is allowed, as the voltage across the

capacitor, Cb, must be zero. The expression for the short-circuit mechanical resonant

frequency is given by


l 1 (2.74)


which is the same resonant frequency as described above when 0 = 0. Note that both

cases describe a situation where no electrical energy is stored. The general expression for

sensitivity can then be rewritten in terms of o)re, as

V 1
P = j-q-C,%RP 2 2 (2.75)
oK j CbR Cw o + + j (Ceb 2C)
W res, b ressc

Well below resonance, we have










jo 1. (2.76)


This term can then be dropped from the expression, yielding
V C 1
=jco (2.77)
P Ceb 2
Ceb RpCeb

To simplify this further, the transduction factor, q, can be written in terms of the

effective piezoelectric modulus, dA, by recalling Eq. (2.62), yielding

V -jcodA C, 1
P C, Ceb dA )2C'
jeO) 1+ a +
C, Ceb RCeb (278)
(2.78)
-dA 1
Ceb 1+ dA21+ 1
C C,Ceb jo RpCeb

Furthermore, the blocked electrical capacitance, Ceb, can be written in terms of the free

electrical capacitance, Ce and the coupling coefficient, k, using Eq. (2.61) and Eq.

(2.63) to yield the an expression for sensitivity, given by

V -d A I

1
P Ceb 1+ dA2 + I


A 1d(2.79)
I C 1CJ j^
Ceb d 1
SCasC 1 k2 j R Cef 1- k2
-d 1


(l1- k2 jR Ce1 -k2)

which can be simplified further to yield














-d-
C 1 1
eb 2 +





eb (1- k2) jR
-dA
Cef +
j coRp


(2.80)


1+-
j ) CfR

Now under the condition of a high value for the loss resistor, Rp, Eq. (2.80) can

be reduced approximately to


V -dA
P Cef

This corresponds to the alternative transduction coefficient, 0', given by

-d,
ci
Cef


(2.81)


(2.82)


and thus under the conditions of a high loss resistor and operation well below a S I, the

sensitivity can be given by

V


P

Case 2: At the short-circuit mechanical resonant frequency

The general expression for sensitivity is once again given by

V 1
S= joC, RP 2 2C,
j j CebR + + j p ( Ceb +Ca)+
)resc )ress


(2.83)


(2.84)









Now, since we are only considering the system at the short-circuit resonance, where

(0 = ),,se we have



o = -1, (2.85)


yielding an exact expression for the sensitivity as

V 1 1
V= j1o CRp1 (2.86)
P -jc CebRP -1+ jR(Ceb +2CaL)+1


We therefore find that the sensitivity at resonance reduces to

V 1 #'
V -I (2.87)
P q k2 '

as

k2 = '. (2.88)

This can be understood from the equivalent circuit of Figure 2-13. At a = o)reS',

the impedance of the mass is canceled by the impedance of the compliance, permitting a

direct transduction of energy from the acoustic to electrical energy domain. Note that

0),o exists whether or not R, = 0, therefore Eq. (2.86) is always valid whenever

o = )se In other words, Eq. (2.86) holds whenever the operating frequency coincides

with the frequency that is defined by the short-circuit resonance, but makes no

requirement for an actual short-circuit load.

As the load across Ceb is increased, the resonance frequency will shift towards the

open-circuit resonance. Recalling Eq. (2.64), the open and short-circuit acoustic

compliances are related by the coupling factor and define the limits of the resonant

frequency as all resistive loads fall somewhere between open and short-circuit. The









larger the coupling factor, then the larger the range between open and short-circuit

resonant frequencies.


General expression vs. simplified cases

Figure 2-14 plots the open-circuit and near short-circuit sensitivity versus

frequency overlaid with the values of 0' and 1/0 The material properties used in

calculating these results are shown in Table 2-3 and Table 2-4. Notice that below

resonance, the general open-circuit sensitivity expression asymptotes to the value of q'.

Additionally, notice that the general expression reaches the value of 1/0 at a frequency

below where the peak amplitude occurs. This is because 1/0 is the simplified expression

for the sensitivity at the short-circuit mechanical resonant frequency, whereas the general

expression represents the open-circuit frequency response. Thus the peak amplitude

occurs at the open-circuit resonant frequency. The frequency at which the general

expression for sensitivity reaches the value of 1/\ coincides with the short-circuit

resonant frequency, as expected. Also note that this short-circuit sensitivity calculation

was performed with an almost-short circuit condition, primarily to illustrate the effect on

the resonant frequency. To achieve the almost-short circuit condition a resistive load of

1 Q was used as the value is much less than the impedance due to Ceb. One final feature

of the frequency response is the low frequency rolloff that is visible in the open circuit

case. This rolloff is due to the dielectric loss resistor, Rp, and the resulting time constant

due to the combination of that resistor and the electrical capacitance, as evidenced in Eq.

(2.80) for the low frequency regime.










Table 2-3: General material properties used in modeling.


Material E[GPa] v p[

Silicon 150 0.27
Platinum 170 0.38
TiO2 283 0.28
[135, 136]
Table 2-4: PZT properties used in modeling.
EA [GPa] v, p [kg/m3] d

30 0.3 7600 1000
[9, 10, 16, 31]
0
10 1

-1
10-
+ Near short circuit frequent

-2 Open circuit frequency re
10
S10
1 >o
a^ ^^_B-^nnn^


kg/nm3
2300
21440
2150



S[pm/V]
-50


cy response
sponse


10

-


10


10
0 1 2 3 4
10 10 10 10 10
Frequency [Hz]
Figure 2-14: Sensitivity vs. frequency for a piezoelectric composite circular plate.


Effects of including the radiation impedance

Since the clamped circular plate is vibrating in a medium, the radiation impedance

of the plate must be taken into account, and consists of a radiation mass and radiation

resistance. The radiation mass accounts for the inertial mass of the fluid that is vibrating

in unison with the plate, while the radiation resistance accounts for the acoustic radiation


t[nm]
3
0.170
0.100


tan 8

0.02


uo [MPa]
30


e


am,


---------------


5;









of energy away from the plate. In addition to the radiation resistance, energy is lost via

structural radiation to the supports. The radiation resistance and radiation mass is

determined to first-order by approximating the backplate as a piston in an infinite baffle,

as given by Blackstock [2] in terms of specific acoustic impedance, Z as

2J, (2ka) 2K, (2ka) ,
ZP =Pc 2J 2ka+ = 2K2ka pc[R(2ka) +jX (2ka)], (2.89)
2ka 2ka

where R1 (2ka) and X1 (2ka) indicate that they are functions of 2ka, a = R' is the

radius of the piston, J, is a Bessel function of the first kind of order one, and K, is a first-

order Struve function. The Maclaurin expansions of Eq. (2.89) are also given by [2]as

(ka)2 (ka)4 (ka)6
R,-- -+ -- (2.90)
1.2 1.22 3 1.22.32 4

and

4 2ka (2ka)+ 2 .... (2.91)
X, (2.91)2 2
I r 3 3 -5 3 -5 .7

For small values of ka, where ka << 1, the resistance and reactance can be approximated

by keeping only the first terms of Eq. (2.90) and Eq. (2.91).

The radiation impedance as given in Eq. (2.89) is in terms of specific acoustic

impedance. This is converted to an acoustic impedance via the effective area, Aff, of the

piezoelectric plate. As the circular composite diaphragm does not deflect uniformly over

the entire surface (i.e. the deflection is a function of radial distance), the volume

displaced by the deflection of the diaphragm is less than that of a circular piston of the

same area. The effective area, Af, is therefore defined to represent the equivalent area

that a circular piston with uniform deflection would need to have to create the same








volumetric displacement as the composite diaphragm. This is necessitated by the need to

maintain continuity of volume velocity across the interface between the mechanical and

acoustical domains. It can be calculated by integrating the distributed deflection of the

diaphragm over the entire surface and then dividing by the center deflection of the

diaphragm. By using the center deflection as the reference value, the distributed

deflection is then considered to be lumped to the center. For an axisymmetric deflection,

such as would occur for the fundamental mode, the effective area is given by

w (r)27rrdr
Aw() [m2] (2.92)


The radiation resistance can then be approximated, for low values of ka as

S(ka) 2 P [ kg]
RaDrad 2Af ,ka << 1 (2.93)

while the radiation mass is approximated as

M 8ra k c ka < 1 (2.94)


These two elements add in series to create a radiation impedance, Zra, that is defined by

ZaDad = RaD +sM d (2.95)

The equivalent circuit of the composite plate including the radiation impedance is shown

in Figure 2-15.

Q R d aDrad M8a C as I
+ +
P V
.e TCe=-

Figure 2-15: Equivalent circuit of piezo-composite plate including the radiation
impedance.






58


When the radiation impedance of the diaphragm is added to the equivalent circuit,

the analysis becomes only slightly more complex. The acoustical input impedance, Zn ,

defined as P/Q, can be represented in terms of the equivalent circuit parameters, and is

now given by

1 R
Z = sMaD +sM ,,d +R -d + -+2 P (2.96)
C a Drad a 1+ sC R


Then, the general expression for sensitivity is now given via the voltage divider method

as




S R RP
1+ CebRp 1+sCR
VP = P = P (2.97)
Z sMaD + SMD ad + RDrad + 1 + 2 Rp
sC l+sCebRp

or



1r -_ -SCa e R- (2.98)
P 1 R
sMaD + SMaDad + RaDad + p + 02 R
sC 1+ sCebR,

From Figure 2-15, it can be seen that the radiation mass, Marad, adds directly to the

acoustical mass of the plate, Ma>, as is evidenced in Eq. (2.97), while Rrad provides

damping to this second-order system. Note that here, the acoustical resistance, Rra ,

damps the resonance, while the electrical resistance, R leads to a low frequency rolloff

The different effects arise because Rrad is in series with the reactive elements, C and









MaD, while Rp is in parallel with the capacitance, Cb. The general expression for the

undamped resonance frequency is now given by


S 2 ( +1
Sreb (2.99)
(aD + MaDrad )Cas

When a short is placed across Ce, it is effectively removed. This is

mathematically equivalent to assigning Ceb-> o as the impedance of a capacitor is

inversely proportional to the capacitance. The short-circuit resonance frequency then

reduces to


Sres (2.100)
J)esse D + MaDrad


Upon comparison with Eq. (2.74), it can be seen that the short-circuit resonance

frequency has now been shifted downward by the radiation mass, as is also the case with

the open-circuit resonance frequency. At the radial frequency of the short-circuit

resonance, the short-circuit input impedance reduces to

Z,n =RDad, (2.101)

because Ceb and R are effectively removed by the short-circuit. The volume

velocity, Q, is then given by

P
Q =- (2.102)
RaDrad

Via the piezoelectric transduction, the current, I, in the piezoelectric material, is then

given by

OP
I = Q = (2.103)
RaDrad









The output voltage in this case, however, remains at zero due to the short circuit placed

across the output capacitance. To achieve real power at the output, a finite resistive load

is necessary. Adding a resistive load, Rload, across the output yields the equivalent circuit

shown in Figure 2-16.

R'd MSld I Cas :l I
+ +

P CT eb load

Figure 2-16: Equivalent circuit with resistive load.

The input impedance, Z4n, is then given by

1 + loadp
Z1 = S (M +M + +d) + RI + + 2 RloadR (2.104)
SC Rload Rp +RloadRpCeb

Now, if we choose Rload such that Road <<1/SCeb and Road << R, then the parallel

combination of these three elements can be approximated with just Rloa. In addition, the

resonant frequency will be very close to the short-circuit resonant frequency. At this

frequency, the input impedance reduces to

Z,, = RaDrad + 2Rload, (2.105)

and the volume velocity, Q, is given by

P P
Q =- (2.106)
Z, RaDrad + 2Rload

Through piezoelectric transduction, a current, I, is created in the piezoelectric given by


I = Q = PO (2.107)
RaDrad + 2Rload









As the resistance, Rlod, is much less than 1/sCeb most of the current goes through it,

leading to a voltage drop given by


V = IRoad = P ad (2.108)
RaDrad + 2Rload

Thus the power absorbed by the load resistance is purely real and is given by

P2 2ROd
Re{)}= Re{IV} = 2 Rload (2.109)
(RaDrad +2 Rload

An optimal solution to this equation is found by setting

d = 0 = _-p22 (-R d (2.110)
dRIoad (Drad + 2Rload)

Solving for the optimal load resistance yields

R
Road a- d, (2.111)


which is just the impedance matching condition at the interface. Note that is for the

special case of Roa << 1/sCb Rloa < R, and operation at resonance.

Equivalent circuit parameters of piezoelectric composite circular plates

In order to obtain the equivalent circuit parameters, an analytical model was

developed for the piezoelectric composite circular plate by Wang et al. [62, 63]. Using

this approach, analytical modeling was accomplished by dividing the problem of Figure

2-6 into two portions, an inner circular plate, surrounded by an annular composite ring

with matching boundary conditions at the interface, as shown in Figure 2-17. The

boundary conditions consist of equal moments and forces at the interface as well as equal

slope and transverse displacement. After solving for the deflection in each region, the

deflection equation for each region can then be combined [62, 63].



























Outer Region






; Ji-ni
I I I I





Region

Figure 2-17: Conceptual 3-D and cross-sectional schematic of the circular composite
plate. (Not to scale.)

The deflection equation can then be utilized to determine the potential and kinetic

energy stored in the plate, leading to expressions for the acoustic compliance and mass of

the composite plate. Similarly, expressions can be found for the electro-acoustic

transduction coefficient and the blocked electrical capacitance. Using the parameters

shown in Table 2-3 and Table 2-4, plots were obtained for the lumped elements as a

function of both R1,/R and tp /t As many of the material parameters for PZT are

highly dependent on actual processing conditions and techniques, typical values were

chosen as a 'best guess' estimate [8, 12-14, 16-48]. The effective acoustic short-circuit









compliance and mass are functions of both R1/IR and t/t and are shown in Figure 2-18

and Figure 2-19, respectively.

-12
x 10
3.5


3


2.5


S2-
E

O 1.5 t =0.6 pm
P

1-.5

t =3.0 am
0.5


0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
R /R2
R1R2

Figure 2-18: Effective acoustic short-circuit compliance as a function of R1 /R and
tp /t For these calculations, t = 3j/m and R, = 2mm, while the piezoelectric
layer has thicknesses of tP = 0.6, 1.2, 1.8, 2.4, 3.0 [pjm].

The acoustic compliance is found to increase with increasing R, /P and decrease

with increasing tp /t however, the acoustic mass is found to generally decrease with

R/R2t and increase with t /tI It is also useful to look at the physical limits and effects

of R /R As R1R / 0, the piezoelectric ring covers the entire surface of the

diaphragm and the compliance is at a relative minimum for a given thickness, while the

mass is at relative maximum. Furthermore, note that the acoustic mass is most sensitive









to changes in the piezoelectric layer thickness under this condition. Meanwhile, as

R, IR2 1, the piezoelectric ring would have infinitesimal width and so would be

essentially nonexistent. Under this condition, the piezoelectric layer thickness has no

effect on the compliance, as would be expected.

3500


3000


2500- tp=3.0 p m


2000 -


1500 /


1000 t =0.6 tm


500
0 0.2 0.4 0.6 0.8 1
R1/R2

Figure 2-19: Effective acoustic mass as a function of R1/R2 and tt, For these
calculations, t = 3/um and R, = 2mm, while the piezoelectric layer has
thicknesses of tp = 0.6, 1.2, 1.8, 2.4, 3.0 [pum].

Shown in Figure 2-20 is a graph of the resonant frequency as a function of both

R, /R and t /t From this figure, it can be seen that the resonant frequency increases

as the thickness of the piezoelectric layer increases or generally as the inner radius of the

piezoelectric layer decreases. The resonant frequency trend with respect to thickness

arises because the effective acoustic mass is found to increase with the thickness while

the effective acoustic compliance decreased, but by a much larger amount. Again, it









should be noted that at the limit of R/R2 -> 1, the resonant frequency is no longer

affected by the thickness of the piezoelectric layer. Also, for a given tp, there is a

maximum resonant frequency at R /R2 -0.4.

14000


12000
/12000- t p=3.0 \tm


N 10000


8000

0 t =0.6 am
On P
( 6000


4000-


2000
0 0.2 0.4 0.6 0.8 1
R /R2
R1R2
Figure 2-20: Short-circuit resonant frequency as a function of R,/R2 and tp/t For
these calculations, t = 3/um and R, = 2mm, while the piezoelectric layer has
thicknesses of tp = 0.6, 1.2, 1.8, 2.4, 3.0 [pum].

The electro-acoustic transduction coefficient, q, was also found as function of these

relative dimensions, and is shown in Figure 2-21. The transduction coefficient represents

the conversion of the voltage across the piezoelectric to the acoustic pressure produced

by the resulting motion of the composite plate. It should be noted that the transduction

coefficient is negative, implying a 1800 phase shift between pressure and voltage, as









computed relative to an assumed poling direction. As can be seen in Figure 2-21, a

maximal magnitude occurs when R1/R2 is around 0.45.




-1


-2 /
t =0.6 tm

(-3 -3
t =3.0 \tm

-4


-5\


-6
0 0.2 0.4 0.6 0.8 1
R1/R2
Figure 2-21: Electro-acoustic transduction coefficient as a function of R, /R and t /t .
For these calculations, t = 3/um and R, = 2mm, while the piezoelectric layer
has thicknesses of tp = 0.6, 1.2, 1.8, 2.4, 3.0 [jum].

Additionally, the magnitude of the transduction coefficient increases with

increasing piezoelectric thickness. Looking at the limit as R, /R -> 1, the transduction is

seen to decrease to zero, as would be expected of a structure with no piezoelectric

material. Furthermore, as RR, -> 0 and the piezoelectric material covers the entire

diaphragm, the transduction factor is seen to go to zero as well. Physically, this results

from cancellations between different regions of the diaphragm that are undergoing

opposite polarities of stress, such that the net electric displacement (charge) is zero.









Another parameter of interest is the coupling coefficient, k. It is defined as the

ratio of energy converted by the transducer to the energy supplied to the transducer, thus

providing a measure of the coupling. For the purposes of harvesting energy, this

parameter is more important than the transduction coefficient as the primary interest is in

maximizing the coupled energy, rather than maximizing only the output voltage. A plot

of k as a function of R1/ R and tpt, is shown in Figure 2-22.

0.014


0.012


0.01 t =3.0 pm


0.008


0.006


0.004
t =0.6 atm

0.002


0
0 0.2 0.4 0.6 0.8 1
R1/R2
Figure 2-22: Electromechanical coupling coefficient, k, as a function of R/R2 and
tp/t, For these calculations, t = 3pm and RP = 2mm, while the
piezoelectric layer has thicknesses of tp = 0.6, 1.2, 1.8, 2.4, 3.0 [jim].

From the figure, the maximum k of 0.013 is obtained for an R,/R2 ratio of 0.95

and a tp/t ratio of unity. Physically, the increasing trend in k as R1/RP increases, arises

from the stress distribution within the diaphragm. There is a stress concentration near the

clamped boundary of the diaphragm that provides for a high level of coupling. By









concentrating the piezoelectric material in this high stress region, the averaged coupling

factor over the ring will be higher than a piezoelectric ring that is spread over a larger

area.

Acoustic Energy Harvester Dynamic Behavior

When the piezoelectric diaphragm is mounted in the wall of a Helmholtz resonator,

the equivalent circuit for the diaphragm (Figure 2-16) is combined with that of the

Helmholtz resonator (Figure 2-4). By using the same lumped element connection rules

previously described, the complete circuit can be obtained, as shown in Figure 2-23.

S RaN MaN RaDrad MaDrad Ma CaD I


? I Iload-

Figure 2-23: Equivalent circuit for acoustic energy harvester with resistive load.

Notice that the equivalent circuit for the diaphragm is in parallel with the cavity

compliance, C, This occurs because the pressure in the cavity is responsible both for

compressing the air in the cavity as well as deflecting the diaphragm, i.e. the cavity and

the diaphragm both see the same pressure. Additionally, it is assumed that Rloa << R<

and therefore the parallel combination can be approximated as Road HR = Rlo .

Acoustical Input Behavior

An expression for the acoustical input impedance can be obtained from the

equivalent circuit and is given by


S(MD +MaDrad ) + RaDrad + R
Z = R sM + -C SCaCaD o + RloadSCb .(2.112)
1 1 R + 2R/,,d
+ (MaD +M Drad ) + +a Drad load
sC c sCD 1 + Rload SCeb









From this equation, it can be seen that the total input impedance is simply the impedance

of the Helmholtz resonator neck in series with a parallel combination of the Helmholtz

resonator cavity impedance and the piezoelectric composite diaphragm impedance with a

resistive load attached.

Many energy harvesting devices with varying geometries were designed for this

dissertation, but for illustrative purposes, I will only explore the behavior of one, which

serves as a typical example. More explicitly, eight different devices were designed and

created to give a range in performance, but to illustrate the qualitative and typical

behavior of the devices, only a single "representative" device was plotted for the

remainder of this chapter. A plot of the acoustical input impedance versus frequency is

shown in Figure 2-24 for the piezoelectric composite diaphragm, both with and without

the Helmholtz resonator, to elucidate the behavior of the individual components. For this

plot and those that follow, the Helmholtz resonator has a neck length of L = 3.18 mm,

and a radius of R = 2.36 mm along with a cavity volume of V = 1950 mm3, in addition

to a diaphragm with a thickness of t, = 3 /m and a outer and inner radius of

R, =1.95 mm and R, = 1.85 mm, respectively. These dimensions were chosen as typical

values in the range of what was expected of the final device and package design. From

this plot, it can be seen that, by itself, the piezoelectric composite diaphragm has a single

resonance near 3.6 kHz, where the impedance reaches a local minimum. When combined

with the Helmholtz resonator, two minima are seen. The lower resonance that occurs

near 1.8 kHz is dominated by the Helmholtz resonator that has an uncoupled resonance of

2 kHz, as evidenced by the peak in Figure 2-5. The upper resonance at 3.9 kHz is






70


dominated by the piezoelectric composite diaphragm. Additionally, an antiresonance,

where the impedance reaches a local maxima, occurs between the two resonances.



S0 -- with Helmholtz resonator
10 ---- without Helmholtz resonator





7
10











1000 2000 3000 4000 5000 6000 7000
Frequency [Hz]


Figure 2-24: Magnitude of the acoustical input impedance for the piezoelectric composite
diaphragm and in combination with the Helmholtz resonator. (L = 3.18 mm,
R 2.36 mm, V = 1950 mm3, t = 3 im, R, =1.95 mm, R, =1.85 mm)

The acoustical input power can be obtained from the input acoustic pressure, P,

and is given by


P2
Re{ rI,- Re{Z}
Re1 ; e{Z


(2.113)


The input acoustic pressure, P, is measured inside the plane-wave tube near the end-face

of the tube. The microphone that measures this pressure is placed as close as possible


(- 1/16 ) to the end-face so that it serves as a measure of the pressure that is incident on

the energy harvesting device. In the case of the device that includes a Helmholtz









resonator, the incident pressure is the pressure incident upon the resonator neck, while for

the diaphragm only device, it is the pressure incident upon the face of the diaphragm.

A plot of the magnitude and phase of the input power is shown in Figure 2-25.

This plot assumes an ideal acoustic source, i.e. that the applied acoustic pressure is

constant and independent of the input impedance. In reality, this is not the case, and the

applied acoustic pressure will vary as a function of frequency due to a frequency

dependent impedance as seen by the speaker. The important point to take away from this

plot, however, is that there are frequencies at which the input power will be maximized.

These frequencies occur when the input impedance is minimized thus allowing for

maximum power flow from an ideal source. These do not necessarily correspond to

frequencies where the output power is maximized, as will be shown in the next section.

Electrical Output Behavior

The electrical output impedance can be found in a similar manner and is given by


Zo, = Rload// // (MD +M rad + +RDd + I//(R, +MI ) .(2.114)
sCeb SCOD SC,

Expanding this equation out yields the full expression for the output impedance, given by

F 1
S(R, +Ma)
1 1 1 sC ac
(RiCad (MaD + MaD rad) + +RaDrad 1 +RC +
Seb 02 Sca -I + R, +M a
7 sC-c .(2.115)
lout ( 1


sCb sC )s 1
SCe 0 2 M + MaDrad e) + S + rd
SC'







72


10-5
S- with Helmholtz resonator
---- without Helmholtz resonator

10--- -----------




10 10 I I I I I I
0 1000 2000 3000 4000 5000 6000 7000
Frequency [Hz]



50 -



C3
S0 -

a -50


0 1000 2000 3000 4000 5000 6000 7000
Frequency [Hz]


Figure 2-25: Magnitude and phase of the acoustical input power for an input acoustic
pressure of P = 94 dB. (L = 3.18 mm, R = 2.36 mm, V = 1950 mm3,
t, = 3 jm, R2 =1.95 mm, R, =1.85 mm)

A graphical plot of the electrical output impedance is shown in Figure 2-26 in

terms of real and imaginary components. The solid curve in the figure corresponds to the

expression given by Eq. (2.115), while the dotted curve represents the impedance of the

piezoelectric composite diaphragm by itself. Notice that only the real component shows

any significant difference between the two cases. Physically, this is due to the "poor"

coupling that is typical of indirect transducers. The result is that acoustical components

have minimal effect on the electrical impedance.






73




with Helnhollz resonalor
S1011 .....- i withool Helmholtz esPnator


10' tn


1000 2000 3000 4000 5000 000 7000
Frequency 1Hz]

0 -DO

o /
-10000 /
/
-15000

1000 201 3000 4000 5000 GOfl 7000
Frequency 1Hz]

Figure 2-26: Electrical output impedance for the piezoelectric composite diaphragm by
itself and in combination with the Helmholtz resonator. (L = 3.18 mm,
R = 2.36 mm, v = 1950 mm3, t = 3 um, R, =1.95 mm, R, =1.85 mm)

The output voltage can easily be found from the equivalent circuit to be


Rload
(P0) 1
+ Rload
V = Cb .(2.116)

RaN + sMN 1 R
RaN + sa +1 (MaD + MaDrad ) + I + RaDrad + oa + + sMaA
SSCaD) 1 + Rload eb
S\aC

The output voltage is displayed graphically in Figure 2-27 for the piezoelectric

composite diaphragm by itself and in combination with the Helmholtz resonator. For

both cases, the magnitude has a similar shape to the output impedance, although a few

differences remain. In the output impedance curve, the second resonant peak is higher

than the first, whereas for the output voltage curve shown below, the first resonant peak







74


is higher than the second. The voltage curve shown was computed for a load resistance,

Road,, equal to the electrical output resistance at the diaphragm resonance frequency, and

an input acoustic pressure of 1 Pa.

10"2 -- -- -.
A |- with HeliiholIr resonalor
---- without Helmholtz resonator


10 i~ i I-i


1 '
1--------------------------


0 1000 2000 30 400D 5000 600 71OD
Frequency [Hz]
"- .----------- -----------




a

-2100 I
0 1000 2000 A000 6000 GDOOi 700o
Frequency [Hz]

Figure 2-27: Magnitude and phase of the output voltage for the piezoelectric composite
diaphragm and in combination with the Helmholtz resonator. (L = 3.18 mm,
R = 2.36 mm, V = 1950 mm3, t = 3 im, R, =1.95 mm, R, =1.85 mm,
P = 94 dB1)

The electrical power delivered to the resistive load, RiQod, can be found from


V2
Re{H}=- (2.117)
Rload

A plot of the electrical output power delivered to the resistive load is shown in Figure

2-28 for the piezoelectric composite diaphragm by itself and in combination with the

Helmholtz resonator, for the case of an incident acoustic pressure of 1 Pa.


1 dB re 20uPa throughout this document.













I I .- without Helmholiz resonator





0 1000 200D 300 400A 00D 6000 7000
Frequency [Hz]
200

100 _- .. ... -------


0 -

,100 L---L-- LL-----------------
1I I
0 1000 Io 3000 4Dfo0 000 Go00 "DO
Frequency [Hz]


Figure 2-28: Magnitude and phase of the electric output power delivered to the load
resistor. (L = 3.18 mm, R = 2.36 mm, V = 1950 mm3, ts = 3 pm,
R, =1.95 mm, R, = 1.85 mm, P= 94 dB)

The overall power conversion efficiency, F, can then be found as the ratio of

output electrical power to input acoustical power, given by


Re{f} =Re{Ho} (2.118)
Re{n

The magnitude and phase of the efficiency is shown in Figure 2-29 for the

piezoelectric diaphragm by itself and in combination with the Helmholtz resonator. As

seen in the plot, the efficiency of the composite diaphragm reaches a peak at the

diaphragm resonance, while the efficiency of the coupled system reaches a peak at the

second resonance, which is dominated by the diaphragm resonance. It is important to

note that the peak magnitude is similar both with and without the Helmholtz resonator,

suggesting that little benefit is gained through the use of the Helmholtz resonator, as this







76


Helmholtz resonator was not optimally designed for impedance matching. This need not

always be the case, as the relative benefits of impedance matching are somewhat offset in

varying amounts by the increased resistive losses due to the Helmholtz resonator. The

benefits can thus be enhanced by maintaining the improved impedance matching while

minimizing the additional losses. A perfect impedance match between the plane wave

tube and the Helmholtz resonator would improve the efficiency by approximately 40%

over the values shown in the figure. This value is estimated based on comparison to a

Helmholtz resonator that is perfectly impedance matched to the plane-wave tube.




L 34
C10 ------------ ------------,--------



S10' --


0 10iD 200 3000 4i000 f6L EOJD 7000
Frequency [Hz]


-- with Helmholtz resonalor
without Helrnholtz rEsonator
-- ---------------- ----


S-100 -
SC
0 1000 000 3000 4000 600f OOD 7000
Frequency [Hz]

Figure 2-29: Magnitude and phase of the energy harvester efficiency. (L = 3.18 mm,
R =2.36 mm, V =1950 mm3, t =3 /um, R, =1.95 mm, R, =1.85 mm,
P = 94 dB)

Operation at an Optimal Frequency

To get a better feel for these devices in their intended application as harvesters of

energy, it helps to look at the theoretical behavior for a single frequency. For the

purposes of illustration, a device with the geometry of Device 8, as given by Table 2-5 on









page 81 was used and the frequency at which the efficiency is maximized was chosen.

This optimal frequency is different between the diaphragm mounted on the Helmholtz

resonator and the diaphragm strictly by itself. For the diaphragm mounted on the

Helmholtz resonator, a sinusoidal signal at 1966 Hz was applied, while for the diaphragm

mounted by itself, a sinusoidal signal at 13.11 kHz was used. For both cases, the signal

was swept over a range of acoustic pressures while the open circuit output voltage was

determined. A plot of the open circuit voltage versus acoustic pressure for both cases is

shown in Figure 2-30.


10
10 Membrane
Membrane/HR



-1 "
10



> 1-2

10-





115 120 125 130 135 140 145 150 155 160
Input Acoustic Pressure [dB]
Figure 2-30: Open-circuit voltage as a function of the applied acoustic pressure for the
piezoelectric diaphragm by itself and packaged with a Helmholtz resonator.
(Device = 8 from Table 2-5, f =1966 Hz, f, =13.11 kHz)

Note the steadily increasing output voltage as the acoustic pressure increases, as

would be expected for a linear system such as this. This model, however, does not take

into account any nonlinearities that may occur at high acoustic levels. Such









nonlinearities may include, but are not limited to, piezoelectric saturation effects on the

output voltage and large deflection effects on the diaphragm compliance, in addition to

acoustic nonlinearities introduced by the Helmholtz resonator orifice at high amplitudes.

These nonlinearities will act to limit the maximum voltage in a real world application.

Operating at these same optimal frequencies, a load resistor was placed across the

output terminals and the power was determined through that resistor. Repeating this over

a range of load resistances yields the plot shown in Figure 2-31. Note that while the input

power remains constant, the output power has a peak at a particular load resistance. This

corresponds to the optimal resistance as given by Eq. (2.12). The input and output power

using an optimal load are also shown on this graph and are indicated by an 'x' for both

the diaphragm with the Helmholtz resonator and without.



-4
10


-in Membrane

-6 H. Membrane/HR
10 in
out -Membrane
S- out -Membrane/HR

10-8
10-




-10

2 3 4
10 10 10
Load Resistance [Ohms]

Figure 2-31: Input and output power as a function of the load resistance. Shown for the
piezoelectric diaphragm by itself and packaged with a Helmholtz resonator.
(Device = 8 from Table 2-5, f =1966 Hz, fm, =13.11 kHz, and P= 114 dB)









Using the optimal values for the load resistance in each case, the input power and

output power was determined as a function of the applied acoustic input pressure, and is

shown in Figure 2-32. Once again, it is important to note that nonlinear effects are not

included in the model used to calculate the output power.


2 Hn. -Membrane
10 in
Membrane/HR
0 H out- Membrane
0in
10 nout -Membrane/HR
out

-2 ^ ^ ^
10


( 104
0 -
10
-8-
10-6


10-

10
115 120 125 130 135 140 145 150 155 160
Acoustic Input Pressure [dB]

Figure 2-32: Input and output power as a function of the applied acoustic input pressure,
while using an optimal load resistor. (Device = 8 from Table 2-5, f =1966
Hz, f,, =13.11 kHz, Road,Mem = 486 Q, Rioad,MemHR = 3242 Q)

Device Configurations

The devices were designed based on the theoretical plots, along with estimates of

the output power based on the lumped element model. Eight configurations were chosen

based on their resonant frequency and maximum power before the onset of nonlinear

behavior. The criteria for choosing the designs were to keep the resonant frequencies









inside of the testable range, below 6.4 kHz, while maximizing the power. Four outer

radii, R2, were chosen, consisting of 0.9, 1.2, 1.5, and 1.8 mm, while the inner radii were

all chosen to be 0.95 times their outer counterpart. The second four configurations had

the same dimensions as the first four, with the addition of a central pillar of silicon

designed to act as a rigid mass. It should be noted that the addition of the central mass

invalidates the structural model that was used, however, the model is still expected to

provide some guidance over the behavior of these particular devices as well. The central

mass was added in order to reduce the resonant frequency of the diaphragms. It was

fixed at 0.2 times the outer radius of each diaphragm. This value was chosen such that

the central mass was located primarily over a region of the diaphragm that experiences

little stress. By restricting the mass to this region, only a minimal increase in the

diaphragm stiffness was expected, while the mass of the diaphragm could be greatly

increased, thereby lowering the resonant frequency. The addition of this central pillar

was predicted to lower the resonant frequency of each device by an average of 57.76 %.

Additionally, the diaphragm thickness was chosen to be 3 um. Only one thickness for the

diaphragm was chosen as all the devices were going to be made on the same wafer.

The chosen device configurations are summarized in Table 2-5. Overall, the

geometries were chosen as to create a proof of concept set of devices. Manual parameter

adjustment was performed in order to improve performance, however, strong constraints

were placed on the geometry in order to ensure fabrication ability and improve the

chances of first-run success. Also listed in the table are the predicted resonant

frequencies and maximum power before the onset of nonlinearities. Note that this does

not define an upper limit for the output power but merely provides a figure of merit upon










which to compare devices. Also, for the purposes of predicting the behavior of these

specific devices and comparing them to experimental data, the lumped element

parameters were computed using the equations from Chapter 2 and the Matlab code given

in Appendix A. The calculated parameters are given in Table 2-6.

Table 2-5: Chosen device configurations for fabrication.
Device t, [pm] t [pUm] R2 [um] Rj [Pm] R,,m [Um] fr [kHz] Powo,,[nW]
1 3 0.5 900 830 0 65.68 15.56
2 3 0.5 900 830 180 28.08 36.12
3 3 0.5 1200 1115 0 48.14 1.73
4 3 0.5 1200 1115 240 20.37 3.99
5 3 0.5 1500 1400 0 38.01 0.302
6 3 0.5 1500 1400 300 15.98 0.667
7 3 0.5 1800 1685 0 31.41 0.070
8 3 0.5 1800 1685 360 13.15 0.162


Table 2-6: Lumped element parameters used for theoretical models.
k d, CaD aD MaDma. Cb
Pa V kg kg
Device V] Pal [] [ Pa] nm] m4 [nF]
1 -1.81 -5.80E-07 0.001 1.01E-14 5.58E-15 5663.3 0 17.4
2 -1.81 -5.80E-07 0.001 1.01E-14 5.58E-15 5663.3 19649 17.4
3 -9.29 -7.16E-06 0.008 1.93E-13 2.07E-14 2602.5 0 26.9
4 -9.29 -7.16E-06 0.008 1.93E-13 2.07E-14 2602.5 11052 26.9
5 -12.93 -1.94E-05 0.016 7.43E-13 5.74E-14 1521.8 0 38.2
6 -12.93 -1.94E-05 0.016 7.43E-13 5.74E-14 1521.8 7073.6 38.2
7 -16.92 -4.71E-05 0.028 2.42E-12 1.43E-13 1009.8 0 51.3
8 -16.92 -4.71E-05 0.028 2.42E-12 1.43E-13 1009.8 4912.2 51.3
















CHAPTER 3
FABRICATION AND PACKAGING


Overview

This chapter focuses on the device fabrication and process flow, as well as the

device packaging design and implementation. First, a detailed discussion of the process

flow is presented, including fabrication methods, equipment, and recipes. This is

followed by an overview of the packaging schemes that were employed. The packages

were designed to flush mount the device in order to expose the diaphragm directly to the

acoustic input, thereby enabling direct measurement of the diaphragm parameters. Two

variations in package design were employed, namely a quarter-wave resonator package

and a sealed cavity package. A graphical overview of this chapter is provided by the

roadmap of Figure 3-1.

Process Flow

The devices were batch fabricated on 4" silicon-on-insulator (SOI) wafers. The

process sequence is given in detail in Appendix B. All of the processing steps up

through the deposition of the top electrode were performed at Sandia National

Laboratories. The remainder of the steps were formed at the University of Florida, with

the exception of the PZT etch step which was performed at the Army Research

Laboratory.























Packaging
Scheme

I Sealed Cavity
jj1i ^- n^J Package
Chip to Package Package
h Mounting



S Polin I


Figure 3-1: Roadmap for Chapter 3.

A diaphragm thickness of 3 um was desired in order to achieve a testable resonant

frequency and therefore required a top silicon thickness in that range, as the top silicon

layer of the SOI wafer ultimately forms the bulk of the diaphragm. Unfortunately, the

only available SOI wafers contained a top silicon layer of 12 um thickness. In order to

reduce this thickness, a timed KOH (Potassium Hydroxide) etch [137] was performed,

however a residue was left behind on most of the wafers, that was difficult to remove.

The composition of the residue was not identifiable, however a 1 minute dip in 10%

Nitric acid, followed by a 2:1 Piranha etch for 5 minutes had no noticeable effect.











........................................................




...................................................
. . . . . . . . . . . . .





.. .. .. .. .. .. .. .. .. .. .. .. .. .. .


^^ ^^^-- m ^ ^ --- -a- I^^ ^
d ) ..... .... ...... ... .. .. .. h )


TiO2 I1 PZT
] Si 77 Bottom Electrode Pt
=I- Buried Oxide (BOX) SiO2 E Photoresist
f7M Top Electrode Pt
Figure 3-2: Condensed process sequence in cross-section. a) Deposit 100 nm of Ti on
SOI Wafer and oxidize to form TiO2. b) Deposit Ti/Pt (40/180 nm) and
liftoff with "ElectrodeBot" mask. c) Spin coat PZT 52/48 solution & pyrolize
(6 layers for 400 nm total). d) Deposit Pt (180 nm) and liftoff with
"ElectrodeTop" mask. e) Wet Etch PZT in 3:1:1 of (NH4)HF2/HC1/DI water.
f) Spin & pattern thick photoresist on bottom with "Cavity" mask. g) DRIE
(deep reactive ion etch) to buried oxide layer. h) Ash resist and BOE
(buffered oxide etch) backside to remove to buried oxide layer.

Two of the original KOH-etched wafers were replaced with two new wafers that

were etched in an HNA bath to avoid the residue issue [138]. The advantage of the

HNA etch was a smooth, mirror finished surface to the wafer with no visible residue.

The disadvantage was a less consistent etch rate than the KOH procedure. For

comparison the HNA etch had an average etch rate of 2.6 /ummin with a standard


deviation of 0.33 /ummin, while the KOH etch had an etch rate of 1.69 /ummin with a


standard deviation of 0.03 m//min.


.. . .. . . . . . . .









Following the wafer thinning, the process flow proceeded as given by Figure 3-2.

The first step was to deposit a 100 nm thick layer of Ti, and oxidize in a tube furnace to

create a TiO2 layer that serves as a diffusion barrier for the PZT [139-142] as shown in

Figure 3-2a.

A liftoff process was then used to pattern the bottom electrodes with 40 nm of Ti

followed by 180 nm of Pt as shown in Figure 3-2b. The Ti layer served as an adhesion

layer in this step. The next step was the deposition of a previously mixed 52/48 sol-gel

solution of PZT [139, 143, 144] as shown in Figure 3-2c. The solution was spin-cast at

2500 RPM for 30 sec, then pyrolized at 350 C in air for 2 min on a hot plate. Repeating

the spin and pyrolize step 6 times yielded a total PZT thickness of approximately 400 nm.

The wafers were then furnace annealed at 650 C for 30 min to achieve a perovskite phase

of PZT that has the most attractive piezoelectric properties.

Wafer bow measurements were performed after deposition and oxidation of the

TiO2 as well as after annealing of the PZT. By recording the wafer curvature in terms of

radius or bow, the stress in the layers can be determined [145-147]. Table 3-1 shows the

results and calculations from these measurements for all of the fabricated wafers. Stress

levels were first calculated for the combination of TiO2 and backside SiO2 that was

formed during the titanium oxidation step. Then the stress contribution from the TiO2

was calculated. Finally, using the wafer curvature measurements after deposition and

annealing of the PZT, the stress in the PZT layer was calculated.

In order the calculate the stress from the wafer curvature, a relationship known as

Stoney's equation [145-147] was used and is given by









E h3
s- ,(3.1)
6(1- v)Rh (1 hh


where 1/R is the curvature, h,,h, is the substrate and film thickness respectively, E. is

Young's modulus of the substrate and v, is Poisson's ratio of the substrate.

Table 3-1: Wafer bow measurements and the resulting calculated stress.
Initial After TiO2 Dep. + Ox. TiO2 + TiO2 After PZT PZT
SiO2
W r Radius Bow Radius Bow Stress Stress Stress Radius Bow Stress
Wafer
# [m] [pu] [m] [pu] [MPa] [MPa] [MPa] [m] [pu] [MPa]
1 -43.8 18.51 -137.8 6.15 1367.1 844.7 1217.0 -149.9 5.65 12.8
1A -40.1 20.5 -103.2 8.4 1335.3 1239.7 1809.6 -1144 1.52 176.3
2 -38.9 20.45 -99.0 8.03 1371.1 848.8 1223.1 -1338 0.72 187.3
3 -60.8 13.05 -291.4 3.12 1140.6 678.6 967.9 n/a n/a n/a
4 -44.2 18.24 -121.2 6.93 1259.8 772.9 1109.4 279.3 -2.82 249.9
5A -63.7 12.84 -717.3 1.48 1254.5 776.3 1114.5 156.3 -4.74 159.4
6 -43.0 18.86 -109.3 7.6 1235.4 769.5 1104.2 809.9 -0.77 214.5
7 -44.3 18.3 -106.1 7.84 1154.2 714.8 1022.2 662.9 -1.27 233.5


Before the above equation can be used to find the stress in the TiO2 layer, the

thickness of the SiO2 layer that was formed during the oxidation step must be determined.

For silicon oxidation, the relationship between oxide thickness and oxidation time is

approximated by [148]

t + At = B(t+r), (3.2)

where A and B are coefficients which depend on material properties and operating

conditions, to is the total oxide thickness, t is the oxidation time, and r is a time shift

due to an initial oxide thickness. For short oxidation times, where t < A2/4B, this

equation reduces to


t (t+r). (3.3)
A